A vaccine targeting mutant IDH1 in newly diagnosed glioma

Abstract

Mutated isocitrate dehydrogenase 1 (IDH1) defines a molecularly distinct subtype of diffuse glioma^1-3. The most common IDH1 mutation in gliomas affects codon 132 and encodes IDH1(R132H), which harbours a shared clonal neoepitope that is presented on major histocompatibility complex (MHC) class II^4,5. An IDH1(R132H)-specific peptide vaccine (IDH1-vac) induces specific therapeutic T helper cell responses that are effective against IDH1(R132H)⁺ tumours in syngeneic MHC-humanized mice^4,6-8. Here we describe a multicentre, single-arm, open-label, first-in-humans phase I trial that we carried out in 33 patients with newly diagnosed World Health Organization grade 3 and 4 IDH1(R132H)⁺ astrocytomas (Neurooncology Working Group of the German Cancer Society trial 16 (NOA16), ClinicalTrials.gov identifier NCT02454634). The trial met its primary safety endpoint, with vaccine-related adverse events restricted to grade 1. Vaccine-induced immune responses were observed in 93.3% of patients across multiple MHC alleles. Three-year progression-free and death-free rates were 0.63 and 0.84, respectively. Patients with immune responses showed a two-year progression-free rate of 0.82. Two patients without an immune response showed tumour progression within two years of first diagnosis. A mutation-specificity score that incorporates the duration and level of vaccine-induced IDH1(R132H)-specific T cell responses was associated with intratumoral presentation of the IDH1(R132H) neoantigen in pre-treatment tumour tissue. There was a high frequency of pseudoprogression, which indicates intratumoral inflammatory reactions. Pseudoprogression was associated with increased vaccine-induced peripheral T cell responses. Combined single-cell RNA and T cell receptor sequencing showed that tumour-infiltrating CD40LG⁺ and CXCL13⁺ T helper cell clusters in a patient with pseudoprogression were dominated by a single IDH1(R132H)-reactive T cell receptor.

Fig. 1. Patient characteristics at baseline and SOC treatment.

cTMZ, concomitant TMZ (75 mg m⁻² body surface area (BSA)) daily during radiotherapy; TMZ, monotherapy with TMZ (three cycles); XRT, radiotherapy (30 × 2 Gy, if not specified otherwise in Supplementary Table 3); Low, low grade methylation (Meth.) class; High, high grade methylation class; ND, not determined; WHO, WHO grade of tumour. n = 32 patients. Brain illustration taken from Adobe Stock Standard under License ID 222738500.

Fig. 2. Cellular and humoral immunogenicity of IDH1-vac.

a, b, Semi-quantitative analysis of T cell (a) and B cell (b) immune responses (IR) in all patients in the IDS measured by IFNγ enzyme-linked immunosorbent spot (ELISpot) assay (a) or IDH1 peptide enzyme-linked immunosorbent assay (ELISA) (b) (n = 30 patients). Patients are classified as T cell responders (n = 24 patients) and non-responders (n = 6 patients) on the basis of specific spot count cut-off of 50 as defined in the study protocol. Response for each visit (V) is shown. c, Flow cytometric effector phenotyping of peripheral IDH1-vac-induced T cells (single live CD3⁺ cells) from available patient samples with high MSS (n = 5 patient samples). Relative values after re-stimulation with IDH1(R132H) peptide compared to negative control peptide (myelin oligodendrocyte glycoprotein; MOG) are shown. FC, fold-change. Gating strategy is shown in Extended Data Fig. 4c. d, Correlation of intratumoral IDH1(R132H) peptide presentation at baseline (quantified by PLA signal) with the magnitude and sustainability of specific peripheral T cell responses (quantified by the MSS; see Extended Data Fig. 5). r, Pearson correlation coefficient. Patient ID numbers are shown in a–d.

Fig. 3. Efficacy of IDH1-vac, pseudoprogression and T cell response.

a, Swimmer plot depicting disease progression and interventions for each patient in the SDS (n = 32 patients). b, Simon and Makuch plot of overall and progression-free survival probabilities according to the time-dependent covariate IDH1-vac-induced immune response in the IDS (n = 30 patients). x-axes show time since first diagnosis. c, Exemplary MRI fluid-attenuated inversion recovery (FLAIR) and T1-weighted with contrast enhancement (CE) sequences of PsPD of patient ID01 at visit 12 compared to clinical screening MRI. d, Frequencies of PsPD, stable disease (SD), and progressive disease (PD) according to T cell response types for patients in the IDS. For definition of transient and sustained responses, see Methods. e, Magnitude of best T cell response defined by maximum specific ELISpot count with negative control subtracted according to disease progression. Individual values, median (solid lines), and quartiles (dotted lines) are shown in violin plots. SD, 95% CI 80–183; PsPD, 95% CI 103–228; PD, 95% CI 11–88. Two-sided Kruskal–Wallis test, Dunn’s multiple comparison. f, Mutation-specificity scores and molecular profile of each patient in the IDS (n = 30 patients). Methylation class low grade or ND, n = 20; methylation class high grade, n = 10, of which CDKN2A/B^−/− n = 4, CDKN2A/B^+/− n = 4, and CDKN2A/B^+/+ n = 2 patients. g, Simon and Makuch plot of overall and progression-free survival probabilities according to the time-dependent covariate MSS in molecularly defined methylation class high grade gliomas. n = 10 patients. x-axes show time since first diagnosis.

Fig. 4. Molecular T cell phenotype of IDH1-vac-associated PsPD.

a, IFNγ ELISpot counts of LILs from PsPD of patient ID08 at visit 07 after ex vivo stimulation with indicated reagents (see Methods). Peripheral blood mononuclear cells (PBMCs) stimulated with cytomegalovirus and adenovirus (CMV/AdV) peptides were used as positive assay control. Data shown as individual values and the mean of three technical replicates (for negative control, IDH1(R132H)). Technical unicates for CMV/AdV, dendritic cells (DCs) only, LILs only, CMV/AdV PBMCs. b, UMAP plot depicting molecular clusters defined by single-cell transcriptome of LILs (n = 16,720 cells) from PsPD of patient ID08. c, CXCL13 expression in LILs from PsPD of patient ID08 within clusters as in b. d, Bubble plot mapping top TCR clones in CD4⁺ and CD8⁺ T cells defined by single-cell TCR sequencing onto transcriptomic clusters defined in b. e, T cell activation measured by luciferase NFAT reporter assay after overexpression of a top-five CD4⁺ TCR (TCR14 in d) in human Jurkat T cells and co-culture with peptide-loaded autologous PBMCs. Data depicted as individual values and the mean of three technical replicates. Representative of three independent experiments.

Extended Data Fig. 1. Trial design and recruitment.

a, Patient disposition CONSORT flow diagram. Forty-four patients were enrolled and screened across 7 trial sites, of which 11 were excluded. Of 33 allocated patients, 32 received the intervention. Twenty-eight patients completed the study, because four discontinued the intervention owing to disease progression. The safety dataset (SDS) for analysis contained all patients who received the intervention (n = 32); 30 of these were evaluable for immunogenicity and comprised the immunogenicity dataset (IDS) (n = 30). b, Study flow chart. The trial population comprised three treatment groups (arms) according to the standard therapy received. IDH1-vac was administered at V03 (week 1), V04 (week 3), V05 (week 5) V06 (week 7), V07 (week 11), V08 (week 15), V09 (week 19), and V10 (week 23). Blood for primary endpoint immunogenicity testing was drawn at V03 (baseline), V05, V07, V10, V12 (week 35, safety follow-up (SFU)), and V13 (week 47, EOS). MRI scans (represented by brain images) were performed at clinical screening, V07, V10, V12, and V13. XRT, radiotherapy; TMZ, temozolomide; cTMZ, concomitant TMZ, RT + TMZ; aTMZ, adjuvant TMZ. TMZ cycle numbers indicated.

Extended Data Fig. 2. Serum cytokine levels during treatment with IDH1-vac.

Heat maps depicting longitudinal (V03–V13) cytokine concentrations in sera from transient (n = 16) and sustained (n = 9) T cell responder patients, measured by multiplex bead technology. For definition of transient and sustained responses, see Methods. White, sample not available; red, concentration out of depicted range.

Extended Data Fig. 3. Relationship between MHC alleles and T cell response.

a, Venn diagram of T cell non-responders and B cell non-responders in the IDS. b, c, Allele prevalence of MHC class I supertype families (b), and MHC class II DRB1* alleles with a total prevalence of three or more, and paralogues (c). Grey, numbers of alleles present or absent (for paralogues) in patients with T cell responses to IDH1-vac (T cell response); black, numbers of alleles present or absent (for paralogues) in patients without T cell responses to IDH1-vac (T cell non-response). n (total alleles) = 64 for 32 patients in the SDS. d, IDH1 and IDH1(R132H) 20-mer p123–142 affinities to six MHCII DRB1* alleles and DRB3*, DRB4*, and DRB5* MHCII paralogues were assessed in vitro. Four alleles and DRB4* paralogue are shown as examples in the graphs. CLIP, KLAT, and PADRE represent positive control peptides. GB, good binder; B, intermediate binder; WB, weak binder; NB, non-binder. n = 1 of 2 independent experiments. e, Correlation analysis of in vitro MHC class II affinities with MSS. White, no affinity-tested alleles present; blue, intermediate binder(s) present; green, weak binder(s) present; black, non-binders present. n = 30 patients in the IDS.

Extended Data Fig. 4. T cell immunogenicity and standard treatment in the IDS.

a, b, T cell immune responses assessed by MSS (top) and by immune response criteria for positivity over time (bottom) according to concomitant use of steroids until EOT (a) and primary SOC treatment (b). Top, individual values and median. n(steroids) = 2; n(no steroids) = 28. n(RT) = 5; n(RT + TMZ) = 22; n(TMZ) = 3. Two-sided Kruskal–Wallis test with Dunn’s multiple comparison (b). c, Gating strategy for flow cytometric effector sub-phenotyping of peripheral IDH1-vac-induced T cells shown in Fig. 2c. FM2, fluorescence minus two; FMO, fluorescence minus one; MOG, negative control.

Extended Data Fig. 5. Definition of the MSS and PLA.

a, Exemplary MSS of patients ID12 and ID05. Mathematical differences between specific spot counts for IDH1(R132H) (red) and wild-type IDH1 (blue) are calculated (light blue) for all immunogenicity testing visits. MSS is defined as the sum of the three largest differences. For MSS of the IDS please refer to Supplementary Table 7. b, Exemplary PLA of patients ID12 and ID05. For relative PLA values of the IDS please refer to Fig. 2. Scale bars, 80 μm. PLA was performed once per patient. c, Relative PLA signals of primary and recurrent tissues from patient ID25.

Extended Data Fig. 6. Probabilities of progression and death in the SDS.

Overall (left) and progression-free (right) survival estimates with number of patients at risk are shown for all patients of the SDS (a), according to SOC treatment (b), extent of resection (c), and WHO grade (d). CR, complete resection; ST, subtotal resection. n(all patients) = 32; n(RT, aTMZ) = 4; n(RT) = 2; n(TMZ + RT, aTMZ) = 22; n(TMZ + RT) = 1; n(TMZ) = 3; n(CR) = 17; n(ST) = 12; n(biopsy) = 3; n(grade 3) = 21; n(grade 4) = 11. Time, months from first diagnosis.

Extended Data Fig. 7. Determinants of pseudoprogression.

a, Time points of PsPD diagnosed by MRI and onset of early immune response are shown in weeks and study visits. Black diamond, onset of immune response; red asterisk, PsPD. Visits with immune monitoring (V03, V05, V07, V10, and V12) are depicted. MRI was performed at clinical screening, V07, V10, V12 and V13. The onset of an early immune response was defined as antibody titre ≥ 1:333 (B cell), and/or T cell response detectable after negative control subtraction. n = 12 patients with PsPD. b, Violin plots showing MSS according to disease progression. Solid line, median; dotted lines, quartiles. n(SD) = 15; n(PsPD) = 11; n(PD) = 4. Two-sided Kruskal–Wallis test with Dunn’s multiple comparison. c, Incidence of disease progression in patients with molecularly defined astrocytomas according to CNV-L, methylation class, and CDKN2A/B status. d, Correlation between predicted and actual occurrence of molecular markers in groups of disease outcomes. Two-sided Fisher’s exact test (PsPD versus SD + PD). n(all patients) = 23 (c, d).

Extended Data Fig. 8. Probabilities of progression and death in the molecular dataset.

Overall (left) and progression-free (right) survival estimates with number of patients at risk are shown according to methylation class (a), CDKN2A/B status and grade (b), and CDKN2A/B status and CNV-L (c). H, methylation class high; L, methylation class low; het, heterozygous CDKN2A/B deletion; 0, no CDKN2A/B deletion; homo, homozygous CDKN2A/B deletion. n(all patients) = 24; n(methylation class H) = 10; n(methylation class L) = 14; n(CDKN2A/B het and 0 + grade 3) = 13; n(CDKN2A/B het and 0 + grade 4) = 7; n(CDKN2A/B het and 0 + CNV-L high) = 4; n(CDKN2A/B het and 0 + CNV-L low) = 16; n(CDKN2A/B homo) = 4.

Extended Data Fig. 9. Outcomes, standard treatment and peripheral immune cell composition in the IDS.

a, Frequencies of T cells, monocytes, regulatory T cells (Tregs), and monocytic myeloid-derived suppressor cells (mo-MDSC) within PBMCs were determined at V07 by flow cytometry and are shown according to disease course at EOS (top) and SOC treatment (bottom). CD3⁺, T cells. n(SD) = 15; n(PsPD) = 11; n(PD) = 4; n(RT) = 5; n(RT + TMZ) = 22; n(TMZ) = 3. Red line, median; dotted lines, quartiles. Two-sided Kruskal–Wallis test with Dunn’s multiple comparison. b, Gating strategy for flow cytometric analysis.

Extended Data Fig. 10. Longitudinal TCRB deep sequencing of PBMC.

Tree maps of longitudinal TCRB deep sequencing of PBMCs. Black outlines highlight time points of PsPD. TCRB deep sequencing was performed if PBMCs were available for exploratory analyses. n = 25 patients of the SDS.

Extended Data Fig. 11. Ex vivo cytotoxic T cell responses, CD8⁺ T cell clone-specific cytokine production and TCR-transgenic cell reactivity of PsPD LILs from patient ID08.

a, Top, visualization of TNF- and IFNγ-expressing T cells by UMAP (Fig. 4). Bottom, relative percentages of cells expressing IFNγ and TNF among CD8⁺ cells, each expressing one of the top ten TCRs. b, c, Flow cytometric gating strategy (b) and quantification (c) of ex vivo cytotoxic T cell responses of PsPD CD8⁺ LILs from patient ID08 upon re-stimulation with IDH1(R132H). d, CD8⁺ T cell clonotype-retrieved TCR-transgenic cell reactivity in luciferase NFAT reporter assays. TCR1, overall top abundant ID08 PsPD CD8⁺ T cell clonotype; TCR3, ID08 PsPD CD8⁺ T cell clonotype with top TNF/IFNγ percentage (a). InfHA (influenza HA peptide; PKYVKQNTLKLAT) and its respective TCR-transgenic cells were used as positive control. Technical triplicates with mean, experiment performed once.

Extended Data Fig. 12. Single-cell RNA-TCR-seq and TCRB deep sequencing of PsPD from patient ID08.

a, Bottom left, abundance of T cells relative to non-T cells in PsPD from patient ID08 versus control IDH1(R132H)⁺ astrocytoma and glioblastoma tissues. Dashed line, median; dotted lines, quartiles. Top and right, gating strategy of ID08-PsPD T cell sorting. b, Heat map of RNA expression of single-cell cluster-defining genes (see Fig. 4). c, Co-visualization of top 5 CD8⁺ and CD4⁺ TCR clonotypes and corresponding RNA single-T cell clusters by UMAP (see Fig. 4). d, Combined visualization of TCRB deep sequencing and scRNA-TCR-seq of PBMCs from patient ID08 at V07 versus LILs from PsPD of patient ID08. The frequency of TCR14 is highlighted. e, Longitudinal T cell response of patient ID08 assessed by IFNγ ELISpot.