Targeting the glucocorticoid receptor-CCR8 axis mediated bone marrow T cell sequestration enhances infiltration of anti-tumor T cells in intracranial cancers

Abstract

Brain tumors such as glioblastomas are resistant to immune checkpoint blockade therapy, largely due to limited T cell infiltration in the tumors. Here, we show that mice bearing intracranial tumors exhibit systemic immunosuppression and T cell sequestration in bone marrow, leading to reduced T cell infiltration in brain tumors. Elevated plasma corticosterone drives the T cell sequestration via glucocorticoid receptors in tumor-bearing mice. Immunosuppression mediated by glucocorticoid-induced T cell dynamics and the subsequent tumor growth promotion can be abrogated by adrenalectomy, the administration of glucocorticoid activation inhibitors or glucocorticoid receptor antagonists, and in mice with T cell-specific deletion of glucocorticoid receptor. CCR8 expression in T cells is increased in tumor-bearing mice in a glucocorticoid receptor-dependent manner. Additionally, chemokines CCL1 and CCL8, the ligands for CCR8, are highly expressed in bone marrow immune cells in tumor-bearing mice to recruit T cells. These findings suggested that brain tumor-induced glucocorticoid surge and CCR8 upregulation in T cells lead to T cell sequestration in bone marrow, impairing the anti-tumor immune response. Targeting the glucocorticoid receptor-CCR8 axis may offer a promising immunotherapeutic approach for the treatment of intracranial tumors.

Introduction

Brain tumors include a wide variety of malignancies that start either within the brain, such as glioblastoma (GBM) (accounting for approximately 50% of primary brain malignancies in adults) [1] or brain metastasis (BrM) from other primary sites (constituting about 90% of brain malignancies). BrM commonly originates from melanoma, lung cancer, or breast tumors [2]. GBM is considered as one of the deadliest forms of brain cancer, with a recurrence rate of over 90% [3]. On average, patients diagnosed with GBM have a survival time of less than 15 months [4], even with conventional therapies such as surgical removal followed by radiation and chemotherapy using the drug temozolomide (TMZ). Additionally, GBM often leads to a decrease in systemic T cell numbers and function, making it less responsive to immune checkpoint blockade therapy [5]. Conversely, a significant number of T cells are found to be sequestered in the bone marrow (BM) of patients with brain tumors, accompanied by the tumor-induced loss of sphingosine 1 phosphate receptor 1 (S1P1) from the T-cell surface [6, 7].

Individuals with GBM and brain metastases are often complicated with peritumoral vasogenic edema, which is a significant contributor to the high morbidity and mortality associated with neurological tumors [8]. To date, the synthetic corticosteroid dexamethasone (DEX) is the preferred drug for treating edema in neurosurgical and neuro-oncological cases [9]. Although clinically effective in managing vasogenic edema, the administration of DEX has well documented to be associated with a range of systemic side effects, including reduced survival rates when used concurrently with radiotherapy and TMZ treatment [10]. Moreover, patients may naturally produce elevated levels of glucocorticoid, potentially leading to the failure of immunotherapy [11]. High levels of glucocorticoid in the tumor microenvironment lead to dysfunction of tumor-infiltrating lymphocytes (TIL) and promote a depleted phenotype [11, 12]. Similarly, in the diet-restricted mouse models, elevated serum glucocorticoid levels induced the accumulation of effector memory T cells in the BM [13]. However, whether glucocorticoid mediate T cell accumulation in BM in subjects with intracranial tumors remains to be determined.

In this study, we show that an increase in glucocorticoid caused by brain tumors leads to an enhanced expression of CCR8 on T cells, which in turn results in the sequestration of these cells in the BM, impairing the immune defense against the brain tumor. These new findings suggest that the glucocorticoid receptor (GR)-CCR8 axis may be targeted to achieve a more effective immunotherapy when treating intracranial malignancies.

Results

Intracranial tumors cause systemic immunosuppression in mice

To investigate the impact of intracranial tumors on the functions of thymus and spleen, we established a murine glioma model by intracranially (IC) injecting GL261 murine glioma cells into the brains of C57BL/6 mice. After the tumors becoming sizeable (around day 18), we analyzed the thymus and spleen. Both organs contracted and a decrease in the numbers of thymocytes at DN1, DN2, DN3, DN4, DP, SP4, and SP8 stages in tumor-bearing mice (Fig. 1A–D). In the meantime, significant T-cell lymphopenia was observed in the CD4 and CD8 compartments of spleen and peripheral blood (Fig. 1E–G). We also investigated whether the reduction of T cells in the thymus and spleen is specific to the glioma or intracranial environment by injecting AT3 breast carcinoma cells intracranially into C57BL/6 mice and the numbers of T cell in thymic, splenic and peripheral blood were assessed. Similarly, breast cancer in the brain led to a significant decrease of T cells in thymus, spleen and peripheral blood (Fig. 1H–L). These above results suggested that both primary and metastatic intracranial tumors can cause systemic immunosuppression.

Fig. 1.

Intracranial tumors cause systemic immunosuppression in mice. A–G GL261 cells (1 × 10⁴) were stereo-tactically injected into C57BL/6 mice. The weights of the thymus (A) and spleen (B) were measured in the PBS group (n = 5) and tumor-bearing mice (n = 6). mean ± SEM. Flow cytometric analysis and quantification of thymocytes at the DN (CD4^-CD8^-), DP (CD4⁺CD8⁺), SP4 (CD4⁺CD8^-) and SP8 (CD4^-CD8⁺) stages of T cells (C) and DN1 (CD44⁺CD25^-), DN2 (CD44⁺CD25⁺), DN3 (CD44^-CD25⁺) and DN4 (CD44^-CD25^-) subpopulations in thymus (D) were performed, in the PBS group (n = 5) and tumor-bearing mice (n = 6). mean ± SEM. Flow cytometric analysis was conducted to quantify of total T cells, CD4⁺ T cells, and CD8⁺ T cells in spleen (E) and blood (F, G) in the PBS group (n = 5) and tumor-bearing mice (n = 6). mean ± SEM. H–L AT3 cells (1 × 10⁴) were injected stereo-tactically into C57BL/6 mice. The weights of the thymus (H) and spleen (I) in PBS group and tumor-bearing mice. mean ± SEM, n = 4. J Flow cytometric analysis and quantification of DN, DP, SP4 and SP8 thymocytes, in PBS group and tumor-bearing mice. mean ± SEM, n = 6. Flow cytometric analysis was conducted to quantify the total T cells, CD4⁺ T cells, and CD8⁺ T cells in spleen (K) and blood (L) in PBS group and tumor-bearing mice. mean ± SEM, n = 4. P values were calculated using two-tailed unpaired t tests

T cells accumulate in the BM of mice with intracranial tumors

In comparison to other compartments examined, both CD4⁺ and CD8⁺ T cells exhibited a significant increase in the BM of mice bearing either GL261 or AT3 tumors (Fig. 2A, B). This accumulation of immune cells in the BM was specific to T cells, since there was no increase in other cell types such as neutrophils, monocytes, macrophages, and B cells (Fig. 2C). When CD4⁺ T cells in BM were induced to differentiate into different subsets (Th1, Th2, Th17 or Treg), T cells from tumor bearing-mice gave rise to more Th1 and Th2 cells than those in control (Fig. 2D). Both naïve and central memory T cells was increasingly accumulated in the BM of tumor-bearing mice for the two types of tumor cells, although there was no significant change in the numbers of effector T cells (Fig. 2E–G). When GL261 glioma cells were subcutaneously injected into C57BL/6 mice, there were no detectable changes in the counts of T cells in spleen (Fig. S2A). Furthermore, the populations of T cells, neutrophils, monocytes, macrophages, and B cells within the BM remained unaltered (Fig. S2A–C), indicating that the intracranial localization is required for tumor cells to induce T cell sequestration in BM.

Fig. 2.

T cells accumulate in the BM of mice with intracranial tumors. A, B, D–G GL261 cells or AT3 cells (1 × 10⁴) were stereotactically inoculated into mice. A Flow cytometric analysis and quantification of total T cells, CD4⁺ T cells, and CD8⁺ T cells in the BM of the following groups: PBS group (n = 5), GL261-bearing mice (n = 7), and AT3-bearing mice (n = 5). mean ± SEM. B Images of BM from tumor bearing mice for 18 days, showing CD3 (red) and hoechst (blue). C Flow cytometric analysis and quantification of neutrophils, monocytes, macrophages and B cells in BM in mice subcutaneously injection of PBS and GL261. mean ± SEM, n = 4. D Subsets of CD4⁺ T cells in the BM of mice from the PBS group (n = 5), GL261-bearing mice (n = 7), or AT3-bearing mice (n = 5). mean ± SEM. E–G CD4⁺ and CD8⁺ naïve T cells (Tn), central memory T cells (Tm), and effector T cells (Te) in the BM of mice from the PBS group (n = 5), GL261-bearing mice (n = 6), or AT3-bearing mice (n = 5). mean ± SEM. P values were calculated by two-tailed unpaired t tests

T cell redistribution caused by intracranial tumor is mediated by glucocorticoid

The accumulation of T cells in BM indicated that the BM might have provided T cells a survival advantage in the presence of brain tumor. Our next aim was to elucidate whether cortisol plays an essential role as a soluble mediator in the migration of intracranial tumor-induced T cell sequestration in BM. We found that the level of glucocorticoid was elevated in the blood of tumor-bearing mice when compared to those injected with PBS (Fig. 3A). Whereas the concentration of glucocorticoids remained unchanged in BM of intracranial GL261-bearing mice, it was increased in AT3-bearing mice (Fig. 3B). Heightened levels of glucocorticoid are known to promote T cell death [14]. While the baseline frequency of dead (7-Aminoactinomycin D⁺, 7-AAD⁺) T cells was higher in BM than in spleen in the control mice, this trend was reversed in tumor-bearing mice (Fig. 3C, D). These data suggested that brain tumor could remold this compartment to promote T cell survival. T cells in the BM also expressed higher levels of the anti-apoptotic protein BCL-2 in the presence of brain tumor (Fig. 3E, F). These data suggest that BM is conducive to T cell survival.

Fig. 3.

T cell redistribution caused by intracranial tumors is mediated by glucocorticoid. A Corticosterone concentrations in plasma in mice of PBS group (n = 5), GL261group (n = 5), and AT3 group (n = 5). mean ± SEM. B Corticosterone concentrations in BM of the mice in PBS (n = 8), GL261 (n = 4) and AT3 group (n = 5). mean ± SEM. C Frequency of 7-AAD⁺ cells in spleen and BM of mice with PBS (n = 5) or GL261 (n = 7) for 18 days, mean ± SEM. D Frequency of 7-AAD⁺ cells in spleen and BM of mice with PBS (n = 5) or AT3 (n = 4) for 18 days, mean ± SEM. E BCL-2 expression by BM T cells in mice with PBS or GL261 for 18 days, mean ± SEM, n = 4. F BCL-2 expression in BM T cells in mice with PBS or AT3 for 18 days, mean ± SEM, n = 5. G–J Adrenalectomy reverses the accumulation of T cells in the BM of tumor-bearing mice. Two weeks post-adrenalectomy, mice were intracranially injected with 1 × 10⁴ GL261 cells. Flow cytometric analysis and quantification of total T cells, CD4⁺ T cells and CD8⁺ T cells in spleen (G) and BM (H) of SO mice with PBS (n = 6), SO mice with tumor (n = 6), ADX mice with PBS (n = 7) and ADX mice with tumor (n = 7). mean ± SEM. I Gross image (left) and weight (right) of tumors in sham operation group (SO) (n = 6) or adrenalectomized group (ADX) (n = 7). mean ± SEM. J Flow cytometric analysis and quantification of total T cells, CD4⁺ T cells and CD8⁺ T cells in tumors from the SO group (n = 6) and ADX group (n = 7). mean ± SEM. K–N Inhibiting active corticosterone synthesis reverses the accumulation of T-cells in the BM of tumor-bearing mice. Mice were given drinking water supplemented with 800 μg/mL metyrapone (MTY) and were intracranially inoculated with 1 × 10⁴ GL261 cells the following day. Flow cytometric analysis and quantification of total T cells, CD4⁺ T cells and CD8⁺ T cells in spleen (K) and BM (L) in DMSO treated mice with PBS injection (n = 5), DMSO treated mice with tumor injection (n = 8), MTY treated mice with PBS injection (n = 5) or MTY treated mice with tumor injection (n = 8). mean ± SEM. M Gross image (top) and weight (bottom) of brain of tumor mice in DMSO and MTY groups. mean ± SEM, n = 8. N Flow cytometric analysis and quantification of total T cells, CD4⁺ T cells and CD8⁺ T cells in tumors from the DMSO group (n = 8) and MTY group (n = 8). mean ± SEM. P values in (A, B, I, J, M, N) were determined by two-tailed, unpaired Student’s t test. P values in C-H, K, L were calculated by two-way ANOVA with Tukey’s multiple comparison test

To investigate the role of GC in regulating T cell redistribution during tumorigenesis, we intracranially injected GL261 into mice that had their adrenal glands removed (adrenalectomy, ADX). In this setting, T cells did not subside in spleen, nor did they accumulate in the BM (Fig. 3G, H). Furthermore, we found that tumor growth was inhibited (Fig. 3I) and the numbers of T cells per gram of tumor increased in the ADX group (Fig. 3J). These findings imply that ADX may inhibit T cell accumulation in the BM, improve T cell infiltration into the tumor, and thus restrict tumor growth.

We next enumerated the different types of immune cells to evaluate the effect of glucocorticoid insufficiency on the tumor microenvironment and only found a modest increase in the population of monocytes in the tumors of the ADX group (Fig. S3A). At the same time, examination of the immune cell infiltration and microglia in the brain tissue following tumor removal revealed an increase in the number of immune cells increased in the ADX group (Fig. S3B, C). Importantly, while brain tumors resulted in a significant decrease in the numbers of CD4⁺ and CD8⁺ T cells in the brains, the CD4⁺ T cells were greatly restored in ADX mice (Fig. S3D, E). Compared to healthy control mice, tumor-bearing mice had higher dendritic cell infiltration in their paraneoplastic tissues (Fig. S3F), implying that dendritic cells may be in high demand as antigen-presenting cells during glioma development. Moreover, the expression level of MHC-II in dendritic cells was increased in para-cancerous tissue compared with healthy brain tissue and the deprivation of GC further increased its expression level (Fig. S3G). The expression level of CD83, a marker of mature dendritic cells, was significantly decreased (Fig. S3H). There were no significant changes in the expression levels of dendritic cell activation markers CD80 and CD86 (Fig. I, J). These findings imply that ADX may not change the maturation and activation status of dendritic cells, but could improve their antigen presenting ability of dendritic cells. These findings imply that ADX may help keep the immune milieu intact, and enable the mice to mount a normal T cell response to contain the tumor progression.

Although steroids are mainly synthesized in the adrenal cortex, it has been shown that tumor cells and the monocyte-macrophage lineage cells are capable of extraadrenal steroidogenesis, thereby inducing T cell dysfunction and promoting tumor growth [15]. Therefore, we treated tumor-bearing mice with metyrapone, an inhibitor of 11beta-hydroxysteroid dehydrogenase (11beta-HSD), to block GC activation. As expected, metyrapone administration offset T cell decline in the spleen and accumulation in BM (Fig. 3K, L). Importantly, metyrapone significantly inhibited tumor growth (Fig. 3M). However, unlike in ADX mice, the numbers of tumor-infiltrating T cells did not increase in the metyrapone group (Fig. 3N), presumably due to defective GC release from thymus when 11beta-HSD is inhibited.

GR signaling is essential for T cell sequestration in BM

Glucocorticoid function through glucocorticoid receptor (GR) encoded by Nr3c1. Importantly, we found that the effect of glucocorticoid-induced T cell sequestration in BM was abolished by the administration of the antagonist of glucocorticoid receptor, mifepristone (RU486) and AL 082D06 (D06, an antagonist with higher specificity to GR) (Fig. 4A, B). Consistently, D06 injections inhibited tumor growth and enhanced the infiltration of T cells in tumor (Fig. 4C–E), though the levels of activation markers (TNF-α and IFN-γ) and exhaustion markers (PD-1 and Tim-3) of T cells remained unchanged after treatment with GR inhibitor (Fig. 4F, G). To determine the target cells of GR action, we implanted GL261 cells into nude mice, which lack T cells, and assessed tumor growth (Fig. 4H). The tumor suppressive effect of D06 was eliminated in nude mice (Fig. 4I, J), indicating that it is the GR action on T cells that promoted tumor growth.

Fig. 4.

GR inhibitor, D06, suppresses T cell sequestration in the BM and inhibits tumor growth. A–G GR inhibitors inhibit the accumulation of T-cells in the BM of tumor-bearing mice. GL261 cells (1 × 10⁴) were intracranially inoculated into mice and then treated with or without mifepristone (RU486) (5 mg/kg) or AL 082D06 (D06) (5 mg/kg) every other day for nine cycles (A). B Flow cytometric analysis was conducted to quantify the total T cells, CD4⁺ T cells, and CD8⁺ T cells in the BM. The study included mice treated with PBS (n = 5), mice treated with RU486 (n = 5), mice treated with D06 (n = 4), tumor-bearing mice treated with PBS (n = 4), tumor-bearing mice treated with RU486 (n = 4), and tumor-bearing mice treated with D06 (n = 5). mean ± SEM. C Bioluminescent images were collected on day 15 after GL261 implantation and treatment with oil (n = 4), RU486 (n = 3), or D06 (n = 5). mean ± SEM. D Weight of tumor in the oil group (n = 4), RU486 group (n = 3), or D06 group (n = 5). mean ± SEM. E Flow cytometric analysis was conducted to quantify the total T cells, CD4⁺ T cells, and CD8⁺ T cells in the tumor. The study included the oil group (n = 4), RU486 group (n = 3), and D06 group (n = 5). mean ± SEM. F Flow cytometric examination of the percentage of tumor T cells positive for IFN-γ, and TNF-α in oil (n = 6), RU486 (n = 6), and D06 group (n = 5). mean ± SEM. G Flow cytometric examination of the percentage of PD-1^-Tim3^-, PD-1⁺Tim3^-, or PD-1⁺Tim3⁺ tumor CD8⁺ T cells in the oil (n = 6), RU486 (n = 6), or D06 group (n = 5). mean ± SEM. H–J GR inhibitors D06 suppress the growth of GL261 by relying on T cells. GL261 cells (1 × 10⁴) were intracranially inoculated into nude mice and treated with or without AL 082D06 (D06) (5 mg/kg) every other day for nine cycles (H). I Gross image (left) and weight (right) of tumors in the oil group (n = 5) or D06 group (n = 7). mean ± SEM. J Bioluminescent images were collected on day 7 and day 14 after GL261 implantation and treatment with oil (n = 6) and D06 (n = 5). mean ± SEM. P values in (C–G, I, J) were determined by two-tailed, unpaired Student’s t test. P values in (B) were calculated by ordinary two-way ANOVA with Tukey’s multiple comparison test

To further elucidate the role of GR in T cell redistribution in brain tumor-bearing mice, we generated T cell-specific GR conditional knockout mice by crossing CD4^Cre mice with Nr3c1^fl/fl mice (referred to as GR cKO mice) (Fig. S4A). We confirmed the reduction of GR expression in T cells of GR cKO mice, when compared with Nr3c1^fl/fl mice (referred to as control mice), by conducting qPCR and western blotting experiments (Fig. S4B, C).

To further assess whether the GR signaling in T cells is responsible for their migration to the BM, we intracranially implanted GL261 tumor cells in GR cKO mice and control mice showed a significant reduction in tumor weight 18 days after implantation (Fig. 5A). Subsequently, we found that the number of T cells in the BM of GR cKO mice was not different from that in the PBS group after tumor inoculation (Fig. 5B). Tumor-infiltrating T cells were slightly increased in GR cKO mice (Fig. 5C). The percentage of IFN-γ positive T cells was higher in GR cKO mice, while TNF-α producing cells were not altered appreciably (Fig. 5D). Additionally, the positive rate of programmed cell death-1 (PD-1), an immunosuppressive molecule, was significantly lower in TILs of GR cKO mice (Fig. 5D). Meanwhile, GR cKO mice showed a prolonged survival (Fig. 5E). These findings supported the notion that GR action on T cells promoted their peripheral collapse and BM sequestration in the presence of intracranial cancer.

Fig. 5.

GR signaling is required for T cell accumulation in the bone marrow. A–E GL261 cells (1 × 10⁴) were intracranially implanted into control mice or GR cKO mice. A Gross image (left) and weight (right) of tumors in control and GR cKO mice. mean ± SEM, n = 10. Total T cells, CD4⁺ T cells, and CD8⁺ T cells in the BM (B) and tumor (C) of different groups of mice were quantified using flow cytometric analysis. Control mice were given PBS (n = 5), GR cKO mice (n = 6), control mice with tumor (n = 6), and GR cKO mice with tumor (n = 6). mean ± SEM. D Flow cytometric examination of the percentage of tumor T cells positive for PD-1, IFN-γ, and TNF-α in control mice (n = 10) and the GR cKO group (n = 6). E The survival rate of GL261 tumors in control mice (n = 5) and GR cKO mice (n = 5). Survival in (E) was assessed by a two-tailed generalized Wilcoxon test. P values in (A, C, D) were determined by two-tailed, unpaired Student’s t test. P values in (B) were calculated by ordinary two-way ANOVA with Tukey’s multiple comparison test

CCR8 directs T cells to BM

Since this mobilization of T cells to the BM might be driven by chemokines, we tested whether increased chemokine signals from the BM are responsible for T cell accumulation. We determined the mRNA levels of Cxcl9, Cxcl10, Cxcl11 and Cxcl12 in BM and found them to be generally upregulated during tumor progression (Fig. S5A). However, when we tested the role of the CXCR3-CXCL10/CXCL11 axis in T cell sequestration by applying AMG 487, we observed no difference between the control and treatment groups (Fig. S5B–D).

Given that GR functions as a transcription factor regulating the expression of many genes, it is possible that other downstream target genes are responsible for the observed alteration in T cell distribution. We then examined the expression levels of chemokine receptors related to T cell migration to BM in the tumor-bearing mice. We observed that following tumor establishment, there was an increase in the expression of chemokine receptor 8 (CCR8) on T cells in BM (Fig. 6A). CCR8 is a seven transmembrane G-protein coupled receptor with a high affinity for mouse CCL1 and CCL8 [16]. Ccr8 was further shown to be upregulated by GC in T hybridoma A1.1 cells (Fig. S6). Furthermore, GC-induced elevation of Ccr8 expression in splenic T cells was totally abrogated in GR cKO mice (Fig. 6B). Together, these results clearly demonstrate that GC-induced upregulation of Ccr8 occurred through the GR signaling in T cells. We also isolated BM and determined the expression of Ccl1 and Ccl8, ligands of CCR8, which were reported to be produced by monocytes [17], macrophages [18], T cells [19], endothelial cells [20], and stromal cells [21]. We purified different subsets of immune cells in the BM by flow cytometry-based cell sorting and indeed observed significant upregulations of Ccl1 in monocytes, macrophages and T cells (Fig. 6C), and of Ccl8 in monocytes and macrophages (Fig. 6D).

Fig. 6.

CCR8 directs T cell sequestration in BM. A CCR8 expression by T cells in BM and spleen of mice with PBS (n = 4) or with GL261 (n = 7) inoculation by flow cytometry. Mean ± SEM. B The expression of Ccr8 in the spleen of control or GR cKO mice stimulated with or without dexamethasone (DEX) (10 nmol/L) for 48 h was assayed by qPCR, Mean ± SD, n = 4. The expression of Ccl1 (C) and Ccl8 (D) in the subsets of BM cells of mice with PBS or GL261 was assayed by qPCR. Mean ± SD, n = 4. E–G CCL1-CCR8 inhibitors R243 blocks T-cell accumulation in the BM of tumor-bearing mice. GL261 cells (1 × 10⁴) were inoculated intracranially into mice and treated with or without R243 (5 mg/kg) every day for 18 days (E). F Flow cytometric analysis and quantification of total T cells, CD4⁺ T cells, and CD8⁺ T cells in BM of control mice treated with PBS (n = 5), control mice treated with R243 (n = 5), tumor-bearing mice treated with PBS (n = 7), and tumor-bearing mice treated with R243 (n = 7). Mean ± SEM. G Gross image (left) and weight (right) of tumors in PBS treated mice (n = 7) and R243 treated mice (n = 8). Mean ± SEM. H Flow cytometric analysis and quantification of total T cells, CD4⁺ T cell, CD8⁺ T cells, and Tregs in tumor of mice treated with PBS (n = 4) or mice treated with R243 (n = 4), Mean ± SEM. I Survival rate of tumor-bearing mice treated with PBS (n = 7) or R243 (n = 7). Wilcoxon tests were applied to analyze the Kaplan-Meier survival curves. J ChIP-seq tracks show GR-bound peaks at the Ccr8 locus of splenic Treg cells treated with ethanol or dexamethasone. ChIP-seq data was download in Gene Expression Omnibus (GEO) under the accession code GSE183808. K Chromatin immunoprecipitation (ChIP)-qPCR analysis of the binding sites of GR in the regulatory region of the Ccr8 gene. Activated spleen T cells were treated for 48 h with or without dexamethasone (DEX, 10 nmol/L). The final DNA extractions were qPCR amplified by 7 pairs of primers covering the four peaks of the Ccr8 regulatory region (Chr9: 119 919 000 ~ Chr10: 119 925 000) (n = 4). P values in (A, C, D, G, H, K) were determined by two-tailed, unpaired Student’s t test. P values in (B, F) were calculated by ordinary two-way ANOVA with Tukey’s multiple comparison test. Survival in (I) was assessed by a two-tailed generalized Wilcoxon test

Next, we investigated whether CCR8 is required for brain tumor-induced T cell redistribution and immunosuppression against tumors. CCR8 antagonist R243 was administered to tumor-bearing mice (Fig. 6E). After repeated injections, the accumulation of T cells in the BM of tumor-bearing mice was blocked (Fig. 6F), tumor outgrowth was inhibited (Fig. 6G), and infiltration of T cells into the tumor was enhanced (Fig. 6H). Consistently, the mice in the R243 group exhibited an extended survival (Fig. 6I). Therefore, the above results suggest that T cells sequestration in the BM in mice with intracranial tumors was mediated by CCR8.

To further dissect how GR regulates Ccr8 expression in T cells, we analyzed the GR chromatin immunoprecipitation (ChIP)-sequencing data (GSE-183808), which were derived from splenic Treg cells treated with dexamethasone or ethanol as control. There were three GR-bound peaks (peak1, peak2, peak3) at the Ccr8 promoter region, and one peak (peak4) at an exon-intron junction region (Fig. 6J), but not in control Treg cells. To test whether these GR bound regions function as enhancers that regulate Ccr8 expression, ChIP-qPCR analysis of T cells was conducted. The data revealed that for each of the 4 peaks, the binding to GR was significantly induced by dexamethasone treatment (Fig. 6K). These results suggested that Ccr8 is transactivated by GR to direct T cell migration.

Discussion

The anti-tumor immune response is commonly compromised in patients suffering from brain tumors. These patients often exhibit T cell sequestration in BM. The relationship between the suppression of tumor immunity and the T cell relocation is not fully understood and the underlying mechanisms remain to be elucidated. We here identified glucocorticoid as a soluble mediator of the crosstalk between T cells and brain tumor. By acting on GR, glucocorticoid increased CCR8 expression in T cells. Together with elevated production of CCL1 and CCL8, two CCR8 ligands, in BM immune cells, peripheral T cells are recruited to BM and gain survival advantage (Fig. 7). This T cell sequestration in BM undermines the host immune response against brain tumors and potentially compromises immune checkpoint blockade therapy. Targeting GR-CCR8 may represent a novel strategy to improve immunotherapy of human brain tumors.

Fig. 7.

A schematic representation of the mechanism of T cell accumulation in BM of brain tumor bearing mice. Intracranial tumors cause elevated levels of glucocorticoid (GC). Glucocorticoid acts on T cells and increases the expression of the target gene, CCR8. Under the chemotactic effect of CCL1 and CCL8 released from the BM, CCR8 expressing T cells accumulate in bone marrow and are precluded from engaging in anti-tumor immunity

GC play a critical role in the regulation of development [22–24], migration [25, 26], and function [27, 28] of T cells. The physical and emotional burdens endured by cancer patients may activate stress-response mechanisms through the hypothalamic-pituitary-adrenal axis [29]. Such episode was reported to be associated with HPA axis dysregulation and earlier mortality in metastatic breast cancer [30]. Extreme stress in mice alters corticosteroid signaling on the brain-immune interface, which decreased immunosurveillance in the cerebrospinal fluid [31]. We here present a cascade of neuro-endocrine-immunoregulatory events in mice bearing intracranial tumors, which consists of a significant release of glucocorticoid, T cell migration to the BM and immunosuppression. Our findings are consistent with the proposed neuro-endocrine-immunoregulatory network in cancer patients [32]. Due to mechanical stress imposed by brain tumors, the surrounding brain tissue also suffers from impeded local vascular perfusion and consequently reduced intracranial immune infiltration [33, 34]. We speculate that any of the following mechanisms may contribute to the release of glucocorticoid caused by intracranial tumors: (1) the mechanical stress of the tumor tissue compresses the cerebral nerves and activates the HPA axis; (2) the tumor tissue causes cerebral edema, which increases the body’s demand for glucocorticoid and triggers the HPA axis; and (3) metabolites secreted by the tumor activate the HPA axis. Future studies may reveal the exact mechanisms.

A previous study utilizing the approach of parabiosis showed that tumor-free partners of glioma-bearing parabionts do not exhibit T cell BM sequestration while sharing vasculature system and thus reached the conclusion that BM T cell sequestration is mediated by non-steroid soluble mediators [35]. However, when tumor cells were injected into GR cKO mice, the BM accumulation of T cells was entirely inhibited. Our results were consistent with the report showing that corticotropin-releasing hormone neurons in the central medial amygdala were activated in tumor-bearing mice [36]. Glucocorticoid-driven migration of T cells and other types of cells to BM was also observed in other pathophysiological conditions [13, 36–39]. BM T cell sequestration under caloric restriction was shown to be dependent on HPA axis [13]. Our results suggest that inhibiting systemic glucocorticoid activation can block T cell bone marrow accumulation, but the numbers of tumor-infiltrating T cells did not increase. It might be because endogenous GCs are crucial for T cell development. Active GC is primarily derived from TECs in the thymus [12], not the adrenal gland, and MTY may prevent TECs from producing active GC, causing aberrant T cell and TCR pool counts in mice and, ultimately, reducing the quantity of tumor-infiltrating T cells [23]. We previously reported that the CXCL10-CXCR3 axis was responsible for the bulk of T cell chemotaxis to the BM during Dex treatment [37]. Meanwhile, GC signaling stimulated GR binding to a glucocorticoid response element (GRE) in the CXCR4 promoter, which facilitated CXCR4 transcription and improved human hematopoietic stem cell homing and engraftment [28, 38]. Interestingly, neither of CXCR3 and CXCR4 appeared to be required for BM T cell sequestration in mice bearing brain tumors. Instead, we identified CCL1/CCL8-CCR8 as a novel GC-driven chemotactic axis for T cell migration to BM. Sphingosine-1-phosphate (S1P)-S1P1 contributes to enhanced recruitment and/or retention of T cells in BM in brain tumor animals or those on a restricted diet [8, 13]. An increase in red blood cells (RBCs), which are the main source of S1P in the blood, was a remarkable component of BM remodeling following food restriction or glioblastoma [13]. Glucocorticoid stimulate the production of red blood cells by encouraging the early erythroid burst forming unit-erythrocyte progenitors to self-renew [39, 40]. Thus we hypothesize that glucocorticoid regulating of the S1P-S1P1R axis also played a role in T cell sequestration in the BM.

Several studies have suggested the potential involvement of CCL1-CCR8 axis in tumor progression [41]. In urothelial and renal carcinoma, higher CCL1 release by tumors and increased CCR8⁺ myeloid cell presence in cancer tissues and peripheral blood suggest that the CCL1-CCR8 axis plays a role in cancer-related inflammation [42]. In melanoma and breast cancer, CCL1 generated by lymphatic epithelial cells (LECs) of the lymph node subcapsular sinus acts on CCR8 of the tumor cells, allowing the tumor cells to enter the lymph node and metastasize [43]. CCL1 acts as autocrine factor in adult T-cell leukemia cells to prevent apoptosis and promote cell proliferation [44]. We thus speculate that the CCL1 and CCL8 secreted by BM immune cells may be the cause of the increased accumulation of T cells in BM of brain tumor-bearing mice. The absence of BM T cell sequestration in tumor-free parabionts implies a more complex mechanism of T cell sequestration in the BM. While GC act on GR in the T cells of peripheral immune organs, the BM microenvironment may be influenced by both neural and endocrine signals to acquire the ability for T cell accumulation in the context of brain tumors.

Compared to other tumor types, CNS tumors display relatively low numbers of tumor-infiltrating lymphocytes (TILs) that commonly exhibit an exhausted phenotype [45–47]. We here showed that targeting the GC-GR-CCR8 axis in GL261-bearing mice significantly inhibited tumor growth. Blocking this chemotactic signaling route may thus stop T cells from accumulating in the BM of individuals with brain cancer and brain metastases, thereby slowing the development of tumors.

We conclude that the GR-CCR8 axis plays a major role in brain tumor-induced T cell redistribution. This finding has implications in the development of immunotherapeutic strategies for patients with brain cancer. Targeting the GR-CCR8 axis may block T cell sequestration in BM and maintain the anti-tumor immune defense. Our finding also calls for prudence when managing patient stress and when using glucocorticoids in brain cancer treatments.

Materials and methods

Mice

Female C57BL/6 mice, 6–8 weeks old weighing 19–21 g, were obtained from Charles River (Beijing, China). CD4^Cre Nr3c1^fl/fl mice were produced by mating CD4^cre mice (Stock No.017336, Jackson Laboratory, Bar Harbor, ME) with Nr3c1^flox mice (Stock No.21021, Jackson Laboratory). These mice were housed in a specific pathogen-free facility at the Laboratory Animal Center of Soochow University with individually ventilated cages at 21 °C to 23 °C with a 12:12 h light/dark cycle. The relative humidity in the animal facility was kept between 40%-60%. The animal protocols for the experimentation and care were in full compliance with the care and use guidelines for experimental animals, and all protocols were approved by the Institutional Animal Care and Use Committee of Soochow University (SUDA20210916A02). Mice were randomly grouped using a random number table, with 4–10 mice per group. The sample size was determined based on previous studies, pilot studies, expected level of heterogeneity of the sample, and significance threshold (chosen at 0.05).

Cell lines

Murine GL261 malignant glioma cells were as described [48], AT3 breast cancer cells were obtained from SUNNCELL (Wuhan, China). Both cell lines are of the C57BL/6 genetic background and were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 1 mM sodium pyruvate, 25 mM HEPES, 4 mM L-glutamine, and 4.5 mg/mL glucose (EallBio, Beijing, China), along with 10% fetal bovine serum (Invitrogen, Waltham, MA) and 100 IU/mL penicillin/streptomycin (Invitrogen). Murine A1.1 T hybridoma cells were obtained from the American Type Culture Collection. The A1.1 cells were cultured in RPMI (EallBio) supplemented with 10% fetal bovine serum and 100 IU/mL penicillin/streptomycin.

Tumor models

For the intracranial implantation, tumor cells in PBS were loaded into a 10 µL microinjector. Mice were anesthetized with isoflurane, and the needle was positioned using a stereotactic frame 2 mm to the right of the bregma and 4 mm below the surface of the skull. A total volume of 2 µL containing 1 × 10⁴ GL261 or AT3 cells was delivered at a single injection site [6]. For subcutaneous implantation, a total volume of 100 µL containing 5 × 10⁵ GL261 cells was delivered into the subcutaneous tissues of the left flank in each mouse. The animals were euthanized either at the end of the experiment or when the rodent health endpoints set by the Institutional Animal Care and Use Committee were reached.

Pharmacological treatments of mice

To inhibit the activity of 11beta-hydroxysteroid dehydrogenase, we added 800 μg/mL metyrapone (MCE, Macao, China) to the drinking water from the day before tumor inoculation. In order to suppress GR activity, we intraperitoneally injected with 5 mg/kg RU486 (MCE) or 5 mg/kg AL 082D06 (MCE) (dimethyl sulfoxide, DMSO) every other day, starting the day after bearing-tumor. To block the CXCL10/CXCL11-CXCR3 interaction, we injected 5 mg/kg AMG487 (MCE) intraperitoneally into mice every other day from the day after bearing-tumor. Additionally, to block the CCL1/CCR8 interaction and inhibit CCR8 signaling and chemotaxis, we injected 5 mg/kg R243 (MCE, China) intraperitoneally daily from the day after bearing-tumor [49].

Adrenalectomy

After C57BL/6 J mice underwent anesthesia, small incisions were made in the dorsal skin directly above each adrenal gland [50]. Both adrenal glands were then removed using curved forceps heated by an alcohol lamp. Sham-operated mice underwent a similar surgical procedure, but their adrenal glands were not removed. As the adrenal glands also secrete aldosterone to regulate water-electrolyte metabolism balance, ADX and sham mice were given a 1% w/v saline solution in their drinking water. Tumor injection was carried out two weeks after the surgery.

Bioluminescence imaging

Bioluminescence imaging (BLI) was utilized to track the progression of tumor metastasis by monitoring luciferase activity. This was done using an IVIS Lumina XRMS Series III Imaging System (PerkinElmer, Waltham, MA) equipped with the XGI-8 system. Before conducting BLI, the mice were anesthetized with 2.5% isoflurane and injected with 150 mg/kg D-luciferin (Beyotime, Jiangsu, China) intraperitoneally 10 min prior to imaging. At the specified time point, the mice were positioned in the supine posture in the imaging instrument.

Murine tissue harvest

The spleen, thymus, femur, and tibia were collected at specific or humane endpoints as outlined in the study protocol. For animals with subcutaneous tumors, humane endpoints were determined by tumor size exceeding 20 mm in one dimension. In the case of animals with intracranial tumors, humane endpoints were indicated by the inability to walk two steps forward on cue. The spleen and thymus were weighed, processed in RPMI with 2% fetal bovine serum, filtered through a 70 μm filter, counted, stained with antibodies, and analyzed using a flow cytometer. BM cells were flushed from one femur and one tibia. Erythrocyte lysis was performed on BM prior to antibody labeling, while the thymus was labeled after producing a single-cell suspension. Blood samples were lysed with erythrocyte lysis buffer and directly labeled with antibodies. Freshly recovered tumors were processed in RPMI Medium 1640 with 5% fetal bovine serum, including 0.5 mg/mL type I collagenase, 1 mg/mL type IV collagenase, and 0.1 mg/mL DNase I (Yuanye Bio-Technology, Shanghai, China). Following a 30-min incubation at 37 °C, the cell suspension was filtered through a 70 μm cell strainer and then stained with an antibody cocktail.

Flow cytometric analysis

The antibodies used in the experiments including: anti-mouse CD45-BV510 (clone 30-F11, BD Biosciences, La Jolla, CA), anti-mouse CD45-BV605 (clone 30-F11, Biolegend, La Jolla, CA), anti-mouse CD45-Percp/cy5.5 (clone 30-F11, Biolegend), anti-mouse CD3-BV421 (clone 17A2, Biolegend), anti-mouse CD3-PE (clone 17A2, Biolegend), anti-mouse CD3-FITC (clone 17A2, Biolegend), anti-mouse CD4-BV421 (RM4-5, Invitrogen), anti-mouse CD4-AF700 (clone GK1.5, Biolegend), anti-mouse CD8-APC (clone 53-6.7, Biolegend), anti-mouse CD44-BV510 (clone IM7, Biolegend), anti-mouse CD25-PE-CY7 (clone 3C7, Biolegend), anti-mouse CD62L-BV421 (clone MEL-14, Biolegend), anti-mouse NK1.1-BV605 (clone PK136, Invitrogen), anti-mouse CD11b-APC-CY7 (clone M1/70, Biolegend), anti-mouse MHCII-PE-CY7 (clone M5/114.15.2, eBioscience), anti-mouse CD11c-BV421 (clone N418, Biolegend), anti-mouse F4/80-PE-CY7 (clone BM8, Biolegend), anti-mouse Ly-6G/Ly-6C(Gr-1)-BV605 (clone RB6-8C5, Biolegend), anti-mouse Ly-6C-BV605 (clone HK1.4, Biolegend), anti-mouse CD19-BV605 (clone 6D5, Biolegend), and anti-mouse CCR8-APC (clone SA214G2, Biolegend). These antibodies were used at a 1:400 dilution. Additionally, anti-mouse BCL-2-APC (clone BCL/10C4, Biolegend), anti-mouse PD-1-BV510 (clone 29 F.1A12, Biolegend), anti-mouse IFNγ–BV510 (cloneXMG1.2, Biolegend) and anti-mouse TNFα-PE (clone MP6-XT22, Biolegend) were used at a 1:300 dilution. 7-AAD-Percp-cy5.5 (eBioscience, San Diego, CA) and Fvs-AF700 (BD Biosciences) were used for dead cell staining. Gating strategy for flow cytometric analysis of immune cells was shown in Figure S1.

Immune cells were incubated with 1 mg/mL ionomycin (Beyotime) and 25 μg/mL phorbol myristate acetate (Merck/Sigma, Burlington, MA) in the presence of protein transport inhibitor (BD Biosciences) for 4 h to facilitate intracellular cytokine staining. Intracellular staining was carried out according to the manufacturer’s instructions using the eBioscience Foxp3/Transcription Factor Staining Buffer Kit.

Frozen tissue section and immunofluorescent staining

Freshly collected bone were treated in 4% PFA for 24 h at room temperature (RT), then treated in 10% EDTA (replaced daily) for 14 days at RT. Tissues were transferred to 30% (wt/vol) sucrose solution (in 1× PBS) for 24 h, embedded in optimum cutting temperature compound and then frozen at −80 °C. Frozen tissue sections (5 μm) were obtained with the CM1950 Cryostats (Leica Biosystems), adhered to poly-l-lysine-coated slides and stored at −80 °C. Frozen tissue sections were recovered to RT and fixed with 4% PFA for 15 min, permeabilized with 0.5% Triton X-100 for 5 min, blocked with 10% FBS for 60 min and stained at a 1:300 dilution with mouse CD3 antibody (clone E4T1B, Cell Signaling Technology) for overnight at 4 °C. Tissue sections were then washed with PBS, followed by nuclear counterstaining with chromosomal dye Hoechst 33258 (2 μg per mL, 5 min at RT). After 3 thorough washes with 1× PBS, tissue sections were stained with fluorochrome-conjugated secondary antibody diluted in PBS for 30 min at room temperature in the dark. Images were captured with confocal microscope (Leica TCS SP8, installed at Suzhou Institute of Systems Medicine, China).

ELISA for GC

The concentration of corticosterone in mouse plasma and bone were determined using an ELISA kit (Thermo Fisher, #EIACORT) following the guidelines provided by the manufacturer.

Cell sorting

Mice were euthanized and BM cells were gently extracted using a syringe. Red blood cells were lysed, then stained with fluorescently-labeled antibodies for 25 min at a low temperature. Flow-cytometric cell sorting was carried out using a CytoFLEX SRT instrument. Macrophages were characterized as CD11b⁺F4/80⁺. Monocytes were characterized as CD11b⁺Ly6G^lowLy6C^hi. T cells were identified as CD3⁺ and further categorized based on CD4 and CD8 expression. Stromal cells and fibroblasts were characterized as CD45^-.

T cell activation

T cells were isolated from the spleens of C57BL/6 mice. CD3⁺ T cells were purified by the Mouse CD3 T cell Isolated Kit (cat no. 480024, Biolegend). T cells (5 × 10⁵ cells/mL) were plated into 48-well plates with 1 mL medium per well. T cells were activated by treatment with 10% bovine serum albumin, αCD3 antibody (clone No.145-2c11), αCD28 antibody (clone No.37.51), and IL-2 (10 ng/mL, PeproTech, AF-200-02-1000) concurrently. The mRNA expression of Ccr8 were measured by reverse transcription polymerase chain reaction (RT-PCR), respectively, at 48 h time point post activation.

Quantitative real-time polymerase chain reaction

Total RNA was extracted from each sample using cell/tissue total RNA isolation kit (Vazyme, Nanjing, China) and reverse transcribed using RT Mix (Vazyme) following the manufacturer’s instructions. Gene expression was measured by a QuantStudio 6 Flex System using SYBR Green Master Mix (Vazyme). The total amount of mRNA was standardized to endogenous β-actin mRNA. The primers for the target gene were designed and listed in Supplementary Table 1 of the supporting information.

Chromatin immunoprecipitation quantitative PCR assays (ChIP-qPCR)

ChIP was performed using a SimpleChIP® Enzymatic Chromatin IP Kit (#9003, Cell Signaling Technology, Danvers, MA) following the manufacturer’s instructions. Activated T cells (4 × 10⁶) were collected, followed by ChIP assays with anti-GR (Cell Signaling Technology, #9003) or IgG. The diluted chromatin (1/50) was used as the input. The DNA was extracted, and qPCR was performed. The value of IP signal relative to input was calculated with 2^{([CT]Input -[CT]IP sample)}. The relative enrichment was represented as the fold change over EtOH group. Primer sequences to detect the GR binding site along the Ccr8 enhancer are listed in Table S2.

Western blotting

Cells were collected and lysed in RIPA buffer (Beyotime) with a protease inhibitor complex (BBI, Berlin, Germany) for 20 min on ice. The resulting mixtures were then spun at 12,000 × g for 15 min and heated at 95 °C for 10 min in sodium dodecyl sulfate sample buffer. The protein concentration of the supernatants was determined using BCA protein assays (Beyotime). Subsequently, the protein samples were separated on polyacrylamide gels, transferred to polyvinylidene fluoride membranes, and incubated in 2% bovine serum albumin diluted in TBST for 1 h at room temperature. The primary antibodies were then incubated overnight at 4 °C, washed three times in TBST, and the secondary antibody conjugated to horseradish peroxidase was incubated for 2 h at room temperature, followed by three washes in TBST. Finally, the membranes were developed using chemiluminescent detection (Millipore) according to the provided manufacturer’s instructions.

Statistical analysis

Significant differences between two groups were analyzed using GraphPad Prism 8 using unpaired two-tailed Student’s t test or in case of multiple groups, two-way ANOVA with multiple testing (Tukey). All data represent mean ± SEM or mean ± SD as indicated. Kaplan-Meier plots were generated using the survival calculation tool from GraphPad Prism, and significance was calculated using the two-tailed log-rank test at P < 0.05. Refer to figure legends for the specific statistical tests used in each experiment. Sample sizes were determined based on previous experience with similar experiments. Values of *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Author contributions

JZ performed most of the experiments, analyzed the data, and wrote the manuscript. YS and XX produced the mouse models and assisted with the experiments. WB, YL and TY assisted with the experiments. LC, JF, PL, YC and ZL provided reagents and advice. CS and YS supervised the project, designed the experiments, and together edited the manuscript. All authors have read and approved the article. All of the schematic representation are prepared using the BioRender online website (https://www.biorender.com).

Competing interests

The authors declare no competing interests. YS is editorial board member of Cellular & Molecular Immunology, but he has not been involved in the peer review or the decision-making of the article.

Supplementary information

The online version contains supplementary material available at 10.1038/s41423-024-01202-5.