Browse the corpus

Nous-209 neoantigen vaccine for cancer prevention in Lynch syndrome carriers: a phase 1b/2 trial. Cancer interception is a preventative approach aiming to reduce cancer incidence by targeting precancers and early-stage cancers. Lynch syndrome (LS) is a prevalent hereditary cancer syndrome affecting ~1 in 300 individuals, with an overall lifetime cancer risk as high as 80%. LS is caused by germline mutations in the DNA mismatch repair genes, leading to microsatellite instability (MSI) and accumulation of shared mutations. When these occur in coding regions, they generate frameshift peptides (FSPs). Nous-209 is a neoantigen-directed immunotherapy based on a heterologous prime boost using great ape adenovirus and modified vaccinia virus Ankara encoding 209 FSPs shared across MSI neoplasms. We present the results from cohort 1 of a phase 1b/2 single-arm trial of Nous-209 for cancer interception in LS carriers (n = 45). Safety and immunogenicity were coprimary endpoints. Safety was assessed in 45 participants. Vaccination was safe with no intervention-related serious adverse events (AEs). The most common AEs were injection-site reactions (any grade in 91% of participants after prime and 76% after boost with no grade 3) and fatigue (any grade in 80% after prime and 53% after boost with 4% grade 3 after prime or after boost). Neoantigen-specific immune responses were observed after vaccination in 100% of evaluable participants (n = 37), with induction of potent T cell immunity (mean response at peak of ~1,100 interferon-γ spot-forming cells per million peripheral blood mononuclear cells). The immune response was durable and detectable at 1 year in 85% of participants. Both CD8+ and CD4+ T cells were induced, recognizing multiple FSPs. Peptide-human leukocyte antigen predictions allowed the identification of >100 immunogenic FSPs with demonstration of cytotoxic activity in vitro. Immunogenic FSPs were found in independent datasets of LS MSI colorectal precancers and cancers. These results highlight Nous-209 ability to efficiently stimulate immunity against neoantigens in LS, supporting its development for cancer interception (ClinicalTrials.gov identifier: NCT05078866 ).

Lynch syndrome (LS) is one of the most prevalent hereditary cancer syndromes affecting ~1 in 300 people1. Persons with LS carry heterozygous germline mutations in one of four DNA mismatch repair (MMR) genes (MLH1, MSH2/EPCAM, MSH6 and PMS2), conferring 50–80% lifetime risk of colorectal cancer (CRC), 40–60% risk of endometrial cancer and increased risk of multiple other tumor types2. In the US, it is estimated that nearly 1 million individuals are affected by LS3. However, LS remains vastly underdiagnosed and current preventive options are limited to surveillance and prophylactic surgeries. The results of the CAPP2 study supported the role of aspirin 600 mg daily taken for 2 years as chemoprevention for at-risk LS carriers4. However, the use of high-dose aspirin in clinical practice is limited because of perceived risks of side effects; therefore, novel strategies for cancer interception remain an unmet need for this population.

study supported the role of aspirin 600 mg daily taken for 2 years as chemoprevention for at-risk LS carriers4. However, the use of high-dose aspirin in clinical practice is limited because of perceived risks of side effects; therefore, novel strategies for cancer interception remain an unmet need for this population. The development of LS-associated cancers typically results from the acquisition of MMR deficiency because of a second inactivating somatic hit on the alternative wild-type allele of the MMR gene, thus leading to subsequent accumulation of mutations within microsatellite (MS) regions5. The most common type of mutations in MS regions are frameshift insertions and deletions (indels)6. When they involve regions coding MS, indels result in the synthesis of frameshift peptide (FSP) neoantigens (neoAgs), which are expected to be highly immunogenic and dissimilar to native proteins. Recurrent and shared neoAgs across neoplastic lesions and different individuals have been identified in LS carriers, some of which have immunogenic features7. In this context, LS represents an ideal condition to leverage immune-interception strategies such as neoAg-based cancer vaccines targeting both precancers and cancers8.

nt and shared neoAgs across neoplastic lesions and different individuals have been identified in LS carriers, some of which have immunogenic features7. In this context, LS represents an ideal condition to leverage immune-interception strategies such as neoAg-based cancer vaccines targeting both precancers and cancers8. The development of ‘off-the-shelf’ vaccines for LS carriers represents indeed a promising approach to cancer interception and prevention. By targeting shared neoAgs generated through the accumulation of genetic mutations, such vaccines could stimulate the immune system to recognize and eliminate precancers before they develop into cancers. The discovery of neoAgs has revived the interest in the field of cancer vaccines that are emerging as promising approaches for targeting personalized or shared neoAgs9. The first phase 1/2a clinical trial leveraging peptide-based neoAg vaccination in LS carriers involved the use of three recurrent FSPs (TAF1B, HT001 and AIM2) in 22 participants with a history of stage III/IV MMR-deficient (MMRd) CRC10. The intervention was well tolerated and induced both humoral and cellular immune responses. Another phase 1/2 study evaluated a neoAg-based dendritic cell vaccine in 23 LS carriers and reported the induction of neoAg-specific immune responses and no incidence of LS-related cancers in transforming growth factor-β receptor 2–responsive participants over a 10-year follow-up11.

oral and cellular immune responses. Another phase 1/2 study evaluated a neoAg-based dendritic cell vaccine in 23 LS carriers and reported the induction of neoAg-specific immune responses and no incidence of LS-related cancers in transforming growth factor-β receptor 2–responsive participants over a 10-year follow-up11. Recently, novel delivery platforms have been developed, including DNA, RNA and viral vectored vaccines. Several clinical trials have now demonstrated the safety, immunogenicity and initial evidence of antitumor efficacy of neoAg-based vaccines across cancer types and settings12–15.

oral and cellular immune responses. Another phase 1/2 study evaluated a neoAg-based dendritic cell vaccine in 23 LS carriers and reported the induction of neoAg-specific immune responses and no incidence of LS-related cancers in transforming growth factor-β receptor 2–responsive participants over a 10-year follow-up11. Recently, novel delivery platforms have been developed, including DNA, RNA and viral vectored vaccines. Several clinical trials have now demonstrated the safety, immunogenicity and initial evidence of antitumor efficacy of neoAg-based vaccines across cancer types and settings12–15. Therefore, the selection of the vaccine platform is key in determining the magnitude, quality and breadth of T cell responses to achieve effective and durable antitumor immunity. Among these new platforms, viral vector-based vaccines represent a powerful platform capable of inducing strong and durable T cell responses in human16,17. Their ability to deliver large gene inserts allows the targeting of many neoAgs simultaneously. This is a crucial aspect for cancer interception linked to the need to induce a broad T cell response for high coverage and for addressing the heterogeneity of their potential future tumors. Vaccination based on great ape adenovirus (GAd) and modified vaccinia Ankara (MVA) vectors encoding 209 shared frameshift mutations (Nous-209) was used to target shared neoAgs in persons with metastatic tumors with microsatellite instability (MSI) and demonstrated safety and potent immunogenicity18. Moreover, early signs of clinical efficacy, observed when the vaccine is used in combination with an anti-programmed cell death protein 1 agent, are supported by evidence of vaccine-induced T cells infiltrating tumor biopsies after treatment, along with the expansion and diversification of the T cell receptor-β repertoire in persons showing clinical response18. Here, we present results from a phase 1b/2 open-label, multicenter study to evaluate the safety and immunogenicity of Nous-209 monotherapy in 45 healthy LS carriers. Comprehensive characterization of T cell responses, including breadth, longevity and functional cytotoxic activity, is reported here to demonstrate the ability of Nous-209 to promote the induction of T cell responses targeting FSPs identified in colorectal precancers and cancers from an independent LS dataset.

NCT05078866 is a phase 1b/2 single-arm, open-label, clinical trial testing Nous-209 for cancer immune interception in LS carriers. Eligible participants included adults of age ≥ 18 years with a clinical diagnosis of LS and with no evidence of active or recurrent invasive cancers for at least 6 months before screening. Safety and immunogenicity were the primary endpoints of this study. Changes in the number of colorectal adenomas, advanced neoplasia and/or carcinomas were prespecified secondary endpoints. The trial enrolled two cohorts: cohort 1 (initial vaccination) reported in this study, and cohort 2, consisting of a subset of participants from cohort 1 who were revaccinated at 1 year to assess the benefit of annual boost. The safety and immunogenicity of cohort 2 will be reported in a separate manuscript.

endpoints. The trial enrolled two cohorts: cohort 1 (initial vaccination) reported in this study, and cohort 2, consisting of a subset of participants from cohort 1 who were revaccinated at 1 year to assess the benefit of annual boost. The safety and immunogenicity of cohort 2 will be reported in a separate manuscript. Nous-209 was administered intramuscularly (IM) as a ‘priming’ dose of GAd20-209-FSPs on week 0 day 1, followed by a ‘boost’ with MVA-209-FSPs at week 8. Blood samples were collected serially at baseline and weeks 3, 8, 9, 24 and 52/68 (Fig. 1a). Before vaccination, all potentially eligible participants underwent standard-of-care screening colonoscopies or flexible sigmoidoscopy. Participants identified to have colorectal adenomas with high-grade dysplasia or diagnosis of invasive carcinoma at baseline were excluded from the study. More specifically, there were five screening failures. Two of them were excluded because of diagnosis of gastric cancer and one with diagnosis of bladder cancer within the previous 6 months.Fig. 1Nous-209 is safe and well tolerated in LS carriers.a, Overall schematic of Nous-209 administration and immunogenicity assessment. The initial priming GAd20-209-FSPs (GAd) was given on day 1 followed by booster MVA-209-FSPs (MVA) at week 8. Research blood samples were collected for immunogenicity evaluation over time. b, CONSORT flow diagram of study enrollment and analysis cohorts. c, Maximum frequency and severity of AEs observed following vaccinations with GAd-209-FSP (week 0) and MVA-209-FSP (week 8).Source data

y booster MVA-209-FSPs (MVA) at week 8. Research blood samples were collected for immunogenicity evaluation over time. b, CONSORT flow diagram of study enrollment and analysis cohorts. c, Maximum frequency and severity of AEs observed following vaccinations with GAd-209-FSP (week 0) and MVA-209-FSP (week 8).Source data a, Overall schematic of Nous-209 administration and immunogenicity assessment. The initial priming GAd20-209-FSPs (GAd) was given on day 1 followed by booster MVA-209-FSPs (MVA) at week 8. Research blood samples were collected for immunogenicity evaluation over time. b, CONSORT flow diagram of study enrollment and analysis cohorts. c, Maximum frequency and severity of AEs observed following vaccinations with GAd-209-FSP (week 0) and MVA-209-FSP (week 8). Source data

a, Overall schematic of Nous-209 administration and immunogenicity assessment. The initial priming GAd20-209-FSPs (GAd) was given on day 1 followed by booster MVA-209-FSPs (MVA) at week 8. Research blood samples were collected for immunogenicity evaluation over time. b, CONSORT flow diagram of study enrollment and analysis cohorts. c, Maximum frequency and severity of AEs observed following vaccinations with GAd-209-FSP (week 0) and MVA-209-FSP (week 8). Source data A total of 45 healthy LS participants enrolled between November 2022 and November 2023 were vaccinated (Fig. 1b), with the first participant of the study enrolled on November 10, 2022 and the last participant enrolled on November 30, 2023. Baseline demographics and clinical characteristics are summarized in Table 1. All participants carried a pathogenic germline MMR mutation with the majority in MSH2 (47%), while the remainder were in MSH6 (24%), MLH1 (18%) and PMS2 (11%). The median age was 50 years (range, 24–71) and 42% of participants (19/45) were cancer survivors (Extended Data Table 1) and had no evidence of active malignancy within the 6 months before study enrollment. Most study participants were male (56%) and self-reported white (91%). The coprimary endpoints of the trial were safety (rate of adverse events, AEs) and immunogenicity against the FSPs encoded by the Nous-209 vaccine.Table 1Demographics and clinical characteristics of trial participantsCategoryn (%)Age18–3911 (24.4)40–6427 (60)≥657 (15.6)GenderFemale20 (44.4)Male25 (55.6)RaceAsian2 (4.4)Not reported2 (4.4)White41 (91.1)EthnicityHispanic or Latino10 (22.2)Not Hispanic or Latino33 (73.3)Unknown2 (4.4)Performance statusECOG 044 (97.8)ECOG 11 (2.2)MMR geneMLH18 (17.8)MSH221 (46.7)MSH611 (24.4)PMS25 (11.1)Cancer historyPrevivor26 (57.8)Survivor19 (42.2)

5.6)GenderFemale20 (44.4)Male25 (55.6)RaceAsian2 (4.4)Not reported2 (4.4)White41 (91.1)EthnicityHispanic or Latino10 (22.2)Not Hispanic or Latino33 (73.3)Unknown2 (4.4)Performance statusECOG 044 (97.8)ECOG 11 (2.2)MMR geneMLH18 (17.8)MSH221 (46.7)MSH611 (24.4)PMS25 (11.1)Cancer historyPrevivor26 (57.8)Survivor19 (42.2) Demographics and clinical characteristics of trial participants

5.6)GenderFemale20 (44.4)Male25 (55.6)RaceAsian2 (4.4)Not reported2 (4.4)White41 (91.1)EthnicityHispanic or Latino10 (22.2)Not Hispanic or Latino33 (73.3)Unknown2 (4.4)Performance statusECOG 044 (97.8)ECOG 11 (2.2)MMR geneMLH18 (17.8)MSH221 (46.7)MSH611 (24.4)PMS25 (11.1)Cancer historyPrevivor26 (57.8)Survivor19 (42.2) Demographics and clinical characteristics of trial participants AEs were reported by 98% (44/45) of participants during the first 9 weeks after vaccination with GAd-209-FSP prime (Fig. 1c and Table 2). Overall, no treatment-related serious AEs (SAEs) were observed and vaccination was well tolerated. The most common treatment-related AEs included systemic reactogenicity symptoms and local injection-site reaction. Specifically, injection-site reactions of any grade occurred in 91% of participants following GAd-209-FSP prime vaccination and 76% following MVA-209-FSP boost vaccination; no grade 3 injection-site reactions were observed. With respect to systemic reactogenicity, the most common events included fatigue (any grade: 36 of 45 participants (80%) after GAd-209-FSP prime, 24 of 45 (53%) after MVA-209-FSP boost; grade 3: 2 of 45 (4%) participants after GAd-209 prime or after MVA-209-boost) and myalgia (any grade: 34 of 45 participants (76%) after GAd-209-FSP prime, 11 of 45 (24%) after MVA-209-FSP boost; grade 3: 2 of 45 participants (4%) after GAd-209-FSP prime or after MVA-209-FSP boost) (Fig. 1c, Supplementary Table 1 and Extended Data Fig. 1). All grade 3 symptoms occurred after the GAd prime dose. In all cases, symptoms were transient, lasting 1–4 days, did not require hospitalization and were managed with self-administered acetaminophen in some cases (Extended Data Table 2). All AEs reached full resolution and, indeed, all participants were able to receive the booster MVA dose as scheduled.Table 2Frequency of definite, probable or possible treatment-related AEs according to any grade and observation window after Nous-209 vaccination in all trial participants (n = 45). #Percentage of AEs of grade ≥3 and <3 calculated from AEs of any gradeData cutoffAEsAEsAEsAny gradeGrade <3Grade ≥3nn (%)#n (%)#Day 1Visit 1 (GAd administration)147142 (96.6)5 (3.4)Week 1Visit 2 (2–6 days after GAd)201198 (98.5)3 (1.5)Weeks 2–7Visit 322 (100)0Week 8Visit 4 (MVA administration)6565 (100)0Week 9Visit 5 (2–6 days after MVA)9595 (100)0Weeks 10–16Visit 655 (100)0Weeks 16–24Visit 7000Weeks 24–36Visit 8000Weeks 36–52Visit 9000Total515507 (98.4)8 (1.6)

administration)147142 (96.6)5 (3.4)Week 1Visit 2 (2–6 days after GAd)201198 (98.5)3 (1.5)Weeks 2–7Visit 322 (100)0Week 8Visit 4 (MVA administration)6565 (100)0Week 9Visit 5 (2–6 days after MVA)9595 (100)0Weeks 10–16Visit 655 (100)0Weeks 16–24Visit 7000Weeks 24–36Visit 8000Weeks 36–52Visit 9000Total515507 (98.4)8 (1.6) Frequency of definite, probable or possible treatment-related AEs according to any grade and observation window after Nous-209 vaccination in all trial participants (n = 45). #Percentage of AEs of grade ≥3 and <3 calculated from AEs of any grade

administration)147142 (96.6)5 (3.4)Week 1Visit 2 (2–6 days after GAd)201198 (98.5)3 (1.5)Weeks 2–7Visit 322 (100)0Week 8Visit 4 (MVA administration)6565 (100)0Week 9Visit 5 (2–6 days after MVA)9595 (100)0Weeks 10–16Visit 655 (100)0Weeks 16–24Visit 7000Weeks 24–36Visit 8000Weeks 36–52Visit 9000Total515507 (98.4)8 (1.6) Frequency of definite, probable or possible treatment-related AEs according to any grade and observation window after Nous-209 vaccination in all trial participants (n = 45). #Percentage of AEs of grade ≥3 and <3 calculated from AEs of any grade Vaccine immunogenicity was the coprimary endpoint of the study and was evaluated using an ex vivo enzyme-linked immunosorbent spot (ELISpot) assay against 16 peptide pools that covered the entire repertoire of the 209 FSPs encoded by Nous-209, with each pool covering multiple FSPs (Fig. 2a). Peripheral blood mononuclear cells (PBMCs) were isolated at different time points before (baseline) and after vaccination. The presence of Nous-209-induced neoAg-specific responses before and after vaccination was evaluable in 37 participants. Nous-209 elicited a positive interferon-γ (IFNγ) T cell response at peak (week 9) in 100% (37/37) of evaluable participants determined according to the predefined protocol criteria for ELISpot positivity (Methods). Total responses at peak after vaccination reached a mean of ~1,100 spot-forming cells (SFCs) (Fig. 2a). Furthermore, 10% of participants (4/37) showed positive spontaneous T cell responses to neoAg peptide pools at baseline before vaccination. In these four subjects, Nous-209 either boosted or induced de novo T cell responses to the vaccine neoAg. Long-term immune responses at 6 months and 1 year after vaccination were evaluable in 33 participants, showing maintenance of elevated and long-lasting T cell immunity with a decline in immune response occurring after the initial peak response at 1 year. Positive T cell responses were indeed still detectable ex vivo in 97% and 85% of evaluable participants at 6 months and 1 year, respectively (Fig. 2b). Responses were directed against different peptide pools with the induction of a broad polytope response to vaccine neoAg (Fig. 2c,d). More specifically, the breadth of the cellular immune response was evaluated by counting the number of Nous-209 peptide pools targeted across visits with an average of eight immunogenic pools per participant. A total of 13% of participants showed reactivity against 1–3 pools, 54% showed reactivity against 4–9 pools and 33% showed reactivity against 10–16 pools (Fig. 2d).

response was evaluated by counting the number of Nous-209 peptide pools targeted across visits with an average of eight immunogenic pools per participant. A total of 13% of participants showed reactivity against 1–3 pools, 54% showed reactivity against 4–9 pools and 33% showed reactivity against 10–16 pools (Fig. 2d). In exploratory analyses, we observed no statistically significant associations between the number of positive pools at peak and gender, age, self-reported race/ethnicity, mutated MMR gene or cancer previvor versus survivor status. However, the increase in positive pools at peak from baseline was significantly higher in females compared to males (P = 0.045; Supplementary Table 2).Fig. 2Nous-209 elicits potent and broad neoAg-specific T cell responses in LS carriers.a, T cell responses assessed by ex vivo IFNγ ELISpot at baseline and peak immune response. Shown are the numbers of SFCs per 106 PMBCs corresponding to the sum of the responses to 16 separate pools of peptides encompassing the 209 FSP sequences. Each line represents data from an individual subject (n = 37), with circles representing the mean of three technical replicates per time point. Comparison of baseline versus peak immune response was performed using a two-tailed Mann–Whitney U-test; P < 0.0001. b, Kinetics of immune response over time showing durability of T cell response after vaccination (mean ± s.e.m.; n = 37 subjects at baseline and peak, n = 33 subjects at 6 months and 1 year). c, Representation of immunogenic peptide pools eliciting IFNγ reactivity by ex vivo ELISpot annotated for participant ID, mutated MMR gene, age, gender and cancer history (previvor or survivor). Right, bar plot showing the cumulative number of reactive pools across visits per each participant. d, Breadth of immune response assessed as the number of positive pools eliciting IFNγ reactivity across visits. Pie chart showing the frequency (%) of LS carriers with a number of reactive pools ranging from 1–3 (pink), 4–9 (dark pink) or 10–16 (violet).Source data

of reactive pools across visits per each participant. d, Breadth of immune response assessed as the number of positive pools eliciting IFNγ reactivity across visits. Pie chart showing the frequency (%) of LS carriers with a number of reactive pools ranging from 1–3 (pink), 4–9 (dark pink) or 10–16 (violet).Source data a, T cell responses assessed by ex vivo IFNγ ELISpot at baseline and peak immune response. Shown are the numbers of SFCs per 106 PMBCs corresponding to the sum of the responses to 16 separate pools of peptides encompassing the 209 FSP sequences. Each line represents data from an individual subject (n = 37), with circles representing the mean of three technical replicates per time point. Comparison of baseline versus peak immune response was performed using a two-tailed Mann–Whitney U-test; P < 0.0001. b, Kinetics of immune response over time showing durability of T cell response after vaccination (mean ± s.e.m.; n = 37 subjects at baseline and peak, n = 33 subjects at 6 months and 1 year). c, Representation of immunogenic peptide pools eliciting IFNγ reactivity by ex vivo ELISpot annotated for participant ID, mutated MMR gene, age, gender and cancer history (previvor or survivor). Right, bar plot showing the cumulative number of reactive pools across visits per each participant. d, Breadth of immune response assessed as the number of positive pools eliciting IFNγ reactivity across visits. Pie chart showing the frequency (%) of LS carriers with a number of reactive pools ranging from 1–3 (pink), 4–9 (dark pink) or 10–16 (violet). Source data

a, T cell responses assessed by ex vivo IFNγ ELISpot at baseline and peak immune response. Shown are the numbers of SFCs per 106 PMBCs corresponding to the sum of the responses to 16 separate pools of peptides encompassing the 209 FSP sequences. Each line represents data from an individual subject (n = 37), with circles representing the mean of three technical replicates per time point. Comparison of baseline versus peak immune response was performed using a two-tailed Mann–Whitney U-test; P < 0.0001. b, Kinetics of immune response over time showing durability of T cell response after vaccination (mean ± s.e.m.; n = 37 subjects at baseline and peak, n = 33 subjects at 6 months and 1 year). c, Representation of immunogenic peptide pools eliciting IFNγ reactivity by ex vivo ELISpot annotated for participant ID, mutated MMR gene, age, gender and cancer history (previvor or survivor). Right, bar plot showing the cumulative number of reactive pools across visits per each participant. d, Breadth of immune response assessed as the number of positive pools eliciting IFNγ reactivity across visits. Pie chart showing the frequency (%) of LS carriers with a number of reactive pools ranging from 1–3 (pink), 4–9 (dark pink) or 10–16 (violet). Source data Full deconvolution of T cell responses against all 209 FSPs and relative minimal epitopes was not feasible given the high number of encoded neoAgs and the limited availability of PBMC samples. To further dissect immune responses after Nous-209 at the level of individual FSPs, we pursued an alternative approach that first determined the reactive peptide pools in each participant and then predicted the top epitopes included in the immunogenic pools on the basis of each participant’s human leukocyte antigen (HLA) class I genotyping, when full deconvolution of T cell response was not feasible (Fig. 3a). Following this strategy, we restricted the analysis to top FSPs and epitopes with the best participant’s class I HLA-binding prediction (half-maximal effective concentration < 500 nM), which were considered the best target candidates of Nous-209 immune response. This work of deconvoluting the peptide pools, while not comprehensive for all 16 available pools, provides insights into the immunogenicity of specific individual FSP, thus leading to the identification of 115 immunogenic FSPs in our trial cohort of LS carriers (Supplementary Table 3).Fig. 3Immunogenic FSPs are present in cancer and precancer lesions of LS carriers from an independent cohort.a, Schematic flowchart illustrating the steps involved in the deconvolution of reactive peptide pools and mapping of immunogenic FSPs in LS carriers on the basis of HLA-binding prediction and samples availability for each participant. b, Bar plot showing the number of Nous-209 FSPs found in independent dataset of cancer (n = 11) and precancer (n = 12) lesions by lookup approach (cutoff: VAF ≥ 10% and mutated tumor reads ≥ 3). c, The pie chart represents the percentage of the identified immunogenic FSPs (n = 115) found in the total number of cancer and precancer samples analyzed. d, Ex vivo IFNγ ELISpot responses on PBMCs before or after depletion of CD8+ T cells in presence of some identified reactive pools or FSPs. CD4 and CD8 indicate the subtype-specific CD4 and CD8 T cell responses identified in five LS carriers. Bars represent the pre-CD8 and post-CD8 T cell depletion response for each subject, respectively.

pot responses on PBMCs before or after depletion of CD8+ T cells in presence of some identified reactive pools or FSPs. CD4 and CD8 indicate the subtype-specific CD4 and CD8 T cell responses identified in five LS carriers. Bars represent the pre-CD8 and post-CD8 T cell depletion response for each subject, respectively. Data are shown as the mean number of SFCs per 106 PBMCs ± s.e.m. Dots represent three technical replicates.Source data a, Schematic flowchart illustrating the steps involved in the deconvolution of reactive peptide pools and mapping of immunogenic FSPs in LS carriers on the basis of HLA-binding prediction and samples availability for each participant. b, Bar plot showing the number of Nous-209 FSPs found in independent dataset of cancer (n = 11) and precancer (n = 12) lesions by lookup approach (cutoff: VAF ≥ 10% and mutated tumor reads ≥ 3). c, The pie chart represents the percentage of the identified immunogenic FSPs (n = 115) found in the total number of cancer and precancer samples analyzed. d, Ex vivo IFNγ ELISpot responses on PBMCs before or after depletion of CD8+ T cells in presence of some identified reactive pools or FSPs. CD4 and CD8 indicate the subtype-specific CD4 and CD8 T cell responses identified in five LS carriers. Bars represent the pre-CD8 and post-CD8 T cell depletion response for each subject, respectively. Data are shown as the mean number of SFCs per 106 PBMCs ± s.e.m. Dots represent three technical replicates. Source data

a, Schematic flowchart illustrating the steps involved in the deconvolution of reactive peptide pools and mapping of immunogenic FSPs in LS carriers on the basis of HLA-binding prediction and samples availability for each participant. b, Bar plot showing the number of Nous-209 FSPs found in independent dataset of cancer (n = 11) and precancer (n = 12) lesions by lookup approach (cutoff: VAF ≥ 10% and mutated tumor reads ≥ 3). c, The pie chart represents the percentage of the identified immunogenic FSPs (n = 115) found in the total number of cancer and precancer samples analyzed. d, Ex vivo IFNγ ELISpot responses on PBMCs before or after depletion of CD8+ T cells in presence of some identified reactive pools or FSPs. CD4 and CD8 indicate the subtype-specific CD4 and CD8 T cell responses identified in five LS carriers. Bars represent the pre-CD8 and post-CD8 T cell depletion response for each subject, respectively. Data are shown as the mean number of SFCs per 106 PBMCs ± s.e.m. Dots represent three technical replicates. Source data To demonstrate the ability of Nous-209 to elicit an immune response against FSPs present in colorectal precancers and cancers of LS carriers, we leveraged available genomic and transcriptomic data from previously published work in an independent cohort of 12 MSI precancers (eight adenomas and four advanced adenomas) and 11 MSI CRCs7. Overall, we identified the presence of a median of 32 FSPs per lesion that are encoded by Nous-209 (Fig. 3b). Interestingly, when restricting the analysis to the 115 immunogenic FSP identified in our trial cohort, we specifically found that 94 of 115 (82%) were present in colorectal precancers and cancers (Fig. 3c). To distinguish CD4+ and CD8+ T cell-mediated responses, CD8+ T cells were depleted. Ex vivo IFNγ ELISpot assays were performed before and after depletion of CD8+ T cells in presence of the highest reactive pools or FSPs identified as immunogenic. Responses were defined as CD8+ mediated if IFNγ production was reduced after depletion and as CD4+ mediated if no substantial change was observed. The results showed that the immunogenic FSPs or reactive pools were able to stimulate both CD4+ and CD8+ T cells in the five tested participants (Fig. 3d).

d as immunogenic. Responses were defined as CD8+ mediated if IFNγ production was reduced after depletion and as CD4+ mediated if no substantial change was observed. The results showed that the immunogenic FSPs or reactive pools were able to stimulate both CD4+ and CD8+ T cells in the five tested participants (Fig. 3d). To analyze the phenotype and killing activity of vaccine-induced CD8+ T cells, we generated HLA class I dextramer-detecting neoAg-reactive CD8+ T cells against an FSP from SPEF2 that was identified among the most recurrent immunogenic FSP in multiple LS carriers (Extended Data Fig. 2a). The SPEF2 FSP sequence includes an 8-mer peptide (IAKKRIKL) predicted in silico to be a strong binder to HLA-B*08:01, which is shared by several trial participants and was confirmed as immunogenic by ELISpot in 11 participants (participants 16, 18, 19, 27, 29, 30, 31,39, 40, 10 and 32; Extended Data Fig. 2a–c). In two participants (participants 10 and 32) SPEF2 FSP was mapped as immunogenic after in vitro stimulation (Extended Data Fig. 3). In participant 18, class I dextramer staining detected SPEF2 FSP vaccine-induced CD8+ T cells directly ex vivo at week 9 after MVA boost, representing ~1.64 % of all circulating CD8+ cells, which persisted at 6 months and 1 year after vaccination (Extended Data Fig. 2d), consistent with the longevity of T cell responses observed using ex vivo ELISpot assays. Vaccine-induced SPEF2 FSP dextramer-positive CD8+ T cells were further characterized by examining the phenotype of naive (CD45RA+CCR7+), central memory (CD45RA−CCR7+), effector memory (CD45RA−CCR7−) and terminally differentiated (TemRA; CD45RA+CCR7−) memory subsets. The analysis showed that, after Nous-209, antigen-specific T cells display a TemRA phenotype, thus demonstrating the induction of a long-term memory response by Nous-209 (Extended Data Fig. 2d). This was confirmed in additional participants who mounted an immune response against SPEF2 FSP after Nous-209 vaccination (Supplementary Fig. 1). Lastly, the cytotoxic potential was investigated by flow cytometry analysis with intracellular cytokine staining in participant 18 PBMCs stimulated ex vivo with SPEF2 peptide. We observed antigen-specific IFNγ secretion and detection of CD8+ T cells expressing the degranulation marker CD107a, thus indicating a cytotoxic function that was stimulated after Nous-209 vaccination (Extended Data Fig. 4a).

ntracellular cytokine staining in participant 18 PBMCs stimulated ex vivo with SPEF2 peptide. We observed antigen-specific IFNγ secretion and detection of CD8+ T cells expressing the degranulation marker CD107a, thus indicating a cytotoxic function that was stimulated after Nous-209 vaccination (Extended Data Fig. 4a). A second neoAg, CDC7, found recurrently immunogenic in participants with HLA-A*03:01, was selected to further evaluate the ability of Nous-209-induced T cells to eliminate tumor cells presenting the neoAg. We conducted a direct killing assay using a microfluidic three-dimensional coculture system (Extended Data Fig. 4b). This approach modeled T cell-mediated cytotoxicity against tumor cells expressing CDC7 FSP using PBMCs from participant 12. Our in vitro assay was based on the MMRd HCT116 human colon cancer cell line that was genetically modified to express the FSP from CDC7 and their matching HLAs (CDC7-HCT116 herein; Methods) and cocultured with in vitro-stimulated effector PBMCs obtained from the same participant before and after vaccination with Nous-209. HCT116 cells were cocultured with PBMCs at an effector-to-target (E:T) ratio of 5:1. To assess both functional and cytotoxic responses, PBMCs were divided into two groups after in vitro stimulation; half of the cells were analyzed for IFNγ secretion using ELISpot and the remaining cells were used for the microfluidic coculture assay. After 48 h, target cell viability and apoptosis were assessed using luminescence-based assays. Our results showed a significant reduction in CDC7-HCT116 tumor cell survival when cocultured with T cells after Nous-209 vaccination compared to both baseline PBMCs from the same participant and all controls (Extended Data Fig. 4c,d). To ensure rigorous evaluation, we used multiple control conditions, including HCT116 cells transduced with CDC7 and HCT116 cells transduced only with the HLA of interest but without the CDC7 neoAg. These controls helped determine whether tumor cell killing was CDC7 dependent rather than influenced by HLA expression alone or intrinsic effects of CDC7 transduction. Quantitative viability assays revealed a marked decrease in tumor cell survival after Nous-209 vaccination compared to baseline and all control conditions (P < 0.0001; Extended Data Fig. 4d). In addition, a significant increase in caspase 3/7-mediated apoptosis was observed in CDC7-HCT116 cells following coculture with PBMCs after vaccination with Nous-209 (P < 0.0001; Extended Data Fig.

survival after Nous-209 vaccination compared to baseline and all control conditions (P < 0.0001; Extended Data Fig. 4d). In addition, a significant increase in caspase 3/7-mediated apoptosis was observed in CDC7-HCT116 cells following coculture with PBMCs after vaccination with Nous-209 (P < 0.0001; Extended Data Fig. 4d), thus confirming antigen-specific T cell-mediated cytotoxicity against tumor cells expressing the CDC7 cognate pHLA. To strengthen these findings, we assessed the IFNγ secretion capacity of the stimulated PBMCs in response to CDC7 neoAg stimulation. At baseline, there was no significant difference between CDC7-stimulated PBMCs and the negative control (DMSO; P > 0.05), indicating that CDC7-specific responses were not detectable before vaccination. In contrast, PBMCs after Nous-209 vaccination from participant 12 exhibited significantly higher IFNγ secretion compared to both the negative control (DMSO; P < 0.0001) and baseline PBMCs (P < 0.0001), thus reinforcing the antigen-specific activation of Nous-209-induced T cells (Extended Data Fig. 4e). These findings demonstrate that Nous-209 vaccination induces a potent and durable neoAg-specific CD8+ T cell response capable of both cytokine secretion and direct tumor cell killing. The presence of long-lived TemRA CD8+ T cells, their ability to degranulate and their direct cytotoxic function provide compelling evidence that Nous-209 effectively primes the immune system to target and eliminate neoAg-expressing tumor cells in LS carriers.

le of both cytokine secretion and direct tumor cell killing. The presence of long-lived TemRA CD8+ T cells, their ability to degranulate and their direct cytotoxic function provide compelling evidence that Nous-209 effectively primes the immune system to target and eliminate neoAg-expressing tumor cells in LS carriers. As a prespecified secondary endpoint of our study, we aimed to characterize the number, size and pathology features of colorectal neoplasia among 43 of 45 vaccinated participants who underwent their standard-of-care screening lower endoscopy at the end of the study. Two vaccinated participants withdrew from the study before undergoing the end-of-study colonoscopy and were not evaluable for this analysis. Overall, the majority of participants (n = 31) had no colorectal adenomas on their end-of-study colonoscopies. Interestingly, of the total 23 adenomas detected in 12 participants, none were classified as advanced adenomas (diameter > 10 mm, presence of high-grade dysplasia and/or villous histology), while two participants had advanced adenomas at baseline colonoscopy (Fig. 4a,b and Supplementary Table 4). Compared to baseline, no statistically significant differences were observed in the overall proportion of study participants with detectable adenomas or advanced adenomas or in the total counts at their end-of-study colonoscopy (Extended Data Tables 3 and 4). Genomic characterization of available precancers removed at baseline (n = 10) and end-of-study (n = 12) colonoscopies was performed by next-generation sequencing assessing both the MSI status and the presence of Nous-209 FSPs. Although not statistically significant, the analysis showed a decrease in the proportion of MSI-High (MSI-H) and MSI-Low (MSI-L) precancers observed at baseline compared to end of study. More specifically, MSI-H and MSI-L precancers were observed in 30% and 30% of participants at baseline and 17% and 8% of participants at end-of-study colonoscopy, respectively. After vaccination, two precancers were MSI-H and one was MSI-L, while the remaining nine were microsatellite stable (MSS) (Extended Data Fig. 5a). Regarding the presence of Nous-209 FSPs after vaccination, none of the MSS precancers displayed any of the Nous-209 FSPs, which were only identified in the two MSI-H lesions and in the MSI-L lesion (Extended Data Fig. 5b).Fig.

was MSI-L, while the remaining nine were microsatellite stable (MSS) (Extended Data Fig. 5a). Regarding the presence of Nous-209 FSPs after vaccination, none of the MSS precancers displayed any of the Nous-209 FSPs, which were only identified in the two MSI-H lesions and in the MSI-L lesion (Extended Data Fig. 5b).Fig. 4Colorectal neoplasia burden observed at end-of-study colonoscopy inversely correlates with breadth of immune response.a, Number of participants who underwent screening colonoscopy at baseline and end of study (EoS; n = 43) who had no adenomas (adenomas absent), at least one adenoma (adenomas present) and advanced adenomas (advanced adenomas present) detected. b, Number of adenomas per trial participant at baseline and end of study; comparison of baseline versus EoS was performed using a two-tailed Mann–Whitney U-test; NS, not significant. c, Number of reactive pools measured at 6 months (n = 34 evaluable subjects) between the participants with and without adenomas. Data are shown as the mean ± s.e.m. Comparison of participants with and without adenomas was performed using a one-tailed Mann–Whitney U-test; *P = 0.0381.Source data

st; NS, not significant. c, Number of reactive pools measured at 6 months (n = 34 evaluable subjects) between the participants with and without adenomas. Data are shown as the mean ± s.e.m. Comparison of participants with and without adenomas was performed using a one-tailed Mann–Whitney U-test; *P = 0.0381.Source data a, Number of participants who underwent screening colonoscopy at baseline and end of study (EoS; n = 43) who had no adenomas (adenomas absent), at least one adenoma (adenomas present) and advanced adenomas (advanced adenomas present) detected. b, Number of adenomas per trial participant at baseline and end of study; comparison of baseline versus EoS was performed using a two-tailed Mann–Whitney U-test; NS, not significant. c, Number of reactive pools measured at 6 months (n = 34 evaluable subjects) between the participants with and without adenomas. Data are shown as the mean ± s.e.m. Comparison of participants with and without adenomas was performed using a one-tailed Mann–Whitney U-test; *P = 0.0381. Source data

a, Number of participants who underwent screening colonoscopy at baseline and end of study (EoS; n = 43) who had no adenomas (adenomas absent), at least one adenoma (adenomas present) and advanced adenomas (advanced adenomas present) detected. b, Number of adenomas per trial participant at baseline and end of study; comparison of baseline versus EoS was performed using a two-tailed Mann–Whitney U-test; NS, not significant. c, Number of reactive pools measured at 6 months (n = 34 evaluable subjects) between the participants with and without adenomas. Data are shown as the mean ± s.e.m. Comparison of participants with and without adenomas was performed using a one-tailed Mann–Whitney U-test; *P = 0.0381. Source data In exploratory analyses performed among 34 participants with evaluable immune response data at 6 months, we observed an association between the breadth of immune response (number of reactive FSP pools) at 6 months and the presence versus absence of any detectable adenomas on the end-of-study colonoscopy (mean: 1.5 versus 4 pools, respectively; Fig. 4c). A similar trend was observed at peak, whereas no statistically significant difference was detected at 12 months, likely because of the expected contraction of the immune response over time. No CRC was identified at colonoscopy. Furthermore, three of 45 vaccinated participants were subsequently diagnosed with invasive noncolorectal carcinomas, including one participant with MMRd gastric cancer, one with MMR-unknown non-small cell lung cancer and one with MMR-proficient prostate cancer.

Vaccine immunogenicity was the coprimary endpoint of the study and was evaluated using an ex vivo enzyme-linked immunosorbent spot (ELISpot) assay against 16 peptide pools that covered the entire repertoire of the 209 FSPs encoded by Nous-209, with each pool covering multiple FSPs (Fig. 2a). Peripheral blood mononuclear cells (PBMCs) were isolated at different time points before (baseline) and after vaccination. The presence of Nous-209-induced neoAg-specific responses before and after vaccination was evaluable in 37 participants. Nous-209 elicited a positive interferon-γ (IFNγ) T cell response at peak (week 9) in 100% (37/37) of evaluable participants determined according to the predefined protocol criteria for ELISpot positivity (Methods). Total responses at peak after vaccination reached a mean of ~1,100 spot-forming cells (SFCs) (Fig. 2a). Furthermore, 10% of participants (4/37) showed positive spontaneous T cell responses to neoAg peptide pools at baseline before vaccination. In these four subjects, Nous-209 either boosted or induced de novo T cell responses to the vaccine neoAg. Long-term immune responses at 6 months and 1 year after vaccination were evaluable in 33 participants, showing maintenance of elevated and long-lasting T cell immunity with a decline in immune response occurring after the initial peak response at 1 year. Positive T cell responses were indeed still detectable ex vivo in 97% and 85% of evaluable participants at 6 months and 1 year, respectively (Fig. 2b). Responses were directed against different peptide pools with the induction of a broad polytope response to vaccine neoAg (Fig. 2c,d). More specifically, the breadth of the cellular immune response was evaluated by counting the number of Nous-209 peptide pools targeted across visits with an average of eight immunogenic pools per participant. A total of 13% of participants showed reactivity against 1–3 pools, 54% showed reactivity against 4–9 pools and 33% showed reactivity against 10–16 pools (Fig. 2d).

Full deconvolution of T cell responses against all 209 FSPs and relative minimal epitopes was not feasible given the high number of encoded neoAgs and the limited availability of PBMC samples. To further dissect immune responses after Nous-209 at the level of individual FSPs, we pursued an alternative approach that first determined the reactive peptide pools in each participant and then predicted the top epitopes included in the immunogenic pools on the basis of each participant’s human leukocyte antigen (HLA) class I genotyping, when full deconvolution of T cell response was not feasible (Fig. 3a). Following this strategy, we restricted the analysis to top FSPs and epitopes with the best participant’s class I HLA-binding prediction (half-maximal effective concentration < 500 nM), which were considered the best target candidates of Nous-209 immune response. This work of deconvoluting the peptide pools, while not comprehensive for all 16 available pools, provides insights into the immunogenicity of specific individual FSP, thus leading to the identification of 115 immunogenic FSPs in our trial cohort of LS carriers (Supplementary Table 3).Fig. 3Immunogenic FSPs are present in cancer and precancer lesions of LS carriers from an independent cohort.a, Schematic flowchart illustrating the steps involved in the deconvolution of reactive peptide pools and mapping of immunogenic FSPs in LS carriers on the basis of HLA-binding prediction and samples availability for each participant. b, Bar plot showing the number of Nous-209 FSPs found in independent dataset of cancer (n = 11) and precancer (n = 12) lesions by lookup approach (cutoff: VAF ≥ 10% and mutated tumor reads ≥ 3). c, The pie chart represents the percentage of the identified immunogenic FSPs (n = 115) found in the total number of cancer and precancer samples analyzed. d, Ex vivo IFNγ ELISpot responses on PBMCs before or after depletion of CD8+ T cells in presence of some identified reactive pools or FSPs. CD4 and CD8 indicate the subtype-specific CD4 and CD8 T cell responses identified in five LS carriers. Bars represent the pre-CD8 and post-CD8 T cell depletion response for each subject, respectively.

To analyze the phenotype and killing activity of vaccine-induced CD8+ T cells, we generated HLA class I dextramer-detecting neoAg-reactive CD8+ T cells against an FSP from SPEF2 that was identified among the most recurrent immunogenic FSP in multiple LS carriers (Extended Data Fig. 2a). The SPEF2 FSP sequence includes an 8-mer peptide (IAKKRIKL) predicted in silico to be a strong binder to HLA-B*08:01, which is shared by several trial participants and was confirmed as immunogenic by ELISpot in 11 participants (participants 16, 18, 19, 27, 29, 30, 31,39, 40, 10 and 32; Extended Data Fig. 2a–c). In two participants (participants 10 and 32) SPEF2 FSP was mapped as immunogenic after in vitro stimulation (Extended Data Fig. 3). In participant 18, class I dextramer staining detected SPEF2 FSP vaccine-induced CD8+ T cells directly ex vivo at week 9 after MVA boost, representing ~1.64 % of all circulating CD8+ cells, which persisted at 6 months and 1 year after vaccination (Extended Data Fig. 2d), consistent with the longevity of T cell responses observed using ex vivo ELISpot assays. Vaccine-induced SPEF2 FSP dextramer-positive CD8+ T cells were further characterized by examining the phenotype of naive (CD45RA+CCR7+), central memory (CD45RA−CCR7+), effector memory (CD45RA−CCR7−) and terminally differentiated (TemRA; CD45RA+CCR7−) memory subsets. The analysis showed that, after Nous-209, antigen-specific T cells display a TemRA phenotype, thus demonstrating the induction of a long-term memory response by Nous-209 (Extended Data Fig. 2d). This was confirmed in additional participants who mounted an immune response against SPEF2 FSP after Nous-209 vaccination (Supplementary Fig. 1). Lastly, the cytotoxic potential was investigated by flow cytometry analysis with intracellular cytokine staining in participant 18 PBMCs stimulated ex vivo with SPEF2 peptide. We observed antigen-specific IFNγ secretion and detection of CD8+ T cells expressing the degranulation marker CD107a, thus indicating a cytotoxic function that was stimulated after Nous-209 vaccination (Extended Data Fig. 4a).

As a prespecified secondary endpoint of our study, we aimed to characterize the number, size and pathology features of colorectal neoplasia among 43 of 45 vaccinated participants who underwent their standard-of-care screening lower endoscopy at the end of the study. Two vaccinated participants withdrew from the study before undergoing the end-of-study colonoscopy and were not evaluable for this analysis. Overall, the majority of participants (n = 31) had no colorectal adenomas on their end-of-study colonoscopies. Interestingly, of the total 23 adenomas detected in 12 participants, none were classified as advanced adenomas (diameter > 10 mm, presence of high-grade dysplasia and/or villous histology), while two participants had advanced adenomas at baseline colonoscopy (Fig. 4a,b and Supplementary Table 4). Compared to baseline, no statistically significant differences were observed in the overall proportion of study participants with detectable adenomas or advanced adenomas or in the total counts at their end-of-study colonoscopy (Extended Data Tables 3 and 4). Genomic characterization of available precancers removed at baseline (n = 10) and end-of-study (n = 12) colonoscopies was performed by next-generation sequencing assessing both the MSI status and the presence of Nous-209 FSPs. Although not statistically significant, the analysis showed a decrease in the proportion of MSI-High (MSI-H) and MSI-Low (MSI-L) precancers observed at baseline compared to end of study. More specifically, MSI-H and MSI-L precancers were observed in 30% and 30% of participants at baseline and 17% and 8% of participants at end-of-study colonoscopy, respectively. After vaccination, two precancers were MSI-H and one was MSI-L, while the remaining nine were microsatellite stable (MSS) (Extended Data Fig. 5a). Regarding the presence of Nous-209 FSPs after vaccination, none of the MSS precancers displayed any of the Nous-209 FSPs, which were only identified in the two MSI-H lesions and in the MSI-L lesion (Extended Data Fig. 5b).Fig.

Cancer interception aims to halt or reverse carcinogenesis at early stages by targeting aberrant cells within precancers. Vaccine-based interception strategies that induce tumor-specific T cells are especially promising in the context of hereditary cancer predisposition syndromes, where the immune microenvironment of precancers may be less immunosuppressive compared to that of advanced cancers, where immune suppression often blocks T cell activity against tumor antigens, thus preventing effective immune responses. Indeed, for solid tumors, there has been notable recent progress in the development of novel neoAg-based cancer vaccines targeting micrometastatic disease in the adjuvant setting. For example, a recent randomized phase 2 trial of adjuvant personalized mRNA vaccine combined with immune checkpoint blockade in participants with resected high-risk melanoma demonstrated significantly improved recurrence-free survival versus checkpoint blockade alone12. Similarly, among participants with resected pancreatic adenocarcinoma, a phase 1 study of a personalized neoAg mRNA vaccine combined with checkpoint blockade in the adjuvant setting demonstrated a strong correlation between induction of neoAg-specific T cell responses and delayed tumor recurrence14. In both of these examples, the approaches relied on a pipeline of personalized vaccine design and production informed by somatic mutation profiling of surgical tumor specimens. However, when transitioning from adjuvant therapy to cancer interception, personalized vaccine approaches have limited practical application, primarily because of the absence of bulk tumor tissue for sequencing. Therefore, we propose that off-the-shelf vaccine strategies are optimally suited for clinical development towards cancer interception. By focusing on recurrent and shared antigen targets, off-the-shelf vaccines avoid the potential pitfalls and resource-intensity of personalized vaccine design and production.

equencing. Therefore, we propose that off-the-shelf vaccine strategies are optimally suited for clinical development towards cancer interception. By focusing on recurrent and shared antigen targets, off-the-shelf vaccines avoid the potential pitfalls and resource-intensity of personalized vaccine design and production. Indeed, recent efforts have demonstrated the feasibility and preliminary signals of activity of an adjuvant off-the-shelf vaccine against the most common and recurrent somatic KRAS mutations observed in colorectal and pancreatic cancers19.

equencing. Therefore, we propose that off-the-shelf vaccine strategies are optimally suited for clinical development towards cancer interception. By focusing on recurrent and shared antigen targets, off-the-shelf vaccines avoid the potential pitfalls and resource-intensity of personalized vaccine design and production. Indeed, recent efforts have demonstrated the feasibility and preliminary signals of activity of an adjuvant off-the-shelf vaccine against the most common and recurrent somatic KRAS mutations observed in colorectal and pancreatic cancers19. LS carriers represent a well-defined, high-risk population of individuals who are likely to benefit from novel interception strategies against colorectal and other cancers8. Defects in the DNA MMR pathway lead to the development of MSI in LS-associated neoplasms (precancers and cancers) and the accumulation of many shared FSP. These peptides are non-self-antigens and foreign to the immune system, thus representing potentially the most immunogenic and safest type of neoAg. The rationale behind the use of a preventive neoAg-based vaccination study with Nous-209 relies on the induction of a broad T cell response against shared mutations targeting precancers and early-stage tumors to prevent cancer progression in this population. We reasoned that targeting a large number of neoAg is the most effective approach to achieve tumor control across a broad population, maximizing the likelihood of engaging and eradicating tumor cells. Therefore, a vaccine designed to target a wide range of shared neoAg could be particularly valuable in anticipating and targeting mutations that may emerge during tumorigenesis. One key distinction between the vaccine platform used in this study and other neoAg vaccines is that the latter have generally targeted a more limited number of neoAgs, as not all platforms allow for the inclusion of 209 neoAgs for an overall length of 6,000 aa.

targeting mutations that may emerge during tumorigenesis. One key distinction between the vaccine platform used in this study and other neoAg vaccines is that the latter have generally targeted a more limited number of neoAgs, as not all platforms allow for the inclusion of 209 neoAgs for an overall length of 6,000 aa. In this study, we report the safety and immunogenicity of Nous-209 assessed in a phase 1b/2 trial in healthy LS carriers. Nous-209 was safe in all treated participants with no treatment-related SAEs reported and with mostly mild reactions to the vaccines, consistent with expected safety profile of heterologous prime and boost vaccines with GAd and MVA vectors20,21. NeoAg-specific T cell response was elicited in all evaluable participants after Nous-209 by ex vivo IFNγ ELISpot, with induction of broad and polytopic responses targeting multiple FSPs. Although we primarily focused on characterizing CD8+ T cell responses, induction of tumor-specific CD4+ T cells was also observed following vaccination, in line with previous data on the capacity of the platform to induce both component of the T cell immunity13,22,23. This is a desirable feature, as several pieces of evidence suggest that promoting both subsets of T cells is beneficial for effective antitumor immunity24,25. Vaccination resulted in the generation of antigen-specific CD8+ with effector memory phenotype. In particular, the observed predominance of TemRA (CCR7⁻CD45RA+) cells among the dextramer+ CD8+ population possibly reflects a repeated antigen exposure in persons with LS, who may experience ongoing presentation of frameshift-derived neoAgs in MSI precancer lesions, contributing to a skewing toward more differentiated effector memory phenotype.

predominance of TemRA (CCR7⁻CD45RA+) cells among the dextramer+ CD8+ population possibly reflects a repeated antigen exposure in persons with LS, who may experience ongoing presentation of frameshift-derived neoAgs in MSI precancer lesions, contributing to a skewing toward more differentiated effector memory phenotype. Overall, these results are consistent with the findings from a phase 1 study in metastatic MMRd cancer showing that Nous-209 elicits broad T cell responses in most treated participants18. Interestingly, a recurrent immunogenic FSP (derived from SPEF2) was characterized after vaccination both in LS carriers and in participants with metastatic disease18, with evidence of in vitro induction of cytotoxic T cells determined by the expression of degranulation marker CD107a in LS and trafficking in the tumor bed in vivo in metastatic disease, concomitant with clinical response. Taken together, these findings, including the in vitro cytotoxic activity against a tumor cell line expressing a second recurrent antigen targeted by Nous-209, are supportive for the tumor killing activity of the vaccine-induced T cells.

n the tumor bed in vivo in metastatic disease, concomitant with clinical response. Taken together, these findings, including the in vitro cytotoxic activity against a tumor cell line expressing a second recurrent antigen targeted by Nous-209, are supportive for the tumor killing activity of the vaccine-induced T cells. The selection of our Nous-209 FSPs was initially validated primarily in sporadic metastatic MMRd tumors26,27 and more recently confirmed in a large dataset of 58 incident MMRd CRCs from LS carriers detected during routine surveillance28, thus providing further evidence of the vaccine’s potential to target early-stage tumors. An important new finding from our study is the demonstration of the presence of Nous-209 FSPs in the MSI precancers within an independent dataset. This observation underscores the potential of Nous-209 to target precancers and highlights its relevance in the context of cancer interception. At the yearly colonoscopy following vaccination, the frequency of study participants in which at least one adenoma was detected (~28%) was similar to the frequency observed at baseline before vaccination and in line with the expectations29,30. Interestingly, genomic analysis of available precancerous lesions from baseline and end-of-study colonoscopies showed a trend toward reduced frequency of MSI precancers after vaccine exposure, in line with the mechanism of action of Nous-209 targeting frameshift mutations occurring as a consequence of the impairment of the MMR system. Moreover, when focusing on advanced adenomas, their baseline detection rate was 4.65%, consistent with previous reports31, whereas, following Nous-209 administration, no participants with advanced adenomas were found. Advanced adenomas are considered CRC precursors and previous reports have shown that the MMRd status is more frequently observed in advanced adenomas than in adenomas7,32. Therefore, although this trial was not powered to detect significant differences in advanced adenomas among participants, we consider the observed reduction in the frequency of advanced adenomas as a potentially encouraging early signal of efficacy. Nonetheless, these data will help inform the development of further larger and randomized studies of Nous-209 as a strategy for immune interception in LS.

advanced adenomas among participants, we consider the observed reduction in the frequency of advanced adenomas as a potentially encouraging early signal of efficacy. Nonetheless, these data will help inform the development of further larger and randomized studies of Nous-209 as a strategy for immune interception in LS. We acknowledge several limitations in our study, including the small sample size that limits the power of correlative analysis. In addition, most participants were self-reported non-Hispanic white, which limits the assessment for differences in safety and immunogenicity across races and ethnicities. By addressing barriers to accrual, future studies with larger cohorts that are representative of all who are affected by LS will be able to confirm the generalizability of our findings. We further acknowledge that deeper characterization of immune responses at the level of individual immunogenic FSPs was limited by the availability and volume of collected blood specimens. While our study provides important safety and immunogenicity signals, a larger randomized study will be essential to evaluate the clinical activity of Nous-209 toward a reduction in cancer incidence in LS. Lastly, except for cardioprotective aspirin, nonsteroidal anti-inflammatory drug use was not allowed during study participation. Given the potential use of aspirin at different dosing as a chemoprevention in LS participants, future studies may need to consider combinatorial strategies.

uction in cancer incidence in LS. Lastly, except for cardioprotective aspirin, nonsteroidal anti-inflammatory drug use was not allowed during study participation. Given the potential use of aspirin at different dosing as a chemoprevention in LS participants, future studies may need to consider combinatorial strategies. Overall, this clinical trial provides important proof-of-concept data of the safety and the robustness of induced immunogenicity of Nous-209 in LS carriers, representing a neoAg vaccine-based approach for LS and supporting its clinical development as a valuable intervention for cancer immune interception.

Our study population comprised individuals aged 18 years or older with a diagnosis of LS, as determined by documented carrier status of a deleterious or pathogenic or suspected to be deleterious or pathogenic (known or predicted to be detrimental or result in loss of function, respectively) germline mutation in MLH1, MSH2/EPCAM, MSH6 or PMS2, identified by a Clinical Laboratory Improvement Amendments-approved laboratory test. Consistent with the primary objectives to evaluate the safety, tolerability and immunogenicity of Nous-209 amongst healthy LS carriers, eligible trial participants had no evidence of active or recurrent invasive cancers for at least 6 months before screening and received no cancer-directed treatment (surgery, systemic therapy, hormonal therapy or radiation) within 6 months before screening. We excluded participants who had histologic evidence of high-grade dysplasia and/or invasive cancer at baseline screening. Eligible participants had adequate organ function and Eastern Cooperative Oncology Group (ECOG) performance status 0–1. Except for cardiopreventive aspirin (<100 mg daily), participants consented to refrain from the use of aspirin, nonsteroidal anti-inflammatory drugs or cyclooxygenase inhibitors for the duration of the study treatment. At the study entry, a total of nine participants reported taking aspirin at a low dose (<81 mg orally daily) for cardiovascular or cancer prevention. Three of them decided to discontinue its use upon recruitment and six continued, with three stating the use for cardiovascular and three stating the use for cancer-preventive reasons. Participants consented to refrain from receiving other vaccinations within the first 10 weeks of initiating study treatment and from receiving adenoviral-based vaccines for the duration of study participation (including postintervention follow-up from week 9 through week 52).

ting the use for cancer-preventive reasons. Participants consented to refrain from receiving other vaccinations within the first 10 weeks of initiating study treatment and from receiving adenoviral-based vaccines for the duration of study participation (including postintervention follow-up from week 9 through week 52). We excluded individuals with active infection, including human immunodeficiency virus, hepatitis B virus (HBV) or hepatitis C virus (HCV) except those with documented laboratory evidence of cleared HBV or HCV infection, individuals with a history of organ allograft or other history of immunodeficiency or individuals with a intercurrent condition requiring systemic treatment with corticosteroids (>10 mg daily of prednisone equivalents) or other immunosuppressive medications within 14 days of study treatment. Females who were pregnant or breastfeeding or planning to become pregnant or men attempting or planning to conceive children within 6 months of the end of study treatment were excluded. Detailed inclusion and exclusion criteria are available in the study protocol (Supplementary Information).

14 days of study treatment. Females who were pregnant or breastfeeding or planning to become pregnant or men attempting or planning to conceive children within 6 months of the end of study treatment were excluded. Detailed inclusion and exclusion criteria are available in the study protocol (Supplementary Information). The trial was a phase 1b/2 single-arm, open-label, multicenter, prospective study originally designed with the coprimary endpoints of safety and immunogenicity following initial vaccination with Nous-209 monotherapy. To achieve a goal of at least 36 individuals evaluable for the primary immunogenicity endpoint, up to 45 participants were enrolled between November 2022 and November 2023 at four institutions (The University of Texas MD Anderson Cancer Center (MDACC), The University of Puerto Rico, Fox Chase Cancer Center and City of Hope) within the National Cancer Institute (NCI) iCAN PREVENT clinical trial consortium. At baseline, all participants underwent standard-of-care screening lower endoscopy (flexible sigmoidoscopy or colonoscopy). Confirmed eligible participants received initial Nous-209 vaccination as a single 1-ml IM injection of GAd20-209-FSPs (nominal concentration of 2 × 1011 viral particles per ml) at week 0 (prime), followed by a single 1-ml IM injection of MVA-209-FSPs (nominal concentration of 2 × 108 infectious units per ml) at week 8 (boost).

d eligible participants received initial Nous-209 vaccination as a single 1-ml IM injection of GAd20-209-FSPs (nominal concentration of 2 × 1011 viral particles per ml) at week 0 (prime), followed by a single 1-ml IM injection of MVA-209-FSPs (nominal concentration of 2 × 108 infectious units per ml) at week 8 (boost). Following key preactivation amendments, protocol version 5 was approved for study initiation in August 2022. In October 2023, protocol version 5.3 was approved, allowing for the addition of a revaccination cohort (cohort 2) in which a subset of eligible participants who completed initial Nous-209 vaccination at week 0 and wk 8 (cohort 1) were then randomized to receive either an MVA-209-FSP IM injection at week 52 or a GAd20-209-FSP IM injection at week 52 followed by an MVA-209-FSP IM injection at week 60. Safety and immunogenicity outcomes related to cohort 2 will be reported in a future manuscript. Separate reporting of cohorts 1 and 2 was permitted by protocol.

domized to receive either an MVA-209-FSP IM injection at week 52 or a GAd20-209-FSP IM injection at week 52 followed by an MVA-209-FSP IM injection at week 60. Safety and immunogenicity outcomes related to cohort 2 will be reported in a future manuscript. Separate reporting of cohorts 1 and 2 was permitted by protocol. Our study was designed and developed by academic authors in collaboration with the Division of Cancer Prevention of the NCI and Nouscom. All authors confirm that the study and analyses were conducted in accordance with the general principles of the Declaration of Helsinki and Good Clinical Practice guidelines of the International Council for Harmonization. Written informed consent was obtained from all study participants. The NCI Central Institutional Review Board and The University of Texas MDACC Institutional Review Board approved this study (protocol nos. MDA21-06-01 and 2022-0065, respectively). Safety monitoring was performed regularly by the Data Safety Monitoring Board of MDACC. The trial was registered on ClinicalTrials.gov (NCT05078866).

ntral Institutional Review Board and The University of Texas MDACC Institutional Review Board approved this study (protocol nos. MDA21-06-01 and 2022-0065, respectively). Safety monitoring was performed regularly by the Data Safety Monitoring Board of MDACC. The trial was registered on ClinicalTrials.gov (NCT05078866). The coprimary endpoints of the trial were safety (rate of grade 2 and 3 AEs) and immunogenicity following initial Nous-209 vaccination. Our primary safety and tolerability endpoint was assessed during the prime and boost vaccination phase (week 0 through week 9 + 7 days). AEs were monitored throughout the study (up to 52 weeks + 14 days following initial vaccination) in all participants who received at least the GAd-209-FSP vaccination at week 0 and were graded according to version 5.0 of the NCI Common Toxicity Criteria for AEs. After each vaccine injection, participants were asked to record symptom reactivity events daily on a memory aid (vaccine report card) for up to 7 days (or up to 8 days after symptom resolution). Protocol-defined injection-site reactions included pain, tenderness, erythema or redness, induration or swelling, itching and bruising, while systemic reactogenicity symptoms included fever, chills, malaise, fatigue, myalgia or muscle aches, headache, nausea, vomiting, anorexia and arthralgia or joint pain.

esolution). Protocol-defined injection-site reactions included pain, tenderness, erythema or redness, induration or swelling, itching and bruising, while systemic reactogenicity symptoms included fever, chills, malaise, fatigue, myalgia or muscle aches, headache, nausea, vomiting, anorexia and arthralgia or joint pain. Our coprimary immunogenicity endpoint was assessed at week 9 and was defined as reactivity to at least one of the 16 FSP pools using an ELISpot assay. Notably, in the case of detection of reactivity pools at baseline, an increase of at least 80% in the preexisting reactivity (measured at baseline) was considered as a positive response to the vaccine. Per protocol, evaluable participants underwent screening lower endoscopy (colonoscopy or flexible sigmoidoscopy) at week 52 ± 14 days (cohort 1) or week 68 ± 14 days (cohort 2) for standard-of-care endoscopic assessment in accordance with local institutional practices for high-risk screening populations and for collection of research biopsies. For all participants, clinical endoscopic biopsies of abnormal mucosa and/or resected polyp specimens (if any) were submitted for routine clinical pathology assessment at each participating study site. As prespecified secondary endpoints, we recorded polyp burden (count, size, histology and presence or absence of high-grade dysplasia) and neoplasia incidence. Additional secondary endpoints are detailed in the study protocol (Supplementary Information).

e clinical pathology assessment at each participating study site. As prespecified secondary endpoints, we recorded polyp burden (count, size, histology and presence or absence of high-grade dysplasia) and neoplasia incidence. Additional secondary endpoints are detailed in the study protocol (Supplementary Information). PBMCs from whole blood were isolated at different time points and cryopreserved at each of the clinical sites before shipment to the central laboratory for immunogenicity assessment at Nouscom. To maintain the functionality of PBMCs, isolation and freezing procedures were completed within a maximum of 8 h from blood collection. PBMCs were isolated using Leucosep Bio-One polypropylene tubes (prefilled; Greiner, Merck) following the manufacturer’s instructions. Cryopreserved cells were thawed, washed, counted and rested overnight before use in immunological assays. A set of 976 recombinant, lyophilized peptides, with the majority of them being 15 aa in length, overlapping by 11 aa and spanning the entire sequence of Nous-209, were produced by JPT Peptide Technologies. Individual FSPs were covered by its specific pool of overlapping peptides and then arranged in 16 peptide pools for immunogenicity assessment, as described below. Lyophilized peptides were reconstituted at 40 mg ml−1 in sterile DMSO (Sigma, D2650), aliquoted and stored at −80 °C. To prepare pools 1–16, the peptides were mixed to a final concentration of 0.4 mg ml−1 for each peptide.

peptides and then arranged in 16 peptide pools for immunogenicity assessment, as described below. Lyophilized peptides were reconstituted at 40 mg ml−1 in sterile DMSO (Sigma, D2650), aliquoted and stored at −80 °C. To prepare pools 1–16, the peptides were mixed to a final concentration of 0.4 mg ml−1 for each peptide. IFNγ ELISpot assays were performed ex vivo in triplicate with 2 × 105 PBMCs per well in R10. PBMCs were resuspended in R10 medium, stimulated with a set of peptides designed to cover the 209 FSPs encoded by the vaccine and arranged into 16 peptide pools (P1–P16) at a final concentration of 3 µg ml−1. Cells were plated in ELISpot plates (human IFNγ ELISpot PLUS kit, Mabtech) and incubated for 18–20 h at 37 °C in a humidified CO2 incubator. At the end of incubation, the ELISpot assay was developed according to the manufacturer’s instructions. Spontaneous cytokine production (background) was measured by incubating PBMCs with medium alone, supplemented with the peptide diluent DMSO (negative control, Sigma-Aldrich), whereas CEFX (JPT Peptide Technologies), a pool of known peptide epitopes for a range of HLA subtypes and different infectious agents, was used as positive control. Results are expressed as SFCs per 106 PBMCs in stimulated cultures after subtracting the DMSO background. A response was considered positive if (1) the number of SFCs per 106 PBMCs ≥ 50 and (2) it was at least twice the DMSO background value. A subject was classified as a responder if reactivity to at least one of the 16 FSP peptide pools is induced after vaccination. If a subject exhibited preexisting reactivity to a peptide pool at baseline, the vaccine was expected to enhance this response by at least 80% in at least one of the 16 FSP peptide pools; in this case, such participants were considered as responders. NeoAg vaccine peptide pools were deconvoluted to identify immunogenic peptides by IFNγ ELISpot assays ex vivo or after in vitro stimulation. ELISpot plates were analyzed on the CTL ImmunoSpot S6 universal analyzer.

at least one of the 16 FSP peptide pools; in this case, such participants were considered as responders. NeoAg vaccine peptide pools were deconvoluted to identify immunogenic peptides by IFNγ ELISpot assays ex vivo or after in vitro stimulation. ELISpot plates were analyzed on the CTL ImmunoSpot S6 universal analyzer. To characterize neoAg-induced CD4+ and CD8+ T cell responses, CD8+ T cells were selectively depleted from the total using anti-CD8 microbeads (Miltenyi Biotech, 130-045-201) following the manufacturer’s instructions. The CD8− cell population was then stimulated with either individual peptides or peptide pools (final concentration 3 μg ml−1) and T cell responses against specific peptides were assessed using an IFNγ ELISpot assay. The depletion efficiency of CD8+ T cells was confirmed by flow cytometry. T cell responses were classified as CD8+ mediated if a significant reduction in IFNγ spot count was observed following CD8+ T cell depletion. Conversely, responses were classified as CD4+ mediated if no substantial change in spot count was detected after depletion

ciency of CD8+ T cells was confirmed by flow cytometry. T cell responses were classified as CD8+ mediated if a significant reduction in IFNγ spot count was observed following CD8+ T cell depletion. Conversely, responses were classified as CD4+ mediated if no substantial change in spot count was detected after depletion PBMCs were thawed and rested for 1 h at 37 °C in R10. The samples were then incubated with anti-human Fc block (Pharmingen BD) at a 1:50 dilution in fluorescence-activated cell sorting (FACS) buffer (1× PBS and 0.5% FBS) for 20 min at 4 °C. After washing with FACS buffer, the samples were stained with HLA class I dextramer HLA-B*0801/IAKKRIKL 8-mer peptide (SPEF2 FSP) conjugated to PE at room temperature for 30 min, protected from light. The live/dead near-IR dead cell stain kit (Thermo Fisher Scientific, L10119) was added at 1:100 dilution and cells were incubated for 15 min at room temperature. The following surface staining antibodies were added: CD4 (BioLegend, clone A161A1, 357406), CD8 (BioLegend, clone SK1, 344710), CD45RA (BioLegend, clone HI100, 304142) and CCR7 (CD197; BioLegend, clone G043H7, 353226). After a 30-min incubation at room temperature, the samples were washed and resuspended in FACS buffer until acquisition. For intracellular staining, PBMCs were stimulated with the vaccine single peptide (from SPEF2, 11-mer) (4 µg ml−1), DMSO (control) and PMA/ionomycin cell stimulation cocktail (positive control; Affymetrix) in the presence of anti-human CD107a (LAMP1; BioLegend, clone H4A3, 328618). After overnight coculture, BD GolgiPlug transport inhibitors (BD, 51-2301KZ) were added. Following 3 h of incubation at 37 °C, the samples were washed with FACS buffer and incubated with Fc block for 20 min at 4 °C. The samples were then washed and stained with live/dead dye in staining buffer for 15 minutes at room temperature. For surface staining, cells were labeled with CD4 (BioLegend, clone A161A1, 357406) and CD8 (BioLegend, clone SK1, 344710) antibodies. After a 15-min incubation at room temperature, the samples were washed twice before fixation and permeabilization with CytoFix/CytoPerm (BD Cytofix/Cytoperm kit) for 15 min at 4 °C. The samples were washed twice and resuspended in 1× perm/wash buffer (BD Cytofix/Cytoperm kit) for intracellular staining with IFNγ (BioLegend, clone 4S.B3, 502532). After 30 min of incubation at room temperature, the samples were washed and then resuspended in perm/wash buffer until acquisition.

t) for 15 min at 4 °C. The samples were washed twice and resuspended in 1× perm/wash buffer (BD Cytofix/Cytoperm kit) for intracellular staining with IFNγ (BioLegend, clone 4S.B3, 502532). After 30 min of incubation at room temperature, the samples were washed and then resuspended in perm/wash buffer until acquisition. Data were acquired on a BD FACS Canto II and analyzed using FlowJo (version 10.1). The gating strategies are provided in Supplementary Fig. 2.

t) for 15 min at 4 °C. The samples were washed twice and resuspended in 1× perm/wash buffer (BD Cytofix/Cytoperm kit) for intracellular staining with IFNγ (BioLegend, clone 4S.B3, 502532). After 30 min of incubation at room temperature, the samples were washed and then resuspended in perm/wash buffer until acquisition. Data were acquired on a BD FACS Canto II and analyzed using FlowJo (version 10.1). The gating strategies are provided in Supplementary Fig. 2. To investigate neoAg-specific immune responses, we genetically modified the MMRd HCT116 colon cancer cell line, which harbors an MLH1 mutation, to express multiple HLA alleles and neoAg minigenes (MGs) derived from the CDC7 gene mutation. These MGs encode neoAg peptides of varying amino acid lengths (specifically 50-mer, 15-mer and 9-mer) designed to be presented by both MHC class I and class II alleles, as informed by previous studies7. The codon-optimized CDC7 50-mer construct was cloned into a lentiviral backbone obtained from VectorBuilder (plasmid VB240607-1533chz). The design of the construct included an N-terminal signal peptide (MSPMRVTAPRTLILLLSGALALTETWAGS), the mutated CDC7 epitope containing the HLA-A*03:01-restricted 15-mer (TSRILNLQVLKKILR) with its minimal 9-mer core (TSRILNLQV), an MHC I trafficking domain (MCLRLRTKLEKALSALFIWPQHSYKIVGIVAGLAVLAVVVIGAVVATVMCRRKSSGG) and a flexible C-terminal linker (KGGSYSQAASSDSAQGSDVSLTA). This configuration ensured efficient routing of the CDC7 construct through the secretory pathway and enhanced antigen processing and presentation for recognition by neoAg-specific T cells. HCT116 cells were transfected to express HLA-A11:02, A24:01, A03:01 and B07:02, in addition to their endogenous alleles (HLA-A02:01, A01:01, and B45:01). The transfection process followed a two-step protocol. In the first step, HLA allele expression cassettes were cloned into PiggyBac transposon vectors under constitutive promoters and cotransfected with a transposase helper plasmid into HCT116 cells using Lipofectamine 3000 according to the manufacturer’s instructions. The fluorescent reporters (red fluorescent protein, blue fluorescent protein and enhanced green fluorescent protein (EGFP)) were included to enable tracking of transgene expression and selection of stable clones by drug selection, followed by flow cytometry sorting to ensure the expression of all transgenes within single cells. In the second step, HCT116 cells carrying the HLA constructs were transduced with ready-to-use lentiviral particles generated by VectorBuilder carrying the CDC7 50-mer construct.

ection of stable clones by drug selection, followed by flow cytometry sorting to ensure the expression of all transgenes within single cells. In the second step, HCT116 cells carrying the HLA constructs were transduced with ready-to-use lentiviral particles generated by VectorBuilder carrying the CDC7 50-mer construct. Cells were plated 1 day before infection, exposed to viral supernatant supplemented with 8 μg ml−1 polybrene and spinoculated at 800g for 90 min at 32 °C to enhance transduction efficiency. After overnight incubation, the medium was replaced with fresh complete DMEM and cells were expanded for 5–7 days. Puromycin selection was applied to enrich for stable integrants, which were further purified by FACS on the basis of EGFP expression. Stable clones were subsequently validated by PCR and sequencing for integration, by western blot and flow cytometry for expression and by functional assays to confirm HLA surface expression and CDC7 MG presentation.

ed to enrich for stable integrants, which were further purified by FACS on the basis of EGFP expression. Stable clones were subsequently validated by PCR and sequencing for integration, by western blot and flow cytometry for expression and by functional assays to confirm HLA surface expression and CDC7 MG presentation. For the functional evaluation of neoAg-specific immune responses, the 15-mer peptide CDC7 (TSRILNLQVLKKILR), which is restricted to MHC I, was used for in vitro validation. The peptide was synthesized by JPT Peptide Technologies and used to stimulate PBMCs derived from LS trial participant 12 collected at baseline and at week 8 after Nous-209. PBMCs were cultured in R10 medium (RPMI 1640 with L-glutamine (Corning, 10040CV), 10% heat-inactivated FBS (HyClone, SH30070.03), 10 mM HEPES buffer (Corning, 25060-CI) and 1× penicillin–streptomycin (Corning, 30002CI)), supplemented with 330 U per ml recombinant human IL-7. Stimulation was carried out using 4 µg ml−1 CDC7 peptide per well, with concanavalin A and DMSO serving as positive and negative controls, respectively. On days 3, 7 and 10, the cells were replenished with R10 medium containing 10 U per ml interleukin 2 (IL-2). After 12 days of culture, the expanded T cells were divided into two populations; one half was used for microfluidic coculture assays to assess tumor cell-targeting potency, while the other half was subjected to IL-2 withdrawal for subsequent IFNγ ELISpot analysis. On day 13, the second population of stimulated PBMCs was seeded in triplicate (3 × 105 cells per well) into 96-well ELISpot plates precoated with human IFNγ capture antibodies. The cells were restimulated with 3 µg ml−1 CDC7 peptide and incubated for 16–20 h. IFNγ secretion was measured using the ELISpot assay, following the manufacturer’s protocol (Mabtech). SFCs were quantified using the ImmunoSpot S6 UNIVERSAL analyzer (Cellular Technology Limited).

ated with human IFNγ capture antibodies. The cells were restimulated with 3 µg ml−1 CDC7 peptide and incubated for 16–20 h. IFNγ secretion was measured using the ELISpot assay, following the manufacturer’s protocol (Mabtech). SFCs were quantified using the ImmunoSpot S6 UNIVERSAL analyzer (Cellular Technology Limited). To evaluate the tumor-targeting potency of neoAg-stimulated T cells, experiments were performed using the OrganoPlate three-lane 64 microfluidic system (MIMETAS). A collagen I extracellular matrix (ECM) gel was prepared by mixing collagen I (5 mg ml−1; AMSbio), 1 M HEPES (Thermo Fisher Scientific) and NaHCO3 (Sigma) in a 1:1:8 ratio on ice. Then, 2 μl of the gel mixture was loaded into each gel inlet and polymerized in a humidified incubator at 37 °C for 15 min. Genetically modified HCT116 cells expressing CDC7 MG derived from the CDC7 gene mutation, encoding peptides of varying amino acid lengths, were used in these experiments. Cells were labeled with the fluorescent dye CellTracker red CMTPX (Thermo Fisher Scientific) and seeded into the top perfusion inlets adjacent to the ECM gel channels at a density of 10,000 cells per µl (2 µl per chip). The plate was incubated on its side for 3–4 h to ensure cell attachment to the ECM gel. NeoAg-stimulated PBMCs were labeled with the fluorescent dye CellTracker green CMFDA (Thermo Fisher Scientific) and seeded into the bottom perfusion channels at a 5:1 E:T ratio. Following cell seeding, 50 µl of culture medium was added to all perfusion inlets and outlets to ensure complete filling of the channels without air bubbles. The coculture system was maintained on a rocker platform (14° inclination, 8-min intervals) for 48 h to enable continuous medium perfusion. After the 48-h incubation, genetically modified HCT116 cells and control cells were collected from the microfluidic chamber outlets. Migrated T cells were depleted from the collected samples using CD3 MicroBeads (Miltenyi Biotec) and column-based magnetic separation according to the manufacturer’s instructions. This process ensured the efficient removal of CD3+ T cells, leaving a purified tumor cell population for downstream analyses. Tumor cell viability was subsequently assessed using the CellTiter-Glo luminescent cell viability assay (Promega). Apoptotic cell death was quantified using a caspase 3/7 luminescence assay, performed in accordance with the manufacturer’s protocols.

lls, leaving a purified tumor cell population for downstream analyses. Tumor cell viability was subsequently assessed using the CellTiter-Glo luminescent cell viability assay (Promega). Apoptotic cell death was quantified using a caspase 3/7 luminescence assay, performed in accordance with the manufacturer’s protocols. For viability, equal volumes of reagent and culture medium were added to samples, inducing cell lysis and generating a stable luminescent signal proportional to intracellular adenosine triphosphate content. Luminescence was recorded using a plate luminometer, expressed as relative luminescence units, normalized to untreated control wells and reported as percentage viability relative to control. For apoptosis, equal volumes of Caspase-Glo 3/7 reagent and sample were combined, resulting in caspase-mediated cleavage of a luminogenic DEVD substrate and light emission proportional to caspase 3/7 activity. Luminescence was recorded on a plate luminometer, normalized to control wells and expressed as percentage apoptosis relative to baseline.

s of Caspase-Glo 3/7 reagent and sample were combined, resulting in caspase-mediated cleavage of a luminogenic DEVD substrate and light emission proportional to caspase 3/7 activity. Luminescence was recorded on a plate luminometer, normalized to control wells and expressed as percentage apoptosis relative to baseline. Genomic DNA was extracted from five serial slides from formalin-fixed paraffin-embedded (FFPE) blocks of endoscopic resections of colorectal adenomas from on-study colonoscopies. First, we deparaffinized the tissue sections with xylene and 100% alcohol. Then, tissues were collected and incubated with lysis buffer in the presence of proteinase K using the Roche microRNA isolation kit. Lysates were centrifuged for 30 min at 4 °C at 15,000 rpm and cell pellets were used for extraction of genomic DNA using the AllPrep DNA/RNA FFPE kit (Qiagen), following the manufacturer’s protocol. DNA quality was assessed using TapeStation analyzer; then, Twist exome capture, library preparation and raw sequencing were performed by the Advanced Technology Genomics Core at The University of Texas MDACC using the Illumina NovaSeqX platform. Alignment of WES data was performed using BWA-mem (version 0.7.19) with default parameters to human genome reference hg38. Duplicate reads were marked with GATK (version 4.6.2.0). Base quality recalibration was performed with GATK Apply BQSR.

s Core at The University of Texas MDACC using the Illumina NovaSeqX platform. Alignment of WES data was performed using BWA-mem (version 0.7.19) with default parameters to human genome reference hg38. Duplicate reads were marked with GATK (version 4.6.2.0). Base quality recalibration was performed with GATK Apply BQSR. Tumor exome data were processed starting from the raw data (FASTQ files), which were downloaded from the National Center for Biotechnology Information under BioProject PRJNA954699. A preliminary quality control of the raw sequence data was performed by filtering out reads of low quality with Trimmomatic (version 0.33)33. The remaining reads were aligned on the GRCh37 human genome BWA-mem (version 0.7.17-r1188)34. Multimapping reads were filtered out using SAMtools (version 1.9)35. Optical duplicates were marked using Picard’s MarkDuplicates tool with Picard tools. DNA alignments were further optimized at regions around indels and base scores were recalibrated after the optimization step using GATK software (version 3.7)36. Frameshift mutations within the Nous-209 neoAgs were identified from aligned sequencing data (BAM files) using a lookup-based approach. A mutation was considered present if a minimum of three reads supported the variant and the variant allele frequency (VAF) exceeded 10%, consistent with previously established criteria.27

meshift mutations within the Nous-209 neoAgs were identified from aligned sequencing data (BAM files) using a lookup-based approach. A mutation was considered present if a minimum of three reads supported the variant and the variant allele frequency (VAF) exceeded 10%, consistent with previously established criteria.27 MSI status of previously published datasets7 and colorectal adenomas from on-study colonoscopies was determined using MSIsensor2 (version 0.1; https://github.com/niu-lab/msisensor2.git)37 in ‘tumor-only’ mode. The pipeline involved indexing the reference genome (hg19), scanning 2,793 MS sites and calculating MSI scores on the basis of the proportion of unstable loci. A sample was classified as MSI-H if the MSI score exceeded the predefined threshold of 20%, MSI-L if the MSI score was between 10% and 20% and MSS if the score was <10%, according to the recommended cutoffs from MSIsensor2. MHC class I binding affinity predictions were performed using the Immune Epitope Database and Analysis Resource MHC I prediction tool (https://www.iedb.org/). Peptide sequences of interest were analyzed for their potential to bind HLA class I molecules using the Consensus method (version 2.18) with peptide lengths of 8, 9 and 10 aa. The analysis included the prediction of binding affinities (half-maximal inhibitory concentration values) for HLA alleles. Downstream immunogenicity analyses were performed, prioritizing peptides on the basis of the results of MHC class I binding affinity predictions.

rsion 2.18) with peptide lengths of 8, 9 and 10 aa. The analysis included the prediction of binding affinities (half-maximal inhibitory concentration values) for HLA alleles. Downstream immunogenicity analyses were performed, prioritizing peptides on the basis of the results of MHC class I binding affinity predictions. The trial was conducted using Simon’s minimax two-stage design, wherein the immunogenicity response rate was defined by the number of evaluable participants with immunogenicity by ELISpot assay among all treated participants. On the basis of prior evidence38, our primary efficacy endpoint was evaluated with respect to a predefined target immunogenicity response rate of ≥75%; by contrast, we considered a immunogenicity response rate of ≤55% to be unacceptable. In the first stage, 24 participants were enrolled and accrual halted to fully evaluate immunogenicity at week 9; with 16 or more responses observed in the first stage, additional participants enrolled in the study to reach a total of 36 evaluable participants. Upon study completion, Nous-209 vaccination was considered effective if >24 participants demonstrated immunogenicity. Under these operating characteristics, if the true immunogenicity response rate is 0.55, the probability of stopping the trial early was 83% at an expected sample size 26. Immunogenicity rates are reported with 95% exact confidence intervals evaluated using the Clopper–Pearson method. Our study applied a Bayesian toxicity monitoring plan in which treatment would be considered unsafe if the estimated rate of unacceptable toxicity (grade 3 of higher treatment-related AEs except for vaccine reactogenicity symptoms) was ≥30% with a probability of ≥70%. Assuming a β prior probability of toxicity with parameters (0.3, 0.7) and considering the Simon’s two-stage design, the trial had a 99.6% chance of stopping early if the true toxicity rate was 50% when the true response rate was 55%. SAS 9.4 was used for exploratory statistical analyses of associations between clinical and demographic factors and immunogenicity. Frequencies and percentages are reported for categorical variables. Summary statistics such as number of nonmissing observations, mean, median, s.d., minimum and maximum are provided for continuous data. The chi-squared test and Fisher’s exact test were used to evaluate the association between categorical variables and vaccine responses.

centages are reported for categorical variables. Summary statistics such as number of nonmissing observations, mean, median, s.d., minimum and maximum are provided for continuous data. The chi-squared test and Fisher’s exact test were used to evaluate the association between categorical variables and vaccine responses. Wilcoxon’s rank sum test or Kruskal–Wallis test was used to compare the distributions of continuous response variables (such as the number positive pools at specific time points), across different demographic or clinical groups. Statistical analyses for cell viability, apoptosis and cytokine secretion were conducted using Prism 10 (GraphPad Software). For cell viability and apoptosis assays, data were analyzed across the following groups: genetically modified HCT116 cells cocultured with CDC7-stimulated PBMCs (patient 12 at baseline and after Nous-209), nongenetically modified HCT116 cells cocultured with CDC7-stimulated patient 12 PBMCs at baseline and after Nous-209 (control group) and genetically modified HCT116 cells cultured alone (CDC7-HCT116 control). Statistical comparisons were performed using a two-sided unpaired t-test, with significance indicated in the figure (P values < 0.0001). For the CTL-related IFNγ ELISpot secretion assay, paired t-tests were used to compare PBMCs at baseline and after Nous-209, with comparisons to the negative control (DMSO) and results expressed as the mean ± s.d.; statistical significance was defined as P < 0.05. Levels of significance are indicated as follows: *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001. Significant differences are annotated in the respective figures for clarity. All other ELISpot data were presented as the mean ± s.e.m., with significance tested using a two-tailed, Mann–Whitney statistical analysis.

evels of significance are indicated as follows: *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001. Significant differences are annotated in the respective figures for clarity. All other ELISpot data were presented as the mean ± s.e.m., with significance tested using a two-tailed, Mann–Whitney statistical analysis. Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

The trial was a phase 1b/2 single-arm, open-label, multicenter, prospective study originally designed with the coprimary endpoints of safety and immunogenicity following initial vaccination with Nous-209 monotherapy. To achieve a goal of at least 36 individuals evaluable for the primary immunogenicity endpoint, up to 45 participants were enrolled between November 2022 and November 2023 at four institutions (The University of Texas MD Anderson Cancer Center (MDACC), The University of Puerto Rico, Fox Chase Cancer Center and City of Hope) within the National Cancer Institute (NCI) iCAN PREVENT clinical trial consortium. At baseline, all participants underwent standard-of-care screening lower endoscopy (flexible sigmoidoscopy or colonoscopy). Confirmed eligible participants received initial Nous-209 vaccination as a single 1-ml IM injection of GAd20-209-FSPs (nominal concentration of 2 × 1011 viral particles per ml) at week 0 (prime), followed by a single 1-ml IM injection of MVA-209-FSPs (nominal concentration of 2 × 108 infectious units per ml) at week 8 (boost).

The coprimary endpoints of the trial were safety (rate of grade 2 and 3 AEs) and immunogenicity following initial Nous-209 vaccination. Our primary safety and tolerability endpoint was assessed during the prime and boost vaccination phase (week 0 through week 9 + 7 days). AEs were monitored throughout the study (up to 52 weeks + 14 days following initial vaccination) in all participants who received at least the GAd-209-FSP vaccination at week 0 and were graded according to version 5.0 of the NCI Common Toxicity Criteria for AEs. After each vaccine injection, participants were asked to record symptom reactivity events daily on a memory aid (vaccine report card) for up to 7 days (or up to 8 days after symptom resolution). Protocol-defined injection-site reactions included pain, tenderness, erythema or redness, induration or swelling, itching and bruising, while systemic reactogenicity symptoms included fever, chills, malaise, fatigue, myalgia or muscle aches, headache, nausea, vomiting, anorexia and arthralgia or joint pain. Our coprimary immunogenicity endpoint was assessed at week 9 and was defined as reactivity to at least one of the 16 FSP pools using an ELISpot assay. Notably, in the case of detection of reactivity pools at baseline, an increase of at least 80% in the preexisting reactivity (measured at baseline) was considered as a positive response to the vaccine.

point was assessed at week 9 and was defined as reactivity to at least one of the 16 FSP pools using an ELISpot assay. Notably, in the case of detection of reactivity pools at baseline, an increase of at least 80% in the preexisting reactivity (measured at baseline) was considered as a positive response to the vaccine. Per protocol, evaluable participants underwent screening lower endoscopy (colonoscopy or flexible sigmoidoscopy) at week 52 ± 14 days (cohort 1) or week 68 ± 14 days (cohort 2) for standard-of-care endoscopic assessment in accordance with local institutional practices for high-risk screening populations and for collection of research biopsies. For all participants, clinical endoscopic biopsies of abnormal mucosa and/or resected polyp specimens (if any) were submitted for routine clinical pathology assessment at each participating study site. As prespecified secondary endpoints, we recorded polyp burden (count, size, histology and presence or absence of high-grade dysplasia) and neoplasia incidence. Additional secondary endpoints are detailed in the study protocol (Supplementary Information).

PBMCs from whole blood were isolated at different time points and cryopreserved at each of the clinical sites before shipment to the central laboratory for immunogenicity assessment at Nouscom. To maintain the functionality of PBMCs, isolation and freezing procedures were completed within a maximum of 8 h from blood collection. PBMCs were isolated using Leucosep Bio-One polypropylene tubes (prefilled; Greiner, Merck) following the manufacturer’s instructions. Cryopreserved cells were thawed, washed, counted and rested overnight before use in immunological assays.

A set of 976 recombinant, lyophilized peptides, with the majority of them being 15 aa in length, overlapping by 11 aa and spanning the entire sequence of Nous-209, were produced by JPT Peptide Technologies. Individual FSPs were covered by its specific pool of overlapping peptides and then arranged in 16 peptide pools for immunogenicity assessment, as described below. Lyophilized peptides were reconstituted at 40 mg ml−1 in sterile DMSO (Sigma, D2650), aliquoted and stored at −80 °C. To prepare pools 1–16, the peptides were mixed to a final concentration of 0.4 mg ml−1 for each peptide.

IFNγ ELISpot assays were performed ex vivo in triplicate with 2 × 105 PBMCs per well in R10. PBMCs were resuspended in R10 medium, stimulated with a set of peptides designed to cover the 209 FSPs encoded by the vaccine and arranged into 16 peptide pools (P1–P16) at a final concentration of 3 µg ml−1. Cells were plated in ELISpot plates (human IFNγ ELISpot PLUS kit, Mabtech) and incubated for 18–20 h at 37 °C in a humidified CO2 incubator. At the end of incubation, the ELISpot assay was developed according to the manufacturer’s instructions. Spontaneous cytokine production (background) was measured by incubating PBMCs with medium alone, supplemented with the peptide diluent DMSO (negative control, Sigma-Aldrich), whereas CEFX (JPT Peptide Technologies), a pool of known peptide epitopes for a range of HLA subtypes and different infectious agents, was used as positive control. Results are expressed as SFCs per 106 PBMCs in stimulated cultures after subtracting the DMSO background. A response was considered positive if (1) the number of SFCs per 106 PBMCs ≥ 50 and (2) it was at least twice the DMSO background value. A subject was classified as a responder if reactivity to at least one of the 16 FSP peptide pools is induced after vaccination. If a subject exhibited preexisting reactivity to a peptide pool at baseline, the vaccine was expected to enhance this response by at least 80% in at least one of the 16 FSP peptide pools; in this case, such participants were considered as responders. NeoAg vaccine peptide pools were deconvoluted to identify immunogenic peptides by IFNγ ELISpot assays ex vivo or after in vitro stimulation. ELISpot plates were analyzed on the CTL ImmunoSpot S6 universal analyzer.

To characterize neoAg-induced CD4+ and CD8+ T cell responses, CD8+ T cells were selectively depleted from the total using anti-CD8 microbeads (Miltenyi Biotech, 130-045-201) following the manufacturer’s instructions. The CD8− cell population was then stimulated with either individual peptides or peptide pools (final concentration 3 μg ml−1) and T cell responses against specific peptides were assessed using an IFNγ ELISpot assay. The depletion efficiency of CD8+ T cells was confirmed by flow cytometry. T cell responses were classified as CD8+ mediated if a significant reduction in IFNγ spot count was observed following CD8+ T cell depletion. Conversely, responses were classified as CD4+ mediated if no substantial change in spot count was detected after depletion

PBMCs were thawed and rested for 1 h at 37 °C in R10. The samples were then incubated with anti-human Fc block (Pharmingen BD) at a 1:50 dilution in fluorescence-activated cell sorting (FACS) buffer (1× PBS and 0.5% FBS) for 20 min at 4 °C. After washing with FACS buffer, the samples were stained with HLA class I dextramer HLA-B*0801/IAKKRIKL 8-mer peptide (SPEF2 FSP) conjugated to PE at room temperature for 30 min, protected from light. The live/dead near-IR dead cell stain kit (Thermo Fisher Scientific, L10119) was added at 1:100 dilution and cells were incubated for 15 min at room temperature. The following surface staining antibodies were added: CD4 (BioLegend, clone A161A1, 357406), CD8 (BioLegend, clone SK1, 344710), CD45RA (BioLegend, clone HI100, 304142) and CCR7 (CD197; BioLegend, clone G043H7, 353226). After a 30-min incubation at room temperature, the samples were washed and resuspended in FACS buffer until acquisition. For intracellular staining, PBMCs were stimulated with the vaccine single peptide (from SPEF2, 11-mer) (4 µg ml−1), DMSO (control) and PMA/ionomycin cell stimulation cocktail (positive control; Affymetrix) in the presence of anti-human CD107a (LAMP1; BioLegend, clone H4A3, 328618). After overnight coculture, BD GolgiPlug transport inhibitors (BD, 51-2301KZ) were added. Following 3 h of incubation at 37 °C, the samples were washed with FACS buffer and incubated with Fc block for 20 min at 4 °C. The samples were then washed and stained with live/dead dye in staining buffer for 15 minutes at room temperature. For surface staining, cells were labeled with CD4 (BioLegend, clone A161A1, 357406) and CD8 (BioLegend, clone SK1, 344710) antibodies. After a 15-min incubation at room temperature, the samples were washed twice before fixation and permeabilization with CytoFix/CytoPerm (BD Cytofix/Cytoperm kit) for 15 min at 4 °C. The samples were washed twice and resuspended in 1× perm/wash buffer (BD Cytofix/Cytoperm kit) for intracellular staining with IFNγ (BioLegend, clone 4S.B3, 502532). After 30 min of incubation at room temperature, the samples were washed and then resuspended in perm/wash buffer until acquisition.

To investigate neoAg-specific immune responses, we genetically modified the MMRd HCT116 colon cancer cell line, which harbors an MLH1 mutation, to express multiple HLA alleles and neoAg minigenes (MGs) derived from the CDC7 gene mutation. These MGs encode neoAg peptides of varying amino acid lengths (specifically 50-mer, 15-mer and 9-mer) designed to be presented by both MHC class I and class II alleles, as informed by previous studies7. The codon-optimized CDC7 50-mer construct was cloned into a lentiviral backbone obtained from VectorBuilder (plasmid VB240607-1533chz). The design of the construct included an N-terminal signal peptide (MSPMRVTAPRTLILLLSGALALTETWAGS), the mutated CDC7 epitope containing the HLA-A*03:01-restricted 15-mer (TSRILNLQVLKKILR) with its minimal 9-mer core (TSRILNLQV), an MHC I trafficking domain (MCLRLRTKLEKALSALFIWPQHSYKIVGIVAGLAVLAVVVIGAVVATVMCRRKSSGG) and a flexible C-terminal linker (KGGSYSQAASSDSAQGSDVSLTA). This configuration ensured efficient routing of the CDC7 construct through the secretory pathway and enhanced antigen processing and presentation for recognition by neoAg-specific T cells. HCT116 cells were transfected to express HLA-A11:02, A24:01, A03:01 and B07:02, in addition to their endogenous alleles (HLA-A02:01, A01:01, and B45:01). The transfection process followed a two-step protocol. In the first step, HLA allele expression cassettes were cloned into PiggyBac transposon vectors under constitutive promoters and cotransfected with a transposase helper plasmid into HCT116 cells using Lipofectamine 3000 according to the manufacturer’s instructions. The fluorescent reporters (red fluorescent protein, blue fluorescent protein and enhanced green fluorescent protein (EGFP)) were included to enable tracking of transgene expression and selection of stable clones by drug selection, followed by flow cytometry sorting to ensure the expression of all transgenes within single cells. In the second step, HCT116 cells carrying the HLA constructs were transduced with ready-to-use lentiviral particles generated by VectorBuilder carrying the CDC7 50-mer construct.

For the functional evaluation of neoAg-specific immune responses, the 15-mer peptide CDC7 (TSRILNLQVLKKILR), which is restricted to MHC I, was used for in vitro validation. The peptide was synthesized by JPT Peptide Technologies and used to stimulate PBMCs derived from LS trial participant 12 collected at baseline and at week 8 after Nous-209. PBMCs were cultured in R10 medium (RPMI 1640 with L-glutamine (Corning, 10040CV), 10% heat-inactivated FBS (HyClone, SH30070.03), 10 mM HEPES buffer (Corning, 25060-CI) and 1× penicillin–streptomycin (Corning, 30002CI)), supplemented with 330 U per ml recombinant human IL-7. Stimulation was carried out using 4 µg ml−1 CDC7 peptide per well, with concanavalin A and DMSO serving as positive and negative controls, respectively. On days 3, 7 and 10, the cells were replenished with R10 medium containing 10 U per ml interleukin 2 (IL-2). After 12 days of culture, the expanded T cells were divided into two populations; one half was used for microfluidic coculture assays to assess tumor cell-targeting potency, while the other half was subjected to IL-2 withdrawal for subsequent IFNγ ELISpot analysis. On day 13, the second population of stimulated PBMCs was seeded in triplicate (3 × 105 cells per well) into 96-well ELISpot plates precoated with human IFNγ capture antibodies. The cells were restimulated with 3 µg ml−1 CDC7 peptide and incubated for 16–20 h. IFNγ secretion was measured using the ELISpot assay, following the manufacturer’s protocol (Mabtech). SFCs were quantified using the ImmunoSpot S6 UNIVERSAL analyzer (Cellular Technology Limited).

To evaluate the tumor-targeting potency of neoAg-stimulated T cells, experiments were performed using the OrganoPlate three-lane 64 microfluidic system (MIMETAS). A collagen I extracellular matrix (ECM) gel was prepared by mixing collagen I (5 mg ml−1; AMSbio), 1 M HEPES (Thermo Fisher Scientific) and NaHCO3 (Sigma) in a 1:1:8 ratio on ice. Then, 2 μl of the gel mixture was loaded into each gel inlet and polymerized in a humidified incubator at 37 °C for 15 min. Genetically modified HCT116 cells expressing CDC7 MG derived from the CDC7 gene mutation, encoding peptides of varying amino acid lengths, were used in these experiments. Cells were labeled with the fluorescent dye CellTracker red CMTPX (Thermo Fisher Scientific) and seeded into the top perfusion inlets adjacent to the ECM gel channels at a density of 10,000 cells per µl (2 µl per chip). The plate was incubated on its side for 3–4 h to ensure cell attachment to the ECM gel. NeoAg-stimulated PBMCs were labeled with the fluorescent dye CellTracker green CMFDA (Thermo Fisher Scientific) and seeded into the bottom perfusion channels at a 5:1 E:T ratio. Following cell seeding, 50 µl of culture medium was added to all perfusion inlets and outlets to ensure complete filling of the channels without air bubbles. The coculture system was maintained on a rocker platform (14° inclination, 8-min intervals) for 48 h to enable continuous medium perfusion. After the 48-h incubation, genetically modified HCT116 cells and control cells were collected from the microfluidic chamber outlets. Migrated T cells were depleted from the collected samples using CD3 MicroBeads (Miltenyi Biotec) and column-based magnetic separation according to the manufacturer’s instructions. This process ensured the efficient removal of CD3+ T cells, leaving a purified tumor cell population for downstream analyses. Tumor cell viability was subsequently assessed using the CellTiter-Glo luminescent cell viability assay (Promega). Apoptotic cell death was quantified using a caspase 3/7 luminescence assay, performed in accordance with the manufacturer’s protocols.

Genomic DNA was extracted from five serial slides from formalin-fixed paraffin-embedded (FFPE) blocks of endoscopic resections of colorectal adenomas from on-study colonoscopies. First, we deparaffinized the tissue sections with xylene and 100% alcohol. Then, tissues were collected and incubated with lysis buffer in the presence of proteinase K using the Roche microRNA isolation kit. Lysates were centrifuged for 30 min at 4 °C at 15,000 rpm and cell pellets were used for extraction of genomic DNA using the AllPrep DNA/RNA FFPE kit (Qiagen), following the manufacturer’s protocol. DNA quality was assessed using TapeStation analyzer; then, Twist exome capture, library preparation and raw sequencing were performed by the Advanced Technology Genomics Core at The University of Texas MDACC using the Illumina NovaSeqX platform. Alignment of WES data was performed using BWA-mem (version 0.7.19) with default parameters to human genome reference hg38. Duplicate reads were marked with GATK (version 4.6.2.0). Base quality recalibration was performed with GATK Apply BQSR.

Tumor exome data were processed starting from the raw data (FASTQ files), which were downloaded from the National Center for Biotechnology Information under BioProject PRJNA954699. A preliminary quality control of the raw sequence data was performed by filtering out reads of low quality with Trimmomatic (version 0.33)33. The remaining reads were aligned on the GRCh37 human genome BWA-mem (version 0.7.17-r1188)34. Multimapping reads were filtered out using SAMtools (version 1.9)35. Optical duplicates were marked using Picard’s MarkDuplicates tool with Picard tools. DNA alignments were further optimized at regions around indels and base scores were recalibrated after the optimization step using GATK software (version 3.7)36. Frameshift mutations within the Nous-209 neoAgs were identified from aligned sequencing data (BAM files) using a lookup-based approach. A mutation was considered present if a minimum of three reads supported the variant and the variant allele frequency (VAF) exceeded 10%, consistent with previously established criteria.27

MSI status of previously published datasets7 and colorectal adenomas from on-study colonoscopies was determined using MSIsensor2 (version 0.1; https://github.com/niu-lab/msisensor2.git)37 in ‘tumor-only’ mode. The pipeline involved indexing the reference genome (hg19), scanning 2,793 MS sites and calculating MSI scores on the basis of the proportion of unstable loci. A sample was classified as MSI-H if the MSI score exceeded the predefined threshold of 20%, MSI-L if the MSI score was between 10% and 20% and MSS if the score was <10%, according to the recommended cutoffs from MSIsensor2.

MHC class I binding affinity predictions were performed using the Immune Epitope Database and Analysis Resource MHC I prediction tool (https://www.iedb.org/). Peptide sequences of interest were analyzed for their potential to bind HLA class I molecules using the Consensus method (version 2.18) with peptide lengths of 8, 9 and 10 aa. The analysis included the prediction of binding affinities (half-maximal inhibitory concentration values) for HLA alleles. Downstream immunogenicity analyses were performed, prioritizing peptides on the basis of the results of MHC class I binding affinity predictions.

The trial was conducted using Simon’s minimax two-stage design, wherein the immunogenicity response rate was defined by the number of evaluable participants with immunogenicity by ELISpot assay among all treated participants. On the basis of prior evidence38, our primary efficacy endpoint was evaluated with respect to a predefined target immunogenicity response rate of ≥75%; by contrast, we considered a immunogenicity response rate of ≤55% to be unacceptable. In the first stage, 24 participants were enrolled and accrual halted to fully evaluate immunogenicity at week 9; with 16 or more responses observed in the first stage, additional participants enrolled in the study to reach a total of 36 evaluable participants. Upon study completion, Nous-209 vaccination was considered effective if >24 participants demonstrated immunogenicity. Under these operating characteristics, if the true immunogenicity response rate is 0.55, the probability of stopping the trial early was 83% at an expected sample size 26. Immunogenicity rates are reported with 95% exact confidence intervals evaluated using the Clopper–Pearson method. Our study applied a Bayesian toxicity monitoring plan in which treatment would be considered unsafe if the estimated rate of unacceptable toxicity (grade 3 of higher treatment-related AEs except for vaccine reactogenicity symptoms) was ≥30% with a probability of ≥70%. Assuming a β prior probability of toxicity with parameters (0.3, 0.7) and considering the Simon’s two-stage design, the trial had a 99.6% chance of stopping early if the true toxicity rate was 50% when the true response rate was 55%. SAS 9.4 was used for exploratory statistical analyses of associations between clinical and demographic factors and immunogenicity. Frequencies and percentages are reported for categorical variables. Summary statistics such as number of nonmissing observations, mean, median, s.d., minimum and maximum are provided for continuous data. The chi-squared test and Fisher’s exact test were used to evaluate the association between categorical variables and vaccine responses.

Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at 10.1038/s41591-025-04182-9.

Supplementary InformationSupplementary Figs. 1 and 2 and clinical study protocol. Reporting Summary Supplementary Tables 1–4Supplementary Table 1: AEs by Common Toxicity Criteria for AEs grade and attribution. Supplementary Table 2: Number of positive (reactive) pools and relative increase at peak from the baseline in females compared to males. Supplementary Table 3: List of immunogenic FSPs identified upon peptide pool deconvolution (pools 1 to 16). Supplementary Table 4: Adenoma counts and characterization at baseline and at end of study. Source Data for Supplementary Fig. 1. Supplementary Figs. 1 and 2 and clinical study protocol. Reporting Summary Supplementary Table 1: AEs by Common Toxicity Criteria for AEs grade and attribution. Supplementary Table 2: Number of positive (reactive) pools and relative increase at peak from the baseline in females compared to males. Supplementary Table 3: List of immunogenic FSPs identified upon peptide pool deconvolution (pools 1 to 16). Supplementary Table 4: Adenoma counts and characterization at baseline and at end of study. Source Data for Supplementary Fig. 1.

Source Data Fig. 1Statistical source data. Source Data Fig. 2Statistical source data. Source Data Fig. 3Statistical source data. Source Data Fig. 4Statistical source data. Source Data Extended Data Fig. 2Statistical source data. Source Data Extended Data Fig. 3Statistical source data. Source Data Extended Data Fig. 4Statistical source data. Source Data Extended Data Fig. 5Statistical source data. Statistical source data. Statistical source data. Statistical source data. Statistical source data. Statistical source data. Statistical source data. Statistical source data. Statistical source data.