Cancer cells accelerate exhaustion of persistently activated mouse CD4 T cells.

Introduction
Tumors employ immunosuppressive mechanisms to evade eradication by the immune system. One such mechanism is the induction of T cell exhaustion, characterized by a decline in T cell effectiveness due to chronic antigen exposure. Additionally, tumor cells exhibit increased expression of PD-L1, a protein that serves as a ligand for programmed cell death protein 1 (PD-1), which is overexpressed on the surface of exhausted tumor-infiltrating lymphocytes (TILs). This interaction contributes to immune evasion by inhibiting T cell function and promoting an immunosuppressive tumor microenvironment.1 Overexpression of PD-L1 protein can be stimulated by soluble factors such as interferon γ (IFN-γ) or tumor necrosis factor α (TNFα) secreted by immune cells; therefore, direct contact between cells is not needed.2,3 Persistent exposure to antigens leads to robust immunosuppressive effects, including lymphocyte exhaustion.4 Exhausted T cells are characterized by both phenotypic alterations, such as expression of surface markers (e.g., PD-1, TIGIT, TIM-3, LAG-3), and functional changes, such as inhibited secretion of IFN-γ, TNFα, and IL-2 or altered expression of transcription factors (e.g., BATF, NFAT1, Eomes – increased; VHL – decreased).5 The metabolic state of the exhausted T cells is also impaired. For instance, cancer cells compete with T cells for scarce substrates – glucose and oxygen – available in the tumor microenvironment (TME). The depletion of glucose and oxygen within the local environment leads to T cell hypoxia and energy deprivation and ultimately compromises the efficacy of the immune response. Moreover, the hypoxic environment and concomitant overexpression of hypoxia-inducible factor (HIF) expression in exhausted cells induce oxidative stress, reduced cytotoxic function, and ultimately leads to cell death.6 Notably, fatty acid, cholesterol, and lipid accumulation can enhance T cell exhaustion by endoplasmic reticulum stress induction and overall metabolic disruption.7,8 Further, tumors consume high levels of specific amino acids, glutamine, and arginine, which play crucial roles in the proliferation and effector function of infiltrating T cells. Furthermore, the function of TILs is compromised by the elevated expression of indoleamine pyrrole 2,3-dioxygenase 1 (IDO1), which leads to tryptophan depletion and kynurenine secretion.9 The exhaustion process in TILs is marked by global epigenetic changes involving the differential availability of thousands of loci. Chromatin accessibility for transcription factors and other chromatin related machineries is determined by histone H3 trimethylation or acetylation at the H3K9/H3K27 positions.10 These epigenetic activities are maintained by highly overexpressed DNA methyltransferases 1/3 (DNMT1/DNMT3) or enhancer of zeste homolog 2 (EZH2), a subunit of the polycomb-repressive complex 2.11 Studies have demonstrated that knockout of the DNMT3A protein with simultaneous anti-PD-1 treatment can secure T cell rejuvenation.12
The majority of studies have focused on the exhaustion of CD8+ cytotoxic T cells (CTLs) due to their well-studied ability to directly eliminate tumors. However, several reports have shown that CD4+ T cells not only play a role as helper cells to support CD8+ CTLs but can also independently exert anti-tumor effects.13,14 It is noteworthy that CD4+ T cells can manifest elevated expression of specific inhibitory receptors such as PD-1, TIM-3, and CTLA-4 which are expressed also by exhausted CD8+ T cells.15,16 Additionally, CD4+ T cell maturation, cytokine secretion, and mobility are reduced.17 Our previous studies have confirmed that effector CD4+ T cells, characterized by the overexpression of the PD-1/PD-L1 pathway, can also become exhausted. In humans, exhausted CD4+ T cells can be identified by the overexpression of PD-L1 protein as well as by the expression of PD-1. Additionally, metabolic, transcriptional, and functional changes occur in CD4+ T cells, leading to inhibition of proliferation and cytotoxic functions. The epigenetic signature of the CD274 (PD-L1) promoter is also noteworthy. The levels of PD-L1 protein may be influenced by the interactions between subunits of the SWItch/Sucrose Non-Fermentable chromatin remodeling complex (SWI/SNF) and PRC2 proteins. These complexes predominantly act in an antagonistic manner to maintain specific homeostasis, which allows the proper execution of cellular processes.18
The process of T cell exhaustion has been thoroughly investigated in a number of species including non-human primates, mice, and humans.19,20 Due to the similarities between the mouse and human immune systems, T cell hyporesponsiveness has been extensively studied using mouse models.21
In vitro and ex vivo studies exploiting mouse models, offer a cost-effective and time-saving solution for evaluation of the biological mechanisms underlying exhaustion.22 Most mouse studies focus on the exhaustion of effector CD8+ T cell populations.23–25 The regulatory T cell (Treg) subpopulation of CD4+ T cells, among others, has been the focus of most studies. The prevalence of Tregs in the tumor microenvironment (TME) may accelerate the exhaustion.26,27 However, the field requiring in-depth insight is the specific mechanism underlying the exhaustion of other CD4+ T cell subpopulations in mice.
The present study hypothesized that the mouse immune system would serve as an adequate model of the designed exhaustion studies. Therefore, the objective of the research was twofold: first, to create a mouse model of CD4+ T cell exhaustion; and second, to confirm that it corresponds to the human exhaustion model. Here, we demonstrated that the overexpression of PD-L1 protein is not exclusive to the exhaustion of human CD4+ T cells but also to mouse CD4+ T cells. Hence, the fundamental mechanism of exhaustion does not differ significantly between these species. Consequently, mice can serve as a valuable model organism for more detailed investigations into CD4+ T cell exhaustion and studies on immunomodulatory drugs to fight against cancer. Additionally, we confirmed that the human CD4+ T cell exhaustion model described in our study by Jancewicz et al.18 accurately reflects the patterns observed in vivo in mice.18

Materials and methods

Animals
Female BALB/c (strain code: 028) and C57BL/6 (strain code: 027) wild-type mice (Charles River), age 8–12 weeks, were housed under standard pathogen-free conditions at the Animal Laboratory of the National Institutes of Health (NIH). For experiments focused on IFN-γ dependence, C57BL/6 mice with IFN-γ receptor knockout were used (Jackson Laboratory). Subsequent to the acquisition of the animals, they were accommodated for a minimum period of one week at NIH animal facility.

Sex as a biological variable
In our previous studies, we demonstrated human model of CD4+ T cell exhaustion, conducted on male population (18). To exclude the influence of hormones on the developed model, only female mice are used in the present article. Analogous mechanisms will be demonstrated in both sexes by comparing the two results.

Study approval
All animal procedures reported in this study performed by NCI-CCR-affiliated personnel were approved by the Institutional Animal Care and Use Committee (IACUC) at National Institutes of Health in Bethesda MD, USA, and in accordance with federal regulatory requirements and standards (reference number: METB-033). All components of the NIH intramural ACU program were accredited by the AAALAC International. The study adheres to The Declaration of Helsinki. The study was conducted according to the ARRIVE guidelines.

Mouse tumor cells
MC38 (Kerafast, RRID: CVCL_B288) and CT26 (ATCC, RRID: CVCL_7256) colorectal carcinoma cell lines were obtained from the Vaccine Branch, National Cancer Institute, Bethesda. Cells were cultured in cRMPI (complete RPMI), which contained RPMI 1640 (Gibco) supplemented with 10% heat-inactivated fetal bovine serum (Gibco), 25 mm HEPES, 2 mm L-glutamine, 10 mm NAA-amino acids, 1 mm sodium pyruvate, 100 U/ml penicillin/streptomycin, and 50 µM β-mercaptoethanol. Cells were grown in a humidified incubator at 37°C with 5% CO2. Prior to the experiments, cell lines were tested for mycoplasma contamination.

IFN-γ treatment
The CT26 and MC38 cell lines were cultured under standard conditions until 70% confluency was obtained. The cells were then treated with recombinant mouse IFN-γ (Merck) at a final concentration of 5 ng/ml. After 24 h of incubation, cells were harvested and stained with anti-H-2Kd, anti-H-2Kb, anti-MHC II, or anti-PD-L1 antibodies for flow cytometry analysis.

Splenocytes and CD4+ T cell isolation
The experiment was replicated thrice, with each repetition utilizing a different spleen from a distinct mouse. This approach yielded a total sample population of 3 BALB/c mice and 3 C57BL/6 mice. Naïve BALB/c and C57BL/6 wild-type mice were euthanized via CO2 inhalation. The spleens were aseptically harvested and placed in a cold cRPMI medium. All steps were performed on ice, unless otherwise noted. The spleen and media were transferred onto a 70 µm nylon mesh and disassociated using gentle pressure from a sterile plunger. The cell strainer was washed twice with medium. Subsequently, the cells were centrifuged, and the supernatant was discarded. The pellet was resuspended in ACK Lysis Buffer (Gibco) and incubated for 2 minutes at room temperature. The medium was added, and the suspension was filtered through a 40 µm cell strainer. The mesh was then washed twice with medium. The isolated splenocytes were pelleted by centrifugation. After determining the cell count, downstream applications were continued.
CD4+ T cells were isolated using the CD4+ T cell Isolation Kit, mouse (Miltenyi Biotec) according to the manufacturer’s protocol, and then activated with Dynabeads™ Mouse T-Activator CD3/CD28 for T-Cell Expansion and Activation (Invitrogen) at a 1:1 ratio. Activated CD4+ T cells were cultured vertically in T75 flasks at a density of 0.5 × 106 cells/mL in cRMPI medium (Gibco) supplemented with 10 U/ml interleukin-2 (Miltenyi Biotec). The cells were split 1/3 every 2–3 days by adding fresh medium with IL-2.

Mixed cultures of CD4+ T cells and cancer cells
Isolated and activated CD4+ T cells were cultured for 8 days following the procedure described above. On day 7, CT26 and MC38 cells were seeded at approximately 70% confluency in 6-well plates. On day 8, CD4+ T cells were restimulated with beads again in a 1:1 ratio, and 2.1 × 106 T cells were transferred to 6-well plates. Cells were continuously cultured in cPRMI medium without exogenous IL-2.
The lymphocytes were divided into three groups: control without cancer cells (T15), co-culture with cancer cells (TNT), and paracrine signaling (TPS), which were maintained on a 6-well plate with a Transwell membrane (Corning) with 0.4 µm pore dimensions. After 48 h of incubation under standard conditions, CD4+ T cells were gently detached from cancer cells, counted using a Neubauer hemocytometer chamber or Cell Counter Countess (Invitrogen), and collected for further analysis.
Under the same conditions as described above, the in vitro experiment was performed using IFNR knockout CD4+ T cells. The cells were isolated from C57BL/7 IFNR-KO mice and the experiment was repeated twice in triplicates, resulting in the sacrifice of two additional mice.

CD4+ T cell exhaustion in vivo
For the in vivo experiments, the CT26 and MC38 cell lines were used. BALB/c and C57BL/6 mice were inoculated with 2 × 105 CT26 or MC38 cells, respectively, in the left abdominal flank. The cells were resuspended in 100 µL of sterile phosphate-buffered saline (PBS). Mice were randomly divided into three groups, each group contained 8 mice of each strain, with final exclusion of 2 mice from each group. Each mouse was considered as an individual within the group, ultimately resulting in 6 repetitions of the experiment. A total of 18 BALB/c mice and 18 C57BL/6 mice were considered as statistically relevant. Animals were euthanized when the tumor size reached approximately 150 (small tumor group), 500 (medium tumor group), or 1500 (large tumor group) mm3. The tumors were measured every 3–4 days and the volume was calculated using the improved ellipsoid formula.
All animals were euthanized before the tumor size reached 2000 mm3 as required by the ACUC policy. After sacrifice, tumors were harvested and dissociated using the Tumor Dissociation Kit for mice (Miltenyi Biotec) with a gentleMACS Octo Dissociator (Miltenyi Biotec) according to the manufacturer’s instructions. Tumor-infiltrating lymphocytes (TILs) were separated by centrifugation on a Percoll density gradient (GE Healthcare) and immediately stained using the appropriate flow cytometry protocol. At the same time, spleens were harvested from each mouse bearing a tumor. Splenocytes were then isolated from these organs. Following a thorough analysis via flow cytometry, the splenocytes were utilized as a control.

Flow cytometry
At least 5 × 105 T cells were used for staining. Cell viability was assessed using the LIVE/DEAD™ Fixable Blue Dead Cell Stain Kit for UV excitation (Invitrogen), diluted in PBS. Antibody staining was performed using PBS containing 1% FBS. To facilitate manipulation, cells were fixed with 1% paraformaldehyde for 15 min after surface staining. A Cytofix/Cytoperm (BD Biosciences) permeabilization kit was used for intracellular staining of FoxP3 and IDO1. Flow cytometry was conducted in the Vaccine Branch Flow Cytometry Core Facility (NCI) using BD FACS Symphony™ A5 (Becton Dickinson). Data were analyzed using the FlowJo™ software (RRID: SCR_008520). In the course of the experiment, the FMO (Fluorescence minus one) sample, constituted of all the groups studied, was utilized as the autofluorescence level of cells in order to analyze the levels of exhaustion markers. The antibodies used in the experiments are listed in Table 1.

Cytokine array
The supernatants were collected from restimulated, co-cultured, and paracrine signaling groups of exhausted C57BL/6 CD4+ T cells. To detect multiple soluble factors simultaneously, the Proteome Profiler Mouse Cytokine Array Panel A (R&D Systems) was used following the manufacturer’s protocol. The supernatants from each group were incubated with membranes separated from the kits with the same LOT number. Chemiluminescent signals were visualized using a western blot imaging system. The experiment was performed twice.

Statistics
In order to ascertain the normality of the distribution, the differences, and the significance of the samples in each experiment, an ordinary two-way ANOVA with multiple comparisons and Dunnett’s, Bonferroni’s or Tukey’s post hoc correction was used. Statistical p-value <0.05 was considered significant (*<0.05; **<0.01; ***<0.001, ****<0.0001). Results were showed as mean and ± SEM. Statistical analyses were performed using the GraphPad Prism version 10 (RRID: SCR_002798). The heatmaps were estimated using hierarchically clustered columns (ward.d2 method) with the pheatmap package in R. The BioRender web software (RRID: SCR_018361) was used to create schematic illustrations.

Results

Persistently stimulated CD4+ T cells and cancer cells mutually inhibit each other’s growth in the mixed cultures
For co-culture experiments, CT26 and MC38 cancer cell lines were used, derived from BALB/c (CT26) or C57BL/6 (MC38) mouse colorectal cancer. CD4+ T cells were isolated from BALB/c or C57BL/6 mice and activated using anti-CD3/CD28 moAb-coated stimulatory beads. The incubation time was determined experimentally and was limited to 8 days for the first stimulation and 2 days for restimulation. CD4+ T cells were cultured with cancer cells and collected for further analyses (Figure 1a). A significant inhibition of cancer cell growth was observed after culturing with restimulated T cells (Figures 1b,c). Simultaneously, co-culture and incubation separated by transwell membranes resulted in a three- to four-fold reduction in the number of CD4+ T cells compared to incubation in the absence of cancer cells (Figures 1b,c). These findings indicate that persistently stimulated CD4+ T cells and cancer cells mutually inhibit each other’s growth in mixed cultures.

Co-culture of murine CD4+ T cells and PD-L1-positive cancer cells results in upregulation of PD-L1 protein on CD4+ T cells
Harvested CD4+ T cells, after restimulation and incubation with tumor cells, were carefully counted and 0.5 × 106 cells were prepared for flow cytometry analysis. The cells were stained using the LIVE/DEAD™ Fixable Blue Dead Cell Stain Kit for UV excitation (Invitrogen) to expose the live population. For precise phenotype and exhaustion determination, only the CD45+CD3+CD4+CD8neg subset was examined (Supplemental Data Figure S2). The present study considered three different cell phenotypes: naïve/resting (CD62L+CD44neg), effector memory (CD62LnegCD44+) and central memory (CD62L+CD44+) CD4+ T cell phenotypes.28 The analysis revealed that the proportion of CD4+T cells with CD62L−CD44+ effector phenotype decreased significantly in T cells co-cultured in the presence of cancer cells in comparison to the T cells cultured in the absence of cancer cells, whereas the proportion of CD4+T cells with central memory phenotype CD62L+CD44+ increased in the presence of cancer cells (Supplemental Data Figure S3). The analyzed groups of CD4+ T cells were also stained with antibodies against the most characteristic markers of the exhaustion process: PD-1 (CD279), PD-L1 (CD274), PD-L2 (CD273), TIM-3 and LAG-3. The gating strategy is shown in Supplemental Data Figure S1, S2, S4, and S5. Compared to day 8, a notable increase in the expression levels of all markers was noted following restimulation in all experimental groups, thereby indicating the initiation of the exhaustion phase in the lymphocytes (Supplemental Data Figure S4 and S5). Among all the markers associated with programmed cell death, the greatest alterations were observed in PD-L1 protein expression. The co-culture group in C57BL/6 mice exhibited a trend of greater than 10-fold increase in the % of PD-L1 protein-expressing cells among the CD45+/CD4+ subset (Figure 2a), while the BALB/c group demonstrated a more than two-fold increase (Figure 2b). We also noticed higher number of PD-L1-positive CD4+ T cells in the transwell-cultured groups in both mouse strains. Simultaneously, there was a statistically significant elevation in the Median Fluorescence Intensity (MFI) of the stained PD-L1 marker in the co-cultured group compared to that in the restimulated group, as observed in both mouse strains tested (Figures 2c,d). In contrast, the observed levels of PD-1 and PD-L2, as well as TIM-3 and LAG-3 marker expression levels in the co-culture group, remained largely unaltered, exhibiting only a negligible tendency toward reduction. Moreover, T cells grown in the transwell plate separated from tumor cells by a permeable membrane showed upregulation of PD-L1, especially in the cells isolated from BALB/c mice. These findings confirmed the suspected involvement of soluble factors or secreted particles in the exhaustion process. In summary, the results presented herein confirmed the findings from our previous investigations involving human subjects.18 These findings suggests that PD-L1 expression increases in CD4+ T cells upon co-culture with tumor cells, potentially suggesting a link with exhaustion.

Tumor growth is accompanied by the increase of expression of PD-L1, CTLA-4, and IDO1 on tumor infiltrating CD4+ T cells
To substantiate the conclusion that the outcomes observed in our previous experiments accurately represent the physiological alterations occurring in living organisms during the process of CD4+ T cell exhaustion, we performed in vivo experiments using BALB/c and C57BL/6 mouse models. Animals were inoculated with 2 × 105 CT26 or MC38 cells into the abdominal flank and tumor growth was carefully monitored during the experiment. Mice were sacrificed when the tumors reached approximately 150, 500, or 1500 mm3 (±100), and the groups were considered small, medium, and large tumors (Figure 3a). All animals under investigation were euthanized per animal welfare ACUC guidelines before the tumor size exceeded 2,000 mm3 to reduce unnecessary suffering. After euthanasia, tumors were harvested, and TILs were isolated and immediately stained for flow cytometry analysis with the whole antibody panel (see Material and Methods 2.9, see Supplemental Data Fig S6-S9 for gating strategy and FACS analysis). From each mouse, splenocytes were also isolated and stained, which served as a control. Careful analysis of TIL FACS staining confirmed our in vitro results and hypothesis that persistent stimulation of CD4+ T cells in the TME leads to their exhaustion, manifested as overexpression of the PD-L1 protein (Figures 3b,c). TILs infiltrating small tumors were characterized by the lowest level of PD-L1 as well as other markers, such as TIM-3, LAG-3, CTLA-4, and IDO1. As tumors grew up to medium size, PD-L1 expression increased, which may indicate activation of the exhaustion signaling axis. The PD-L1 level in the large tumor group was not as high as in the medium group, suggesting the dynamic nature of exhaustion process. PD-L1 may be transiently upregulated during intermediate tumor stages but suppressed in late stages due to T cell loss or alternative immunosuppressive mechanisms. Interestingly, IDO1 protein levels were higher in large tumors. It can be reasonably assumed that as the tumor grows, this molecule enhances its immunosuppressive effect, inhibiting the influx of new TILs into the TME and making the immune cells less responsive to the tumor. Significantly, we detected variations in PD-L1 and IDO1 expression among CD4+ TILs, contingent on the tumor cell line. C57BL/6 mice harboring MC38 tumor cells exhibited elevated levels of exhaustion markers within the TILs, a finding likely due to intrinsic differences between the strains.

To obtain more comprehensive data, the critical comparator was provided by analyzing matched splenocytes from the same tumor-bearing mice (Figure 3d, Supplemental Data Figure S10). The splenic CD4+ T cell population exhibited significantly lower expression of exhaustion markers in comparison to TILs, especially in C57BL/6 + MC38 model, thereby confirming that the exhaustion phenotype is more TME-specific, as splenic T cells remain relatively unaffected despite bearing the same tumor antigen exposure history. However, the detection of upregulation of CTLA-4 and IDO1 markers in splenocytes from medium and large groups suggests systemic immune response by tumor-derived factors.

Cell-to-cell contact is not required for CD4+ T cell activation and exhaustion
Transwell experiments suggested that soluble factors can mediate the observed effects. To identify small soluble molecules that may also be involved in this process, a Cytokine Array (R&D Biosystems) was used. After polyclonal stimulation with C57BL/6 CD4+ T cells isolated from wild-type mice and cultured with or without cancer cells, the supernatant was collected and used according to the manufacturer’s protocol. The raw data of the arrays are available in Supplemental Data Figure S12. The results demonstrated a strong reduction in the expression levels of the majority of the immune system stimulating cytokines, including IL-1A, IL-4, and IL-13. Additionally, the secretion of IL-2, which is responsible for stimulating T cells to growth, was also inhibited. Notably, IFN-γ and TNFα secretion was inhibited (Figure 4a). These findings provide evidence for TIL anergy and confirm that a similar exhaustion mechanism is observed in CD4+ T cells and CD8+ T cells.3 Conversely, the majority of CXCL family chemokines, in addition to IL-5 and M-CSF, demonstrated enhanced secretion in the co-culture group (Figure 4b). One or a combination of chemokines may play a crucial role in the mechanism of CD4+ T cell exhaustion. Furthermore, the altered composition of the tumor microenvironment can result in uncontrolled influx/polarization of immune cells, which can enhance immunosuppression.

IFN-γ plays significant, but not essential role in PD-L1 expression on CD4+ T cells
IFN-γ is a key signaling molecule that induces PD-L1 expression, primarily on tumor cells but also on T cells. In order to investigate the mechanism of PD-L1 expression in this experiment, it was first necessary to confirm whether PD-L1 could be induced on tumor cell lines by exogenous addition of IFN-γ. Numerous studies have shown that IFN-γ upregulates both MHC class I and MHC class II expression, with the latter being particularly relevant due to the dependence of CD4+ T cells on MHC-II. Therefore, our experiments were designed to evaluate the impact of IFN-γ on these molecules.
To determine the response of the CT26 and MC38 tumor lines to IFN-γ treatment, cancer cells were treated with 5 ng/ml IFN-γ. Following a 24-hour period, the expression of PD-L1, MHC class I, and MHC class II on the surface of cancer cells was evaluated. Both cell lines responded to the treatment, exhibiting an increase in the expression of MHC class I markers, H-2Kd for the CT26 line and H-2Kb for the MC38 line (Figure 5a) and PD-L1 expression (Figure 5b). No alterations were observed in MHC class II expression (Figure 5c). The analysis of MHC I and PD-L1 expression in cancer cell lines following IFN-γ treatment confirms their proper response to IFN-γ stimulation. Notwithstanding the fact that differences in T cell responses to cancer may be influenced by varying levels of MHC II expression, these experiments indicate that CD4+ T cells can mediate cancer elimination independently of MHC II expression, due to lack MHC II induction upon IFN-γ exposure.

The exhaustion of persistently stimulated CD4+ T cells was adapted for the purpose of identifying the influence of IFN-γ on the exhaustion process. C57BL/6 mice with a global knockout of IFN-γ receptors were utilized. CD4+ T cells were isolated from mouse spleens and stimulated continuously, according to a previously established protocol. They were then cultured in the presence or absence of cancer cells, with or without a transwell membrane. The observed trend of mutual growth inhibition of CD4+ T cells and cancer cells in the co-culture group remained present despite the lack of IFN-γ receptors (Figures 5d,e). Persistent stimulation with anti-CD3/CD28 moAb coated beads resulted in upregulation of PD-L1 protein expression in co-culture in comparison to the restimulated group, despite the abrogation of the IFN-γ effect on the cells of the KO (Figure 5f, Supplemental Data Figure S11). Although the majority of studies have focused on cancer cells,29,30 it is proved that IFN-γ can also induce PD-L1 protein expression on CD4+ T cells. Nevertheless, it suggests that this is not the sole factor influencing CD4+ T cell exhaustion.

CT26 and MC38 cancer cell lines can be considered as different types of “hot” tumor models
The CT26 and MC38 mouse cancer cell lines are regarded as “hot” tumors, exhibiting high immunogenicity. Nevertheless, the immunological response varies between these lines (for further details, please refer to Section 2.3). To establish differences in the immune landscape between the CT26 and MC38 cell lines, tumors were harvested when they reached the desired size (~150, 500, or 1.500 mm3) and TILs were isolated. Flow cytometry analysis focused on CD3+ T cells (percentage of parent – live/CD45+ subset), CD4+ T cells (percentage of parent – live/CD45+/CD3+ subset), and CD8+ T cells (percentage of parent – live/CD45+/CD3+ subset). The number of CD3+ T cells was significantly higher in small tumors isolated from the MC38 TME than in those isolated from the CT26 TME. However, in medium and large tumors, the distribution of CD3+ T cells between the two cell lines was less diverse. Concurrently, a greater proportion of cells from the MC38 TME exhibited late activation and activation (CD69+, CD25+) and effector memory (CD44+) phenotypes; however, this expression was downregulated in large tumors. Additionally, a greater proportion of T cells exhibited expression of the checkpoint molecules CTLA-4 and PD-L1, and immunosuppressive IDO1. This was more pronounced in MC38 TILs than in CT26 TILs, irrespective of the tumor size (Figure 6a, Supplemental Data Fig S13). Moreover, the observed strength of staining, expressed as MFI, indicated a more robust immune response of TILs to the presence of MC38 cells, with an accelerated expression of exhaustion markers (PD-L1 and CTLA-4) on both CD4+ and CD8+ T cells (Figure 6b). The MC38 tumor TME exhibited higher levels of FoxP3 and IDO1 markers within TIL populations than the CT26 tumor TME. These findings collectively suggest that the overall immune pattern in both cell lines is similar, yet TILs in the MC38 TME are more susceptible to the exhaustion process and the immunosuppressive effects are accelerated. The conclusion is primarily based on data derived from small- and medium-sized tumors, whereas larger tumors do not exhibit the same degree of clustering. This is presumably attributable to partial or complete dysregulation of the immune system in the TME of aged tumors.

It is worth noting that the pattern of TILs activation and exhaustion within CD4+ and CD8+ T cells remains similar, yet the dynamics differ between these two populations, as well as between tumor cell lines. The expression of all important markers, including activation markers (CD69), metabolic suppression (IDO1), and exhaustion state markers (CTLA-4, PD-1, and PD-L1), was markedly elevated in CD8+ T cells relative to that on CD4+ T cells during the early stages of tumor development. This expression is further increased as the tumor grows, potentially influencing the overall anti-tumor response. Interestingly, FoxP3, a Treg marker, was detected in the CD4+ T cell subset in tumors driven by MC38 cells. This finding suggests a stronger immunosuppressive tumor microenvironment in this context.

Discussion
Lymphocyte dysfunction, also known as exhaustion, represents a significant challenge in modern medicine. The defining characteristic of this phenomenon is the overexpression of immune checkpoint molecules – such as PD-1, TIM-3, LAG-3 or CTLA-4 - leading to diminished effector function and proliferative capacity of T cells. Checkpoint inhibitors (CPIs) have emerged as a promising therapeutic strategy, effectively blocking these inhibitory pathways and reinvigorating T cell-mediated anti-tumor responses.31 However, first- or second-generation CPIs, or combinations thereof, are mostly used to rejuvenate cytotoxic CD8+ T cells, with the majority of functionality and toxicity tests conducted in mice.32,33 It should be noted that CD4+ T cells can also become exhausted.17 Our findings suggest that prolonged stimulation of mouse CD4+ T cells may result in their exhaustion, and this process is significantly accelerated by the presence of tumor cells (Figure 1), revealing an equilibrium where neither population achieves complete dominance at that moment. This phenomenon aligns with other studies, where CD4+ T cells initially exhibit potent cytotoxic activity, but progressively develop functional exhaustion marked by reduced proliferative capacity.34,35 It can be related to the high expression of inhibitory receptors on the surface of T cells, and TME which induce tolerance to the presence of tumors and leads to T cells exclusion.36 On the other hand, tumor cells have not yet succeeded in evading the attack of immune cells, which exhibit an effector phenotype and secrete an appropriate set of cytokines into the TME.37 Potentially, also non-immunological mechanisms may contribute to T cell dysfunction. Among these, metabolic competition, nutrient depletion, and hypoxia represent critical components of the immunosuppressive tumor microenvironment (TME). It is noteworthy that our co-culture conditions resulted in rapid acidification of the medium, suggesting the onset of anaerobic metabolism by tumor cells and/or immune cells under stress. This phenomenon of increased acidity is consistent with the presence of a hypoxic, high-lactate microenvironment, which is commonly observed in rapidly growing tumors.38,39 Our findings indicate a significant role for the tryptophan – IDO1–kynurenine axis in the mediation of T cell exhaustion. Tryptophan starvation has been observed to impede mTOR signaling, thus prompting the emergence of T cell anergy and apoptosis.40 All those observation can result in mutual cell growth inhibition.
Our previous reports on human CD4+ T cell exhaustion demonstrated the involvement of the PD-1/PD-L1 signaling pathway and overexpression of PD-L1 protein on the surface of CD4+ T cells as major contributors.18
In vitro studies conducted in mouse cell lines (Figure 2) and in vivo studies performed on a mouse model (Figure 3) have shown that PD-L1 protein overexpression is also noted on mouse CD4+ T cells. This is consistent with the results of previous studies on human cells. The involvement of this protein in the process of tumor-induced exhaustion was demonstrated mostly by the intensity of PD-L1 receptor expression on the surface of CD4+ T cells cultured in the presence of cancer cells. This phenomenon occurs not only during cell-to-cell contact, but is also mediated by factors secreted into the microenvironment, suggesting that this signaling is paracrine-dependent. The existence of clinical data from patients with extranodal natural killer/T‐cell lymphoma confirm, that exhausted CD4+ T cells can express PD-L1 on their surface, even on mononuclear cells isolated from peripheral blood (PBMCs), which can serve as a potential biomarker.41 It is noteworthy that in the co-culture groups, along with the progressive exhaustion of CD4+ T cells, there was no increase in the expression level of the PD-1 receptor, which is considered to be an important prognostic marker in CD8+ T cells associated with exhaustion.42–44 Zhao et al. discussed how PD-1+ CD4+ T cells within the tumor microenvironment exhibit signs of exhaustion, noting that while PD-1 expression may initially be transient during activation, prolonged antigen exposure leads to its sustained upregulation, resulting in a dysfunctional T cell phenotype.45 This is supported by our in vivo findings, which demonstrate the initiation of PD-1 expression. It suggests that PD-1 may not be the principal mediator responsible for exhaustion in CD4+ T cells. The function of the PD-L2 protein also remains unclear. No significant upregulation of PD-L2 expression was observed in the co-culture; on the contrary, a lower percentage of cells showed positive expression of this marker (Figure 2). It has been demonstrated that PD-L2 protein can be expressed on activated T cells (both CD4+ and CD8+), and its blocking or downregulation can lead to negative regulation of T cells by decreasing the levels of secreted cytokines such as IL-10 or IFN-γ.46 Furthermore, evidence indicates that PD-L2 demonstrates preferential expression in Th2-polarized CD4+ subsets, accompanied by elevated IL-4 cytokine production.47 In contrast, our Th1-skewed co-culture conditions (IL-2 supplementation) appear to exhibit a capacity for suppressing PD-L2 induction. Therefore, this regulatory process may have influenced the levels of soluble factors in our cytokine assay (Figure 4), which also demonstrated a reduction in the levels of the aforementioned cytokines. The investigation of LAG-3 and TIM-3 expression on exhausted CD4+ T cells also resulted in no changes being observed. While these markers are well-established within the context of CD8+ T cell exhaustion,5 it is conceivable that CD4+ T cells may employ distinct inhibitory pathways. Our findings concur with those of previous studies, which demonstrated that CD4+ exhaustion is frequently characterized by PD-L1/CTLA-4, as opposed to TIM-3/LAG-3.17 This is evidenced by the elevated expression of CTLA-4 observed in both of our experimental models. The upregulation of CTLA-4 in response to persistent TCR signaling serves to prevent autoimmunity.48 Consequently, this heightened expression can lead to a state of T cell exhaustion and functional inhibition. Furthermore, the phenomenon of metabolic immunosuppression is characterized by the presence of an overexpressed IDO1 protein, a process that is driven by the accumulation of both T cell and tumor-derived metabolites.49,50 A seminal study has identified the kynurenine pathway as a direct cause of T cell exhaustion, including CD4+ T cells, highlighting the significance of this pathway in immune-related diseases.51,52 It is noteworthy that the findings, with the exception of those related to CTLA-4 and IDO1, are confined to the TME, thereby indicating that T cell exhaustion is TME-specific (Figure 3). Splenic T cell exhaustion markers, including PD-L1, PD-1, TIM-3, and LAG-3, demonstrate stability. In contrast, CTLA-4 and IDO1 demonstrate significantly higher expressions in the spleen among medium and large tumor groups of tumor-bearing mice (Figure 3). The potential causes of this phenomenon include the development of systemic immune activation by tumor-derived factors (such as kynurenine metabolites), expansion of Tregs, or the secretion of soluble mediators by the tumor, including exosome.26,53,54
The cytokine secretion assay provides robust confirmation of the state of exhaustion among CD4+ T cells, as illustrated in Figure 4. Significantly reduced expression levels of IFN-γ, TNFα, and IL-2 are distinctive indicators of this condition.55 Concurrently, GM-CSF is downregulated, which typically promotes the anti-tumor response through the activation and recruitment of macrophages, dendritic cells, or neutrophils,56 as well as a reduction in the level of the chemoattractant CCL3, which plays analogous roles.57 Conversely, elevated secretion of CXC-group chemokines, including CXCL1, CXCL2, CXCL9, and CXCL10, has been observed, suggesting the promotion of tumor growth, enhanced angiogenesis, and immunosuppression.58,59 Furthermore, TIMP1 protein may serve as a prognostic indicator in the context of cancer progression,60 indicating that its presence in samples containing tumor cells is associated with an unfavorable prognosis. Additionally, it is shown that the overexpression of CCL5 (RANTES) in combination with IL-6 may result in an increase in breast cancer aggressiveness.61 Although the cytokine IL-6 was not overrepresented in the supernatant from the mouse experiment, previous studies on human material have confirmed high levels of transcription of this gene, suggesting that a similar relationship can be assumed here as well. These observations suggest that tumor-CD4+ T cell homeostasis is undergoing a shift, which favors the tumor. This indicates that T cell function may be impaired as a result of exhaustion.
Several studies have shown that IFN-γ can induce and enhance both MHC class I and MHC class II expression.62,63 In the context of anticancer functions, this is significant because high expression of MHC class II regulates and activates MHC class II-dependent cells, including CD4+ T cells.64 Furthermore, it has been demonstrated that IFN-γ induces PD-L1 receptor expression on cancer cells (Figure 5), while IFN-γ receptor 1 (IFNGR1) depletion results in reduced PD-L1 expression.65 The results of these studies imply that the expression of PD-L1 is reliant upon the secretion and availability of IFN-γ within the TME. This may render the results obtained in our experiments particularly noteworthy. Our findings indicate that IFN-γ has an activating effect on MHC class I and PD-L1, yet the level of MHC class II expression on CT26 and MC38 tumor cells remained unchanged upon IFN-γ stimulation (Figure 5). The underlying factors contributing to this finding include post-transcriptional modifications, such as ubiquitination and phosphorylation of the primary MHC class II activator (CIITA), as well as constrained expression in reaction to signaling pathways like MEK/ERK.66,67 Furthermore, knockout of the IFN-γ receptor in mice from which CD4+ T cells were obtained for the in vitro experiment did not result in the complete inhibition of PD-L1 expression on the surface of CD4+ T cells. These results suggest that PD-L1 expression on CD4+ T cells is not regulated solely by the IFN-γ-dependent signaling pathway, despite the well-established role of this cytokine in T cell exhaustion. It is likely that PD-L1 expression on CD4+ T cells is controlled by other factors secreted by tumor cells or other TME components, such as TGF-β or VEGF68,69 as well as by intracellular mechanisms, such as metabolic changes or deregulated equilibrium of epigenetic factors.70–72
The observed differences in the immune response of CD4+ T cells to the presence of CT26 and MC38 cells prompted us to analyze the immune profiles of these tumors (Figure 6). Most studies have indicated that both tumor lines are classified as “hot/inflamed tumors”, which are defined by high lymphocytic infiltration (TILs) and a favorable response to immunotherapy. However, the MC38 cell line can also be categorized as an “intermediate tumor”.73 Zhong et al. 2020 reported that, in comparison to MC38, the CT26 cell line is distinguished by markedly elevated infiltration of CD4+ and CD8+ TILs, enhanced cytolytic activity, and increased immunosuppression, which is characterized by high expression of IDO1, FoxP3, and TGFB1.74 In contrast, it has also been documented that the proportion of CD3+ T cells (comprised of both CD4+ and CD8+ T cells) within MC38 tumors was found to be approximately threefold higher in untreated mice when compared to those observed in CT26 tumors.75 Further, Kristensen et al. observed that while both CT26 and MC38 tumors are considered “hot” due to their substantial infiltration of CD4+ and CD8+ T cells, MC38 tumors had a greater overall abundance of these T cell subsets compared to CT26 tumors. Although CT26 retains features of a “hot” tumor, its immune microenvironment may be more immunosuppressive due to the influence of regulatory T cells or inhibitory cytokines.76 These conflicting outcomes may be attributed to discrepancies in cell culture conditions or methods of mouse housing and diet. Previous studies have demonstrated that dietary intake can influence the immune response to colorectal cancer, affecting tumor growth and T cell infiltration.77 Our findings support the hypotheses that the MC38 cell line exhibits an augmented immune response, characterized by moderately elevated CD3+CD4negCD8+ infiltration, compared to the CT26 line. Most crucially, the MC38 cell line exhibits a distinctive immunological profile, with infiltrating CD4+ and CD8+ TILs displaying accelerated activation and effector functions (CD25high and CD44high). Concurrently, these cells demonstrate an earlier transition to exhaustion (PD-1high, PD-L1high, and CTLA4high in small tumors) and immunosuppression (FoxP3mod and IDO1high in small tumors).
The observed variability between MC38 in C57BL/6 mice and CT26 in BALB/c mice may be due to differences in both the cancer model and the specific mouse strains used. C57BL/6 mice have been observed to manifest pronounced Th1 and Th17 immune responses, as evidenced by an augmented production of pro-inflammatory cytokines, such as IL-12 and IFN-γ, which play a pivotal role in counteracting intracellular pathogens and tumors.78 In contrast, BALB/c mice tend to exhibit a Th2-dominated response, leading to higher levels of IL-4, IL-5, and IL-13 cytokines.79 Those differences can affect TIL response for tumor presence with more protective profile of C57BL/6 and increased susceptibility of BALB/c mouse strain.80 The collective influence of these factors illuminates the distinguishing characteristics of the MC38 + C57BL/6 and CT26 + BALB/c models. Specifically, the MC38 tumors demonstrate higher initial TIL infiltration, followed by accelerated exhaustion, while the CT26 tumors exhibit lower infiltration, resulting in a more gradual immune suppression. Importantly, this may indicate a potential opportunity for the early application of CPIs in C57BL/6 mice. However, it is important to recognize that it is relatively easy to miss the moment when targeted therapy may be effective due to the complex network of signaling pathways that can lead to the exhaustion of TILs.
A thorough review of the data presented on heatmaps with TILs from diverse tumors reveals both similarities and dissimilarities between the activation and exhaustion of CD4+ and CD8+ T cells (Table 2). In both populations, markers of activation (CD69) and effector and central memory function (CD44 and CD62L) are observed to decrease with tumor growth. Consequently, cytotoxicity and T cell efficacy are found to be high only in the early stages of the tumor, while as the tumor grows, immune function is reduced, resulting in a diminished anti-tumor response. Conversely, a significantly high proportion of both T cell populations with elevated expression of the immunosuppressive protein IDO1 and checkpoint inhibitor molecules (PD-L1 and CTLA-4) was observed in small tumors. This effect is further amplified as the tumor grows. The only exception is cells expressing the PD-1 marker, which show a decrease in percentage within a medium-sized tumor but then reappear in large tumors. These findings lead to functional impairment of both immune cell populations, thereby enhancing exhaustion with tumor growth. The aforementioned changes occur with greater dynamism within the CD8+ population than within the CD4+ population, resulting in a more expeditious shutdown of cytotoxic cell function and, consequently, rendering CD8+ T cells unable to perform their intended function. Notably, PD-L1 receptor overexpression is a significant factor in both T cell populations. Such overexpression suggests the existence of a mechanism that suppresses T cell functionality. Therefore, exhaustion is generated not only by PD-L1-positive tumor cells but also by CD4+ and CD8+ T cells that express PD-L1 on their surface. Specifically, a greater proportion of PD-L1-positive T cells interacts with PD-1-positive CD4+ T cells, resulting in negative feedback, leading to their exclusion and immunosuppression development. This hypothesis is supported by studies demonstrating, among other findings, that insufficient CD4+ T cell infiltration is correlated with a worse prognosis among patients with sarcoma and neuroblastoma.81,82 Additionally, FoxP3-positive Tregs were observed exclusively in tumors derived from MC38 cells, with their number increasing in larger tumors. This observation suggests a more immunosuppressive TME in MC38-derived tumors.

Concluding remarks
Despite extensive research, the precise mechanisms driving T cell exhaustion remain unclear. Our findings enhance the understanding of this process in mouse CD4+ T cells, which have been studied less extensively than their CD8+ counterparts. Persistently activated lymphocytes exhibit high PD-L1 expression, a trait further amplified by the presence of tumor cells. Unlike other exhaustion markers, this upregulation was not observed in CD8+ T cells, highlighting a distinct difference between the two populations.
Additionally, cancer alters the composition of the tumor microenvironment (TME), prompting exhausted CD4+ T cells (and/or tumor cells) to secrete a unique set of cytokines. Notably, these changes occur even in the absence of direct cell-to-cell contact. Our findings confirm that a mouse model of immune exhaustion is immunologically analogous to a previously established human model, providing valuable insights for evaluating therapeutic strategies aimed at reversing exhaustion. In particular, the use of PD-L1 inhibitors presents a promising avenue for further investigation.
Although both the BALB/c + CT26 and C57BL/6 + MC38 mouse models exhibited similar overall immune responses to tumor development, subtle differences were observed, including an accelerated T cell exhaustion phenotype within the MC38 TME. These findings suggest that PD-L1 expression on CD4+ T cells is highly conserved across species. However, given the unique characteristics of different tumor types and host environments, careful selection of research models remains crucial.

Supplementary Material

Supporting Data Values.xlsx

Sup Data Fig 13.jpg

Sup Data Fig 12.jpg

Supplemental Data captions and legends.docx

Sup Data Fig 4.jpg

Sup Data Fig 1.jpg

Sup Data Fig 7.jpg

Sup Data Fig 8.jpg

Sup Data Fig 2.jpg

Sup Data Fig 3.jpg

Sup Data Fig 9.jpg

Sup Data Fig 5.jpg

Sup Data Fig 6.jpg

Sup Data Fig 10.jpg

Sup Data Fig 11.jpg