You are here

Article Text

Article menu
Immune cell therapies and immune cell engineering
Original research
Microenvironmental alkalization promotes the therapeutic effects of MSLN-CAR-T cells
  1. Min Wu1,2,
  2. Ling Mao1,
  3. Xuejia Zhai1,3,
  4. Jie Liu1,
  5. Junhan Wang1,
  6. Langhong Li1,
  7. Jiangjie Duan1,2,
  8. Jun Wang1,2,
  9. Shuang Lin4,
  10. Jianjun Li3 and
  11. Shicang Yu1,2
  1. 1Department of Stem Cell and Regenerative Medicine, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, Chongqing, China
  2. 2Jin-feng Laboratory, Chongqing, Chongqing, China
  3. 3Deaprtment of Oncology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, Chongqing, China
  4. 4International Joint Research Center for Precision Biotherapy, Ministry of Science and Technology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, Chongqing, China
  1. Correspondence to Professor Shicang Yu; yushicang{at}163.com; Professor Jianjun Li; jianjunli{at}tmmu.edu.cn

Abstract

Triple-negative breast cancer (TNBC) is characterized by high invasion, prone metastasis, frequent recurrence and poor prognosis. Unfortunately, the curative effects of current clinical therapies, including surgery, radiotherapy, chemotherapy and immunotherapy, are still limited in patients with TNBC. In this study, we showed that the heterogeneous expression at the protein level and subcellular location of mesothelin (MSLN), a potential target for chimeric antigen receptor-T (CAR-T) cell therapy in TNBC, which is caused by acidification of the tumor microenvironment, may be the main obstacle to therapeutic efficacy. Alkalization culture or sodium bicarbonate administration significantly promoted the membrane expression of MSLN and enhanced the killing efficiency of MSLN-CAR-T cells both in vitro and in vivo, and the same results were also obtained in other cancers with high MSLN expression, such as pancreatic and ovarian cancers. Moreover, mechanistic exploration revealed that the attenuation of autophagy-lysosome function caused by microenvironmental alkalization inhibited the degradation of MSLN. Hence, alkalization of the microenvironment improves the consistency and high expression of the target antigen MSLN and constitutes a routine method for treating diverse solid cancers via MSLN-CAR-T cells.

  • tumor microenvironment
  • breast cancer
  • immunotherapy
  • chimeric antigen receptor (CAR)

Data availability statement

Data are available on reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/jitc-2024-009510

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Key messages

    • The intricate microenvironment of solid tumors, characterized by surface antigen heterogeneity and loss, T-cell infiltration and exhaustion, contributes to the limited effectiveness of chimeric antigen receptor-T (CAR-T) therapy.

    • Acidic tumor microenvironment promotes the heterogeneous expression of methothelin (MSLN) and inhibits the therapeutic potency of MSLN-CAR-T cells in diverse cancer models.

    • The combination of microenvironmental alkalinization and CAR-T-cell therapy may represent a standard approach for treating solid cancers.

    Introduction

    Breast cancer (BC) is one of the most common malignant tumors in women worldwide and is estimated to account for 24.5% of all cancers.1–3 The incidence of triple-negative breast cancer (TNBC), which accounts for 10%–20% of all breast cancer cases, is gradually increasing.4 Patients with this subtype of breast cancer have a poor prognosis due to rapid progression.2 5 Since there is no obvious drug target, patients with TNBC are not sensitive to endocrine or molecular targeted therapy, and surgeries involving postoperative chemotherapy still cannot completely remove the lesion, with an in situ recurrence rate as high as 25%.1 2 6 Therefore, there is an urgent need to explore new options and strategies for TNBC treatment.

    In recent years, chimeric antigen receptor-T (CAR-T) cells have achieved encouraging clinical effects in the treatment of hematological tumors. Single-chain antibodies (scFvs) and T-cell activation domains were used to modify the T cells. When CARs are expressed on the membrane of T cells, genetically modified CAR-T cells can recognize tumor-specific antigens (TSAs) or tumor-associated antigens (TAAs), which are expressed on target cells, and initiate killing functions to eliminate cancer cells without major histocompatibility complex (MHC) interactions.7 8 To enhance the precision of the immune response, selecting appropriate TSAs or TAAs that are homogeneously expressed on the membrane of cancer cells so that CAR-T cells can specifically bind to tumor surface antigens and kill cancer cells without affecting normal tissues or organs is an important prerequisite for CAR-T-cell therapy. MSLN is a membrane glycoprotein that is expressed in normal mesothelial cells but is highly expressed in mesothelioma, breast cancer, ovarian cancer, lung cancer, pancreatic cancer and other cancers.9 10 Moreover, several antibody-based therapeutic agents, as well as vaccines and T-cell therapies directed at the MSLN, are undergoing clinical evaluation, but the results are not satisfactory.11–17 How to improve the therapeutic efficacy of MSLN-CAR-T-cell therapy for TNBC and other solid tumors needs to be further explored.

    In immunotherapy, the complexity of the solid tumor microenvironment (TME) may influence T-cell infiltration, T-cell exhaustion and tumor phenotypes, such as surface antigen heterogeneity and antigen loss.18–21 Tumor cells produce many hydrogen ions, lactic acid and pyruvate during glycolysis, which contribute to the acidic TME.18 22 23 The closer to the center of the cancer nest, the lower the pH is, whereas the closer to the interstitium, the greater the pH is.24–26 The autophagy-lysosome pathway plays an important role in biological processes, such as macromolecular degradation, antigen presentation and plasma membrane repair.27 The acidic environment within lysosomes is maintained within a narrow pH range (4.5-5.0), which is optimal for the digestion of autophagic cargo macromolecules, and fluctuations in the internal or TME pH may cause changes in lysosomal activity.28 Unfortunately, whether the TME can directly affect lysosomal activity, MSLN heterogeneity and CAR-T-cell therapy efficacy is currently unknown. We hypothesized that changes in the pH of the TME may affect lysosomal activity to influence the expression of the MSLN protein and ultimately alter the ability of MSLN-CAR-T cells to kill TNBC cells.

    In this study, bioinformatics and tissue microarray analysis revealed that MSLN expression on the cell membrane was significantly greater in TNBC patient samples than in non-TNBC patient samples and normal tissue samples and that MSLN expression exhibited a markedly heterogeneous distribution. Importantly, the expression level and cell membrane localization of MSLN were positively correlated with pH, and inhibiting the function of the autophagy-lysosome system via TME alkalization restored the cytomembrane density of MSLN. Moreover, alkaline culture conditions enhanced the killing effect of MSLN-CAR-T cells against TNBC cells in vitro. Furthermore, the administration of exogenous sodium bicarbonate (NaHCO3) upregulated MSLN expression and increased its cell membrane localization, thereby promoting the in vivo efficacy of CAR-T-cell therapy in TNBC and in pancreatic and ovarian cancers.

    Results

    The heterogeneous expression of MSLN limits the effectiveness of MSLN-CAR-T cells against TNBC

    First, RNA sequencing (RNA-seq) analysis, membrane protein analysis and immunohistochemistry analysis were used to identify candidate antigens of CAR-T-cell therapy for TNBC (figure 1A). By database mining, the gene expression profiles of patients with TNBC and patients considered non-TNBC were compared, and 19 genes were found to be highly expressed in 4 TNBC patient cohorts (figure 1B, online supplemental table S1). Analysis of the Human Protein Atlas database (https://www.proteinatlas.org/) revealed that MSLN, PROM1, GABRP, CRYAB, ART3 and UGT8, among 19 highly expressed proteins, were located on the cell membrane (online supplemental table S1). Furthermore, the analysis of 6 membrane-localized proteins in The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) datasets (Breast cancer: n=1099; normal: n=292) revealed that only MSLN was highly expressed in breast cancer tissues compared with normal tissues (normal vs breast cancer: 0.26±0.05 vs 0.34±0.10, p=5.95×10-4) (figure 1C). MSLN might be a potential target for CAR-T-cell therapy in TNBC. Next, immunohistochemistry (IHC) staining of MSLN in tissue chips (BR1901, TNBC: n=54; non-TNBC: n=41) verified that the expression of MSLN was significantly greater in TNBC samples than in non-TNBC samples and that MSLN was located mainly on the cell membrane (figure 1D,E; online supplemental figure S1). Moreover, flow cytometry revealed that MSLN was significantly more highly expressed on TNBC cell membranes than on non-TNBC and normal cell membranes (figure 1F, online supplemental figure S2). At present, some clinical studies have used MSLN-CAR-T cells as a treatment target for breast cancer, but the effect was unsatisfactory.29 30 Our results suggested that MSLN might be more suitable for CAR-T-cell therapy in patients with TNBC than in patients considered non-TNBC.

    Supplemental material

    Figure 1
    Figure 1

    The heterogeneous expression of mesothelin (MSLN) limits the effectiveness of MSLN-chimeric antigen receptor-T (CAR-T) cells against triple-negative breast cancer (TNBC). (A) Schematic diagram of the screening process. TNBC, non-TNBC, normal messenger RNA (mRNA) and tissue immunohistochemistry (IHC) data were used to screen potential targets. (B) Venn diagram showing the genes highly expressed in patients with TNBC compared with patients considered non-TNBC in different datasets (GSE65216, n=82, log fold change (logFC)>2.5, p<0.05; GSE7904, n=41, logFC>3.0, p<0.05; GSE20194, n=54, logFC>1.5, p<0.05; GSE19615, n=231, logFC>1.0, p<0.05). Details of the 19 differentially expressed genes between patients with TNBC and patients considered non-TNBC in different datasets are presented in online supplemental table S1. (C) Differential analysis of six genes (MSLN, PROM1, GABRP, CRYAB, ART3 and UGT8) in breast cancer tissue (n=1099) and normal tissues (n=292) in the The Cancer Genome Atlas and Genotype-Tissue Expression datasets. The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests. (D) MSLN expression detected by IHC in non-TNBC and TNBC tissue array samples. Representative images. Upper panel, non-TNBC. Down panel, TNBC. Scale bars, 50 μm and 20 μm. (E) Average H-scores of MSLN expression determined by IHC in non-TNBC (n=41) and TNBC (n=54) tissue array samples. The H-scores according to online supplemental figure S1. The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests. (F) Statistical analysis of the expression of MSLN in normal (MDA-KB2), non-TNBC (SK-BR-3 and BT474) and TNBC cell lines (MDA-MB-231, HCC38, MDA-MB-453 and HCC70) according to online supplemental figure S2. The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests. (G) Schematic diagram of the second-generation MSLN-CAR lentiviral vector. (H, I) MDA-MB-231 (H) or HCC38 (I) cells (2×104) were cocultured with T cells or MSLN-CAR-T cells at effector-to-target (E:T) ratios of 1:1, 1:5 and 1:10 for 6 hours, and the killing efficiency was evaluated via the lactate dehydrogenase (LDH) assay. All the experiments were performed at least three times. Quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests. (J) Schematic diagram of the tumor model generated using MSLN-specific CARs. T or MSLN-CAR-T cells were administered to NOG(NOD.Cg-PrkdcscidIl2rgtm1Sug/JicCrl) mice 4 days after MDA-MB-231-Luc or HCC38-Luc tumor inoculation via the tail vein. In vivo imaging was used to observe the changes in tumors on different days. (K, L) In vivo imaging was used to observe the changes in tumor volume at 7, 11, 18, 21 and 25 days after MDA-MB-231 tumor inoculation, n=5 per group (K). Quantitative analysis of the relative fluorescence intensity (RFI). The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-way analysis of variance (ANOVA) (L). (M, N) In vivo imaging was used to observe the tumor volume at 6, 9, 13, 16 and 20 days after HCC38 tumor inoculation, n=5 per group (M). Quantitative analysis of the relative fluorescence intensity (RFI). The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-way ANOVA (N). (O, P) MSLN expression in xenograft tissue detected by IHC. Representative image, scale bars, 100 μm and 20 μm (O). The quantitative analysis of MSLN expression detected by IHC in xenograft tissue. The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests (P).

    The killing effect of MSLN-CAR-T cells on TNBC was subsequently investigated in vitro and in vivo (online supplemental figure S3A). A second-generation MSLN-CAR vector (figure 1G) was used, and the percentage of CAR-T cells expressing the CARs was approximately 33% (online supplemental figure S3B). The MSLN-positive TNBC cell lines MDA-MB-231 and HCC38 were cocultured with MSLN-CAR-T cells. Approximately 35% of the MDA-MB-231 cells were killed by MSLN-CAR-T cells at an effector-to-target ratio (E:T) of 1:10, which was approximately 10% greater than that of the T cells after 6 hours of killing (figure 1H). Similar results were also observed in HCC38 cells (figure 1I). MSLN-CAR-T cells displayed strong killing effects against TNBC in vitro. Later, the volume of xenografts in tumor-bearing NOG mice was evaluated to determine the in vivo killing effects of MSLN-CAR-T cells (figure 1J). Unfortunately, no significant differences in tumor size were found between the T-cell-treated group and the MSLN-CAR-T-cell-treated group (figure 1K–N). Therefore, we wanted to further explore the reasons behind this phenomenon. Many factors interfere with the poor efficacy of CAR-T-cell therapy in solid tumors, such as tumor antigen target loss, TME changes, immunosuppression and T-cell infiltration and exhaustion.31 Because MSLN expression in the xenografts was low and heterogeneous, although T-cell infiltration was increased in the CAR-T-cell therapy group (figure 1O,P; online supplemental figure S4A–C), we hypothesized that the loss of tumor antigen-targeted MSLN might limit the effectiveness of MSLN-CAR-T-cell therapy in vivo, which needs to be further explored.

    The expression of MSLN correlated with the distance of TNBC cells from the stroma

    First, we reanalyzed MSLN expression in human TNBC tissue samples. Single-cell and tissue spatial transcriptome datasets of patients with TNBC were downloaded from the GEO database, and MSLN-positive and MSLN-negative cells were indeed found in the same tumor nest (GSE75688: 2% positive vs 98% negative, GSE118389: 8% positive vs 92% negative, CID4465: 17% positive vs 83% negative, CID44971: 6% positive vs 94% negative) (online supplemental figure S5A-J; figure 2A,B). Interestingly, the expression of MSLN decreased stepwise from the leading edge to the center of the CID4465 TNBC tissue (figure 2A; online supplemental figure S5E, J).

    Figure 2
    Figure 2

    The expression of mesothelin (MSLN) correlated with the distance of triple-negative breast cancer (TNBC) cells from the stroma. (A) Spatial expression and distribution of MSLN messenger RNA (mRNA) in the TNBC tissue spatial transcriptome datasets CID4465 and CID44971. Left panel, MSLN mRNA in TNBC cells. Right panel, MSLN mRNA in TNBC tissue (including other cell types). The cell type and expression analysis were performed according to online supplemental figure S5. (B) mRNA expression of MSLN in TNBC single-cell transcriptome datasets. GSE75688: 88 malignant cells; GSE118389: 459 malignant cells; CID4465: 123 malignant cells; CID44971: 774 malignant cells. The solid blue circles indicate the proportion of MSLNs, and the hollow circles indicate the proportion of MSLNs. (C) Differential expression ratios of MSLN in tissue array samples determined by immunohistochemistry (IHC) staining (TNBC tissue samples, n=54; non-TNBC tissue samples, n=41). The MSLN staining grade of the tumor cells was divided into 0, 1, 2 and 3 according to online supplemental figure S1. (D) Representative images of the spatial distribution of MSLN in TNBC tissue array samples. Left: scale bars, 200 μm. Right: scale bars, 20 μm. (E) Model of TNBC cells at different distances from the stroma. (F) MSLN expression in tumor cells at different distances from the stroma in TNBC tissue samples. n=10. The MSLN staining grade of the tumor cells was divided into 0, 1, 2 and 3 according to online supplemental figure S1. (G) Quantitative analysis of the expression of MSLN in TNBC cells at different distances from the stroma, n=10. The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests.

    The distribution of MSLN in TNBC cells was subsequently assessed in a tissue microarray via IHC. Consistent with the results derived from the single-cell and tissue spatial transcriptome datasets, the expression of MSLN also showed obvious heterogeneity, even within the same cancer nest (figure 2C), and was highly expressed around the leading edge of the carcinoma nest (figure 2D). Moreover, we measured the relationship between the expression of MSLN and the distance between cancer cells and the stroma (figure 2E). TNBC cells closer to the stroma presented higher MSLN expression (figure 2F,G). These results suggested that there was significant heterogeneity in the expression of MSLN in the same TNBC tissue sample, which may be an important reason for the failure of MSLN-CAR-T-cell therapy.

    The pH of the culture medium affects the expression and subcellular localization of MSLN

    The Warburg effect is a metabolic characteristic of solid tumors32 that reduces the pH of the TME to approximately 6.5.25 33 Previous studies have shown that the pH gradient is correlated with the distance from the cancer nest,25 similar to the results of the correlation analysis of MSLN expression in TNBC tissue (figure 2A and D). We then analyzed whether the pH of the TME could affects the expression of MSLN in vitro. Different TNBC cell lines were treated with acidic (pH=6.5, acidic conditions of the TME), normal (pH=7.4, normal tissue conditions) or alkaline (pH=8.5, alkaline conditions) culture media. Compared with that under normal culture conditions, the expression of MSLN under acidic conditions was significantly downregulated and markedly upregulated under alkaline conditions (figure 3A; online supplemental figure S6A). Moreover, the expression level of MSLN was time dependent under alkaline culture conditions (figure 3B; online supplemental figure S6B). Furthermore, the cell membrane expression levels of MSLN in TNBC cell lines were also obviously increased under alkaline culture conditions (figure 3C–E; online supplemental figure S6C–E). Consistently, the fluorescence intensity of MSLN in TNBC cell lines detected by immunofluorescence staining was also obviously increased under alkaline conditions (figure 3F,G; online supplemental figure S6F, G). Next, we analyzed the fluorescence intensity of MSLN between the cell membrane and cytoplasm by a distance model (1 μm) (figure 3H). Compared with that under normal conditions, the membrane expression level of MSLN was significantly lower in TNBC cells cultured under acidic conditions, while the membrane expression level of MSLN was markedly greater in TNBC cells cultured under alkaline conditions (figure 3I; online supplemental figure S6H). These results indicate that the pH of the TME is responsible for the heterogeneity and membrane expression of MSLN in TNBC cells.

    Figure 3
    Figure 3

    The pH of the tumor microenvironment affects the expression levels and subcellular localization of mesothelin (MSLN). Different triple-negative breast cancer (TNBC) cell lines (MDA-MB-231 and HCC38) were treated with acidic (pH=6.5), normal (pH=7.4) or alkaline (pH=8.5) culture media, after which the cells were collected or analyzed after 48 hours of treatment. All the experiments were performed at least three times, and representative data from one experiment are shown. The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests. (A) The expression of MSLN in different culture media was detected by western blot analysis. (B) The expression of MSLN at different time points was detected by western blot analysis. (C) The expression of MSLN in cell membranes was analyzed via a membrane separation assay via western blot analysis. (D) The percentage of MSLN cells positive for different culture conditions was measured via flow cytometry in the PE (phycoerythrin) channel. (E) Quantitative analysis of MSLN expression in TNBC cells cultured under different conditions; n=3 or 4. (F) The expression of MSLN under different pH culture conditions was detected by immunofluorescence staining. A representative image is shown. MSLN: green, DIL(1,1′-Dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate): red, DAPI (4’,6-Diamidino-2’-phenylindole): blue. Scale bars, 20 μm. Upper panel, MDA-MB-231 cells; lower panel, HCC38 cells. (G) Quantitative analysis of the expression of total MSLN in TNBC cells cultured under different pH conditions, n=100. (H, I) Expression of MSLN within 1 μm from the cell membrane to the cytoplasm under different pH culture conditions. Schematic diagram (H). Quantitative analysis, n=100 (I).

    Alkalization of culture media inhibits the degradation of MSLN in lysosomes

    To clarify the relationship between the expression of MSLN and the pH of the TME, we examined MSLN messenger RNA (mRNA) under different pH conditions and found that there were no significant differences (figure 4A). Thus, we hypothesized that the pH of the TME might affect MSLN degradation. Cellular proteins can be degraded through the autophagy-lysosomal or ubiquitin-proteasome pathways, and pH can impact the activity of degradation enzymes in these two organelles.34 35 Chloroquine (CQ), an inhibitor of lysosomal activity, upregulated MSLN expression in a dose-dependent and time-dependent manner (online supplemental figure S7A, B), but MG132, an inhibitor of proteasome activity, did not change the expression of MSLN (figure 4B). Immunofluorescence staining revealed that MSLN colocalized with LAMP1 and LC3B (figure 4C), and flow cytometry also revealed that the cell membrane expression levels of MSLN in TNBC cell lines were obviously increased by CQ treatment (figure 4D), suggesting that the autophagy-lysosomal pathway may alter the recycling and degradation of MSLN. We found that the intracellular pH (pHi) increased with increasing TME pH (figure 4E; online supplemental figure S7C). Moreover, the expression levels of p62, a protein involved in lysosomal degradation, were lower under acidic conditions and greater under alkaline conditions (figure 4F), and the flow cytometry activity assay also revealed that lysosomes under acidic conditions had greater activity and alkaline conditions had less activity (figure 4G), which suggested lower degradation of MSLN under alkaline conditions. Similar phenotypes were observed in the MDA-MB-231 and HCC38 cells (figure 4A–G). These results indicated that the acidic TME causes the degradation and heterogeneity of MSLN through lysosomes and that the alkalization of culture media can increase MSLN by inhibiting degradation.

    Figure 4
    Figure 4

    Alkalization of culture media inhibits the degradation of mesothelin (MSLN) in lysosomes. Different triple-negative breast cancer (TNBC) cell lines (MDA-MB-231 and HCC38) were treated with acidic (pH=6.5), normal (pH=7.4) or alkaline (pH=8.5) culture media and then collected after 48 hours. All the experiments were performed at least three times, and representative data from one experiment are shown. Quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests. (A) Messenger RNA (mRNA) expression of MSLN detected by quantitative PCR. (B) MDA-MB-231 and HCC38 cells were treated with chloroquine (CQ) (20 μM) and MG132 (20 μM) for 6 hours, and the expression of MSLN was detected by western blot analysis. (C) Localization of MSLN relative to LAMP1-positive lysosomes or LC3B-positive autophagosomes. The cells were detected via immunofluorescence staining. MSLN: green. Up, LAMP1: red. Down, LC3B: red. DAPI: blue. Scale bars, 20 μm. (D) The membrane expression of MSLN after treatment with different concentrations of CQ was evaluated via the PE channel flow cytometry. (E) Intracellular pH changes detected by 2',7'-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein acetoxymethyl ester staining, n=5 or 6. (F) The p62 protein was detected by western blot analysis under different culture conditions. (G) Lysosomal activity was detected via a lysosomal intracellular activity assay kit. The dashed lines represent the control activity at pH=6.5.

    Alkalization of culture medium promotes the killing effect of MSLN-CAR-T cells on TNBC in vitro

    According to previous evidence, the membrane expression of MSLN is upregulated under alkaline conditions, but whether this strategy can improve the killing efficiency of MSLN-CAR-T cells is unclear. TNBC cells treated with culture media at different pH values (6.5, 7.4 and 8.5) had no obvious variation of proliferation, nor did those treated with T cells or CAR-T cells (online supplemental figure S8A, B). Better killing effect of MSLN-CAR-T cells was demonstrated at pH 8.5 and pH 7.4 than at pH 6.5 (figure 5A,B; online supplemental figure S8C). However, there was no significant difference in the killing efficacy of untreated T cells against TNBC cells under the same pH culture conditions (figure 5A,B; online supplemental figure S8C). The killing efficiency of MSLN-CAR-T cells against HCC38 cells greatly improved to 80% under alkaline conditions (figure 5B). We also found that MSLN-CAR-T cells could release more granzyme B under normal and alkaline conditions after 6 hours of the killing process (figure 5C,D). Pre-killing and post-killing MSLN-CAR-T-cell subsets at different pH values were analyzed, and the proportions of CD8+ cells, CD8+ effector memory T (Tem) cells and CD4+ Tem cells increased at pH 7.4 and 8.5 (figure 5E; online supplemental figures S9 and S10). These cells are known to be responsible for extending the killing effect. In conclusion, alkaline culture conditions improved the killing efficiency of MSLN-CAR-T cells against TNBC cells in vitro.

    Figure 5
    Figure 5

    Alkalization of the culture medium promoted the killing effect of mesothelin (MSLN)-chimeric antigen receptor-T (CAR-T) cells on triple-negative breast cancer (TNBC) cells in vitro. (A) MDA-MB-231 cells pretreated with culture medium at different pH values for 48 hours were then cocultured with T cells or MSLN-CAR-T cells at effector-to-target (E:T) ratios of 1:0.5, 1:1, 1:5 or 1:10 for 6 hours, after which the killing efficiency was evaluated via the lactate dehydrogenase (LDH) assay. (B) HCC38 cells pretreated with culture medium at different pH values for 48 hours were then cocultured with T cells or MSLN-CAR-T cells at E:T ratios of 1:0.5, 1:1, 1:5 or 1:10 for 6 hours, after which the killing efficiency was evaluated via the LDH assay. (C, D) Granzyme B released under different pH conditions was detected via flow cytometry. Representative results (C). Quantitative analysis, n=4 (D). (E) The proportions of CD4+, CD4+ naïve, CD4+ central Memory T (Tcm), CD4+ effector memory T (Tem), CD8+, CD8+ naïve, CD8+ Tcm and CD8+ Tem cells before and after the killing of MSLN-CAR-T cells were detected via flow cytometry, and the data were drawn with a heatmapper (http://www.heatmapper.ca/expression/) according to online supplemental figures S9 and S10. All the experiments were performed at least three times, and representative data from one experiment are shown. Quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests.

    Microenvironmental alkalization enhances MLSN expression in TNBC in vivo

    Many studies have reported that oral NaHCO3 can increase the pH of tumor tissue36–38 and has a limited inhibitory effect on tumor growth.39 Therefore, the progression of a TNBC xenograft model in NOD-SCID mice treated with NaHCO3 was observed (figure 6A). We found that NaHCO3 (100 mM, 200 mM) increased the pH of the tumor tissue (figure 6B) but had no effect on tumor growth (figure 6C), body weight (online supplemental figure S11A) or overall survival (online supplemental figure S11B), while the expression of MSLN increased with increasing NaHCO3 (figure 6D). Local regions of MSLN expression were observed in the 100 mM group. In the 200 mM group, the expression of MSLN increased, and the expression of MSLN in the tissue surrounding the tumor was significantly greater than that in the tumor interior, accompanied by high MSLN expression in the cell membrane (figure 6D). However, MSLN was only slightly expressed in some scattered areas in the control group (figure 6D). To further verify the therapeutic effect of CAR-T cells in vivo, NOG mice were treated with 200 mM NaHCO3 simultaneously (figure 6E). NaHCO3 (200 mM) had little effect on MDA-MB-231 tumor growth (figure 6F,G) or body weight (online supplemental figure S11C), but upregulated the expression of MSLN (figure 6I,J). The pH decreased with increasing distance from the tumor center (approximately 7.1–6.5), but the addition of the 200 mM xenograft model maintained the entire tumor at a relatively consistent pH (approximately 7.2) (figure 6H).

    Figure 6
    Figure 6

    Microenvironmental alkalization with sodium bicarbonate (NaHCO3) enhances mesothelin (MSLN) expression in vivo. (A) Schematic diagram of the tumor model treated with NaHCO3. MDA-MB-231 cells were then subcutaneously inoculated into the left upper limb of NOD-SCID mice. When the tumor volume reached approximately 100 mm3, the tumor-bearing mice were treated with 0, 100 mM or 200 mM NaHCO3. (B) The internal tissue pH was measured via a microelectrode, n=3. (C) The volume of xenografts in tumor-bearing mice treated with different concentrations of NaHCO3, n=5. P values were determined by unpaired two-way analysis of variance (ANOVA). (D) MSLN expression was detected via immunohistochemistry. Scale bars, 100 μm and 20 μm. (E) Schematic diagram of the tumor model treated with NaHCO3. MDA-MB-231 cell tumor-bearing NOG mice treated with or without 200 mM NaHCO3. (F, G) Changes in the MDA-MB-231 xenograft size detected by in vivo imaging at 4, 7, 18, 21 and 25 days after tumor inoculation. Representative images (F). The relative fluorescence intensity was statistically analyzed. The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-way ANOVA, n=4 (G). (H) The internal tissue pH was measured via a microelectrode, n=3. (I, J) MSLN expression was detected via immunohistochemistry. Representative images. Scale bars, 100 μm and 20 μm (J). Quantitative analysis, n=65. The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests (J).

    Microenvironmental alkalization promotes the killing effects of MSLN-CAR-T cells against TNBC in vivo

    Next, MSLN-CAR-T cells were injected into MDA-MB-231 and HCC38 tumor-bearing mice via the tail vein, and 0 or 200 mM NaHCO3 was orally administered (figure 7A). Using serial in vivo imaging, we found that the volume of xenografts in the NaHCO3 group decreased, demonstrating that alkalization treatment had obviously increased the killing ability of MSLN-CAR-T cells in vivo (figure 7B–E). The survival time of the tumor-bearing mice in the NaHCO3 plus MSLN-CAR-T-cell group was also longer than that of the other treatment groups (MDA-MB-231, p =0.0021, median survival: 43.0 day s vs 60.0 days. HCC38, p=0.0021, median survival: 36.0 days vs 53.0 days)(figure 7F,G). We also detected the persistence of CAR-T cells in the peripheral blood of tumor-bearing mice via flow cytometry. CAR-T cells were detected 30 days after treatment in the NaHCO3 group (figure 7H,I). The expression of MSLN in cancer cells was slightly greater than that in the control group (figure 7J–M), probably because many MSLN-positive cells had been killed. We also observed an increase in the number of infiltrating T cells (figure 7N–Q), which may have been promoted by the administration of NaHCO3. These results indicate that microenvironmental alkalization can improve the killing efficiency of MSLN-CAR-T cells against TNBC cells in vivo.

    Figure 7
    Figure 7

    Exogenous administration of sodium bicarbonate (NaHCO3) promoted the killing effects of mesothelin (MSLN)-chimeric antigen receptor-T (CAR-T) cells against triple-negative breast cancer (TNBC) in vivo. (A) Schematic diagram of the tumor therapy model using MSLN-specific CARs. MDA-MB-231-Luc cells or HCC38-Luc cells were then subcutaneously inoculated into the left upper limb of NOG mice. MSLN-CAR-T-cell therapy with or without 200 mM NaHCO3 was administered 3 days after tumor inoculation. (B, D) Changes in xenografts in MDA-MB-231 tumor-bearing mice detected by in vivo imaging at 10, 14, 18, 20 and 22 days after tumor inoculation (B). The relative fluorescence intensity was statistically analyzed. The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-way analysis of variance (ANOVA), n=5 (D). (C, E) Changes in xenografts in HCC38 tumor-bearing mice detected by in vivo imaging at 8, 11, 14, 18 and 31 days after tumor inoculation (C). The relative fluorescence intensity was statistically analyzed. P values were determined by unpaired two-way ANOVA, and quantitative data are presented as the means±SDs, n=5 (E). (F, G) Survival analysis of the MDA-MB-231 (F) and HCC38 (G) NOG models, n=10. P values were determined by the log-rank (Mantel-Cox) test. (H, I) The persistence of CAR-T cells in the blood of MDA-MB-231 (H) and HCC38 (I) cells was measured via flow cytometry in the fluorescein isothiocyanate channel, n=3. The quantitative data are presented as the means±SDs, and p values were determined via unpaired two-tailed t-tests. (J, K) MSLN expression in xenografts derived from MDA-MB-231 cells was detected by immunohistochemistry (IHC) staining. Scale bars, 100 μm and 20 μm (J). Quantitative analysis, n=30 (K). (L, M) MSLN expression in xenografts derived from HCC38 cells was detected by IHC. Scale bars, 100 μm and 20 μm (L). Quantitative analysis, n=25 (M). (N, O) The expression of CD3 in MDA-MB-231 xenografts was detected by IHC. Scale bars, 100 μm and 20 μm (N). For the quantitative analysis, n=20 (O). (P, Q) The expression of CD3 in xenografts derived from HCC38 cells was detected by IHC. Scale bars, 100 μm and 20 μm (P). The quantitative data were presented as the means±SDs and p values were determined via unpaired two-tailed t-tests, n=25 (Q).

    Microenvironmental alkalization also enhances the killing effects of MSLN-CAR-T cells on other solid tumors in vivo

    To verify the consistency of microenvironmental alkalization in MSLN-CAR-T-cell therapy, we treated pancreatic cancer and ovarian cancer cells, which also displayed high MSLN expression, with NaHCO3 combined with CAR-T-cell therapy. Pretreatment of AsPC-1, PANC-1 and SK-OV-3 cells with medium at different pH values for 48 hours had no effect on proliferation (online supplemental figure S12A) but increased the expression of MSLN (online supplemental figure S12B) and the proportion of MSLN-positive cells (online supplemental figure S12C, D). The MSLN-CAR-T cells displayed greater killing efficacy against AsPC-1, PANC-1 and SK-OV-3 cells when cocultured at pH 8.5 and pH 7.4 than when cocultured at pH 6.5 (figure 8A, F and K). When MSLN-CAR-T cells were injected into tumor-bearing NOG mice via the tail vein, we found that the mice fed 200 mM NaHCO3 presented small xenograft sizes according to serial in vivo imaging, although the onset time was inconsistent (figure 8B–D, G–I and L–N). Moreover, the tumor-bearing mice in the MSLN-CAR-T-cell plus NaHCO3 treatment group also exhibited longer survival (AsPC-1, p=0.0495, median survival: 29.5 days vs 49.5 days; PANC-1, p=0.0471, median survival: 34.0 days vs 45.0 days; SK-OV-3 cells, p=0.0499, median survival: 33.0 days vs 42.0 days) (figure 8E, J and O). These results indicate that microenvironmental alkalization combined with CAR-T-cell therapy can also be used in other solid cancers with relatively high MSLN expression. Alkalization of the microenvironment improves the consistency of target antigen expression and may be a routine method for treating diverse solid cancers via MSLN-CAR-T cells.

    Figure 8
    Figure 8

    Microenvironmental alkalization combined with chimeric antigen receptor-T (CAR-T)-cell therapy can also be used for other solid cancers with high mesothelin (MSLN) expression. (A, F and K) AsPC-1 (A), PANC-1 (F) and SK-OV-3 (K) cells pretreated with medium at different pH values for 48 hours were cocultured with T cells or MSLN-CAR-T cells at effector-to-target (E:T) ratios of 1:1, 1:5 or 1:10 for 6 hours at different pH values, and a lactate dehydrogenase (LDH) cytotoxicity assay was used to evaluate the killing efficacy. All the experiments were performed at least three times, and representative data from one experiment are shown. The quantitative data are presented as the means±SDs, and p values were determined via two-tailed t-tests. (B) Schematic diagram of MSLN-specific CAR-T-cell therapy combined with sodium bicarbonate (NaHCO3) therapy in AsPC-1 tumor-bearing mice. (C, D) In vivo imaging was used to observe the changes in the volume of xenografts derived from AsPC-1 cells at 6, 11, 18, 21 and 26 days after tumor inoculation. Representative images (C). The relative fluorescence intensity (RFI) was statistically analyzed, n=4. The quantitative data are presented as the means±SDs, and p values were determined via two-way analysis of variance (ANOVA (D). (E) Survival analysis of AsPC-1 tumor-bearing NOG model mice, n=4. P values were determined by the log-rank (Mantel-Cox) test. (G) Schematic diagram of MSLN-specific CAR-T-cell therapy combined with NaHCO3 therapy in PANC-1 tumor-bearing mice. (H, I) In vivo imaging was used to observe changes in the volume of xenografts derived from PANC-1 cells at 6, 9, 13, 16 and 20 days after tumor inoculation. Representative images (H). The relative fluorescence intensity (RFI) was statistically analyzed, n=5, and the quantitative data are presented as the means±SDs. P values were determined via two-way ANOVA (I). (J) Survival analysis of the PANC-1 tumor-bearing NOG model, n=5. P values were determined via the log-rank (Mantel-Cox) test. (L) Schematic diagram of MSLN-specific CAR-T-cell therapy combined with NaHCO3 therapy in SK-OV-3 tumor-bearing mice. (M, N) In vivo imaging was used to observe the changes in the volume of xenografts derived from SK-OV-3 cells at 6, 9, 13, 16 and 20 days after tumor inoculation. Representative images (M). The relative fluorescence intensity (RFI) was statistically analyzed, n=5. The quantitative data are presented as the means±SDs, and p values were determined via two-way ANOVA (N). (O) Survival analysis of the SK-OV-3 tumor-bearing NOG model mice, n=4. P values were determined by the log-rank (Mantel-Cox) test.

    Discussion

    The antitumor effects of CAR-T-cell therapy on B-cell malignancies are remarkable. However, CAR-T-cell therapy for solid tumor treatment has many limitations, such as antigen deficiency, antigen deletion and loss, T-cell infiltration failure, T-cell exhaustion and microenvironmental immunosuppression.40–42 In this study, we revealed that the expression of MSLN, a potential antigen target for CAR-T-cell therapy in TNBC, was heterogeneous and may be the reason for the failure of CAR-T-cell therapy. Microenvironmental alkalization can attenuate the heterogeneity of MSLN surface expression in TNBC cells, thereby promoting the therapeutic effect of MSLN CAR-T cells in vitro and in vivo, and the alkaline microenvironment may reduce the activity of lysosomes to inhibit MSLN degradation. The effect of alkaline microenvironment treatment on MSLN-CAR-T-cell immunotherapy efficacy was also verified in other solid tumor types, and this strategy might be further assessed in immunotherapy research to broaden the applicability of CAR-T-cell therapy.

    Target antigen selection is a major determinant of the safety and efficacy of CAR-T-cell therapy.42 As a type of cancer that is prone to recurrence and difficult to treat, TNBC lacks ER (estrogen receptor), PR (progesterone receptor) and HER2 (human epidermalgrowth factor receptor 2) expression, and available TAAs for CAR-T-cell therapy are limited.43 Here, we confirmed that MSLN could be a suitable CAR-T-cell therapeutic target in TNBC. However, there is overwhelming evidence that tumors are heterogeneous with respect to both morphology and function, which may lead to patients with heterogeneous TAA expression.44–48 For example, the intertumor heterogeneity of HER2 results in the molecular classification of breast cancer into four subtypes: luminal A, luminal B, HER2+ and basal-like.47 The proportions of CEA-positive cells in paired primary colorectal cancer primary foci, liver metastases and lymph node metastases are different.49 50 Circulating BCMA-expressing cells vary in terms of clonotypic postgerminal center B cells, plasmablasts and both normal and malignant plasma cells.51 Loss of the CD19 antigen during CAR-T-cell therapy results in tumor recurrence.52 In line with the findings of a series of recent studies,53 we found that MSLN was heterogeneously expressed in TNBC cells and tissue samples, even in different areas of one tumor. Moreover, in vivo MSLN-CAR-T-cell therapy also revealed heterogeneous expression of MSLN. This heterogeneity of MSLN protein levels may account for the poor efficacy and off-target toxicity of MSLN-CAR-T cells in vivo.12 54–56

    The physicochemical characteristics of the TME are important determinants of intratumoral heterogeneity.45 57–60 The pH of the TME may regulate tumor cell proliferation, metabolism and senescence through many pathways, such as membrane permeability, enzyme activity, substance uptake and factor release.32 61 To our surprise, the pH of the TME strongly contributed to the heterogeneity of MSLN expression. By treating tumor cells with a normal or alkaline pH, the MSLN expression level and the killing efficacy of MSLN-CAR-T cells were significantly increased in vitro. These findings provide us with a new avenue for increasing the efficacy of CAR-T-cell therapy. Indeed, pH alterations in the TME are necessary in vivo. Reportedly, NaHCO3 can regulate the pH value of tumor cells via the sodium bicarbonate cotransporter SLC4A7, which has no significant difference in expression according to tumor type in the Human Protein Atlas database . Oral administration of NaHCO3 can increase the pH of the TME and significantly prevent metastasis without affecting the growth of primary tumors in prostate cancer, colorectal cancer, liver cancer or spontaneous metastasis mouse models.22 36 62–64 In fact, sodium bicarbonate is widely used in the clinic to treat conditions such as metabolic acidosis, alkalization of the urine, excessive gastric symptoms and non-specific drug poisoning therapeutic. Some clinical trials have used high doses of NaHCO3 to study its safety and pain relief in patients with cancer (NCT02531919, NCT01350583 and NCT01846429). We verified that oral administration of NaHCO3 can upregulate MSLN expression in vivo by changing the intratumoral pH. More importantly, combination treatment with MSLN-CAR-T cells and NaHCO3 changed tumor progression from cold to hot, upregulated the expression of MSLN and increased the therapeutic efficacy of MSLN-CAR-T-cell therapy in vivo. Moreover, NaHCO3 improved CAR-T-cell efficacy in other cancer cells, which also presented high MSLN expression. It is feasible to use NaHCO3 for precision MSLN-CAR-T-cell clinical therapy, but it is unknown whether the expression of other CAR-T-cell TAAs could also be promoted by alkalinization. The specific application scheme needs to be further explored.

    The autophagy and lysosome systems participate in many biological processes, including antigen presentation, macromolecular degradation, intracellular pathogen destruction, exosome release and plasma membrane repair, and play crucial roles in cancer development and progression, such as cancer cell energy metabolism, immune escape, proliferation, invasion, metastasis and tumor-associated angiogenesis.27 65–68 Estrogen receptors can be degraded by lysosomes in breast cancer cells, and lysosome inhibitors can reduce androgen receptor degradation in prostate cancer cells.69 New techniques involving lysosomal degradation (lysosome-targeting chimera, LYTAC; autophagy-targeting chimera, AUTAC) have been applied for cancer therapy.70 Our data on MSLN expression support the functions of autophagy and the lysosome system, and the degradation of MSLN was inhibited because microenvironmental alkalization increased the intracellular pH and inhibited lysosomal activity. These findings suggested that autophagy and lysosome inhibitors could also be used to increase the killing efficiency of MSLN-CAR-T cells. The relationship between the TME and targeted therapy was revealed. However, autophagy inhibition reportedly sensitizes PDAC (Pancreatic ductal carcinoma) to dual ICBs (immune checkpoint blockades) by preventing the degradation of MHC-I.71 Whether proteasome inhibitors affect other protein targets requires additional data.

    Many studies have reported that the immunosuppressive TME in solid tumors reduces the efficacy of therapy and promotes cancer progression.42 72 73 Immunosuppressive blocking antibodies, nanoparticles, tandem blocking signal CAR-T cells and other methods have been used to reduce the immunosuppressive properties of the TME.74–78 Our study confirmed that the infiltration of MSLN-CAR-T cells was limited in vivo, but treatment with NaHCO3 effectively increased the infiltration of MSLN-CAR-T cells. Moreover, T-cell exhaustion is a major factor limiting the clinical efficacy of CAR-T-cell therapy.79 80 Many CAR-T cells manifest tonic signaling during in vitro manufacturing, leading to early exhaustion that limits potency and cytokine production inhibition and gene depletion has been used to slow T-cell exhaustion.81–84 We demonstrated that microenvironmental alkalization did not cause T-cell exhaustion but increased the toxicity of CAR-T cells. Unfortunately, we did not extend CAR-T-cell therapy to observe the persistence of CAR-T cells or the corresponding factors or long-term effects; however, the proportion of Tem cells in vitro and the persistence of CAR-T cells in the peripheral blood in vivo were increased, which suggested the long-term possibility of killing these cells. The increase in the cell proportion may also explain the greater effect of MSLN-CAR-T-cell therapy combined with microenvironmental alkalization. Moreover, immunosuppression-associated checkpoint proteins, such as V-domain immunoglobulin suppressor of T-cell activation (VISTA) and programmed death-ligand 1 (PD-L1), are worthy of attention. VISTA can bind to V-set and Ig domain-containing 3 and P-selectin glycoprotein ligand 1 and maintains T-cell and myeloid quiescence at acidic pH values, such as in the TME, but not at physiological pH.85 Microenvironmental alkalization treatment may effectively relieve the immunosuppression caused by VISTA. The high expression of PD-L1 in TNBC86 also suggests the possibility of combining PD-L1 antibodies with alkalizational immunotherapy.

    Overall, microenvironmental alkalization treatment significantly increased the therapeutic efficacy of MSLN-CAR-T cells, and oral administration of NaHCO3 was also safe and feasible in clinical practice. We believe that the combination of microenvironmental alkalinization and CAR-T-cell therapy can soon be used in clinical treatment.

    Materials and methods

    Study design

    This study hypothesized that that alkalization of the TME attenuates the heterogeneous expression of TAA MSLN and enhances the therapeutic efficacy of MSLN-CAR-T cells in preclinical solid tumor models. In this study, tissue sample information analysis and tissue validation were used to identify antigen targets, and single-cell RNA-seq (scRNA-seq), tissue spatial RNA-seq (tsRNA-seq) and tissue expression data were used to analyze the heterogeneity of MSLN expression. All the analyzed samples were molecularly characterized as part of the routine diagnostic workup, and the specific screening process was as follows. Flow cytometry, western blot analysis, immunofluorescence and lysosomal activity assays under different conditions were used to demonstrate the existence and mechanism of heterogeneity. In vitro killing of TNBC cells by CAR-T cells pretreated under different conditions was used to verify the therapeutic effect of alkalization. The control and treatment groups, the number of biological replicates (sample sizes) and the statistical analyses for each experiment are specified in the figure legends. NaHCO3 was used to change the TME in vivo. The mice in each treatment group were divided into two or three groups of NOD-SCID (NOD SCID/Sja) and NOG (NOD.Cg-PrkdcscidIl2rgtm1Sug/JicCrl) mice: different concentrations of NaHCO3 and treated with T or CAR-T cells. Every group was composed of four to five animals each to ensure statistical power. The animals were randomly assigned to the control or treatment group and housed together to minimize environmental differences and experimental bias. Investigators were not blinded during the follow-up of the mice or evaluation of the in vivo experiments. The data were collected at predetermined time points, with changes and sample collection as described below.

    Analysis of bulk tissue RNA-seq datasets

    The breast cancer bulk tissue RNA-seq datasets GSE65216 (331 breast cancer tissue samples, 14 breast cancer cell lines and 11 normal tissue samples), GSE7904 (43 breast cancer tissue samples, 7 normal breast tissues and 12 normal tissue samples), GSE20194 (230 primary breast cancer tissue samples) and GSE19615 (115 primary breast cancer tissue samples) from the GEO database (http://www.ncbi.nlm.nih.gov/geo), a TCGA dataset (1099 breast cancer tissue samples and 113 normal tissue samples) and a GTEx dataset (179 normal samples) (https://www.genome.gov/Funded-Programs-Projects/Genotype-Tissue-Expression-Project) were downloaded. The differential gene expression profiles of patients with TNBC versus patients considered non-TNBC were analyzed via the R package (limma V.3.20.9). A Venn diagram and columnar scatter plot of differential expression were generated on the Xiantao website (https://www.xiantao.love/).

    Analysis of single-cell RNA-seq and tissue spatial RNA-seq datasets

    The scRNA-seq data from GSM5354528-CID4465 and GSM5354531-CID44971 (the TNBC subset of GSE176078) were downloaded from the GEO. The tsRNA-seq data were downloaded from the Zenodo data repository, and the definition of malignant cells was established according to the original literature.87 The breast cancer sample RNA-seq data and corresponding sample information from the GSE118389 (805 TNBC cells, 459 malignant cells) and GSE75688 (203 TNBC cells, 88 malignant cells) cohorts were downloaded from the GEO.

    The scRNA-seq and tsRNA-seq data were processed in Seurat V.4, and the ‘NormalizeData’ function was used to normalize the expression matrix. Then, ‘FindVariableFeatures’, ‘ScaleData’ and ‘RunPCA’ were used to process the datasets. Clusters were calculated via the FindClusters function and visualized via the uniform manifold approximation and projection dimensional reduction method. The malignant cell subpopulations were subsequently extracted via the ‘subset (object, idents=c1)’ function. The ‘DotPlot’ function, ‘FeaturePlot’ function and ‘DimPlot’ function with ‘cells highlight=whichCells (object, idents=c (1))’ were used to visualize malignant cell subpopulations and the expression of target antigens. The ‘AddModuleScore’ function was used to score the spots in the tsRNA-seq data and determine the spatial location of malignant cells.

    Cell lines and culture conditions

    HEK293T cells; the human breast cancer cell lines HCC38, MDA-MB-231, HCC70 MDA-MB-453 BT-474 and SK-BR-3; the human normal breast cell line MD-KB2; the human pancreatic cancer cell line AsPC-1; the human pancreatic cancer cell line PANC-1 and the human ovarian cancer cell line SK-OV-3 were obtained from the cell bank of the Chinese Academy of Sciences (Shanghai, China). SK-OV-3, AsPC-1, PANC-1, MDA-MB-231, MDA-MB-453, MD-KB2 and HEK293T cells were maintained in Dulbecco’s modified Eagle’s medium (Gibco/Life Technologies, Shanghai, China), and SK-BR-3, HCC70 and HCC38 cells were maintained in Roswell Park Memorial Institute 1640 medium (Gibco/Life Technologies). All culture media contained 10% fetal bovine serum (FBS) (Gibco/Life Technologies) and 1% penicillin-streptomycin (Gibco/Life Technologies). All the cells were cultured at 37℃ in an atmosphere of 5% CO2 (Thermo, 411). We used HCl or NaOH to adjust the pH of the medium (HEPES, Gibco/Life Technologies) with a pH meter measurement (Sartorius, PB-10). The medium was subsequently filtered through 0.22 μm filters (Millipore) and used for cell culture.

    Tissue immunohistochemistry and H&E staining

    Animal tissue samples were fixed with 4% paraformaldehyde for >48 hours, embedded in paraffin and cut into 3 μm thick sections. Human breast cancer tissue arrays were purchased from Alina (XIAN, BR1901, Medical Ethics Committee, Tongxu County People’s Hospital, Henan Province, China). IHC and H&E staining were performed as previously described.88 Briefly, the sections were incubated with anti-MSLN (Abcam, ab133489, 1:100) or anti-CD3 (Abcam, 52959, 1:100) antibodies and stained via an Envision System (Dako, K3468). Whole-slide images were captured via a KF-PRO-120 digital slide scanner (KFBIO, Ningbo, China), and the micrographs were exported by generating whole-slide images via K-Viewer V.1.6.0.28 (KFBIO). The score for each tissue sample (H-score)=percentage of positive tumor cells an IHC staining grade of (1)×1+percentage of positive tumor cells with an IHC staining grade of (2)×2+percentage of positive tumor cells with an IHC staining grade of (3)×3, as shown in online supplemental figure S1, and the average score of 3 different fields for each sample was used as the H-score for statistical analysis. Image-Pro Plus V.6.0 software was used for quantitative analysis.

    Immunofluorescence staining

    Briefly, the cells were cultured on glass coverslips, fixed in 4% paraformaldehyde for 20 min and blocked in 10% goat serum (ZSGB-BIO, ZLI-9021) for 30 min at room temperature. The coverslips were subsequently incubated with the corresponding primary antibodies (MSLN, 1:100; Abcam, ab133489; LAMP1, 1:100; ABclonal, A21194; LC3B, 1:100; ABclonal, A19665) and secondary antibodies (Alexa Fluor 488, 1:500; Invitrogen, A-27034; Alexa Fluor 546, 1:500; Invitrogen, A-11010; Alexa Fluor 647, 1:500; Invitrogen, A-32733). DIL (1,1′-Dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate, Beyotime, C1036, China) was used to stain the cell membrane for 30 min. Finally, the slides were counterstained with DAPI (4′,6-Diamidino-2-phenylindole dihydrochloride, Beyotime, C1005, China) and observed via a laser confocal scanning microscope (Zeiss, LSM980). Image-Pro Plus V.6.0 software was used for quantitative analysis.

    Real-time PCR

    Total RNA was isolated from TNBC cells via RNAiso Plus (TAKARA, 9108), and total RNA was quantified and reverse transcribed (TAKARA, RR047A). Next, MSLN and GAPDH were analyzed via the SYBR PrimeScript PCR Kit II (TaKaRa, RR820A) and quantified via a Bio-Rad CFX96 detection system (Bio-Rad). The following primers were used: MSLN forward, 5’-CAGAGGAGGCTCAGAGAGCTA-3’, MSLN reverse: 5’-GGCTGGAAATGTTAGGTGGGT-3’, GAPDH forward: 5’-CAATGACCCCTTCATTGACC-3’, GAPDH reverse: 5’-GACAAGCTTCCCGTT

    CTCAG-3’. With GAPDH serving as an internal control, the expression of genes was calculated via the comparative threshold cycle method. Each experiment was independently repeated at least three times.

    Western blot analysis

    The cells were lysed in RIPA buffer (Beyotime, P0013B), and the total protein concentration was quantified via a BCA kit (Thermo, 23225). Equal amounts of protein were separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis, transferred to polyvinylidene fluoride membranes and incubated with anti-MSLN (1:1000; Abcam, ab133489), anti-GAPDH (1:1000; Cell Signaling Technology, 2118), Na+K+ATPase (1:1000; Cell Signaling Technology, 2356), LC3B (1:1000; ABclonal, A19665) and P62 (1:1000; ABclonal, A19700) antibodies at 4℃ overnight. The membranes were subsequently incubated with secondary antibodies conjugated to HRP (1:3000; Cell Signaling Technology, 7074S), and the bands were visualized with Super Signal West Dura reagents (Thermo Scientific, 34075) and a FUSION FX 6 (VILBER). GAPDH was used as an internal reference. The membrane protein samples that were separated via a membrane-cytoplasmic kit (Thermo Scientific, 89842) were also examined via western blot analysis, and the Na+/K+ ATPase was used as an internal reference.

    Construction of the CAR vector

    The second-generation MSLN-CAR vector was purchased from Suzhou ICARTAB (LIC012A). Specifically, the MSLN-CAR vector comprises the MSLN-specific target scFv with a CD8 signal sequence, two intracellular costimulatory signaling regions (CD3ζ and 4-1BB), a CD8 hinge region and a transmembrane sequence.

    Lentivirus packaging and production

    The viruses were filtered and collected from the supernatants of HEK293T cells transfected with the MSLN-CAR vector and three helper packaging plasmids (psPAX2, pRRE and pRSV-Rev) for 48 or 72 hours via Lipo3000 (Invitrogen, L3000015). Lenti-X (TaKaRa, 2 109 257A) was added to the viral supernatant at a 3:1 ratio. The cells were resuspended in fresh medium at a 1:3 ratio and set aside. An HIV ELISA kit (PerkinElmer, NEK050001KT) was used for viral titer determination. Luciferase virus was purchased from Tsingke.

    Isolation of primary peripheral blood mononuclear cells

    Fresh blood was collected from healthy volunteers. Peripheral blood mononuclear cells were isolated from normal donor blood via Ficoll (GE, 17144003). Primary cells were cultured in T-cell medium supplemented with Lonza X-VIVO 15 (Lonza, 04-418Q) with 10% FBS (Gibco/Life Technologies, 10 099-141) and 100 U/mL interleukin-2 (Proteintech, HZ-1015) (online supplemental figure S3A).

    Activation and transduction of T cells

    T cells were activated via microbeads coated with antihuman CD3 and CD28 antibodies (1:3; B&L Biological Technology, GMP-TL601). MSLN-CAR-T cells were generated via the lentiviral transduction of T cells. Briefly, activated T cells were inoculated into RetroNectin-coated culture plates and transduced with lentiviral supernatants (MOI=7) supplemented with polybrene (8 μg/mL; Beyotime, ST1380-10 g). The plates were subsequently centrifuged at 900×g at 32℃ for 90 min and incubated at 37℃ for 24 hours (online supplemental figure S3A). Finally, MSLN-CAR-T cells were cultured in normal medium. CAR expression on CAR-T cells was evaluated by flow cytometry.

    Flow cytometry

    The cells were collected via centrifugation and washed twice with phosphate-buffered saline (PBS). Next, the cells were stained, precipitated and washed according to the antibody instructions. Finally, the cells were incubated at 4℃ in PBS supplemented with 2% FBS for flow cytometry (BD, FACSAria II). The expression of MSLN on cancer cells was detected by using a human MSLN PE(phycoerythrin)-conjugated antibody (RD, FAB32652P). The expression of CAR on CAR-T cells was detected via EGFR-fluorescein isothiocyanate (FITC) (ICARTAB, 21092702) and the protein L (Pierce, 2997)-streptavidin APC (allophycocyanin, Invitrogen, S21374). The immunophenotypes of the T cells were tested via antibodies, including PE mouse antihuman CD3 (BioLegend, 300308), APC mouse antihuman CD4 (BioLegend, 300514), PerCP (Peridinin chlorophyll complex) /Cyanine5.5 mouse antihuman CD8 (BioLegend, 344710), FITC antihuman CD62L (BioLegend, 304838), PE antihuman CD45RO (BioLegend, 304206) and PE anti-Granzyme B (Miltenyi, 130-101-351) antibodies. The data were analyzed via FlowJo V.7.6.1.

    Lactate dehydrogenase assay

    A total of 2×104 TNBC cells were cultured in 96-well plates in media at different pH values. Then, the T cells and MSLN-CAR-T cells were collected and added to 96-well plates at 1:1, 1:5 and 1:10 ratios (target cell:effector cell) for 6 hours. Lactate dehydrogenase (LDH) levels were measured via a kit (Beyotime, C0017). Briefly, 10 μL of 10× lysis buffer was added for 1 hour, and the mixture was centrifuged at 400×g for 5 min. Afterward, 60 μL of the supernatant was transferred to a new plate, and 60 μL of LDH working solution (Beyotime, C0017) was added. The plate was incubated in the dark for 30 min, after which the optical density (OD) was detected at 490 nm and 620 nm (Thermo Fisher Scientific, VLB000D0). Four to five replicates were used for each sample. Cytotoxicity (%) was calculated from the OD of the sample and control.

    Intracellular pH measurement

    The intracellular pH of the cancer cells was measured by the pH-sensitive fluorescent probe 2',7'-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM; Beyotime, S1006) according to previous methods.89 Briefly, cancer cells were pretreated with medium at different pH values for 48 hours, after which 1×104 cells were seeded in 96-well microplates. Next, the cells were incubated in HBSS (Hank's Balanced Salt Solution) supplemented with 3 μM BCECF-AM at 37℃ for 30 min. After washing with PBS three times, the fluorescence was determined by using a microplate reader (Thermo Fisher Scientific, VLB000D0) with an emission wavelength of 535 nm and dual-excitation wavelengths of 490 nm and 430 nm. Four to five replicates were used for each sample. The intracellular pH was calculated according to the standard curve.

    Lysosomal activity assay

    The cells were collected by centrifugation and washed twice with PBS. Next, the cells were stained and incubated at 37℃ with 5% CO2 for 1 hour, precipitated and washed according to the antibody instructions (Abcam, ab234622). Finally, the cells were incubated at 4℃ in PBS supplemented with 2% FBS for flow cytometry (BD, FACSAria II) with an excitation filter at a wavelength of 488 nm. The data were analyzed via FlowJo V.7.6.1.

    NaHCO3 treatment

    MDA-MB-231 cells (1×106) were subcutaneously injected into the right flank of female NOD-SCID mice (SLAC ANIMAL). When the tumor volume (V=length×width×width/2) reached approximately 100 mm3, the tumor-bearing mice were randomly divided into three groups and treated with 0 mM, 100 mM or 200 mM NaHCO3 (Solarbio, S5240) orally for 40 days. The experiment was terminated when the tumor volume reached 2000 mm3. The size of the tumors was measured via Vernier calipers. MDA-MB-231 cells (1×106) were subcutaneously injected into the right flank of female NOG mice (CIEA, Charles River) at a 50:50 ratio of Matrigel (Biocoat, 356231) to PBS. Tumor-bearing mice were divided into two groups. One group was the control group, and the other group was orally administered 200 mM NaHCO3. Luciferin (2 μg/20 g; Promega, P1043) was intraperitoneally injected and used for imaging, and tumor treatment was evaluated via in vivo mouse imaging (BLT, AniView100).

    Intratumoral pH detected by a microelectrode

    Mice with tumors approximately 1 cm in length were selected and anesthetized by gas anesthesia. The tumor was located, the probe of the microelectrode (WPI, PH-OPTICA) was inserted 25 µm by a locator (RWD, 71000) multiple times and the pH of the tumor tissue was acquired. Approximately 20 measurements were performed to obtain the pH of the tumor center.

    In vivo killing effect of MSLN-CAR-T cells

    MDA-MB-231, HCC38, PANC-1, AsPC-1 and SK-OV3 cells expressing luciferase (1×106) were subcutaneously injected into the right flank of female NOG mice (CIEA, Charles River) at a 50:50 ratio of Matrigel (Biocoat, 356231) to PBS. NaHCO3 (200 mM) was added to the drinking water, and CAR-T cells (5×106) were injected via the tail vein 3 or 4 days after tumor engraftment. One day before treatment, the mice were imaged, distributed into groups of roughly equivalent tumor burdens, randomly assigned to treatment groups and kept blinded until the end of the study. Luciferin (2 μg/20 g; Promega, P1043) was intraperitoneally injected and used for imaging, and tumor treatment was evaluated via in vivo mouse imaging (BLT, AniView100). Survival data were recorded when the mice died in different treatment groups, and p values were determined via the log-rank (Mantel-Cox) test.

    In vivo CAR-T-cell persistence assay

    Peripheral blood was collected from the tail vein of tumor-bearing mice and placed into anticoagulant tubes. Then, the red blood cells were lysed for 15 min on ice, and the remaining cells were resuspended via centrifugation and stained with human CD3-FITC (BioLegend, 300454) according to the manufacturer’s instructions. Finally, the cells were incubated at 4℃ in PBS supplemented with 2% FBS for flow cytometry (BD, FACSAria II). The data were analyzed via FlowJo V.7.6.1.

    Statistical analysis

    Statistical analysis was performed via R software (V.4.0.2), SPSS statistical software (V.22.0) and GraphPad Prism statistical software (V.7.0). The R package (limma V.3.20.9) was used for analysis of differential gene expression profiles, and two-tailed t-tests and two-tailed independent-sample t-tests were used for analyzing differentially expressed genes in TNBC and non-TNBC samples. The quantitative data that conformed to a normal distribution are expressed as the mean±SD. P values were determined by two-tailed t-tests, and p<0.05 was considered to indicate statistical significance. Two-way analysis of variance was used to analyze the differential effects of CAR-T-cell therapy. P<0.05 was considered to indicate statistical significance. The log-rank (Mantel-Cox) test was used to analyze the survival of the model mice.

    Supplemental material

    Data availability statement

    Data are available on reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.

    Ethics statements

    Patient consent for publication

    Ethics approval

    The Institute Research Medical Ethics Committee of the Army Medical University granted approval for this study (Ethics Committee of the First Affiliated Hospital of Army Medical University, PLA [(A) KY202299)]. All the animal studies were performed in accordance with the National Institutes of Health (NIH) guidelines, with the approval of the Third Military Medical University on Laboratory Animal Care (Laboratory Animal Welfare and Ethics Committee of the Army Medical University, AMUWEC20232128).

    Acknowledgments

    We would like to thank Professor Yu Shi, Professor You-Hong Cui and Professor Liang Yi for their constructive suggestions.

    References

    1. ↵
      1. Goldhirsch A,
      2. Winer EP,
      3. Coates AS, et al
      . Personalizing the treatment of women with early breast cancer: highlights of the St Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2013. Ann Oncol 2013;24:2206–23. doi:10.1093/annonc/mdt303
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Sung H,
      2. Ferlay J,
      3. Siegel RL, et al
      . Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209–49. doi:10.3322/caac.21660
      OpenUrlCrossRefPubMed
    3. ↵
      1. Torre LA,
      2. Bray F,
      3. Siegel RL, et al
      . Global cancer statistics, 2012. CA A Cancer J Clinicians 2015;65:87–108. doi:10.3322/caac.21262
      OpenUrl
    4. ↵
      1. Malorni L,
      2. Shetty PB,
      3. De Angelis C, et al
      . Clinical and biologic features of triple-negative breast cancers in a large cohort of patients with long-term follow-up. Breast Cancer Res Treat 2012;136:795–804. doi:10.1007/s10549-012-2315-y
      OpenUrlCrossRefPubMed
    5. ↵
      1. Uscanga-Perales GI,
      2. Santuario-Facio SK,
      3. Ortiz-López R
      . Triple negative breast cancer: Deciphering the biology and heterogeneity. Med Univ 2016;18:105–14. doi:10.1016/j.rmu.2016.05.007
      OpenUrlCrossRef
    6. ↵
      1. Perou CM,
      2. Sørlie T,
      3. Eisen MB, et al
      . Molecular portraits of human breast tumours. Nature New Biol 2000;406:747–52. doi:10.1038/35021093
      OpenUrl
    7. ↵
      1. Lin NU,
      2. Claus E,
      3. Sohl J, et al
      . Sites of distant recurrence and clinical outcomes in patients with metastatic triple-negative breast cancer: high incidence of central nervous system metastases. Cancer 2008;113:2638–45. doi:10.1002/cncr.23930
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. van der Merwe PA,
      2. Dushek O
      . Mechanisms for T cell receptor triggering. Nat Rev Immunol 2011;11:47–55. doi:10.1038/nri2887
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Hassan R,
      2. Bera T,
      3. Pastan I
      . Mesothelin: a new target for immunotherapy. Clin Cancer Res 2004;10:3937–42. doi:10.1158/1078-0432.CCR-03-0801
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. von Minckwitz G,
      2. Schneeweiss A,
      3. Loibl S, et al
      . Neoadjuvant carboplatin in patients with triple-negative and HER2-positive early breast cancer (GeparSixto; GBG 66): a randomised phase 2 trial. Lancet Oncol 2014;15:747–56. doi:10.1016/S1470-2045(14)70160-3
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Fujisaka Y,
      2. Kurata T,
      3. Tanaka K, et al
      . Phase I study of amatuximab, a novel monoclonal antibody to mesothelin, in Japanese patients with advanced solid tumors. Invest New Drugs 2015;33:380–8. doi:10.1007/s10637-014-0196-0
      OpenUrlPubMed
    12. ↵
      1. Haas AR,
      2. Tanyi JL,
      3. O’Hara MH, et al
      . Phase I Study of Lentiviral-Transduced Chimeric Antigen Receptor-Modified T Cells Recognizing Mesothelin in Advanced Solid Cancers. Mol Ther 2019;27:1919–29. doi:10.1016/j.ymthe.2019.07.015
      OpenUrl
    13. ↵
      1. Hassan R,
      2. Bullock S,
      3. Premkumar A, et al
      . Phase I Study of SS1P, a Recombinant Anti-Mesothelin Immunotoxin Given as a Bolus I.V. Infusion to Patients with Mesothelin-Expressing Mesothelioma, Ovarian, and Pancreatic Cancers. Clin Cancer Res 2007;13:5144–9. doi:10.1158/1078-0432.CCR-07-0869
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Hassan R,
      2. Cohen SJ,
      3. Phillips M, et al
      . Phase I clinical trial of the chimeric anti-mesothelin monoclonal antibody MORAb-009 in patients with mesothelin-expressing cancers. Clin Cancer Res 2010;16:6132–8. doi:10.1158/1078-0432.CCR-10-2275
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Le DT,
      2. Wang-Gillam A,
      3. Picozzi V, et al
      . Safety and survival with GVAX pancreas prime and Listeria Monocytogenes-expressing mesothelin (CRS-207) boost vaccines for metastatic pancreatic cancer. J Clin Oncol 2015;33:1325–33. doi:10.1200/JCO.2014.57.4244
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Lv J,
      2. Li P
      . Mesothelin as a biomarker for targeted therapy. Biomark Res 2019;7:18. doi:10.1186/s40364-019-0169-8
    17. ↵
      1. Maruggi G,
      2. Zhang C,
      3. Li J, et al
      . mRNA as a Transformative Technology for Vaccine Development to Control Infectious Diseases. Mol Ther 2019;27:757–72. doi:10.1016/j.ymthe.2019.01.020
      OpenUrlCrossRef
    18. ↵
      1. Moskophidis D,
      2. Lechner F,
      3. Pircher H, et al
      . Virus persistence in acutely infected immunocompetent mice by exhaustion of antiviral cytotoxic effector T cells. Nature New Biol 1993;362:758–61. doi:10.1038/362758a0
      OpenUrl
    19. ↵
      1. Ferreras C,
      2. Fernández L,
      3. Clares-Villa L, et al
      . Facing CAR T Cell Challenges on the Deadliest Paediatric Brain Tumours. Cells 2021;10:2940. doi:10.3390/cells10112940
    20. ↵
      1. Mittal D,
      2. Gubin MM,
      3. Schreiber RD, et al
      . New insights into cancer immunoediting and its three component phases--elimination, equilibrium and escape. Curr Opin Immunol 2014;27:16–25. doi:10.1016/j.coi.2014.01.004
      OpenUrlCrossRefPubMed
    21. ↵
      1. Majzner RG,
      2. Mackall CL
      . Tumor Antigen Escape from CAR T-cell Therapy. Cancer Discov 2018;8:1219–26. doi:10.1158/2159-8290.CD-18-0442
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Estrella V,
      2. Chen T,
      3. Lloyd M, et al
      . Acidity generated by the tumor microenvironment drives local invasion. Cancer Res 2013;73:1524–35. doi:10.1158/0008-5472.CAN-12-2796
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Schornack PA,
      2. Gillies RJ
      . Contributions of cell metabolism and H+ diffusion to the acidic pH of tumors. Neoplasia 2003;5:135–45. doi:10.1016/s1476-5586(03)80005-2
      OpenUrlCrossRefPubMed
    24. ↵
      1. Helmlinger G,
      2. Yuan F,
      3. Dellian M, et al
      . Interstitial pH and pO2 gradients in solid tumors in vivo: high-resolution measurements reveal a lack of correlation. Nat Med 1997;3:177–82. doi:10.1038/nm0297-177
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Persi E,
      2. Duran-Frigola M,
      3. Damaghi M, et al
      . Systems analysis of intracellular pH vulnerabilities for cancer therapy. Nat Commun 2018;9:2997. doi:10.1038/s41467-018-05261-x
    26. ↵
      1. Koltai T
      . The Ph paradigm in cancer. Eur J Clin Nutr 2020;74:14–9. doi:10.1038/s41430-020-0684-6
      OpenUrl
    27. ↵
      1. Tang T,
      2. Yang Z-Y,
      3. Wang D, et al
      . The role of lysosomes in cancer development and progression. Cell Biosci 2020;10:131. doi:10.1186/s13578-020-00489-x
    28. ↵
      1. Hu M,
      2. Chen J,
      3. Liu S, et al
      . The Acid Gate in the Lysosome. Autophagy 2023;19:1368–70. doi:10.1080/15548627.2022.2125629
      OpenUrlCrossRef
    29. ↵
      1. Zhang Q,
      2. Liu G,
      3. Liu J, et al
      . The antitumor capacity of mesothelin-CAR-T cells in targeting solid tumors in mice. Mol Ther Oncolytics 2021;20:556–68. doi:10.1016/j.omto.2021.02.013
      OpenUrl
    30. ↵
      1. Zhai X,
      2. Mao L,
      3. Wu M, et al
      . Challenges of Anti-Mesothelin CAR-T-Cell Therapy. Cancers (Basel) 2023;15:1357. doi:10.3390/cancers15051357
      OpenUrl
    31. ↵
      1. Dieci MV,
      2. Criscitiello C,
      3. Goubar A, et al
      . Prognostic value of tumor-infiltrating lymphocytes on residual disease after primary chemotherapy for triple-negative breast cancer: a retrospective multicenter study. Ann Oncol 2015;26:1518. doi:10.1093/annonc/mdv241
      OpenUrlCrossRefPubMed
    32. ↵
      1. Reshkin SJ,
      2. Cardone RA,
      3. Harguindey S
      . Na+-H+ Exchanger, pH Regulation and Cancer. Recent Pat Anticancer Drug Discov 2012;8:85–99. doi:10.2174/15748928130108
      OpenUrl
    33. ↵
      1. Tarin D,
      2. Price JE
      . Influence of microenvironment and vascular anatomy on “metastatic” colonization potential of mammary tumors. Cancer Res 1981;41:3604–9.
      OpenUrlAbstract/FREE Full Text
    34. ↵
      1. Rusilowicz-Jones EV,
      2. Urbé S,
      3. Clague MJ
      . Protein degradation on the global scale. Mol Cell 2022;82:1414–23. doi:10.1016/j.molcel.2022.02.027
      OpenUrl
    35. ↵
      1. Mindell JA
      . Lysosomal acidification mechanisms. Annu Rev Physiol 2012;74:69–86. doi:10.1146/annurev-physiol-012110-142317
      OpenUrlCrossRefPubMedWeb of Science
    36. ↵
      1. Robey IF,
      2. Baggett BK,
      3. Kirkpatrick ND, et al
      . Bicarbonate increases tumor pH and inhibits spontaneous metastases. Cancer Res 2009;69:2260–8. doi:10.1158/0008-5472.CAN-07-5575
      OpenUrlAbstract/FREE Full Text
    37. ↵
      1. Abumanhal-Masarweh H,
      2. Koren L,
      3. Zinger A, et al
      . Sodium bicarbonate nanoparticles modulate the tumor pH and enhance the cellular uptake of doxorubicin. J Control Release 2019;296:1–13. doi:10.1016/j.jconrel.2019.01.004
      OpenUrl
    38. ↵
      1. Ando H,
      2. Ikeda A,
      3. Tagami M, et al
      . Oral administration of sodium bicarbonate can enhance the therapeutic outcome of Doxil® via neutralizing the acidic tumor microenvironment. J Control Release 2022;350:414–20. doi:10.1016/j.jconrel.2022.08.031
      OpenUrl
    39. ↵
      1. Yang M,
      2. Zhong X,
      3. Yuan Y
      . Does Baking Soda Function as a Magic Bullet for Patients With Cancer? A Mini Review. Integr Cancer Ther 2020;19:1534735420922579. doi:10.1177/1534735420922579
    40. ↵
      1. Stern LA,
      2. Jonsson VD,
      3. Priceman SJ
      . CAR T Cell Therapy Progress and Challenges for Solid Tumors. Cancer Treat Res 2020;180:297–326. doi:10.1007/978-3-030-38862-1_11
      OpenUrl
    41. ↵
      1. Jo Y,
      2. Ali LA,
      3. Shim JA, et al
      . Innovative CAR-T Cell Therapy for Solid Tumor; Current Duel between CAR-T Spear and Tumor Shield. Cancers (Basel) 2020;12:2087. doi:10.3390/cancers12082087
      OpenUrl
    42. ↵
      1. Hou AJ,
      2. Chen LC,
      3. Chen YY
      . Navigating CAR-T cells through the solid-tumour microenvironment. Nat Rev Drug Discov 2021;20:531–50. doi:10.1038/s41573-021-00189-2
      OpenUrlCrossRef
    43. ↵
      1. Xie Y,
      2. Hu Y,
      3. Zhou N, et al
      . CAR T-cell therapy for triple-negative breast cancer: Where we are. Cancer Lett 2020;491:121–31. doi:10.1016/j.canlet.2020.07.044
      OpenUrlCrossRef
    44. ↵
      1. Dentro SC,
      2. Leshchiner I,
      3. Haase K, et al
      . Characterizing genetic intra-tumor heterogeneity across 2,658 human cancer genomes. Cell 2021;184:2239–54. doi:10.1016/j.cell.2021.03.009
      OpenUrlCrossRefPubMed
    45. ↵
      1. de Sousa VML,
      2. Carvalho L
      . Heterogeneity in Lung Cancer. Pathobiology 2018;85:96–107. doi:10.1159/000487440
      OpenUrl
    46. ↵
      1. Hassan R,
      2. Thomas A,
      3. Alewine C, et al
      . Mesothelin Immunotherapy for Cancer: Ready for Prime Time? J Clin Oncol 2016;34:4171–9. doi:10.1200/JCO.2016.68.3672
      OpenUrl
    47. ↵
      1. Roulot A,
      2. Héquet D,
      3. Guinebretière J-M, et al
      . Tumoral heterogeneity of breast cancer. Ann Biol Clin (Paris) 2016;74:653–60. doi:10.1684/abc.2016.1192
      OpenUrlCrossRef
    48. ↵
      1. Marusyk A,
      2. Polyak K
      . Tumor heterogeneity: causes and consequences. Biochim Biophys Acta 2010;1805:105–17. doi:10.1016/j.bbcan.2009.11.002
      OpenUrlCrossRefPubMed
    49. ↵
      1. Zhu Y,
      2. Zhang Q,
      3. Wei C, et al
      . High-Precision Quantitative Analysis Reveals Carcinoembryonic Protein Expression Differs Among Colorectal Cancer Primary Foci and Metastases to Different Sites. Technol Cancer Res Treat 2021;20:15330338211037175. doi:10.1177/15330338211037175
    50. ↵
      1. Sutnar A,
      2. Pesta M,
      3. Liska V, et al
      . Clinical relevance of the expression of mRNA of MMP-7, MMP-9, TIMP-1, TIMP-2 and CEA tissue samples from colorectal liver metastases. Tumour Biol 2007;28:247–52. doi:10.1159/000110897
      OpenUrl
    51. ↵
      1. Ndacayisaba LJ,
      2. Rappard KE,
      3. Shishido SN, et al
      . Characterization of BCMA Expression in Circulating Rare Single Cells of Patients with Plasma Cell Neoplasms. Int J Mol Sci 2022;23:13427. doi:10.3390/ijms232113427
    52. ↵
      1. Fu XH,
      2. Wang Y,
      3. Wang HJ, et al
      . CD19 antigen loss after treatment of Bispecific T-cell Engager and effective response to salvage bispecific CAR-T therapy in B cell acute lymphoblastic leukemia: a case report and literature review. Zhonghua Xue Ye Xue Za Zhi 2020;41:282–6. doi:10.3760/cma.j.issn.0253-2727.2020.04.004
      OpenUrl
    53. ↵
      1. Weidemann S,
      2. Gagelmann P,
      3. Gorbokon N, et al
      . Mesothelin Expression in Human Tumors: A Tissue Microarray Study on 12,679 Tumors. Biomedicines 2021;9:397. doi:10.3390/biomedicines9040397
      OpenUrl
    54. ↵
      1. O’Rourke DM,
      2. Nasrallah MP,
      3. Desai A, et al
      . A single dose of peripherally infused EGFRvIII-directed CAR T cells mediates antigen loss and induces adaptive resistance in patients with recurrent glioblastoma. Sci Transl Med 2017;9:eaaa0984. doi:10.1126/scitranslmed.aaa0984
    55. ↵
      1. Beatty GL,
      2. Haas AR,
      3. Maus MV, et al
      . Mesothelin-specific chimeric antigen receptor mRNA-engineered T cells induce anti-tumor activity in solid malignancies. Cancer Immunol Res 2014;2:112–20. doi:10.1158/2326-6066.CIR-13-0170
      OpenUrlAbstract/FREE Full Text
    56. ↵
      1. Morgan RA,
      2. Chinnasamy N,
      3. Abate-Daga D, et al
      . Cancer regression and neurological toxicity following anti-MAGE-A3 TCR gene therapy. J Immunother 2013;36:133–51. doi:10.1097/CJI.0b013e3182829903
      OpenUrlCrossRefPubMedWeb of Science
    57. ↵
      1. Arneth B
      . Tumor Microenvironment. Medicina (Kaunas) 2019;56:15. doi:10.3390/medicina56010015
    58. ↵
      1. Du T,
      2. Gao J,
      3. Li P, et al
      . Pyroptosis, metabolism, and tumor immune microenvironment. Clin Transl Med 2021;11:e492. doi:10.1002/ctm2.492
    59. ↵
      1. Quail DF,
      2. Joyce JA
      . Microenvironmental regulation of tumor progression and metastasis. Nat Med 2013;19:1423–37. doi:10.1038/nm.3394
      OpenUrlCrossRefPubMed
    60. ↵
      1. Vitale I,
      2. Manic G,
      3. Coussens LM, et al
      . Macrophages and Metabolism in the Tumor Microenvironment. Cell Metab 2019;30:36–50. doi:10.1016/j.cmet.2019.06.001
      OpenUrlCrossRef
    61. ↵
      1. Aredia F,
      2. Scovassi AI
      . Manipulation of Intracellular pH in Cancer Cells by NHE1 Inhibitors. Protein Pept Lett 2016;23:1123–9. doi:10.2174/0929866523666161013125536
      OpenUrl
    62. ↵
      1. Ibrahim Hashim A,
      2. Cornnell HH,
      3. Coelho Ribeiro M de L, et al
      . Reduction of metastasis using a non-volatile buffer. Clin Exp Metastasis 2011;28:841–9. doi:10.1007/s10585-011-9415-7
      OpenUrlCrossRefPubMedWeb of Science
    63. ↵
      1. Silva AS,
      2. Yunes JA,
      3. Gillies RJ, et al
      . The potential role of systemic buffers in reducing intratumoral extracellular pH and acid-mediated invasion. Cancer Res 2009;69:2677–84. doi:10.1158/0008-5472.CAN-08-2394
      OpenUrlAbstract/FREE Full Text
    64. ↵
      1. Berndt H
      . Epidemiology and clinical aspects of primary liver tumors. Z Arztl Fortbild (Jena) 1991;85:1135–8.
      OpenUrlPubMed
    65. ↵
      1. Lawrence RE,
      2. Zoncu R
      . The lysosome as a cellular centre for signalling, metabolism and quality control. Nat Cell Biol 2019;21:133–42. doi:10.1038/s41556-018-0244-7
      OpenUrlCrossRefPubMed
    66. ↵
      1. Zhang Z,
      2. Yue P,
      3. Lu T, et al
      . Role of lysosomes in physiological activities, diseases, and therapy. J Hematol Oncol 2021;14:79. doi:10.1186/s13045-021-01087-1
      OpenUrlCrossRef
    67. ↵
      1. Lamming DW,
      2. Bar-Peled LL
      . Lysosome: The metabolic signaling hub. Traffic 2019;20:27–38. doi:10.1111/tra.12617
      OpenUrlCrossRefPubMed
    68. ↵
      1. O’Neill KI,
      2. Kuo L-W,
      3. Williams MM, et al
      . NPC1 Confers Metabolic Flexibility in Triple Negative Breast Cancer. Cancers (Basel) 2022;14:3543. doi:10.3390/cancers14143543
      OpenUrl
    69. ↵
      1. Sampayo RG,
      2. Toscani AM,
      3. Rubashkin MG, et al
      . Fibronectin rescues estrogen receptor α from lysosomal degradation in breast cancer cells. J Cell Biol 2018;217:2777–98. doi:10.1083/jcb.201703037
      OpenUrlAbstract/FREE Full Text
    70. ↵
      1. Kocak M,
      2. Ezazi Erdi S,
      3. Jorba G, et al
      . Targeting autophagy in disease: established and new strategies. Autophagy 2022;18:473–95. doi:10.1080/15548627.2021.1936359
      OpenUrlCrossRefPubMed
    71. ↵
      1. Yamamoto K,
      2. Venida A,
      3. Yano J, et al
      . Autophagy promotes immune evasion of pancreatic cancer by degrading MHC-I. Nature New Biol 2020;581:100–5. doi:10.1038/s41586-020-2229-5
      OpenUrl
    72. ↵
      1. Hawkins ER,
      2. D’Souza RR,
      3. Klampatsa A
      . Armored CAR T-Cells: The Next Chapter in T-Cell Cancer Immunotherapy. Biologics 2021;15:95–105. doi:10.2147/BTT.S291768
      OpenUrlCrossRefPubMed
    73. ↵
      1. Wang JX,
      2. Choi SYC,
      3. Niu X, et al
      . Lactic Acid and an Acidic Tumor Microenvironment suppress Anticancer Immunity. Int J Mol Sci 2020;21:8363. doi:10.3390/ijms21218363
    74. ↵
      1. Perrot I,
      2. Michaud H-A,
      3. Giraudon-Paoli M, et al
      . Blocking Antibodies Targeting the CD39/CD73 Immunosuppressive Pathway Unleash Immune Responses in Combination Cancer Therapies. Cell Rep 2019;27:2411–25. doi:10.1016/j.celrep.2019.04.091
      OpenUrlCrossRef
    75. ↵
      1. Narayan V,
      2. Barber-Rotenberg JS,
      3. Jung I-Y, et al
      . PSMA-targeting TGFβ-insensitive armored CAR T cells in metastatic castration-resistant prostate cancer: a phase 1 trial. Nat Med 2022;28:724–34. doi:10.1038/s41591-022-01726-1
      OpenUrlCrossRef
    76. ↵
      1. Riley RS,
      2. June CH,
      3. Langer R, et al
      . Delivery technologies for cancer immunotherapy. Nat Rev Drug Discov 2019;18:175–96. doi:10.1038/s41573-018-0006-z
      OpenUrlCrossRef
    77. ↵
      1. Qi F-L,
      2. Wang M-F,
      3. Li B-Z, et al
      . Reversal of the immunosuppressive tumor microenvironment by nanoparticle-based activation of immune-associated cells. Acta Pharmacol Sin 2020;41:895–901. doi:10.1038/s41401-020-0423-5
      OpenUrl
    78. ↵
      1. MacGlashan D
      . Blocking antibodies in immunotherapy: Quality versus quantity. J Allergy Clin Immunol 2019;144:1177–9. doi:10.1016/j.jaci.2019.09.024
      OpenUrl
    79. ↵
      1. Wherry EJ
      . T cell exhaustion. Nat Immunol 2011;12:492–9. doi:10.1038/ni.2035
      OpenUrlCrossRefPubMed
    80. ↵
      1. Gumber D,
      2. Wang LD
      . Improving CAR-T immunotherapy: Overcoming the challenges of T cell exhaustion. EBioMedicine 2022;77:103941. doi:10.1016/j.ebiom.2022.103941
    81. ↵
      1. Zhang X,
      2. Zhang C,
      3. Qiao M, et al
      . Depletion of BATF in CAR-T cells enhances antitumor activity by inducing resistance against exhaustion and formation of central memory cells. Cancer Cell 2022;40:1407–22. doi:10.1016/j.ccell.2022.09.013
      OpenUrlCrossRef
    82. ↵
      1. Funk CR,
      2. Wang S,
      3. Chen KZ, et al
      . PI3Kδ/γ inhibition promotes human CART cell epigenetic and metabolic reprogramming to enhance antitumor cytotoxicity. Blood 2022;139:523–37. doi:10.1182/blood.2021011597
      OpenUrl
    83. ↵
      1. Si J,
      2. Shi X,
      3. Sun S, et al
      . Hematopoietic Progenitor Kinase1 (HPK1) Mediates T Cell Dysfunction and Is a Druggable Target for T Cell-Based Immunotherapies. Cancer Cell 2020;38:551–66. doi:10.1016/j.ccell.2020.08.001
      OpenUrlCrossRefPubMed
    84. ↵
      1. Alizadeh D,
      2. Wong RA,
      3. Yang X, et al
      . IL15 Enhances CAR-T Cell Antitumor Activity by Reducing mTORC1 Activity and Preserving Their Stem Cell Memory Phenotype. Cancer Immunol Res 2019;7:759–72. doi:10.1158/2326-6066.CIR-18-0466
      OpenUrlAbstract/FREE Full Text
    85. ↵
      1. Yuan L,
      2. Tatineni J,
      3. Mahoney KM, et al
      . VISTA: A Mediator of Quiescence and a Promising Target in Cancer Immunotherapy. Trends Immunol 2021;42:209–27. doi:10.1016/j.it.2020.12.008
      OpenUrlCrossRefPubMed
    86. ↵
      1. Mittendorf EA,
      2. Philips AV,
      3. Meric-Bernstam F, et al
      . PD-L1 Expression in Triple-Negative Breast Cancer. Cancer Immunol Res 2014;2:361–70. doi:10.1158/2326-6066.CIR-13-0127
      OpenUrlAbstract/FREE Full Text
    87. ↵
      1. Wu SZ,
      2. Al-Eryani G,
      3. Roden DL, et al
      . A single-cell and spatially resolved atlas of human breast cancers. Nat Genet 2021;53:1334–47. doi:10.1038/s41588-021-00911-1
      OpenUrlCrossRefPubMed
    88. ↵
      1. Wang J,
      2. Xu S-L,
      3. Duan J-J, et al
      . Invasion of white matter tracts by glioma stem cells is regulated by a NOTCH1-SOX2 positive-feedback loop. Nat Neurosci 2019;22:91–105. doi:10.1038/s41593-018-0285-z
      OpenUrlCrossRefPubMed
    89. ↵
      1. Yuan K,
      2. Mei J,
      3. Shao D, et al
      . Cerium Oxide Nanoparticles Regulate Osteoclast Differentiation Bidirectionally by Modulating the Cellular Production of Reactive Oxygen Species. Int J Nanomedicine 2020;15:6355–72. doi:10.2147/IJN.S257741
      OpenUrlCrossRefPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • MW, LM, XZ and JL contributed equally.

    • Contributors MW, LM, XZ and JL performed the experiments, analyzed the data, and wrote and reviewed the manuscript. MW, LM and JL performed the in vitro experiments and analyzed the data. MW and XZ designed and performed the animal surgeries and in vivo experiments and analyzed the data. Junhang Wang and LL performed the cell culture experiments and analyzed the data. SL performed the animal tissue experiments. JD and Jun Wang discussed the data and reviewed the manuscript. SY designed the study, discussed the data, and wrote and reviewed the manuscript. JL discussed the data and wrote and reviewed the manuscript. SY, Jun Wang and JD obtained the grants. All the authors provided their consent for publication. SY is the guarantor of the study.

    • Funding This study was supported by grants from the National Key Research and Development Program of China (2022YFA1205003 to Shicang Yu), Major Projects of the National Natural Science Foundation of China (92059204 to Shicang Yu), General Projects of the National Natural Science Foundation of China (81903009 to Jun Wang) and Major Projects of Technological Innovation and Application Development Foundation in Chongqing (CSTB2022TIAD-STX0012 to Jiangjie Duan).

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.