Mitochondrial metabolic reprogramming of macrophages and T cells enhances CD47 antibody-engineered oncolytic virus antitumor immunity

Abstract

Background

Although immunotherapy can reinvigorate immune cells to clear tumors, the response rates are poor in some patients. Here, CD47 antibody-engineered oncolytic viruses (oAd-αCD47) were employed to lyse tumors and activate immunity. The oAd-αCD47 induced comprehensive remodeling of the tumor microenvironment (TME). However, whether the acidic TME affects the antitumor immunotherapeutic effects of oncolytic viruses-αCD47 has not been clarified.

Methods

To assess the impact of oAd-αCD47 treatment on the TME, we employed multicolor flow cytometry. Glucose uptake was quantified using 2NBDG, while mitochondrial content was evaluated with MitoTracker FM dye. pH imaging of tumors was performed using the pH-sensitive fluorophore SNARF-4F. Moreover, changes in the calmodulin-dependent protein kinase II (CaMKII)/cyclic AMP activates-responsive element-binding proteins (CREB) and peroxisome proliferator-activated receptor gamma coactivator-1α (PGC1α) signaling pathway were confirmed through western blotting and flow cytometry.

Results

Here, we identified sodium bicarbonate (NaBi) as the potent metabolic reprogramming agent that enhanced antitumor responses in the acidic TME. The combination of NaBi and oAd-αCD47 therapy significantly inhibited tumor growth and produced complete immune control in various tumor-bearing mouse models. Mechanistically, combination therapy mainly reduced the number of regulatory T cells and enriched the ratio of M1-type macrophages TAMs (M1.TAMs) to M2-type macrophages TAMs (M2.TAMs), while decreasing the abundance of PD-1⁺TIM3⁺ expression and increasing the expression of CD107a in the CD8⁺ T cells. Furthermore, the combination therapy enhanced the metabolic function of T cells and macrophages by upregulating PGC1α, a key regulator of mitochondrial biogenesis. This metabolic improvement contributed to a robust antitumor response. Notably, the combination therapy also promoted the generation of memory T cells, suggesting its potential as an effective neoadjuvant treatment for preventing postoperative tumor recurrence and metastasis.

Conclusions

Tumor acidic microenvironment impairs mitochondrial energy metabolism in macrophages and T cells inducing oAd-αCD47 immunotherapeutic resistance. NaBi improves the acidity of the TME and activates the CaMKII/CREB/PGC1α mitochondrial biosynthesis signaling pathway, which reprograms the energy metabolism of macrophages and T cells in the TME, and oral NaBi enhances the antitumor effect of oAd-αCD47.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Oncolytic virus immunotherapy can lyse tumors and activate immunity to clear tumors, but the response rates are poor in some patients. The oncolytic adenovirus (oAd)-αCD47 induced comprehensive remodeling of the tumor microenvironment (TME). Acidic TME leads to immunoresistance and treatment resistance. However, whether acidic TME interferes with the antitumor efficacy and mechanism of oncolytic viruses-αCD47 remains to be investigated.

WHAT THIS STUDY ADDS

• Tumor acidic microenvironment impairs mitochondrial energy metabolism in macrophages and T cells and induces oAd-αCD47 immunotherapeutic resistance; sodium bicarbonate (NaBi) improves the acidity of the TME and activates the calmodulin-dependent protein kinase II /cyclic AMP activates-responsive element-binding proteins/peroxisome proliferator-activated receptor gamma coactivator-1α mitochondrial biosynthesis signaling pathway, thereby reprogramming the energy metabolism of macrophages and T cells in the TME, and oral NaBi enhances the antitumor effect of oAd-αCD47.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Oncolytic viruses have potent immune-stimulatory potential, sufficient metabolic support of immune cells is key for a sustained complete antitumor immune response and promotes long-term memory effects. NaBi oral therapy can provide sufficient mitochondrial energy metabolism for antitumor immune cells, providing theoretical support for combination therapy with multiple immunotherapies. Our study presents a novel and effective combination strategy for immunotherapy, with good cancer treatment effects.

Introduction

Oncolytic virotherapy is an effective tumor immunotherapy that enhances antitumor immune response by lysing tumor cells and spreading tumor-associated epitopes.¹ To date, only four oncolytic viruses (OVs) have been approved for marketing, including ECHO-7 enteroviruses (Rigvir) for the treatment of various malignant tumors, recombinant human adenovirus type 5 (H101) for head and neck tumors, human granulocyte-macrophage colony-stimulating factor modified second-generation herpes simplex viruses (Imlygic) for advanced melanoma, and third-generation herpes simplex viruses (Delytact) for primary brain cancer.² Imlygic was the first oncolytic agent approved by the US Food and Drug Administration, with a response rate of only 16% in patients with melanoma.³ This is partly because OVs are quickly cleared by viral antibodies, reducing the quantity of viruses to reach tumor lesions.⁴ Furthermore, OVs as a form of cancer immunotherapy can stimulate a T-cell antitumor immune response. Therefore, similar to other immunotherapies, its antitumor effects can be suppressed by the tumor immune microenvironment (TIME).

TIME is the basis of tumor immune escape, immune tolerance, and acquired drug resistance of immunotherapy, as well as the key factor leading to oncolytic virus drug resistance.⁵ The presence of vast quantities of immunosuppressive cells in the TIME, such as tumor-associated macrophages (TAMs), regulatory T cells (Tregs), and myeloid-derived immunosuppressive cells (MDSCs), leads to antitumor T-cell tolerance. TAMs are the key component of the TIME which exerts an immunosuppressive role that promotes tumor growth.⁶ However, TAMs can also exert antitumor functions due to their phenotypic plasticity.⁶ TAMs are divided into M1-type macrophages (M1.TAMs) that have pro-inflammatory and antitumor functions, and M2-type macrophages (M2.TAMs) that induce immunosuppression and promote tumor progression.⁷ Therefore, TAM-targeting therapy is an attractive target for cancer immunotherapy.

The “do not eat me” signaling protein CD47 is highly expressed on the surface of most tumor cells. CD47 binds to signal regulatory protein α (SIRPα)—a signal regulatory protein on macrophages—directly silencing the phagocytic signal of macrophages and inhibiting the antigen presentation function of macrophages, leading to tumor evasion from immune surveillance.8,10 Blocking the CD47-SIRPα axis promotes the phagocytosis of TAMs in tumor cells and induces macrophage transition to M1.TAMs to further enhance antitumor immunity, which is an effective antitumor immunotherapy strategy.^{8 11} Several drugs targeting CD47 signaling are currently undergoing clinical trials. However, because the CD47 molecule is also highly expressed in normal erythrocytes, monoclonal antibody therapy targeting CD47 can induce associated hemolysis and cause serious adverse effects.^{12 13} Previous studies have demonstrated that OVs encoding CD47 antibodies (OVs-αCD47) have a significant antitumor activity.14,16 OVs-αCD47 can block CD47 signaling and enhance phagocytosis of macrophages in vitro, significantly avoiding the hemolytic side effects caused by CD47 monoclonal antibodies.¹⁶ While OVs-αCD47 effectively induces macrophage-mediated phagocytosis of tumor cells in vitro, the potential inhibitory effects of the tumor microenvironment (TME) on CD47 antibody-mediated phagocytosis remain to be elucidated. Recent studies have shown that abnormal metabolism of tumor cells represents a key mechanism of antitumor immune function suppression. Tumor cells deplete essential nutrients in the TME and produce numerous toxic metabolic by-products that affect immune cell metabolism.¹⁷ Most importantly, abnormal immune cell metabolism eventually leads to antitumor immunosuppression or immune tolerance, thereby resisting antitumor immunotherapy.^{17 18} For example, the production of high concentrations of lactate (LA) by tumor cells can induce the formation of an acidic TME in solid tumors.¹⁹ The acidic TME inhibits the antitumor function of immune cells and promotes tumor progression, metastasis, and drug resistance.^{20 21} Tumor-infiltrating T cells (TILs) display significant metabolic deficits in an acidic environment.^{22 23} Similarly, macrophages in an acidic environment are shaped into M2.TAMs, thereby producing more arginase (Arg1) to suppress antitumor immune responses.^{24 25} Moreover, monocytes tend to differentiate into metabolism-deficient macrophages in an acidic environment, impairing macrophage functions.²⁶ These findings suggest that the accumulation of LA can inhibit the function of antitumor immune cells in the TME. Thus, whether a strongly acidic TME interferes with the antitumor immunotherapeutic effects of OVs-αCD47 remains unknown.

In addition, we explored the antitumor effect of oncolytic adenovirus expressing CD47 high-affinity blocking nanobody (oAd-αCD47) and its impact on the TIME. It was found that LA-induced acidic microenvironment can interfere with the antitumor immune response of oAd-αCD47, mainly by reducing the mitochondrial content of CD8⁺ T cells and TAMs, thereby inhibiting the killing function of CD8⁺ T cells and the phagocytic function of TAMs. We proposed that sodium bicarbonate (NaBi) can act as an immunomodulator to reprogram metabolic disorders of CD8⁺ T cells and TAMs in an acidic environment. NaBi enhanced the antitumor effect of oAd-αCD47 in various tumor-bearing models. In addition, we investigated the alterations in the immune microenvironment and the metabolic profiles of macrophages and CD8+T cells following the combination therapy. We further explored the potential mechanisms through which NaBi modulates the metabolism of immune cells. Our findings highlight a novel role for NaBi in regulating the metabolism of antitumor immune cells and offer valuable insights for the clinical application of OVs in combination therapies.

Materials and methods

Cells culture and oncolytic virus

A murine colon carcinoma cell line MC38, melanoma cell line B16-F10, and breast cancer cell line 4T1 were obtained from the American Type Culture Collection (USA). B16-F10-Luc and 4T1-Luc were provided by our laboratory. All cells were cultured in the Roswell Park Memorial Institute 1640 medium (Gibco, New York, USA) supplemented with 10% (v/v) fetal bovine serum (FBS, Gibco). All cell lines were maintained at 37°C in a humidified atmosphere containing 5% CO₂. Both the OVs (oAd-NC) and the CD47-engineered OVs (oAd-αCD47) were constructed in our laboratory.¹⁶

Animal experiments

Six-week-old female BALB/c mice and C57BL/6 J mice were purchased from Huafukang Biology Technology (Beijing, China). All mice live in a sterile environment. Animal experiments were approved by the Animal Protection and Utilization Committee of Sichuan University (Approval No.20240104005).

For subcutaneous tumor models, 1×10⁶ MC38, B16-F10-luc, or 4T1-luc cells were subcutaneously injected into the right flank of the mice. Mice started oral NaBi 200 mM/mice 3 days before tumor inoculation until the end of the experiment. When the tumor size reached 100–150 mm³, the mice received intratumoral treatment. The MC38 tumor model received a single dose of 5×10⁸ plaque-forming units (pfu) of oAd-NC or oAd-αCD47, whereas B16-F10 and 4T1 tumor models received 2.5×10⁸ pfu of oAd-αCD47. The tumor volume and the lifetime of tumor-bearing mice were recorded. The final tumor size and pulmonary metastasis were measured 6 min after the intraperitoneal injection of 150 mg/kg body weight of D-luciferin (Promega). The bioluminescence images were obtained using an In Vivo Imaging System (IVIS, PerkinElmer).

To deplete TAMs, pathological activated neutrophil-myeloid-derived suppressor cells MDSCs (PMN-MDSCs), MDSCs, and CD8⁺ T cells, clophosome (FormuMax, F70101C-AC), anti-Ly6G mAbs (Bio X Cell, clone 1A8), and anti-CD8 mAbs (Bio X Cell, clone 53–6.7) were administered intraperitoneally at a dose of 200 µg/mouse every 4 days for three consecutive treatments, beginning 4 days post-tumor inoculation. The depletion efficacy was confirmed by flow cytometric analysis of macrophage, neutrophil, and CD8⁺ T-cell populations in peripheral blood.

For tumor rechallenge experiments, mice were inoculated with tumor cells twice the initial dose. For B16-F10 and 4T1 rechallenge experiments, we performed bilateral tumor rechallenge tests in which MC38 or CT26 tumor cells were inoculated on the right side and B16-F10 or 4T1 tumor cells on the left side. The tumor volume and the lifetime of tumor-bearing mice were recorded.

To study tumor recurrence and metastasis, mice were subcutaneously injected with 1×10⁶ B16-F10-luc cells and subsequently administered NaBi or normal drinking water. Following oAd-αCD47 treatment and incomplete or complete surgical resection of tumor foci, mouse models of tumor recurrence and metastasis were established. The NaBi and water-drinking groups were further divided into two subgroups: one group continued to receive 200 mM NaBi, while the other received normal water post-surgery. Tumor growth curves, mouse survival times, and lung metastasis were monitored to assess tumor recurrence and metastasis.

Tumor diameters were measured using Vernier calipers, and the size was evaluated using the following equation: Tumor volume (mm³)=(length×width² /2.

Metabolic analysis

Changes in mitochondrial morphology and content of T cells and macrophages were analyzed by Transmission Electron Microscopy and MitoTracker FM dye kit. Specifically, 2NBDG (Cayman Chemical) and MitoTracker FM dye (Thermo Fisher) were used to measure glucose uptake and mitochondrial content. Immune cells were stained on the cell surface and incubated at 37°C for 30 min with 20 µM 2-NBDG and MitoTracker FM in a medium containing 1% FBS for flow cytometric analysis.

Flow cytometric analysis

Fresh tumor tissues were collected and digested to prepare a single-cell suspension. Dead cells in the single-cell suspension were excluded using a Zombie Red Fixable Viability Kit (BioLegend, 1:2,000) and then blocked by an Fc blocker for antibody staining.

T lymphocytes: CD3⁺CD4⁺CD8⁺PD-1⁺ TIM3⁺CD69⁺CD107a⁺; Tregs/Teffs: CD3⁺CD4⁺CD8⁺CD25⁺FoxP3⁺/FoxP3⁻; TAMs: CD45⁺CD11b⁺Gr1⁻F4/80⁺CD206⁺Arg1⁺; PMN.MDSCs/monocyte-derived MDSCs (Mo.MDSCs): CD45⁺CD11b⁺ Ly6G⁺/Ly6C⁺ were for tumor immune cell composition analysis. The intracellular protein Arg1 or nuclear proteins, includingregulatory T cells (Tregs), peroxisome proliferator-activated receptor gamma coactivator-1α (PGC1α), CREB, and phosphorylated form of CREB (pCREB), were treated using a Fixation/Permeabilization Kit (BD Biosciences) or a FoxP3 Fixation and Permeabilization Kit (eBioscience), respectively, and then stained 30 min for fluorescence-activated cell sorting (FACS) analysis.

For intracellular cytokine analysis, red blood cells from single-cell suspensions of the spleen and blood were excluded using a Red Blood Cell Lysis Buffer (BioLegend). Lymphocytes of all samples were stimulated for 2 hours with phorbol 12-myristate 13-acetic acid (PMA) and ionomycin for surface molecular staining, and then treated with the Fixation/Permeabilization Kit to detect interferon-gamma (IFN-γ) and tumor necrosis factor-alpha (TNF-α). All fluorescently labeled antibodies used (antibody dosage, 1:200) for flow cytometric analysis were obtained from BioLegend.

Immunofluorescence and flow cytometry analysis of phagocytosis in vivo

Tumor tissues were embedded and 5 µm sections were prepared. Sections were fixed, blocked, and incubated overnight at 4°C using the following primary antibodies: rat anti-F4/80 (1:200, R&D Systems) and rabbit anti-CK19 (1:100, Abcam). Sections were subsequently stained with fluorescent secondary antibodies of different colors at 25°C for 1 hour. The nucleus was stained with 4′,6-diamidino-2-phenylindole for 5 min and then mounted in SlowFade Gold. To assess macrophage phagocytosis by flow cytometry, single-cell suspensions derived from tumor tissue were stained with CD45 and F4/80 fluorescent antibodies for 30 min. Subsequently, the CK19 antibody was applied for 1 hour at room temperature, followed by a 1-hour incubation with a fluorescent secondary antibody. Macrophage phagocytosis was determined by evaluating the expression of CK19 within the CD45⁺F4/80⁺ population.

Intratumoral pH measurement

A pH-sensitive fluorophore SNARF-4F (Invitrogen, Carlsbad, California, USA) was exploited for pH imaging of MC38 tumors. MC38 tumor-bearing mice were intravenously injected with SNARF-4F (2 nmol per mouse in 100 µL of sterile saline). Mice were euthanized after 20 min. Tumor and spleen tissues were harvested and imaged using IVIS.

Analysis of LA concentration in the cell culture supernatant and tumor tissues

The concentration of LA in cell culture supernatants and tumor tissues was measured on the automated computerized SpectraMax iD3 (Molecular Devices). Briefly, 1 mL of cell culture suspension and tumor tissue suspensions were centrifuged at 600 g for 5 min, then the supernatant was collected and centrifuged at 900 g for 5 min and stored on ice until analysis. Analysis was performed within 5 hours of sample collection using the Lactate Assay Kit (Sigma-Aldrich).

Measurement of intracellular calcium concentration

Cells were loaded with 2 µM Fluo-4-AM (Invitrogen, Carlsbad, California, USA) at 37°C for 30 min. Cells were washed twice with phosphate-buffered saline (PBS) containing 2% FBS, and the fluorescence intensity of Fluo-4-AM was analyzed by FACS.

Western blot analysis

The proteins of T cells and bone marrow-derived macrophage (BMDM) were extracted and lysed using 1×radioimmunoprecipitation assay lysis solution containing a cocktail of protease inhibitors (Thermo Scientific). The proteins were separated on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to a polyvinylidene fluoride membrane (Millipore). The membrane was incubated at 4°C overnight with the following primary antibodies: rabbit anti-CREB, anti-pCREB, anti-calmodulin-dependent protein kinase II (CaMKII), anti-phosphorylated form of CaMKII (pCaMKII) (all 1:1,000; Abcam, UK), and rabbit anti-β-actin (1:5,000; Abcam), followed by incubation with a horseradish peroxidase-conjugated goat anti-rabbit IgG secondary antibody (1:10,000; Santa Cruz Biotechnology) at 25°C for 1 hour. The membrane was analyzed using an ECL Prime Western Blotting kit (GE HealthCare) and VersaDoc 4000 MP (Bio-Rad Laboratories).

Statistical analysis

Statistical analysis was performed using GraphPad Prism V.8.0 and SPSS. The Student’s t-test was used to compare two means with assuming equal variance. Continuous data were evaluated using analysis of variance models. Animal survival was simulated using Kaplan-Meier survival curves and analyzed using a log-rank (Mantel-Cox) test. A p value of <0.05 was considered statistically significant. All experiments were repeated two to three times.

Results

oAd-αCD47 enhances CD8⁺ T cell and macrophage antitumor activity but fails to produce sustained antitumor effects

To investigate the antitumor effect and immunological mechanism of oAd-αCD47, MC38 tumor-bearing mice were used, which are sensitive to immunotherapy with higher infiltration of TAMs.^{27 28} Single-dose intratumoral injection of oAd-αCD47 significantly inhibited tumor growth compared with PBS and oAd-NC (figure 1A,B). Despite initial tumor regression, tumor growth resumed following discontinuation of oAd-αCD47 therapy, indicating that oAd-αCD47 treatment may not achieve complete tumor eradication (figure 1A). Modification of the TME is crucial for the effectiveness of oncolytic virus immunotherapy.² First, we analyzed the composition of immune cells in the TME on day 4 after OV treatment. Compared with PBS and oAd-NC groups, oAd-αCD47 treatment reduced CD4⁺ T cells and increased the proportion of granulocyte-derived MDSCs (PMN.MDSCs), while no significant changes were observed in CD8⁺ T cells, Mo.MDSCs, and M1.TAMs (figure 1C and online supplemental figure S1A). oAd-αCD47 treatment significantly increased the ratio of CD8⁺ T cells to CD4⁺ T cells, CD8⁺ T cells to Tregs cells, and M1. TAMs to M2.TAMs compared with PBS and oAd-NC groups (figure 1D). Our findings suggest that CD8⁺ T cells and M1-TAMs are key contributors to the efficacy of oAd-αCD47 therapy. To confirm this, we conducted in vivo depletion experiments targeting CD8⁺ T cells, TAMs, and PMN-MDSCs (Ly6G⁺ cells) (online supplemental figure S1B,C). Depletion of CD8⁺ T cells or TAMs significantly reduced the antitumor efficacy of oAd-αCD47 compared with the control group receiving oAd-αCD47 alone. In contrast, depletion of PMN-MDSCs had no notable impact on the therapeutic effects of oAd-αCD47 (figure 1E). These results indicated that oAd-αCD47 treatment induced an antitumor immune response mainly with CD8⁺ T cells and M1.TAMs as the core. Importantly, oAd-αCD47 stimulated marked phagocytosis of tumor cells by macrophages compared with PBS and NC groups (figure 1F and online supplemental figure S1D). The rate of tumor growth can respond to the antitumor effect of oAd-αCD47. As depicted in online supplemental figure S1E, the average tumor growth rate on day 3 post-oAd-αCD47 treatment was a modest 2.4%, potentially attributable to macrophage-mediated phagocytosis and the establishment of an antitumor microenvironment. However, the mean tumor growth rate accelerated significantly from 15.2% to 48.0% between days 14 and 18 following oAd-αCD47 treatment. We speculated that there might be a change in the immune cell composition of the TME, resulting in altered immune homeostasis. Therefore, we analyzed the composition of TIME on day 14 after oAd-αCD47 treatment. oAd-αCD47 and oAd-NC treatments promoted the infiltration of CD4⁺ and CD8⁺ T cells and decreased the proportion of M2.TAMs, thus increasing the ratio of CD8⁺ T cells to Tregs, Teffs (effector T cells) to Tregs, and M1.TAMs to M2.TAMs compared with the PBS group (figure 1G and online supplemental figure S1F). However, neither immune cell composition nor immune homeostasis significantly differed between oAd-αCD47 and oAd-NC groups (figure 1G and online supplemental figure S1G). These data suggested that antitumor immune cells remained dominant on day 14 after oncolytic virotherapy. Moreover, the total number of immune cells increased on day 14 after oAd-αCD47 treatment compared with day 4, and the proportion of major CD8⁺ T killer cells increased from 1% to 12.84% (online supplemental figure S1H). In conclusion, the TIME on day 14 contained numerous antitumor effector cells, especially CD8⁺ T cells but tumor growth was not significantly controlled. Therefore, we hypothesized that the effector function of antitumor effector cells may be suppressed by other immunosuppressive factors in the TME, leading to drug resistance.

Figure 1. CD47 antibody-engineered oncolytic virus controls tumor progression but cannot produce sustained antitumor effects. (A) Mean tumor growth curves and individual tumor growth volumes were calculated after virus treatment in the MC38 tumor model (n=7). (B) Representative tumor size and weight in each experimental group. (C) The percentage of tumor-infiltrating immune cell subsets in tumor tissues was analyzed by FACS on day 4 after virus treatment (n=5). (D) Bar graphs of ratios of CD8⁺ T cells to CD4⁺ T cells, CD8⁺ T cells, or Teffs to Tregs, M1.TAMs to M2.TAMs in MC38 tumors on day 4 post-treatment (n=5). (E) Mean tumor growth curves were calculated after treatment with different depleting antibodies in the MC38 tumor model (n=6). (F) Phagocytic efficiency of macrophages was assayed by flow cytometry. (G) The percentage of the immune cell population in MC38 tumor tissues was analyzed by FACS on day 14 after virus treatment (n=5). Data are presented as the mean±SEM. P values were determined using one-way analysis of variance. FACS, fluorescence-activated cell sorting; MDSC, myeloid-derived immunosuppressive cell; Mo.MDSC, monocyte-derived MDSC; M1.TAM, M1-type macrophages TAM; M2.TAM, M2-type macrophageso TAM; oAd, oncolytic adenovirus; PBS, phosphate-buffered saline; PMN.MDSC, pathological activated neutrophil-myeloid-derived suppressor cells MDSC; TAM, tumor-associated macrophage; Teff, effector T cell; Treg, regulatory T cell.

LA accumulation in the TME leads to metabolic defects in oAd-αCD47-activated T cells and TAMs, which are reprogrammed by NaBi

Both the number and functional status of antitumor immune cells determine the antitumor effect of oAd-αCD47. Tumor cells evade immune surveillance by inducing the overexpression of T-cell co-suppressive markers.²⁹ The expression analysis of co-suppressive molecules indicated that the number of CD8⁺ T cell and CD4⁺ T-cell subsets with low expression of PD-1⁺Tim3⁺ in tumor tissues was significantly lower than that in control groups on days 4 and 14 after oAd-αCD47 treatment (online supplemental figure S2A,B). Thus, the expression of programmed cell death protein 1 (PD-1) and T-cell immunoglobulin domain and mucin domain 3 (TIM3) on the surface of T cells was not the major factor affecting oAd-αCD47 efficacy. Recent evidence suggests that metabolic defects are another crucial predictor of antitumor effector cell dysfunction.³⁰ Therefore, we used MitoTracker staining to label the mitochondria to analyze the mitochondrial content of effector cells as a marker of metabolic adequacy. Interestingly, it was found that the mitochondrial content was significantly lower in CD8⁺ T cells in tumor tissues than in CD8⁺ T cells in the spleen (figure 2A). In addition, the mitochondrial content increased transiently in CD8⁺ T cells on day 4 after oAd-αCD47 treatment compared with control groups (figure 2A and online supplemental figure S2C). However, with tumor-induced metabolic disorders, the mitochondrial content was significantly lower in CD8⁺ and CD4⁺ T cells on day 14 than on day 4 after oAd-αCD47 treatment (figure 2A and online supplemental figure S2C). Moreover, compared with CD8⁺ T cells in the spleen, less glucose was taken up by CD8⁺ T cells in tumor tissues on day 14 after treatment, and no significant difference in glucose uptake of CD4⁺ T cells was detected in spleen and tumor tissues (figure 2B and online supplemental figure S2D). Considering that oAd-αCD47 treatment activates macrophages to phagocytose tumors, the mitochondrial content of macrophages is one of the important factors that determine effective phagocytic function.^{26 31} Therefore, the metabolic profile of TAMs and their subtypes was also analyzed. The results showed that the mitochondrial content of TAMs was higher in the oAd-αCD47 treatment group than in PBS and oAd-NC groups on day 4 after treatment, with no significant difference among the three groups on day 14 (figure 2C). Additionally, TAMs in tumor tissues had a lower mitochondrial content than macrophages (MФ) in normal spleen tissues (figure 2C). Subsequently, the results of TAM subtype analysis showed that the mitochondrial content of M2.TAMs but not M1.TAMs increased in tumor tissues after 4 days of oAd-αCD47 treatment compared with controls (figure 2D and online supplemental figure S2E). Furthermore, more glucose was taken up by TAMs, M2.TAMs, and M1.TAMs in tumor tissues from tumor-bearing mice compared with that taken by macrophages in splenocytes from normal mice on day 4 after treatment; however, the glucose uptake of TAMs, M2.TAMs, and M1.TAMs did not significantly differ from that of macrophages in normal mice spleen cells on day 14 after treatment (online supplemental figures S2F–H). Mitochondria are vital in regulating innate and adaptive immunity and play an important role in the antitumor immune function of T cells and macrophages.³¹ Our data suggested that the mitochondrial content of tumor-infiltrating CD8⁺ T cells but not CD4⁺ T cells was positively correlated with the expression of antitumor cytokines such as TNF-α and IFN-γ and negatively correlated with tumor size (online supplemental figures S1I–K). These findings suggested that mitochondrial-sufficient CD8⁺ T cells can release more antitumor cytokines to inhibit tumor growth. We also found that mitochondria-enriched TAMs corresponded to smaller tumor sizes (online supplemental figure S2L).

Figure 2. NaBi reprograms the metabolism of T cells and macrophages in an LA environment. Analysis of mitochondrial content (A) and glucose uptake (B) in CD8⁺ T cells on days 4 and 14 after treatment in tumor-bearing mice. Representative flow cytogram of MitoTracker for 2NBDG staining in CD8⁺ T cells and spleen-T cells and tabulated flow cytometric data are shown (n=5). (C) Representative cytometry and statistical plots of mitochondrial content in TAMs and spleen-MФ at indicated time points (n=5). (D) Statistics of mitochondrial content in M1.TAMs at indicated time points (n=5). (E and F) The concentration of LA in the supernatant of T cells (E) and BMDMs (F) cultured in fresh RPMI 1640 medium or MC38 cell culture supernatant for 48 hours (n=3). (G and H) Representative flow histograms and statistics of changes in mitochondrial content of T cells (G) and BMDM (H) cultured in RPMI 1640 medium containing different compositions for 48 hours in vitro (n=3). (I) The release of IFN-γ and TNF-α of T cells cultured in RPMI 1640 medium with different compositions for 48 hours in vitro (n=3). (J) BMDM were cultured in RPMI 1640 with different components for 48 hours, then mixed with oncolytic adenovirus-αCD47-infected tumor cell culture supernatant with or without CD47 antibody for 1 hour. This was followed by incubation with MC38 cells stained with the cell membrane red dye-DiD for 4 hours. Fluorescence and flow images of MC38 cells phagocytosed by BMDM were measured by confocal microscopy and fluorescence-activated cell sorting (MC38: BMDM=1:5; n=3). Data represent the mean±SEM. P values were measured using one-way analysis of variance. BMDM, bone marrow-derived macrophage; IFNγ, γ-interferon; LA, lactate; M1.TAM, M1-type macrophages TAM; MFI, mean fluorescence intensity; NaBi, sodium bicarbonate; PBS, phosphate-buffered saline; RPMI, Roswell Park Memorial Institute; TAM, tumor-associated macrophage; TIL, tumor-infiltrating T cells; TNF, tumor necrosis factor.

LA is the final metabolite of tumor glycolysis and can acidify the TME, thereby impairing the mitochondrial function of immune cells, ultimately leading to immunosuppression.³² Thus, we assessed the changes in TME acidity with tumor progression time after oAd-αCD47 treatment. The results revealed that the pH value of the TME was lower than that of the spleen from normal mice and TME acidity further increased at day 14 after treatment compared with day 4 (online supplemental figure S2M). The LA content was higher in MC38 culture supernatants than in T cell and BMDM culture supernatants (figure 2E,F). We then added exogenous LA or MC38 culture supernatants to the culture medium of T cells and BMDMs measured the mitochondrial content and morphology after 48 hours. The results showed that the mitochondrial content of T cells and BMDMs was significantly reduced after the addition of LA or MC38 culture supernatants, which was restored by NaBi (figure 2G,H). Furthermore, we did not detect any significant alterations in mitochondrial morphology, including mitochondrial size, matrix electron density, and cristae structure, among the various treatment groups (online supplemental figure 2N). However, NaBi rescued LA-induced T-cell dysfunction, restoring the levels of TNF-α and IFN-γ in T cells (figure 2I). CD47 blockade promoted phagocytosis of BMDMs, which was inhibited by LA or MC38 culture supernatants. The addition of NaBi to cultures successfully reversed the inhibitory effect of LA or MC38 culture supernatants on the phagocytic capacity of BMDMs (figure 2J).

Overall, although oAd-αCD47 activated CD8⁺ T cells and TAMs to suppress tumor progression, the acidic TME formed by LA accumulation induced mitochondrial depletion of CD8⁺ T cells and TAMs, resulting in the suppression of the antitumor capacity of effector cells. More importantly, NaBi successfully alleviated LA-induced mitochondrial defects in T cells and BMDMs and restored their antitumor function.

NaBi enhances the antitumor effect of oAd-αCD47 in vivo

In vitro experiments demonstrated that NaBi can reprogram metabolically dysregulated T cells and BMDMs in an LA environment. Therefore, NaBi might be a suitable immunomodulator to enhance oAd-αCD47-induced antitumor immunity. First, we evaluated the dynamic pH changes in the TME in tumor-bearing mice treated with long-term oral administration of NaBi. The result showed that the pH value of tumor tissues from mice drinking water containing NaBi was significantly higher than that from mice drinking normal water (online supplemental figure S3A,B). The LA levels in tumor tissue decreased about twofold after NaBi treatment compared with the water group (online supplemental figure S3C). Next, we explored whether NaBi could enhance the antitumor effect of oAd-αCD47 based on the protocol shown in figure 3A. In the MC38 murine colon carcinoma model, it was found that oral administration of NaBi did not significantly affect tumor progression but treatment with the combination of NaBi and oAd-αCD47 (combination therapy) significantly inhibited tumor growth compared with NaBi or oAd-αCD47 alone (figure 3B). In the combination treatment group, approximately 42.9% (3/7) of tumor-bearing mice showed complete tumor regression and eventually achieved long-term survival benefits (figure 3B,C). Next, we further explored the antitumor effect of the combination therapy in murine melanoma (B16-F10) and breast cancer (4T1) models with immune tolerance and highly acidified TME characteristics.^{33 34} In the B16-F10 model, NaBi enhanced the antitumor effect of oAd-αCD47, not only inhibiting tumor progression but also causing complete tumor regression in 33.3% (3/9) of tumor-bearing mice (figure 3D–E and online supplemental figure S3D). Moreover, combination therapy significantly prolonged the survival of tumor-bearing mice and reduced the rate of pulmonary metastatic nodules (figure 3F and online supplemental figure S3E). Compared with monotherapy, combination therapy significantly inhibited subcutaneous tumor progression and lung metastasis in the 4T1 model, and the tumor was completely eliminated in 50% (5/10) of mice, significantly prolonging the survival of tumor-bearing mice (figure 3G–I and online supplemental figure S3F–G). These data suggested that NaBi can enhance the efficacy of oAd-αCD47 treatment in vivo and trigger a durable antitumor response.

Figure 3. NaBi promotes oAd-αCD47 to exert a superior antitumor effect. (A) NaBi combined with oAd-αCD47 antitumor immunization schedule. (B) The mean volume of tumor growth and individual tumor growth in each group of tumor-bearing mice after treatment in the MC38 model; (one-way ANOVA). (C) Survival times of MC38 tumor-bearing mice; (log-rank test). (D–F) Representative images showing changes in tumor bioluminescence (D), tumor volume (E), and survival times (F) of B16-F10 tumor model mice; (one-way ANOVA for tumor volume and log-rank test for survival times). (G–I) Tumor bioluminescence images (G), tumor volume (H), and survival times (I) of the 4T1 breast cancer model; (one-way ANOVA for tumor volume and log-rank test for survival curves). Data are presented as the mean±SEM. ANOVA, analysis of variance; NaBi, sodium bicarbonate; oAd, oncolytic adenovirus; TME, tumor microenvironment.

Combination therapy remodels the immunosuppressive microenvironment to promote activation of CD8⁺ T cells and TAMs

NaBi with oAd-αCD47 combination showed excellent therapeutic effects, implying that the TIME may have been remodeled. Therefore, we used multicolor flow cytometry to analyze changes in TIME after combination therapy. Previous studies reported that NaBi can potentiate the antitumor effects of immunotherapy.^{20 35} however, the exact immunological mechanism remains elusive. Therefore, the antitumor mechanism by which NaBi potentiates oAd-αCD47 warrants further elucidation. Based on the results of the flow-analyzed gating strategy of online supplemental figure S4A, we conclude that compared with the normal drinking water group, NaBi treatment enriched the infiltration of CD4⁺ T cells and Teffs in the TME, while CD8⁺ T cells and M1.TAMs did not change significantly (figure 4A and online supplemental figure S4B). In addition, NaBi treatment reduced the proportion of immunosuppressive Tregs and M2.TAMs but no significant difference was found between PMN.MDSCs and Mo.MDSCs (figure 4A,B and online supplemental figure S4B). Compared with oAd-αCD47 monotherapy, combination therapy reduced Tregs levels and promoted the increase in M1.TAMs, whereas other cellular components did not significantly differ between the two groups (figure 4A–C and online supplemental figure S4B). Combination therapy eventually led to an activated anti-TIME, mainly reflected by the increased ratio of M1.TAMs to M2.TAMs in the TME (online supplemental figure S4C). Compared with the water group, Arg1 expression in M2.TAMs were lower in monotherapy and combination treatment groups, implying that the tumor-promoting function of M2.TAMs were inhibited (online supplemental figure S4D). Moreover, NaBi monotherapy increased the expression of the killer molecule CD107a in CD8⁺ T cells compared with the water group, while the abundance of CD107a expression was higher in CD8⁺ T cells after combination therapy (figure 4D). The CD8⁺ T-cell activation molecule CD69 did not significantly differ among the four groups (online supplemental figure S4E). More importantly, NaBi monotherapy reduced the expression of a co-inhibitory molecule PD-1⁺Tim3⁺ in CD8⁺ and CD4⁺ T cells, whereas PD-1⁺Tim3⁺ expression was lowest in CD8⁺ T cells in the combination therapy group (figure 4E and online supplemental figure S4F). To investigate the relative importance of CD8⁺ T cells and TAMs in the antitumor efficacy of the NaBi combination therapy with oAd-αCD47, we conducted in vivo depletion experiments. The results demonstrated that depletion of either CD8⁺ T cells or TAMs significantly compromised the antitumor effects of the combination therapy (figure 4F). These findings underscore the critical roles of CD8⁺ T cells and TAMs in mediating the antitumor response induced by the combination therapy. Combination therapy shaped the immunosuppressed TME into an immune-activated TME, in which antitumor effector cells dominated tumor progression and played an immune activation function.

Figure 4. Combination therapy reshapes the tumor microenvironment by reducing Tregs, increasing M1.TAMs, and promoting T-cell activation. MC38 tumor tissues were collected and used for the analysis of immune microenvironment composition by fluorescence-activated cell sorting staining after virus administration. (A) The percentage of tumor-infiltrating immune cell subsets in tumor tissues was examined. (B) The percentage of Tregs gated on CD4⁺ T cells. (C) Flow histograms and statistics of M1.TAMs were shown under the F4/80⁺Gr1⁻ (TAMs) gate. (D) Representative flow histogram and abundance of CD107a expression on CD8⁺ T cells. (E) Uniform Manifold Approximation and Projection (UMAP) clustering diagrams and statistical plots after treatment tumor-bearing mice. The percentage of PD-1 and Tim3 were determined in CD8⁺ T cells. (F) Mean tumor growth curves and individual tumor growth volumes were calculated following treatment with different depleting antibodies in the MC38 tumor model (n=6). Data are presented as the mean±SEM. P values were calculated using one-way analysis of variance. Samples from five to seven mice per group were used for analysis. MDSC, myeloid-derived immunosuppressive cell; Mo.MDSC, monocyte-derived MDSC; M1.TAM, M1-type macrophages TAM; M2.TAM, M2-type macrophageso TAM; NaBi, sodium bicarbonate; oAd, oncolytic adenovirus; PD-1, programmed cell death protein 1; PMN.MDSC, pathological activated neutrophil-myeloid-derived suppressor cells MDSC; TAM, tumor-associated macrophage; Teff, effector T cell; TIM3, T-cell immunoglobulin domain and mucin domain 3; Treg, regulatory T cell.

Combination therapy ameliorates metabolic disturbance of CD8⁺ T cells and TAMs to enhance antitumor function in vivo

Next, we analyzed changes in CD8⁺ T cells and TAM metabolism after combination therapy in vivo. The results indicated that NaBi acted as an immunometabolite modulator to increase the mitochondrial content of CD8⁺ and CD4⁺ T cells without altering glucose uptake (figure 5A and online supplemental figure S5A). Furthermore, we verified whether the metabolic alteration could improve the function of TILs. After restimulation with PMA and ionomycin, IFN-γ and TNF-α expression in TILs increased in both the NaBi alone group and combination therapy group (figure 5B). However, increased expression of IFN-γ and TNF-α in CD4⁺ T cells occurred only in combination therapy group (online supplemental figure S5B). Macrophages within TILs exhibited analogous metabolic alterations, characterized by a notable increase in mitochondrial content in TAMs following NaBi treatment (figure 5C). Furthermore, our analysis of TAM subtypes revealed a significant enhancement of mitochondrial content in both M1.TAMs and M2.TAMs (figure 5D and online supplemental figure S5C). The mitochondrial content of M1.TAMs were higher than that of M2-TAMs in the combination therapy group but an increase in the mitochondrial content of M1.TAMs and M2.TAMs were comparable in the NaBi group (figure 5E and online supplemental figure S5D). This data suggested that the antitumor function of M1.TAMs and the phagocytic function of TAMs were enhanced after combination therapy (figure 5F,G and online supplemental figure S5E). Regardless of the treatment, glucose uptake by TAMs, M1-TAMs, and M2-TAMs remained relatively unchanged (online supplemental figure S5F). Collectively, NaBi augmented the antitumor immune response of oAd-αCD47 by rectifying metabolic dysregulation within CD8+T cells and TAMs, thereby enhancing antitumor activity and tumor phagocytosis.

Figure 5. Combination therapy increases the mitochondrial content of TILs and TAMs to fight tumors in vivo. (A) Flow analysis diagram and statistical diagram for the mitochondrial content and glucose uptake in CD8⁺ T cells after treatment in tumor-bearing mice. (B) After tumor-infiltrating T cells were stimulated by phorbol 12-myristate 13-acetic acid and ionomycin for 2 hours, the expressions of TNF-α and IFN-γ in CD8⁺ T cells were detected. Typical flow analysis plots are shown. (C) Flow histograms and statistical plots of mitochondrial abundance in TAMs are shown. (D) Statistical plot of mitochondrial content in M1.TAMs. (E) The ratio of mitochondrial content of M2.TAMs or M1.TAMs in the combined treatment group to those in the untreated group. (F) Typical fluorescent microscopic images of TAMs phagocytosis of MC38 tumors cells. CK19 (green), F4/80 (red), and nucleus (blue). Scale bars: 50 µm. (G) Phagocytic efficiency of macrophages was assayed by flow cytometry. Data are presented as the mean±SEM (six to eight mice per group). P values were calculated using one-way analysis of variance. IFNγ, γ-interferon; M1.TAM, M1-type macrophages TAM; M2.TAM, M2-type macrophageso TAM; MFI, mean fluorescence intensity; NaBi, sodium bicarbonate; oAd, oncolytic adenovirus; TAM, tumor-associated macrophage; TIL, tumor-infiltrating T cell; TNF, tumor necrosis factor.

NaBi increases the mitochondrial content of T cells and BMDMs by promoting the expression of PGC1α mediated by the Ca²⁺-CaMKII-CREB signaling pathway in an LA environment

PGC1α has a critical role in directing mitochondrial proliferation and biosynthesis.³⁶ We also found that TILs and M1.TAMs with high PGC1α expression had sufficient mitochondrial content (figure 6A). This underscored the importance of PGC1α in regulating the mitochondrial content of immune cells in antitumor immunity. Specifically, the expression of PGC1α in CD8⁺ T cells and M1.TAMs were analyzed after treatment and the results showed that PGC1α was significantly increased in both NaBi and combination therapy groups (figure 6B). The transcription of PGC1α is regulated by a cAMP-dependent signaling pathway. CREB, which in turn promote PGC1α transcription.^{37 38} No significant changes were observed in the expression of CREB in CD8⁺ T cells but the expression of the pCREB increased dramatically after NaBi treatment (online supplemental figure S6A). Conversely, CREB expression was reduced in M1.TAMs and pCREB did not change significantly in NaBi and combination treatment groups (online supplemental figure S6B). Nonetheless, the results consistently led to an increase in the ratio of pCREB to CREB after NaBi treatment (figure 6C), indicating that the activation of CREB was increased to promote the expression of PGC1α in the TME. Previous studies reported that high concentrations of LA inhibited the activation of CaMKII signaling by modulating changes in the intracellular calcium levels, thereby regulating PGC1α to control mitochondrial content.^{39 40} Furthermore, we measured abnormal cellular calcium levels in T cells and BMDMs in the LA-enriched medium compared with cells in the normal medium; however, this phenomenon was reversed by NaBi (figure 6D). Immunoblot analysis revealed that there was no significant difference in CaMKII and CREB expression levels between T cells and BMDMs cultured in a normal medium and an LA-enriched medium. However, the pCaMKII and pCREB was reduced under high LA incubation but this trend was reversed by NaBi (figure 6E). In both T cells and macrophages, the pCREB/CREB ratio was consistently reduced in the high lactate environment, resulting in decreased PGC1α expression. However, NaBi treatment effectively reversed this trend (figure 6F,G). Consequently, NaBi can enhance the mitochondrial content of T cells and BMDMs within the LA-rich environment by activating the Ca2+-CaMKII-CREB signaling pathway and promoting PGC1α expression.

Figure 6. NaBi induces an increase in mitochondrial content by enhancing the expression of PGC1α. (A) Correlation analysis of mitochondrial content with PGC1α expression in CD8⁺ T cells or M1.TAMs in all tumor samples (n=23); (two-tailed t-test). (B) Expression analysis of PGC1α in CD8⁺ T cells or M1.TAMs (n=5~7); (one-way ANOVA). (C) The ratio of pCREB to CREB in CD8⁺ T cells or M1.TAMs (n=5~7); (one-way ANOVA). T cells and BMDMs were cultured for 48 hours in RPMI 1640 containing 15 mM LA or 15 mM LA and 15 mM NaBi. (D) The changes in intracellular Ca²⁺ were detected by a Ca²⁺ fluorescent probe (Fura-4, AM). (E) Western blot analysis of CaMKII and pCaMKII, CREB and pCREB in T cells and BMDMs; (20 µg total protein, n=3~4, one-way ANOVA). (F) The ratio of pCREB to CREB in T cells and BMDMs; (G) the expression of PGC1αin T cells and BMDM; (n=3~4, one-way ANOVA). Data presented as the mean±SEM. ANOVA, analysis of variance; BMDM, bone marrow-derived macrophage; CaMKII, calmodulin-dependent protein kinase II; CREB, cyclic AMP-responsive element-binding proteins; LA, lactate; M1.TAM, M1-type macrophages TAM; MFI, mean fluorescence intensity; NaBi, sodium bicarbonate; oAd, oncolytic adenovirus; pCaMKII, phosphorylated form of CaMKII; pCREB, phosphorylated form of CREB; PGC1α, peroxisome proliferator-activated receptor gamma coactivator-1α; RPMI, Roswell Park Memorial Institute.

NaBi combined with oAd-αCD47 promotes memory immune response

The combination of NaBi and oAd-αCD47 induced complete tumor regression in partial tumor-bearing mice. Therefore, we investigated whether combination therapy could stimulate long-term memory immune response. First, we analyzed the lymphocyte composition of the peripheral circulatory system including the lymph nodes (LN) and spleen. We found that NaBi monotherapy increased the infiltration of CD8⁺ T and CD4⁺ T cells primarily in LN but not in the spleen (figure 7A and online supplemental figure S6C). The proportion of CD8⁺ T and CD4⁺ T cells in LN and spleen was comparable between the combination treatment group and oAd-αCD47 monotherapy group, but the CD8⁺ T-cell component in the spleen was increased after oncolytic virotherapy (figure 7A and online supplemental figure S6C). Furthermore, combination treatment primarily enhanced the expression of TNF-α⁺IFN-γ⁺ in CD8⁺ T cells in the spleen, while there was no significant difference in TNF-α⁺IFN-γ⁺T cells in LN (figure 7B and online supplemental figure S6D). The CD8⁺ T cells in the spleen of mice treated with combination therapy had a stronger ability to express perforin (figure 7C). Meanwhile, the expression of effector memory marker CD44 increased on CD8⁺ T and CD4⁺ T cells following the combination treatment in the spleen (figure 7D). Notably, evidence shows that combination therapy induced memory CD4⁺ and CD8⁺ T-cell production with antitumor activity. Subsequently, mice with complete tumor regression after combination treatment were selected as tumor rechallenge subjects. In the MC38 rechallenge experiment, all mice in the rechallenge group (Re) did not develop tumors, whereas mice in the naive group (Na) without any treatment developed tumors (figure 7E). For B16-F10 bilateral rechallenge experiments, mice developed MC38 tumors in both Na and Re groups, while mice had not grown B16-F10 tumor in Re group, and the opposite was true in Na group (figure 7F and online supplemental figure S6E). For 4T1 bilateral rechallenge experiments, both 4T1 and CT26 tumors grew normally in the Na group, whereas some tumors occurred in the Re group, including 1/6 tumor incidence in 4T1 and 3/6 tumor incidence in CT26 (figure 7G). Combination therapy elicited the potent memory immune response, which was essential for controlling distal tumor or tumor metastasis. Tumor inoculation and treatment were performed according to the protocol shown in figure 7H. Left-sided tumors and right-sided tumors had a slower progression under the combination therapy group compared with monotherapy (figure 7I and figure 7J). Interestingly, NaBi treatment alone did not affect right-sided tumor growth but significantly suppressed left-sided tumor progression and tumorigenesis in five mice (figure 7I and figure 7J). Although not statistically significant, a trend towards an increase in CD8⁺ T-cell populations was observed in the NaBi-treated group. Nevertheless, the mitochondrial content of CD8⁺ T cells in the NaBi-treated group was significantly higher than that in the control group (figure 7K,L). In conclusion, NaBi increased the capacity of the oAd-αCD47 to induce effector memory antitumor and suppressed distant tumor growth.

Figure 7. NaBi promotes the memory immune responses induced by oAd-αCD47. (A) Percentage of CD3⁺ T cells expressing CD4 or CD8 in the spleen (n=6~7, one-way ANOVA). (B) TNF-α and IFN-γ released from CD8⁺ T cells or CD4⁺ T cells in the spleen (n=6~7, one-way ANOVA). (C) Expression of perforin in CD8⁺ T cells in the spleen (n=6~7, one-way ANOVA). (D) Percentage of CD44⁺ T cells in CD8⁺ T cells or CD4⁺ T cells in the spleen (n=6~7, one-way ANOVA). (E) Tumor growth curves and tumor size plots of individual mice in the MC38 rechallenge model; (one-way ANOVA). (F and G) After the B16-F10 (F) or 4T1 (G) tumor regressed, the mice in the naive group or the rechallenge group were inoculated with B16-F10 or 4T1 and MC38 or CT26 tumors on the left and right sides, respectively, and the effect of memory response was explored based on the tumor volume and tumor formation rate; (one-way ANOVA). (H) Schedule of inoculation and treatment for bilateral MC38 tumor models. (I and J) Left and right individual tumor growth curves (I) and average tumor growth curves (J); (one-way ANOVA). (K) The proportion of CD8⁺ T cells in the left tumor (n=6~7, one-way ANOVA). (L) Representative flow analysis diagram and statistical plots of mitotracker staining of CD8⁺ T cells in the left tumor (n=6~7, one-way ANOVA). Data are presented as the mean±SEM. ANOVA, analysis of variance; IFN, interferon; Na, naive group; NaBi, sodium bicarbonate; oAd, oncolytic adenovirus; Re, rechallenge group; TNF, tumor necrosis factor.

NaBi combined with oAd-αCD47 neoadjuvant therapy inhibits tumor recurrence and metastasis after surgical resection

Studies have shown that cancer recurrence and metastasis are responsible for more than 90% mortality rates.⁴¹ Although surgical resection is one of the major treatment options for cancer, residual tumors and circulating tumor cells (CTCs) remain the dominant factors for recurrence and metastasis.⁴² A randomized clinical trial revealed that preoperative adjuvant oncolytic virotherapy could substantially increase recurrence-free survival in patients with advanced melanoma.^{43 44} Although oncolytic viral therapy combined with surgical resection has demonstrated superior response rates, regulating the acidic TME before and after surgery may be a suitable strategy to further improve outcomes. Therefore, we investigated the feasibility of NaBi combined with oAd-αCD47 as a neoadjuvant strategy before surgical resection and the efficacy of NaBi maintenance therapy after surgical resection. We used a highly aggressive B16-F10 mouse model of incomplete tumor resection to simulate tumor recurrence. NaBi was administered to tumor-bearing mice before and after oAd-αCD47 combined surgical resection (figure 8A). In vivo imaging and tumor growth curves of mice were adopted to check tumor recurrence after treatment (figure 8B,C). Consequently, both preoperative NaBi combined with oAd-αCD47 treatment (①vs③) and postoperative NaBi maintenance therapy (①vs② and ③vs④) inhibited tumor progression and recurrence to a certain extent (figure 8D). The most optimal treatment effect, however, was observed in a neoadjuvant treatment approach of NaBi combined with oAd-αCD47 before surgery and continued maintenance of NaBi after surgery (figure 8D), including about 28.6% (2/7) mice without detectable tumors, prolonged survival in mice and the lowest tumor LN metastasis (figure 8E,F). Studies have shown that surgery can promote cancer metastasis by releasing CTCs into circulatory system.⁴¹ To evaluate the potential of NaBi in removing CTCs, a metastasis model was constructed by intravenously injecting B16-F10 cells into mice after complete surgical resection of the tumor, simulating the shedding of CTCs from the primary tumor into the systemic circulation (figure 8G). As anticipated, mice treated with the combination of NaBi and oAd-αCD47 as neoadjuvant therapy prior to surgery, followed by maintenance NaBi treatment, exhibited the highest survival rates and the lowest incidence of lung metastasis (figure 8H,I). Notably, NaBi therapy reduced the expression of immune checkpoint Tim3⁺ on the surface of CD8⁺ T cells in blood unlike the group that did not receive NaBi treatment; however, no significant difference was noted in the expression of PD1⁺ among all groups (figure 8J). In addition, we observed higher mitochondrial mass and more IFN-γ and TNF-α release from CD8⁺ T cells in the two groups treated with NaBi postoperatively (figure 8K,L). These findings suggest that the neoadjuvant treatment regimen of NaBi combined with oAd-αCD47 preoperatively and NaBi maintenance postoperatively can prevent recurrence and hematogenous metastasis of tumor lesions; this is an effective approach to improve overall survival.

Figure 8. NaBi combined with oAd-αCD47 neoadjuvant therapy inhibits tumor recurrence and metastasis after surgical resection. (A) Schedule of inoculation and treatment for B16-F10 recurrence models. (B–F) Changes in tumor bioluminescence images over time (B), individual tumor growth curves (C), average tumor growth curves (D), survival times (E), and tumor lymph node metastasis rate (F) of B16-F10 tumor recurrence models; (one-way ANOVA for tumor volume and log-rank test for survival curves). (G) Schedule of inoculation and treatment for B16-F10 metastasis models. (H and I) Changes of mice survival curves (H) and tumor lung metastases (I) in B16-F10 tumor metastasis models; (log-rank test for survival curves and one-way ANOVA for lung metastases). (J) The percentage of Tim3 and PD-1 were determined in CD8⁺ T cells in peripheral blood (n=6); (one-way ANOVA). (K) The percentage of mitotracker was determined in CD8⁺ T cells in peripheral blood (n=5~6); (one-way ANOVA). (L) Representative flow cytometry contour plot and statistical plots of CD8⁺ T cells expressing IFN-γ and TNF-α in peripheral blood (n=5~6); (one-way ANOVA). Data are presented as the mean±SEM. ANOVA, analysis of variance; IFN, interferon; i.v., intravenous; NaBi, sodium bicarbonate; oAd, oncolytic adenovirus; PD-1, programmed cell death protein 1; TIM3, T-cell immunoglobulin domain and mucin domain 3; TNF, tumor necrosis factor.

Discussion

A swift immune response, characterized by a significant influx of effector T cells and a reconfiguration of the immunosuppressive TME, is crucial for effective antitumor growth. However, the challenging inhibitory microenvironment within solid tumors often hinders T-cell infiltration post-immunotherapy. Consequently, engineering a favorable immune activation microenvironment is imperative. Moreover, the presence of numerous tumor-promoting TAMs in the inflammatory TME, and induction of TAMs polarization into M1.TAMs with an antitumor phenotype by modulating “do not eat me” signaling is an effective approach to rapidly activate immune responses.^{10 11} This study demonstrates that an oncolytic virus can overcome the immunosuppressive microenvironment and T-cell depletion within tumors. Consistent with previous findings, OVs expressing CD47-blocking antibodies can stimulate the differentiation of TAMs into M1-TAMs and induce phagocytosis of tumor cells.^{15 16} However, we found that CD47 antibody-dependent phagocytosis was disrupted by the TME. Studies have shown that the accumulation of fatty acid oxidase in the TME suppresses CD47 antibody-mediated phagocytosis in recurrent gliomas.⁴⁵ Elsewhere, another study demonstrated that tumor cell overexpression of dihomoglycoside GD2 disrupts macrophage phagocytosis whereas blocking GD2 and CD47 signaling significantly improves the antitumor effect.⁴⁶ Furthermore, we found that CD47 antibody-dependent phagocytosis is inhibited by LA accumulated in the TME. These findings prolong the immunosuppressive effects of LA, providing a theoretical reference for improving CD47 antibody treatment by regulating the acidic microenvironment (figure 9).

Figure 9. Schematic diagram of combination therapy to remodel the acidic microenvironment of tumors to promote antitumor effect. Tumor acidic microenvironment impairs mitochondrial energy metabolism in macrophages and T cells and induces oAd-αCD47 immunotherapeutic resistance; NaBi improves the acidity of the TME and activates the calmodulin-dependent protein kinase II/cyclic AMP-responsive element-binding proteins/PGC1α mitochondrial biosynthesis signaling pathway, thereby reprogramming the energy metabolism of macrophages and T cells in the TME, and oral NaBi enhances the antitumor effect of oAd-αCD47. MDSC, myeloid-derived immunosuppressive cell; M1.TAM, M1-type macrophages tumor-associated macrophage; M2.TAM, M2-type macrophages tumor-associated macrophage; NaBi, sodium bicarbonate; oAd, oncolytic adenovirus; pCaMKII, phosphorylated form of CaMKII; pCREB, phosphorylated form of CREB; PGC1α, peroxisome proliferator-activated receptor gamma coactivator-1α; TME, tumor microenvironment.

LA accumulation promotes tumor progression, and antitumor immune response can be improved by inhibiting LA production or neutralizing accumulated LA.⁴⁷ The alkaline environment can neutralize the acidity of LA, making LA exist as LA ions. Studies indicate that LA ions can stimulate the stemness of CD8⁺ T cells to improve antitumor immunity.⁴⁸ Therefore, the acidic microenvironment by NaBi was neutralized to improve oAd-αCD47-mediated antitumor immune responses of both macrophages and T cells. NaBi is clinically used to antagonize metabolic acidosis in patients. Oral NaBi can enter solid tumors to relieve acidity of TME, and combined with immune checkpoint (PD-1 and CTLA-4) inhibitors improve antitumor immune response.^{35 49 50} In recent studies, NaBi was found to suppress leukemia recurrence and progression by reprogramming T-cell metabolism.⁵¹ In addition, our study underscores the potential of NaBi to serve as an immunometabolic modulator, and the combination of NaBi and oAd-αCD47 is more effective than monotherapy. By analyzing the T cell and TAMs compartment through multicolor flow cytometry, NaBi improved the metabolic capacity of T cells and TAMs and promoted the generation of T cells with memory phenotype to maintain a durable immune response. The combination treatment induced a complete tumor regression in various tumor-bearing mice. These complete responders develop a long-term memory response to suppress tumor growth on reoccurrence. Therefore, not all the immune memory clones induced by combination therapy are virus-specific but only specific T cells that can effectively destroy tumors on reoccurrence.

According to previous studies, NaBi promotes the treatment effect of immune checkpoint inhibitors by increasing the pH of the TME.³⁵ However, whether NaBi affects the composition and function of the immune system remains unclear. Our results indicate that NaBi weakens the formation of an immunosuppressive microenvironment by reducing the number of Tregs and M2.TAMs. Furthermore, NaBi downregulated the Arg1, PD-1, and Tim3 expression to establish an immunosuppressive effect on the TME. However, NaBi monotherapy cannot control tumor growth, possibly because it cannot recruit TILs to destroy tumors. To achieve optimal antitumor effects, immunotherapy thus requires simultaneous regulation of the relationship between antitumor immune cells and the immunosuppressive microenvironment.

Excessive LA production is considered one of the hallmarks of tumor metabolic imbalance.¹⁷ Our findings indicate that impaired mitochondrial mass in T cells and macrophages in an LA-enriched environment causes antitumor dysfunction. Moreover, this study reinforces the notion that mitochondria are homeostatic regulators of continuous killing properties of antitumor immune cells.^{31 52} Importantly, we confirmed that LA disrupts mitochondrial homeostasis in immune cells but can be rescued by NaBi. This is extremely important for the oAd-αCD47 to activate high-efficiency TAMs to phagocytose tumors; this is because while blocking the “don’t eat me” signal promotes the phagocytosis of TAMs, the loss of mitochondria that provides phagocytic energy also results in phagocytic impairment. Our findings are also consistent with previous studies that LA-induced dysregulation of intracellular Ca²⁺ levels can suppress PGC1α expression, causing loss of mitochondrial mass.^{37 39 40} It was reported that the expression of PGC1α was decreased in tumor-infiltrating effector cells.^{26 30} Herein, we observed that NaBi reinforced the functions of antitumor immune cells by upregulating the expression of PGC1α (figure 9).

NaBi combined with oAd-αCD47 treatment induced the formation of immune memory, as evidenced by tumor-free surviving mice showing delayed or complete protection from tumor growth when challenged again with tumor cells. Interestingly, in our model, mice that developed a complete response to B16-F10 (4T1) tumors after combination therapy were also protected from the MC38 (CT26) tumor challenge. The observed cross-tumor immunity might be attributed to the presence of a shared tumor antigen recognized by both MC38 (CT26) and B16-F10 (4T1). Furthermore, common mutations between these tumor cell lines could contribute to the delayed growth of both tumors. Thus, the presence of similar antigenic epitopes in different tumors originating from mice of the same genus after combination therapy may further improve immune memory.

Combination therapy is the future development direction of antitumor immunotherapy. The use of NaBi in this work to improve the effector function of oncolytic virus-induced antitumor immune cells proved to be highly valuable. Whereas OVs have potent immune-stimulatory potential, sufficient metabolic support of immune cells is key for a sustained complete antitumor immune response and promotes long-term memory effects. Our findings suggest that NaBi could serve as a promising metabolic modulator in immunotherapy. While most existing immunotherapy strategies for OVs primarily involve cytokines, immune checkpoint inhibitors, and adoptive cell therapy, our results highlight the potential of metabolic modulators to synergize with OVs and enhance immunotherapy outcomes. Furthermore, NaBi may be beneficial in combination with other immunotherapy modalities, such as chimeric antigen receptor T-cell immunotherapy, adoptive cell immunity, and tumor vaccine therapy. Additionally, NaBi combined with oAd-αCD47 neoadjuvant treatment strategy is important in preventing tumor recurrence and metastasis following surgical treatment. This is because residual tumor cells after surgical resection will evade immune surveillance through dormancy, eventually resulting in tumor recurrence and metastasis.^{53 54} However, NaBi activates these dormant cancer cells and improves tumor cell clearance.⁵³ In conclusion, our study presents a novel and effective combination strategy for immunotherapy, with good cancer treatment effects.

Data availability statement

Data are available upon reasonable request.