Advances in Vaccine Development for Cancer Prevention and Treatment in Lynch Syndrome

Abstract

Lynch Syndrome (LS) is one of the most common hereditary cancer syndromes, and is caused by mutations in one of the four DNA mismatch repair (MMR) genes, namely MLH1, MSH2, MSH6 and PMS2. Tumors developed by LS carriers display high levels of microsatellite instability, which leads to the accumulation of large numbers of mutations, among which frameshift insertion/deletions (indels) within microsatellite (MS) loci are the most common. As a result, MMR-deficient (MMRd) cells generate increased rates of tumor-specific neoantigens (neoAgs) that can be recognized by the immune system to activate cancer cell killing. In this context, LS is an ideal disease to leverage immune-interception strategies. Therefore, the identification of these neoAgs is an ongoing effort for the development of LS cancer preventive vaccines. In this review, we summarize the computational methods used for in silico neoAg prediction, including their challenges, and the experimental techniques used for in vitro validation of their immunogenicity. In addition, we outline results from past and on-going vaccine clinical trials and highlight avenues for improvement and future directions.

1. Introduction

Our improved understanding of tumor microenvironment (TME), especially regarding the different roles of immune cells and tumor-suppressive regulators, has revolutionized cancer treatment and led to the application of combinatory therapies incorporating immune checkpoint inhibitors (ICIs) and targeted agents. Additionally, this understanding has renewed the scientific and medical interest in the development of cancer vaccines, thus leading to rapid advances in therapeutic vaccine platforms and, more recently, in preventive vaccines for non-viral-induced cancers. These different immunotherapy approaches are particularly promising for “hot tumor” types, which are characterized by an increased burden of neoantigens (neoAgs) and high levels of immune cell infiltration. Among them, tumors developed by Lynch Syndrome (LS) carriers, which display microsatellite instability (MSI) secondary to MMR deficiency (MMRd) and as a by-product accumulate hundreds to thousands of neoAgs, are ideal for the application of immunoprevention strategies.

1.1. Lynch Syndrome Prevalence and Etiology

Among hereditary cancer syndromes, LS is one of the most common with a worldwide prevalence of 2 – 3% of all colorectal cancers (CRC), and 1.8% of endometrial cancers (EC) diagnosed (Abu-Ghazaleh et al. 2022; Holter et al. 2022). Traditionally, the indication of genetic testing to detect LS depends on age of diagnosis of cancer (younger than 50 years of age), family history of LS-associated cancers, and application of clinical criteria such as the Amsterdam II criteria (Vasen et al. 1999) and the Bethesda guidelines (Umar et al. 2004) as well as the recognition of the presence of MSI in colorectal and endometrial cancers. However, the application of these tools by providers in the clinic is inconsistent. In addition, these have limited sensitivity, and therefore a substantial number of LS cases are missed from diagnosis. As a consequence, the prevalence of LS is potentially higher than reported (Muller et al. 2019) (Perez-Carbonell et al. 2012; Barnetson et al. 2006).

From the genetics standpoint, LS is an autosomal dominant disease secondary to heterozygous germline mutations in one of the MMR genes: MLH1, MSH2, MSH6, PMS2 and the 3’ terminal deletion of EPCAM (Te Paske et al. 2022). Depending on the MMR gene mutated, the lifetime risk for cancer may vary. For example, LS carriers with MLH1 and MSH2 mutations have up to 80% lifetime risk of developing CRC, while PMS2 carriers have up to 30% in aged individuals (Dominguez-Valentin et al. 2020). In addition, MSH2 and MSH6 female carriers experience up to 60% risk of developing EC. At lower incidences, LS patients are also at risk of ovarian, gastric, small bowel, pancreatic, biliary tract, ureter, bladder, and glioblastomas after age 40 (Moller et al. 2017; Giardiello et al. 2014; Engel et al. 2012).

Tumors in LS patients are characterized by bi-allelic inactivation of one of the MMR genes by combining the germline mutation with a somatic hit, which leads to a hypermutator phenotype with the accumulation of large numbers of mutations. Most of these mutations occur in MS loci, thus giving rise to high levels of MSI (MSI-H) (Vilar and Gruber 2010; Yurgelun et al. 2012). Among the increased numbers of mutations, one of the most common types are frameshift insertion and deletion (indels). These indels generate tumor-specific mutated antigens known as neoAgs, which are processed into short peptides and presented on the cell surface of cancer cells by major histocompatibility complexes (MHC I and II) to T-cell receptors (TCRs) of CD8⁺ T-cells or CD4⁺ T-helper cells to stimulate natural killers, CD68⁺ macrophages, and other immune cells for cancer cell killing (Schwitalle et al. 2004; Hause et al. 2016). However, later activation of immune-checkpoint regulators allows malignant cells to evade this immune response and to continue growing uncontrollably towards tumor formation (Le et al. 2015). Thus, the development of immunotherapies to prevent and treat LS-derived tumors through re-activation of CD8⁺ and CD4⁺ T-cells that recognize neoAgs is an important field of study.

1.2. Opportunities for vaccine development

Several MS loci are highly prone to frameshift indels in MMRd neoplasms, and consequently, the probability of LS patients sharing the same mutations in these MS sites is high (Schwitalle et al. 2004; Woerner et al. 2003; Kloor and von Knebel Doeberitz 2016). Therefore, neoAgs generated from these recurrently mutated coding MS are distinctly common among multiple LS patients (Roudko et al. 2020; Ballhausen et al. 2020). For this reason, identifying and cataloging shared neoAgs based on their predicted immunogenicity and their affinities to different human leukocyte antigen (HLA) molecules is an important on-going effort by different research groups with the ultimate goal of developing a universal LS cancer-preventive vaccine.

The purpose of this review is to provide a summary of the biology behind carcinogenesis in LS, to present the different in silico approaches for neoAg discovery discussing their respective challenges and limitations, and to review the immunological assays for validating the immunogenicity of these predicted neoAgs. We also present the current state of LS vaccine development, including the past and on-going pre-clinical efforts and clinical trials testing different types of neoAg-based vaccines to highlight avenues for improvement and future directions.

2. The immune microenvironment of Lynch Syndrome cancers and pre-cancers

2.1. DNA mismatch repair deficiency, neoAg production and recognition by TCRs

During DNA replication, the DNA polymerase can accidentally allow slippage and introduce errors such as base-to-base mismatches and indels of nucleotides. This event is especially prone to occur in MS, which are repetitive sequences of mono-, di-, tri-, tetra- or pentanucleotide repeats that span across the genome (Olave and Graham 2022). The MMR system is evolutionarily conserved and corrects these nucleotide mismatches, thus maintaining the integrity of the genome and the stability of the MS sequences. The MMR system consists of a protein complex comprised of two heterodimers; the first dimer, MutSα, is formed by MSH6 and MSH2, which scans the newly replicated DNA strands as a sliding clamp. When it identifies a base mismatch or an indel, it recruits the second heterodimer MutLα formed by MLH1 and PMS2 (Figure 1A). The MutSα-MutLα complex then activates the downstream pathway to complete the correction of the error by involving several proteins, including exonuclease-1 (EXO1), DNA polymerase δ, single-stranded binding-factor replication protein A (RPA), proliferating cell nuclear antigen (PCNA), and non-histone chromatin component high-mobility group box 1 (HMGB1) (Randrian, Evrard, and Tougeron 2021; Jiricny 2006; Kunkel and Erie 2015).

Figure 1.

Comparison of the MMR mechanism in a normal cell vs a dysplastic/cancer cell.

In LS carriers, the loss of MMR activity, due to the inactivation of one of these key proteins (MLH1, MLH2, MSH6 or PMS2), will lead to the accumulation of indels throughout the genome. These indels cause these specific MS loci to either expand or contract the number of repeats, thus resulting in frameshifts of the open reading frame, indels of single amino acids, or single-nucleotide variants (SNVs), including nonsense and missense mutations (Turajlic et al. 2017). When mutated MS loci are located in a coding region, it can inactivate critical tumor suppressor genes, activate oncogenes, or generate a frameshift mutated neoAgs (Figure 1B) (Bohaumilitzky et al. 2020).

When MS with indels are translated by ribosomes in the cytoplasm of the cell, the resulting mutant proteins undergo proteolysis by the proteasome and then are processed into small peptides that act as mutated neoAgs. These small peptides are then taken into the endoplasmic reticulum (ER) by the Transporter-associated with Antigen Processing (TAP) protein, where they will be coupled with the MHC-I and β2-microglobulin (β2M) complex. The peptide-MHCI-B2M-bound complex is then exported to the cell surface. These neoAgs are unique to the MMR defective cell, different from the wild type protein, and foreign to the host. Therefore, once expressed on the cell surface, they are recognized by CD8⁺ T-cells expressing TCRs that are specific to that peptide-MHCI complex for the initiation and activation of cancer cell killing cascades (Mardis 2019) (Figure 2). CD4⁺ T-cells, which are activated in LS polyps and after therapeutic neoAg-vaccination, also play an important role during neoAg recognition (Chang et al. 2018; Kloor et al. 2020; Pastor and Schlom 2021). They communicate with CD8⁺ T-cells, using dendritic cells (DCs) as messengers, to reprogram them into effector and memory phenotypes (Schoenberger et al. 1998; Saxena et al. 2021). In addition, CD4⁺ T-cells can indirectly eradicate cancer cells through induction of cytotoxic macrophages (Bogen et al. 2019), and have also been shown to have direct antitumor effector activity through recognition of HLA class II-epitopes presented on cancer cells (Saxena et al. 2021).

Figure 2.

Steps necessary for neoAg presentation to the immune system on the cell surface via the MHC complex.

2.2. Immune architecture during tumorigenesis in LS

High production of neoAgs in MSI cancers leads to high infiltration of immune cells within the colonic mucosa. This infiltration is particularly prominent in LS-derived tumors of the colon, endometrium, and stomach (Hause et al. 2016; Cortes-Ciriano et al. 2017). Among the infiltrated immune cells, CD3⁺, CD8⁺ T-cells, CD68⁺, as well as Type 1 helper (Th1) cells are the most frequently observed (Galon et al. 2006; Walkowska et al. 2019) and they can be used as prognostic markers due to their connection to immune activation (Dahlin et al. 2011; Huh, Lee, and Kim 2012).

Evident immune responses occur even in LS healthy carriers who do not have a personal history of cancer (Schwitalle et al. 2008). This is explained by the fact that neoAg production can occur in pre-cancers (e.g. colonic adenomas), which leads to activation of CD8⁺ and CD4⁺ T-cells as well as proinflammatory cytokines like interferon gamma (IFNy) and tumor necrosis factor α (TNFα) (Chang et al. 2018; Koornstra et al. 2009). Furthermore, the normal mucosa of LS carriers without a personal cancer history (i.e., previvors) was shown to contain elevated numbers of CD3⁻, FOXP3⁻, and CD8⁺ T-cells compared to non-LS controls (Bohaumilitzky et al. 2022) due to the potential presence of histologically normal appearing colon crypts displaying MMRd (Kloor et al. 2012).

However, despite increased presence of immune cells during all stages of colorectal carcinogenesis, certain cancerous cells can still evade the immune system and continue their pathway towards cancer progression. One possible mechanism of immune evasion by dysplastic MMRd cells is through the acquisition of loss of function in antigen processing and presenting genes, such as B2M, which has been reported to be mutated in 17% of LS CRCs, and 20 – 60 % of all MMRd CRCs (Clendenning et al. 2018; Kloor et al. 2005; Zhang et al. 2022). B2M mutations have been associated with down-regulation of genes involved in antigen presentation MHC-I folding, assembly and loading, including HLA-F, UBE2D2, SEC31A and ITGB5 (Walkowska et al. 2019). Inactivating mutations in genes that regulate HLA class II antigen expression, like RFX5 and CIITA, have also been observed in more than 25% of MSI-H CRCs, which is another mechanism for immune evasion (Kloor and von Knebel Doeberitz 2016; Surmann et al. 2015). Immune resistance can also be achieved through the up-regulation of immune checkpoint molecules, including the anti-programmed cell death protein 1 (PD-1), the cytotoxic T lymphocyte associated antigen 4 (CTLA-4) and the lymphocyte activation gene 3 (LAG-3), which prevents the activation of the TCR on cytotoxic T-cells that recognize the peptide-MHC-I complex (Walkowska et al. 2019; Willis et al. 2020; Keir, Francisco, and Sharpe 2007; Llosa et al. 2015). Therefore, as the tumors progress from early to metastatic stages, the level of immune infiltration, particularly by CD8⁺ T-cells, decreases (de Miranda et al. 2012). Even though it has not been well studied, CD4⁺ T-cells have also been shown to acquire exhausted phenotypes in different cancer types, with increased expression of inhibitory receptors like PD-1, CTLA-4, LAG-3, and TIGIT (Miggelbrink et al. 2021). In CRC, the immune checkpoint T-cell immunoglobulin mucin-3 (TIM-3) was shown to induce T-cell exhaustion pathways in CD4⁺ tumor infiltrated lymphocytes (Sasidharan Nair et al. 2020).

Taken together, the pronounced immune activation during early stages of carcinogenesis not only makes LS carriers ideal candidates for immunotherapy, including checkpoint inhibitors for safe and durable effects (Therkildsen et al. 2021), but it is also in agreement with the application of immune interception strategies such as cancer preventive neoAg-based vaccines before immune evasion occurs and tumors become more aggressive.

3. Identification and prediction of neoAg repertoires

3.1. Methods for neoAg discovery

The first step towards the development of neoAg-based vaccines is the identification of specific antigenic peptide sequences, which should be immunogenic and recognized by T-cells in LS carriers. To do so, it is crucial to pinpoint the biochemical characteristics of the peptides that dictate how these are processed and presented to the immune system. The most common methodology currently used is performing whole-exome (WES) and mRNA sequencing of both the tumor and matching normal adjacent tissues and blood samples. This is followed by in silico analysis to identify the mutations and the putative neoAgs potentially generated from these mutations that are likely processed by mutated cells into short peptides, and have the ability to bind with the host MHC class I/II molecules, thus leading to TCR engagement and apoptosis (Verdon and Jenkins 2021).

Besides in silico-based prediction, there is also a direct method for identification of peptides bound to MHC through immunopeptidomics, which involves immunoprecipitation of HLA-peptide complexes from cancer/pre-cancer specimens followed by identification of amino acid sequences of eluted peptides using liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) in order to match MS/MS spectra against customized protein sequence databases. This methodology has successfully identified candidate neoAgs from mouse and human tumor cell lines and fresh tumor material (Yadav et al. 2014; Lu et al. 2022). Moreover, this technology allows the detection of naturally processed and presented HLA-bound peptides, including those with post-translational modifications, and can identify non-canonical or cryptic peptides derived from alternative open reading frames, novel exon-exon junctions, intronic sequences, long non-coding RNAs, and 5′ untranslated regions (Garcia-Garijo, Fajardo, and Gros 2019). Immunopeptidomics approaches can filter the list of candidate neoAgs to be screened, thus resulting in fewer false positives compared to in silico predictions. However, like any other experimental methods, it has its limitations and potential drawbacks. The main issue with the direct MS/MS-based approach is the sensitivity for identifying low and underrepresented peptides (Abelin et al. 2017). This method requires a large amount of sample input and the signals generated by low-abundance peptides may not be strong enough to be detected by the mass spectrometer, thus resulting in several putative immunogenic peptides being undetected. Another potential problem is MS-based analysis of MHC-bound peptides may miss some peptides and fail to identify crucial epitopes due to the highly polymorphic nature of MHC molecules, which can lead to certain peptides not having a strong binding affinity to the specific MHC molecule being analyzed (Laumont et al. 2016). Other methods such as immunoprecipitation, fractionation, and enrichment strategies, complementary to direct MS-based analysis, have been employed to increase the sensitivity of peptide detection and identification (Ma and Johnson 2012; Faridi, Purcell, and Croft 2018).

3.2. Current bioinformatics tools for neoAg discovery and their challenges in LS

3.2.1. Sequencing data analysis for mutation calling

A prototypical neoAg discovery pipeline uses either WES or whole-genome sequencing (WGS) data from paired tumor and normal tissues (Figure 3). After data quality control and alignment to the reference genome, mutation calling is an important step, which relies on utility of several software tools, including Mutect, VarScan2, VarDict, SomaticSniper, and Muse, among others (Cibulskis et al. 2013; Fan et al. 2016; Larson et al. 2012; Lai et al. 2016; Reble et al. 2017). In silico pipelines previously used to predict neoAgs from MMRd tumors have primarily combined mutation calls derived from Mutect, VarScan2, Somatic Sniper, and Muse (Roudko et al. 2020).

Figure 3.

Prototypical steps of a neoAg prediction pipeline. Blue boxes represent input/output data, and green boxes represent the analysis to be performed with specified bioinformatics tools options.

3.2.2. HLA typing and epitope prediction

After mutation calling, the next step is HLA typing, which can be performed from WES or RNAseq data. The most common tools used are PHLAT and Optitype (Bai, Wang, and Fury 2018; Szolek et al. 2014) followed by application of prediction of tumor-specific neoAgs on the basis of epitope-HLA allele interactions from the entire pool of mutated peptides. Among several algorithms, the most commonly used for HLA-epitope affinity are NetMHCpan, MHCflurry, MHCnuggets, and VaxRank (Reynisson et al. 2020; Rubinsteyn et al. 2017; Shao et al. 2020; O'Donnell et al. 2018). These tools use different types of deep learning models to predict the epitope and neoAg affinity to specific HLA alleles. To date, NetMHCpan has the been the main algorithm used for the prediction of neoAgs in MMRd and LS-derived tumors (Roudko et al. 2020; Leoni et al. 2020; Ballhausen et al. 2020).

3.2.3. Immunogenicity ranking of the predicted neoAgs

Finally, predicted neoAgs should be ranked based on different factors including expression, variant allele frequency (VAF) of the mutation, MHC binding affinity, stability of the neoepitope–MHC complex, and level of dissimilarity compared to their wild-type counterparts (foreignness). Different pipelines that combine all the above steps, from sequencing data processing and quality control to neoAg ranking, are publicly available. Some of the most widely used are pVAC-Seq, Antigen.garnish, pTuneos, and Openvax (Hundal et al. 2016; Kodysh and Rubinsteyn 2020; Zhou et al. 2019; Richman, Vonderheide, and Rech 2019). However, none of these publicly available pipelines have been developed strictly to predict neoAgs from frameshift mutations in MSI-H cancers. For this purpose, it is important to consider modifying the upstream sections of the pipeline and some of the implemented functions within the different packages that constitute it to allow for an efficient detection of frameshift mutations, from which neoAgs can be correctly generated for subsequent prediction of their immunogenic features.

It is important to note that these tools are very accurate at predicting neoAg presentation by MHC complexes and their affinities to different HLA alleles. However, for a neoAg to be immunogenic it should also be effectively recognized by TCRs to trigger immune responses. To date, neoAg prediction efforts applied to LS tumors lack the inclusion of tools that can predict if the neoAg has matched cognate TCR sequence. Recent bioinformatics methods are just now emerging, due to the complexity and plasticity of these mechanisms, to predict TCR-epitope pairing (Singh et al. 2017; Xie et al. 2023). The in silico tool most commonly used for this purpose is NetCTL/NetCTLpan, which combines MHC binding, C-terminal cleavage affinity and the transporters associated with antigen processing (Stranzl et al. 2010). However, several groups are applying machine learning algorithms to generate a more robust methodology tackle this problem (Raybould et al. 2022; Hudson et al. 2023).

3.2.4. Challenges of neoAg prediction from MSI tumors

It is important to note that mutations in MS can be challenging to detect using next generation sequencing (NGS) technologies due to the difficulty in distinguishing true homopolymer repeat changes from sequencing and PCR amplification errors as well as other sources of noise (Treangen and Salzberg 2011). To tackle this problem, Ballhausen et al. proposed the use of DNA sequencing by fragment analysis, targeting mutations in 41 coding microsatellites that are commonly mutated in MMRd tumors followed by fragment length in silico analysis to detect the mutations, and prediction of neoAg-HLA affinity using NetMHCpan (Ballhausen et al. 2020). Nonetheless, there is an unmet need to improve current neoAg detection and prediction pipelines for the detection of indels within MS loci and subsequent prediction of frameshift peptides derived from these MS indel events.

3.3. Landscape of recurrent neoAgs predicted in LS

Prediction of neoAgs in LS-derived and sporadic MMRd tumors has shown that several coding MS (cMS) loci are highly prone to acquire mutations, and the probability of the same mutation occurring at the exact same locus in tumors from different patients is very high compared to microsatellite stable (MSS) tumors (Roudko et al. 2020; Ballhausen et al. 2020; Schwitalle et al. 2004). This leads to a high percentage of neoAgs generated from these shared mutations to be highly recurrent. Table 1 shows the list of conserved neoAgs in LS, with their respective frequency among the LS population, which have been confirmed for their immunogenicity using in vitro assays with human samples. These results support the feasibility of the development of a universal cancer-vaccine for MMRd and LS-related tumors.

Table 1.

Conserved neoantigens in LS tumors with immunogenicity validated in human samples. Freq, frequency.

Gene	Neopeptide sequence	Freq (%)	Organ	HLA restriction	References
CASP5	QLRCWNTWAKMFFMVFLIIWQNTMF	46	Colon	HLA-A*02-01	(Schwitalle et al. 2008; Schwitalle et al. 2004; Reuschenbach et al. 2010)
OGT	SLYKFSPFPLPPFPPIFFH	22	Colon	HLA-A*02-01	(Schwitalle et al. 2008; Ripberger et al. 2003; Linnebacher et al. 2001)
Sec63	NLHLCYYHSQSNRNKSRQMESLGMKLQ	62	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
TGFBR2	SLVRLSSCVPVALMSAMTTSSSQKNITPAILTCC	76	Colon	HLA-A*02-01	(Schwitalle et al. 2008; Bauer et al. 2013; Linnebacher et al. 2001; Reuschenbach et al. 2010)
TGFBR2	KCIMKEKKSLVRLSSCVPVALMSAMTTSSSQKNITPAILTCC	54-88	Colon, Endometrial	HLA-A*02-01	(Ballhausen et al. 2020)
TAF1B	TILKKAGIGMCVKVSSIFFINKQKP	78	Colon	HLA-A*02-01	(Schwitalle et al. 2008; Kloor et al. 2020; Bauer et al. 2013; Reuschenbach et al. 2010)
ASTE1/HT001	GRRNRIPAVLRTEGEPLHTPSVGMRETTGLGC	10-85	Colon, Endometrial, Stomach	HLA-A*02-01	(Schwitalle et al. 2008; Roudko et al. 2020; Bauer et al. 2013; Kloor et al. 2020)
MSH3	RATFLLALWECSLPQARLCLIVSRTLLLVQS	40	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
AIM2	HREVKRTNSSQLV	56	Colon	HLA-A*02-01	(Schwitalle et al. 2008; Bauer et al. 2013; Kloor et al. 2020; Reuschenbach et al. 2010)
U79260	LRHSLTLSPGWSAVARSRL	81	Colon	HLA-A*02-01	(Schwitalle et al. 2008; Linnebacher et al. 2001)
U79260	GWSAVARSRLTATSASQVQV	81	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
U79260	TATSASQVQVILLPQPPEWL	81	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
U79260	ILLPQPPEWLGLQARAAAPS	81	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
AC1	SCQLNLGRKEHAKIFTFF	67	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
AC1	KEHAKIFTFFFQLDTMDGNP	67	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
AC1	FQLDTMDGNPGELTLELQTL	67	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
AC1	GELTLELQTLQIKQSQNALL	67	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
AC1	QIKQSQNALLPAGPLTQTPV	67	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
ZNF294	MVRLDLLMRYLKAIKRMKNVYLQKERRLKAGN	52	Colon	HLA-A*02-01	(Schwitalle et al. 2008; Reuschenbach et al. 2010)
FLJ20378	ETEFCSCCPGWSAVAQSWLTATSTSRVQAILLPQPPE	88	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
DAMS	KAPAGQETLSLQSR	85	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
DAMS	KLQLVRKHCLYNQDKHLAQGVQPFLTYRLA	85	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
UST3	NLLCVKCSTCPTYVKGSPSCPLRDLQ	70	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
UST3	SPSCPLRDLQTLWPILALISMSSIWGTMFS	70	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
UST3	MSSIWGTMFSCCRLSLVQSSSWPTVLHLGH	70	Colon	HLA-A*02-01	(Schwitalle et al. 2008)
MARCKS	RSAFPSRSLSS	73	Colon	HLA-A*02-01	(Reuschenbach et al. 2010)
LTN1	SLKSSKKKMVRLDLLMRYLKAIKRMKNVYLQKERRLKAGN	71-83	Colon, Endometrial	HLA-A*02-01	(Ballhausen et al. 2020)
SLC35F5	VAKISFFFALCGFWQICHIKKHFQTHKLL	22-67	Colon, Endometrial, Stomach	HLA-A*02-01	(Ballhausen et al. 2020; Roudko et al. 2020)
SLC22A9	ELEAAQKKNLLCVKCSTCPTYVKGSPSCPLRDLQTLWPILALISMSSIWGTMFSCCRLSLVQSSSWPTVLHLGH	36-76	Colon, Endometrial, Stomach	HLA-A*02-01	(Ballhausen et al. 2020; Roudko et al. 2020)
SLC22A9	ELEAAQKKTFSV	50-76	Colon, Endometrial	HLA-A*02-01	(Ballhausen et al. 2020)
TTK	SSKTFEKKGEKNDLQLFVMSDTTYKIYWTVILLNPCGNLHLKTTSL	41-50	Colon and Endometrial	HLA-A*02-01	(Ballhausen et al. 2020; Roudko et al. 2020)
MYH11	KSKLRGPPHRKLRSDAPGEETRPLSFLLEGLEDVELLKMQMVLRRKRTLETQTSMEPRPVNKQLSTVLHHGKKTKNQNKQTKKTQQQPRTKQNPADCT	29-40	Colon, Endometrial	HLA-A*02-01	(Ballhausen et al. 2020)
SEC31A	INYCQKKLMLLRLNLRKMCGPF	36-74	Colon, Endometrial, Stomach	Unspecified	(Roudko et al. 2020)
SETD1B	MENSHPPTTTTSSPRRSPALRARGGTTIGEVTS	42-64	Colon, Endometrial	Unspecified	(Roudko et al. 2020)
OR7E24	MSYFPILFFFSSKGVRATQSHRISQVSQNSSSWDSQRIQNCSRSSLGCSCPCTWSRCWGTCSSSWLSALTPTSTPPCTSSSPTCPWLTSVSPPPRSPR	8-29	Colon, Endometrial, Stomach	Unspecified	(Roudko et al. 2020)
RNF43	PQRKRRGVPPSPPLALGPRMQLCTQLARFFPITPPVWHILGPQRHTP	36-62	Colon, Endometrial, Stomach	Unspecified	(Roudko et al. 2020)
CKAP2	SLMEQIPHL	4 - 28	Colon, Endometrial, Stomach	HLA-A*02-01	(Leoni et al. 2020)

Of note, the nonsense-mediated RNA decay (NMD) pathway has been shown to play an important role in regulating the mutational profile of malignancies by preferentially suppressing the expression of specific mutations through degradation of mRNA encoding certain immunogenic neoAgs, primarily those with inhibitory function in expression of tumor suppressor genes. This opens the possibility of potential use of NMD-modulators against different tumor types, including those with high neoAg burden, like LS-derived tumors (Tan, Stupack, and Wilkinson 2022; Xie et al. 2023).

4. In vitro validation of immunogenicity of predicted neoAgs

Although recent advances in systems biology tools and immunopeptidome analyses have been very useful in the identification and prediction of candidate neoAgs, it is still essential to validate their immunogenicity and potential T-cell activation using in vitro immunological assays. Therefore, the optimal approach for neoAg discovery involves a combination of genomic, proteomic, validation and functional steps.

4.1. ELISpot, cytokine, and immune cell profiling for immunogenicity of neoAgs

T-cells isolated from patients or healthy donors are commonly used to assess the immune response to a candidate neoepitope through in vitro immunological assays for cytokine levels, T cell-mediated apoptosis, and T-cell profiling. To stimulate T-cells in vitro, they can be cultured with antigen-presenting cells and cytokines or exposed to synthetic peptides that mimic the target neoAg. The gold standard method for assessing antigenic response is the enzyme-linked immunosorbent spot (ELISpot) assay, which directly measures T-cell recognition of a candidate neoAg and production of secreted IFN-γ that can be quantitively measured by ELISAs or IFN-γ-secreting cells as spot forming units using an ELISpot plate reader. This method has been used in the past for validation of immunogenic neoAgs in various types of cancers including colorectal, melanoma, NSCLC, and glioblastoma (Tran et al. 2016; Zheng et al. 2022; Gebert et al. 2021; Hilf et al. 2019; Podaza et al. 2020). In some cases, T-cells from a patient with a high mutational burden can be used directly without in vitro stimulation and expansion of neoAg-specific T-cells to test the immunogenicity of neoAgs. This approach, known as ‘direct ex vivo’ analysis, involves isolating T-cells from a patient to screen a library of potential neoAgs for their reactivity (Fehlings et al. 2019; Shelton et al. 2021). Both in vitro stimulation and direct ex vivo assays present distinct benefits and drawbacks. While in vitro stimulation enables the proliferation of T-cells and potentially yields a stronger immune response to the antigen of interest, it may not accurately reflect the patient's physiological immune response. In contrast, direct ex vivo analysis offers an alternative to in vitro stimulation assays, allowing for the identification of T-cell responses in their natural state, wherein resident memory T-cells interact with neoAgs, thus triggering cytokine secretion that may accurately represent the in vivo immune response to a tumor (Albert-Vega et al. 2018).

To replicate the complex microenvironment of malignancies, researchers have developed more sophisticated methods for assessing the immunogenicity of candidate neoepitopes, such as multiplex cytokine assays and immune cell profiling (Keskin et al. 2019). These techniques allow for the measurement of multiple cytokines simultaneously such as IFN-γ, TNF-α, IL-2, and IL-10, and the identification of specific immune cell populations such as T-cells, natural killer cells, DCs, and macrophages in response to a neoAg. Furthermore, recent advances in single-cell genomics have enabled the characterization of the TCR repertoire of neoAg-specific T-cells, providing valuable insights into the clonality and diversity of the immune response towards a specific neoepitope (Lu et al. 2021; Lowery et al. 2022).

4.2. Clonal expansion of neoAg-specific TCR expressing T-cells

To consider neoAg-based vaccine development, the identification of candidate neoAg and their cognate TCRs capable of eliciting an immune response and promoting antitumor immunity is crucial. In this context, the primary focus of cancer vaccines and cancer prevention strategies is centered around stimulating a cytotoxic T-lymphocyte (CTL) response against the tumor. However, it is important to note that both CD8⁺ and CD4⁺ T-cells are essential players in recognizing neoAgs and thereafter ensuing immune response.

A commonly employed method for isolating cytotoxic CD8⁺ and CD4⁺ T-cells specific to neoAg involves the use of fluorescently labeled multimeric MHC molecules, such as tetramers, pentamers, and dextramers. These molecules exhibit high affinity and specificity for TCRs on the surface of T-cells. It is worth noting that the application of peptide-MHC class I (pMHC-I) multimers that are used to specifically stain CD8⁺ T-cells, is more prevalent than pMHC-II multimers. This method allows for the precise identification and isolation of neoAg-specific T-cells through fluorescence-activated cell sorting (FACS) analysis (Dolton et al. 2018; Chang 2021). Barcoded multimer technologies are also an innovative approach for the characterization of neoAg-specific T-cells. By using DNA barcodes attached to multimers, researchers can track individual T-cells and measure their response to neoAgs with high specificity and sensitivity (Bentzen et al. 2016). These technologies can be further enhanced by combining them with single-cell genomics tools to identify the clonal population of expanded T-cells (Ma et al. 2021). In addition, neoAg-specific T-cells are analyzed for proliferation, cytokine production, or expression of activation markers like 4-1BB/CD137, OX40 (CD134) or CD25 (Leko et al. 2021; Halstead et al. 2002; Watts 2005). OX40 is primarily expressed on CD4⁺ T-cells. On the other hand, CD137 expression is highly restricted to transiently activated CD8⁺ T-cells, but can also be detected on stimulated CD8⁺ T-cells of all phenotypes, including naïve T-cells and early and late memory effector T-cells (Wolfl et al. 2007).

4.3. Functional validation of neoAg-specific T-cells for killing assays

Activated cytotoxic T-cells following exposure to neoAgs through engagement with TCR should be able to trigger apoptosis of neoAg-expressing dysplastic and cancer cells, which is a determining factor for the use of neoAgs in vaccine development. Various model systems have been developed to study T-cell-mediated killing, such as transgenic mouse models and patient-derived xenograft (PDX) models. PDX models with PIK3CA-mutant tumors have been used to investigate immunogenicity and therapeutic targeting of neoAgs in complex tumor microenvironments (Chandran et al. 2022). However, these models have limitations in fully recapitulating human LS-derived cancers. Our group has demonstrated a potential alternative model system in rhesus macaques with a germline MLH1 mutation, which spontaneously develop MMRd CRC and share genomic similarities with human LS and sporadic MMRd colorectal tumors (Ozirmak Lermi et al. 2022). This model system could be useful in studying immune response to neoAgs and testing new therapies for LS-associated cancers. Microfluidic models, on the other hand, create controlled microenvironments that mimic aspects of neoAg-presenting cells, thus allowing for a more physiologically relevant context for studying T-cell behavior (Shelton et al. 2021; Duzagac et al. 2021). In vitro culture systems can also be used to study T-cell-mediated killing of neoAg-presenting cells, such as DCs or other antigen-presenting cells (Zhu et al. 2021). The choice of model depends on the research question being asked and the resources available for the study. However, it is crucial to select a model that can best mimic the in vivo context of neoAg-based T-cell-mediated killing.

5. Current state of cancer vaccines and interception strategies in LS

Since the late 1800s, with the discovery of William Coley that bacterial infections after surgery led to tumor regression, the concept of cancer vaccines has evolved substantially (Coley 1891). Researchers have since focused on developing targeted cancer vaccines, including those aimed at inducing a specific immune response against predictable neoAgs. This approach has been particularly explored in cancers with a high mutational burden, such as MMRd tumors (Gambini, Ferrero, and Kuhn 2022). LS, resulting from MMRd, is an ideal population for developing preventive cancer vaccines due to the high mutational burden present even in pre-cancers and the predictability of induced mutations, which are usually generated from frameshift mutations in MS loci. LS-associated cancers are prone to immune evasion, indicating the need for immune interventions to augment immune surveillance and enhance the elimination of cancer cells or precancerous cells that present neoAgs (Bohaumilitzky et al. 2020). Different studies suggest that immune surveillance plays a significant role in controlling and suppressing the outgrowth of LS-associated tumors (Willis et al. 2020; Bohaumilitzky et al. 2020; Muller, Yurgelun, and Kupfer 2020). The immune system exposure towards frameshifted peptides (FSPs), which generates foreign amino acid sequences, can provide direct clues to the presence of non-neoplastic or early dysplastic MMRd cells in LS carriers (Willis et al. 2020). Since potent interventions such as immune checkpoint blockade (ICB) are not suitable for healthy, tumor-free individuals due to unacceptable side effects, alternative approaches such as non-steroidal anti-inflammatory drugs (NSAID) used as immune stimulant together with vaccines are needed to improve a durable natural immune stimulation present in LS carriers.

5.1. NeoAg vaccine types and delivery strategies in pre-clinical and clinical trials

NeoAg vaccines encompass a diverse range of molecular types, including RNA, DNA, protein, and peptides. RNA vaccines encode target antigens, which are incorporated by cells for expression; while DNA vaccines transcribe and translate DNA molecules into antigenic proteins (Hernandez-Sanchez et al. 2022; Dolgin 2021). These DNA and RNA platforms are considered inherently more immunogenic, as these molecules naturally activate the innate immune system. DNA vaccines are designed to be retained in the host nucleus without integration into the host genome, while mRNA is directed to the cytoplasm for translation. Protein and peptide neoAg vaccines directly deliver pre-made antigens or short peptides matching the neoAg sequence (Ottensmeier and Savelyeva 2017). Among these types, peptide-based vaccines are the most widely used in clinical trials. They offer the advantage of being easy to manufacture and can be designed for specific neoAgs. On the other hand, viral vector vaccines are a type of vaccine that uses a non-pathogenic and non-immunogenic virus, such as adenovirus, as a carrier or vector to deliver genetic material from a target virus into human cells (D'Alise et al. 2019). These vaccines have the advantage of efficient antigen delivery, endogenous post-translational modification, and may elicit strong memory responses towards viral encoded antigens. However, they can also cause immune interference (Liu, Liang, and Huang 2021). Researchers have discovered that an adenovirus-based vaccine that targets tumor neoepitopes can promote the stemness of neoAg-specific CD8⁺ T-cells, thus leading to tumor rejection and enhanced responses to anti-PD-1 therapy by improving immunogenicity and antitumor efficacy. The vaccine was first tested on murine models, which showed an increase in the number of polyfunctional neoAg-specific CD8⁺ T-cells and an accumulation of Tcf1⁺ stem-like progenitors in draining lymph nodes and effector CD8⁺ T-cells in tumors. The vaccine efficacy was further confirmed in the first 12 patients enrolled into a First-in-Human Phase I clinical trial for patients with metastatic MMRd gastrointestinal (GI) tumors, where the expansion and diversification of TCRs were observed in post-treatment biopsies of patients with clinical response (D'Alise et al. 2022).

Adjuvants, immunostimulatory agents, may be necessary to enhance the immune response elicited by the vaccine. This is particularly important for peptide vaccines, which can sometimes induce unpredictable, weak, or temporary immune responses. Adjuvants such as polyinosinic-polycytidylic acid and poly-L-lysine (poly-ICLC) are used to augment the immune response as it mimics double-stranded RNA and activates Toll-like receptor 3, thus triggering an immune response (Ott et al. 2020; Liu et al. 2018). In addition, sustaining the desired immune response may require multiple administrations of the vaccine.

The molecular composition of neoAg vaccines is not the only factor determining their efficacy; the delivery method also plays a significant role. To maximize their effectiveness, researchers are actively exploring advanced delivery systems, including liposomal and nanoparticle approaches. These innovative systems provide several advantages, such as safeguarding vaccine components and promoting their uptake by immune cells (Hodge et al. 2020). One notable example is the use of lipid nanoparticle-encased pseudouridylated mRNA as an adjuvant, which not only stimulates T-helper and germinal B-cells but also protects the mRNA from degradation by extracellular RNases (Wang et al. 2021). These mRNA-based vaccines exhibit the potential to elicit strong immune responses while carrying a minimal risk of genomic integration. They also allow for post-translational modifications of antigens and have a larger protein cargo capacity (Wang, Wang, and Liu 2022). Consequently, liposomal mRNA and RNA vaccines have gained considerable popularity within the field of vaccination. Clinical trials have been conducted on various liposomal RNA vaccines, such as FixVac/bnt111 and personalized RNA mutanome vaccines that target individual neoAgs expressed in a patient's tumor, showing promising results in inducing CD4⁺ and CD8⁺ T-cell immunity against the vaccine antigens and causing durable objective responses (SE and Pharmaceuticals 2021; Sahin et al. 2017; GmbH and SE 2013). Autogene Cevumeran (RO7198457) is another well-known RNA-based neoAg lipid nanoparticle vaccine in development, currently undergoing phase I clinical trials for various solid tumors including colorectal, bladder, and renal cancers (Genentech and SE 2017). The vaccine encodes tumor-specific neoAgs using synthetic long peptides and mRNA to encode neoAgs specific to a patient's tumor, delivered to antigen-presenting cells (APCs) to trigger an immune response. The trial is also assessing the optimal dose and schedule of the vaccine as well as its effectiveness in combination with other immunotherapies.

DC vaccines, based on antigen-presenting cells, have shown promise as a delivery strategy as well. However, they face challenges in entering the HLA-I pathway and presenting antigens effectively to CD8⁺ T-cells in vivo (Xiong et al. 2022). Researchers have explored DC vaccines with extracellular vesicles (EVs) derived from tumor cells to overcome limitations in the recognition of neoepitopes by T-cells. EVs, acting as versatile carriers, deliver a diverse repertoire of tumor antigens, including neoaAgs, to DCs, improving antigen delivery and enhancing cross-presentation (Zhou et al. 2022; Wylie et al. 2021). This integration of EVs with DC vaccines offers a promising solution to efficiently enter the HLA-I pathway and optimize antigen presentation to CD8⁺ T-cells, ultimately enhancing the efficacy of neoAg-based vaccines. Ongoing clinical trials in patients with LS or MSI CRC aim to enhance the immune response against well-known tumor-associated antigen carcinoembryonic antigen (CEA) and specific frameshift-derived neoAgs, providing valuable insights for DC-based neoAg vaccines (NCT. 2010). An overview encompassing the predominant neoAg vaccine types, including their associated delivery platforms, while highlighting the specific advantages and disadvantages, is summarized in Figure 4.

Figure 4.

NeoAg vaccine types and delivery platforms with their advantages and disadvantages.

To maximize the effectiveness of neoAg vaccines, it is important to include epitopes that are recognized by both CD8⁺ and CD4⁺ T-cells. The inclusion of these "helper" epitopes in therapeutic cancer vaccines has demonstrated remarkable improvements in antitumor responses, such as enhancing CD8⁺ T-cell expansion, reducing coinhibitory receptor expression, and increasing migratory potential (Zhang et al. 2019). To ensure effective cross-presentation, physically linking the epitopes for delivery to the same APC is speculated to be crucial, as CD4⁺ T-cell help is necessary for cross-presentation. Targeting epitopes across multiple HLA alleles is also important to minimize immune escape and relapse after immunotherapy due to loss of heterozygosity. These factors, encompassing CD8⁺ and CD4⁺ T-cell recognition and targeting multiple HLA alleles, play a critical role in tailoring neoAg vaccines (Hodge et al. 2020).

Combining vaccines with other chemoprevention strategies is another promising approach to prevent cancer by targeting both the immune system and cellular pathways associated with cancer development. NSAIDs, especially aspirin (ASA), have been widely studied for their potential in GI cancer prevention, reducing the production of prostaglandin E2 (PGE2) by inhibiting the cyclooxygenase 1 and 2 enzymes (Burn et al. 2011). ASA is effective in reducing LS -associated tumor penetrance and is widely used for GI cancer prevention (Burn et al. 2020). However, a recent clinical trial raised questions regarding the overall pan-cancer rates and mortality among healthy elderly populations, thus indicating the potential of ASA to increase these rates (McNeil et al. 2021). Recent studies have shown that Naproxen (NAP), a propionic acid NSAID, has greater cancer-preventive activity than ASA in LS mouse models, thus suggesting potential chemopreventive benefits of NAP regimen over ASA (Martin-Lopez et al. 2018). Our group conducted a Phase Ib, placebo-controlled, randomized clinical trial of two dose levels of naproxen sodium (220 and 440 mg) administered daily for 6 months in 80 LS carriers (Reyes-Uribe et al. 2021). The results showed that the level of PGE2 in the colorectal mucosa was significantly reduced after treatment with naproxen compared to placebo. Moreover, naproxen was found to activate various resident immune cell types, without increasing lymphoid cellularity, and changed the expression patterns of the intestinal crypt towards epithelial differentiation and stem cell regulation. Pre-clinical animal studies have demonstrated the efficacy of combining NAP with a mouse-specific, recurrent FSP vaccine in reducing tumor burden in a LS mouse model (Gebert et al. 2021). In this study, four shared FSP neoAgs were identified and shown to induce CD4⁺/CD8⁺ T-cell responses in mice. Vaccination with the combination of these FSPs significantly enhanced adaptive immunity and prolonged overall survival in mice at-risk of intestinal cancer. Importantly, when combined with daily naproxen, the FSP vaccine exhibited even greater efficacy in terms of immune response, tumor growth delay, and survival. These findings highlight the potential of combining FSP vaccination with NSAIDs like naproxen to enhance the immune response and effectively reduce the tumor burden in LS.

5.2. Clinical and pre-clinical trials evaluating neoAg vaccination in LS and MMRd-cancers

Cancer vaccines have not been as successful as vaccines against infectious diseases as cancer cells have many ways to evolve and evade immune attacks (Goldman and DeFrancesco 2009). The success of an immune response depends on several critical steps, including antigen presentation, T-cell activation, migration to the tumor site, and killing of cancer cells (Chen and Mellman 2013; Saxena et al. 2021). However, preventive vaccination before the onset of an immunosuppressive tumor microenvironment allows for the targeting and elimination of arising tumor cells. Preventive vaccines to enhance the immune system's capacity to detect and attack pre-cancer cell clones hold high potential for reducing tumor incidence in LS carriers (Kloor and von Knebel Doeberitz 2016). Evidence of this potential can be seen in the immunogenicity of certain neoAgs derived from the MMRd-induced landscape of FSP, as demonstrated in a canine clinical trial of a preventive vaccine (Katsnelson 2021). Local immune responses in LS-associated CRC, adenomas (pre-cancers), and even entirely normal-appearing colonic mucosa have also been demonstrated (Chang et al. 2018; Bohaumilitzky et al. 2022). Such findings suggest that the adaptive immune system plays a vital role in the variable penetrance of the disease by eliminating precancerous MMRd cells.

A first-in-human clinical trial investigated the safety and immunological effectiveness of a FSP vaccine in 22 participants with history or current MMRd colorectal cancer (GmbH and KG 2011). The vaccine was based on a combination of three recurrent frameshift-derived neoAgs: AIM2, HT001 and TAF1B. All participants who received the full cycle of vaccinations developed pronounced humoral and T-cell responses against the vaccine peptides. T-cell responses were mainly CD4⁺ responses, while CD8⁺ responses were detected in 9 out of 22 patients, indicating that certain HLA class I genotypes were required to develop CD8⁺ responses. No systemic vaccine-related side effects occurred in vaccinated patients, although the vaccine induced local side effects that are typical from vaccinations. Our collaboration with Nouscom, s.r.l. resulted in the development of a Phase I in LS carriers using Nous-209, an adenoviral-vaccine encoding 209 shared frameshift peptide antigens against MMRd tumors (Institute 2022). These antigens were selected based on the analysis of frameshift mutations that occur within cMSs in MMRd colorectal, gastric, and endometrial cancers compared to adjacent normal samples from the cancer genome atlas (TCGA) (Leoni et al. 2020). The selection criteria included an antigen prevalence of ≥5% for each tumor type, while being present in ≤25% of alleles in matched healthy tissues and ≤2% of alleles in healthy tissue samples. Additionally, the selected neoAgs were required to have a length of ≥8 amino acids and to map to protein coding regions A. Phase I clinical trial combining the Nous-209 vaccine with anti-PD-1 antibody pembrolizumab has been recently completed in patients with metastatic colorectal cancer, gastric and gastro-esophageal cancer (SRL 2019). The trial enrolled a total of 21 participants in the US, and short-term results show that out of 12 evaluable subjects, 7 had partial responses, 2 disease stability, and 3 progression. The combination treatment was reported to be safe, highly immunogenic, and demonstrated promising early signs of clinical efficacy with no dose-limiting toxicities (D'Alise et al. 2022; Overman et al. 2021). However, because the study used combination therapy, it is unclear how much of the response was attributable to pembrolizumab vs. Nous-209. A list of clinical trials discussed in this section and their main characteristics are summarized in Table 2.

Table 2.

Summary of clinical trials assessing the efficacy of neoAg vaccines.

Clinical trial	Phase	Agent	Disease	Delivery platform	NCT ID	Status	Ref.
Cancer Preventive Vaccine Nous-209 for Lynch Syndrome Patients	Ib/II	GAd-209-FSP MVA-209-FSP	LS	Adenoviral Tumor-specific neoAg Priming Vaccine and MVA Tumor-specific neoAg Boosting Vaccine	NCT05078866	Recruiting	N/A
Nous-209 Genetic Vaccine for the Treatment of Microsatellite Unstable Solid Tumors	I/II	GAd-209-FSP MVA-209-FSP	Solid Tumor, Adult	Adenoviral Tumor-specific neoAg Priming Vaccine and MVA Tumor-specific neoAg Boosting Vaccine	NCT04041310	Recruiting	N/A
INO-5401 + INO-9012 in Combination with Atezolizumab in Locally Advanced Unresectable or Metastatic/Recurrent Urothelial Carcinoma	I/II	INO-5401 INO-9012 Atezolizumab	Locally Advanced Unresectable or Metastatic/Recurrent Urothelial Carcinoma	CELLECTRA™ 2000	NCT03502785	Active, not recruiting	N/A
Messenger RNA (mRNA)-Based, Personalized Cancer Vaccine Against neoAgs Expressed by the Autologous Cancer	I/II	NCI-4650	Melanoma, CRC, GI Cancers	N/A	NCT03480152	Terminated	(Cafri et al. 2020)
A Study of Autogene Cevumeran (RO7198457) as a Single Agent and in Combination With Atezolizumab in Participants with Locally Advanced or Metastatic Tumors	I	Autogene cevumeran with Atezolizumab	Locally Advanced or Metastatic Tumors (Melanoma, NSCLC, Bladder Cancer and others)	N/A	NCT03289962	Active, not recruiting	N/A
IVAC MUTANOME Phase I Clinical Trial	I	IVAC MUTANOME	Melanoma	N/A	NCT02035956	Completed	(Sahin et al. 2017)
Dendritic Cell Vaccination in Patients With Lynch Syndrome or Colorectal Cancer With MSI	I/II	DC vaccination	CRC	Peptide-loaded DC	NCT01885702	Active, not recruiting	N/A
Vaccination Against MSI Colorectal Cancer	I/II	FSP-vaccination	CRC	N/A	NCT01461148	Completed	(Kloor et al. 2020; Reuschenbach et al. 2014)

6. Future directions

Although neoAg-based cancer vaccines hold great promise, the success of this approach is hindered by several factors such as weak immunogenicity to certain neoAgs, mutational heterogeneity of tumors, and inter-individual variability in the expression of neoAgs on MHC molecules. This inter-individual variability in the expression of neoAg on MHC molecules poses a significant challenge in designing effective neoAg-based vaccines that can serve a wide range of patients, as the unique set of presented neoAgs can result in variable vaccine efficacy across patients. However, advances in single-cell sequencing and bioinformatic approaches are allowing more accurate profiling of the most frequently recurring and shared mutated neoAgs in LS-associated tumors, thus identifying the most immunogenic for incorporation into vaccine platforms. NeoAg-based vaccines have also shown promising results in the treatment of GI cancers. However, there are several future directions that need to be explored in order to improve their efficacy and applicability. First, identifying clinically relevant neoAgs that elicit a strong and specific immune response against tumors, which can be achieved through advanced sequencing technologies to analyze tumor tissues and match them with HLA types. Secondly, finding the optimal vaccine formulation and dosing strategy, while minimizing toxicity, which can be achieved by conducting clinical trials with larger sample sizes. All these improvements require extensive genomic profiling efforts of tumors to identify the appropriate neoAgs. Lastly, overcoming immunosuppression in the tumor microenvironment through combining neoAg vaccines with other immunoprevention strategies, such as NSAIDs, can enhance their effectiveness by promoting T-cell activation and overcoming immune checkpoints.

In summary, neoAg-based vaccines hold great promise for the treatment of MMRd and LS-associated tumors. However, further research is needed to optimize their efficacy and safety. To target individual tumors, personalized approaches are still needed to identify clinically relevant neoAgs, but it is crucial to explore inclusive approaches that target shared neoAg present in multiple patients with the same type of cancer or disease as these can potentially provide a more universal and effective solution. By targeting shared neoAgs, neoAg-based vaccines could potentially benefit a larger patient population and enable more efficient vaccine development. Moreover, combining neoAg vaccines with other immunoprevention strategies can help overcome immunosuppression and enhance their effectiveness. These combined efforts can pave the way for the development of effective and safe neoAg-based vaccines for treating cancers.

Highlights:

• Lynch syndrome patients develop DNA mismatch repair deficient tumors that display microsatellite instability and have a high burden of frameshift mutations generating neoantigens, which are recognized by CD8⁺ and CD4⁺ T-cells to activate cancer cell killing.

• DNA mismatch repair deficiency causes recurrent mutations in ‘prone’ microsatellite loci that are organ-specific, thus leading to the presence of shared neoantigens in tumors across Lynch syndrome patients.

• Different in silico methods have been developed to predict and identify neoantigen sequences and their immunogenicity using next-generation sequencing data.

• Lynch Syndrome patients constitute a well-defined and prevalent population that has the potential to benefit from immune-interception strategies to prevent Lynch Syndrome-associated colorectal cancer and other tumor types such as endometrial, ovarian, small bowel and urothelial through past and current clinical trial efforts using neoantigen-based vaccines.

• Neoantigen-based vaccines hold great promise for the treatment of mismatch repair deficient gastrointestinal cancers. However, further research is needed to optimize their efficacy and safety.

Abbreviation List:

APC

antigen-presenting cell

ASA

Aspirin

CEA

carcinoembryonic antigen

CRC

colorectal cancer

DC

Dendritic cell

ELISpot

Enzyme-linked immunosorbent spot

GI

gastrointestinal

HLA

human leukocyte antigen

ICB

immune checkpoint blockade

ICIs

immune checkpoint inhibitors

Indels

insertion/deletion

LS

Lynch Syndrome

MAGE

melanoma-associated antigen

MHC

major histocompatibility complex

MS

microsatellite

MSI

microsatellite instability

MSI-H

microsatellite instability-high

MSS

microsatellite stability

MMR

Mismatch repair

MMRd

Mismatch repair-deficient

MUC1

mucin 1

NAP

Naproxen

NSAIDs

Non-steroidal anti-inflammatory drugs

neoAg

neoantigen

neoags

neoantigens

PGE2

prostaglandin E2

TAA

tumor-associated antigens

TCR

T-cell receptor

TME

Tumor microenvironment