Targeting SNCA in the treatment of malignant ascites in gastrointestinal cancer

Highlights

• •SNCA⁺ cells systemically increase in gastric cancer patients with malignant ascites.
• •SNCA⁺ cells also systemically increase in mouse tumor ascites models.
• •Blocking SNCA induces potent anti-tumor immunity in in vitro and in vivo settings.
• •Anti-SNCA therapy provides a significantly better prognosis in mouse ascites models.
• •Anti-SNCA therapy synergistically improves anti-PD1 efficacy in mouse ascites models.

Abstract

Peritoneal tumor dissemination and subsequent malignant tumor ascites (MTA) occur unexpectedly and repeatedly in patients with gastrointestinal (GI) cancers, and worsen quality of life and prognosis of the patients. Various treatments have been clinically developed for these patients, while most of the MTA cases are refractory to the treatments. Thus, effective treatments are urgently needed to improve the clinical outcomes. In this study, we identified α-synuclein (SNCA) as an immunological determinant of MTA progression in GI cancer through translational research using mouse tumor models and clinical specimens collected from gastric cancer patients. We found that the SNCA⁺ subsets were significantly increased in CD3⁺ T cells, CD56⁺ NK cells, and CD11b⁺ myeloid cells within MTA and peripheral blood cells (PBCs) of MTA cases, albeit almost absent in PBCs of healthy donors, and spleen of naive mice. Of note, the SNCA⁺ T-cell subset was rarely seen in patients that intraperitoneal lavage fluid without tumor cells was collected before surgery as a tumor-free control, suggesting a possible cancer-induced product, especially within the peritoneal cavity. In vivo treatment with anti-SNCA blocking mAb significantly induced anti-tumor effects in mouse MTA models, and synergistically improved anti-PD1 therapeutic efficacy, providing a significantly better prognosis. These suggest that SNCA is involved in severe immunosuppression in the MTA cases, and that blocking SNCA is effective in dramatically improving the immune status in the hosts. Targeting SNCA will be a promising strategy to improve clinical outcomes in the treatment of GI cancer patients, especially with MTA.

Introduction

Peritoneal tumor dissemination and subsequent malignant tumor ascites (MTA) are the leading causes of death in gastrointestinal (GI) cancers, since MTA occurs unexpectedly and repeatedly in patients, particularly in advanced stages, leading to poor quality of life and poor prognosis [1]. Various treatments, such as hyperthermic intraperitoneal (i.p.) chemotherapy [2], neoadjuvant systemic and hyperthermic i.p. chemotherapy combined with cytoreductive surgery [3], and pressurized i.p. aerosol chemotherapy [4], have been clinically developed for these patients. However, most of the MTA cases are refractory to these treatments [5,6]. Thus, development of effective treatments is urgently needed to improve the clinical outcomes in GI cancers.

Epithelial-to-mesenchymal transition (EMT) in tumor cells has been considered as a pivotal biological step for peritoneal metastasis [7]. EMT-associated key molecules, such as LIMK1 [8], LGALS4/GAL4 [9], and GABRD [10], have been recently identified in gastric cancer (GC), and blocking these molecules with the specific inhibitors has been shown to suppress peritoneal metastasis in mouse tumor models. However, EMT is only a part of the cancer metastasis process [7], and tumor cells evolve not only intrinsically but also extrinsically through interplay with numerous environmental components in the host [11]. The most influential host factor is the immune system. In the i.p. immunity, macrophages (Møs) [12] and fibroblasts [13] have received great attention, and many mechanisms underlying the i.p. tumor dissemination and progression have been demonstrated so far. However, most efforts have not yet led to the clinical practice. On the other hand, blocking immune inhibitory checkpoint (IC) molecules, including CTLA4, PD1, LAG3, and TIGIT, with the specific mAbs has been successful in clinical cancer therapy as a representative immunomodulatory method [14]. However, only a limited number of patients receive clinical benefit, and the therapeutic efficacy in MTA cases remains unclear because such patients are often excluded before the treatment. Thus, development of alternative drugs is needed to better control the immune system in clinical settings.

We have been investigating both cancer cells and host immunity using clinical specimens obtained from patients with various cancers, and have elucidated several molecular mechanisms underlying treatment resistance through translational research using various mouse tumor models [[15], [16], [17]]. In this study, we sought to identify an immunological determinant of the MTA progression using MTA and i.p. lavage fluid (ILF) as a tumor-free control obtained from GC patients, and mouse MTA models that we previously established [17].

Materials and methods

Cell lines

Murine colorectal cancer Colon26 cells were purchased from Cell Resource Center for Biomedical Research at Tohoku University in Japan, and MC38 cells were kindly provided by the NCI at the NIH in USA. These cells were tested for Mycoplasma negativity using a Hoechst-staining detection kit (MP Biomedicals), and were expanded and frozen in liquid nitrogen to avoid changes due to long-term culture. The cells were cultured in 10 % FBS/DMEM medium (GIBCO).

Mice

Five-week-old female BALB/c and C57BL/6 mice were purchased from Charles River Laboratories in Japan, and were maintained under pathogen-free conditions. The mice were used according to the protocol (No. T17-055) approved by the Animal Care and Use Committee at the National Cancer Center Research Institute.

Preparation of anti-SNCA mAb

We used clone 9E4, which recognizes amino acids 118–126 of SNCA and inhibits SNCA activity, since this is a murine version of Prasinezumab/PRX002 that has been evaluated in clinical trials for Parkinson's disease [18,19]. The mAb was kindly provided by Chiome Bioscience within a collaborative research project. They synthesized using vectors expressing the DNA fragment of 9E4 VH and VL based on publicly available amino acid sequence information (#JP, 2018-046836, A), and purified using a Hitrap protein G HP column (Cytiva) followed by gel filtration on HiLoad 26/600 Superdex 200 pg (Cytiva). The purity was confirmed by SDS-PAGE, and the binding activity was confirmed by ELISA using human and mouse SNCA (Abcam).

In vivo therapy

We used mouse MTA models that we previously established [17]: Colon26 or MC38 cells were both subcutaneously (s.c., 5 × 10⁵) and i.p. (2–5 × 10⁵) implanted into syngeneic mice to observe therapeutic effects on both s.c. and i.p. tumors simultaneously. The mAbs (200 µg/mouse) were intravenously (i.v.) injected into the mice, since we previously demonstrated that i.v. administration was better than i.p. administration (that is usually performed in mouse therapeutic experiments) in the anti-PD1 therapy using the MTA model [17], albeit the reason remains unclear. Also, considering the clinical practice (biweekly treatment with mAbs), mAbs were administered at one-week intervals in the case of multiple doses in contrast to the general intervals of several days in mouse therapeutic experiments. The following antibodies were used: Anti-PD1 mAb (clone 29F.1A12; BioLegend), anti-LAG3 mAb (clone C9B7W; BioXCell), anti-TIGIT mAb (clone 1G9; BioXCell), anti-CTLA4 mAb (clone 9H10; BioLegend), anti-IL1B mAb (clone B122; BioXCell), anti-SNCA mAb, and mouse IgG (mIgG, Clone MOPC-21; BioXCell) as a control. To deplete CD8⁺ T cells and NK cells, mice were i.p. injected with anti-CD8 mAb (Clone 2.43, 200 µg; BioXCell) or anti-asialo GM1 polyclonal Ab (20 µL; BioLegend) before and after the treatment. The depletion efficacy (>80 %) was validated 1–2 days after injection. Therapeutic efficacy was evaluated by tumor volume (0.5 × Length × Width², mm³) and/or mouse survival time. Two weeks after tumor implantation, MTA was collected by washing peritoneal cavity of the mice with 5 ml of 4 % citrated PBS, and spleen was dispersed into single cell suspensions followed by treatment with ACK Lysing Buffer (ThermoFisher). To assess tumor-killing activity of CD8⁺ T cells, bulk spleen cells (SPCs) were pre-stimulated with the H-2L(d)-restricted tumor antigen AH1 peptide (1 µg/mL; MBL) or the H-2K(b)-restricted tumor antigen gp70 peptide (1 µg/mL; MBL) for 6 days, and the recovered CD8⁺ T cells were tested for cytotoxic activity (target = Colon26 or MC38, 4 h) as described before {Kudo-Saito, 2009 #52}. Cytotoxic activity of NK cells was similarly assessed using Yac-1 cells as a target.

Preparation of clinical samples

We collected MTA of gastric cancer (GC) patients (n = 16, median age 66, male x 8, female x 8), ILF without tumor cells by cytological diagnosis (n = 15, median age 69, male x 10, female x 5) as a tumor-free control, and peripheral blood of GC patients (MTA cases x 12, ILF cases x 3) and healthy donors (n = 3) as a normal control at the National Cancer Center Hospital according to the protocol (No. 2016-067 and No. 2017-046) approved by the IRB, and the Keio University Hospital according to the protocol (No. 20180064) approved by the IRB. ILF was harvested by washing peritoneal cavity with 300 ml saline at initiation of surgical tumor resection. Peripheral blood cells (PBCs) were prepared by centrifugation followed by ACK treatment. Informed consent was obtained from all subjects. All activities were conducted in accordance with the ethical principles of the Declaration of Helsinki.

Flow cytometric analysis

After Fc blocking, cells were stained with the following immunofluorescence-conjugated antibodies: In clinical study, anti-CD45-APC-Cy7 (BioLegend), anti-CD3-BUV496 (BD), anti-CD4-PerCP-Cy5.5 (BioLegend), anti-CD8-BUV395 (BD), anti-CD56- BUV650 (BioLegend), anti-CD11b-BV510 (BioLegend), anti-GZMB-PE-Cy7 (BioLegend)anti-Ki67-FITC (BioLegend), anti-LAG3-BV650 (BioLegend), anti-TIGIT-PE-Cy7 (BioLegend), anti-CTLA4-BV785 (BD), anti-FOXP3-BV421 (BioLegend), anti-PDL1-FITC (BD), anti-HLA-DR-APC-Cy7 (BioLegend), anti-PD1-BV605 (BioLegend), anti-SNCA-PE (Abcam), and the appropriate isotype control. In mouse study, anti-CD45-PE-Cy5 (BioLegend), anti-CD3e-BUV496 (BD), anti-CD4-BV785 (BioLegend), anti-CD8a-BUV395 (BD), anti-DX5-APC-Cy7 (BioLegend), anti-CD11b-BV711 (BioLegend), anti-CTLA4-PerCP-Cy5.5 (BioLegend), anti-FOXP3-BV421 (BioLegend), anti-Gr1-BV605 (BioLegend), anti-I-A(d)-BUV496 (BD), anti-PDL1-BV785 (BioLegend), anti-SNCA-PE (LSBio), and the appropriate isotype control. For intracellular staining, cells were treated with Cytofix/Cytoperm solution (BD) before the staining. Data were acquired using a BD LSR Fortessa X-20 cytometer (BD), and were analyzed by FlowJo software (BD). Before defining the specific molecular expressions, debris was firstly excluded by FSC/SSC followed by gating CD45⁺ leukocytes, and immunofluorescence intensity was compared to isotype controls to gate cell fractions positive and negative for specific molecular expression (Supplementary Figs. S1, S2). CD3⁺ cells were defined as T cells. CD56⁺ cells (DX5⁺ cells in mice) were defined as NK cells. CD11b⁺ cells within the large fraction of CD45⁺ leukocytes were defined as myeloid cells. The GZMB⁺Ki67⁺ subset of T/NK cells was defined as anti-tumor effector cells. The LAG3⁺TIGIT⁺ subset of T cells was defined as exhausted T cells (Texs). The CTLA4⁺FOXP3⁺ subset of T cells was defined as regulatory T cells (Tregs). The HLA-DR⁻PDL1⁺ subset of myeloid cells (Gr1⁺I-A⁻PDL1⁺ subset in mice) was defined as myeloid-derived suppressor cells (MDSCs). Data were presented as percentages of specific cells in clinical studies, and as cell numbers in the ILF (×10⁴/ml) and spleen (×10⁶/spleen) by multiplying total leukocyte numbers by the percentages in mouse studies.

Statistical analysis

Data are shown as means ± SDs unless otherwise specified. Experiments were repeated at least three times to confirm the reproducibility. Significant differences (P value <0.05) were evaluated using GraphPad Prism 7 software (MDF). To compare between two groups, the data were analyzed by the unpaired two-tailed Student's t test. Non-parametric groups were analyzed by the Mann–Whitney test. To compare multiple groups, the data were analyzed by one-way ANOVA, followed by the Bonferroni post-hoc test for pairwise comparison of groups on the basis of the normal distributions. In the combination therapy, significance to the single treatment was evaluated using a two-way ANOVA with Bonferroni post-hoc test. Survival was analyzed by Kaplan–Meier method and the Mantel–Cox Log-Rank test.

Results

SNCA⁺ cells increase in peritoneal cavity of gastric cancer patients

We compared immune cell populations between MTA and tumor-free ILF obtained from GC patients by flow cytometry. Although the percentage of CD45⁺ cells was significantly lower in MTA than those in ILF (probably due to the presence of tumor cells), the number of CD45⁺ leukocytes per 1 ml were almost similar in both, and the majority of CD45⁺ cells were CD3⁺ T cells, with a predominance of the CD4⁺ subset in both specimens (Fig. 1A). The CD4⁺ T cells similarly contained approximately 30 % of the potentially immunoregulatory CTLA4⁺FOXP3⁺ subset (Treg) in both specimens, but more abundantly contained the potentially exhausted LAG3⁺TIGIT⁺ subset (Tex) in MTA than that in ILF (P = 0.0149; Fig. 1B). CD8⁺ T cells were significantly less in MTA than those in ILF (P = 0.0077), and contained more Treg (P = 0.0591) and Tex (P = 0.0033), although also containing significantly more the potentially cytotoxic GZMB⁺Ki67⁺ subset (P = 0.0298; Fig. 1B). These suggest that T-cell immunity in peritoneal cavity of the MTA cases is suppressive and exhausted. No significant differences in large CD11b⁺ myeloid cells (possibly including monocytes, macrophages and granulocytes; Fig. 1A) and CD11b⁺PDL1⁺HLA-DR⁻ myeloid-derived suppressor cell-like subset (MDSC; Fig. 1C) were seen between MTA and ILF. However, the CD11b⁺CTLA4⁺ myeloid subset was significantly more contained in MTA than that in ILF (P = 0.0246; Fig. 1C). We previously identified the CD11b⁺CTLA4⁺ myeloid subset that can promote tumor progression and metastasis directly and indirectly via inducing immune dysfunction leading to anti-PD1 resistance {Imazeki, 2021 #35}. This suggests that the CD11b⁺CTLA4⁺ myeloid subset is also involved in MTA progression of GC patients.

Fig. 1.

Profiling of immune cells in peritoneal cavity of gastric cancer patients by flow cytometric analysis. We analyzed cells isolated from malignant tumor ascites (MTA; n = 16) and intraperitoneal (i.p.) lavage fluids (ILF; n = 15) as a control with no tumor cells in the peritoneal cavity obtained from gastric cancer (GC) patients by flow cytometry. Data are depicted as percentages of a specific cell subset in a specific cell population, except for the number of CD45⁺ cells shown in panel A. (A) General cell populations. (B) Anti-tumor effector cells, exhausted T cells, and regulatory T cells. (C) CD11b⁺ myeloid subsets expressing immune inhibitory checkpoint (IC) molecules. (D) SNCA⁺ subsets in CD11b⁺ myeloid cells, CD56⁺ NK cells and CD3⁺ T cells. P values were analyzed by the Mann–Whitney test.

We previously demonstrated that blocking FSTL1, which can enhance cancer refractoriness directly by inducing EMT and indirectly via inducing apoptosis in CTLs, is effective in various anti-PD1 resistant mouse tumor models, including the mouse lung cancer 3LL model [[20], [21], [22], [23]]. When we explored new factors by comprehensive analysis of gene expressions in splenic T cells obtained from the tumor models, SNCA was found to be remarkably increased by tumor implantation but dramatically decreased by anti-FSTL1 therapy, albeit only slight decrease by anti-PD1 therapy (Fig. S3A). SNCA is a neuropathological molecule involved in various neurodegenerative disorders, such as Parkinson's disease and Alzheimer's disease [24], and has been reported to bind to LAG3 [25]. We verified the existence of the SNCA⁺ subset in various cell populations in SPCs of tumor-bearing mice, albeit almost none in naive mice (Fig. S3B), and confirmed that treatment with anti-SNCA blocking mAb can induce potent anti-tumor immunity in the in vitro CTL induction system (Fig. S3C) and the in vivo therapeutic experiments (Fig. S3D, E). In clinical settings, the SNCA⁺ myeloid subset was also abundantly contained in CD11b⁺ myeloid cells, CD56⁺ NK cells, and CD3⁺ T cells contained in MTA (Fig. 1D). Of note, the SNCA⁺ T-cell subset was significantly increased only in MTA, but not in LIF (P = 0.0037). This suggests that the SNCA⁺ cells, particularly the T-cell subset, are involved in MTA progression of GC patients.

SNCA⁺ cells increase in peripheral blood of GC patients

We also analyzed PBCs, and found that CD4⁺ T cells were significantly more, but CD8⁺ T cells were significantly less in patients than those in healthy donors (P < 0.05; Fig. 2A). The Tex and the SNCA⁺ subset were significantly increased in CD3⁺ T cells of patients than those in healthy donors, but were more abundantly seen in patients with MTA than those in ILF-provided patients (P < 0.05; Fig. 2B, C). This is consistent with the i.p. data. The majority of CD45⁺ cells were CD11b⁺ myeloid cells in any cases including healthy donors (Fig. 2A), but the SNCA⁺ subset was more abundantly seen in patients than those of healthy donors (P < 0.05; Fig. 2C). However, no difference in the SNCA⁺ myeloid subset was seen between two groups. These suggest that the SNCA⁺ cells, particularly the T-cell subset, are systemically expanded in GC patients with MTA, and are involved in MTA progression.

Fig. 2.

Profiling of immune cells in peripheral blood of GC patients by flow cytometric analysis. We analyzed peripheral blood cells (PBCs) obtained from GC patients with MTA (n = 12), ILF-provided GC patients (n = 3), and healthy donors (designated H; n = 3) by flow cytometry. Data are depicted as percentages of a specific cell subset in a specific cell population. (A) General cell populations. (B) T cells expressing IC molecules. (C) SNCA⁺ subsets in CD11b⁺ myeloid cells, CD56⁺ NK cells and CD3⁺ T cells. P values were analyzed by the Mann–Whitney test. Representative data from the dot plot panels of PBCs are shown below.

Blocking CTLA4 is effective in the mouse Colon26-MTA model

Tex and Treg were significantly increased in MTA of GC patients. We next evaluated anti-tumor efficacy induced by IC inhibitory mAbs targeting CTLA4, LAG3, and TIGIT using a mouse Colon26-implanted MTA model that we previously established [17], because mouse GC models that spontaneously produce MTA were not available to us. Anti-CTLA4 and anti-LAG3 treatments significantly inhibited s.c. tumor growth as compared to the control treated with mouse IgG (mIgG; P < 0.001), whereas the impact on mouse survival was very small despite statistical significance (P < 0.05; Fig. 3A). We combined anti-IL1B mAb with anti-CTLA4 mAb, since we previously identified IL1B as a key effector molecule produced from the CTLA4⁺ myeloid subset that is expanded under cancer metastasis [16]. Indeed, anti-IL1B therapy synergistically enhanced anti-CTLA4 therapeutic efficacy, and significantly prolonged mouse survival (P = 0.013 versus anti-CTLA4 monotherapy; Fig. 3B). These suggest that the anti-CTLA4/IL1B combination regimen is effective in hosts with MTA.

Fig. 3.

Blocking CTLA4 is effective in the mouse Colon26-MTA models. BALB/c mice were both subcutaneously (s.c., 5 × 10⁵) and i.p. (2 × 10⁵) implanted with Colon26 cells, and began receiving treatments on day 3–4 after tumor implantation (n = 10). (A) Blocking IC molecules has little impact on mouse survival. Mice were intravenously (i.v.) injected with mAb specific for PD1, TIGIT, FSTL1, LAG3, or CTLA4, or mouse IgG (mIgG) as a control (200 µg/mouse) on days 3 and 10. (B) Combination with anti-IL1B therapy increases the anti-CTLA4 therapeutic efficacy on mouse survival. Mice were i.v. injected with anti-CTLA4 mAb, anti-IL1B mAb, and/or mouse IgG (mIgG) as a control (200 µg/mouse) on days 4, 11 and 18. Tumor volume data show means ± SDs. P values were analyzed by the Mann–Whitney test. Mouse survival was analyzed by the Kaplan–Meier method and the Mantel–Cox Log-Rank test. Representative data of three independent experiments.

Anti-tumor efficacy induced by anti-SNCA therapy in the mouse Colon26-MTA model

We next evaluated anti-tumor efficacy induced by anti-SNCA blocking mAb using the same tumor model, since SNCA⁺ cells were systemically increased in GC patients, particularly with MTA. Anti-SNCA treatment significantly suppressed s.c. tumor growth (Fig. 4A), and this therapeutic efficacy was abrogated by depletion of CD8⁺ T cells or NK cells, suggesting the requirement of both cells for the therapeutic mechanisms (Fig. 4B). The SNCA therapy synergistically enhanced anti-PD1 therapeutic efficacy (P = 0.006 versus anti-SNCA monotherapy), resulting in tumor disappearance in 60 % of the treated mice, and mouse survival was greatly prolonged (P = 0.029 versus anti-SNCA monotherapy; Fig. 4C). In the anti-SNCA/PD1-treated mice, SNCA⁺ cells were significantly reduced in MTA and spleen as compared to the control mice (P < 0.05; Fig. 4D). Anti-tumor effector cells, including CD3⁺CD8⁺ T cells and DX5⁺CD3⁻ NK cells, were significantly increased, but immunosuppressive CD3⁺CD4⁺CTLA4⁺FOXP3⁺ Tregs and CD11b⁺Gr1⁺I-A(d)-PDL1⁺ MDSCs were significantly reduced in MTA (P < 0.05; Fig. 4E) and spleen (P < 0.05; Fig. 4F), as compared to those of the control mice. Of note, NK cells were dramatically increased in spleen of the anti-SNCA/PD1-treated mice, although no difference was observed in its cytotoxic activity (Fig. 4G). Anti-SNCA monotherapy had a significant impact on the numbers and cytotoxic activity of CD8⁺ T cells, albeit no synergy by combining anti-PD1 therapy. These suggest that blocking SNCA is effective in inducing potent anti-tumor immunity through reducing immunosuppressive cells in hosts with MTA.

Fig. 4.

Anti-tumor efficacy induced by anti-SNCA therapy in the mouse Colon26-MTA models. The Colon26 tumor models began receiving treatments (200 µg/mouse) on day 3–4 after tumor implantation (n = 10). Spleen and MTA were harvested for assays on day 15, and samples from 3 out of 10 mice were pooled and assayed as n = 3. (A) Comparison of anti-tumor efficacy between mAb administration routes. Mice were i.p. or i.v. injected with anti-SNCA mAb or mIgG as a control on day 3. (B) Both CD8⁺ T cells and NK cells are required for the anti-SNCA therapeutic efficacy. Mice were i.p. injected with anti-CD8 mAb to deplete CD8⁺ T cells or anti-asialo GM1 Ab to deplete NK cells before and after anti-SNCA therapy. (C) Anti-SNCA therapy synergizes with anti-PD1 therapy in inducing anti-tumor effect. Mice were i.v. injected with anti-SNCA mAb, anti-PD1 mAb, and/or mIgG as a control 2 or 3 times (days 4, 11 and 18). (D) SNCA⁺ subsets in CD11b⁺ myeloid cells and CD3⁺ T cells in MTA and spleen (n = 3). (E) Immune cells in MTA (n = 3). (F) Immune cells in spleen (n = 3). (G) Potent CTLs are generated in the anti-SNCA-treated mice. Splenic CD8⁺ T cells pre-treated with tumor antigen AH1 peptide were cocultured with Colon26 cells at 40:1. NK cells were cocultured with Yac-1 cells at 20:1. *P < 0.01, **P < 0.05 versus control by the Mann–Whitney test. Mouse survival was analyzed by Kaplan–Meier method and the Mantel–Cox Log-Rank test. Graphs show means ± SDs. Representative data of three independent experiments.

SNCA blockade therapy synergizes with anti-PD1 therapy in inducing anti-tumor immunity in the mouse MC38-MTA model

We then validated the anti-SNCA therapeutic efficacy using another mouse MTA model implanted with MC38 cells. Anti-SNCA therapy significantly prolonged mouse survival time (P < 0.001 versus control), and synergistically enhanced anti-PD1 therapeutic efficacy (P < 0.001 versus anti-SNCA monotherapy; Fig. 5A). One difference in this model from the Colon26 model was that SNCA⁺ cells in MTA and spleen were not reduced by anti-SNCA therapy with/without anti-PD1 combination, although SNCA⁺ cell increase caused by anti-PD1 therapy was significantly suppressed by anti-SNCA combination (P < 0.05; Fig. 5B). The pattern of increase (CD8⁺ T cells and NK cells) and decrease (Tregs and MDSCs) in other cell populations in MTA (Fig. 5C) and spleen (Fig. 5D) upon the treatments was almost similar to the C26 model. Another difference was that NK cytotoxic activity was significantly enhanced by anti-SNCA therapy with/without anti-PD1 combination (P < 0.001 versus control; Fig. 5E).

Fig. 5.

SNCA blockade therapy synergizes with anti-PD1 therapy in inducing anti-tumor immunity in the mouse MC38-MTA models. C57BL/6 mice were i.p. implanted with MC38 cells (5 × 10⁵), and were i.v. injected with anti-SNCA mAb, anti-PD1 mAb, and/or mIgG as a control (200 µg/mouse) on days 4, 11 and 18 after tumor implantation (n = 10). Spleen and MTA were harvested for assays on day 15, and samples from 3 out of 10 mice were pooled and assayed as n = 3. (A) Significant improvement in the prognosis of mice receiving anti-SNCA/PD1 combination therapy. (B) SNCA⁺ subsets in CD11b⁺ myeloid cells and CD3⁺ T cells in MTA and spleen (n = 3). (C) Immune cells in MTA (n = 3). (D) Immune cells in spleen (n = 3). (E) Potent CTLs are generated in the anti-SNCA-treated mice. Splenic CD8⁺ T cells pre-treated with tumor antigen gp70 peptide were cocultured with MC38 cells at 40:1. NK cells were cocultured with Yac-1 cells at 20:1. *P < 0.01, **P < 0.05 versus control by the Mann–Whitney test. Mouse survival was analyzed by Kaplan–Meier method and the Mantel–Cox Log-Rank test. Graphs show means ± SDs. Representative data of three independent experiments.

Collectively, these results suggest that anti-SNCA therapy is effective in hosts with MTA through reduction of immunosuppressive cells followed by increase of anti-tumor effector cells, and is greatly helpful to enhance the anti-PD1 therapeutic efficacy. Targeting SNCA may contribute to improvement of clinical outcomes in the treatment of GI cancer patients, especially with MTA.

Discussion

Peritoneal tumor dissemination and subsequent MTA seriously worsen quality of life and prognosis of GI cancer patients, and thus unique strategies are needed to improve the clinical outcomes. In this study, we identified SNCA expressed in immune cells as a diagnostic and therapeutic target to predict possible unresponsiveness to anti-PD1/PDL1 therapy and effectively treat GI cancer patients with MTA. In cancer, SNCA has been reported in several papers as an antigen expressed in cancer cells such as melanoma. However, there are no reports showing SNCA expression in immune cells other than neural cells within the host. Our study revealed that SNCA is expressed in human and mouse immune cells, and that such SNCA⁺ immune cells, especially within CD3⁺ T cells, were significantly and remarkably increased systemically in both ascites and peripheral blood in gastric cancer patients with MTA, although very few in healthy donors and gastric cancer patients without MTA. Also, in cancer therapy, SNCA is only recognized as a tumor antigen, and no therapeutic approaches targeting it have been developed. However, our study revealed that anti-SNCA therapy significantly induced a strong antitumor effect and significantly extended mouse survival in mouse MTA models with increased SNCA⁺ cells. These suggest that anti-SNCA therapy may be effective for treating gastric cancer patients with MTA and/or increased SNCA⁺ cells. Although anti-PD1/PDL1 therapy has attracted worldwide attention, its efficacy is limited to only a small proportion of patients. Also, although strategies that can predict or enhance its efficacy have been developed for many years, no reliable effective diagnostic and therapeutic methods have yet been established. However, our study revealed that anti-PD1 therapy is ineffective in mouse MTA models with increased SNCA⁺ cells, but combination with anti-SNCA therapy significantly enhanced each monotherapeutic efficacy, thereby greatly improving the mouse prognosis. These suggest that the increase in SNCA⁺ immune cells may be a useful biomarker for predicting unresponsiveness to anti-PD1/PDL1 therapy, and that the combination regimen may be useful for treating gastric cancer patients with MTA, for whom there are few effective treatments. Targeting SNCA will be a promising strategy to improve clinical outcomes in the treatment of GI cancer patients, especially with MTA.

SNCA is a neuropathological molecule that produces fibrils in the central nervous system, and is involved in Parkinson's disease and other synucleinopathies [26]. LAG3 has been identified as a receptor for SNCA to initiate endocytosis, transmission, and toxicity in a mouse Parkinson's disease model [25], although there are many other molecules that LAG3 binds to, such as MHC class II, CLEC4G, LGALS3, and FGL1 [27]. Anti-LAG3 mAb has attracted attention as a second-generation IC inhibitor following anti-CTLA4/PD1/PDL1 mAbs in cancer immunotherapy [28]. In our study, however, only a slight impact of anti-LAG3 therapy was seen on mouse survival in mouse MTA models, whereas anti-SNCA therapy greatly induced anti-tumor effects in the same models. This implies that SNCA itself may play an immunological role, or that other molecules besides LAG3 may cooperate with SNCA in the mechanism underlying MTA progression. The further study is needed to clarify at the molecular and cellular levels how the SNCA⁺ cells facilitate the onset, recurrence, and exacerbation of MTA in the hosts.

SNCA blockade has been experimentally evaluated for neurodegenerative diseases, such as Parkinson's disease [19] and Lewy body dementia [18], but not for cancer. Some of the anti-SNCA mAbs, such as Prasinezumab/PRX002, ABBV-0805, BIIB054, LU AF82422, and MEDI1341, have been clinically developed for Parkinson's disease, and its safety has been already confirmed [29]. Repositioning these mAbs to cancer therapy by combining anti-PD1/PDL1 therapy, which has been widely used in clinical settings, may enable faster development of the mAbs for cancers, including GI cancers. As an alternative strategy, combination therapy with anti-CTLA4 and anti-IL1B mAbs, which significantly improved mouse prognosis in mouse MTA models, may also be quickly applied to the treatment of GI cancers in clinical settings, since both mAbs have already been clinically developed for cancer.

Clinical analysis of patient's samples using advanced technologies have been increasing in recent years. For example, a single-cell transcriptome profiling of peritoneal carcinomatosis obtained from GC patients revealed 12 genes as a prognostic signature [30]. Another comprehensive multi-omic analysis of MTA obtained from GC patients also identified a combination of genes as a key prognostic factor [31]. In the study, they found drugs that were effective in mouse xenograft models transplanted with patient-derived tumor cells. However, these studies used immunodeficient mice, which lack a complete immune system, in the in vivo therapeutic experiments, focusing on the cancer side, but not host immunity side. Accumulating evidence suggests that tumor cells evolve not only intrinsically but also extrinsically via the tumor-host interactions [11]. Thus, it is important to analyze and understand both tumor cells and host factors including most influential immune factors within the peritoneal cavity in order to successfully control cancer and MTA of the patients.

We were able to identify SNCA as an immunological determinant in MTA progression in GI cancer by understanding the differences of the i.p. immune status between MTA and tumor-free ILF of GC patients. However, many issues remain to be resolved, such as the relationship between the level of SNCA⁺ immune cells and clinical responses/prognosis of patients, and the relationship between the level of SNCA⁺ immune cells and tumor features (mutational burden, DNA mismatch repair status, PDL1 expression levels, HER2 positivity, etc.), which have been identified as biomarkers to predict possible anti-PD1 responses in cancer. In conjunction with basic research, the further clinical study with a larger number of patients would greatly contribute to improving clinical outcomes in the treatment of various types of cancers, including GI cancers. We hope that this study will be the first step to clarify the significance of SNCA in cancer, beyond the boundaries of the field of neurological diseases.

Funding sources

This study was supported by the Japan Agency for Medical Research and Development P-CREATE (106209 to C. K.-S.), and Chiome Bioscience Inc. (No. C2017-143 to C. K.-S.).

CRediT authorship contribution statement

Chie Kudo-Saito: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Hiroshi Imazeki: Writing – original draft, Resources, Investigation, Data curation. Hiroki Ozawa: Writing – original draft, Resources, Investigation, Data curation. Hirofumi Kawakubo: Supervision, Resources, Methodology. Hidekazu Hirano: Resources. Narikazu Boku: Supervision, Resources. Ken Kato: Supervision. Hirokazu Shoji: Supervision, Resources, Project administration, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

C.K.-S. received honoraria from Ono Pharmaceutical and Bristol Myers Squibb, and a research grant from Chiome Bioscience. H.I. received honoraria from Ono Pharmaceutical. H.H. received honoraria from Novartis, Ono Pharmaceutical, Taiho Pharmaceutical, Teijin Pharma, and Nichi-Iko. N.B. received honoraria from Eli-Lilly, Bristol Myers Squibb, Ono Pharmaceutical, Taiho Pharmaceutical, and Daiichi-Sankyo, and research grants from Ono Pharmaceutical and Takeda Pharmaceutical. K.K. received honoraria and research grants from Ono Pharmaceutical and Bristol-Myers Squibb. H.S. received honoraria from Ono Pharmaceutical and Bristol Myers Squibb, and research grants from Ono Pharmaceutical and Takeda Pharmaceutical. Other authors have no competing