Datasets:

sergioq2

/

raft-inmunotherapy-cancer

like 0

483(7390):474- 8, 2012. PMCID: PMC3478770. 74. Noushmehr H, Weisenberger DJ, Diefes K, Phillips HS, Pujara K, Berman BP, Pan F, Pelloski CE, Sulman EP, Bhat KP, Verhaak RG, Hoadley KA, Hayes DN, Perou CM, Schmidt HK, Ding L, Wilson RK, Van Den Berg D, Shen H, Bengtsson H, Neuvial P, Cope LM, Buckley J, Herman JG, Baylin SB, Laird PW, Aldape K, Cancer Genome Atlas Research N. Identification of a CpG island methylator phenotype that defines a distinct subgroup of glioma. Cancer Cell. 17(5):510- 22, 2010. PMCID: PMC2872684. 75. Rohle D, Popovici -Muller J, Palaskas N, Turcan S, Grommes C, Campos C, Tsoi J , Clark O, Oldrini B, Komisopoulou E, Kunii K, Pedraza A, Schalm S, Silverman L, Miller A, Wang F, Yang H, Chen Y, Kernytsky A, Rosenblum MK, Liu W, Biller SA, Su SM, Brennan CW, Chan TA, Graeber TG, Yen KE, Mellinghoff IK. An inhibitor of mutant IDH1 delays growth and promotes differentiation of glioma cells. Science.

340(6132):626- 30, 2013. 76. Losman JA, Looper RE, Koivunen P, Lee S, Schneider RK, McMahon C, Cowley GS, Root DE, Ebert BL, Kaelin WG, Jr. (R) -2-hydroxyglutarate is sufficient to promote leukemogenesis and its effects are reversible. Science.

339(6127):1621- 5, 2013. 77. Duncan CG, Barwick BG, Jin G, Rago C, Kapoor -Vazirani P, Powell DR, Chi JT, Bigner DD, Vertino PM, Yan H. A heterozygous IDH1R132H/WT mutation induces genome- wide alterations i n DNA methylation. Genome Res.

22(12):2339- 55, 2012. PMCID: PMC3514664. 78. Dang L, White DW, Gross S, Bennett BD, Bittinger MA, Driggers EM, Fantin VR, Jang HG, Jin S, Keenan MC, Marks KM, Prins RM, Ward PS, Yen KE, Liau LM, Rabinowitz JD, Cantley LC, Tho mpson CB, Vander Heiden MG, Su SM. Cancer -associated IDH1 mutations produce 2- hydroxyglutarate. Nature.

462(7274):739- 44, 2009. PMCID: PMC2818760. 79. Jin G, Pirozzi CJ, Chen LH, Lopez GY, Duncan CG, Feng J, Spasojevic I, Bigner DD, He Y, Yan H. Mutant IDH 1 is required for IDH1 mutated tumor cell growth. Oncotarget.

3(8):774- 82, 2012. PMCID: PMC3478455. 80. Andronesi OC, Kim GS, Gerstner E, Batchelor T, Tzika AA, Fantin VR, Vander Heiden MG, Sorensen AG. Detection of 2- hydroxyglutarate in IDH -mutated glioma patients by in vivo spectral -editing and 2D correlation magnetic resonance spectroscopy. Sci Transl Med.

4(116):116ra4, 2012. 81. Wen PY, Macdonald DR, Reardon DA, Cloughesy TF, Sorensen AG, Galanis E, Degroot J, Wick W, Gilbert MR, Lassman AB, Tsien C, M ikkelsen T, Wong ET, Chamberlain MC, Stupp R, Lamborn KR, Vogelbaum MA, van den Bent MJ, Chang SM. Updated response assessment criteria for high- grade gliomas: response assessment in neuro- oncology working group. J Clin Oncol. 28(11):1963 -72, 2010. 82. Boisselier B, Gallego Perez -Larraya J, Rossetto M, Labussiere M, Ciccarino P, Marie Y, Delattre JY, Sanson M. Detection of IDH1 mutation in the plasma of patients with glioma. Neurology.

79(16):1693- 8, 2012. Pro00054746: RESIST Study Version: 20171011 PT-PhI-II v05.22.12 Duke Cancer Institute Page 54 CONFIDENTIAL 83. Yung WK, Albright RE, Olson J, Fredericks R, Fi nk K, Prados MD, Brada M, Spence A, Hohl RJ, Shapiro W, Glantz M, Greenberg H, Selker RG, Vick NA, Rampling R, Friedman H, Phillips P, Bruner J, Yue N, Osoba D, Zaknoen S, Levin VA. A phase II study of temozolomide vs. procarbazine in patients with glioblastoma multiforme at first relapse. Br J Cancer. 83(5):588- 93, 2000. PMCID: PMC2363506. Pro00054746: RESIST Study Version: 20171011 PT-PhI-II v05.22.12 Duke Cancer Institute Page 55 CONFIDENTIAL 18 APPENDICES 18.1 Protocol Synopsis and Summary Please see separate documents (available upon request). Pro00054746: RESIST Study Version: 20171011 PT-PhI-II v05.22.12 Duke Cancer Institute Page 77 CONFIDENTIAL 18.2 Special or Representative SOPs and FORMs Please see separate document s (available upon request). Pro00054746: RESIST Study Version: 20171011 PT-PhI-II v05.22.12 Duke Cancer Institute Page 78 CONFIDENTIAL 18.3 DSMBplus Monitoring Plan Please see separate document (available upon request). Pro00054746: RESIST Study Version: 20171011 PT-PhI-II v05.22.12 Duke Cancer Institute Page 79 CONFIDENTIAL 18.4 Standard Radiation Therapy Radiotherapy typically begins within ≤ 5 weeks of surgery. One treatment of 1.8- 2.0 Gy/fraction should be given daily 5 days per week for a total of 59.4- 60.0 Gy over <7 weeks. 3D conformal and intensity - modulated RT is permitted. All portals should be treated during each treatment session. Doses are specified as the target dose that shall be to the center of the target volume. The gross target volume (GTV) for both the initial volume (GTV1) and the conedown volume (GTV2) should be based on the postoperat ive CT/MRI (and preferably the MRI; the preoperative scans may be used if postoperative scans are not available). This initial target volume (GTV1) should include the contrast - enhancing lesion (and should include the surgical resection cavity) and surrounding edema (if it exists) demonstrated on CT/MRI plus a 2.0- cm margin (this 2.0- cm margin- extended volume will be considered the initial planning target volume, or PTV1). The initial target volume should be treated to 46 Gy at 2Gy/fraction or 45- 50.4 Gy at 1.8Gy/fraction. If no surrounding edema is present, the initial planning target volume (PTV1) should include the contrast -enhancing lesion (and should include the surgical resection cavity) plus a 2.5- cm margin. Please note that clinical judgment may be us ed to modify PTV1 to exclude sensitive structures such as the optic chiasm, non- cranial contents, or anatomic regions in the brain where natural barriers would likely preclude microscopic tumor extension, such as the cerebellum, the contralateral hemispher e, directly across from the tentorium cerebri, the ventricles, etc. After 46 Gy, the tumor volume (GTV2) for the conedown treatment should include the contrast -enhancing lesion (without edema) on the pre- surgery CT/MRI scan plus a 1.5- 2-cm margin (PTV2). T reat to 14 Gy at 2Gy/fraction or 14.4- 9.0 Gy at 1.8Gy/fraction to a total of 60.0 or 59.4Gy, respectively. Dose is prescribed to the isodose line such that at least 95% of the target volume receives he prescribed dose. The optic apparatus should be limit ed to a maximum of 54Gy and no more than 5% of the volume of the brainstem should receive >54Gy. Radiation should be delayed or interrupted if the platelet count is < 20,000. Radiation should not begin or resume until the platelet count is ≥ 20,000. Hemat ologic toxicities should be rated on a scale of 0- 5 as defined in the NCI Common Terminology Criteria for Adverse Events (CTCAE) version 3.0. If radiotherapy has to be temporarily interrupted for technical or medical reasons unrelated to the temozolomide administration, then treatment with daily temozolomide should continue. If radiotherapy has to be permanently interrupted then treatment with daily temozolomide should stop. The following should be recorded at entry into this study: daily treatment record, all isodose distributions (in color), dose volume histograms including the cumulative dose to the target volumes, optic chiasm, optic nerves and brain stem, and the radiotherapy summary. Pro00054746: RESIST Study Version: 20171011 PT-PhI-II v05.22.12 Duke Cancer Institute Page 80 CONFIDENTIAL 18.5 Temozolomide Therapy 18.5.1 Concurrent TMZ and RT For subjects who progr ess to Grade 3 at the time of surgery, TMZ should be administered concomitant with standard external beam RT under the direction of the study Neuro- Oncologists (listed on title page), or their designees, respectively, at Duke University or another institut ion at the standard dose per discretion of the treating neuro- oncologist. During standard RT, s ubjects will receive 75mg/m2/day TMZ for 42 days (6 weeks). Per the new NCI Temozolomide (NSC 362856) Action Letter: o Liver function tests (or CMP) should be performed:  prior to treatment initiation. If abnormal, the decision to initiate temozolomide treatment should carefully consider the benefits and risks for the individual patient;  after each treatment cycle. o For patients on a 42 day treatment cycle, liver function tests should be repeated midway during this cycle; o For patients with significant liver function abnormalities the benefits and risks of continuing treatment should be carefully considered. If any interruption occurs or dose reduction is required, TMZ may be resumed when re- treatment criteria are met as outlined in Table 5 and Table 6. Dose reduction or alteration during this period due to toxicity is standardly performed as outlined in Table 5. Table 5: Temozolomide Dosing Interruption or Discontinuation during Concomitant Radiotherapy Toxicity TMZ Interruptiona TMZ Discontinuation Absolute Neutrophil Count >0.5 and <1.0 x 109/L <0.5 x 109/L Platelet Count >10 and <100 x 109/L <10 x 109/L Common toxicity criteria (CTC2) Non - hematological Toxicity (except for alopecia, nausea, vomiting) CTC Grade 2 CTC Grade 3 or 4 a: Treatment with concomitant TMZ can be continued when all of the following conditions are met: absolute neutrophil count >1.0 x 109/L; platelet count > 100 x 109/L; CTC non- hematological toxicity <Grade 1 (except for alopecia, nausea, vomiting). The dose may be modified as outlined in Table 6 at the discretion of the treating oncologist. Table 6. Dose Modifications of Temozolomide during Radiation Therapy Dose at Toxicity Modified Dose 75 mg/m2 (during XRT) 60 mg/m2 60 mg/m2 (during XRT) 50 mg/m2 18.5.2 Guidelines for 21-day TMZ Therapy Subjects with stable histologic grade at time of surgery (Grade 2) will be treated with adjuvant TMZ at a targeted dose of 50- 100mg/m2/day for 21 days every 28 days for up to 12 cycles . Subjects that progress to Grade 3 at the time of surgery will recei ve SOC RT with TMZ x 6 weeks following surgery. Once RT is complete, subjects that progressed to Grade 3 at the time of surgery will have a 3 ± 1 week break before 2 CTC, common toxicity criteria (NCI) Pro00054746: RESIST Study Version: 20171011 PT-PhI-II v05.22.12 Duke Cancer Institute Page 81 CONFIDENTIAL resum ing TMZ ( Table 7). Then, subjects should receive TMZ at a target ed dose of 50-100 mg/m2/day for 21 days every 28 days for up to 12 cycles . A CBC with auto differential (Lab Code 2010300) should be obtained at the end of each cycle for all subjects receiving 21-day TMZ . Table 7. Criteria to Resume Temozolomide after Radiation Absolute Neutrophil Count ≥ 1.0 x 109/L Platelet Count ≥ 100 x 109/L Resolution of CTC Non -hematologic toxicities to Grade 2 or less. Because of the risk of opportunistic infections in patients receiving TMZ, patients may receive antibiotic prophylaxis at the discretion of the primary Neuro- Oncologist, consisting of inhaled pentamidine or oral levofloxacin. Antiemetic prophylaxis with m etoclopramide or a 5- hydroxytryptamine3 antagonist will also be recommended before the initial doses of concomitant TMZ and may be used during the adjuvant 21- day course of TMZ. All subjects will start 21- day TMZ cycle 1 at a dose of 50mg/m2/day, and the dose may be increased up to 100mg/m2/day in subsequent cycles at the discretion of the treating physician. During 21-day TMZ treatment (28 day cycles ), dose adjustments to TMZ, if needed, may be conducted as outlined below: Table 8. Temozolomide Dose Delay, Reduction , or Discontinuation During 21-day Treatment Toxicity Delay TMZ Dose a Reduce TMZ by 25% (down to minimum of 50 mg/m2) b Discontinue TMZ Absolute Neutrophil Count >0.5 and <1.0 x 109/L >0.5 and <1.0 x 109/L <0.5 x 109/L Platelet Count >10 and <100 x 109/L >10 and <100 x 109/L <10 x 109/L CTC Non -hematological Toxicity (except for alopecia, nausea, vomiting) CTC Grade 3 CTC Grade 3 CTC Grade 4 c a: If dose is delayed, treatment with TMZ can resume when the following conditions are met: absolute neutrophil count >1.0 x 109/L; platelet count >100 x 109/L; CTC non- hematological toxicity resolved to baseline (except for alopecia, nausea, vomiting). b: For subjects whose dose was escalated above 50 mg/m2 at the discretion of treating physician c: TMZ is to be discontinued if dose reduction to < 50 mg/m2 is required or if the same Grade 3 non- hematological toxicity (except for alopecia, nausea, vomiting) recurs after dose reduction or delay .

1504 | CANCER DISCOVERY JUL Y 2023 AACRJournals.orgNEWS IN BRIEF Personalized mRNA Vaccine Immunogenic against PDAC A personalized mRNA vaccine designed to target an individual’s tumor-specific neoantigens is showing early promise against pancreatic ductal adenocar - cinoma (PDAC), an aggressive cancer characterized by a low mutational burden and with few good immunotherapy op- tions (Nature 2023;618:144–50). “It’s a very compelling study, and it really sets the stage for future mRNA vaccine immunotherapy in pancreas cancer,” says Neeha Zaidi, MD, of The Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins in Baltimore, MD, who was not involved in the research but wrote an accompanying commentary (Nature 2023;618:37–8). In a phase I trial, researchers at Memorial Sloan Kettering Cancer Center (MSKCC) in New York, NY, found that sequential administration of the anti–PD-L1 drug atezolizumab (Tecentriq; Genentech) followed by eight weekly infusions of autogene cevumeran (BNT122; BioNTech)—an mRNA vaccine formulated in lipoplex nanoparticles that contains up to 20 patient-specific neoantigens—and then a biweekly chemotherapy regimen extended recurrence-free survival in patients with PDAC whose tumors had been removed. Following vaccination, eight of 16 trial participants developed neoantigen- specific CD8+ T cells.

Among these im- munologic responders, four exhibited T cells against only one neoantigen, while the others mounted T-cell responses against three to eight vaccine-encoded mutations. Data from one study participant who developed a small liver lesion fol- lowing vaccination suggest that these T cells serve an antitumor immune surveillance function. A biopsy revealed a dense infiltrate of vaccine-induced T cells, along with DNA from rare cells harboring the same TP53 mutation as the original pancreatic tumor. Because the lesion soon disappeared, the T cells presumably helped prevent nascent micrometastases from taking root. Blood samples from the vaccine- responsive individuals showed that their neoantigen-specific T cells expanded in number, eventually representing up to 10% of their circulating T cells. That amount declined during 5 months of subsequent chemotherapy, but the cell population rebounded after patients received a booster dose of autogene cevumeran. “The magnitude of the response is high—and it’s in a tumor type that historically was thought not to have such [antigenic] targets at all,” says MSKCC’s Vinod Balachandran, MD, who led the study. “When we see a high magnitude response, it strongly correlates with clinical benefit,” he adds. After a median follow-up of 18 months, none of the patients who developed vaccine-induced T-cell responses showed signs of disease recurrence, whereas patients who did not develop an immune response experienced a relapse after a median of 13.4 months following surgery. “These data continue to provide strong evidence that tumor-specific memory CD8+ T cells correlate with long-term durable remission in pancreatic cancer,” says Stephanie Dougan, PhD, of Dana-Farber Cancer Institute in Boston, MA, who was not involved in the study. “Whether boost- ing these endogenous responses with a vaccine will convert more patients into long-term survivors is worth testing in a randomized clinical trial.” BioNTech’s cofounder and CEO Uğur Şahin, MD, agrees. The first PDAC trial “was really for immunological learning,” he says. Now, the company expects to launch a phase II adjuvant trial of autogene cevumeran, with or without a checkpoint inhibitor, for the same patient population later this year. –Elie Dolgin ■ doi: 10.1158/2159-8290.CD-NB2023-0038 Women with Breast Cancer Can Attempt Pregnancy Women with hormone-responsive breast cancer can stop endocrine ther - apy to attempt pregnancy, according to findings from the POSITIVE trial (N Engl J Med 2023;388:1645–56).

Investigators found that temporarily interrupting therapy did not raise the short-term risk of disease recurrence, and nearly 64% of women gave birth while therapy was paused. Clinical guidelines recommend adjuvant endocrine therapy for 5 to 10 years to reduce the risk of recur - rence following treatment for early- stage hormone receptor–positive breast cancer. Prior to this trial, there were little data on whether interrupting therapy to pursue pregnancy would increase the risk of relapse, which pre- sented a clinical dilemma for women in their prime childbearing years. “This is good news that is most relevant for women diagnosed with cancer before age 40 who may not have started a family or want to complete their families,” says the study’s lead author, Ann Partridge, MD, MPH, of Dana-Farber Cancer Institute in Bos- ton, MA. “We found that temporarily interrupting therapy for pregnancy did not worsen outcomes from any breast cancer events—including distant events, which are most concerning because they are not typically curable.” The trial followed 497 women age 42 or younger who had received adjuvant endocrine therapy for 18 to 30 months. In total, 74% of study participants had at least one pregnancy and 63.8% had at least one live birth. After a median of 3 years, the incidence of breast cancer was 8.9% in the treat- ment interruption group versus 9.2% in an external control group. Research- ers allowed for the possibility that the findings are biased by a “healthy mother” effect—meaning that women with a lower risk of recurrence were more likely than others to participate in the trial and become pregnant. “The results of the POSITIVE trial provide a framework with concrete recommendations and critical safety data to guide conversations with our patients about the short-term safety of interrupting endocrine therapy to pursue pregnancy,” says Jennifer Specht, MD, of Fred Hutchinson Cancer Center in Seattle, WA. “With this data, we can provide advice to pa- tients about the risks of breast cancer recurrence and important reassurance about pregnancy outcomes.”Downloaded from http://aacrjournals.org/cancerdiscovery/article-pdf/13/7/1504/3345622/1504.pdf by guest on 29 November 2024

Citation: Savsani, K.; Dakshanamurthy, S. Novel Methodology for the Design of Personalized Cancer Vaccine Targeting Neoantigens: Application to Pancreatic Ductal Adenocarcinoma. Diseases 2024 ,12, 149. https:// doi.org/10.3390/diseases12070149 Academic Editor: Maurizio Battino Received: 15 May 2024 Revised: 3 July 2024 Accepted: 8 July 2024 Published: 11 July 2024 Copyright: ©2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). diseases Article Novel Methodology for the Design of Personalized Cancer Vaccine Targeting Neoantigens: Application to Pancreatic Ductal Adenocarcinoma Kush Savsani1 and Sivanesan Dakshanamurthy2,* 1Department of Surgery, Virginia Commonwealth University, Richmond, VA 23219, USA 2Lombardi Comprehensive Cancer Center, Georgetown University School of Medicine, Washington, DC 20007, USA *Correspondence: [email protected] Abstract: Personalized cancer vaccines have emerged as a promising avenue for cancer treatment or prevention strategies. This approach targets the specific genetic alterations in individual patient’s tumors, offering a more personalized and effective treatment option. Previous studies have shown that generalized peptide vaccines targeting a limited scope of gene mutations were ineffective, em- phasizing the need for personalized approaches. While studies have explored personalized mRNA vaccines, personalized peptide vaccines have not yet been studied in this context. Pancreatic ductal adenocarcinoma (PDAC) remains challenging in oncology, necessitating innovative therapeutic strate- gies. In this study, we developed a personalized peptide vaccine design methodology, employing RNA sequencing (RNAseq) to identify prevalent gene mutations underlying PDAC development in a patient solid tumor tissue. We performed RNAseq analysis for trimming adapters, read alignment, and somatic variant calling. We also developed a Python program called SCGeneID, which validates the alignment of the RNAseq analysis. The Python program is freely available to download. Using chromosome number and locus data, SCGeneID identifies the target gene along the UCSC hg38 reference set. Based on the gene mutation data, we developed a personalized PDAC cancer vaccine that targeted 100 highly prevalent gene mutations in two patients. We predicted peptide-MHC binding affinity, immunogenicity, antigenicity, allergenicity, and toxicity for each epitope. Then, we selected the top 50 and 100 epitopes based on our previously published vaccine design methodology. Finally, we generated pMHC-TCR 3D molecular model complex structures, which are freely available to download. The designed personalized cancer vaccine contains epitopes commonly found in PDAC solid tumor tissue. Our personalized vaccine was composed of neoantigens, allowing for a more precise and targeted immune response against cancer cells. Additionally, we identified mutated genes, which were also found in the reference study, where we obtained the sequencing data, thus validating our vaccine design methodology. This is the first study designing a personalized peptide cancer vaccine targeting neoantigens using human patient data to identify gene mutations associated with the specific tumor of interest. Keywords: personalized cancer vaccines; neoantigens; pancreatic ductal adenocarcinoma; peptide based personalized cancer vaccine; MHC; HLA; TCR 1. Introduction Personalized cancer vaccines are a rising innovation in the field of vaccine design [ 1]. These vaccines induce an antigen-specific CD8+and CD4+T-cell response to enhance anti-tumor activity based on a patient’s individual tumor. Technological innovation has led to the ability to rapidly sequence and analyze patient genome data, which led to the selection of gene targets and on-demand production of personalized therapy [ 2]. A phase I clinical trial synthesized personalized mRNA vaccines against PDAC from solid tumors, which led to improved disease-free survival [ 3]. The trial analyzed a patient Diseases 2024 ,12, 149. https://doi.org/10.3390/diseases12070149 https://www.mdpi.com/journal/diseases Diseases 2024 ,12, 149 2 of 14 population who underwent surgical resection of PDAC tumors. Future development of personalized cancer vaccines directly to demonstrate significant efficacy in patients without major surgical intervention. Pancreatic ductal adenocarcinoma (PDAC) is the most common form of pancreatic cancer and is projected to be the second-leading cause of cancer mortality by 2030 [ 4,5]. Current clinical therapies involve neoadjuvant therapy followed by possible surgical resection [ 6]. However, patients with PDAC suffer from poor prognosis, with a median survival rate of 22.1 months and an actual survival rate of 17.0% [ 7]. PDAC is often diagnosed late, and as a result, surgical resection may not be a viable option for many patients [ 8]. As the cancer progresses and possible treatment options decrease, survival outcomes also significantly worsen. The five-year survival rate for patients diagnosed with late-stage PDAC is less than 10% [8]. PDAC progresses as a complex activation of driver genes and inactivation of tumor suppressor genes [ 9]. Commonly mutated genes observed in PDAC include KRAS, TP53, CDNK2A, DPC4/SMAD4, and BRCA2. Studies of key mutations in these genes are con- ducted with the goal of developing targeted gene therapies. One particular mutation, the KRAS G12D mutation, is present in over 40% of PDAC patients [ 10]. However, this specific mutation has been found not to be significantly associated with overall survival outcomes. The TP53 gene is mutated in about 50% of PDAC patients [ 11]. These mutations include gain-of-function point mutations and null mutations as a result of deletions. Mutations of the CDNK2A gene have been found to be significantly associated with poorer survival outcomes for patients with PDAC compared to mutations of KRAS and TP53 [12,13]. Several PDAC vaccines are under development, and clinical trials are being conducted using a variety of immunologic targeting methods [ 14]. These methods include cell-based, protein-based, microorganism-based, DNA-based, exosome-based, and peptide-based vac- cines. Peptide-based vaccines have been growing in popularity due to their ability to be quickly and cheaply developed and their flexibility in patient populations [ 15]. For PDAC, the first peptide vaccine to undergo clinical trials was a KRAS-targeting peptide co-administered with GM-CSF to promote a greater immune response [ 16]. The vaccine successfully induced specific immune responses in 58% of patients, contributing to a longer survival time for treated patients. Other peptide vaccines targeting survivin, gas- trin, VEGFR-1, VEGFR-2, and WT1 have been ineffective in inducing immune response or contributing to significantly improved survival [ 14,16–20]. However, the design of personalized-based peptide cancer vaccines is completely absent. This study focuses on the development of a design protocol to create personalized peptide vaccines with application to PDAC. The protocol identifies genetic variants using RNAseq analysis and designs a personalized peptide vaccine using a vaccine development protocol and omics pipeline previously developed by our group [21]. 2. Materials and Methods 2.1. Patient Genomic Data We obtained patient genomic data from the Gene Expression Omnibus (GEO) database [ 22],a publicly accessible repository of comprehensive microarray, next-generation sequencing, and other forms of high-throughput functional genomic data. For this study, we specifically collected raw Illumina sequencing data pertaining to human patient solid tumor samples. These samples were part of a detailed study focused on analyzing long-term heterogeneity in patients with pancreatic ductal adenocarcinoma (PDAC) [ 23]. This study included genomic data from a cohort of 19 patients, consisting of 10 long-term and 9 short-term survivors, providing a diverse basis for examining genetic variations linked to survival outcomes. For the objectives of this study, we selected one patient classified as a short-term survivor and one patient classified as a long-term survivor to design personalized vaccines, serving as a proof-of-concept for our approach. This selection was strategic, allowing us to explore the potential of personalized medicine in cases with poorer prognoses and to evaluate the efficacy of targeted therapies based Diseases 2024 ,12, 149 3 of 14 on genomic insights. The design and development of the vaccine were personalized to the unique genetic profile of the chosen patient, focusing on the anomalies most likely to influence tumor behavior and treatment response. To confirm that our personalized vaccine design was rigorous and potentially effective, we compared the targeted genetic components of the vaccine to key genes previously identified as significant in the sur- vival of PDAC patients by Bhardwaj et al. [ 23]. This comparison enabled us to validate our personalized vaccine design approach and increase the therapeutic relevance of the vaccine design. This properly controlled process of data selection and comparison with established genetic markers supports our vaccine design methodology, which is detailed further in the section below. 2.2. RNAseq Analysis of Patient Data We performed an RNAseq analysis using the Partek Flow genomic analysis suite, as shown in Figure 1, which outlines our comprehensive RNAseq workflow to obtain and confirm variant data. Initially, we imported the raw sequence data in a fastq format into Partek Flow. This format is widely used for storing the output from high-throughput sequencing instruments and contains both nucleotide sequence data and corresponding quality scores. Following data importation, the first computational step involved trimming the Illumina sequencing adapters. These adapters, which are artificial sequences added during library preparation, can interfere with the analysis if not removed, as they may be misinterpreted as part of the genomic sequence. After trimming, we aligned the reads to a reference genome using the Burrows–Wheeler Aligner (BWA) algorithm version 0.7.18. BWA is a software tool that efficiently aligns relatively short sequences (such as those from Illumina sequencers) against a long reference sequence, such as a complete genome. This alignment is important for locating the genomic origins of each read and is fundamental to identifying variations from the reference sequence. In the post-alignment, we executed somatic variant calling using the Strelka algorithm, which was specifically designed to detect somatic variants with high sensitivity and accuracy in tumor-normal paired samples. This step was important for identifying potentially significant genetic mutations that could be relevant in the context of disease, herein cancer. To ensure the reliability of our findings, we manually inspected each significant gene variant using the Integrative Genomics Viewer (IGV) (Partek Inc., Chesterfield, Missouri). IGV is an interactive visualization tool that allows us to visually explore genomic data, thus facilitating the validation of computa- tional predictions through a critical human-oversight step. We excluded gene variants of inadequate quality from further analysis. This quality control step is key to avoiding false positives that could skew the results of downstream applications, such as vaccine develop- ment. Finally, we focused our efforts on analyzing single nucleotide polymorphisms (SNPs) that held potential for inclusion in our vaccine development process. SNPs, being the most common type of genetic variation among cancer patients, provide valuable insights into genetic variability, which can be exploited to design targeted vaccines. Diseases 2024 , 12, x FOR PEER REVIEW 4 of 15 Figure 1. RNAseq analysis work flow using Partek Flow suite. Created using BioRender.com. 3. Gene Annotation Con firmation Using SCGeneID Python Program Development and Applic ation of SCGeneID After obtaining and processing genomic data through Partek Flow, we advanced to the next step by developing a Python program named ‘SCGeneID’. The code for this in-novative tool is comprehensively detailed in Supplementary File S1 and is freely available for download. SCGeneID was speci fically designed to enhance ou r analytical capabilities in gene annotation by using both chromosome number and locus information. Using the hg38 reference set accessible via the UCSC Genome Browser [24], SCGeneID systemati- cally identi fies corresponding gene names based on their chromosomal location. The tool operates by exploiting web-scra ping techniques to extract relevant genomic data directly from the browser. Once the data are retrieve d, SCGeneID processes this information to generate a detailed output that includes a table forma tted with chromosome numbers, locus details, and the names of associated ge nes. This functionality not only streamlines the gene identi fication process but also warrants a ccuracy by referencing updated ge- nomic data. The application of SCGeneID in our study was twofold. Primarily, it served to externally validate the a lignment accuracy and overall reliability of our RNAseq anal- ysis process. By cross-verifying the gene anno tations provided by Partek Flow with those extracted by SCGeneID, we could con firm the consistency and validity of our results. Ad- ditionally, as shown in Figure 2, we employed a modi fied version of SCGeneID to specif- ically extract a list of genes from a given variant file. This adaptation was particularly im- portant for our personalized vaccine as it allo wed us to focus on particular genomic vari- ants of interest, facilitating a more targeted approach in our subsequent analyses. Figure 1. RNAseq analysis workflow using Partek Flow suite. Created using BioRender.com. Diseases 2024 ,12, 149 4 of 14 3. Gene Annotation Confirmation Using SCGeneID Python Program Development and Application of SCGeneID After obtaining and processing genomic data through Partek Flow, we advanced to the next step by developing a Python program named ‘SCGeneID’. The code for this innovative tool is comprehensively detailed in Supplementary File S1 and is freely available for download. SCGeneID was specifically designed to enhance our analytical capabilities in gene annotation by using both chromosome number and locus information. Using the hg38 reference set accessible via the UCSC Genome Browser [ 24], SCGeneID systematically identifies corresponding gene names based on their chromosomal location. The tool operates by exploiting web-scraping techniques to extract relevant genomic data directly from the browser. Once the data are retrieved, SCGeneID processes this information to generate a detailed output that includes a table formatted with chromosome numbers, locus details, and the names of associated genes. This functionality not only streamlines the gene identification process but also warrants accuracy by referencing updated genomic data. The application of SCGeneID in our study was twofold. Primarily, it served to externally validate the alignment accuracy and overall reliability of our RNAseq analysis process. By cross-verifying the gene annotations provided by Partek Flow with those extracted by SCGeneID, we could confirm the consistency and validity of our results. Additionally, as shown in Figure 2, we employed a modified version of SCGeneID to specifically extract a list of genes from a given variant file. This adaptation was particularly important for our personalized vaccine as it allowed us to focus on particular genomic variants of interest, facilitating a more targeted approach in our subsequent analyses. Diseases 2024 , 12, x FOR PEER REVIEW 5 of 15 Figure 2. SCGeneID Python program work flow. Created using BioRender.com.

4. Personalized Vaccine Design Protocol We employed a vaccine design protocol that has been previously outlined in our pub- lished studies [21]. This protocol integrates cu tting-edge bioinformatics tools to predict and select epitopes from mutations identi fied in genomic data. 4.1. Epitope Prediction and Selection Initially, we used the IEDB NetMHC 4.1 tool to predict epitopes. NetMHC 4.1 is spe- cifically designed to return potential epitop es along with their predicted binding a ffinity for the top 27 expressed HLA alleles in the human population. The binding a ffinity indi- cated by the IC 50 value measured in nanomolar (nM) determines the strength of the inter- action between the epitope and the HLA molecules, which is a critical factor in the im-mune response e fficacy. 4.2. Clinical Checkpoint Parameters Subsequently, we computed several epitope-speci fic clinical checkpoint parameters. The immunogenicity of each epitope was determined using the IEDB Class I Immunogen- icity Tool, which assesses the potential of an epitope to trigger an immune response. The antigenicity, which evaluates the capability of the epitope to be recognized by antibodies, was determined using VaxiJen v2.0. 4.3. Data Filtering and Selection Criteria With the binding a ffinity, immunogenicity, and antigenicity data computed for each epitope and its associated HLA allele, we employed stringent filters to select the most promising epitopes. These filters were applied based on the criteria outlined in Table 1, Figure 2. SCGeneID Python program workflow. Created using BioRender.com.

4. Personalized Vaccine Design Protocol We employed a vaccine design protocol that has been previously outlined in our published studies [ 21]. This protocol integrates cutting-edge bioinformatics tools to predict and select epitopes from mutations identified in genomic data. Diseases 2024 ,12, 149 5 of 14 4.1. Epitope Prediction and Selection Initially, we used the IEDB NetMHC 4.1 tool to predict epitopes. NetMHC 4.1 is specifically designed to return potential epitopes along with their predicted binding affinity for the top 27 expressed HLA alleles in the human population. The binding affinity indicated by the IC 50value measured in nanomolar (nM) determines the strength of the interaction between the epitope and the HLA molecules, which is a critical factor in the immune response efficacy. 4.2. Clinical Checkpoint Parameters Subsequently, we computed several epitope-specific clinical checkpoint parameters. The immunogenicity of each epitope was determined using the IEDB Class I Immunogenic- ity Tool, which assesses the potential of an epitope to trigger an immune response. The antigenicity, which evaluates the capability of the epitope to be recognized by antibodies, was determined using VaxiJen v2.0. 4.3. Data Filtering and Selection Criteria With the binding affinity, immunogenicity, and antigenicity data computed for each epitope and its associated HLA allele, we employed stringent filters to select the most promising epitopes. These filters were applied based on the criteria outlined in Table 1, fo- cusing on identifying epitopes that are strong binders, highly immunogenic, and antigenic. Table 1. Restriction criteria to quantitatively filter and qualitatively assess each epitope. Parameter Restriction Binding affinity ( b)Strong binder 0 nM ≤b≤50 nM Normal binder 50 nM < b≤500 nM Weak binder 500 nM < b≤5000 nM Immunogenicity ( i) i≥0 Antigenicity ( a) a≥0.4 Toxicity Non-toxic Allergenicity Non-allergenic 4.4. Physicochemical Property Assessment In addition to these functional assessments, we analyzed various physicochemical properties of the epitopes using ProtParam https://web.expasy.org/protparam/ (Accessed on 23 September 2023). This analysis included determining parameters such as half-life, instability index, isoelectric point, aliphatic index, and GRAVY score. Although these parameters were informative for understanding the physical and chemical characteristics of the epitopes, they were not used in the epitope selection process. Further, we assessed toxicity using ToxinPred and screened for allergenic potential using AllerTOP v2.0, ensuring that only non-toxic and non-allergenic epitopes were considered for further analysis. 4.5. Epitope Selection and Workflow Integration After applying the filtration restrictions (Table 1), we selected the top 50 and 1 00 epito pes that met all the specified criteria, warranting a robust selection of candidates for potential vaccine design. We employed binary filters on toxicity and allergenicity to ensure the selection of epitopes that were both non-toxic and non-allergenic. 4.6. Methodological Workflow Figure 3 shows the comprehensive workflow of our methodology, starting from the collection of Illumina sequencing data, performing RNAseq analysis, and the selection of top epitopes for the development of peptide vaccines. This streamlined workflow integrates Diseases 2024 ,12, 149 6 of 14 multiple stages of data processing and epitope evaluation, indicating the robustness of our approach in vaccine design. Diseases 2024 , 12, x FOR PEER REVIEW 7 of 15 Figure 3. Peptide Based Personalized Cancer Vaccine Design methodological overall work flow. 5. Results We obtained Illumina sequencing data from two patients out of the 19 available in the GEO accession project [23]. The sequencing data represent the genetic landscape of the patient’s solid tumor sample. We performe d RNAseq analysis to determine prevalent Figure 3. Peptide Based Personalized Cancer Vaccine Design methodological overall workflow. Diseases 2024 ,12, 149 7 of 14 5. Results We obtained Illumina sequencing data from two patients out of the 19 available in the GEO accession project [ 23]. The sequencing data represent the genetic landscape of the patient’s solid tumor sample. We performed RNAseq analysis to determine prevalent mutations. Using these mutations, we determined strong and normal binding MHC class I epitopes that are immunogenic, antigenic, non-toxic, and non-allergenic. We selected the top 50 and top 100 epitopes from these data for a peptide vaccine. 5.1. Determination of Genetic Variants with RNAseq Analysis We performed RNAseq analysis on Illumina sequencing data to obtain a list of genetic variants identified in a solid PDAC tumor. For Patient 1, the RNAseq analysis performed using Partek Flow resulted in 100,819 mutations. These mutations included single-nucleotide polymorphisms, multi-nucleotide polymorphisms, deletions, and insertions. Isolating the single-nucleotide polymorphisms, we identified 189 unique variants, which we could use to develop the peptide vaccine. For Patient 2, the RNAseq analysis resulted in 87,128 mutations., of which we identified 125 unique variants we could use to develop the peptide vaccine. 5.2. Confirmation of Genetic Variants and Sequencing Alignment Using SCGeneID We confirmed the alignment of the sequencing data to the hg38 human reference genome using our SCGeneID program version 1. Using SCGeneID, we qualitatively identified the corresponding genes to all mutations obtained through RNAseq analy- sis loci against the hg38 human reference genome. We found 100% similarity between the genes identified through Partek Flow and genes identified using SCGeneID for both Patients 1 and 2 . Therefore, we were confident that the variant genes identified using Partek Flow were correctly aligned to the reference genome. 5.3. Collection of 9-Mer and 10-Mer Top Epitopes from Genetic Variants From the pool of identified genetic variants, we curated lists of the top 50 and t op 10 0 epitopes, prioritized based on their binding affinity and immunogenic properties. The top epitopes for Patients 1 and 2 can be found in Supplementary Files S2–S5. All selected epitopes consisted of 9 or 10 amino acids, representing an epitope capable of binding to an MHC class I molecule. All the top 50 epitopes were classified as having strong binding affinity to their associated HLA allele. The top 100 epitopes included both strong and normal binders. We found no epitopes in the top 100, which were classified as weak binders. Table 2 shows the top 50 epitopes for Patient 1, along with their associated genes, mutations, and binding HLA alleles. 5.4. Population Coverage Analysis of Top 100 Epitopes We also performed a population coverage analysis to assess the extent of the global population that could potentially benefit from the personalized vaccine. The analysis for Patient 1 showed that the vaccine could cover 69.64% of the global population. Table 3 provides this coverage along with average hit rates and PC 90data for various world subregions. While the population coverage may appear relatively low at first glance, it is essential to consider the context of this study. The vaccine was uniquely designed based on the gene expression profile of a specific individual, making it personalized and tailored to the specific mutations and characteristics of their tumor. Consequently, the expectation for widespread coverage across diverse populations is not high. As the patient cohort from whom the vaccine was developed predominantly comprised individuals with European ancestry, the vaccine’s performance in these regional subgroups aligned with the genetic background of the patients involved. Diseases 2024 ,12, 149 8 of 14 Table 2. Top 50 epitopes along with their strong-binding associated HLA allele. Gene Mutation Epitope HLA Alleles GNAS R201C AMSNLVPPV HLA-A*02:01 SMAD4 Y353C QSIKETPCW HLA-B*58:01 TP53 R248Q CTYSPALNK HLA-A*03:01 KRAS G12D KSFEDIHHY HLA-B*58:01 SMAD4 Y353C MPIADPQPL HLA-B*39:01 SMAD4 Y353C CLSDHAVFV HLA-A*02:01 SMAD4 Y353C KIYPSAYIK HLA-A*03:01 TP53 R248Q LEDSSGNLL HLA-B*40:01 KRAS G12D LARSYGIPF HLA-B*15:01 TP53 R248Q APAAPTPAA HLA-B*07:02 SMAD4 Y353C LLDEVLHTM HLA-A*02:01 TP53 R248Q KTYQGSYGF HLA-B*58:01 SMAD4 Y353C APAISLSAA HLA-B*07:02 SMAD4 Y353C LQSNAPSSM HLA-B*15:01 TP53 R248Q LLGRNSFEV HLA-A*02:01 KRAS G12D KSALTIQLI HLA-B*58:01 SMAD4 Y353C KETPCWIEI HLA-B*40:01 GNAS R201C NQFRVDYIL HLA-B*39:01 TP53 R248Q LQIRGRERF HLA-B*15:01 SMAD4 Y353C LPHHQNGHL HLA-B*07:02 SMAD4 Y353C LQVAGRKGF HLA-B*15:01 SMAD4 Y353C CILRMSFVK HLA-A*03:01 KRAS G12D CLLDILDTA HLA-A*02:01 SMAD4 Y353C LRRLCILRM HLA-B*27:05 GNAS R201C LIDCAQYFL HLA-A*02:01 5.5. 3D-Structure Modeling of Epitope-MHC and TCR Interaction Complex TCR (T-cell receptor) and pMHC (peptide-major histocompatibility complex) interac- tions play a fundamental role in immunogenicity, which involves the ability of a peptide to initiate an immune response against tumor cells. TCRs on the surface of T cells recog- nize antigens that are presented by MHC molecules on the surface of antigen-presenting cells. This recognition is specific to the peptide being presented by the MHC. The cor- rect configuration and interaction of a TCR with a pMHC complex is essential for the T cell to become activated and initiate an immune response. Thus, to explore the bind- ing of our designed peptide vaccines, we initiated TCR-pMHC peptide interaction mod- eling. We found the PDB files for the HLA alleles HLA-B*58:01 on the RCSB protein data bank https://www.rcsb.org/ (accessed on 23 September 2023). Using MDockPeP https://zougrouptoolkit.missouri.edu/mdockpep/ (accessed on 23 September 2023) and CABS-dock [ 25,26], we attached a top epitope to the binding grooves of the HLA allele. We created two models of the peptide-MHC binding complex (Figure 4). TCR binding models were created using the same method as Kim et al. [ 21]. We used TCRModel https://tcrmodel.ibbr.umd.edu/ (accessed on 23 September 2023) to create 3D models of a TCR complex binding to our peptide-MHC complexes. Subsequently, we used PyMOL version 2.5.5 to edit all of the 3D models. In Figure 4, yellow color represents HLA alleles, and red represents epitopes. The 3D models we obtained were KSFEDIHHY, a mutation of Diseases 2024 ,12, 149 9 of 14 the KRAS gene, binding to the MHC Class I molecule HLA-B*58:01 as well as KTYQGSYGF, a mutation of the TP53 gene, binding to the MHC Class I molecule HLA-B*58:01. All pMHC- TCR 3D molecular model structures generated in this study can be found in Figure 5 and Supplementary Files S6–S9. Table 3. Population coverage of the personalized PDAC vaccine for regional subgroups. Population/Area Coverage Average Hit pc90 Central Africa 39.22 1.84 0.16 Central America 1.4 0.06 0.41 East Africa 41.73 2.16 0.17 East Asia 55.26 2.8 0.22 Europe 81.05 4.98 0.53 North Africa 43.55 2.29 0.18 North America 70.36 4.09 0.34 Northeast Asia 47.97 2.21 0.19 Oceania 38.93 1.59 0.16 South Africa 23.99 0.93 0.13 South America 36.87 1.86 0.16 South Asia 37.28 1.66 0.16 Southeast Asia 55.59 2.34 0.23 Southwest Asia 43.73 2.33 0.18 West Africa 42.65 2.14 0.17 West Indies 63.52 3.47 0.27 Average 45.19 2.3 0.23 Standard deviation 17.8 1.13 0.11 Diseases 2024 , 12, x FOR PEER REVIEW 10 of 15 recognize antigens that are presented by MH C molecules on the surface of antigen-pre- senting cells. This recognition is speci fic to the peptide being presented by the MHC.

The correct con figuration and interaction of a TCR with a pMHC complex is essential for the T cell to become activated and initiate an i mmune response. Thus, to explore the binding of our designed peptide vaccines, we initiated TCR-pMHC peptide interaction modeling. We found the PDB files for the HLA alleles HLA-B*58:01 on the RCSB protein data bank https://www.rcsb.org/ (accessed on 23 September 2023). Using MDockPeP https://zougrouptoolkit.missour i.edu/mdockpep/ (accessed on 23 September 2023) and CABS-dock [25,26], we a ttached a top epitope to the binding grooves of the HLA allele. We created two models of the peptide-MHC binding complex (Figure 4). TCR binding models were created using the same method as Kim et al. [21]. We used TCRModel https://tcrmodel.ibbr.umd. edu/ (accessed on 23 September 2023) to create 3D models of a TCR complex binding to our peptide-MHC complexes. Subsequently, we used PyMOL version 2.5.5 to edit all of the 3D models. In Figure 4, yellow color represents HLA alleles, and red represents epitopes. The 3D models we obtained were KSFEDIHHY, a mutation of the KRAS gene, binding to the MHC Class I molecule HLA-B*58:01 as well as KTYQGSYGF, a mutation of the TP53 gene, binding to the MHC Class I molecule HLA- B*58:01. All pMHC-TCR 3D molecular model st ructures generated in this study can be found in Figure 5 and Supplementary Files S6–S9. (A) ( B) Figure 4. (A) The peptide KSFEDIHHY, a mutation of th e KRAS gene, bound to MHC Class I mol- ecule HLA-B*58:01. ( B) The peptide KTYQGSYGF, a mutation of the TP53 gene, bound to MHC Class I molecule HLA-B*58:01. Figure 4. (A) The peptide KSFEDIHHY, a mutation of the KRAS gene, bound to MHC Class I molecule HLA-B*58:01. ( B) The peptide KTYQGSYGF, a mutation of the TP53 gene, bound to MHC Class I molecule HLA-B*58:01. Diseases 2024 ,12, 149 10 of 14 Diseases 2024 , 12, x FOR PEER REVIEW 11 of 15 Figure 5. The peptides KSFEDIHHY and KTYQGSYGF boun d to HLA-B*58:01 and their respective TCR complex. 6. Discussion We developed a personalized peptide-based vaccine for two patients with pancreatic ductal adenocarcinoma (PDAC). This proc ess began with RNA sequencing (RNAseq) analysis, which enabled the identi fication of speci fic genetic mutations driving the devel- opment of PDAC in the patients. Based on this analysis, we developed a personalized cancer vaccine using our previously published peptide vaccine development strategy. Our approach involved targeting 100 epitopes that were prevalent in the PDAC patient and identi fied as viable candidates for peptide vaccine design. By focusing on the speci fic gene targets present in each patient, we intended to improve the speci ficity of the vaccine, en- suring that it e ffectively targeted the unique genetic alterations present in the patient’s tumor. This method not only enhances the potential e fficacy of the vaccine by adapting it to the individual’s genetic landsc ape but also mini mizes potential o ff-target effects, thus optimizing the therapeutic outcome. The final filtered epitopes are predicted to be immunogenic and antigenic, have a high or normal binding a ffinity, and are non-toxic and non-allergenic. The binding a ffinity restriction used in this study di ffers from other previous in silico vaccine design method- ologies using the same NetMHC tool. Our pr evious methods of peptide vaccine design used quantitative filters on the percentile rank of the binding a ffinity value. However, the percentile rank compares the epitopes to a test set of data in IEDB and, therefore, is not an accurate nor absolute assessment of binding a ffinity necessary for this study. Using the IC50 value instead is an absolute measure of the binding a ffinity of the epitopes. We are also able to specify the strength of the binding a ffinity based on the IC 50 value, which pro- vides more qualitative measures for comparison when transitioning to murine studies. By tailoring the vaccine to each patient’s speci fic genetic makeup, we expect to enhance its effectiveness and improve clinical outcomes. This approach represents a signi ficant step forward in the field of immunotherapy for PDAC, o ffering a more targeted and personal- ized treatment option that has the potential to transform the management of this challeng- ing disease. Figure 5. The peptides KSFEDIHHY and KTYQGSYGF bound to HLA-B*58:01 and their respective TCR complex. 6. Discussion We developed a personalized peptide-based vaccine for two patients with pancre- atic ductal adenocarcinoma (PDAC). This process began with RNA sequencing (RNAseq) analysis, which enabled the identification of specific genetic mutations driving the devel- opment of PDAC in the patients. Based on this analysis, we developed a personalized cancer vaccine using our previously published peptide vaccine development strategy. Our approach involved targeting 100 epitopes that were prevalent in the PDAC patient and identified as viable candidates for peptide vaccine design. By focusing on the specific gene targets present in each patient, we intended to improve the specificity of the vaccine, ensuring that it effectively targeted the unique genetic alterations present in the patient’s tumor. This method not only enhances the potential efficacy of the vaccine by adapting it to the individual’s genetic landscape but also minimizes potential off-target effects, thus optimizing the therapeutic outcome. The final filtered epitopes are predicted to be immunogenic and antigenic, have a high or normal binding affinity, and are non-toxic and non-allergenic. The binding affinity restriction used in this study differs from other previous in silico vaccine design methodologies using the same NetMHC tool. Our previous methods of peptide vaccine design used quantitative filters on the percentile rank of the binding affinity value. However, the percentile rank compares the epitopes to a test set of data in IEDB and, therefore, is not an accurate nor absolute assessment of binding affinity necessary for this study. Using the IC 50value instead is an absolute measure of the binding affinity of the epitopes. We are also able to specify the strength of the binding affinity based on the IC 50value, which provides more qualitative measures for comparison when transitioning to murine studies. By tailoring the vaccine to each patient’s specific genetic makeup, we expect to enhance its effectiveness and improve clinical outcomes. This approach represents a significant step forward in the field of immunotherapy for PDAC, offering a more targeted and personalized treatment option that has the potential to transform the management of this challenging disease. Diseases 2024 ,12, 149 11 of 14 The top epitopes selected using our novel methodology are widely recognized in the literature as common drivers and tumor-suppressor genes in PDAC [ 9,27,28]. Addi- tionally, these specific epitopes have been identified in trials involving the sequencing of human tumor samples [ 29,30]. The consistent presence of our top epitopes in both our reference study and other clinical trials of PDAC patients serves as strong validation of our personalized cancer vaccine design methodology. Using RNA sequencing analysis by Partek Flow, along with our peptide cancer vaccine design processes, we created a peptide vaccine derived from the individual’s tumor tissue genetic data. This integrative approach not only emphasizes the relevance of our vaccine targets but also enhances the precision medicine framework by adapting the therapeutic strategy to the genetic individualities of each patient’s tumor. This could potentially lead to improved clinical outcomes by specifically targeting the molecular abnormalities driving the cancer. Previous studies on the development of peptide vaccines have primarily concen- trated on creating generalized vaccines that could be used for a large and broad popu- lation [ 15,31–33]. These generalized vaccines target a limited set of gene mutations to increase sensitivity but often at the expense of specificity. The development of effec- tive global peptide vaccines poses additional challenges. The vast global diversity of HLA alleles complicates the creation of a peptide vaccine that can effectively target a comprehensive population [ 31]. Each individual’s HLA type influences how well their immune system can recognize and respond to the peptides presented by the vaccine, making it difficult to design a universally effective vaccine. The development of person- alized peptide vaccines has historically been limited by the cost and time to produce the peptides [ 32]. However, implementing a novel design method described in this study offers a unique and innovative solution to quickly design neoantigen personalized peptide-based vaccines. Recently, with the advent of advanced sequencing technology, neoantigen peptide vaccines are becoming a more viable solution for patients [ 34]. How- ever, the design process has been complicated by a multitude of software required to design a personalized vaccine. Our methodology using Partek Flow provides a sim- ple and streamlined RNAseq analysis procedure to obtain the list of neoantigens. Our program, SCGeneID, is useful for identifying and confirming proper alignment and identification of genes from the RNAseq analysis process. Overall, our methodology employs only two tools throughout the entire design process, significantly simplifying the development of personalized cancer vaccines. This streamlined approach not only reduces the complexity and duration of vaccine design but also enhances the precision with which these vaccines can be personalized to individual genetic profiles. 7. Limitations While this study presents a promising personalized cancer vaccine strategy targeting neoantigens in pancreatic ductal adenocarcinoma (PDAC) patients, there are several limita- tions that should be acknowledged. Firstly, the pilot trial size of two patients is relatively small, which could limit the generalizability of the methodology. A larger sample size would provide more robust data and a better account for variations and accommodate the heterogeneity inherent in the genetic landscape of PDAC more effectively. The purpose of this paper was to demonstrate a successful method for designing personalized peptide cancer vaccines. However, in future outcome-oriented studies, the use of a larger sample size would be favorable. While the absence of experimental confirmation may appear as a limitation, the significance of this innovative methodological framework for personalized cancer vaccines, being the first of its kind, corroborates the importance of this work. This framework enables the efficient prioritization of the most promising personalized vaccine candidates, thus accelerating the vaccine design process, enhancing the probability of success in subsequent preclinical and clinical evaluations, and also helping to optimize re- sources by focusing on the candidates for further preclinical studies. Peptide vaccines have their weaknesses in functionality as well. If a patient’s cancer significantly downregulates MHC, the probability of a peptide binding to an MHC receptor significantly decreases. Diseases 2024 ,12, 149 12 of 14 8. Future Directions We have developed an automation of the peptide vaccine design process using web scraping and API tools [ 24,25]. Implementation of such software would further simplify the personalized cancer vaccine process. Additionally, the use of these programs would allow for the prediction of MHC class II epitopes as well. Furthermore, moving the RNAseq analysis process from a cloud-based solution using Partek Flow to a hardware process using Python or R would allow for complete automation of the personalized vaccine design process. Given such a scaled program and processes, the only limitation to the vaccine design process would be the time to sequence a patient’s tumor tissue. 9. Conclusions We developed a personalized cancer vaccine targeting specific gene mutations preva- lent among PDAC patients by implementing our novel personalized vaccine design work- flow. This study addresses the limitations of generalized vaccines specifically for pancreatic ductal adenocarcinoma (PDAC). By analyzing the genetic alterations driving PDAC in a patient’s tumor tissue, we identified 100 gene mutations as targets for our personalized vaccine strategy. The gene targets were identified and validated using our SCGeneID program, which used the chromosome number and nucleotide position data. By inte- grating SCGeneID into our workflow, we not only enhanced the precision of our gene annotations but also significantly improved the efficiency of our data analysis process. This development represents a significant step forward in the application of computational tools in personalized vaccine design, providing a robust method for accurate gene identification and the validity of complex genomic analyses. The top 50 epitopes consisted of only high-affinity binding epitopes, indicating the potential efficacy of the vaccine. The use of IC 50values as an absolute measure of binding affinity provided more accurate and quantitative comparisons. To visualize the interactions between epitopes and HLA alleles, 3D models of TCR-peptide-MHC complexes were created. The personalized cancer vaccine developed in this study may hold great promise for PDAC patients. By targeting the unique genetic alterations in each patient’s tumor, this approach offers a more specific and personalized treatment option. Further research is warranted to simplify the variant identification and epitope ranking process. Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/diseases12070149/s1, File S1: SCGeneID Python program; F ile S2 : Top 50 Epitopes for Patient 1; File S3: Top 100 Epitopes for Patient 1; File S4: Top 50 Epitopes for Patient 2; File S5: Top 100 Epitopes for Patient 2; File S6: 3-dimensional structure of pMHC complex in Figure 4a; File S7: 3-dimensional structure of pMHC complex in Figure 4b; File S8: 3-dimensional structure of pMHC-TCR complex in Figure 5a; File S9: 3-dimensional structure of pMHC-TCR complex in Figure 5b. Author Contributions: Conceptualization, S.D.; methodology, S.D.

and K.S.; software, S.D. and K.S.; validation, K.S.; formal analysis, K.S.; investigation, K.S.; resources, S.D.; data curation, K.S.; writing—orig inal draft preparation, S.D. and K.S.; writing—review and editing, S.D. and K.S.; visu- alization, K.S.; supervision, S.D.; project administration, S.D.; funding acquisition, S.D. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: All the data supporting reported results can be found in the Supplementary Materials. Acknowledgments: The author, S.D. (Sivanesan Dakshanamurthy), acknowledges partial support from the Georgetown University Lombardi Comprehensive Cancer Center (LCCC) Cancer Cell Biology Program Pilot Award. The author K.S. (Kush Savsani) participated in the Lombardi Compre- hensive Cancer Center, Summer Research Program. Diseases 2024 ,12, 149 13 of 14 Conflicts of Interest: The authors declare no conflict of interest. References 1. Shemesh, C.S.; Hsu, J.C.; Hosseini, I.; Shen, B.Q.; Rotte, A.; Twomey, P .; Girish, S.; Wu, B. Personalized Cancer Vaccines: Clinical Landscape, Challenges, and Opportunities. Mol. Ther. J. Am. Soc. Gene Tssher 2021 ,29, 555–570. [CrossRef] [PubMed] 2. Sahin, U.; Türeci, Ö. Personalized vaccines for cancer immunotherapy. Science 2018 ,359, 1355–1360. [CrossRef] [PubMed] 3. Rojas, L.A.; Sethna, Z.; Soares, K.C.; Olcese, C.; Pang, N.; Patterson, E.; Lihm, J.; Ceglia, N.; Guasp, P .; Chu, A.; et al. Personalized RNA neoantigen vaccines stimulate T cells in pancreatic cancer. Nature 2023 ,618, 144–150.

[CrossRef] [PubMed] 4.

Kong, X.; Cheng, D.; Xu, X.; Zhang, Y.; Li, X.; Pan, W. IFN- γinduces apoptosis in gemcitabine-resistant pancreatic cancer cells. Mol.

Med. Rep. 2024 ,29, 76. [CrossRef] 5. Park, W.; Chawla, A.; O’Reilly, E.M. Pancreatic Cancer: A Review. JAMA 2021 ,326, 851–862. [CrossRef] [PubMed] 6. Anderson, E.M.; Thomassian, S.; Gong, J.; Hendifar, A.; Osipov, A. Advances in Pancreatic Ductal Adenocarcinoma Treatment. Cancers 2021 ,13, 5510. [CrossRef] [PubMed] 7. Strobel, O.; Lorenz, P .; Hinz, U.; Gaida, M.; König, A.K.; Hank, T.; Niesen, W.; Kaiser, J.O.R.; Al-Saeedi, M.; Bergmann, F.; et al. Actual Five-year Survival After Upfront Resection for Pancreatic Ductal Adenocarcinoma: Who Beats the Odds? Ann.

Surg. 2022 , 275, 962–971. [CrossRef] 8. Tonini, V .; Zanni, M. Pancreatic cancer in 2021: What you need to know to win. World J. Gastroenterol. 2021 ,27, 5851–5889. [CrossRef] 9.

Hu, H.-F.; Ye, Z.; Qin, Y.; Xu, X.-W.; Yu, X.-J.; Zhuo, Q.-F.; Ji, S.-R. Mutations in key driver genes of pancreatic cancer: Molecularly targeted therapies and other clinical implications. Acta Pharmacol.

Sin. 2021 ,42, 1725–1741. [CrossRef] 10.

Shen, H.; Lundy, J.; Strickland, A.H.; Harris, M.; Swan, M.; Desmond, C.; Jenkins, B.J.; Croagh, D. KRAS G12D Mutation Subtype in Pancreatic Ductal Adenocarcinoma: Does It Influence Prognosis or Stage of Disease at Presentation? Cells 2022 ,11, 3175.

[CrossRef] 11. McCubrey, J.A.; Yang, L.V .; Abrams, S.L.; Steelman, L.S.; Follo, M.Y.; Cocco, L.; Ratti, S.; Martelli, A.M.; Augello, G.; Cervello, M. Effects of TP53 Mutations and miRs on Immune Responses in the Tumor Microenvironment Important in Pancreatic Cancer Progression. Cells 2022 ,11, 2155.

[CrossRef] 12. Sun, H.; Zhang, B.; Li, H. The Roles of Frequently Mutated Genes of Pancreatic Cancer in Regulation of Tumor Microenvironment. Technol. Cancer Res. Treat. 2020 ,19, 1533033820920969. [CrossRef] 13. Wartenberg, M.; Cibin, S.; Zlobec, I.; Vassella, E.; Eppenberger-Castori, S.; Terracciano, L.; Eichmann, M.D.; Worni, M.; Gloor, B.; Perren, A.; et al. Integrated Genomic and Immunophenotypic Classification of Pancreatic Cancer Reveals Three Distinct Subtypes with Prognostic/Predictive Significance. Clin.

Cancer Res. Off. J. Am. Assoc. Cancer Res. 2018 ,24, 4444–4454. [CrossRef] 14. Huang, X.; Zhang, G.; Tang, T.Y.; Gao, X.; Liang, T.B. Personalized pancreatic cancer therapy: From the perspective of mRNA vaccine. Mil.

Med. Res. 2022 ,9, 53. [CrossRef] 15. Liu, W.; Tang, H.; Li, L.; Wang, X.; Yu, Z.; Li, J. Peptide-based therapeutic cancer vaccine: Current trends in clinical application. Cell Prolif. 2021 ,54, e13025. [CrossRef] [PubMed] 16. Gjertsen, M.K.; Bakka, A.; Breivik, J.; Saeterdal, I.; Gedde-Dahl, T.; Stokke, K.T.; Solheim, B.G.; Egge, T.S.; Søreide, O.; Thorsby, E.; et al. Ex vivo ras peptide vaccination in patients with advanced pancreatic cancer: Results of a phase I/II study. Int. J. Cancer 1996 ,65, 450–453. [CrossRef] 17. Suzuki, N.; Hazama, S.; Ueno, T.; Matsui, H.; Shindo, Y.; Iida, M.; Yoshimura, K.; Yoshino, S.; Takeda, K.; Oka, M. A Phase I Clinical Trial of Vaccination With KIF20A-derived Peptide in Combination With Gemcitabine For Patients With Advanced Pancreatic Cancer. J.

Immunother. 2014 ,37, 36–42. [CrossRef] [PubMed] 18. Suzuki, N.; Hazama, S.; Iguchi, H.; Uesugi, K.; Tanaka, H.; Hirakawa, K.; Aruga, A.; Hatori, T.; Ishizaki, H.; Umeda, Y.; et al. Phase II clinical trial of peptide cocktail therapy for patients with advanced pancreatic cancer: VENUS-PC study. Cancer Sci.

2017 , 108, 73–80. [CrossRef] 19.

Nishida, S.; Ishikawa, T.; Egawa, S.; Koido, S.; Yanagimoto, H.; Ishii, J.; Kanno, Y.; Kokura, S.; Yasuda, H.; Oba, M.S.; et al. Combination Gemcitabine and WT1 Peptide Vaccination Improves Progression-Free Survival in Advanced Pancreatic Ductal Adenocarcinoma: A Phase II Randomized Study. Cancer Immunol.

Res. 2018 ,6, 320–331. [CrossRef] 20. Gilliam, A.D.; Broome, P .; Topuzov, E.G.; Garin, A.M.; Pulay, I.; Humphreys, J.; Whitehead, A.; Takhar, A.; Rowlands, B.J.; Beckingham, I.J. An international multicenter randomized controlled trial of G17DT in patients with pancreatic cancer. Pancreas 2012 ,41, 374–379. [CrossRef] 21.

Kim, M.; Savsani, K.; Dakshanamurthy, S. A Peptide Vaccine Design Targeting KIT Mutations in Acute Myeloid Leukemia. Pharmaceuticals 2023 ,16, 932. [CrossRef] [PubMed] 22.

Edgar, R.; Domrachev, M.; Lash, A.E. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 2002 ,30, 207–210. [CrossRef] [PubMed] 23. Bhardwaj, A.; Josse, C.; Van Daele, D.; Poulet, C.; Chavez, M.; Struman, I.; Van Steen, K. Deeper insights into long-term survival heterogeneity of pancreatic ductal adenocarcinoma (PDAC) patients using integrative individual- and group-level transcriptome network analyses. Sci.

Rep. 2022 ,12, 11027. [CrossRef] [PubMed] 24. Kent, W.J.; Sugnet, C.W.; Furey, T.S.; Roskin, K.M.; Pringle, T.H.; Zahler, A.M.; Haussler, D. The Human Genome Browser at UCSC. Genome Res. 2002 ,12, 996–1006. [CrossRef] [PubMed] Diseases 2024 ,12, 149 14 of 14 25. Kurcinski, M.; Pawel Ciemny, M.; Oleniecki, T.; Kuriata, A.; Badaczewska-Dawid, A.E.; Kolinski, A.; Kmiecik, S. CABS-dock standalone: A toolbox for flexible protein–peptide docking. Bioinformatics 2019 ,35, 4170–4172. [CrossRef] [PubMed] 26. Xu, X.; Yan, C.; Zou, X. MDockPeP: An ab-initio protein-peptide docking server. J. Comput. Chem. 2018 ,39, 2409–2413. [CrossRef] 27. Ying, H.; Dey, P .; Yao, W.; Kimmelman, A.C.; Draetta, G.F.; Maitra, A.; DePinho, R.A. Genetics and biology of pancreatic ductal adenocarcinoma. Genes. Dev. 2016 ,30, 355–385. [CrossRef] 28. Saiki, Y.; Jiang, C.; Ohmuraya, M.; Furukawa, T. Genetic Mutations of Pancreatic Cancer and Genetically Engineered Mouse Models. Cancers 2021 ,14, 71. [CrossRef] [PubMed] 29. Millar, D.G.; Yang, S.Y.C.; Sayad, A.; Zhao, Q.; Nguyen, L.T.; Warner, K.; Sangster, A.G.; Nakatsugawa, M.; Murata, K.; Wang, B.X.; et al. Identification of antigenic epitopes recognized by tumor infiltrating lymphocytes in high grade serous ovarian cancer by multi-omics profiling of the auto-antigen repertoire. Cancer Immunol.

Immunother. 2023 ,72, 2375–2392. [CrossRef] 30.

Baleeiro, R.B.; Bouwens, C.J.; Liu, P .; Di Gioia, C.; Dunmall, L.S.C.; Nagano, A.; Gangeswaran, R.; Chelala, C.; Kocher, H.M.; Lemoine, N.R.; et al. MHC class II molecules on pancreatic cancer cells indicate a potential for neo-antigen-based immunotherapy. Oncoimmunology 2022 ,11, 2080329. [CrossRef] 31. Abd-Aziz, N.; Poh, C.L. Development of Peptide-Based Vaccines for Cancer. J. Oncol. 2022 ,2022 , 9749363. [CrossRef] [PubMed] 32. Stephens, A.J.; Burgess-Brown, N.A.; Jiang, S. Beyond Just Peptide Antigens: The Complex World of Peptide-Based Cancer Vaccines. Front. Immunol. 2021 ,12, 696791. Available online: https://www.frontiersin.org/journals/immunology/articles/10.338 9/fimmu.2021.696791/full (accessed on 24 March 2024). [CrossRef] [PubMed] 33. Mizukoshi, E.; Nakagawa, H.; Tamai, T.; Kitahara, M.; Fushimi, K.; Nio, K.; Terashima, T.; Iida, N.; Arai, K.; Yamashita, T.; et al. Peptide vaccine-treated, long-term surviving cancer patients harbor self-renewing tumor-specific CD8+ T cells. Nat.

Commun. 2022 ,13, 3123. [CrossRef] [PubMed] 34. Biswas, N.; Chakrabarti, S.; Padul, V .; Jones, L.D.; Ashili, S. Designing Neoantigen Cancer Vaccines, Trials, and Outcomes. Front. Immunol. 2023 ,14, 1105420. Available online: https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2023 .1105420/full (accessed on 25 March 2024). [CrossRef] [PubMed] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Ludwig Institute for Cancer Research Statistical Analysis Plan A Phase 1/2 Study of Combinati on Immun otherapy and mRNA Vaccine in Subj ects with Non-small Cell Lung Ca ncer (NSCLC) Protocol Number LUD2014 -012-VAC NCT03164772 Amendment 9.1, 01-SEP-2020 SAP Version 3.0 (09DEC2020) Prepared By: Advanced Clinical 10 Parkway North, Suite 350 Deerfield, IL 60015 TABLE OF CONTENTS 1.0 PURPOSE ................................ ................................

............................. 1 2.0 OVERALL STUDY DESIGN AND OBJECTIVES ................................ ....... 1 2.1 Trial Objectives .................................................................................. 1 2.2 Study Endpoints ................................................................................. 2 2.3 Trial Desi gn and Trial Procedures ........................................................ 3 2.4 Treatments and Assignment to Treatments .......................................... 4 2.5 Determination of Sample Size .............................................................. 4 3.0 GENERAL ANALYSIS CONVENTIONS ................................ ................... 5 3.1 General Conventions ........................................................................... 5 3.2 Trial Periods ....................................................................................... 5 3.3 Visit Windows ..................................................................................... 6 3.4 Baseline ............................................................................................. 6 4.0 ANALYSIS POPULATIONS ................................ ................................ .... 6 4.1 Intent -to-Treat Population .................................................................. 6 4.2 Per-Protocol Population (Clinical Efficacy) ............................................. 6 4.3 Per-Protocol Population (DLT Assessments) ......................................... 7 5.0 SUBJECT DISPOSITION ................................ ................................

........ 7 6.0 DEMOGRAPHIC AND BASELINE CHARACTERISTICS ........................... 8 6.1 Demographic Characteristics ............................................................... 8 6.2 Medical History ................................................................................... 8 6.3 Prior and Concomitant Medications and Procedures .............................. 9 7.0 EFFICACY ANALYSIS ................................ ................................ ........... 9 7.1 Response Rate ................................................................................... 9 7.2 Objective Response Rate (ORR) .......................................................... 9 7.3 Disease Control Rate (DCR) .............................................................. 10 7.4 Progression -free Survival (PFS) ......................................................... 10 7.5 Overall Survival (OS) ........................................................................ 11 7.6 Duration of Response (DoR) .............................................................. 11 7.7 Analysis of Time to Event (TTE) Variables .......................................... 11 8.0 TRIAL DRUG EXPOSURE AND COMPLIANCE ................................ ..... 12 9.0 SAFETY ANALYSIS ................................ ................................ ............ 12 9.1 Adverse Events ................................................................................ 12 9.1.1 Treatment -Emergent Adverse Events ................................... 13 9.1.2 Adverse Events of Special Interest (AESIs) .......................... 15 9.2 Clinical Laboratory Data .................................................................... 15 9.3 Vital Signs and Weight ...................................................................... 17 9.4 Electrocardiograms ........................................................................... 17 9.5 ECOG Performance ........................................................................... 18 9.6 Physical Examinations ....................................................................... 18 9.7 Pregnancy Testing ............................................................................ 18 10.0 EXPLORATORY ANALYSIS ................................ ................................ 18 11.0 STA TISTICAL/ANALYTICAL ISSUES ................................ ................... 19 11.1 Handling of Dropouts or Missing Data ................................................ 19 11.2 Safety Data Handling ........................................................................ 19 11.3 Coding Dictionaries ........................................................................... 20 11.4 Pooling of Centers in Multi -Center Trials ............................................. 20 11.5 Multiple Comparisons/Multiplicity ....................................................... 20 11.6 Examination of Subgroups ................................................................ 20 11.7 Interim Analysis and Data Monitoring ................................................ 20 12.0 QUALITY CONTROL ................................ ................................

.......... 20 13.0 TABLES AND LISTING CONVENTIONS ................................ ............... 21 14.0 REFERENCES ................................ ................................

.................... 22 15.0 RECORD RETENT ION ................................ ................................

........ 22 16.0 CHANGE HISTORY ................................ ................................

............ 23 17.0 APPENDICES ................................ ................................ ...................... 24 17.1 Appendix A: Study Flowchart ............................................................. 24 17.2 Appendix B: Drug Administrative Schedules per Cycle ......................... 27 ACRONYMS Below is the list of acronyms that will be used throughout this document. Abbreviation Definition AE adverse event AESI adverse event of special i nterest CI confidence intervals CR Complete response CTC AE V.4.03 Common Terminology Criteria for Adverse Events v. 4.03 CTLA-4 Cytotoxic T lymphocyte-associated antigen 4 DNA Deoxy ribonucleic Acid DCR Disease Control Rate DLT Dose limiting toxicity DoR Duration of response ECG electrocardiogram ECOG Eastern Cooperative Oncology Group FFPE Formalin-fixed paraffin -embedded GCP Good Clinical Practice ICH International Conference on Harmonization ILD Interstitial l ung disease irRECI ST Immune-related Response Evaluation Cri teria in Solid Tumors MedDRA Medical Dictionary for Regulatory Activities NCI CTCAE National Ca ncer Institute Common Terminology Criteria for d NSCLC Non-small cell lung cancer ORR Objective Response Rate OS Overall survival PBMC Peripheral blood mononuclear c ell PD Progressive disease PD-L1 Programmed death li gand 1 PFS Progression free survival PR Partial response Q4W Every 4 weeks RECI ST Response Evaluation Cri teria in Solid T umors RCD Recommended combination dose RNA Ribonucleic Acid SAE Serious Adverse Event SD Stable disease TEAE Treatment Emergent Adverse Event TIL tumor-infiltra ting lymphocyte TME Tumor microenvironment Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09 DEC2020 Page 1 of 27 1.0 PURPOSE This SAP describes the methods to be used in the analysis of trial data from clinical protocol “ A Phase 1/2 Study of Combinati on Immunothe rapy and mRN A Va ccine in Subje cts with Non- small Ce ll Lung Cancer (NSCLC)” in order to answer the trial objective(s), and is based on Amendment 9.1 of the trial protocol ( LUD2014-012- VAC) , dated 01 September 2020. Populations for analysis, data handling rules, statistical methods, and formats for data presentation are described within this document. The statistical analyses and summary tabulations described in this SAP will provide the basis for the results sections of the CSR for this trial. The SAP outlines any differences in data analysis methods relative to those planned in the trial protocol. Any changes to the data analysis methods after the SAP is finalized will be described in the CSR. 2.0 OVERALL STUDY DESIGN AND OBJECTIVE S 2.1 Trial Objectives This study w ill eval uate the safety and e fficacy of the add ition of a v accine the rapy (BI 1361849, previously known as CV9202) to 1 or 2 che ckpoint inhibitors (durval umab and tre melimumab) for NS CLC. The study has two phases; the dose evaluation phase and the dose expansion phase. In the dose expansion phase, there are two cohorts as follows: Arm A: mRNA Vaccine [BI 1361849 (formerly CV9202)] + anti- PD-L1 [durvalumab] Arm B: mRNA Vaccine [BI 1361849] + anti- PD-L1 [durva lumab] + anti -CTLA -4 [tremelimumab]. The p rimary trial objectives are: • Dose evaluation Phase: To evaluate s afety an d toler ability, including Dose Limiting Toxicities ( DLTs) and Recommended Combination Dose ( RCD) of a combination immunotherapy with durval umab or durvalumab and tr emelimum ab together with an RNA vaccine BI 1361849. • E xpansion Phase: To e valuate s afety an d toler ability of a combination immunotherapy with durval umab or durvalumab and tr emelimumab together with an RNA vaccine BI 1361849. The secondary trial objective is: • To evaluate the clinical efficacy of the combination immunotherapy for all subjects in both the dose evaluation and the dose expansion phases. The exploratory trial objective is: Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 2 of 27 • To evaluate Bi ologic Ac tivity and the e ffects on t umor microenvironment, and i mmune response for all subjects in both the dose evaluation and the dose expansion phases. 2.2 Study Endpoints Primary Endpoints The primary endpoint of the combination therapy of durval umab and tremelimumab are: • Dose- Lim iting Toxici ties (DL Ts). DLTs are de fined as any ad verse events that are possibly, probably, or de finitely related to the administra tion of durval umab, tremelimumab, or BI 1361849 components. This applies to Dose Evaluation subject s during the DLT Evaluation period. • Safety and tolerability evaluated using CTCAE 4.03. • Recommended Combination Dose (RCD) determined in Dose Evaluation Phase. The safety endpoints for this study include: • Treat ment-emergent adv erse events (TEA Es) and serious AEs ( SAEs) • Clinical la boratory te sts (hematology and chemistry) • Vital signs and we ight me asurements • ECG • ECOG pe rformance status eval uation • An y other medically indicated assessments, incl uding s ubject intervi ews • All safety analyses will be reported separately for Arm A and Arm B The Clinical efficacy endpoints will be assessed by ir RECI ST and RECI ST 1.1 and will include: • Objective Response rate (ORR) at 8 and 24 weeks. ORR is defined as the pe rcentage of subjects meeting c riteria of Complete Respon se (CR) or P artial Response (PR) over a period of at least 4 weeks.

• Duration of Response (DoR). DoR is defined as the interval between the date of earliest determination of CR or PR to the date of earliest determination of PD, or to the date of death, if PD does not occur. • Disease Control rate (D CR). DCR is defined as the p ercenta ge of subjects mee ting criteria of S table Dise ase (SD), PR, or CR ove r a period of at lea st 4 we eks. • Progressive Free Survival (P FS) at 8 and 24 weeks. PFS is defi ned as the inte rval between t he date of first dose to the date of e arliest determinat ion of Progre ssive D isease Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC.

CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 3 of 27 (PD), or to the da te of de ath, if PD does not occur. Week 8 corresponds to Cycle 3 Day 1 and Week 24 corresponds to Cycle 7 Day 1. • Overall Survival (OS). OS is defi ned as the inte rval between t he date of first dose until the date of death or t he date of last follow-up. • All efficacy analyses will be reported separately for Arm A and Arm B. Pharmacodynamics analyses on biologic activity related to study treatment will be performed for all subjects who receive at least one dose of any study drug. These ana lyses will not be performed by Advanced Clinical but will be conducted by the Sponsor’s designated pharmacodynamics experts/vendors and the results will be summarized separately . 2.3 Trial Design and Trial Procedures This is an open- label multicenter 2-arm study to eval uate the saf ety and preliminary efficacy of the additi on of a vaccine the rapy to 1 or 2 che ckpoint i nhibitors for NS CLC: • Arm A: mRN A Va ccine [BI 1361849 (f ormerly CV9202)] + anti-PD-L1 [du rvalumab]. • Arm B: mRN A Va ccine [BI 1361849] + anti- PD-L1 [du rvalumab ] + anti-CTLA-4 [treme limumab]. For each arm of t he study, there is a dose evalu ation phas e in which t he Re commended Combinati on Dose (RC D) is determined according to a standard 3 + 3 de sign. For A rm A, the RCD of BI 1361849 + durval umab is determined. The st arting dose of durval umab is 1500 mg with pos sible de-escalat ion to 750 mg; the dose fo r BI 1361849 rem ains cons tant at 12 x 80 ug. The 1500 mg Q4W dosi ng of durval umab is recommended only for subje cts with > 30kg body weight. Sub jects w ith a body we ight ≤ 30 kg are not e ligible for enroll ment in the current st udy. If a s ubject’s body weight drops to ≤ 30 kg wh ile on the study, the subje ct will receive wei ght- based dosi ng equivalent to 20 mg/kg of durval umab as long as the body wei ght rem ains ≤ 30 kg. When the wei ght improv es to >30 kg, the s ubject may return to fixed dosi ng of durval umab 1500 mg. For Arm B, the RCD of BI 1361849 + durv alumab from Arm A w ith the addition of tremelimum ab 75 mg is ev aluated. There is no dose escalat ion/de-esc alation for Arm B; if there is unaccepta ble toxici ty in Arm B, the arm w ill be dis continued as presented in Figure 1. The dose eval uation phase is followed by an expan sion phase, in which the cohort at the R CD for each arm is expande d to 20 subje cts (i nclusive of the subje cts from the dose ev aluation cohort i.e., 14 subjects will be added to the 6 treated at the RCD in the dose evaluation phase for each arm (A and B ). Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 4 of 27 2.4 Treatments and Assignment to Treatments En rollment w ill sta rt in the A rm A dose eval uation cohorts in a sequen tial fas hion. A fter the RCD is determined for Arm A, enrollment w ill start for the dose evaluation of Arm B and for the Arm A Expansi on Cohor t, whereby no randomiz ations w ill be pe rformed. Any new subje ct will be assi gned to the A rm B dose ev aluation cohort, unle ss no slot is availa ble, in which ca se the subje ct will be assi gned to the Arm A Ex pansion C ohort. Afte r the dose eval uation and safe ty review for A rm B is comple te, enrollment in to the A rm B Ex pansi on Co hort Group w ill be prioritized ove r Arm A. Figure 1: Enrollment Schema for Arms A and B * A fter the dose evaluation and safety review for Arm B is complete, enrollment into the Arm B Expansion Cohort will be prioritized over Arm A. 2.5 Determination of Sample Size The dose evaluation pha se will utilize a standard 3 + 3 de sign for Arms A and B, which w ill result in the enrollment of 6 to 18 subjec ts. In the expan sion pha se, 20 subje cts per arm are thought to provide suffic ient dat a to adequat ely identify essential safety and pre liminary efficacy signals. Therefore, up to 14 additional s ubjects will be added to the 6 s ubjects treated at the RCD in each arm (A and B). The sample size n=20 for the expansion phase is deemed to provide suffi cient pre cision for t he estimation of inci dence of adve rse events as estimated by the Cloppe r Pearson confi dence intervals (CI) for i ncidence of a dverse events. Overall, approximately 56 subjec ts will be enrolled in up to 8 sites in the US. The enrollment period will take approximately 24 months. Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 5 of 27 3.0 GENERAL ANALYSIS CONVENTIONS 3.1 General Conventions Data collected in this study will be documented using summary tables and subject data listings. Continuous endpoints will be summarized using descriptive statistics, e.g. number of subjects , mean, median, standard deviation, minimum and maximum values. The mean and median will be reported to 1 decimal place more than the level of precision of the data being reported and the SD will be reported to 2 decimal places more than the level of precision of the data being reported, unless otherwise noted. Categorical endpoints will be presented in frequency tables with number and percentage of observations for each level. All efficacy analyses will be presented separately for Arm A and Arm B. Tumor Response will be summarized and analyzed descriptively. A 95% Confidence Interval based on the binomial distribution will be constructed for the estimated PFS rate and ORR at 8 and 24 weeks and DCR. The number and percentage of subjects who died or had a confirmed progression, who survived without a confirmed progression, and who were lost to follow up (unknown survival and/or progression status) will be summarized. PFS rate at 8 and 24 weeks and the corresponding 95% CIs will be calculated based on Kaplan-Meier product limit estimates and will be displayed along with the corresponding number of subjects at risk. (Note – this may include OS as well) PFS and OS will be summarized using the 25th percentile, Median, and 75th percentile as well as the minimum and maximum survival time, calculated by Kaplan -Meier method, and will be displayed graphically. The ORR, DCR and DOR will be summarized descriptively and a 95% Confi dence Interv al based on binom ial distri bution w ill be presented. 3.2 Trial Periods The overall length of study per subject is up to 15 months; comprising of 12 months for treatment and three months for on study follow-up. The screening period is up to one month prior to first drug administration. Study drug w ill be admin istered ove r 12 cycl es with a cy cle le ngth of 28 days. The st udy dr ugs used in this study w ill be administ ered per cycle as shown in Appendix B. The overall length of the study is 39 months. However the length of the surviaval post study follow- up is up to five years from initiation of treatment. Sponsor: Ludwig Institute for Cancer Research.

Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 6 of 27 3.3 Visit Windows The allowed visit windows are defined as shown in Table 1. Visits will be analyzed per the eCRF collection. Table 1: Visit Windows Visit Name Allowed window (Cycle Day ± Window) Screening Day -28 to -1 Weeks 2, 3, 5, 6, 7, 9, 11, 13, 14 +/- 2 day Weeks 17, 21, 25, 27 +/- 3 day Week 28 +/- 2 day Weeks 29, 33, 37, 39, 41, 45 +/- 3 day Week 46 +/- 2 day On study Follow -up 28 (+/ -4) days post last dose 42 (+/-4) days post last dose 91 (+/ -7) days post last dose Post Study Follow -up Every 6 months until 5 years Note: Visit days are defined from baseline (Cycle 1, Day 1 visit). Vacinations may have to be dela yed to all ow for recovery of A Es prior to continu ation of va ccine treatment. After the delayed v accination all subsequent vacci nations must be continued within the schedule relat ive to the fi rst vaccinati on pre sented in the flow chart. 3.4 Baseline Unless specified otherwise, baseline measurements will be the most recent value prior to receiving the first dose of study medication. 4.0 ANALYSIS POPULATIONS 4.1 Intent -to-Treat Population The Intent-To- Treat (ITT) Population is defi ned as all subjects who r eceive at least one dose of any of the st udy drugs. This population will be used for safety and efficacy analyses. 4.2 Per-Protocol Population (Clinical Efficacy) The Per-Protocol (PP) Population (Clinical Efficacy) is defined as all subjec ts who received at least 75% of t he sc heduled doses of durval umab and treme limum ab and at least 4 of the 5 BI 13618 49 vacci nations over the first 2 cycl es, as well as, respe ctive dise ase assessments (irRECIST or RECIST, clinical progression or death), w ithout major protocol viol ations. Efficacy analyses will also be presented for PP population. Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 7 of 27 4.3 Per-Protocol Population (DLT Assessments) The Per-Protocol (PP) Population (DLT Assessments) is de fined as follows: • All subjects in the dose escalation who experience a DLT at any time during the DLT Evaluation Period (as defined in Section 3.1.9 of the protocol). • All subjects with no DLT who receive at least 75% of the scheduled doses of durvalumab and tremelimumab and at least 4 of the 5 BI 1361849 vaccinations over the first 2 cycles as well as respective safety assessments without major protocol violations during the DLT Evaluation Period . 5.0 SUBJECT DISPOSITION Subject disposition will be summ arized overall for a ll subject s who entered the study in a disposition table (i.e., signed the informed consent for the study). In addition, the number of subject s in each popula tion (ITT and PP) and subjects that were removed from a popula tion w ill be su mmar ized. The number and propor tion of subject s who complete the study, as well as those who discontinue the study w ill be summar ized along w ith the r eason for discontinua tion. Primary reason for treatment termination include the following criteria: • Withdrawal of consent for furthe r treat ment • Pregn ancy or intent to become pregnant • DLT at any time • Progre ssive dis ease requiri ng alt ernat ive syst emic tre atment • Significant protocol violat ion or noncomplia nce th at, in the opinion of t he Inve stigator or Sponsor, warrants withdrawal • Development of i ntercurr ent, non- cancer-re lated illne sses or complicat ions that prevent either continuat ion of t herapy or regular follow-up • Best medic al interest of the subject (at the dis cretion of t he Inve stigator) Primary reason for study discontinuation include the following criteria: • Best medic al interest of the subject at the disc retion of t he Inve stigator • In itiation of al ternat ive anti cance r therapy (mar keted or investigational) • Withdrawal of consent for all follow-up • Lost to follow-up • Death Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC.

CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 8 of 27 Subject disposition data will also be presented in data listings. 6.0 DEMOGRAPHIC AND BASELINE CHARACTERISTICS 6.1 Demographic Characteristics Demogra phic and base line char acteristics at study entry ( Screening/baseline visit which will occur up to one month before start of treatment) will be summar ized for the ITT and PP Popula tions. Demographic and baseline variables to be summarized include: • Continuous variables - Age (years) at time of consent - Height (cm) at screening - Weight (kg) at screening - Body mass index (BMI) (kg/m2) at screening • Categorical variables - Gender - Race - Ethnicity Demographic and baseline characteristics by subject will be presented in a table and data listings. 6.2 Medical History Medical History was collected at the Screening/Baseline visit. Medi cal history, including any ongoing and signif icant conditions or dis eases that stopped at or prior to informed consent, must be elicited from e ach subject during sc reening. The medi cal history shall include a complete review o f systems, past medical and su rgical histo ries, and any a llergies. The frequency count and percentage of subjects experiencing any medical conditions will be tabulated by system organ classifications (SOC) and preferred term (PT). If a preferred term or system organ class was reported more than once for a subject , the subject would only be counted once in the incide nce for that preferred term or system organ class. Medical history data will also be listed for Safety Population. Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 9 of 27 6.3 Prior and Concomitant Medication s and Procedures Prior medications and procedures include any medic ation or non-drug therapy or procedure taken or performed w ithin 30 days prior to sc reening a nd before the fir st dose of study drug. Prior medi cations w ill be coded using the Wor ld Health Organiza tion Drug Dic tionary [10]. Concomitant medications are all medi cations, other than the study drug, taken on or af ter the fir st day of st udy drug dos e through 30 days after end of treatment. Concomitant medi cations and procedures will be coded using the Wor ld Health Organiza tion Drug Dic tionary. Medications that started before the first dose of study drug and were ongoing on the date of the first dose will be considered concomitant medications. The number and proportion of the subjects who took each medication, or had qualifying concomitant procedures, will be tabulated by the ATC-2 level and preferred name for concomitant medications. A subject will only be counted once within each ATC-2 code and within each preferred name. Prior and concomitant medications will also be listed for ITT. 7.0 EFFICACY ANALYSIS Tumor Response w ill be summ arized and anal yzed de scriptively for e ach a rm for the ITT and P P populations . 7.1 R esponse Rate • Responses at Week 8 and Week 24 will be presented . The Number and Percent of subjects in each category (Complete Response, Partial Response, Stable Disease, Progressive Disease or Not Evaluable) will be presented . • Best Overall Response (BOR) will also be presented. In the determination of BOR, any subsequent scans on or after the date of a clinical progressi on will be ignored. 7.2 Objective Response Rate (ORR) • Objective Response Rate (ORR) is defined as the percentage of subjects meeting criteria of Complete Response (CR) or Partial Response (PR). • To be assigned a status of a CR or PR , changes in tumor measurements m ust be confirmed by repeat asse ssments that must be perf ormed at l east 4 weeks after the cr iteria for response are first met. If a response is not confirmed, it is considered stable disease. • If a subject had clinical progression (CP) , any recorded response (CR or PR) after the date of CP will be considered invalid. Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 10 of 27 • Subjects who drop out prior to having a response assessment (including clinical progression) for ORR will be consi dered as non-responders and they w ill be included in the denominator when ca lcula ting the propor tion. • A 95% Confi dence Interval based on binom ial distribution w ill be constr ucted for t he estimated ORR at 8 and 24 we eks. 7.3 Disease Control Rate (DCR) • Disease Control Rate (DCR) is defined as the percent of subjects with Stable Disease (SD), Partial Response (PR) and Complete Response (CR). • Subjects who drop out prior to mee ting the responde r criteria for DCR w ill be consi dered as non-responde rs. • A 95% Confi dence Interval based on binom ial distribution w ill be constr ucted for t he estimated DCR at 8 and 24 weeks. 7.4 P rogression -free Survival (PFS) • PFS is the inte rval bet ween t he date of first dose to the date of e arliest determinat ion of Progre ssive D isease (PD) (including clinical progression), or to the da te of de ath, if PD does not occur. • Subjects w ithout documentat ion of progression at the time of t he anal ysis w ill be censored at the date of last res ponse assessment. • Subjects with no tumor res ponse ass essment will be censored at the st art date of t he treatment. • Subjects who disc ontinued tre atment or w ithdrew from the st udy for reas ons other than documented PD (including clinical progression) or d eath will be censored at the date of last response ass essment pri or to discontinuati on or withdra wal. Descriptive analyses of PFS, at Week 8, Week 24 and Overall, will include the following: • Number and percentage of subjects that died or had a progressive disease or clinical progression, • Number and percentage of subjects censored: o Number and percentage of subjects lost to follow up (unknown survival and/or progression status) o Number and percent of subjects survived without progression o Number and percentage of subjects missing tumor response assessment. Sponsor: Ludwig Institute for Cancer Research.

Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 11 of 27 o Number and percent of subjects with initiation of alternative therapy. • The 25th percentile, Median and 75th percentileand the maximum and minimum PFS at Week 8 , Week 24 and Overall and the corr esponding 95% CIs will be displayed. • PFS rate at Week 8,Week 24 and Overall and the corr esponding 95% CIs will be displayed. 7.5 Overall Survival ( OS) • The i nterval between t he date of first dose until the date of death or t he date of last follow- up will be calculated . • The OS rate , including the 95% CI, will be presented . Duration of follow- up will also be presented. OS will also be update d at yearly interv als during the Post Study Follow- up an d can be provided as addendums to the fi nal report. • Subjects who are still alive w ill be censored on the date of last follow-up. Eve ry effort will be made to follow subjec ts for OS after they dis continue the st udy. • OS will be summarized usi ng the 25th percenti le, Media n, and 75th percentile as well as the minimum and m aximum survival time, calc ulated by Kapla n-Meier method, and w ill be dis played gra phicall y 7.6 Duration of Response (DoR) • Dura tion of re sponse (DO R) will be calculated interval bet ween t he date o f earliest determinat ion of CR or PR to the da te of first PD, clinical progression or death whatever occurs first . Patients who had a respon se and did not lose it subsequently and were alive will be censored at the time of last available tumor assessment. • DOR in weeks = (Earlier of Date of PD or death – date of first response + 1) / 7 • The median DOR with 2 -sided 95% confidence intervals w ill be presented for subjects who have a confirmed CR or PR at 8 and 24 weeks. 7.7 Analysis of Time to Event (TTE) Variables Progression Free Survival (PFS), OS and DoR will be analyzed using the Kaplan- Meier product limit estimates along with the corresponding number of subjects at risk . The Duration of follow- up (in months) will be analyzed using the Reverse Kaplan- Meier product limit estimation method. For all time to event analys es the number of events, number censored, 25th percentile, median, 75th percentile, corresponding 95% CIs as well as the minimum and maximum survival time will be presented in tables. In addition, survival curves will be presented graphically. Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 12 of 27 8.0 TRIAL DRUG EXPOSURE AND COMPLIANCE Extent of treatment exposure to durvalumab, will be assessed for each of the two study arms. Durvalumab exposure (Arms A and B) will be summarized by the following parameters: • Duration of durvalumab treatment (week), calculated as: (Last dose date – first dose date + 1) / 7 • Number of durvalumab doses • Total durvalumab dose (mg), defined as the sum of the actual doses (mg) administered • Number of subject s with durvalumab dose de- escalated (only in Arm A) Tremelimumab exposure (Arm B) will be summarized by the following parameters: • Duration of Tremelimumab treatment (week), calculated as: (Last dose date – first dose date + 1) / 7 • Number of tremelimumab doses • Total Tremelimumab dose (mg), defined as the sum of the actual doses (mg) administered BI 1361849 exposure (Arms A and B) will be summarized by the following parameters: • Duration of BI 1361849 treatment (week), calculated as: (Last dose date – first dose date + 1) / 7 • Number of doses • Total BI 1361849 dose ( µg), defined as the sum of the actual doses (mg) administered 9.0 SAFETY ANALYSIS 9.1 Adverse Events Adverse events (AEs) will be coded using MedDRA v20 or later and will be classified by System Organ Class (SOC) and preferred term (PT). Severity of AEs will be assessed according to CTCAE (v4.03).

Prior AE s are those occurring after subject sign off the informed consent and before the administration of the first dose of study treatment. Drug -related ness to the study drugs is captured in five categories as follows: • Definitely related (The AE is c learly related to the inve stigational agent) Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 13 of 27 • Probably rel ated (The AE is likely r elated to the inve stigational agent) • Possibly related (The AE may be r elated to the inve stigational agent ) • Unlikely related (T he AE is doubtf ully r elated to the investi gational agent) • Unrelated (The AE is clearly not r elated to the inve stigational TEAEs are considered as related to the study drug if enterd as definitely , probably or possibly related to study drug as assessed by the Investigator. If a subject has multiple occurre nces of the same system organ class (SOC) or preferred term, then only the most sever e event will be summar ized in the tables for that SO C and pre ferred term. 9.1.1 Treatment -Emergent Adverse Events Treatment -emergent adverse events (TEAEs) are defined as any AEs that started aft er dosi ng or wors ened in seve rity after dosi ng. An overall AE summary for number of subjects will be presented for the following categories: • Any Adverse Event • Any TEAE • Any Deaths within the AE reporting period • Any study drug- related adverse event (TRAE) (i.e., related to durval umab, tremelimumab and/or vaccine) o Any durvalumab related TR AE o Any tremelimumab related T RAE o Any vaccine related T RAE • Any TEAE ≥ CTCAE Grade 3 • Any TR AE ≥ CTCAE Grade 3 • Any treatment -emergent Serious AEs (SAE s) • Any Serious TRAEs • Any TEAE leading to treatment (durval umab, tremelimumab, and/or vaccine) discontinuation o Related to durvalumab o Related to tremelimumab o Related to vaccine Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC.

CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 14 of 27 The following events will be tabulated by SOC and PT. Summaries will be sorted by decreasing frequency of PT within SOC which is sorted alphabetically. • TEAEs by SOC and PT • TEAEs b y PT frequency only • Treatment -emergent SAEs • Treatment -related AEs o Durvaluma b related o Tremelimumab related o Vaccine related • Treatment -related SAEs • TEAEs leading to treatment ( durval umab, tremelimumab, and/ or vaccine ) discontinuation o Durvalumab discontinuation o Tremelimumab discontinuation o Vaccine discontinuation • DLTs A summary of TEAEs by SOC, PT and maximum severity, sorted by decreasing frequency of PT within SOC which is sorted alphabetically, will also be provided for: • TEAEs • TEAEs ≥ CTCAE Grade 3 • Treatment related AEs o Durvalumab related o Tremelimumab related o Vaccine related • TRAEs ≥ CTCAE Grade 3 In tabulation by severity grade, • For a given SOC, only the most severe SOC for each subject will be included. • For a given PT, only the most severe PT for each subject will be included. The following listing s will be provided: • All AEs (flag TEAE) Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 15 of 27 • DLTs • SAEs • TEAEs leading to study treatment discontinuation • TEAEs leading to study drug (durval umab or tremelimumab) dose modifications • Death s in the AE reporting period 9.1.2 Adverse Events of Special Interest (AESIs) AESIs will not be separately presented in this analysis. 9.2 Clinical Laboratory Data Clinical Laboratory data (chemistry, hematology, coagulation, and urinalysis) will be collected as specified in the study flowchart in section 3.2 of the protocol.

The Clinical lab tests are summarized in Table 3 below. Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC.

CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 16 of 27 Table 3. Clinical Laboratory Tests Clinical Chemistry Hematology Coagulation Urinalysis Alanine transaminase (ALT) Album in Alkaline phosphata se Aspartate aminotr ansfera se (AST) Blood Urea Nitrogen (BUN) Calcium Carbon Dioxide (CO 2) Chloride Creatinine Glucose Lactate Dehydrogenase (LDH) Magnesium Potassium Sodium Total bilir ubin Total Protein Lipase Amylase Free T3 Free T4 TSH Hematoc rit Hemoglobin Mean cellular volum e (MCV) Mean corpuscular hemoglobi n (MCH) Mean corpuscular hemoglobin concentr ation (MCHC) Mean Platelet Volume (MPV) Platelet count Red blood cell (RBC) count RBC Distribution Width (RDW) White blood cell (WBC) count Differential (% and absolute count for each) including: Basophils Eosinophils Lymphocytes Monocytes Neutrophils Activated pa rtial thrombopla stin time (aPTT) Prothrombin time (PT) INR Albumin Bilirubin Blood/Hemoglobin Glucose Ketones White Blood Cells Nitrite pH Protein Red Blood Cells Specific gravity Urobilinogen Color Turbidity The clinical lab data will be presented in tables and listings as follows: • All laboratory results will be presented in the listings. Sponsor: Ludwig Institute for Cancer Research.

Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 17 of 27 • Abnormal laboratory values with their clinical significance status will be presented in a listing. • Urinalysis data, differentials (%), and RDW will be presented in listings only • Magnesium, MCV, MCH, MCHC, MPV, differentials (absolute count), aPTT, PT, and INR will be presented in listings and summary tables but NOT in shift tables. • The remaining analytes will be presented i n listings, summary tables and shift tables The following summaries will be presented: • Overall values by time point utilizing continuous descriptive statistics for each arm and cohort. • For categorical parameters, the n and percentage will be displayed for each arm and cohort. • Frequencies and percentages for the shifts in these categories (i.e., low to normal, low to high, high to low, etc.) from baseline to each post- treatment assessment time point. For each continuous laboratory parameter, r esults will be categorized as low, normal, or high based on the laboratory normal ranges. Frequencies and percentages will be presented by Arm A and Arm B and overall Safety population for the shifts in these categories from baseline to selected post -treatment assessment time points (e.g., low to normal, low to high, high to low). Percentages for the shift tables will be calculated based on the number of subjects who had results for both baseline and the corresponding post- treatment assessment time point. • For continuous parameters, descriptive statistics will be presented for the changes from baseline to each selected post -treatment assessment time point. 9.3 Vital Signs and Weight Subjects will be monitored before, duri ng and aft er tremelimum ab and durval umab infusi on w ith assessment of vital signs as presented in section 6.5 of the protocol. • Descri ptive stat istics will be presented each timepoint. • Descriptive statistics will also be presented fo r the changes in vi tal signs from bas eline to each post-tr eatment asse ssment time point. 9.4 Electrocardiograms 12-Lead ECG will be completed at screening/baseline visit, Day 1 of cycles 1, 2, 4, 6, 8, 10 and 12 and the +28 day post last treatment visit. • Baseline and post -baseline assessments will be classified as normal, abnormal – not clinically significant, or abnormal – clinically significant. The baseline value will be the pre-dose assessment for each formulation, and the post -baseline value will be t he worst post-baseline assessment for each formulation. Sponsor: Ludwig Institute for Cancer Research.

Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 18 of 27 • ECG data will be presented in a data listing. 9.5 ECOG Performance ECOG perfomance will be evaluated on a 6 -point scale from grade 0 -5. ECOG PS a ssessments will be summ arized overa ll by time point utilizing des criptive s tatistics. ECOG performance status will be taken except on Cycle 2 Day 8, Cycle 4 Day 8, Cycle 7 Day 22, Cycle 12 Day 8, and the post study follow-up. • ECOG PS assessments will also by summarized as categorical variable by timepoint with the n and percentage being displayed for e ach arm or indication- specific cohort. 9.6 Physical Examinations A full phy sical exa mination will be perf ormed at screening/ baseline. There will also be targ eted physical examinations at other time points except on Cycle 2 Day 8, Cycle 4 Day 8, Cycle 7 Day 22, Cycle 12 Day 8, and the post study follow-up. • Continuous and categorical variables at baseline and various timepoints will be summarized using descriptive statistics. • Clinically significant changes from baseline to postbaseline time points will be summarized using descriptive statistics. 9.7 Pregnancy Testing Serum p regnancy testing (when applica ble) is required pri or to dosi ng up to 1 week before start of treatment. Pregnancy testing is also done on Day 1 of Cycles 1, 3, 5, 7, 9, 11, and On Study Follow-up Days +28 and +91 Days. Also, all s ubjects of childbe aring potential who w ithdraw from st udy must have a serum pregnancy test done at the End of Study vis it, unle ss it w as done within 7 days prior to the End of Study Visi t. • T he results of pregnancy testing (Yes/No) will be summarized for every applicable timepoint with the n and percentage being displayed for each arm or cohort. 10.0 EXPLORATORY ANALYSIS Pharmacodynamics analyses on biologic activity related to study treatment will be performed for all subjects who receive at least one dose of any study drug. These analyses will not be performed by Advanced Clinical but will be conducted by the Sponsor’s designated pharmacodynamics experts/vendors and the results will be summarized in a report which will be included as an appendix to the Clinical Study Report. Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 19 of 27 Per Protocol Amendment 9.1, some of the biological assessment tests will no longer be applicable. Hence, some blood sample collections for immune monitoring were discontinued as indicated in the study flow chart. 11.0 STATISTICAL/ANALYTICAL ISSUES 11.1 Handling of Dropouts or Missing Data In the dose evaluation phase, subjects who are not fully evaluable for DLT per section 4.1.2 of the protocol may be replaced. Subjects who are not fully evaluable for the PFS rate and ORR may be replaced. All ava ilable da ta will be pre sented on the da ta listin gs as collected. Algorithms for imputing partial or missing dates related to AEs and prior/concomitant medications are shown below in Table 4 . Table 4 Imputation Rules for Partially Missing Dates Variable Missing Day Missing Day, Month Missing Day, Month, Year Date of Last Therapy /Date of In itial Diagnos is Assign 1 Assign January 1 if prior to date of inform ed co nsent, otherwise use date of inform ed consent Missing (do not impu te) Adverse Event/Medication Start Date Assign first day of month unless it is t he month of first dose of study medication. Otherwis e, assign d ate of first dose of study med ication. Assign January 1 unles s the year is year of firs t dose of study medicatio n Otherwis e, assign d ate of first dose of stud y medication. Assign date first dose of study medication. Adverse Event/Medication End Date Assign the last day of the month or end of stud y date, whichev er is earlier. Assign December 31 or end of study d ate, whichev er is earlier. If ongoing, end date is missi ng.

Otherwis e, ass ign end of study d ate. 11.2 Safety Data Handling Adver se event toxicity grade w ill be cla ssified using NCI-CTCAE Version 4.03 cr iteria (Grade 1 – Grade 5). If a subject has multiple occurre nces of the same system organ class (SOC) or Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 20 of 27 preferred term, then only the most sever e event will be summar ized in the tables for that SO C and preferred term. Ad verse events of ≥ Grade 3 w ill also be summ arized. A missing toxicity grade will not be impute d. The AE analysis w ill be repeated for SA Es and A Es leading to dose re duction or discontinua tion. 11.3 Coding Dictionaries Medi cal history, A Es, and concurrent procedure s will be coded using Medi cal Dic tionary f or Regula tory Activities (MedDRA V2 0). Prior and conco mitant medi cations w ill be coded using the Wor ld Health Organiza tion Drug Dic tionary (WhoDrugDDEB2_201703). 11.4 Pooling of Centers in Multi- Center Trials There will be no pooling of centers. 11.5 Multiple Comparisons/Multiplicity No adjustments for multiple comparisons or multiplicity will be made. 11.6 Examination of Subgroups No subgroup analysis will be performed in this study. 11.7 Interim Analysis and Data Monitoring Interim Safety R eviews w ill be pe rformed to assess DL Ts in the dose ev aluation cohorts for determinat ion of RC D. Interim an alyses may be performed to anal yze the 8- and 24- week endpoin ts. 12.0 QUALITY CONTROL All data displays and analyses will adhere to the International Conference on Harmonisation (ICH) Harmonized Tripartite Guideline: Structure and Content of Clinical Study Reports (ICH Topic E3). The clinical monitor will verify that the clinical study is conducted and data are generated, documented (recorded), and reported in compliance with the protocol, GCPs, and any applicable regulatory requirements. Ludwig Institute for Cancer Research (LICR) or their designee will review all tables, listings, and figures prior to final database lock. Final SAS datasets, programs and outputs will be transferred to LICR at project completion. Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 21 of 27 13.0 TABLES AND LISTING CONVENTIONS Mock -ups for statistical tables and listings will be provided. Final formats for the statistical tables and listings may deviate from these mock -ups upon agreement with LICR . Footnotes will be used as needed to clarify the information that is presented in the tables and listings. Unless otherwise requested by LICR , the term ‘subject’ wi ll be used in all tables and listings, in accordance with CDISC standards. The numbering of tables, figures and listings will be in the following order: Table 14.x.y.z, Figures 15.x.y.z, Listings 16.x.y.z. The general layout of tables and listings will be as follows: All tables and listings will use landscape orientation. Margins will be at least 2.0 cm at the top and bottom and at least 0.8 cm on the left and right, excluding headers and footers, in accordance Ludwig Institute for Cancer Research . Page x of y Protocol: LUD2014-012-VAC Run Date: DDMMMYY - HH:MM Clinical Study Report Listing 16.2_x (or Table 14.x_x) <Title> <Population> Col 1 Col 2 Col 3 etc. ____________________________________________________________________________ _____________ __________________________________________________________________________________________ _ <Any footnotes> File Name: <pathname for SAS program> Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 22 of 27 with electronic Common Technical Document ( eCTD) guidelines. Font will be Courier New, unless otherwise specified, with an 8-point font size in most cases.

Page numbering will be sequential within each table, listing, and figure. Column headers should be in initial capital letters. Units for numeric data will be included when appropriate. Tables and data listings will be created from different SAS programs. A single program may produce multiple tables or multiple data listings from the same dataset (e.g., all clinical chemistry data listings may be generated by a single program). 14.0 REFERENCES 1.

Cancer Rese arch Institute. Cancerresearcho rg http://www cancer researchorg/cancer - immunotherapy/impac ting- all-cancers /lung- cancer Accessed 03 M arch 2016. 2.

Eisenhaue r EA, The rasse P, Bogaerts J, Schw artz LH, Sargent D, Ford R, et al. Ne w response evaluation cri teria in s olid tumours: revis ed RECI ST guide line ( version 1.1 ). Eur J Ca ncer.

2009;4 5(2): 228-47. 3.

Bohnsack O, Ludajic K, Hoo s A. Ad aptat ion of the i mmun e-related response cri teria: irRECI ST. Ann als of O ncology ESMO 2014 Post er. 2014;25 (Suppl 4):iv361-iv72.

4. Co mmon T erminology Cr iteria for Adver se Events (CTCA E) Version 5.0. Publish ed: November 27, 2017.

Na tional Cancer Ins titute, N ational I nstitutes of Hea lth. U.S. Department of Health and Human S ervices.

5. International Conference on Harmonization (ICH, 1996) E6 Guideline for Good Clinical Practice 6. International Conference on Harmonization (ICH, 1997) E8 Guideline: General Considerations for Clinical Trials 7. Medi cal Dic tionary f or Regula tory Activities (MedDRA), Version 20. 8. World Health Organization (WHO) Drug Dictionary (WhoDrugDDEB2_201703). 15.0 RECORD RETENTION Records related to the activities listed in this plan will be retained according to AC SOP AD-005. Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 23 of 27 16.0 CHANGE HISTORY Version Date Description of Changes 0.1 07OCT2019 First draft version 0.2 10OCT2019 Current version. First version delivered to LICR by AC 0.3 01NOV 2019 Revised by AC , Sent to LICR 0.4 14NOV2019 Version 0.3 Revised by AC, Version 0.4 s ent to LICR 1.0 12DEC2019 Final version 1.1 20AUG2020 1. Section 7 revised to better expound on BOR, PFS, DOR 2. Clinical Laboratory Data (Section 9.2 revised to reflect intended analysis 3. Section 9.3 (Urinalysis) deleted as it is covered in Section 9.2 4. Section 3.4 inserted to explain baseline measurements 2.0 15SEP2020 Alignment with protocol Amendment version 9.1 (01SEP20 20) Revisions to Section 7 as updated in version 1.1 and to align with mock shells, 2.1 02DEC2020 Phase 2 tumor types in Section 9.2 changed to Arm A and Arm B (irRECIST or RECIST, clinical progression or death) inserted in the definition of per protocol analysis population to emphasize disease assessments methods 3.0 09DEC2020 Final version Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 24 of 27 17.0 APPENDICES 17.1 Appendix A: Study Flowchart Cycle 5 Cycle 6 Cycle Day 1 8 (±2) 15 (±2) 1 (±2) 8 (±2) 15 (±2) 1 ( ±2) 15 (±2) 1 (±2) 8 (±2) 1 (±3) 1 (±3) Cumulative Study DayUp to 1 month before Tx s ta rt1 8 15 29 36 43 57 71 85 92 113 141 Treatment Durvalumab (Arms A and B) X X X X X X Tremelimumab (Arm B) X X X X BI 1361849 - 6 components (Arms A and B) X X X X X X X X X X Disease Staging (date/stage at 1st diagnosis and at study entry) X Disease Assessment by ir REC IST/REC ISTa X X X Study Procedures and Examinations Eligibility Assessment and Informed Consent (IC)hX Demogr aphics (incl. DoB; sex ; height; r ace; ethnicity) X M edical histor y X P hy sical Ex am (incl.

weight and EC OG P er f Status)bX X X X X X X X X X X 12-Lead ECGaX X X X X Vital Signs (T, HR, BP , RR)eX X X X X X X X X X X Concomitant Medication(s)/ Procedure(s) X X X X X X X X X X X X X Adverse Events (starting or worsening after IC)f X X X X X X X X X X X X X Blood Hematology (complete blood count, differ ential, platelets)aX X X X X X X X X X X X X Chemistry (glucose, BUN, creat., Na, K, Cl, CO 2, Ca, Mg, protein, albumin, Tbili., AST, ALT, ALP, LDH)aX X X X X X X X X X X C hemistry cont. (Free T3, Free T4, TSH)aX X X X X X X X X X X C hemistr y cont. (Amy lase and lipase)aX X X X X X X X X X X Ur inaly sisa,cX X X X X X X C oagulation par ameter sa,dX Serum pregnancy test (Urine test only on Day 1)a Up to 1 week before Tx s ta rt X X X Specimens for Other Peripheral Blood Assays Blood for exosomal profiling aX X Blood for PaxGene RNA and DNAa,k Note: Discontinued per Amendment 9.1 X XJXJ Blood (P BM C and plasma) for flow cy tometr y and biological assay sa,i,k Note: Discontinued per Amendment 9.1 X XJXJ Blood for humoral responses and other bi oma rkersa,k Note: Discontinued per Amendment 9.1 X XJXJ Biopsy or FFPE slides for tumor microenvironmentgX Ov er all Sur v iv al Progres s i on Free Survi va lSpecimens for Routine Laboratory Procedures5 Study Week 1 11 X (one week after 3rd or 5th BI 1361849 injection) Cycle 4 14J Tumor and Disease Assessments21 17 13Treatment (each cycle = 4 weeks) Long Term Follow-upTumor BiopsyLUD2014-012-VAC Study FlowchartScreening / Baseline Cycle 2 Cycle 3 2 3 7Cycle 1 6J 9 Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 25 of 27 Cycle 8 Cycle 9Cycle 11 Cycle Day 1 (±3) 15 (±3) 22(±2) 1 (±3) 1 (±3) 1 (±3) 15 (±3) 1 (±3) 1 ( ±3) 8 (±2) Cumulative Study Day 169 183 190 197 225 253 267 281 309 316 Treatment Durvalumab (Arms A and B) X X X X X X Tremelimumab (Arm B) BI 1361849 - 6 components (Arms A and B) X X X X Disease Staging (date/stage at 1st diagnosis and at study entry) Disease Assessment by ir REC IST/REC ISTa X X X Study Procedures and Examinations Eligibility Assessment and Informed Consent (IC)h Demogr aphics (incl. DoB; sex ; height; r ace; ethnicity) M edical histor y P hy sical Ex am (incl.

weight and EC OG P er f Status)bX X X X X X X X X X X 12-Lead ECGaX X X X Vital Signs (T, HR, BP , RR)eX X X X X X X X X X X Concomitant Medication(s)/ Procedure(s) X X X X X X X X X X X X X Adverse Events (starting or worsening after IC)f X X X X X X X X X X X X X Specimens for Routine Laboratory Procedures Blood Hematology (complete blood count, differ ential, platelets)aX X X X X X X X X X X X X Chemistry (glucose, BUN, creat., Na, K, Cl, CO 2, Ca, Mg, protein, albumin, Tbili., AST, ALT, ALP, LDH)aX X X X X X X X X X X C hemistry cont. (Free T3, Free T4, TSH)aX X X X X X X X X X X C hemistr y cont. (Amy lase and lipase)aX X X X X X X X X X X Ur inaly sisa,cX X X X X X X C oagulation par ameter sa,dX Serum pregnancy test (Urine test only on Day 1)aX X X X X Specimens for Other Peripheral Blood Assays Blood for exosomal profiling aX X Blood for PaxGene RNA and DNAa,k Note: Discontinued per Amendment 9.1 XJXJX Blood (P BM C and plasma) for flow cy tometr y and biological assay sa,i,k Note: Discontinued per Amendment 9.1 Blood for humoral responses and other bi oma rkersa,k Note: Discontinued per Amendment 9.1 XJXJX Biopsy or FFPE slides for tumor microenvironmentgoptional post- treatment Ov er all Sur v iv al Progres s i on Free Survi va lLUD2014-012-VAC Study Flowchart (cont.) X XTreatment (each cycle = 4 weeks) Cycle 12 Long Term Follow-upTumor BiopsyEvery 6 months until 5 years post initiation of treatment33 37 39 41 Last Study Drug Dose + 91 ( ±7) days End of Study45Cycle 7 28J Tumor and Disease Assessments46Last Study Drug Dose +42 (±4) daysLast Study Drug Dose +28 ( ±4) days29 27Post Study Follow- upOn Study Follow-upf,K Cycle 10 Q8W starting 8 weeks after last disease assessment25 Study Week Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC.

CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 26 of 27 Flowchart Footnotes a - pre-dose, when applicable. Note: Review of results for hematology, chemistry and pregnancy test (when applicable) is required prior to dosing. b - Full phy sical ex amination at baseline; tar geted phy sical ex amination at other time pointsc - Ur inaly sis per for med at Scr eening, Day 1, ev er y 4 weeks and as clinically indicated.d - Coagulation tests: prothrombin time, APTT and INR – only performed at Screening, first On Study Follow-up, and as clinically indicated. e - See Section 6.5 for assessment of vital signs on drug admin days. Note: when durvalumab vital assessments do not precede the BI 1361849 dose, p re-dos e vital as s es s ments mus t be done for BI 1361849. f - See section 7.1.5 for details r egar ding collection of AEs for 90 day s after last study dr ug administr ation g - Biopsy Samples: 1. A fresh core pre-treatment biopsy (minimum 3 cores from lung tissue or 4 cores from another site) obtained within 60 days of study start will be requested prior to study entry; archival sample may be used if a pre-treatment fresh biopsy is not feasible. NOTE: Per Amendment 9.1, biopsies should not be taken from a target lesion unless that is the only suitable lesion, in which case it should be of reasonable size (refer to Section 4.3.1.1). 2. If a fresh core pre-treatment biopsy was obtained, an on-treatment biopsy will be collected 1 week after the 3rd or 5th BI 1361849 treatment, if feasible; this on-treatment biopsy should only be collected if a pre-treatment fresh biopsy was obtained (so that paired biopsies may be examined). 3. Optional post-treatment core biopsies (minimum 3 cores from lung or 4 cores from another site) will be obtained at the time of tumor progression o r at the completion of tr eatment fr om subjects who consent to this pr ocedur e, and if clinically feasible. See Section 4.3.1.1 for details 4. If possible and as deter mined by Inv estigator , a minimum of 8 subjects in each ar m will hav e fr esh tissue sampling fr om which fine needle aspir ation (FNA) can be also obtained for immune profiling h - Standar d of C ar e pr ocedur es may be used for eligibility assessments pr ov ided they meet the cr iter ia specified in either the inclusion cr iter ia or flowchar t i - See Section 3.3 for instructions for blood draws for PBMC collection. Note: PBMC collection is discontinued per Amendment 9.1. J - Sampling at 7(±2) day s after pr eceding v accine administr ation. If v accine administr ation time points ar e delay ed, the P BM C & biological specimen collection time points should also be delay ed to maintain the 5-9 day time fr ame after pr eceding v accination. If the v accination is omitted, blood sampling should occur after the next subsequent vaccine administration. Sample for WBC (hematology ) must be av ailable for the day of the blood samples for immunomonitor ing. Note: Sample collection for PaxGene, PBMC, and humoral response is discontinued per Amendment 9.1 (See Section 4.3.1.2) Visits 6, 14, and 28 are no longer required per Amendment 9.1. k - If treatment is discontinued after <14 weeks of treatment, samples for PBMC, PaxGene and humoral response should be collected within 5-9 days after the last pr eceding v accination or as close as possible to this time fr ame if the 5-9 day window is not feasible. Note: Sample collection for PaxGene, PBMC, and humoral response is discontinued per Amendment 9.1 (See Section 4.3.1.2) Sponsor: Ludwig Institute for Cancer Research. Protocol Number: LUD2014- 012-VAC. CONFIDENTIAL SAP Version: 3.0 Effective Date: 09DEC2020 Page 27 of 27 17.2 Appendix B : Drug Administrative Schedules per Cycle

Ellingsen et al. Journal of Translational Medicine (2022) 20:419 https://doi.org/10.1186/s12967-022-03624-z RESEARCH Characterization of the T cell receptor repertoire and melanoma tumor microenvironment upon combined treatment with ipilimumab and hTERT vaccination Espen Basmo Ellingsen1,2,3* , Gergana Bounova4, Iliana Kerzeli5, Irantzu Anzar6, Donjete Simnica7, Elin Aamdal8, Tormod Guren8, Trevor Clancy6, Artur Mezheyeuski9, Else Marit Inderberg10, Sara M. Mangsbo5,11, Mascha Binder7, Eivind Hovig1,12 and Gustav Gaudernack3 Abstract Background: This clinical trial evaluated a novel telomerase-targeting therapeutic cancer vaccine, UV1, in combina- tion with ipilimumab, in patients with metastatic melanoma. Translational research was conducted on patient-derived blood and tissue samples with the goal of elucidating the effects of treatment on the T cell receptor repertoire and tumor microenvironment. Methods: The trial was an open-label, single-center phase I/IIa study. Eligible patients had unresectable meta- static melanoma. Patients received up to 9 UV1 vaccinations and four ipilimumab infusions. Clinical responses were assessed according to RECIST 1.1. Patients were followed up for progression-free survival (PFS) and overall survival (OS). Whole-exome and RNA sequencing, and multiplex immunofluorescence were performed on the biopsies. T cell receptor (TCR) sequencing was performed on the peripheral blood and tumor tissues. Results: Twelve patients were enrolled in the study. Vaccine-specific immune responses were detected in 91% of evaluable patients. Clinical responses were observed in four patients. The mPFS was 6.7 months, and the mOS was 66.3 months. There was no association between baseline tumor mutational burden, neoantigen load, IFN-γ gene signature, tumor-infiltrating lymphocytes, and response to therapy. Tumor telomerase expression was confirmed in all available biopsies. Vaccine-enriched TCR clones were detected in blood and biopsy, and an increase in the tumor IFN-γ gene signature was detected in clinically responding patients. Conclusion: Clinical responses were observed irrespective of established predictive biomarkers for checkpoint inhibitor efficacy, indicating an added benefit of the vaccine-induced T cells. The clinical and immunological read-out warrants further investigation of UV1 in combination with checkpoint inhibitors. Trial registration Clinicaltrials.gov identifier: NCT02275416. Registered October 27, 2014. https:// clini caltr ials. gov/ ct2/ show/ NCT02 275416? term= uv1& draw= 2& rank=6 Keywords: Cancer, Immunotherapy, Therapeutic Cancer Vaccine, Telomerase, hTERT, Melanoma, Ipilimumab © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.

The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.Background T cells that recognize tumor antigens are the foun - dation of current immunotherapy, on which the effi - cacy of checkpoint inhibitors (CPIs) relies. CPIs have Open AccessJournal of Translational Medicin e *Correspondence: [email protected] 1 Department of Tumor Biology, Institute for Cancer Research, The Norwegian Radium Hospital, Oslo, Norway Full list of author information is available at the end of the article Page 2 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 revolutionized the treatment of many cancers, most notably malignant melanomas [1, 2]. Although many patients experience deep and durable clinical responses to CPIs, unfortunately, many patients progress and require additional treatment. CPIs work by inhibit - ing axes that restrict T cell activation and proliferation, releasing spontaneously primed antitumor immune responses [3]. Therefore, predictive biomarkers for CPI efficacy are largely related to tumor immunogenicity, such as tumor mutational burden [4], neoantigen bur - den [5], PD-L1 positivity [6], tumor-infiltrating lympho - cytes [ 7], and IFN-γ gene signature [8]. Conversely, a lack of CPI efficacy is associated with inadequate antitumor T cell responses and immunosuppressive factors in the tumor microenvironment (TME) [9]. Although incre - mental benefits are achieved by simultaneously inhibiting multiple immune checkpoints [10], the quality and quan - tity of T cells specific for tumor antigens remains a limit - ing factor for further advancements in immunotherapy. Therapeutic cancer vaccines (TCVs) aim to direct the expansion of T cells targeting relevant tumor antigens, providing a new wave of tumor-specific T cells to the TME [11]. This approach thereby serves to supplement and reinforce the anti-tumor immune response, synergiz - ing with immunotherapies that depend on the presence of tumor-specific T cell responses (e.g. CPIs). Strate - gies applied to TCVs include targeting neoantigens and shared tumor-associated antigens (TAAs). Tumor-spe - cific somatic mutations may give rise to aberrant peptides that are sufficiently different from their normal counter - parts, allowing recognition of these neoantigens by the patient’s T cells. In contrast, TAAs are non-mutated anti - gens with a selective expression pattern that is preferen - tially limited to tumors. Personalized neoantigen cancer vaccines require tumor tissue harvesting and next-gener - ation sequencing for in silico neoantigen prediction, and subsequent personalized vaccine production. TAA-based vaccines, however, can forego personalized production and can be delivered directly off-the-shelf to the patient, assuming that the tumor expresses the target antigen. TAA-based vaccines may also be more relevant in can - cers with low tumor mutational burden, where fewer neoantigens are presented. Telomerase reverse transcriptase (hTERT) is a TAA activated in 85–90% of all tumor types [12, 13]. Telom - erase is expressed by cancer cells to maintain telomere replication supporting unconstrained cancer cell pro - liferation and metastasis [14, 15]. Therefore, telomerase activation is considered a hallmark of cancer [16]. Mela - nomas frequently harbor hTERT promoter mutations and gene copy number amplification which are associ - ated with increased hTERT expression [17–19]. These genomic aberrations collectively lead to an increased antigen presence providing a scientific rationale for tar - geting hTERT in melanomas. High tumor telomerase activity is a well-established negative prognostic factor across multiple cancer indications [20–24], whereas anti- telomerase CD4 T cell immune responses are emerging as independent positive prognostic factors validated in several malignancies [25–27]. Based on these character - istics, hTERT is considered a promising TAA for thera - peutic vaccination [28]. UV1 is a TCV composed of three synthetic long pep - tides derived from the active site of hTERT and has been proven to establish robust, long-lasting T cell responses across an HLA-unselected population in three completed phase I clinical trials [29]. Immune responses induced by UV1 have been identified as CD4 Th1-polarized effector memory cells with inflammatory cytokine profiles (tumor necrosis factor-α and IFN-γ). Expanding a population of CD4 Th1 cells targeting a shared tumor antigen could lead to intratumoral re-activation of these cells, inducing an inflammatory TME and immune-mediated cancer cell death [30–32]. CD4 T cells enhance antitumor immunity by licensing dendritic cells for effective antigen presen - tation and by secretion of inflammatory cytokines, pro - moting immune cell infiltration, and effector functions. Immune checkpoints maintain immune responses within a desired physiological spectrum, and their block - ing is expected not only to disinhibit spontaneously primed anti-tumor T cell responses, but also de novo T cell responses induced by vaccination. This provides a therapeutic rationale for combining checkpoint inhibi - tion with TCVs. The cytotoxic T lymphocyte-associated protein 4 (CTLA-4) immune checkpoint competitively inhibits the binding of CD28 on T cells with CD80/CD86 on antigen-presenting cells, thereby reducing T cell acti - vation by preventing the co-stimulation of primed T cells. Ipilimumab is a monoclonal antibody that blocks this immune checkpoint and disrupts negative regula - tion imposed by CTLA-4. We wanted to explore whether combining ipilimumab with the UV1 vaccine would lead to synergy in terms of expanding vaccine-specific T cells, improving anti-tumor immune responses, and clinical outcomes. We have previously reported safety and feasibility data of this phase I/IIa clinical trial evaluating combined UV1 vaccination and ipilimumab in patients with meta - static melanoma [33]. A parallel phase 4 trial evaluating ipilimumab monotherapy at Norwegian hospitals during the same period has since been published, demonstrat - ing clinical outcomes aligned with previously reported data on ipilimumab monotherapy [34, 35]. Considering the combination study yielded comparatively superior progression-free survival, overall survival, and objective responses rate, we sought to further investigate whether Page 3 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 this cohort comprised patients with favorable baseline characteristics and explore the dynamics of the vaccine- induced immune response. Herein, we report an updated survival analysis and extensive translational research from this clinical trial. Methods Study design, patients, and treatments The study design, eligibility criteria, and treatments have been previously described [33]. The UV1/hTERT-MM study was an open-label, single-arm, single-center, phase I/IIa clinical trial (NCT02275416). The primary objective of this study was to assess the safety of ipilimumab com - bined with UV1 vaccination in patients with malignant melanoma. The secondary objectives included immune response assessment, objective response rate (ORR) per RECIST 1.1, overall survival (OS), and progression-free survival (PFS). Key eligibility criteria were age ≥ 18 years and a histologically confirmed diagnosis of unresect - able stage III/IV cutaneous malignant melanoma. An Eastern Cooperative Oncology Group (ECOG) perfor - mance status ≤ 1 and any previous therapies for mela - noma were permitted. The study participants provided written informed consent prior to enrolment.

The study was approved by the competent regulatory authority and independent ethics committee. UV1 consists of three synthetic long peptides derived from the active site of telomerase reverse transcriptase (hTERT 660–689 termed p719-20; hTERT 691–705 termed p725; hTERT 651–665 termed p728). A total of 300 µg of lyophilized peptides in equimolar amounts was reconstituted in water for injection and administered intradermally to the lower abdomen. Granulocyte–mac - rophage colony-stimulating factor (GM-CSF, sargra - mostim) (Leukine, Sanofi Aventis, Bridgewater, NJ, US) was used as a vaccine adjuvant at 75 µg and was injected intradermally at the same site 10–15 min before UV1. Ipilimumab (3 mg/kg) was administered according to the label for up to 4 infusions. Patients received up to nine UV1 vaccinations, initiated one week prior to the first dose of ipilimumab. Immune response assessment Peripheral blood mononuclear cells (PBMCs) were iso - lated from whole blood samples (50 ml in acid dextrose tubes) at baseline and at frequent intervals during the treatment and long-term follow-up periods, as previously described [29]. Briefly, vaccine-specific T cell immune responses were assessed using a standard prolifera - tion assay (3H-Thymidine incorporation). PBMCs were pre-stimulated in vitro for 10–12 days with a mixture of vaccine peptides at 10 μM. After pre-stimulation, the cells were re-stimulated for 48 h with or without 10 μM vaccine peptides using irradiated autologous PBMCs as antigen-presenting cells, and tested in triplicate for pro - liferation by 3H-thymidine incorporation. The stimu - lation index (SI) was calculated by dividing the mean proliferation count in vaccine peptide-stimulated wells by the mean proliferation count in unstimulated wells. A three-fold increase in proliferation towards any of the three peptides, or a mixture of these, was considered an immune response-positive sample. Staphylococcus aureus enterotoxin C3 (SEC3) was used as the positive control in the immune response assay. The reported SI values were those observed during the treatment period, defined as up to 16 weeks after the last vaccination. TCR sequencing DNA was extracted from PBMC samples of all patients at baseline and up to two time points thereafter using the GenElute Mammalian Genomic DNA Kit (Sigma Aldrich), according to the manufacturer´s instructions (Additional file 1: Table S1). In addition, DNA extrac - tion was performed on stimulated PBMC samples after a 10-day in vitro stimulation (N01, N02, N07, and N09) and biopsies from patients N02 and N03. The T cell receptor beta (TRB) locus was amplified from up to 250 ng of genomic DNA, as described pre - viously [36, 37]. Briefly, a TRB repertoire library was generated using two consecutive polymerase chain reac - tions. First, the rearranged TRB locus was amplified and sample-specific barcodes were added to the amplicons. Library concentrations and sizes were determined using a Qubit (Thermo Fisher) and Bioanalyzer (Agilent), respectively. The final library was sequenced on an Illu - mina MiSeq with the MiSeq Reagent Kit v3 (600-cycles) chemistry. TCR profiling and processing Sequencing quality was assessed before and after reper - toire sequencing data processing with IGX Inspect (IGX Platform 3.0.6 August 2021), a quality control application designed for immune receptor sequencing data. Checks included standard fastq quality metrics, such as average read quality, Q30 scores, as well as V and J gene align - ment distributions and read fate, and a percentage break - down of the receptor extraction status of raw reads. All the samples had good sequence quality and a high per - centage of reads with successfully extracted receptors (all > 95%, except for one sample with 90% and one with 70%). Raw fastq files were processed using the IGX Pro - file (IGX Platform 3.0.6 August 2021), a tool that parses immune receptor structural components by aligning germline genes and adaptively correcting errors based on the overall sample quality. The IGX Profile provided Page 4 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 receptor annotation with complementarity-determining region 3 (CDR3) sequences, V and J gene assignments, functionality, alignment scores, and quality information. All receptors with the same CDR3 amino acid sequence were considered instances of the same clone and all anal - yses were performed at the clone level. Count normalization of TCR repertoires was per - formed by downsampling. Identification of significantly expanded clones was performed using EdgeR [38]. The expanded clones already present in the baseline sample were filtered out. A more detailed description of these methods is provided in the Additional file 1. Multiplex immunofluorescence staining Biopsies were harvested at baseline from nine patients and at week 12–15 from five patients (Additional file 1: Table S2). One part of the biopsy was snap-frozen in liq - uid nitrogen and stored at − 80 °C, whereas the other was formalin-fixed and paraffin-embedded (FFPE). FFPE biopsies were used for multiplex immunofluorescence staining. Biopsy Sects. (4 µm thick) were stained using a custom-based 5-color IHC kit (Akoya Biosciences, Marl - borough, MA, USA) and the fully automated Leica Bond RXm (Leica Biosystems, Buffalo Grove, IL, USA). The slides were deparaffinized, rehydrated, and rinsed with distilled H2O. Antigen retrieval and removal of antibod - ies from the previous cycles were performed by boiling at 95 °C at pH 9 (first cycle) or pH 6 (all remaining cycles). For multiplex immunofluorescence staining, a panel of immune markers was developed using antibod - ies against CD4 (rabbit/ERP6855, Abcam, 1:80), CD8α (mouse/144B, Invitrogen/MA5-13,473, 1:100), PD-L1 (rabbit/E1L3N, Akoya, ready to use), and TERT (rabbit/ ab230527, Abcam, 1:400). A cocktail of two antibod - ies was used to identify the melanoma cells: anti-Sox10 (rabbit/EP268-1, Akoya, ready to use) and anti-S100 (mouse/4C4.9, Akoya, ready to use). Staining was devel - oped using amplification HRP-polymer systems and Opal fluorophore dyes (see Additional file 1: Table S3). To visualize the cell nuclei, the tissue was stained with 4′,6-diamidino-2-phenylindole (Spectral DAPI, Akoya). The slides were mounted with Prolong Diamond Anti - fade Mountant (Thermo Fisher, Waltham, MA, USA) and imaged at × 20 magnification using the Vectra® Pola - ris™ Automated Quantitative Pathology Imaging Sys - tem (Akoya Biosciences, Marlborough, MA, USA). Each image was manually reviewed and curated by a patholo - gist to exclude artifacts and staining defects. Whole‑exome and RNA sequencing and downstream analyses Snap-frozen biopsies were disrupted on a TissueLyser LT, followed by DNA extraction using the AllPrep DNA/RNA/miRNA Universal Kit (Qiagen, Hilden, Ger - many). RNA extraction was performed using a GenEl - ute™ Total RNA Purification Kit (Merck). Biopsy DNA and RNA extracts were obtained from nine patients at baseline and five at week 12–15. Whole-exome sequencing (WES) was performed as previously described in Aamdal et al. [33]. Briefly, 1 µg of DNA was used as the starting material for exome library preparation using the Agilent AllExome V5 kit, according to the manufacturer’s protocol. Sequencing was performed pair-ended, generating approximately 90 M PE reads per tumor and 40 M PE reads per nor - mal, using sequencing by synthesis chemistry on a HiSeq4000 system. Variant calling was performed as previously described [33]. Tumor mutational burden (TMB) was defined as the number of non-synonymous variants with an allelic frequency of > 5% per megabase. RNA samples were processed using an Illumina TruSeq stranded mRNA kit with 100 ng as the start - ing material. RNA sequencing was performed on the NextSeq500 using two HighOutput flow cells with 75 bp single-read sequencing. Hierarchical clustering of the genes included in the IFN-γ signature [8 ] was per - formed using Euclidean distance with the Morpheus tool (https:// softw are. broad insti tute.

org/ morph eus). HLA class I expression was assessed for HLA-A, -B, and -C, and HLA-DP , -DQ, and -DR for class II, as pre - viously described [39]. Differentially expressed genes post-treatment compared to baseline were assessed using the NOISeq tool [40] and were calculated for each patient with available biopsies (N01, N02, N03, and N07). Gene set enrichment analysis of the differen - tially expressed genes was performed using WebGestalt [41] and Gene Ontology mapping to the Biological Pro - cesses functional database. The artificial intelligence (AI) prediction platform used for immunogenic neoantigen prediction was the NEC Immune Profiler (NIP) [42]. The NIP software predicted each of the key determinants of antigen pres - entation (AP) for each somatic mutation, by predicting the potential of all tumor-specific mutated peptides to be efficiently presented by each of the patients Class I HLA-A and -B alleles. Statistics The sample size (n) represents the number of patients or samples analyzed.

Survival analyses were performed using the Kaplan–Meier method. All statistical analyses were performed using GraphPad Prism version 9.2.0. (GraphPad Software). Statistical significance was set at p < 0.05. Page 5 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 Results Clinical outcome of combined ipilimumab and UV1 treatment We have previously presented patient demographics and 5-year clinical follow-up data [33]. Between Janu - ary and October 2015, 12 patients with stage IV mel - anoma were enrolled in this study. Patients received a mean of 5.5 UV1 vaccinations (range 3–9) and 3.2 courses of ipilimumab (range 1–4). UV1 was consid - ered safe and well-tolerated, with most adverse events being grade 1–2 injection site reactions. With up to 85.5 months of clinical follow-up, the median PFS was 6.7 months (Fig. 1A), and the median OS was 66.3 months (Fig. 1B). The ORR was 33% (three partial responses and one complete response) (Fig. 1C). The median time to clinical response was 30.2 weeks (range 16.4–155.3), and the median duration of response was 64.1 weeks (range 12-not reached) (Fig. 1D). One patient (N09) had a sustained complete response last - ing more than 5.5 years. A vaccine-induced anti-UV1 immune response was demonstrated in 10/11 evalu - able patients (91%). The median of maximum in vitro T cell proliferation response (stimulation index) across all patients was 11.5 (range 2.3–60.0). Patient N02 exhib - ited the strongest immune response, with a 60-fold increase in T cell proliferation response to in vitro vac - cine peptide stimulation (Fig. 1 E). Baseline tumor microenvironment and TCR repertoire status There was no apparent association between baseline tumor mutational burden (TMB) or neoantigen load and response to therapy (Fig. 2A). One patient had a mis - sense mutation in hTERT (Glu429Lys). However, the mutation did not map to epitopes corresponding to UV1 peptides and did not persist in the post-treatment biopsy (Additional file 1: Figure S1). Alternative lengthening of telomeres (ALT) has been described as an hTERT-inde - pendent telomere maintenance mechanism that may limit the efficacy of hTERT-targeting T cells. Therefore, we examined the presence of loss-of-function mutations in two genes, ATRX and DAXX, known to induce the ALT phenotype [43]. No nonsense or frameshift muta - tions were detected; however, we observed missense mutations in ATRX in patients N09 (Met352Ile) and N01 (Glu2105Lys), and both ATRX and DAXX mutations in patient N06 (Asp2048Asn, Ser377Phe). The ATRX muta - tion present at baseline in patient N01 was not detected in the biopsy specimen harvested at week 12 (Additional file 1: Figure S1). Fig. 1 Treatment outcomes and anti-hTERT immune response induction in patients treated with UV1 and ipilimumab. A Progression-free survival and B overall survival in all patients enrolled (n = 12). C Maximum change in the sum of longest diameter in RECIST 1.1 evaluable patients (n = 9). D Duration of response (n = 4). E Maximum T cell proliferation response in terms of stimulation index (SI) in evaluable patients (n = 11). The dotted line represents positivity threshold (SI ≥ 3). CR complete response, PR partial response, SD stable disease, PD progressive disease, IR immune response Page 6 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 Two patients had tumors polyploid of the hTERT gene. Patient N02 had seven copies of hTERT at baseline, whereas patient N13 had three copies (Fig. 2B). We fur - ther investigated copy number alterations in HLA and TAP genes, as their loss have been described as a mecha - nism of resistance to immunotherapy [44]. Most patients were diploid for HLA and TAP , and we did not observe any loss of heterozygosity for these genes (Fig. 2B). Melanoma cell hTERT expression was determined by co-staining for Sox10/S100 and hTERT and was con - firmed in all evaluable biopsies (baseline sample from patient N11 was not evaluable because of a lack of Sox10/ S100 stained cells). The median fraction of hTERT-posi - tive melanoma cells was 72.7% (range 5.2–99.3) (Fig. 2C). Two patients (N02 and N13) that were polyploid for the hTERT gene also had the highest hTERT + Sox10/S100 density (Fig. 2E). Interestingly, patient N02 displayed the highest in vitro T cell proliferation response against UV1 peptides, but patient N13 did not demonstrate a vaccine- specific T cell response. However, patient N13 had only one post-vaccination sample for the immune response assessment at week 4. There was no overall associa - tion between tumor hTERT density and vaccine-specific peripheral T cell proliferation response (Fig. 2E). Baseline PD-L1 expression and infiltration of CD4 or CD8 T cells were not significantly different between responders and non-responders (Fig. 2F). However, the baseline hTERT intensity in Sox10/S100 positive cells was significantly lower in the clinical responders (Fig. 2F). We performed RNA sequencing of the available biop - sies and found no apparent association between the baseline IFN-γ gene signature and response to therapy (Fig. 2G). Patient N06, with a JAK1 mutation, demon - strated relatively low expression of the IFN-γ signature. Baseline HLA class I and II expression levels were not significantly different between the clinical responders and non-responders (Additional file 1: Figure S2). Across the clinical response categories, we observed a trend towards an inverse relationship between base - line PBMC and biopsy T cell receptor (TCR) repertoire diversity. The complete response patient (N09) had the highest intratumoral TCR diversity at baseline, and com - paratively low PBMC diversity. The diversity and clonal - ity of PBMC samples at baseline were fairly even across clinical response categories (Fig. 2H and I). Evolution of the TCR repertoire and tracking of vaccine‑enriched clonotypes We matched the baseline and post-treatment (week 14–15) tumor biopsy TCR sequencing data available from two patients (N02 and N03, both with PD as BOR). Most intratumoral TCR clonotypes changed between baseline and post-treatment (Fig. 3A). The overlapping TCRs did not have significantly different relative abun - dances after treatment, indicating that treatment did not lead to intratumoral expansion of the T cells (Fig. 3B). Although there was an increase in total unique TCR clo - notypes post-treatment for patient N02, intratumoral TCR clonality increased (Fig. 3C), indicating that fewer TCR clonotypes constituted a larger proportion of the intratumoral TCR repertoire. The opposite was observed for patient N03. We investigated the PBMC TCR rep - ertoire for TCRs unique to the post-treatment biopsies, observing an increasing number of these clonotypes upon treatment, indicating peripheral expansion of these clonotypes. However, these TCRs did not constitute a considerably larger fraction of the total TCR repertoire over time (Fig. 3D). From week 7 PBMC samples, TCR sequencing was performed before and after 10 days in vitro stimulation with the vaccine peptides in patients N01, N02, N07, and N09. We assessed whether vaccine-enriched TCR clo - notypes were detectable in unstimulated PBMCs and a biopsy at later time points (only patient N02 had avail - able calculated vaccine-enriched TCRs and a post-treat - ment biopsy) (Fig. 3E). In patient N02, we observed three clones expanding from week 7 to week 18 (Fig. 3F), and a single vaccine-enriched TCR clone was detected in the biopsy post-treatment (Fig. 3G). For patients N01, N07, and N09, we observed similar findings, with single clones (See figure on next page.) Fig. 2 Baseline tumor microenvironment. A Clinical outcome and genomic characteristics of patients with available biopsies at baseline. Genes in red font are related to immunotherapy resistance [45], and genes in blue telomerase expression. Green squares indicate non-synonymous mutations. B Gene copy alterations of TERT, HLA, and TAP . C Fraction of Sox10/S100 cells also positive for hTERT based on immunofluorescence staining (dotted line represents the median, 72.7%). D Representative immunofluorescence staining of tumor biopsies. CD4 is stained cyan, CD8 is stained blue, hTERT is stained red, and Sox10/S100 is stained green. Image 1 is from patient N04 at baseline, and image 2 from patient N06 at baseline. The hTERT dense area in the top right of image 2 is likely a hair follicle. E Linear regression analysis of immune response (maximum stimulation index (SI)) and hTERT + Sox10/S100 + cell density based on immunofluorescence staining. The association is not significant (Pearson’s correlation, p = 0.47). F Baseline CD4, CD8, and PD-L1 density in responders (R) and non-responders (NR) (Mann–Whitney test, CD4 p = 1.0; CD8 p = 0.71; PD-L1 p = 0.90). Baseline hTERT intensity in melanoma cells was significantly lower in clinical responders (Mann–Whitney test, p = 0.04). G Hierarchical clustering of baseline IFN-γ signature using the Euclidean distance. The grey box indicates missing data (HLA-DQA1 only). H Baseline PBMC and tumor TCR repertoire diversity and I clonality according to RECIST 1.1 response categories. BOR, best overall response; ns, not significant; ND, not detected. *Patient N04 only had week 12 biopsy available for whole-exome and RNA sequencing. The BRAF status was based on a diagnostic biopsy for this patient Page 7 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 Fig. 2 (See legend on previous page.) Page 8 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 expanded in later PBMC samples, reaching up to 12% of the entire repertoire (Additional file 1: Figure S3). To assess whether in vitro stimulation induced the expansion of a few single T cell clones, we compared the TCR clonality of the unstimulated and stimulated PBMC samples (Fig. 3H) and observed no trend towards increased clonality despite strong T cell proliferation responses in the same sample, especially evident in patient N02 (Fig. 3I). Furthermore, there was no overall correlation between in vitro T cell proliferation responses and TCR clonality in the unstimulated samples (Fig. 3J). Evolution of the tumor microenvironment There was no significant overall increase in tumor-infil - trating lymphocytes (TILs) or PD-L1 expression in the five patients with available baseline and post-treatment (week 12–15) biopsy immunofluorescence staining. An increase in CD8 density was observed in patients N01, N02, and N07 (Fig. 4A), and we observed a non-signif - icant trend towards increased delta CD8 density with increased vaccine-specific peripheral T cell responses (Fig. 4B). No loss of heterozygosity for either the HLA/TAP or hTERT genes was observed in post-treatment biop - sies (Fig. 4C). A relative increase in the expression of both HLA class I and II genes was observed in patient N07 post-treatment, and class II only in patient N02 (Fig. 4D). For the genes included in the IFN-γ signature, we observed a relative increase in expression post-treat - ment in responding patient N07, and to a lesser degree in patient N01 (Fig. 4E). The two progressors did not show considerable changes in IFN-γ signature upon treatment. A similar expression pattern was observed for genes related to T cell function and activation, immune check - point molecules, and cytokine activity (Additional file 1: Figure S4). We evaluated mutational contraction and expansion in four patients with available baseline and post-treatment biopsy whole-exome sequencing data. The responding patients (N07 and N01) exhibited a drastic reduction in total single nucleotide variants (SNVs), with only a few overlapping between the two time points. Conversely, the two progressors had a high degree of overlapping SNVs, indicating a limited impact of the treatment and negligible killing of cancer cells harboring these muta - tions (Fig. 4F). The four neoantigens present at baseline in responding patient N01 were not observed post-treat - ment, whereas most neoantigens persisted in non- responding patient N03 (Fig. 4G). Gene set enrichment analysis was performed on dif - ferentially expressed genes post-treatment for the four patients with matching biopsies, observing enriched gene sets related to immune responses in the responding patients, such as “T cell activation” , “Leukocyte prolifera - tion” , although with an FDR > 0.05 (Additional file 1: Fig - ure S5). The two progressors did not show enrichment in gene sets related to adaptive immune responses. Discussion The anti-CTLA-4 monoclonal antibody ipilimumab was the first checkpoint inhibitor (CPI) to receive Food and Drug Administration approval for the treatment of metastatic malignant melanoma. Here, we report trans - lational research and updated clinical follow-up of 12 patients with malignant melanoma enrolled in a clinical trial evaluating ipilimumab and the TCV candidate UV1. Since the completion of this clinical trial, PD-1 inhibitors, either as single agents or in combination with a CTLA-4 or a LAG-3 inhibitor, have replaced ipilimumab mono - therapy as the standard of care for metastatic melanoma. While ipilimumab led to a median OS of approximately 10 months [35], the combination of nivolumab and ipili - mumab further improved clinical outcomes, exhibiting a 6-year overall survival rate of approximately 50% [2]. Despite these advancements, insufficient T cell responses remain a limiting factor for the efficacy of immunother - apy in the treatment of melanoma. TCVs represent a promising approach for boosting T cell responses against tumor antigens without significantly aggravating toxicity. Therapeutic cancer vaccines aiming to mount anti- hTERT immune responses have been evaluated with several platforms, including peptide, mRNA, and DNA- based approaches [28]. UV1 is a multipeptide therapeu - tic vaccine that has demonstrated HLA-independent induction of vaccine-specific T cell responses in patients treated across three completed phase I/IIa clinical tri - als [29]. Effective induction of robust T cell responses is a prerequisite for the potential clinical activity of a TCV. Fig. 3 Evolution of the TCR repertoire on treatment. A Overlapping intratumoral TCR clonotypes at baseline and post-treatment for patient N02 and N03. B Relative abundance (normalized read count) for persisting intratumoral TCRs between baseline and post-treatment. C Intratumoral TCR clonality at baseline and post-treatment. D The number of clonotypes unique to the post-treatment biopsy also detected in PBMC samples, and their fraction of the TCR repertoire. The TCR repertoire fraction was calculated by summing the normalized read count for each TCR and dividing by the total read count for the same sample. E Volcano plot illustrating enriched TCRs after a 10-day in vitro peptide stimulation of PBMCs. Orange dots indicate TCRs with a log fold change above 5, unadjusted p < 0.05, and red dots adjusted p < 0.05. F Vaccine-enriched TCRs identified in unstimulated PBMC samples. TCR clonotypes are labeled according to their rank in terms of log fold change after stimulation. G Vaccine-enriched TCRs identified in tumor biopsies. H Sample clonality pre and post-10-day in vitro vaccine peptide stimulation. I T cell proliferation responses from the same samples. J Unstimulated PBMC sample clonality vs. in vitro T cell proliferation response(See figure on next page.) Page 9 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 Fig. 3 (See legend on previous page.) Page 10 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 Fig. 4 Evolution of the tumor microenvironment on treatment. A CD4, CD8, and PD-L1 density in baseline and post-treatment biopsies. Circles indicate clinical responders and squares non-responders. The difference between baseline and post-treatment was not significant (Mann–Whitney test, CD4 p = 0.69; CD8 p = 0.69; PD-L1 p = 0.84). B Linear regression analysis of difference in CD8 density between post-treatment and baseline versus maximum T cell proliferation response (Pearson’s correlation, p = 0.22). C Copy number status of HLA/TAP and hTERT genes post-treatment. D HLA class I and II expression in baseline and post-treatment biopsies. The p values represent unpaired t test for difference between baseline and post-treatment expression levels. E IFN-γ signature in baseline vs. post-treatment biopsies. F Overlapping single nucleotide variants and G predicted neoantigens between baseline and post-treatment biopsies. IR immune response, TIL tumor-infiltrating lymphocytes, R clinical responder, NR clinical non-responder, BL baseline, PT post-treatment, NS not significant, TPM transcripts per million, ND not detected Page 11 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 While T cell responses after therapeutic vaccination have been well documented in peripheral blood, there is still a need to further elucidate vaccine-specific T cell traffick - ing after peripheral priming and their interaction with the tumor microenvironment. Tumor hTERT protein expression was confirmed in all evaluable biopsies using combined hTERT and mela - noma cell immunofluorescence staining. The relatively high fraction of hTERT positive melanoma cells (median 72.7%) supports the concept of hTERT being a relevant tumor antigen also in otherwise heterogenous tumors. As hTERT activation serves essential tumorigenic functions, the hTERT negative melanoma cells may be bystander cells contributing less to metastasis and are thus less rel - evant for clinical progression. The intensity of hTERT staining in melanoma cells was significantly higher in clinical progressors. The increased staining intensity may be related to a higher tumoral hTERT activity, which is a well-described negative prognostic factor [20–24]. Copy number amplification of the hTERT gene is a mecha - nism of tumor hTERT activation and associates with high tumor hTERT expression [18, 19]. Two biopsies (N02 and N13) were polyploid for the hTERT gene, and interestingly, these two biopsies had the highest hTERT- Sox10/S100 density based on immunofluorescence. Furthermore, patient N02 demonstrated the strongest T cell proliferation response to in vitro peptide stimula - tion, possibly indicating tumoral boosting of the immune response. Regrettably, we had only one PBMC sample (week 4) for immune response assessment of patient N13, which did not show a positive immune response. We did not observe mutations in the UV1 region of hTERT, either at baseline or post-treatment, which could poten - tially render the UV1-specific immune response redun - dant. Inducing immune responses towards epitopes in the hTERT active site theoretically limits tumor immune escape, as mutations in this region could negatively affect telomerase activity and thus impede tumor growth. We observed mutations in ALT-related DAXX and ATRX genes. However, these missense mutations did not induce the ALT phenotype, as hTERT expression was confirmed by immunofluorescence staining of the same biopsies. Baseline CD4 or CD8 T cell infiltration was not asso - ciated with clinical response, and we did not observe a significant influx of TILs in post-treatment samples. The limitations of our study include the small number of patients with evaluable samples, timing of tissue harvest - ing, and intratumoral heterogeneity. As the median time to clinical response was 30.2 weeks, tumor tissue sam - pling at weeks 12–15 may be too early to describe clini - cally relevant T cell infiltration, although TIL influx after ipilimumab treatment of melanoma has been observed after 18 weeks in other studies [46]. We observed a non-significant trend towards tumor CD8 influx with increased peripheral vaccine-specific T cell responses (Fig. 4B). This observation may fit well with the proposed mechanism of action of a therapeutic cancer vaccine, whereby vaccination promotes the infiltration of T cells into the tumor. Nevertheless, this correlation requires further testing in larger cohorts. Increased expression of the IFN-γ gene signature and genes related to T cell activation and cytokine activity was observed in clini - cally responding patient N07 (Fig. 4 and Additional file 1: Figure S4). Conversely, the two non-responding patients exhibited relatively higher expression of the immune checkpoints CD276 and VTCN1 (B7-H3 and B7-H4), the latter being upregulated post-treatment. TCR sequencing is emerging as an important tool for characterizing T cell dynamics and tissue trafficking [47]. By sequencing the rearranged TRB locus, we aimed to elucidate how ipilimumab and hTERT vaccination affected the overall TCR repertoire and whether vaccine- enriched TCR clonotypes were detectable in peripheral blood and tumor biopsies. Our strategy for identifying TCRs related to vaccination consisted of paired TCR sequencing of PBMC samples before and after a 10-day in vitro vaccine peptide stimulation. The TCR clonality of the sample did not increase after the 10-day in vitro stimulation, despite exhibiting strong T cell prolifera - tion responses to vaccine-peptide stimulation. These findings support the concept that the long UV1 vaccine peptides contain multiple epitopes eliciting a diverse T cell response in each patient, rather than single vaccine- specific clonotypes. This hypothesis is further supported by previously published data on diversity among immune responder HLA genotypes and the various HLA restric - tions and epitope specificities of vaccine-specific T cell clones [29]. Nevertheless, we identified TCR clonotypes that were significantly enriched after in vitro stimulation and subsequently detected them in unstimulated PBMCs and tumor tissue. Alternative strategies, such as peptide- MHC multimer or IFN-γ positivity sorting, may be supe - rior for accurately detecting vaccine-specific T cell clones and validate our current approach in future studies. The clinical read-out of our study yielded an ORR of 33%, mPFS of 6.7 months, and mOS of 66.3 months. The clinical outcomes of patients enrolled in a phase 4 clini - cal trial evaluating ipilimumab monotherapy at Norwe - gian hospitals during the same period as our study were recently published (n = 151) [34], demonstrating an ORR of 9%, mPFS of 2.7 months, and mOS of 12.1 months. Conclusions Although the small sample size and lack of a control arm limits interpretations of clinical efficacy, our study provides support for further clinical evaluation of UV1 Page 12 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 vaccination. Anti-telomerase immune responses were established in 91% of patients, and clinical responses were observed in patients with otherwise less favora - ble baseline genomic, transcriptomic, and tumor microenvironmental features predictive of CPI effi - cacy. Currently, five randomized phase II clinical tri - als are evaluating UV1 in combination with various CPIs across multiple indications (NCT05075122, NCT04742075, NCT04382664, NCT04300244, and NCT05344209). Abbreviations CPI: Checkpoint inhibitor; PD-1: Programmed cell death protein 1; PD-L1: Programmed death-ligand 1; IFN-γ: Interferon-γ; TME: Tumor microenviron- ment; TCV: Therapeutic cancer vaccine; TAA : Tumor-associated antigen; hTERT: Human telomerase reverse transcriptase; HLA: Human leukocyte antigen; CTLA-4: Cytotoxic T-lymphocyte-associated protein 4; Th1: T helper 1; ORR: Objective response rate; OS: Overall survival; PFS: Progression-free survival; RECIST 1.1: Response Evaluation Criteria in Solid Tumors 1.1.; ECOG: Eastern Cooperative Oncology Group; GM-CSF: Granulocyte–macrophage colony- stimulating factor; PBMC: Peripheral blood mononuclear cell; SI: Stimulation index; SEC3: Staphylococcus aureus enterotoxin C3; TRB: T cell receptor beta; CDR3: Complementarity-determining region 3; FFPE: Formalin-fixed and paraffin-embedded; WES: Whole-exome sequencing; TIL: Tumor-infiltrating lymphocyte; NR: Not reached; BOR: Best overall response; CR: Complete response; PR: Partial response; SD: Stable disease; PD: Progressive disease; SNV: Single nucleotide variant; ALT: Alternative lengthening of telomeres. Supplementary Information The online version contains supplementary material available at https:// doi.

org/ 10. 1186/ s12967- 022- 03624-z. Additional file 1: Table S1. Samples selected for TCR sequencing. Table S2. Biopsies used for immunofluorescence staining. Table S3. Antibodies and amplification reagents used for multiplex fluorescence IHC. Figure S1. Comparison of mutations in baseline and post-treatment biopsy. Figure S2. Baseline HLA class I and II expression in clinical responders vs. non-responders. Figure S3. Vaccine-enriched TCRs. Figure S4. Gene expression profiles at baseline vs. post-treatment. Figure S5. Gene set enrichment analysis of differentially expressed genes post-treatment. Acknowledgements Not applicable. Author contributions Conception and design: EBE, GG, TG. Development of methodology: EBE, GB, IK, IA, DS, TC, AM, EMI, SMM, MB, EH, GG. Acquisition, analysis, and interpreta- tion: EBE, GB, IK, IA, DS, AM, GG, and all authors. Writing: EBE and all authors. All authors read and approved the final manuscript. Funding The UV1/hTERT-MM trial was sponsored by Ultimovacs ASA (Oslo, Norway). This project received funding from the Norwegian Research Council (grant number 298864) and Eurostars (grant number 284619). Availability of data and materials Relevant data are provided in the publication and supplementary material. Further details are available on reasonable request. Whole-exome sequenc- ing data are deposited in the European Genome-Phenome Archive (https:// ega- archi ve. org/) under accession number EGAS00001005253. The requests to access the dataset should be directed at [email protected] Ethics approval and consent to participate The study participants provided written informed consent prior to enrolment. The trial was conducted in accordance with the ethical principles of the Decla- ration of Helsinki and the International Conference on Harmonization of Good Clinical Practice and approved by an independent ethics committee and the appropriate national and institutional review boards (REK reference 2014/421). Consent for publication Not applicable. Competing interests EBE, SMM, and GG are employees of Ultimovacs ASA or Ultimovacs AB. GG, SMM, and EMI are shareholders in Ultimovacs ASA. GG and EMI are inventors of the UV1 patent. SMM is the founder and shareholder of Immuneed AB, Vivologica AB, and Strike Pharma AB, none of which has had any role in this work. Other authors declare that they have no conflicts of interest. Author details 1 Department of Tumor Biology, Institute for Cancer Research, The Norwegian Radium Hospital, Oslo, Norway. 2 Faculty of Medicine, University of Oslo, Oslo, Norway. 3 Ultimovacs ASA, Oslo, Norway. 4 ENPICOM B.V., DA ’s-Hertogenbosch, Netherlands. 5 Department of Pharmacy, Science for Life Laboratory, Uppsala University, Uppsala, Sweden. 6 NEC Oncoimmunity, Oslo, Norway. 7 Department for Internal Medicine IV - Hematology and Oncology, Universitätsklinikum Halle (Saale), Halle, Germany. 8 Department of Oncology, Oslo University Hos- pital, Oslo, Norway. 9 HistoOne AB, Uppsala, Sweden. 10 Department of Cellular Therapy, Oslo University Hospital, Oslo, Norway. 11 Ultimovacs AB, Uppsala, Sweden. 12 Centre for Bioinformatics, University of Oslo, Oslo, Norway. Received: 10 August 2022 Accepted: 1 September 2022 References 1.

Robert C, Ribas A, Schachter J, Arance A, Grob J-J, Mortier L, Daud A, Car - lino MS, McNeil CM, Lotem M, et al. Pembrolizumab versus ipilimumab in advanced melanoma (KEYNOTE-006): post-hoc 5-year results from an open-label, multicentre, randomised, controlled, phase 3 study. Lancet Oncol.

2019;20:1239–51. 2. Wolchok JD, Chiarion-Sileni V, Gonzalez R, Grob J-J, Rutkowski P , Lao CD, Cowey CL, Schadendorf D, Wagstaff J, Dummer R, et al. Long-Term Outcomes With Nivolumab Plus Ipilimumab or Nivolumab Alone Versus Ipilimumab in Patients With Advanced Melanoma. J Clin Oncol. 2022;40:127–37. 3. Wei SC, Duffy CR, Allison JP . Fundamental mechanisms of immune check - point blockade therapy. Cancer Discov. 2018;8:1069–86. 4. Kim JY, Kronbichler A, Eisenhut M, Hong SH, van der Vliet HJ, Kang J, Shin JI, Gamerith G. Tumor mutational burden and efficacy of immune check - point inhibitors: a systematic review and meta-analysis. Cancers (Basel). 2019;11:1798. 5. McGranahan N, Furness AJS, Rosenthal R, Ramskov S, Lyngaa R, Saini SK, Jamal-Hanjani M, Wilson GA, Birkbak NJ, Hiley CT, et al. Clonal neoanti- gens elicit T cell immunoreactivity and sensitivity to immune checkpoint blockade. Science.

2016;351:1463–9. 6. Taube JM, Klein A, Brahmer JR, Xu H, Pan X, Kim JH, Chen L, Pardoll DM, Topalian SL, Anders RA. Association of PD-1, PD-1 ligands, and other features of the tumor immune microenvironment with response to Anti– PD-1 therapy. Clin Cancer Res. 2014;20:5064–74. 7.

Uryvaev A, Passhak M, Hershkovits D, Sabo E, Bar-Sela G. The role of tumor-infiltrating lymphocytes (TILs) as a predictive biomarker of response to anti-PD1 therapy in patients with metastatic non-small cell lung cancer or metastatic melanoma. Med Oncol.

2018;35:25. 8. Ayers M, Lunceford J, Nebozhyn M, Murphy E, Loboda A, Kaufman DR, Albright A, Cheng JD, Kang SP , Shankaran V, et al. IFN-γ–related mRNA profile predicts clinical response to PD-1 blockade. J Clin Investig. 2017;127:2930–40. Page 13 of 13 Ellingsen et al. Journal of Translational Medicine (2022) 20:419 9. Jenkins RW, Barbie DA, Flaherty KT. Mechanisms of resistance to immune checkpoint inhibitors. Br J Cancer. 2018;118:9–16. 10. Tawbi HA, Schadendorf D, Lipson EJ, Ascierto PA, Matamala L, Castillo Gutié - rrez E, Rutkowski P , Gogas HJ, Lao CD, De Menezes JJ, et al. Relatlimab and Nivolumab versus Nivolumab in untreated advanced melanoma. N Engl J Med. 2022;386:24–34. 11. Saxena M, van der Burg SH, Melief CJM, Bhardwaj N. Therapeutic cancer vaccines. Nat Rev Cancer. 2021;21:360–78. 12. Shay JW, Bacchetti S. A survey of telomerase in human cancer. Eur J Cancer. 1997;33:787–91. 13. Kim N, Piatyszek M, Prowse K, Harley C, West M, Ho P , Coviello G, Wright W, Weinrich S, Shay J. Specific association of human telomerase activity with immortal cells and cancer. Science. 1994;266:2011–5. 14.

Liu Z, Li Q, Li K, Chen L, Li W, Hou M, Liu T, Yang J, Lindvall C, Bjorkholm M, et al. Telomerase reverse transcriptase promotes epithelial-mesen- chymal transition and stem cell-like traits in cancer cells. Oncogene. 2013;32:4203–13. 15.

Hannen R, Bartsch JW. Essential roles of telomerase reverse transcriptase hTERT in cancer stemness and metastasis. FEBS Lett.

2018;592:2023–31. 16. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144:646–74. 17. Huang FW, Hodis E, Xu MJ, Kryukov GV, Chin L, Garraway LA. Highly recurrent TERT promoter mutations in human melanoma. Science. 2013;339:957–9. 18. Pirker CH, Holzmann Klaus, Spiegl-Kreinecker Sabine, Elbling Leonilla, Thal- linger Christiane, Pehamberger Hubert, Micksche Michael, Berger Walter. Chromosomal imbalances in primary and metastatic melanomas: over-rep - resentation of essential telomerase genes. Melanoma Res. 2003;13:483–92. 19. Zhang A, Zheng C, Lindvall C, Hou M, Ekedahl J, Lewensohn R, Yan Z, Yang X, Henriksson M, Blennow E, et al. Frequent amplification of the telomerase reverse transcriptase gene in human tumors. Cancer Res. 2000;60:6230–5. 20. Taga S, Osaki T, Ohgami A, Imoto H, Yasumoto K. Prognostic impact of tel- omerase activity in non-small cell lung cancers. Ann Surg.

1999;230:715–20. 21. Bertorelle R, Briarava M, Rampazzo E, Biasini L, Agostini M, Maretto I, Lonardi S, Friso ML, Mescoli C, Zagonel V, et al. Telomerase is an independent prognostic marker of overall survival in patients with colorectal cancer. Br J Cancer.

2013;108:278–84. 22. Clark GM, Osborne CK, Levitt D, Wu F, Kim NW. Telomerase activity and survival of patients with node-positive breast cancer. J Natl Cancer Inst. 1997;89:1874–81. 23. Juratli TA, Thiede C, Koerner MVA, Tummala SS, Daubner D, Shankar GM, Williams EA, Martinez-Lage M, Soucek S, Robel K, et al. Intratumoral heterogeneity and TERT promoter mutations in progressive/higher-grade meningiomas. Oncotarget. 2017;8:109228. 24. Hugdahl E, Kalvenes MB, Mannelqvist M, Ladstein RG, Akslen LA. Prognostic impact and concordance of TERT promoter mutation and protein expres- sion in matched primary and metastatic cutaneous melanoma. Br J Cancer. 2018;118:98–105. 25. Laheurte C, Dosset M, Vernerey D, Boullerot L, Gaugler B, Gravelin E, Kaulek V, Jacquin M, Cuche L, Eberst G, et al. Distinct prognostic value of circulating anti-telomerase CD4(+ ) Th1 immunity and exhausted PD-1(+ )/TIM-3(+ ) T cells in lung cancer. Br J Cancer. 2019;121:405–16. 26.

Nardin C, Laheurte C, Puzenat E, Boullerot L, Ramseyer M, Marguier A, Jac- quin M, Godet Y, Aubin F, Adotevi O. Naturally occurring telomerase-specific CD4 T cell immunity in melanoma. J Invest Dermatol. 2021. https:// doi. org/ 10. 1016/j. jid. 2021. 07. 160. 27. Dosset M, Castro A, Carter H, Zanetti M. Telomerase and CD4 T cell immunity in cancer. Cancers (Basel). 2020. https:// doi. org/ 10. 3390/ cance rs120 61687. 28. Ellingsen EB, Mangsbo SM, Hovig E, Gaudernack G. Telomerase as a target for therapeutic cancer vaccines and considerations for optimizing their clini- cal potential. Front Immunol. 2021;12: 682492. 29. Ellingsen EB, Aamdal E, Guren T, Lilleby W, Brunsvig PF, Mangsbo SM, Aamdal S, Hovig E, Mensali N, Gaudernack G, Inderberg EM. Durable and dynamic hTERT immune responses following vaccination with the long- peptide cancer vaccine UV1: long-term follow-up of three phase I clinical trials. J Immunother Cancer. 2022;10: e004345. 30. Kreiter S, Vormehr M, van de Roemer N, Diken M, Löwer M, Diekmann J, Boegel S, Schrörs B, Vascotto F, Castle JC, et al. Mutant MHC class II epitopes drive therapeutic immune responses to cancer. Nature.

2015;520:692–6. 31. Borst J, Ahrends T, Babala N, Melief CJM, Kastenmuller W. CD4(+ ) T cell help in cancer immunology and immunotherapy. Nat Rev Immunol. 2018;18:635–47.

32.

Cohen M, Giladi A, Barboy O, Hamon P , Li B, Zada M, Gurevich-Shapiro A, Beccaria CG, David E, Maier BB, et al. The interaction of CD4+ helper T cells with dendritic cells shapes the tumor microenvironment and immune checkpoint blockade response. Nat Cancer.

2022;3:303–17. 33. Aamdal E, Inderberg EM, Ellingsen EB, Rasch W, Brunsvig PF, Aamdal S, Heintz KM, Vodak D, Nakken S, Hovig E, et al. Combining a universal telomer - ase based cancer vaccine with ipilimumab in patients with metastatic melanoma - five-year follow up of a phase I/IIa trial. Front Immunol.

2021;12: 663865.

34. Aamdal E, Jacobsen KD, Straume O, Kersten C, Herlofsen O, Karlsen J, Hus- sain I, Amundsen A, Dalhaug A, Nyakas M, et al. Ipilimumab in a real-world population: a prospective phase IV trial with long-term follow-up. Int J Cancer. 2021. https:// doi. org/ 10. 1002/ ijc. 33768.

35. Hodi FS, O’Day SJ, McDermott DF, Weber RW, Sosman JA, Haanen JB, Gonzalez R, Robert C, Schadendorf D, Hassel JC, et al. Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med. 2010;363:711–23. 36. Simnica D, Schliffke S, Schultheiß C, Bonzanni N, Fanchi LF, Akyüz N, Gösch B, Casar C, Thiele B, Schlüter J, et al. High-throughput immunogenetics reveals a lack of physiological T cell clusters in patients with autoimmune cytope - nias. Front Immunol.

2019. https:// doi. org/ 10. 3389/ fimmu. 2019. 01897.

37. Simnica D, Akyüz N, Schliffke S, Mohme M, v.Wenserski L, Mährle T, Fanchi LF, Lamszus K, Binder M. T cell receptor next-generation sequencing reveals cancer-associated repertoire metrics and reconstitution after chemotherapy in patients with hematological and solid tumors. OncoImmunology. 2019;8:e1644110. 38. Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformat - ics.

2009;26:139–40. 39. Anzar I, Sverchkova A, Samarakoon P , Ellingsen EB, Gaudernack G, Stratford R, Clancy T. Personalized HLA typing leads to the discovery of novel HLA alleles and tumor-specific HLA variants. HLA. 2022;99:313–27. 40. Tarazona S, Furió-Tarí P , Turrà D, Pietro AD, Nueda MJ, Ferrer A, Conesa A. Data quality aware analysis of differential expression in RNA-seq with NOISeq R/Bioc package. Nucleic Acids Res. 2015;43:e140–e140. 41. Liao Y, Wang J, Jaehnig EJ, Shi Z, Zhang B. WebGestalt 2019: gene set analysis toolkit with revamped UIs and APIs. Nucleic Acids Res. 2019;47:W199–205. 42. Malone B, Simovski B, Moliné C, Cheng J, Gheorghe M, Fontenelle H, Var - daxis I, Tennøe S, Malmberg J-A, Stratford R, Clancy T. Artificial intelligence predicts the immunogenic landscape of SARS-CoV-2 leading to universal blueprints for vaccine designs. Sci Rep. 2020;10:22375. 43. Heaphy CM, De Roeland FW., Jiao Yuchen, Klein Alison P , Edil Barish H, Shi Chanjuan, Bettegowda Chetan, Rodriguez Fausto J, Eberhart Charles G, Sachidanand Hebbar G, Offerhaus Johan, Roger McLendon B, Rasheed Ahmed, He Yiping, Yan Hai, Bigner Sueli Mieko, Oba-Shinjo Suely Kazue, Marie Nagahashi, Riggins Gregory J, Kinzler Kenneth W, Vogelstein Bert, Hruban Ralph H, Maitra Anirban, Papadopoulos Nickolas, Meeker Alan K. Altered telomeres in tumors with ATRX and DAXX mutations. Science.

2011;333:425. 44. Montesion M, Murugesan K, Jin DX, Sharaf R, Sanchez N, Guria A, Minker M, Li G, Fisher V, Sokol ES, et al. Somatic HLA class I Loss Is a widespread mechanism of immune evasion which refines the use of tumor mutational burden as a biomarker of checkpoint inhibitor response. Cancer Discov.

2021;11:282–92. 45. Keenan TE, Burke KP , Van Allen EM. Genomic correlates of response to immune checkpoint blockade. Nat Med. 2019;25:389–402. 46. Sharma A, Subudhi SK, Blando J, Scutti J, Vence L, Wargo J, Allison JP , Ribas A, Sharma P . Anti-CTLA-4 immunotherapy does not deplete FOXP3+ regula- tory T cells (Tregs) in human cancers. Clin Cancer Res. 2019;25:1233–8. 47. Rosati E, Dowds CM, Liaskou E, Henriksen EKK, Karlsen TH, Franke A. Overview of methodologies for T-cell receptor repertoire analysis. BMC Biotechnol. 2017;17:61. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub - lished maps and institutional affiliations.

Citation: Hao, Q.; Long, Y.; Yang, Y.; Deng, Y.; Ding, Z.; Yang, L.; Shu, Y.; Xu, H. Development and Clinical Applications of Therapeutic Cancer Vaccines with Individualized and Shared Neoantigens. Vaccines 2024 ,12, 717. https://doi.org/10.3390/ vaccines12070717 Academic Editor: Walter J. Storkus Received: 29 May 2024 Revised: 18 June 2024 Accepted: 24 June 2024 Published: 27 June 2024 Copyright: ©2024 by the authors. Licensee MDPI, Basel, Switzerland.

This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). Review Development and Clinical Applications of Therapeutic Cancer Vaccines with Individualized and Shared Neoantigens Qing Hao1,†, Yuhang Long1,†, Yi Yang1, Yiqi Deng1,2, Zhenyu Ding1, Li Yang1, Yang Shu1,3,4,* and Heng Xu1,4,5,* 1State Key Laboratory of Biotherapy and Cancer Center, Department of Biotherapy, West China Hospital, Sichuan University, Chengdu 610041, China; [email protected] (Q.H.); [email protected] (Y.L.); [email protected] (Y.Y.); [email protected] (Y.D.); [email protected] (Z.D.); [email protected] (L.Y.) 2Colorectal Cancer Center, Department of General Surgery, West China Hospital, Sichuan University, Chengdu 610041, China 3Gastric Cancer Center, Department of General Surgery, West China Hospital, Sichuan University, Chengdu 610041, China 4Institute of General Surgery, West China Hospital, Sichuan University, Chengdu 610041, China 5Research Center of Clinical Laboratory Medicine, Department of Laboratory Medicine, West China Hospital, Sichuan University, Chengdu 610041, China *Correspondence: [email protected] (Y.S.); [email protected] (H.X.) †These authors contributed equally to this work. Abstract: Neoantigens, presented as peptides on the surfaces of cancer cells, have recently been proposed as optimal targets for immunotherapy in clinical practice. The promising outcomes of neoantigen-based cancer vaccines have inspired enthusiasm for their broader clinical applications. However, the individualized tumor-specific antigens (TSA) entail considerable costs and time due to the variable immunogenicity and response rates of these neoantigens-based vaccines, influenced by factors such as neoantigen response, vaccine types, and combination therapy. Given the crucial role of neoantigen efficacy, a number of bioinformatics algorithms and pipelines have been developed to improve the accuracy rate of prediction through considering a series of factors involving in HLA-peptide-TCR complex formation, including peptide presentation, HLA-peptide affinity, and TCR recognition. On the other hand, shared neoantigens, originating from driver mutations at hot mutation spots (e.g., KRASG12D), offer a promising and ideal target for the development of therapeutic cancer vaccines. A series of clinical practices have established the efficacy of these vaccines in patients with distinct HLA haplotypes. Moreover, increasing evidence demonstrated that a combination of tumor associated antigens (TAAs) and neoantigens can also improve the prognosis, thus expand the repertoire of shared neoantigens for cancer vaccines. In this review, we provide an overview of the complex process involved in identifying personalized neoantigens, their clinical applications, advances in vaccine technology, and explore the therapeutic potential of shared neoantigen strategies. Keywords: neoantigen; therapeutic cancer vaccines; cancer immunotherapy 1. Introduction Cancer is largely attributed to the accumulation of somatic genomic alterations, which not only promotes malignant cell outgrowth, proliferation, and survival but also enables the immune system to identify and attack cancer cells by distinguishing them as ‘foreign’ [ 1]. Different types of genomic alterations were identified in tumor cells, including single nucleotide variants (SNV), copy number alterations, and structural variants, thus disrupting normal cellular functions and leading to the malignant transformation of cells. Crucially, these alterations have the potential to generate novel epitopes through protein translation, proteasome degradation and Major Histocompatibility Complex (MHC) class I presentation. These new antigenic epitopes, recognized as ‘non-self’ by the host, are referred to as tumor- specific antigens (TSAs) or neoantigens. Their exclusive presence in malignant tissues Vaccines 2024 ,12, 717. https://doi.org/10.3390/vaccines12070717 https://www.mdpi.com/journal/vaccines Vaccines 2024 ,12, 717 2 of 24 triggers a series of tumor-specific immune responses [ 2], thereby endowing neoantigens as valuable targets for cancer treatment. Recently , extensive clinical studies have underscored the antitumor activity of neoantigen- based vaccines in patients across various cancers, including the severe cancer types (e.g., pancreatic cancer) [ 3–10]. Although in phase I/II trials with small sample size, cancer vaccines have exhibited great success to reduce the tumor volume and prolong the overall survival of cancer patients, thus were considered as the next immunotherapy frontier [ 11]. However, this personalized therapeutic approach has substantial limitations, including the high cost and time-consuming synthesis processes, a lack of consensus on the best practices for neoantigen discovery, an unclear optimal vaccine delivery scheme, and challenges in transferring between individuals. To achieve robust clinical efficacy, further development and investigation are necessary. Notably, in addressing the major rate-limiting step in the manufacture of individualized vaccines, the development of shared vaccines targeting recurrent drive mutation-derived neoantigens presents a viable solution for enhancing clinical efficacy in an efficient and economical manner [12]. Herein, we provide an overview of the complex procedures involved in identify- ing personalized neoantigens, outline the clinical development landscape of therapeutic neoantigen-based cancer vaccines, and discuss the clinical benefits of combining neoantigen- based vaccines with other therapies. Furthermore, we explore the therapeutic promise of research into shared neoantigen strategies and their clinical applications. 2. Prediction and Selection of Neoantigens 2.1. Source of Neoantigen High-throughput sequencing technology and bioinformatics analysis have now en- abled the in silico prediction of potential neoantigens, revolutionizing our approach to cancer research (Figure 1). In conventional practices, tumor tissues and paired normal samples (e.g., peripheral blood) are collected and analyzed by whole exome sequencing (WES) to identify protein-altering somatic mutations, comprising single-nucleotide variants (SNVs), short insertions and deletions/insertion of nucleotides (indels). In addition, tran- scriptomic sequencing-based analysis not only provides insights into the expression levels of mutated genes and helps to further mutation verification, but also detects fusion genes and tumor specific splicing. Somatic mutations occurring in the coding regions could serve as the mechanism for the ‘foreignness’ of cancer cells and alter protein sequences in various ways. Given their abundance in many cancers, SNVs are the earliest targets used to identify neoantigens and remain a primary focus within this field. Although only a single amino acid substitution, SNV-derived neoantigens are capable of eliciting the antitumor immune responses [ 3–5,13]. In contrast, Indels can result in the in-frame or frameshift translations of the open reading frame, producing polypeptide sequences with little or no homology to the wild-type counterpart, thereby enhancing immunogenicity. The ‘dissimilarity’ of anti- gens to their wild-type counterpart has been utilized as a predictive indicator to evaluate recognition by host immunity [ 14,15]. Furthermore, fusion genes, which link two unrelated genes by intra- or inter-chromosomal rearrangements, are another potential resource of highly immunogenic neoantigens [ 16]. Although a series of recurrent fusion genes have been identified in multiple cancer types, particularly common in leukemia [ 11,17,18], the vast majority of fusion genes appear to be individual occurrences, since gene fusions are considered relatively rare events [19,20]. Confronted with the problem of low mutation burden in many cancer types [ 21], investigators are looking beyond the coding region to cryptic peptides derived from non- canonical transcription and translation [ 22,23] (Figure 2). These cryptic peptides, emerging from ‘non-translated’ region, non-canonical reading frames, and initiation at non-AUG start codons, introduce novel protein sequences with minimal overlap with their regular counterparts [ 22]. Recent studies have demonstrated that cryptic antigens constitute a significant portion of the cancer-associated immunopeptidome [ 24,25]. Specifically, alterna- tive splicing events, including back-splicing circularization that generates circular RNAs, Vaccines 2024 ,12, 717 3 of 24 have been identified as sources of cryptic neoantigens, expanding the landscape of tumor- specific antigens [ 23,26–28]. Alternative splicing allows a single gene to produce multiple protein isoforms [ 29], and is more prevalent in the various types of cancers compared to that in the normal tissues [ 26]. Such an expansion offers promising new opportunities for neoantigen-based therapy [ 30]. Furthermore, altered patterns of post-translational modifications, influenced by cancer-specific pathways, might affect the immunological foreignness which potentially leads to the creation of cancer-specific neoantigens [ 31,32]. However, the cryptic neoantigens and their full potential in immunotherapy remain to be fully elucidated, suggesting a clear direction for future research in this promising area. Vaccines 2024 , 12, x FOR PEER REVIEW 3 of 26 regular counterparts [22]. Rece nt studies have demonstrated that cryptic antigens consti- tute a signi ficant portion of the cancer-associa ted immunopeptidome [24,25]. Speci fically, alternative splicing events, including back-splicing circularization that generates circular RNAs, have been identi fied as sources of cryptic neoantigens, expanding the landscape of tumor-speci fic antigens [23,26–28]. Alternative splicing allows a single gene to produce multiple protein isoforms [29], and is more prevalent in the various types of cancers com- pared to that in the normal tissues [26]. Such an expansion o ffers promising new oppor- tunities for neoantigen-based ther apy [30]. Furthermore, altered pa tterns of post-transla- tional modi fications, in fluenced by cancer-speci fic pathways, might a ffect the immuno- logical foreignness which potentially leads to the creation of cancer-speci fic neoantigens [31,32]. However, the cryptic neoantigens and their full potential in immunotherapy re-main to be fully elucidated, suggesting a clear direction for future research in this prom- ising area. Recently, pioneering studies have shown that intra-tumor bacteria are an intrinsic part of the tumor microenvironment (TME) ac ross various human canc er types, residing intracellularly within the cytoplasm of both tumor cells and immune cells [33]. This dis- covery underscores the complex interplay between microbial presence and cancer biol- ogy. Furthermore, T cell response elicited by bacteria could cross-react with tumor anti- gens, suggested that homologous epitopes sh ared by both the bacteria and tumors can contribute to anti-tumor immunity [34,35]. Such findings suggest that exposing microbial epitopes through the elimination of intra-tu mor bacteria could provide alternative tumor epitopes for cancer immunotherapy. Building on this concept, a pivotal study has shed light on this approach by delivering antibiotics encapsulated in liposomes to target intra- tumor bacteria, eliciting an anti-tumor CD8 + T cell response activated by microbial-de- rived neoantigens [36]. Although these cross-re active T cell responses have been initially uncovered in mouse models, they present an innovative strategy for harnessing the im-mune system against cancer, thereby expandin g the repertoire of neoantigens available for cancer immunotherapy. Figure 1. The full design of neoantigen vaccine. Patient tumor samples and peripheral blood mononuclear cells are isolated for DNA and RNA sequencing, followed by the initiation of the neoantigen prediction process. In this workflow, tumor-specific mutations are first identified, along with patient HLA typing. The binding of mutated peptides to HLA alleles and their potential to elicit T cell responses are predicted and prioritized for neoantigen screening. Using the chosen vaccine platform (such as mRNA, peptides, or dendritic cells), personalized neoantigen vaccines are produced on demand under Good Manufacturing Practice (GMP) conditions. After immunization, the immunogenicity of the selected neoantigens is validated using immunological assays with the patient’s peripheral blood T cells. WES: Whole exome sequencing. Recently, pioneering studies have shown that intra-tumor bacteria are an intrinsic part of the tumor microenvironment (TME) across various human cancer types, residing intra- cellularly within the cytoplasm of both tumor cells and immune cells [ 33]. This discovery underscores the complex interplay between microbial presence and cancer biology. Further- more, T cell response elicited by bacteria could cross-react with tumor antigens, suggested that homologous epitopes shared by both the bacteria and tumors can contribute to anti- tumor immunity [ 34,35]. Such findings suggest that exposing microbial epitopes through the elimination of intra-tumor bacteria could provide alternative tumor epitopes for cancer immunotherapy. Building on this concept, a pivotal study has shed light on this approach by delivering antibiotics encapsulated in liposomes to target intra-tumor bacteria, eliciting an anti-tumor CD8+T cell response activated by microbial-derived neoantigens [ 36]. Al- Vaccines 2024 ,12, 717 4 of 24 though these cross-reactive T cell responses have been initially uncovered in mouse models, they present an innovative strategy for harnessing the immune system against cancer, thereby expanding the repertoire of neoantigens available for cancer immunotherapy. Vaccines 2024 , 12, x FOR PEER REVIEW 4 of 26 Figure 1. The full design of neoantigen vaccine. Pati ent tumor samples and peripheral blood mono- nuclear cells are isolated for DNA and RNA sequencing , followed by the initiation of the neoantigen prediction process. In this work flow, tumor-speci fic mutations are first identi fied, along with patient HLA typing. The binding of mutated peptides to HLA alleles and their potential to elicit T cell re- sponses are predicted and prioritized for neoantig en screening. Using the chosen vaccine platform (such as mRNA, peptides, or dend ritic cells), personalized neoantigen vaccines are produced on demand under Good Manufacturing Practice (GMP ) conditions. After immunization, the immuno- genicity of the selected neoantigens is validated using immunological assays with the patient’s pe- ripheral blood T cells. WES: Whole exome sequencing. Figure 2. Cryptic peptide derived from non-canonical tr anscription and translation. Non-canonical transcription is an alternative sp licing event. Non-canonical translation often includes: translation of ‘non-translated’ regions, translation of non- canonical reading frames, and translation of non- AUG start codons. Changes in post-translational modi fication pa tterns may also lead to the produc- tion of cancer-speci fic neoantigens. 2.2. Prediction of Neoantigen The work flows for neoantigen discovery have been extensively reviewed by several research groups [37–41]. Therefore, herein we brie fly overview the key processes of neo- antigen prediction and summarize recent advancements in this field. Several teams have released their open-sou rce neoantigen prediction pipelines, in- cluding MuPeXI [42], INTEGRATE-NEO [43], NeoPredPipe [44], pTuneos [45], ASNEO [46], NeoFuse [47], pVACtools [48], QBRC [49], Seq2Neo [50], nextNEOpi [51], Im- muneMirror [52]. Most of these tools are one-stop solution for neoantigen identi fication, starting with raw sequences data from DNA and/or RNA. The latest software, spanning from 2020 to the present, are detailed in Tabl e 1. SNVs, indels, and gene fusions are in- c l u d e d a s t h e m a i n s o u r c e s i n t h e m a j o r i t y o f t o o l s , s u c h a s n e x t N E O p i , p V A C t o o l s , Seq2Neo, etc. Speci fically, the QBRC neoantigen calling pipeline introduces a novel Figure 2. Cryptic peptide derived from non-canonical transcription and translation. Non-canonical transcription is an alternative splicing event. Non-canonical translation often includes: translation of ‘non-translated’ regions, translation of non-canonical reading frames, and translation of non-AUG start codons. Changes in post-translational modification patterns may also lead to the production of cancer-specific neoantigens. 2.2. Prediction of Neoantigen The workflows for neoantigen discovery have been extensively reviewed by sev- eral research groups [ 37–41]. Therefore, herein we briefly overview the key processes of neoantigen prediction and summarize recent advancements in this field. Several teams have released their open-source neoantigen prediction pipelines, includ- ing MuPeXI [ 42], INTEGRATE-NEO [ 43], NeoPredPipe [ 44], pTuneos [ 45], ASNEO [ 46], NeoFuse [ 47], pVACtools [ 48], QBRC [ 49], Seq2Neo [ 50], nextNEOpi [ 51], ImmuneMir- ror [ 52]. Most of these tools are one-stop solution for neoantigen identification, starting with raw sequences data from DNA and/or RNA. The latest software, spanning from 2020 to the present, are detailed in Table 1. SNVs, indels, and gene fusions are included as the main sources in the majority of tools, such as nextNEOpi, pVACtools, Seq2Neo, etc. Specifically, the QBRC neoantigen calling pipeline introduces a novel indicator to evaluating neoantigen clonal balance, named the Cauchy–Schwarz index of Neoantigens (CSiN). This immunogenicity index is also included in the nextNEOpi tool. Vaccines 2024 ,12, 717 5 of 24 Table 1. Open source neoantigen prediction pipeline. Name Short Description Format of Input Ref. ImmuneMirror Integrative Pipeline and Web Server DNA, RNA-seq [52] nextNEOpiComprehensive pipeline for neoantigen prediction from raw sequencing dataDNA, RNA-seq [51] Seq2NeoA one-stop solution for neoepitope features predictionDNA, RNA-seq [50] QBRCneoantigen calling pipeline from somatic mutationDNA, RNA-seq [49] pVACtoolsA Computational Toolkit, including pVACseq, pVACfuse, pVACviz and pVACvectorVCF [48] NeoFuse Gene fusion-derived neoantigen RNA-seq [47] ASNEO Alternative Splicing-derived neoantigen RNA-seq [46] However, in most cases, these prediction pipelines employ only one or two tools during the key analysis processes of neoantigen prediction, involving mutation calling, HLA typing, and MHC-peptide binding, which could lead to high rates of false positives. In contrast, nextNEOpi and pVACtools stand out by applying multiple tools for these tasks. Employing multi-algorithm consensus approaches likely yields more accurate results, as distinct algorithms typically utilize different detection strategies. 3. Presentation by MHC Molecular The pivotal step for in silico neoantigen prediction relies on the production and presentation of peptides by MHC-I and MHC-II alleles. A number of prediction tools have been developed, including netMHCpan [ 53], MixMHCpred [ 54,55], MHCflurry [ 56], and HLAthena [ 57], which based on different functions [ 58,59]. Notably, netMHCpan and MixMHCpred are capable of predicting MHC class I and II-restricted peptide. In recent years, one of the major advances in MHC-peptide binding prediction has been the training of networks on eluted ligands data detected by immunoprecipitation and followed liquid chromatography-tandem mass spectrometry (LC-MS/MS), which directly provides endogenously processed and presented peptides from cells [ 53,56,57]. In addition, MHCflurry, EDGE [ 60] and MARIA [ 61] consider protease cleavage and flanking sequences as presentation features. Besides these established tools, we summarize the novel methods developed in the past two years for MHC-I and MHC-II binding prediction (Table 2). Table 2. Open Source prediction tools for MHC-peptide binding. Name MHC Class Short Description Ref. DeepMHCI IAn anchor position-aware deep interaction model, with a great performance on non-9-mer peptides[62] LightMHC IA Light Model for pMHC Structure Prediction with Graph Neural Networks[63] PepPPO ICharacterize binding motif via generating repertoires of peptides presented by given MHC-I alleles[64] DeepMHCII IIA novel binding core-aware deep interaction model for accurate MHC-II peptide binding affinity prediction[65] AEGIS IIApply natural language processing algorithms to identify the MHCII immunopeptidome in humans and a preclinincal mouse model[66] TLimmuno2 IIA transfer learning-based, long short-term memory model[67] Computational algorithms for MHC-peptide binding prediction have two categories: allele-specific and pan-specific [ 68]. In allele-specific methods, a separate model is trained Vaccines 2024 ,12, 717 6 of 24 for each MHC allele, which inevitably leads to underperformance for rare MHC alleles with insufficient ligand data. This issue can be addressed by pan-specific MHC-peptide binding predictor or tools like netMHCpan use the homology MHC sequence to infer potential binding properties. Pan-specific methods integrate information about MHC alleles and peptides into a single model, enabling simultaneously learning the binding properties of all MHC alleles. Tools such as DeepMHCI and DeepMHCII utilize this approach. Furthermore, the stability of MHC-peptide complexes (pMHC) serves as another critical indicator for immunogenicity prediction in cancer vaccines. For instance, a high- throughput stability screening method utilize a standard real-time (RT)-PCR instrument to determine temperature denaturation curves and assess stability [ 69], while a neural net- work predictor (i.e., NetMHCstabpan) was also developed [ 70]. Both studies recommend integrating the stability information with MHC-peptide binding to improve the selection of neoantigens in cancer vaccines. 4. Recognition by TCR The recognition and interaction of TCR-pMHC are the crucial determinants of im- munogenicity. The TCR-pMHC prediction models can be broadly categorized into three groups: (a), Similarity-Based TCR Profiling Model: These models, such as TCRdist [ 71], GIANA [72], and GLIPH2 [73], utilize a similarity-weighted Hamming distance to cluster a set of related TCRs and visualize their specific binding motif; (b), Restricted peptide TCR recognition model: Models represented by netTCR [ 74], netTCR-2.0 [ 75], TCRex [ 76] predict TCR sequence recognition for specific peptides, such as viral and few cancer epi- topes; (c) Broad-scope TCR-pMHC interaction prediction model: Models like SETE [ 77], ERGO-II [ 78], and pMTnet [ 79], are designed to predict interactions for given peptides and TCRs, crucial for neoantigen prediction pipelines. For example, pMTnet utilize peptide, MHC alleles and CDR3 sequences of TCR βas inputs to predict the binding rank of each TCR-pMHC, while ERGO-II integrates multiple levels of information, including epitopes, MHC alleles, T cell type, CDR3 αand CDR3 βamino acid, and corresponding V and J genes. Up to now, TCR recognition has made substantial strides. The development of novel tools that utilize various machine learning algorithms to enhance the prediction of TCR- pMHC interactions is detailed in Table 3. Among these algorithms, particularly noteworthy is TEIM [ 80], which use convolutional neural networks (CNN) to learn local interaction between CDR3 βs and peptides, and Panpep [ 81], which adopts neural Turing machines (NTM) to improve TCR binding specificity prediction for any peptide, especially neoanti- gens or exogenous peptides. Unfortunately, current pMHC-TCR interaction predictors are still at a preliminary level, primarily due to the limited validated pMHC-TCR interaction data for adequate training, and a poorly understanding of the underlying binding mecha- nisms. This underscores the importance of continued research and data collection to refine these predictive models. Table 3. Open Source prediction tools for TCR-pMHC interaction. Name Short Description Ref. TABR-BERT A BERT-based transfer learning method [82] TEIMTCR–Epitope Interaction Modelling at Residue Level predicted both pairwise residue distances and contact sites involved in the TCR–epitope interactions[80] PanPepPan-Peptide Meta Learning by combining the concepts of meta-learning and the neural Turing machine, particularly when confronted with unseen epitopes[81] TEPCAMTCR-Epitope identification based on Cross-Attention and Multi-channel convolution[83] TCRmodel2An adapted AlphaFold framework for speedy, accurate modeling of both TCR–pMHC complexes and unbound TCRs[84] Vaccines 2024 ,12, 717 7 of 24 Table 3. Cont . Name Short Description Ref. STAPLERTCR-peptide specificity prediction from full-length TCR-peptide data[85] TCRconv Using contextualized motifs to predicte recognition [86] TEINet A deep learning framework utilizes transfer learning model [87] catELMoPredicting binding between immune cells receptors and antigens based on protein sequence data[88] epiTCRA Random Forest-based method dedicated to predicting the TCR-peptide interactions[89] MixTCRpred A deep learning TCR-epitope interaction predictor [90] EPIC-TRACEA new machine learning model that utilizes the both αandβTCR chains, epitope sequence, and MHC[91] TCRdock Structural based prediction of TCR epitope specificity [92] DeepTRPan peptide-MHC class I binding prediction with a user-friendly web service[93] PiTEA binding affinity prediction consists of two sequence encoders and a stack of linear layers[94] TPBTEA model based on convolutional Transformer for Predicting the Binding of TCR to Epitope[95] AttnTAPThe bi-directional long short-term memory model for robust feature extraction of TCR sequences[96] PhyAugmentationThe deep neural network with physical modeling and data-augmented pseudo-labeling[97] ATM-TCRMulti-head self-attention mechanism to capture biological contextual information[98] 5. Neoantigen and Immunotherapies T cell responses against neoantigens have emerged as the core effectors of cancer therapeutic strategies. Therapeutic cancer vaccines not only amplify pre-existing endoge- nous T cell responses and but also induce de novo ones. Beyond neoantigen-based cancer vaccines, neoantigen-specific T cells also drive the efficacy of immune checkpoint blockade (ICB) [ 99,100] and adoptive T cell therapies [ 101]. Additionally, the success of ICB is closely correlated with the tumor mutation burden (TMB) across a variety of cancers [ 100,102,103]. However, it is the tumor neoantigen burden (TNB)—a measure directly employed in neoantigen evaluation—that may serve as a superior biomarker for immunotherapy out- comes [ 104–106]. TNB provides a more direct assessment of immunogenic potential, thereby potentially improving the predictive accuracy of therapeutic responses. 5.1. Neoantigen-Based Cancer Vaccines The earliest cancer vaccines were developed to prevent liver and cervical cancers, which are caused by hepatitis B virus (HBV) [ 107] and human papillomavirus (HPV), respectively. Nowadays, the focus of cancer vaccines has shifted to a therapeutic strategy against personalized neoantigens. These therapeutic cancer vaccines have entered clinical trials, including dendritic cell (DC)-based [ 7,108], mRNA-based [ 4,10,109], and peptide- based vaccines [ 3,5,6,8,13] (Table 4). In general, these vaccines are designed to combine approximately 20 neoantigens across multiple complementary categories, such as class I- restricted and II-restricted, clone and subclone. This strategic design is aimed at mitigating the risk of off-target effects and addressing immune escape, thereby enhancing efficacy and safety. Additionally, several ongoing clinical trials are summarized in the Supplementary Table S1, highlighting the continued efforts to refine and expand the use of neoantigen- Vaccines 2024 ,12, 717 8 of 24 based vaccines. For some “cold” cancer types (e.g., glioblastoma) with limited somatic mutation rate, combination of neoantigen and tumor associated antigens (TAAs) were also used to increase the potential targets. Table 4. Selected key clinical trial of neoantigens vaccines. Trial (Format) Cancer Type Phase Short Description Ref. NCT01970358 (SLP) Melanoma I/IbOne of the pioneering works on personalized therapeutic neoantigens, demonstrating anti-tumor efficacy in combination with ICIs[3] NCT02035956 (mRNA) Melanoma IOne of the pioneering works on personalized therapeutic vaccines using TAA and neoantigens[4] NCT02287428 (SLP) Glioblastoma I/IbDemonstrates the feasibility of therapeutic neoantigen vaccines in immunologically ‘cold’ tumors[5] NCT02149225 (peptides) Glioblastoma IDemonstrates the feasibility of therapeutic vaccines using TAA and neoantigens in immunologically ‘cold’ tumors[13] NCT02897765 (SLP)Bladder Tumors, Melanoma, Lung CancerIbIn the large cohort, neoantigen-specific CD4+ and CD8+ T cell responses were observed in all vaccinated patients[6] NCT02956551 (Dendritic cell)Non-Small-Cell Lung IProvides new therapeutic opportunities for lung cancer treatment[7] NCT03639714 (GRT-C901 adenovirus-based/GRT- R902 RNA-based)Non-Small Cell Lung Cancer, Colorectal Cancer, Gastroesophageal Adenocarcinoma, Urothelial CarcinomaI/IIDemonstrate exceptional performance in the treatment of patients with advanced metastatic solid tumors[110] NCT03380871 (SLP)Non-Small Cell Lung CancerIbPersonalized neoantigen vaccine with chemotherapy and anti-PD-1 as first-line treatment for non-squamous non-small cell lung cancer[8] NCT03897881 (mRNA) Melanoma IIbReported clinical benefit by prolonged recurrence free survival (RFS) in patients with resected melanoma comparing ICB plus neoantigen mRNA vaccine to ICB alone.[9,111] NCT04161755 (mRNA) Pancreatic Cancer IReported a prolonged RFS in vaccine-responder patients compared to non-responders, demonstrating that neoantigen vaccines offer effective treatment for PDAC[10] NCT03953235 (GRT-C903 adenovirus-based/GRT- R904 RNA-based)Solid Tumor I/IIDemonstrate good tolerability and initial therapeutic potential in advanced solid tumors through vaccines targeting shared neoantigens in combination with ICI[112] NCT04251117 (DNA) Hepatocellular carcinoma I/IIDemonstrate personalized therapeutic vaccines enhancing responses to PD-1 inhibitors through the induction of tumor-specific immunity[113] SLP , synthetic long peptides. Vaccines 2024 ,12, 717 9 of 24 5.1.1. mRNA-Based Vaccines mRNA vaccines emerge as a promising alternative to traditional vaccine methods, showing encouraging outcomes in infectious diseases, like COVID-19 [ 114,115] and various cancer types [ 4,10,109]. They boast numerous benefits, including high potency, rapid development, and low-cost production [ 116], due to the high yields of in vitro transcription (IVT) reactions and advanced industrial setups [ 117]. Given that naked mRNA is rapidly degraded by extracellular RNases and is unable to penetrate cell membranes, effective delivery systems are essential for the successful applications of these vaccines [ 118]. A common approach is to encapsulate the mRNA within lipid nanoparticles (LNPs), that are tiny spheres designed to protect the mRNA molecules and their delivery into cells and tissues [119,120]. In the realm of cancer vaccines, neoantigen-based mRNA vaccines offer several ad- ditional advantages: (a) a single mRNA can incorporate multiple distinct neoantigens, thereby increasing the vaccine’s breadth and potency; (b) mRNAs can encode full-length or long-kmer neoantigen, containing multiple neoepitopes without MHC-restriction. Sahin et al. demonstrated its feasibility in a phase I clinical trial of personalized neoantigen vaccines. In this study, 13 patients with stage III and IV melanoma received a therapeutic mRNA vaccine targeting up to 10 neoantigens. These 10 neoantigens were engineered into two synthetic RNA molecules, each encoding five linker-connected 27mer neoantigens. Neoantigen-specific CD4+ T cell and CD8+ T cell responses are observed in all patients, which led to a sustained progression-free survival [4]. A recent mRNA neoantigen vaccine developed by BioNTech has demonstrated in- spiring clinical benefit [ 10]. Personalized mRNA vaccines were delivered intravenously to patients with pancreatic ductal adenocarcinoma (PDAC). Each mRNA strand encodes up to 10 MHC-I and 10 MHC-II-restricted neoantigens, and encapsulated in LNPs. In the vaccine-responder patients, a prolonged recurrence-free survival (RFS) was observed compared to non-responder patients, suggesting mRNA-based neoantigen vaccine could induce potent antitumor effects. 5.1.2. Peptide-Based Vaccines In recent years, peptide-based therapeutic cancer vaccines targeting neoantigens have achieved significant clinical benefits [ 3,5,6,8,13]. Due to peptide vaccines’ relative ease of manufacturing, low toxicity, and high chemical stability during storage [ 121], peptide- based vaccines are the preferred choice for most trials [ 122]. However, a major limitation of this approach is its inherently low immunogenicity, which hampers broader clinical application [ 123]. To address this issue, researchers have developed several strategies to boost the immunological response. In studies conducted over the last few decades, the size of the vaccinated peptides has proved critical for stimulating protective antitumor immunity. The optimal design involves using synthetic 15–30 mer long neoantigens peptides rather than 8–11 mer short peptides that represent the core epitopes of CD8+ T cells. Short peptides can be direct presentation by all nucleated cells, leading to suboptimal T cells activation. In contrast, long peptides require processing by professional antigen- presenting cells (APCs) residing in draining lymph nodes (dLNs), thereby ensuring the sustained expansion of effector CD8+ T cells [ 124]. Moreover, long peptides engage both CD4+ and CD8+ T cell responses, prolonging the duration of antigen presentation [ 125]. Furthermore, the coadministration of appropriate adjuvants via the subcutaneous route, including Cytokines such as GM-CSF (granulocyte–macrophage colony-stimulating factor), toll-like receptor (TLR) agonist like poly-ICLC (lysine and carboxymethylcellulose) and CpG, could induce powerful antitumor vaccine responses [121]. Ott et al. launched peptide-based neoantigens vaccines in six high-risk melanoma patients [ 3]. As a pioneering work in personalized therapeutic neoantigen clinical trial, they designed their peptide vaccine: synthesized long peptides with poly-ICLC to target up to 20 neoantigens per individual. Four of these patients remained recurrence-free 25 months post-vaccination, while two relapse achieved complete tumor regression fol- Vaccines 2024 ,12, 717 10 of 24 lowing subsequent anti-PD-1 therapy. In long-term follow-ups to four years, these anti- tumor T cell responses induced by neoantigens demonstrated strong efficacy, suggesting neoantigen-specific T cells converted into a memory phenotype and provide long-lasting protection [ 126]. The same peptide-based vaccine scheme also induces neoantigen-specific T cell responses in glioblastoma patients [ 5]. In another glioblastoma trial, Hilf reposted their personalized vaccination phase I results [ 13]. Their vaccines that had poly-ICLC and GM-CSF adjuvants displayed favorable safety and strong immunogenicity, extending the median progression-free survival to 14.2 months, and overall survival to 29.0 months. The immunogenicity of peptide vaccines can be further enhanced through developing novel deliver systems, especially nanoparticle platform [ 127]. Nanoparticles can co-deliver peptides and adjuvants, enhancing their delivery to dLNs, and thus promoting DC inter- nalization, which is necessary for long peptide-based vaccines [ 128,129]. For example, a high-density lipoprotein-mimicking nanodisc was manufactured to co-deliver neoantigen peptides and CpG, and it elicited up to 47-fold greater frequencies of neoantigen-specific T cells than free CpG plus neoantigen peptides [ 130]. The co-delivery scheme can be im- plemented by structure-based programming of Lymph-node-targeted amphiphile (Amph) vaccines [ 131]. The components of Amph vaccines (neoantigens and adjuvants) are modi- fied with diacyl lipids, exhibiting both hydrophilic and lipophilic characteristics and thus can bind with albumin in the plasma. These albumins act as carrier molecules that guide the vaccine to the lymph nodes, enhancing lymph node accumulation and efficient delivery into APCs. These Amph vaccines have been implemented pancreatic and colorectal cancer, where neoantigens derived with KRAS G12D and G12R [ 132]. The phase 1 study shows that among 25 patients, 84% exhibited KRAS-specific T cell responses, and 24% observed tumor biomarker clearance. 5.1.3. DC-Based Vaccines As the most potent APCs, DC can be generated, loaded, and administered to stimulate robust antitumor responses in vivo , thus becoming optimal cell population for vaccina- tion purposes [ 133]. As mentioned above, both mRNA and peptide-based vaccines elicit neoantigen-specific T cells that rely on the uptake of neoantigens by endogenous DCs, a process that is relatively inefficient. DC-based vaccines, however, can efficiently accomplish this step ex vivo through pulsing or transfection, thereby maximizing their efficacy [ 134]. Autologous DCs are initially isolated from patients, subsequently loaded with neoantigens and allowed to mature under optimal conditions. Once matured, these DCs, now equipped with substantial immune-stimulating capabilities, are reintroduced into the patient. The personalized neoantigen DC vaccines have been proved safe and effective in humans as early as 2015. Carreno and his colleagues first reported the clinical results of a small phase I trial in advanced melanoma patients [ 108]. They used DC pulsing with synthetic neoantigens peptides and observed the increased neoantigen-specific T cells in breadth and diversity. In 2021, a single-arm cohort study initiated by West China Hospital applied the personalized neoantigen DC-based vaccines in patients with advanced recurrent lung cancer [ 7]. They vaccinated subcutaneously 12 heavily treated patients with metastatic lung cancer using synthetic neoantigens peptides pulsing DC. All patients tolerated and responded well to this therapy, with an overall response rate (ORR) of 25%, a disease control rate (DCR) of 75%, a median progression-free survival (mPFS) of 5.5 months, and a median overall survival (mOS) of 7.9 months. These findings demonstrated the efficacy of neoantigen DC-based vaccines in cancer patients for the first time. Recently, Liau reposted a phase III clinical trial of the autologous tumor lysate-loaded DC neoantigen vaccines (DCVax-L) in glioblastoma patients [ 135]. The study enrolled 331 patients with newly diagnosed glioblastoma (nGBM) and recurrent glioblastoma (rGBM), with 232 randomized to the DCVax-L group and 99 to the placebo group. Survival rates after 60 months from randomized therapy was 13.0% in the DCVax-L group vs. 5.7% in control, confirming that DCVax-L treatment greatly enhances long-term survival for Vaccines 2024 ,12, 717 11 of 24 nGBM patients. For patients with rGBM, the median OS from relapse was 13.2 months in the DCVax-L group compared to 7.8 months in the control cohort. This is the first trial proven to extend the survival of patients with rGBM through neoantigen therapy. 5.2. TCR-T Adoptive T cell transfer therapy has been significantly reshaped during the devel- opment of neoantigens, evolving towards the generation of engineering T cells equipped with neoantigen-specific T cell receptors (TCRs). This innovative approach, known as TCR-T therapy, bears similarities to CAR-T therapy in that it involves modifying a pa- tient’s autologous T lymphocytes ex vivo. Known TCR sequences that target specific neoantigens are introduced before reinfusing the modified cells back into the patient for cancer treatment. Several clinical trials implementing TCR-T therapy have yielded en- couraging results in treating melanomas, synovial cell sarcoma, ovarian, and pancreatic cancers [136–140] . This approach has a number of advantages: including high antigen sensitivity and near-physiological signaling, which enhance tumor cell detection and killing while also improving T cell persistence [141]. Initially, Robbins reported a series of clinical trial using TCR-T therapy that targeted the HLA-A*02:01-restricted NY-ESO-1 antigen, a well-known type of TAA (also classified as a cancer/testis antigen) [ 137–139]. NY-ESO-1 is expressed in 10% to 50% of metastatic melanomas, breast, prostate, thyroid, and ovarian cancers, as well as 80% of synovial cell sarcomas, with no expression in normal adult tissues except the testis [ 138]. In these trials, objective clinical responses were observed in 11 of 18 patients with synovial cell sarcoma, 11 of 20 patients with melanoma, 16 of 20 patients with multiple myeloma. The success of these trials confirmed the safety and effectiveness of TCR-T therapy and inspired further development of TCR-T targeting personalized neoantigens. For TCR-T targeting neoantigens, the process involves the following steps: neoantigens are first identified by the prediction workflow, followed by the synthesis of peptides to stimulate CD8+ T cells to obtain neoantigen-specific T cells; these cells are then sequenced to identify their TCRs, enabling the generation of neoantigen-specific TCR-engineered T cells. The feasibility of this preparation protocol was demonstrated in a pilot study within just 2 weeks [ 136]. One notable application involved an HLA-C*08:02-restricted TCR-T targeting the mutant KRAS p.G12D in a patient with metastatic pancreatic cancer, where regression of visceral metastases was observed, with the response enduring beyond 6 months [140]. 5.3. Combination Therapies Combining neoantigen vaccines with other immunotherapies—including ICB, adop- tive T cell transfers, surgical excision, radiotherapy or chemotherapy—could enhance therapeutic efficacy for cancer treatment, particularly in treating advanced or aggressive cancers. A recent trial involving a neoantigen vaccine for pancreatic cancer demonstrated significant clinical benefits, which supports this perspective. This trial implemented a multifaceted treatment approach that included surgical excision, ICB, neoantigen vaccines, and four-drug chemotherapy regimen [ 10]. Furthermore, neoantigen vaccines have been effectively used in some clinical trials to delay cancer recurrence post-surgery, yielding impressive results in esophageal cancer, as indicated in trial NCT05023928. The synergy between vaccines and ICB treatments can be explained by their comple- mentary mechanisms. Cancer therapeutic vaccines have the potential to overcome ICB resistance and boost the effectiveness of ICB in combination immunotherapy [ 142,143]. Conversely, ICB has been shown to improve the immunogenicity of neoantigen vaccines and increase the tumor vaccine response rate, as demonstrated in colorectal cancer [ 144]. Furthermore, an inherent challenge with therapeutic neoantigen vaccines is that the antitu- mor T cells they elicit may become dysfunctional and exhausted when facing a high tumor burden. Induction therapy with ICB might help mitigate this issue, improving the overall effectiveness and durability of the immune response [40]. Vaccines 2024 ,12, 717 12 of 24 6. Shared Vaccine Although personalized neoantigen vaccines have demonstrated clinical benefits in various cancers, the time and economic costs of their implementation are unavoidable chal- lenges. To address these limitations, many research groups are simultaneously exploring ‘off-the-shelf’ immunotherapy options. These alternatives, such as shared neoantigen or tumor-associated antigen (TAA) vaccines, offer a more immediate and cost-effective solu- tion. For instance, in some studies, researchers inject TAA vaccines into patients until release of their neoantigen vaccine, which typically takes at least 3 months [ 4,13]. This approach could provide on-going immune protection against cancer during the waiting period. 6.1. A Special Class of Neoantigen: Shared Neoantigen Neoantigens derived from driver mutations in oncogenes or tumor suppressors are attractive targets for immunotherapy. Unlike most ‘passengers’ somatic mutations that do not contribute to oncogenesis, vaccines based on shared neoantigens offer three significant advantages. Firstly, most driver mutations occur in the early stage of tumor development and are considered trunk mutations. Given the high degree of heterogeneity in tumors, trunk mutations exist in the majority of tumor cells, making them more suitable targets for vaccine targets. Secondly, while a substantial portion of patients initially respond to immunotherapy, many subsequently develop mechanisms of immune evasion. These mechanisms include deficits in the antigen presentation machinery, loss of neoantigens, or exploitation of alternate immune checkpoint pathways [ 145]. However, the loss of driver mutations can impair oncogenic pathways, suggesting that selecting neoantigen from driver mutations may reduce the likelihood of immune escape to some extent. Thirdly, some driver mutations recur with high frequencies across different types of cancers, offering the potential for shared targets among patients. This pan-cancer applicability enhances the utility of neoantigen-based vaccines, potentially simplifying the development of broad- spectrum oncological treatments. Further details on the specific neoantigens discussed are provided in Table 5. 6.1.1. KRAS KRAS driver mutations, mostly occurred on 12th or 13th amino acid, which are commonly found in pancreatic ductal adenocarcinoma (PDAC) and colorectal cancer (CRC), present as attractive targets for immunotherapy. These mutations are essential for tumor survival and exhibit constitutive activation throughout the progression of the disease [ 146,147]. Recently, a variety of immunotherapies targeting KRAS mutations have emerged. Adoptive HLA C*08:02-restricted KRAS p.G12D-specific T cells transfer therapy has demonstrated the objective regression in patients with metastatic PDAC and CRC [ 140,148]. Six months after the engineered T cells transfer, the response was still ongoing, and made up more than 2% of all circulating peripheral-blood T cell [ 140]. However, TCR-T scheme is restricted to MHC alleles, and is not conducive to the further application between patients. A clinical trial has been conducted using long peptide vaccines targeting KRAS p.G12D/R mutations in 20 PDAC and 5 CRC patients, where 84% exhibited specific T cell responses to KRAS p.G12D/R and observed tumor biomarker responses [ 132]. The vaccine effectively co- delivered both neoantigens and adjuvants to the APCs. After internalization, neoantigens are presented by the patient’s MHC molecules, featuring simplified manufacturing and off-the-shelf availability. 6.1.2. TP53 TP53 is one of the most frequently mutated genes in human cancer. The loss of p53 protein can lead to the occurrence of cancer, and the development of drugs and therapies targeting TP53 mutations has always been a focal point [ 149]. Recently, Rosenberg and his team have recognized an immunogenic neoantigen (HMTEVVR HC) originating from TP53 Vaccines 2024 ,12, 717 13 of 24 p.R175H mutation within the HLA-A*02:01 context in a patient with metastatic colorectal cancer. This discovery led to the identification of specific T cells and TCR sequences that target this neoantigen [ 150], laying the foundation for subsequent adoptive cell transfer therapy that expresses the same MHC allele and mutation. Building on this foundational work, they conducted TCR-T therapy on a patient with chemotherapy-resistant breast cancer. By transducing her own peripheral blood T cells with an allogeneic HLA-A*02:01-restricted TCR specific for TP53 p.R175H [ 151], they observed objective tumor regression that lasted for six months. Simultaneously, they also treated 12 patients with chemotherapy-resistant epithelial cancer, employing adoptive transfer of ex vivo-expanded autologous tumor-infiltrating lymphocytes (TIL) without any genetic engineering. However, this trial achieved limited clinical responses, with only 2 out of 12 patients showing partial responses. Furthermore, targeting the same mutation can also be engineered into a bispecific antibody, H2-scDb, to stimulate T cell killing of TP53-mutant tumor cells [ 152]. These outcomes highlight the potentially greater efficacy of engineered T cells targeting mutant TP53, compared to the unmodified TIL approach. 6.1.3. IDH1 The IDH gene, responsible for encoding isocitrate dehydrogenase, can undergo muta- tions on 132th amino acid that disrupt cellular metabolism and potentially induce aberrant DNA methylation [ 153]. Such mutations have been identified across a spectrum of tumor types and are an early and decisive event in the development of gliomas. The IDH1 p.R132H mutation recurs in over 70% of diffuse grade II and grade III gliomas. Therefore, Schu- macher et al. made neoantigen peptide library targeting this mutation and demonstrated that, the long peptide p123-142 (GWVKPIIIG HHAYGDQYRAT) is immunogenic neoanti- gen that can elicit class II-restricted CD4+ specific-T cell response in HLA-DRB1*01:01 context [154]. 6.1.4. EGFR The epidermal growth factor receptor (EGFR) is a transmembrane protein that pos- sesses cytoplasmic kinase activity and conveys critical growth factor signals from the extracellular environment to the cell [ 155]. Considering that over 60% of non-small cell lung carcinomas (NSCLCs) exhibit EGFR expression, it has emerged as a crucial target for cancer therapy [155]. Lizee and his colleagues conducted a phase I trial of personalized neoantigen peptide vaccines in 24 stage III/IV NSCLC patients who had previously progressed following multiple conventional therapies, including surgery, radiation, chemotherapy, and tyrosine kinase inhibitors (TKIs) [ 156]. Notably, all seven responders in the trial harbored EGFR mutations, from which two highly shared immunogenic neoantigens were identified: KITDFG RAK from p.L858R, restricted by HLA-A*11:01, and LTSTVQLI Mfrom T790M, restricted by HLA-C*C15:02. Furthermore, it is estimated that approximately 15% of Asian NSCLC patients with EGFR mutations exhibit both the HLA-A*11:01 and p.L858R mutations, pointing to a significant subset of patients who could benefit from tailored therapeutic strategies [157,158]. 6.1.5. PIK3CA Mutations in PIK3CA (the gene encoding phosphatidylinositol 3-kinase alpha, PI3K α) are among the most common genetic alterations driving oncogenesis. PIK3CA mutations, including H1047R, H1047L, E545K, and E542K, show high prevalence in cancers such as BRCA, CESC, and COAD, affecting 24%, 20%, and 16% of cases, respectively [ 159]. Building on this, Chandran established a panel of TCRs that specifically recognize the neoantigen HLA-A*03:01-A LHGGWTTK, derived from the PIK3CA hotspot mutation p.H1047L, highlighting the immunogenicity of the common shared neoantigens in prevalent HLA molecules [ 160]. Following the panel of TCRs, the team validated their functionality through TCR-T adoptive therapy in a mouse model. Tumor regression exclusively in mice Vaccines 2024 ,12, 717 14 of 24 bearing the PIK3CA mutation, whereas those with wild-type PIK3CA did not respond to the treatment. These findings underscore the therapeutic potential of shared neoantigens derived from mutant PIK3CA. 6.1.6. ALK Anaplastic lymphoma kinase (ALK) rearrangements account for 5–6% of all non- small cell lung cancer (NSCLC) cases and are caused by the fusion of ALK with other partner genes. Although several ALK-targeted tyrosine kinase inhibitors (TKIs) have been approved for patients with ALK-positive (ALK+) NSCLC, achieving complete regression is rare due to the development of resistance to these ALK TKIs. This presents a significant challenge in the long-term management of the disease. In response to this issue, Professors Roberto and Rafael B. developed a neoantigen vaccine that targets ALK mutations [ 161]. This innovative approach elicits a strong immune response against ALK, which has shown promising results in preclinical models. Specifi- cally, the vaccine has been successful in eradicating primary tumors in mice and preventing the occurrence of metastatic disease. Additionally, in patients with ALK+ NSCLC, they identified four ALK peptides (AMLDLLHVA, RPRPSQPSSL, IVRCIGVSL, VPRKNITLI) that are presented by HLA-A*02:01 and HLA-B*07:02. The immunogenicity and safety of vaccines targeting these neoantigens were validated in human HLA-transgenic mice, paving the way for the development of a clinical vaccine to treat ALK+ NSCLC. Table 5. Selected immunotherapies targeting shared neoantigens. Gene Frequency Neoantigen Peptides MHC Restriction Immunotherapy KRAS93% PDAC 50% CRCG12D- YKLVVVGAD GVGKSALTI G12R- YKLVVVGAR GVGKSALTI/Neoantigen vaccines [132] G12D-GAD GVGKSA(L) HLA-C*08:02TCR-T therapy [140,148] TP53 20% BRCA R175H-HMTEVVRH C HLA-A*02:01TCR-T therapy [145], bispecific antibody [152] IDH1 70% GLR132H- GWVKPIIIG HHAYGDQYRATHLA-DRB1*01:01Neoantigen vaccines [154] EGFR 60% NSCLCL858R-KITDFGR AK HLA-A*11:01 Neoantigen vaccine [156]T790M-LTSTVQLIM HLA-C*C15:02 PIK3A24% BRCA 20% CESC 16% COADH1047L-AL HGGWTTK HLA-A*03:01TCR-T therapy [160] ALK 5% NSCLCAMLDLLHVA HLA-A*02:01 Neoantigen vaccines [161]IVRCIGVSL RPRPSQPSSL VPRKNITLIHLA-B*07:02 6.2. Tumor Associated Antigens Therapeutic cancer vaccines have historically targeted tumor associated antigens (TAAs), which, unlike tumor specific antigens, may also be expressed in normal cells/tissues but are aberrantly overexpressed in tumor (Figure 3). TAAs have been identified earlier than tumor specific antigens because they can be detected without high throughput sequencing. Building on this discovery, the initial efforts to develop therapeutic cancer vaccines focused on these aberrantly expressed self-antigens. However, in numerous cancer vaccine trials Vaccines 2024 ,12, 717 15 of 24 across various tumor types, most vaccination strategies only induced blood TAA-specific T cell responses; they did not achieve objective clinical benefit [ 162]. The reasons for this could include: (1) TAA-specific T cells have not proliferated in sufficient numbers to recog- nize and eliminate tumor cells. (2) The lack of appropriate maturation signals results in T cells that are either unresponsive or produce regulatory T cells with suppressive effects. (3) TAA-specific T cells do not remain within the tumor long enough to effectively kill the malignant cells [ 163]. Among the few successes, only Provenge®(sipuleucel-T for prostate cancer targeting prostate acid phosphatase) has been approved by the U.S. Food and Drug Administration (FDA). This vaccine has achieved a moderate improvement in treatment outcomes, extending median survival by approximately 4 months [ 164]. Regrettably, even Sipuleucel-T failed due to its limited efficacy and high cost, reflecting the broader challenges in TAA-based vaccine design including overcoming central or acquired immune tolerance, and achieving a sufficient affinity of T cells towards TAAs. Vaccines 2024 , 12, x FOR PEER REVIEW 17 of 26 Figure 3. The distinctions between TAA and TSA. TSA ar ises from altered protein sequences caused by tumor-speci fic mutations. These alterations can be pr esented on the tumor cell surface through the endogenous antigen processing pathway, thereb y triggering direct cytotoxic responses from the host’s T cells. Conversely, TAA ty pically originates from genes th at are abnormally overexpressed in tumor cells but are either unexpressed or mini mally expressed in normal tissues. TAAs are also capable of provoking immune responses, position ing them as promising shared targets for immu- notherapy across various cancer types. Furthermore, following the demise of cancer cells, cellular debris is engulfed by neighboring cells, thereb y enabling the presentation of TAA and TSA to CD4+ T cells via MHC class II. TAA/TSA-speci fic T cells are initially activated by antigen-presenting cells (APCs), such as dendritic cells (DCs), which carry these epitopes. This proc ess typically occurs in peripheral lymphoid organs, and th e activated T cells subsequently di fferentiate into cytotoxic T cells, which ultimately in filtrate the tumor and a ttack the cancer cells. 7. Challenges and Prospects In recent years, extensive preclinical and clinical research have tested various strate- gies of neoantigen discovery and vaccine formulations. Immu notherapy targeting neoan- tigen has achieved impressive e ffects, but several obstacles must be overcome to elicit po- tent antitumor response and achieve full clinical bene fit. One major challenge is the low response rates of predicted neoantigens in practical clinical applications, where only a few candidate neoantigens were recognized by pati ent-derived T cells. To address this issue, iterative re finement of software tools using experimental data and optimization of ma- chine learning models for be tter prediction of MHC-peptide binding a ffinity and prioritize immunogenicity are recommended. Secondly, the antitumor e ffects induced by vaccine- specific T cells are often limited due to the suppressive tumor mi croenvironment. Com- bining neoantigen vaccines with other immu notherapies could potentially overcome this bottleneck, enhancing therapeutic e fficacy. Additionally, the sampling of tumor tissues is Figure 3. The distinctions between TAA and TSA. TSA arises from altered protein sequences caused by tumor-specific mutations. These alterations can be presented on the tumor cell surface through the endogenous antigen processing pathway, thereby triggering direct cytotoxic responses from the host’s T cells. Conversely, TAA typically originates from genes that are abnormally overexpressed in tumor cells but are either unexpressed or minimally expressed in normal tissues. TAAs are also capable of provoking immune responses, positioning them as promising shared targets for immunotherapy across various cancer types. Furthermore, following the demise of cancer cells, cellular debris is engulfed by neighboring cells, thereby enabling the presentation of TAA and TSA to CD4+ T cells via MHC class II. TAA/TSA-specific T cells are initially activated by antigen-presenting cells (APCs), such as dendritic cells (DCs), which carry these epitopes. This process typically occurs in peripheral lymphoid organs, and the activated T cells subsequently differentiate into cytotoxic T cells, which ultimately infiltrate the tumor and attack the cancer cells. Vaccines 2024 ,12, 717 16 of 24 As the field of cancer immunotherapy evolves, the development of therapeutic cancer vaccines presents new paradigms toward personalized neoantigen-based vaccine, particu- larly toward personalized neoantigen-based vaccines. However, the application of TAAs has not been neglected; instead, there is a shift towards utilizing a combination of multiple TAAs. In some trials, TAA vaccines were used to bridge the waiting period during the preparation of neoantigen vaccines [ 4,13]. Additionally, Adot évi et al. developed a clinical trial using a universal cancer peptide–based vaccine (UCPVax) targeting telomerase reverse transcriptase (TERT), a protein found to be overexpressed in more than 85% of cancers [ 165]. This vaccine, consisting of two highly selected 15-mer peptides derived from TERT, was administered to 59 patients with refractory advanced NSCLCs, achieving disease control in 21 cases (39%) and a median overall survival of 9.7 months. Moreover, Kjeldsen et al. de- signed an immunomodulatory vaccine targeting PD-L1 and indoleamine 2,3-dioxygenase (IDO), both critical immunoregulatory molecules in the tumor microenvironment. This vaccine, which includes specific antigens (21-mer peptide DTLLKALLEIASCLEKALQVF from IDO, and 19-mer peptide FMTYWHLLNAFTVTVPKDL from PD-L1), was used in combination with nivolumab to treat patients with metastatic melanoma, achieving an objective response rate of 80% and a complete response rate of 43% [ 166]. These results highlight the safety and immunogenicity of TAA-based vaccine, suggesting its potential as a significant addition to cancer treatment regimens. However, these trials also reveal limitations. Despite the inclusion of multiple peptides in the vaccine designs, the same peptides were uniformly administered to each patient, likely due to cost considerations. This approach, however, ignored the individualized effects of MHC molecule presenta- tion, which are essential for assessing the immunogenicity of TAA/neoantigens. In light of these limitations, clinical trials utilizing individualized MHC-specific TAA peptides may offer a more significant clinical benefit and demonstrate the potential of personalized immunotherapy to enhance treatment efficacy. Moreover, it has been well-established that several cell types (e.g., cancer-associated fibroblasts, and tumor-associated macrophage) in the tumor microenvironment (TME) can facilitate the progression and drug resistance of cancer cells through the secretion of factors or direct interaction [ 167–169], and thus represent a potential therapeutic target for tumor treatment [ 170–172]. With the development and rapid applications of single cell RNA sequencing techniques in investigating the TME, specific cell subtypes that can contribute to tumor progression have been identified, including SPP1+macrophage, FAP+GPX3+cancer associated fibroblast, and CYP4F3+monocyte at pan-cancer level [ 173–175]. Consequently, the aberrantly upregulated genes in TME components can also be the potential resource of TAAs. 7. Challenges and Prospects In recent years, extensive preclinical and clinical research have tested various strategies of neoantigen discovery and vaccine formulations. Immunotherapy targeting neoantigen has achieved impressive effects, but several obstacles must be overcome to elicit potent antitumor response and achieve full clinical benefit. One major challenge is the low response rates of predicted neoantigens in practical clinical applications, where only a few candidate neoantigens were recognized by patient-derived T cells. To address this issue, iterative refinement of software tools using experimental data and optimization of machine learning models for better prediction of MHC-peptide binding affinity and prioritize immunogenicity are recommended. Secondly, the antitumor effects induced by vaccine-specific T cells are often limited due to the suppressive tumor microenvironment. Combining neoantigen vaccines with other immunotherapies could potentially overcome this bottleneck, enhancing therapeutic efficacy. Additionally, the sampling of tumor tissues is difficult to obtained, thus highlighting the importance of sequencing ctDNA from the liquid biopsy for neoantigen prediction and clinical decision-making [176–178]. Beyond their application in cancer immunotherapy, neoantigens also play a significant role in explaining susceptibility to autoimmune diseases. For instance, recent studies Vaccines 2024 ,12, 717 17 of 24 have shown that carboxyethyl modification of a cysteine residue in integrin αIIb disrupts immune tolerance and generates pathogenic neoantigens. These neoantigens, presented by HLA-DRB1*04, stimulate CD4+ T cell responses and induce the production of autoantibod- ies, leading to autoimmune diseases such as ankylosing spondylitis (AS) [179]. Finally, the development of ‘off-the-shelf’ strategies provides opportunities for patients with specific cancer type patients harboring recurrent driver mutations. These vaccines offer a cost-effective and efficient means of treatment, showcasing the broader potential of neoantigen-based therapies in oncology. 8. Conclusions In conclusion, neoantigens have emerged as excellent targets for immunotherapy, particularly in the realm of cancer treatment. Presented as peptides on cancer cells, these antigens offer a personalized approach to immunotherapy that has shown promising clini- cal outcomes. Despite their potential, the application of neoantigen-based cancer vaccines is hindered by substantial costs and variability in effectiveness. Therefore, integrating comprehensive bioinformatics tools with clinical strategies will be critical for optimizing the development and application of neoantigen-based therapies. Future research in this field should focus on refining these tools and expanding the scope of shared neoantigens to enhance the accessibility and effectiveness of cancer vaccines. Ultimately, the continued ex- ploration of neoantigen and TAA-based strategies holds significant promise for improving the prognosis of cancer patients through personalized immunotherapy. Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/vaccines12070717/s1, Supplementary Table S1: Ongoing clinical trials of personalized neoantigen-based vaccines. Author Contributions: Q.H.

and Y.L. wrote the article and made the figures. Y.Y. and Y.D. contributed to the data collection. Z.D., L.Y., Y.S. and H.X. designed and edited the manuscript. All authors contributed to researching the article and to the review and editing of the manuscript. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the National Key R&D Program of China (No. 2023YFC3405200) , Natural Science Foundation of Sichuan Province (No. 2023NSFSC1903 and No. 2023NSFSC1845), 1.3.5 Project for Disciplines of Excellence, West China Hospital, Sichuan University (No. ZYYC23025 and No. ZYYC23013). Conflicts of Interest: The authors declare no conflict of interest. References 1. Schumacher, T.N.; Schreiber, R.D. Neoantigens in Cancer Immunotherapy. Science 2015 ,348, 69–74. [CrossRef] 2. Jiang, T.; Shi, T.; Zhang, H.; Hu, J.; Song, Y.; Wei, J.; Ren, S.; Zhou, C. Tumor Neoantigens: From Basic Research to Clinical Applications. J. Hematol. Oncol. 2019 ,12, 93. [CrossRef] 3. Ott, P .A.; Hu, Z.; Keskin, D.B.; Shukla, S.A.; Sun, J.; Bozym, D.J.; Zhang, W.; Luoma, A.; Giobbie-Hurder, A.; Peter, L.; et al. An Immunogenic Personal Neoantigen Vaccine for Patients with Melanoma. Nature 2017 ,547, 217–221.

[CrossRef] 4. Sahin, U.; Derhovanessian, E.; Miller, M.; Kloke, B.-P .; Simon, P .; Löwer, M.; Bukur, V .; Tadmor, A.D.; Luxemburger, U.; Schrörs, B.; et al. Personalized RNA Mutanome Vaccines Mobilize Poly-Specific Therapeutic Immunity against Cancer. Nature 2017 ,547, 222–226.