An Autologous Dendritic Cell Vaccine Promotes Anticancer Immunity in Patients with Ovarian Cancer with Low Mutational Burden and Cold Tumors

Epithelial ovarian carcinoma (EOC) is one of the leading causes of death from gynecologic malignancies, with a 5-year survival rate of 47.8% (1). The majority of women with EOC achieve complete remission after primary or interval cytoreductive surgery followed by standard-of-care (SOC) chemotherapy with a platinum-taxane doublet and bevacizumab. Homologous recombination (HR) defects are key determinants of platinum sensitivity in patients with EOC and provide rationale for maintenance therapy with PARP inhibitors, which is associated with improved progression-free survival (PFS; ref. 2). However, most patients with EOC ultimately relapse and succumb to their disease, calling for the development of novel therapeutic regimens in complement or substitution of current treatments.

The successful implementation of immune checkpoint inhibitors (ICI) in the clinical management of various solid tumors has raised considerable expectations for women with EOC (3, 4). However, EOC is poorly sensitive to ICIs employed as standalone immunotherapeutic agents (5, 6), correlating with limited tumor mutational burden (TMB; refs. 7, 8) and poor infiltration by CD8⁺ cytotoxic T lymphocytes (CTL; refs. 9, 10).

Thus, novel strategies are needed to overcome the frequent lack of preexisting immunity in patients with EOC, potentially including chemotherapies that induce immunogenic cell death (ICD; ref. 11), adoptive T-cell transfer strategies (12, 13), and/or vaccination approaches (14, 15). As most immunotherapies only benefit a small percentage of patients (4), it is also imperative to discover biomarkers that prospectively identify patients with EOC potentially achieving benefits from specific immunotherapeutic regimens, in the setting of personalized cancer immunotherapy (16).

We have recently published the results of an open label, randomized phase II clinical study (SOV01, NCT02107937) comparing the efficacy of an autologous dendritic cell (DC)-based vaccine (DCVAC) delivered in the context of (parallel or sequential) SOC carboplatin plus paclitaxel chemotherapy versus SOC only in patients with EOC (17). In this setting, DCVAC was well tolerated and significantly extended the PFS of women with EOC, but only when administered after SOC chemotherapy (17). Here, we demonstrate that while a higher-than-median TMB and robust tumor infiltration by CD8⁺ T lymphocytes were associated with improved clinical outcome in patients from SOV01 receiving SOC chemotherapy, the same did not hold true for patients receiving DCVAC. Conversely, the clinical benefits of DCVAC were more pronounced in patients with EOC with lower-than-median TMB and scant CD8⁺ T-cell infiltration by CD8⁺ T cells. Pending validation in independent studies, our findings suggest that DCVAC stands out as a promising clinical tool to jumpstart anticancer immunity in patients with immunologically “cold” EOC.

From November 18, 2013 to May 18, 2015, a total of 99 patients with EOC were randomized to receive SOC chemotherapy after debulking surgery, either as a standalone adjuvant intervention (arm C, n = 31), or in the context of parallel (arm A, n = 34) or sequential (arm B, n = 34) DCVAC administration, as part of a randomized phase II, multicenter clinical study (SOV01, NCT02107937; ref. 17). A concise description of the study design is provided in Supplementary Materials and Methods and Supplementary Fig. S1. Two patients from arm B were excluded from the analysis since they did not meet inclusion criteria (17). Informed written consent was obtained according to the Declaration of Helsinki, and the study was approved by appropriate Ethical Committees (201607 S14P). The primary endpoints were safety and PFS, while the secondary endpoint was overall survival (OS). Of 97 patients randomized to treatment, 82 (85%) had samples and data available for the current analysis (Supplementary Table S1). Clinical findings have previously been reported (17). Baseline patient characteristics were similar across treatment groups (Supplementary Table S2). Pathology staging was performed according to the 8^th tumor–node–metastasis (TNM) classification (2017; ref. 18), and histologic types were determined according to the current World Health Organization (WHO) classification.

IHC was performed according to conventional protocols (19–21). Briefly, tumor specimens were fixed in neutral buffered 10% formalin solution and embedded in paraffin as per standard procedures. In brief, 3-μm–thick tissue sections were deparaffinized and rehydrated in a descending alcohol series (100%, 96%, 70%, and 50%), followed by antigen retrieval with Target Retrieval Solution (Leica) in EDTA at pH 8.0 with a heated water bath (97°C, 30 minutes). Sections were allowed to cool down to room temperature for 30 minutes. Endogenous peroxidase was blocked with 3% H₂O₂ for 15 minutes. Thereafter, sections were treated with Protein Block (DAKO) for 15 minutes and incubated with a primary antibody specific for CD8 for 30 minutes, followed by revelation of enzymatic activity with EnVision⁺/horseradish peroxidase (HRP), Rabbit (DAKO) for 20 minutes (Supplementary Table S3). Finally, sections were counterstained with hematoxylin (DAKO) for 15 seconds. Images of whole tumor sections were acquired using a Leica Aperio AT2 scanner (Leica).

Immunofluorescence (IF) with antibodies specific for lysosomal associated membrane protein 3 (LAMP3; best known as DC-LAMP), CD8 and CD20 (Supplementary Table S3) was performed according to conventional protocols (21). Briefly, 3-μm–thick formalin-fixed, paraffin-embedded (FFPE) tissue sections were deparaffinized and rehydrated in a descending alcohol series (100%, 96%, 70%, and 50%), followed by antigen retrieval with Target Retrieval Solution (Leica) in EDTA pH 8.0 with a heated water bath (97°C, 30 minutes). Sections were allowed to cool down to room temperature for 30 minutes, then treated with Signal Enhancer (Thermo Fisher Scientific) for 30 minutes and Blocking Buffer (Thermo Fisher Scientific) for 1 hour. The DC-LAMP–specific antibody was incubated overnight at 4°C, the CD8-specific antibody for 2 hours at room temperature, and the CD20-specific antibody for 1 hour at room temperature. Thereafter, slides were incubated with appropriate HRP Polymer secondary antibodies for 1 hour at room temperature, followed by Tyramide Signal Amplification (Thermo Fisher Scientific). Finally, sections were treated with TrueBlack Lipofuscin Autofluorescence Quencher (Biotium) for 30 seconds and mounted with ProLong Gold Antifade Reagent containing DAPI (Thermo Fisher Scientific; Supplementary Table S3). The specificity of the staining was determined using appropriate isotype controls. Images of whole tumor sections were acquired using a Leica Aperio AT2 scanner (Leica; Supplementary Fig. S2).

Infiltration of tumor nests by CD8⁺ T cells, DC-LAMP⁺ DCs, and CD20⁺ B cells was quantified in whole tumor sections by the Calopix software (Tribvn), as published previously (19). Data are reported as absolute numbers of CD8⁺, DC-LAMP⁺, or CD20⁺ cells/mm². Quantitative assessments were performed by three independent investigators (J. Fucikova, L. Kasikova, T. Lanickova) and independently reviewed by an expert pathologist (J. Laco, A. Ryska).

Blood samples for immunomonitoring were collected at enrollment, at the end of DCVAC treatment, and at follow-up (6 months after treatment; Supplementary Fig. S1). Lymphocyte subsets including CD45⁺ cells; CD3⁺, CD4⁺, and CD8⁺ T cells; CD16⁺CD56⁺ natural killer (NK) cells; and CD19⁺ B cells were assessed in fresh blood samples using CYTO-STAT tetraCHROME CD45-FITC/CD4-RD1/CD8-ECD/CD3-PC5, CYTO-STAT tetraCHROME CD45-FITC/CD56-RD1/CD19-ECD/CD3-PC5 (Beckman Coulter), and anti–HLA-DR (Beckman Coulter) reagents (Supplementary Table S4). Cells were quantified by flow cytometry using a FACS CANTO II analyzer (BD Biosciences) and analyzed using FlowJo (Tree Star, Inc.; Supplementary Figures S3A and S3B).

The frequency of T cells specific for ERBB2 (best known as HER2), MAGE family member A3 (MAGEA3), and mucin 1, cell surface associated (MUC1) was assessed by flow cytometry according to standard procedures. In brief, peripheral blood mononuclear cells (PBMC) were cultured for 10 days together with overlapping peptides spanning the whole sequence of HER2, MAGEA3, and MUC1 (PepMix; JPT Peptide Technologies) at a final concentration of 1 mg/mL. On days 4 and 7, 20 UI/mL IL2 (Gentaur) was added. On day 9, PBMCs were re-stimulated with peptide mixtures, and 5 mg/mL brefeldin A (BioLegend) was added after 4 hours of incubation. Unstimulated cells were used as negative control. Moreover, PBMCs treated with CEFT pool mixture (PepMix, JPT peptide technologies) as for the aforementioned protocol were employed as a positive control. Eventually, PBMCs were co-stained with CD3-Alexa700, CD4-PerCP-Cy5.5, and CD8-eFluor450 (Thermo Fisher Scientific) conjugates plus the Aqua Blue Live/Dead cell viability dye (Life Technologies; Supplementary Table S4) for 25 minutes. Thereafter, cells were fixed with Fixation/Permeabilization Buffer (BD Biosciences) for 30 minutes, further permeabilized with Permeabilization Buffer (BD Biosciences) for 10 minutes and incubated with IFNγ-PE-Cy7 and GZMB-PE (BioLegend) for 30 minutes. Flow cytometry was performed on an LSRFortessa Analyzer (BD Biosciences), and data were analyzed with FlowJo (Tree Star, Inc.). Upon exclusion of dead cells, IFNγ expression was considered to be antigen-specific if the frequency of IFNγ-producing T cells detected in response to peptide stimulation was at least twice the frequency of IFNγ-producing cells detected in control conditions.

A total of 100 ng of RNA isolated from FFPE tumor sections was used to analyze expression of 770 genes involved in immune responses based on the PanCancer Immune Profiling Panel (NanoString Inc.), as per manufacturer's recommendations. Briefly, RNA was hybridized with biotin-labeled capture probes and fluorescent reporter probes for 21 hours at 65°C. Following hybridization, samples were injected into a NanoString SPRINT cartridge and loaded onto the SPRINT profiler. Following image acquisition, mRNA counts were calculated from RCC files using nSolver analysis software v4.0 (NanoString Inc.)

The NGS library was prepared using TruSight Oncology 500 DNA kit (Illumina) as per manufacturer's recommendations. Denaturation of next-generation sequencing (NGS) libraries was achieved by bead-based normalization, according to the manufacturer protocol. Final denaturation was performed according to the NextSeqSystem Denature and Dilute Guide. NGS libraries were sequenced on the NextSeq 550 using NextSeq 500/550 High Output kit v2.5 (300 cycles) in pair-end mode 2 × 101 bp. The TMB analysis was done using the TSO500 LocallApp. Further details are provided in Supplementary Materials and Methods.

Total RNA was isolated with the RNeasy Mini Kit (Qiagen). Cell lysates in RLT buffer enriched with 1% 2-mercaptoethanol were quickly thawed and processed as per manufacturer instructions, including a DNase I digestion step. RNA concentration and purity were determined using a NanoDrop 2000c (Thermo Fisher Scientific). Purified RNA samples were stored at −80°C until further use. cDNA for the detection of selected 96 genes associated with immune system (Supplementary Table S5), was synthesized from 100 ng of total RNA using the TATAA GrandScript cDNA Synthesis Kit (TATAA Biocenter).

This study was conducted as retrospective exploratory study. PFS was defined as the time from randomization to the date of the first radiologic progression or death, whichever came first. OS was calculated as the time from randomization to death from any cause. Survival analyses were estimated by Cox proportional hazard regressions and the Kaplan–Meier method, where differences between the groups of patients were calculated using the log–rank test. For log–rank tests, the prognostic value of continuous variables was assessed using median cut-off of intratumoral CD8⁺ T-cell density and TMB value. The Mann–Whitney test was used to compare the density of tumor-infiltrating cells among patient groups. The Wilcoxon test was used to compare the frequency of immune markers before and after therapy. The Fisher exact test was used to compare patient distribution across subgroups. The enrichGo function from ClusterProfiles was used to identify enriched Gene Ontology (GO) terms based on hypergeometric distribution. P values were adjusted for multiple comparisons using the Benjamini-Hochberg correction. All analyses were performed with Prism 8.4.2 (GraphPad), SAS software V.9.4, and R (http://www.r-project.org/). P values < 0.05 were considered statistically significant.

The data generated in this study are available upon request to the corresponding author.

We first set to harness the TrueSightOnco500 panel to compare mutational profile and TMB in samples from patients with EOC involved in the SOV01 study. We identified a panel of 20 somatic mutations (TP53, BRCA1, ARID5B, BARD1, MED12, PIK3CA, SDHA, BCORL1, KRAS, SPTA1, TET1, ZFHX3, ZNF703, ARID1A, BCOR, FANCD2, GNAS, KDM5A, MDC1, SLX4) that were recurrent in more than 5% of these patients, which were well balanced across study arms (Fig. 1A). Along similar lines, we did not observe a significant difference in baseline TMB in patients from different study arms (Fig. 1B). To elucidate the impact of TMB on the EOC immune contexture, we investigated the relative expression levels of 770 genes associated with immune responses using the Nanostring PanCancer Immune Profiling Panel. We identified a set of 144 genes that were significantly overrepresented in EOC samples from patients with higher-than-median TMB (TMB^Hi) as compared with patients with lower-than-median TMB (TMB^Lo; Fig. 1C; Supplementary Table S6). Importantly, we chose to stratify the patient cohort based on median TMB rather than using the standard cut-off of 10 somatic mutations per megabase (22), because TMB in our cohort ranged from 0 to 10 somatic mutations per megabase (median = 2.2), which is in line with published data (8, 23). Gene enrichment studies based on the enrichGo function from Cluster Profiler revealed a significant association between such differentially expressed genes (DEG) and immune system activation, especially (but not limited to) positive regulation of adaptive immune response, as well as cytotoxic T-cell– and NK-cell–mediated immunity (Supplementary Fig. S4A). To validate these findings with an independent methodology, we analyzed the relationship between TMB status and the intratumoral abundance of CD8⁺ CTLs, DC-LAMP⁺ mature DCs, and CD20⁺ B cells as assessed by multispectral IF (Fig. 1D; Supplementary Fig. S2). We observed significantly higher densities of CD8⁺ CTLs (P = 0.010), mature DC-LAMP⁺ DCs (P = 0.002), and CD20⁺ B cells (P = 0.005) in tumor samples from TMB^Hi versus TMB^Lo EOCs (Fig. 1E). Taken together, these findings indicate that an elevated TMB is associated with improved immune effector functions in the EOC microenvironment.

To assess the prognostic and predictive value of the TMB in our cohort, we evaluated PFS and OS upon stratifying the patients enrolled in the SOV01 study based on median TMB (Fig. 2A). In arm C (patients receiving SOC chemotherapy), the TMB^Hi status was significantly associated with improved PFS and OS (PFS: P = 0.021; OS: P = 0.040; Fig. 2B). To corroborate our findings in an independent patient cohort, we retrieved TMB values for 206 patients with high-grade serous ovarian carcinoma (HGSOC) from The Cancer Genome Atlas (TCGA) database and stratified them based on median TMB. Also in this setting, TMB^Hi patients exhibited improved OS as compared with their TMB^Lo counterparts (P = 0.013; Supplementary Fig. S4B). These data suggest that an elevated TMB may be indicative of an ongoing immune response that favorably affects disease outcome in patients with EOC. In line with this notion, the neoplastic lesions of TMB^Hi patients were abundantly infiltrated by CD8⁺ T cells (Fig. 1E; Supplementary Fig. S4C).

Conversely, TMB^Hi patients from arms A and B (receiving DCVAC) did not exhibit favorable PFS and OS as compared with their TMB^Lo counterparts (Fig. 2C; Supplementary Fig. S4D). To obtain additional insights into the predictive value of the TMB on DCVAC efficacy, we directly compared PFS and OS in patients with TMB^Lo tumors from all study arms. We found that DCVAC conferred an OS advantage to patients with TMB^Lo EOCs irrespective of whether it was administered in parallel with or sequentially to SOC chemotherapy (Arm A: P = 0.020, Arm B: P = 0.010; Fig. 2D and E), as well a PFS advantage to patients with TMB^Lo EOCs, but only when administered after SOC chemotherapy (Arm A: P = 0.165, Arm B: P = 0.007; Fig. 2D and E). Conversely, TMB^Hi patients treated with DCVAC failed to exhibit improved PFS and OS as compared with TMB^Hi patients receiving SOC chemotherapy (Fig. 2F and G). We obtained similar results when arms A and B of the study were analyzed as a single group of patients (OS: P = 0.004; PFS: P = 0.031; Supplementary Fig. S4E and S4F). Univariate Cox analysis further confirmed the increased risk for death associated with an elevated TMB in patients with EOC receiving DCVAC irrespective of treatment schedule (Table 1).

Taken together, these findings indicate that while an elevated TMB is associated with improved disease outcome in patients with EOC treated with SOC chemotherapy, a low TMB status is associated with improved disease outcome in women with EOC receiving DCVAC.

TMB generally correlates with immune infiltration by CD8⁺ CTLs in multiple solid tumors, including EOC (8, 24). Driven by these premises and our TMB findings, we employed the Nanostring PanCancer Immune Profiling Panel to estimate the prognostic value of 770 immune genes in patients with EOC enrolled in the SOV01 study. Univariate Cox analyses identified 90 genes with a positive prognostic value in patients with EOC treated with SOC chemotherapy only (Fig. 3A and Supplementary Table S7), which were globally associated with T-lymphocyte activation and proliferation (Supplementary Fig. S5A). The same genes, however, did not correlate with improved disease outcome in patients with EOC receiving DCVAC, as demonstrated by unsupervised hierarchical clustering and univariate Cox regression analysis (Fig. 3A; Supplementary Table S7).

To validate these findings with an independent approach, we harnessed IHC for quantifying tumor-infiltrating CD8⁺ CTLs in pretreatment tumor samples (Fig. 3B; Supplementary Fig. S5B) and assessing their impact on PFS and OS in patients from the SOV01 study. Importantly, tumor-infiltration by CD8⁺ CTLs at baseline did not differ across study arms (Fig. 3C). Similarly, the proportion of patients with EOC with higher-than-median intratumoral CD8⁺ CTLs (CD8^Hi) was similar across study arms (Supplementary Fig. S5C). Confirming prior findings by us and others (9, 20, 25), high intratumoral levels of CD8⁺ CTLs had a positive impact on PFS and OS in patients with EOC receiving SOC chemotherapy (PFS: P = 0.003; OS: P = 0.002; Fig. 3D; Supplementary Fig. S6A). Consistent with this notion, patients from arm C relapsing on SOC chemotherapy had reduced intratumoral levels of CD8⁺ T cells, DC-LAMP⁺ DCs, and CD20⁺ B cells as compared with nonrelapsing patients (Supplementary Fig. S6B). Conversely, the intratumoral abundance of CD8⁺ CTLs failed to affect PFS and OS in patients with EOC receiving DCVAC, irrespective of treatment schedule (Fig. 3D; Supplementary Fig. S6A). Importantly, patients with lower-than-median CD⁺ CTLs (CD8^Lo) EOCs receiving DCVAC had improved OS as compared with their CD8^Lo counterparts receiving SOC chemotherapy, irrespective of treatment schedule (Arm A: P = 0.005; Fig. 3E; Arm B: OS: P = 0.011; Fig. 3F). The CD8^Lo status was also associated with improved PFS, but only in patients receiving DCVAC after SOC chemotherapy (Arm B: PFS: P = 0.023; Fig. 3F). In line with these findings, patients from arms A and B relapsing upon DCVAC had increased intratumoral levels of CD8⁺ T cells, DC-LAMP⁺ DCs, and CD20⁺ B cells as compared with nonrelapsing patients (Supplementary Fig. S6C). We obtained similar results when arms A and B of the study were analyzed as a single group of patients (OS: P = 0.005; PFS: P = 0.020; Supplementary Fig. S6D). On the contrary, we were unable to observe a significant difference in PFS and OS between CD8^Hi patients treated with SOC chemotherapy versus DCVAC, irrespective of administration schedule (Fig. 3G and H).

Altogether, these data extend our previous findings on TMB status as they demonstrate that low intratumoral density of CD8⁺ CTLs is linked to superior disease outcome in patients with EOC treated with DCVAC.

To evaluate the development of antitumor immune responses, we compared the transcriptional profile of PBMCs from patients enrolled in SOV01 pre- and post-DCVAC administration. Specifically, we focused on 96 genes linked to various immune cell subsets and functions, including (but not limited to): T_H1 versus T_H2 polarization, B cells, T-cell activity and cytotoxicity, DCs, NK-cell activity, and immunosuppression (Supplementary Table S5). Samples for this study were available from 20, 21, and 17 patients from arms A, B, and C, respectively. In this context, we noted that the PBMCs of CD8^Lo patients expressed increased amounts of multiple genes involved in CD8⁺ CTL and/or NK-cell effector functions such as CD8A, GZMB, GNLY, and IL12A after DCVAC administration (Fig. 4A and B; Supplementary Fig. S7A). The same did not hold true for the PBMCs of patients receiving SOC chemotherapy (irrespective of CD8 status) and those of CD8^Hi patients receiving DCVAC, with the sole exception of CD8A (P = 0.047; Fig. 4A and B; Supplementary Fig. S7A). Moreover, the PBMCs collected from CD8^Lo patients post-DCVAC treatment exhibited increased levels of several genes associated with immune effector functions beyond CD8A, GZMB, GNLY, and IL12A (i.e., IFNG, KLRB1, KLRK1, and PRF1) as compared with their CD8^Lo counterparts receiving SOC chemotherapy (Fig. 4A and B; Supplementary Fig. S7A). However, none of these genes were significantly upregulated in the PBMCs isolated from CD8^Hi patients post-DCVAC treatment as compared with their CD8^Hi counterparts treated with SOC chemotherapy only (Fig. 4A and B; Supplementary Fig. S7A).

To confirm and extend these findings with another technological approach, we analyzed circulating biomarkers of immune responses by flow cytometry. Confirming our transcriptional findings, we detected a significant increase in CD56⁺ NK cells in the circulation of CD8^Lo patients upon DCVAC administration, which persisted until follow-up, irrespective of administration schedule (Fig. 4C). A similar increase could be documented for CD8⁺ T cells in patients from arm A, but not for total CD3⁺ T cells, CD4⁺ T cells, and B cells (Fig. 4C; Supplementary Fig. S7B).

Altogether these findings indicate that DCVAC elicits signs of improved effector functions in the peripheral blood of patients with EOC with scant tumor infiltration by CD8⁺ T cells, but mostly fails to alter the baseline status of these biomarkers in patients with immunologically “hot” EOC.

We next assessed the ability of DCVAC to induce antigen-specific CTL responses against HER2, MAGEA3, and MUC-1 in peripheral blood of 36 patients. We selected these three tumor-associated antigens (TAA) because they are highly expressed by the EOC cell lines (OV90 and SKOV3) used for loading autologous DCs during DCVAC manufacturing (ref. 26; Supplementary Fig. S8). To this aim, we evaluated the ability of circulating CTLs to secrete IFNγ (IFNG, best known as IFNγ) upon PBMC exposure to peptide mixture spanning multiple domains of HER2, MAGEA3, and MUC-1. Neither HER2-, MAGEA3-, nor MUC1-specific CTL responses could be detected in PBMCs from untreated healthy donors (not shown). Conversely, the PBMCs of EOC CD8^Lo (but not CD8^Hi) patients treated with DCVAC (irrespective of treatment schedule) contained an approximately two-fold percentage of CD8⁺ CTLs responding to HER2 or MAGEA3 (but not MUC1) as compared with baseline (Fig. 5A and B). As expected, SOC chemotherapy failed to increase the percentage of CTLs secreting IFNγ in response to stimulation with TAA-derived peptides, irrespective of intratumoral CD8⁺ CTL density. The frequency of CD45⁺CD3⁺CD8⁺ T cells responding to a pool of bacterial and viral peptides also did not change over the course of treatment (Supplementary Figures S9A and S9B).

Since TAA-specific CTL responses have been previously associated with improved disease outcome in other cohorts of cancer patients (27, 28), we tested the prognostic value of TAA-specific CTLs in 36 patients with EOC from the SOV01 study receiving DCVAC. We found that DCVAC-treated patients with higher-than-median circulating frequency of HER2- or MAGEA3-specific (but not MUC-1–specific) CD8⁺ CTLs had a trend toward improved PFS (which reached statistical significance in the case of MAGEA3) as compared with patients with lower-than-median circulating levels of TAA-specific CTLs (Fig. 5C).

While these data need to be confirmed in a larger cohort of DCVAC-treated patients, our findings indicate that DCVAC may boost clinically relevant TAA-specific T-cell responses mostly in patients with immunologically “cold” tumors, which is the patient subset that obtained most clinical benefits from DC-based vaccination in the SOV01 trial.

The composition of the immune infiltrate of human solid tumors, its localization, and functional orientation are major predictors of patient survival and response to immunotherapeutic interventions (4, 16, 24, 29). As compared with other solid malignancies, EOC has a reduced TMB (8, 23), generally correlating with limited tumor infiltration by CD8⁺ CTLs (30) and hence resistance to CTL-reactivating maneuvers, such as ICIs (25, 31). Thus, strategies to induce anticancer immune responses in patients with EOC as well as biomarkers that improve decision making with respect to (immuno)therapeutic approach in women with EOC are highly awaited. In this context, DC-based vaccines including DCVAC (15, 17) stand out as promising tools to reconfigure the immunologic microenvironment of EOC in support of clinical benefit and/or improved sensitivity to ICIs (14). In line with this notion, we have recently reported the results of a randomized, phase II clinical trial (SOV01, NCT02107937) demonstrating that DCVAC is well tolerated and significantly extends PFS over SOC chemotherapy in patients with EOC, but only when administered after (rather than in parallel to) SOC chemotherapy (17).

Here, we demonstrate that, contrarily to patients with EOC receiving SOC chemotherapy (25), women with EOC obtain clinical benefits from DCVAC that manifest with circulating biomarkers of ongoing anticancer immunity when their malignancies have a low TMB and are scarcely infiltrated by CD8⁺ CTLs, irrespective of treatment schedule. Several scenarios can be invoked to explain these apparently counterintuitive findings. For instance, the tumor microenvironment (TME) of patients with TMB^Hi EOC and robust tumor infiltration by CD8⁺ CTLs may exhibit the activation of compensatory immunosuppressive mechanisms that keep effector immune functions at bay, not only preventing potential DCVAC-elicited CTLs from mediating anticancer effects, but also limiting the ability of DCVAC to elicit TAA-specific CTLs (25, 32). This has previously been reported in patients with various solid carcinomas, including ovarian carcinoma, in which the intratumoral density of CD8⁺ CTLs correlates with the abundance of immunosuppressive cell populations such as CD4⁺CD25⁺FOXP3⁺ regulatory T (Treg) cells (33–35). Conversely, EOCs with low TMB and scant infiltration by CD8⁺ CTLs may be relatively naïve with respect to immune infiltration at large, and hence allow (at least initially or to some degree) for the effector functions of DCVAC-driven CTLs. In further support of this interpretation, the transcriptional profile of EOC samples from the TCGA database that contain higher-than-median levels of CD8 is enriched for a variety of transcripts expressed by immunosuppressive cells, such as V-set immunoregulatory receptor (C10orf54; best known as VSIR), CD68, CD274 (best known as PDL1), cytotoxic T-lymphocyte associated protein 4 (CTLA4), ectonucleoside triphosphate diphosphohydrolase 1 (ENTPD1; best known as CD39), forkhead box P3 (FOXP3), hepatitis A virus cellular receptor 2 (HAVCR2; best known as TIM3), indoleamine 2, 3-dioxygenase 1 (IDO1), IL10, lymphocyte activating 3 (LAG3), 5´-nucleotidase ecto (NT5E; best known as CD73), programmed cell death 1 (PDCD1; best known as PD-1), TGFB1, T-cell immunoreceptor with Ig and ITIM domains (TIGIT), and triggering receptor expressed on myeloid cells 2 (TREM2; Supplementary Fig. S10). Moreover, both a high TMB and robust infiltration by CD8⁺ CTLs are expected to correlate with increased intratumoral heterogeneity (ITH), which confers the entire malignant lesion as an ecosystem with increased probability to escape TAA-specific CTLs by virtue of the CTL-dependent selection of TAA-negative cancer cells (36). On the contrary, a limited TMB is generally associated with reduced ITH and hence (at least theoretically) poor adaptability in response to TAA-targeting anticancer immunity. Finally, the cellular biology of TMB^Hi EOCs may differ considerably from that of their TMB^Lo counterparts so to render them resistant to DCVAC-driven CTLs, for instance as a function of altered proliferation (37, 38). All these possibilities remain to be experimentally verified in the context of EOC.

Our study has various limitations. First, it is an explorative retrospective study on a relatively low number of patients with no preplanned statistical analysis, which altogether limits statistical power. Second, the collection of posttreatment tumor biopsies was not included in the clinical trial design, which prevents us from investigating alterations in the EOC microenvironment elicited by DCVAC. Third, blood samples collected in the course of treatment were more frequently available for patients experiencing limited side effects and overall performing well (who accepted to remain in the study until completion), as compared with patients suffering from treatment-related toxicity or progressing rapidly.

Irrespective of these and other caveats, our findings bring up two important considerations. First, they suggest that baseline TMB and T-cell infiltration may perhaps be further evaluated as a tool to guide therapeutic decisions in patients with EOC. Specifically, they suggest that while patients with TMB^Hi EOC and robust tumor infiltration by CD8⁺ CTLs may benefit from SOC chemotherapy alone (and perhaps even ICIs, based on sporadic literature reports; refs. 39, 40), women with TMB^Lo lesions and scant tumor infiltration by CD8⁺ CTLs may instead benefit from strategies that jumpstart anticancer immunity, such as DCVAC or radiotherapy (41). In this context, diagnostic test quantifying tumor infiltration by CD8⁺ CTLs and approved for clinical use in other oncological indications, such as the Immunoscore, could be considered as a tool to guide patient selection and inclusion in future clinical trials (29). Second, our data lend additional support to the key importance of administration for the development of combinatorial anticancer regimens (42, 43). In particular, they add to a growing preclinical literature indicating that measures that initiate anticancer immunity such as DCVAC or radiotherapy should be administered prior to (and not after or concomitant to) ICIs or other therapeutic agents that engage the effector phase of the immune response (e.g., CDK4/6 inhibitors; refs. 44, 45). From this perspective, it is tempting to speculate that DCVAC followed by SOC chemotherapy and/or ICI-based immunotherapy may mediate even more pronounced anticancer effects than the DCVAC-based regimens tested in SOV01, based on the rationale that (i) DCVAC would initially promote at least some degree of tumor reactivity by CD8⁺ CTLs (as documented in this study by an increase in circulating TAA-specific CD8⁺ CTLs), (ii) SOC chemotherapy is more active in patients with “hot” EOCs (as demonstrated in this study and various others; refs. 46, 47), and (iii) that patients with non-EOC tumors responding to ICIs generally exhibit elevated tumor infiltration by T cells at baseline (48–50). Additional clinical trials are needed to address these possibilities as well as the potential value of TMB and CD8⁺ CTL infiltration at baseline as biomarkers to guide decision making in the clinical management of EOC.

J. Fucikova reports personal fees from Sotio Biotech during the conduct of the study; in addition, J. Fucikova has patents for US 10,955, 751 and US 10,918, 704 issued, and a patent for US 2021/0122548 pending. M. Hensler reports patents for US 10,955, 751 and US 10,918, 704 issued, and a patent for US 2021/0122548 pending. L. Kasikova reports personal fees from Sotio Biotech during the conduct of the study; in addition, L. Kasikova has patents for US 10,955, 751 and US 10,918, 704 issued, and a patent for US 2021/0122548 pending. T. Lanickova reports personal fees from Sotio Biotech during the conduct of the study. J. Rakova reports personal fees from Sotio Biotech during the conduct of the study. M. Hraska reports personal fees from Sotio Biotech during the conduct of the study. K. Sochorova reports personal fees from Sotio Biotech and SCTbio, formerly named Sotio Biotech, during the conduct of the study. J. Laco reports grants from Czech government during the conduct of the study, as well as personal fees from AstraZeneca outside the submitted work. A. Ryska reports grants from Czech government during the conduct of the study. A. Ryska also reports personal fees from Amgen, Eli Lilly and Company, Gilead, and Merck Serono; grants, personal fees, and nonfinancial support from MSD; and personal fees and nonfinancial support from Roche and BMS outside the submitted work. A. Coosemans reports personal fees from Sotio Biotech during the conduct of the study, as well as other support from Oncoinvent and Novocure outside the submitted work. I. Vergote reports personal fees from Agenus, Akesobio China, Bristol Myers Squibb, Eisai, F. Hoffmann-La Roche Ltd, Jazzpharma, Karyopharm, Mersana, Novartis, Seagen, Sotio Biotech, and Zentalis. I. Vergote also reports grants and other support from Amgen (Europe) GmbH, Roche, and Genmab; personal fees and other support from AstraZeneca, Deciphera Pharmaceuticals, GSK, Immunogen Inc., MSD, and Novocure; other support from Clovis Oncology Inc., Carrick Therapeutics, Elevar Therapeutics, Millennium Pharmaceuticals, Octimet Oncology NV, and Verastem Oncology; and grants, personal fees, and other support from Oncoinvent AS outside the submitted work. J. Bartunkova reports patents for US 10,955, 751 and US 10,918, 704 issued, and a patent for US 2021/0122548, pending; in addition, J. Bartunkova is a minority shareholder of the company Sotio Biotech. J. Galon reports grants and personal fees from Veracyte during the conduct of the study. J. Galon also reports personal fees from Lunaphore, Catalym, and Northwest Biotherapeutics, as well as grants from Imcheck Biotherapeutics outside the submitted work; in addition, J. Galon has a patent for Immune biomarkers to predict prognosis of cancer licensed and with royalties paid from Veracyte. L. Galluzzi reports personal fees and other support from Sotio Biotech during the conduct of the study. L. Galluzzi also reports other support from Lytix; personal fees and other support from Phosplatin, Onxeo, and Noxopharm; and personal fees from Boehringer Ingelheim, AstraZeneca, OmniSEQ, The Longevity Labs, Inzen, and Luke Heller TECPR2 Foundation outside the submitted work. R. Spisek reports other support from Sotio Biotech outside the submitted work; in addition, R. Spisek has patents for US 10,955, 751 and US 10,918, 704 issued, and a patent for US 2021/0122548 pending. No disclosures were reported by the other authors.

J. Fucikova: Conceptualization, resources, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing–review and editing. M. Hensler: Conceptualization, data curation, software, formal analysis, supervision, writing–review and editing. L. Kasikova: Data curation, software, methodology. T. Lanickova: Data curation, methodology. J. Pasulka: Software. J. Rakova: Data curation. J. Drozenova: Resources. T. Fredriksen: Data curation, software, formal analysis. M. Hraska: Data curation, software, formal analysis. T. Hrnciarova: Data curation, software, formal analysis. K. Sochorova: Data curation, software, formal analysis. D. Rozkova: Data curation, software, formal analysis. L. Sojka: Resources. P. Dundr: Resources. J. Laco: Resources, data curation, software, formal analysis. T. Brtnicky: Data curation, software, formal analysis. I. Praznovec: Resources. M.J. Halaska: Resources. L. Rob: Resources. A. Ryska: Resources. A. Coosemans: Resources, writing–review and editing. I. Vergote: Writing–review and editing. D. Cibula: Writing–review and editing. J. Bartunkova: Conceptualization, resources, data curation, writing–review and editing. J. Galon: Writing–review and editing. L. Galluzzi: Conceptualization, writing–review and editing. R. Spisek: Conceptualization, resources, project administration, writing–review and editing.

This study was sponsored by Sotio Biotech, Prague. J. Laco and A. Ryska were supported by the project BBMRI-CZ LM2018125, by the program PROGRES Q40/11, by the Cooperatio Program, research area DIAG, and by European Regional Development Fund-Project BBMRI-CZ: EF16_013/0001674. L. Galluzzi is supported by a Breakthrough Level 2 grant from the US Department of Defense (DoD), Breast Cancer Research Program (BRCP; grant no. BC180476P1), by the 2019 Laura Ziskin Prize in Translational Research [#ZP-6177, principal investigator (PI): Silva C. Formenti] from Stand Up to Cancer (SU2C), by a Mantle Cell Lymphoma Research Initiative (MCL-RI, PI: Selina Chen-Kiang) grant from the Leukemia and Lymphoma Society (LLS), by a startup grant from the Department of Radiation Oncology at Weill Cornell Medicine (New York, NY), by a Rapid Response Grant from the Functional Genomics Initiative, by industrial collaborations with Lytix and Phosplatin, and by donations from Phosplatin, the Luke Heller TECPR2 Foundation, Onxeo, and Sotio Biotech. L. Rob and M.J. Halaska were supported by Cooperatio program, Maternal and Childhood Care (207035), Third Faculty of Medicine, Charles University (Prague, Czech Republic). P. Dundr was supported by Ministry of Health, Czech Republic, research project RVO 64165.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.