The CCL2-CCR4 Axis Promotes Regulatory T Cell Trafficking to Canine Glioma Tissues

Purpose Spontaneously occurring glioma in pet dogs is increasingly recognized as a valuable translational model for human glioblastoma. Canine high grade glioma and human glioblastomas share many molecular similarities, including accumulation of immunosuppressive regulatory T cells (Tregs) that inhibit anti-tumor immune responses. Identifying in dog mechanisms responsible for Treg recruitment may afford targeting the cellular population driving immunosuppression, the results providing a rationale for translational clinical studies in human patients. Our group has previously identified C-C motif chemokine 2 (CCL2) as a glioma-derived T-reg chemoattractant acting on chemokine receptor 4 (CCR4) in a murine orthotopic model of glioma. Recently, we demonstrated a robust increase of CCL2 in the brain tissue of canine patients bearing high-grade glioma. Methods We performed a series of in vitro experiments using canine Tregs and patient-derived canine glioma cell lines (GSC 1110, GSC 0514, J3T-Bg, G06A) to interrogate the CCL2-CCR4 signaling axis in the canine. Results We established a flow cytometry gating strategy for identification and isolation of FOXP3+ Tregs in dogs. The canine CD4 + CD25high T-cell population was highly enriched in FOXP3 and CCR4 expression, indicating they are bona fide Tregs. Canine Treg migration was enhanced by CCL2 or by glioma cell line-derived supernatant. Blockade of the CCL2-CCR4 axis significantly reduced migration of canine Tregs. CCL2 mRNA was expressed in all glioma cell lines and expression increased when exposed to Tregs but not to CD4 + helper T-cells. Conclusion Our study validates CCL2-CCR4 as a bi-directional Treg-glioma immunosuppressive and tumor-promoting axis in canine high-grade glioma.


Introduction
Despite more than 450 NIH-sponsored phase II and III clinical trials, glioblastoma (GBM) remains a uniformly lethal primary brain tumor of adults [1][2][3][4].Notably, immunotherapeutic approaches that have advanced treatment in several non-CNS cancers have failed to show clinical bene t in GBM [5,6].This lack of progress highlights two key challenges that restrain therapeutic advancement across GBM: 1) the microenvironment of adult GBM is markedly immunosuppressed and 2) therapeutic successes in preclinical rodent models poorly translate into increased survival for human patients [7,8].The lack of therapeutic translation may rest, in part, on the homogeneous genetic backgrounds and lack of environmental immune in uence of rodent models that do not faithfully recapitulate the heterogeneity and complexity of human tumors [8,9].There is an urgent need for novel immunotherapeutic approaches to develop and test strategies by which to overcome the immunosuppressed microenvironment of GBM as well as the incorporation of complementary preclinical models which more closely predict therapeutic e cacy in GBM patients [8,9,10].
While immunosuppression in GBM is multifactorial, the recruitment of CD4 + CD25 + FOXP3 + regulatory T cells (Tregs) inhibit anti-tumor immune responses and present a signi cant therapeutic challenge [9,10].We and others have previously shown that glioma derived C-C Motif Chemokine Ligand 2 and 22 (CCL2 and CCL22) is a critical chemoattractant responsible for Treg recruitment into the GBM microenvironment.Moreover, increased intratumoral CCL2 expression correlated with reduced overall survival in human patients [10,11].Using murine models, we determined that CCL2 induces Treg recruitment into the glioma microenvironment through the CC chemokine receptor type 4 (CCR4) expressed on Tregs.Importantly, targeting of CCR4 reduced intratumoral Treg abundance and prolonged animal survival [10,11].Therefore, the CCL2-CCR4 signaling axis may serve as a therapeutic target for GBM patients.
Canine high-grade gliomas (HGGs) which develop de novo in an outbred, immunocompetent host, are increasingly pursued as a therapeutic model for human GBM [12][13][14][15][16][17].Canine HGGs, as de ned by the Comparative Brain Tumor Consortium of the National Cancer Institute, encompass grade III and IV canine astrocytomas and oligodendrogliomas [12].With a comparable disease incidence, naturally occurring canine HGGs share clinical, imaging, and histopathologic features with GBM [12].Tregs have been identi ed in high-grade canine oligodendrogliomas and astrocytoma [18,19].Moreover, we have demonstrated that CCL2 is robustly increased in canine HGG tumors relative to normal canine brain [14].Importantly, the CCL2-CCR4 axis is targetable and e cacious in canine patients.Blockade of CCR4 with the FDA approved monoclonal antibody reduced tumoral Treg in ltration and improved survival time in dogs affected with bladder and prostate cancer [20][21].Small molecule inhibitors of CCR4 are also in development (AZD2098 and K777) [22,23].
While Tregs have been observed in spontaneously occurring canine HGG, the mechanisms leading to their recruitment and function remains a critical knowledge gap for translational therapies targeting Tregs.Here, we performed a series of in vitro experiments using acutely isolated canine lymphocytes from healthy dogs and canine patient-derived glioma cell lines to interrogate the CCL2-CCR4 axis in dogs.Canine Tregs increased migration toward human recombinant CCL2 and glioma cell line-derived supernatant.However, this effect was abolished in the presence of CCL2 and CCR4 blockade.Moreover, canine glioma cells increased CCL2 mRNA expression when exposed to Tregs, but not CD4 + helper Tcells.These data demonstrate that the CCL2-CCR4 signaling axis is necessary and su cient for canine glioma cell-induced Treg migration and is mechanistic contributor of Treg recruitment to the tumor microenvironment in canine HGG.This work provides mechanistic rationale for preclinical trials testing the e cacy of CCR4 + Treg-targeted therapies in canine HGG.

Isolation of canine T regulatory cells.
This study was approved by Institutional Animal Care and Use Committee and the UC Davis Veterinary Medical Teaching Hospital Clinical Trial Review Board.Three apparently healthy client-owned dogs with a normal physical examination were recruited for this study.Following client consent, peripheral blood samples were taken from the jugular or lateral saphenous veins using standard veterinary technique and collected in ethylenediaminetetraacetic acid (EDTA) tubes.Peripheral blood mononuclear cells (PBMCs) were isolated from whole blood by gradient separation and negative selection with ammonium chloride potassium (ACK) as previously described [24].Puri ed cells were washed twice in PBS with 10% FBS by centrifuging at 600 g for ve minutes at 4°C.After washing, cells were re-suspended in PBS with 10% FBS and prepared for uorescence-activated cell sorting (FACS).Freshly isolated PBMCs were stained with mixtures of Paci c Blue (PB) conjugated anti-dog CD4 (clone YKIX302.9;Thermo Fischer Scieinti c) and Phycoerythrin (PE) conjugated anti-dog CD25 (clone P4A10; Thermo Fischer Scienti c).Cells were washed twice with PBS at room temperature for 10 minutes prior to sorting with the Beckman Coulter "MoFlo Astrios EQ".CD4 + cells were sorted by expression of CD25 into CD4 + CD25high and CD4 + CD25low populations.CD25 expression in the 90th percentile (highest 10%) were de ned as CD4 + CD25high, whereas CD25 expression at or below the 20th percentile (lowest 20%) were de ned as CD4 + CD25low.Cells were immediately re-suspended in culture medium after sorting and seeded into 96-well plates at a concentration of 0.5-2 x 10 4 cells per well.Cells were cultured in 5% reduced serum 1640 RPMI for 8-10 hours prior to downstream assays.Media contained RPMI-1640 complemented with 10% FBS, 10mM HEPES, 100µg/mL streptomycin, 100U/mL penicillin, 2000U/ml recombinant IL-2 and 0.5mM β-mercaptoethanol.
Canine Glioma Cell Lines.
Genomic integrity of each cell line was veri ed by comparison of copy number alterations with parental tumor DNA using Illumina Canine HD SNP array and sequenced to con rm canine origin.All cell lines were routinely tested and con rmed to be mycoplasma free by PCR.GSC1110 and GSC0514 cell lines were cultured in the serum free conditions containing Neurobasal-A media (Thermo Fisher Scienti c; 10888022), B27 supplement (Thermo Fisher Scienti c; 12587010), N2 supplement (Thermo Fisher Scienti c; 17502048), and 0.5X L-glutamine (Thermo Fisher Scienti c; 25030081).bFGF (R&D Systems, Inc., Minneapolis, MN, USA; 233-FB-025), EGF (R&D Systems, Inc., USA; 236-EG-200), and PDGF (Gemini Bio Products, West Sacramento, CA, USA; 300-178P) were added at 25 ng/ml concentration.J3T-Bg and G06A cell lines were cultured in media consisting of Dulbecco's Modi ed Eagle Medium with high glucose and glutamine, supplemented with 10% heat-inactivated fetal calf serum and 100 units/mL of penicillin, 100 µg/mL of streptomycin.
Quantitative Real-Time PCR.

Statistical analysis.
All statistical analyses were performed using Graphpad Prism 9 (GraphPad Software Inc., San Diego CA).Data were tested for normality via Shapiro-Wilks test.Data are reported as mean ± SEM.Statistical signi cance was assessed via unpaired two-tailed student's t test or ANOVA with Tukey's multiple comparisons test.Results were regarded statistically signi cant for p < 0.05.

High CD25 expression correlates with robust FOXP3 expression in CD4 + positive canine T lymphocytes.
There is very limited data on identi cation and isolation of Tregs for downstream assays in dogs [28].We hypothesized that freshly isolated canine CD4 + CD25 + T cells would possess the FOXP3 transcription factor associated with the T regulatory cell, similar to human Tregs.To test this hypothesis, we assessed FOXP3 expression in acutely isolated CD4 + CD25 + T cells from healthy dogs.When the total population of CD4 + CD25 + cells were gated together, FOXP3 immunoreactivity was detected in 26.1% (33.4 ± 5.7, n = 3) of cells (Fig. 1A).However, when examining the expression of FOXP3 as a function of CD25 expression, we observed a positive correlation between CD25 expression and FOXP3 expression (Fig. 1B, C).The geometric mean uorescence intensity (gMFI) of FOXP3 in the CD4 + CD25 high population was signi cantly increased compared to the CD4 + CD25 low and CD8 + cellular populations (p < 0.0001) (Fig. 1D).Furthermore, we observed a 17-fold increase in FOXP3 mRNA expression in CD4 + CD25 high cells compared to CD4 + CD25 low cells (p < 0.05) (Fig. 1E).Immunocytochemistry con rmed intranuclear FOXP3 immunoreactivity in CD4 + CD25 high T cell population (Fig. 1F).Therefore, these data demonstrate that CD4 + CD25 high T cells are indeed FOXP3 + expressing cells and positive selection of this population is a viable approach to isolate canine Tregs in the absence of intranuclear transcription factor FOXP3 immunostaining.
Canine Tregs induces increased CCL2 mRNA expression in canine glioma cell lines.
Glioma cells increase secretion of CCL2 to recruit immune-regulatory cells in models of human glioma [27].To determine if canine Tregs are capable of stimulating CCL2 expression in canine glioma cells, we performed co-culture experiments.

Discussion
Spontaneously occurring glioma in pet dogs is increasingly recognized as a unique model with valuable translational potential [29][30][31].Canine preclinical research and clinical trials may allow fast-track clinical investigations in human patients [29,32].Here, we have demonstrated that CCL2-CCR4 is a relevant, targetable, bi-directional Treg-glioma signaling axis in the dog.Our studies con rmed that canine glioma cells induced Treg chemotaxis through CCL2-CCR4 signaling, which was abolished following blockade.
Given the marked immunosuppression induced by Tregs in the glioma microenvironment, targeting CCL2-CCR4 can potentially improve patient outcomes through synergy with existing conventional and other experimental glioma therapies.As CCR4 monoclonal antibodies are FDA-approved for the treatment of Sézary syndrome, an aggressive human cutaneous T cell lymphoma [28], clinical investigations in canine glioma patients have the potential to support studies in human glioblastoma patients.
Since the discovery of Tregs and their immunosuppressive, pro-cancerous properties, considerable effort has gone into de ning, characterizing, and targeting murine and human Treg cells [31].In recent years, the successful reversal of Treg-mediated immunosuppression has advanced approaches to human oncology [34][35][36].Depleting or inhibiting intratumoral Treg in ux in murine models of adenocarcinoma, lymphoma, colon carcinoma, melanoma, and thymoma reduced tumor growth and, in some cases, induced remission [34][35][36][37][38][39].Phase 1 and II clinical trials investigating Treg-targeted therapies are currently underway in several cancer types [NCT03236129, NCT00986518, NCT05200559, NCT00847106, NCT02009384].The work of several groups has shown that these immunosuppressive cells support gliomagenesis [38].Blockade of Treg mobilization increases the survival of glioma-bearing mice, showing antineoplastic potential of targeting the recruitment capacity of these cells in glioblastoma [10,11,38,39].Canine Treg characterization is extremely limited [40].Given the scarcity of information on the identi cation and isolation of canine Tregs, one of the objectives of our work was to establish selection criteria for the isolation of live dog Tregs to develop preclinical support for clinical studies in dogs and subsequently in human patients.
There is considerable overlap in surface marker expression between Tregs and helper T cells.Moreover, while overexpression and nuclear localization of the forkhead box transcription factor three (FOXP3) by CD4 + CD25 + T cells demonstrate high speci city for Tregs, detection requires cellular permeabilization and is not suitable for isolating live Tregs.Similarly, as in human Tregs, we observed that the total population of CD4 + CD25 + canine T cells harvested from freshly isolated PBMCs was poorly enriched in FOXP3 (Fig. 1A).However, cascaded gating via ow cytometry revealed a positive correlation between CD25 and FOXP3.Over 70% of these cells with the highest CD25 expression (90th percentile) also expressed FOXP3 (Fig. 1B).Therefore, high CD25 expression is a critical marker for identifying canine Tregs and can serve as a surrogate marker for isolating these cells.This gating strategy also proved to be a robust work ow for the selective isolation of canine CD4 + CD25 + expressing high levels of CCR4 and using the enriched Tregs for downstream assays from freshly isolated canine PBMCs.
Our group and others have previously shown that glioma derived CCL2 is one of the crucial chemoattractant responsible for Treg recruitment into the glioma microenvironment in murine models and human tissues [10,11,16].An increased CCL2 expression has been correlated with reduced overall survival in human glioma patients [11].Importantly, antibody-mediated targeting of the Treg CCL2 higha nity receptor, CCR4, reduced intratumoral Treg abundance and prolonged survival in a mouse glioma model [10,11].Our work also demonstrated that CCL2 is robustly increased in high-grade canine astrocytoma compared to normal brains and given the role of CCL2 and CCR4 in both human and mouse models of gliomas, we studied here how this increased expression of CCL2 may affect the canine Tregs and CD4 + helper T-cells.
We performed a series of migration experiments to interrogate the CCL2-CCR4 axis in dogs.Importantly, canine Tregs migration was signi cantly enhanced by canine glioma cell line-derived supernatant, which was mitigated by the presence of an anti-CCL2 antibody.However, the most signi cant reduction in canine Treg migration was observed in the presence of a CCR4 antagonist or when combined with an anti-CCL2 antibody (dual-blockade).This suggests that the CCR4 receptor may be the major molecule mediating Treg migration to the tumor site in response to CCL2, and possibly other CCR4 ligands [41] secreted in the tumor environment.In the veterinary literature, several other chemokines are capable of binding and inducing Treg migration via the CCR4 receptor.For instance, CCL17 was reported to bind the CCR4 receptor to trigger the migration of Tregs toward canine urothelial carcinoma [42].These chemokines have not been the subject of our study, and we focus on the interrogation of the CCL2-CCR4 axis in canine gliomas as the next translational, targetable step of our work.A comprehensive evaluation of Treg-responsive chemokines within canine glioma supernatant and tumor microenvironment will be an important next-step to advance understanding of Treg biology in the context of high-grade glioma.
Here, we have interrogated the CCL2-CCR4 axis in canine glioma.Our data indicate that targeting either CCL2 or CCR4, alone or in combination, may represent a viable therapeutic strategy to diminish Treg accumulation in the glioma microenvironment.The availability of several therapeutic and experimental agents, including FDA-approved anti-CCR4 monoclonal antibody (Mogamulizumab) and small molecule inhibitors such as FLX475 [28, 43] AZD2098 and K777 [21,22,23,43], facilitates preclinical trials in canines to determine safety, e cacy, and whether this approach may have synergy with other immunotherapies.Given the importance of host immune mechanisms in governing the response to immunotherapy, future canine clinical trials will be critical to inform therapeutic e cacy in human patients.
In conclusion, we have established a work ow to identify and positively select canine Tregs from whole blood in dogs for downstream use.We have further established that the CCL2-CCR4 signaling axis is necessary and su cient for canine Treg chemotaxis.Importantly, this work demonstrates that canine Following 24-hour co-culture with Tregs, we observed increased CCL2 mRNA expression in each of four canine glioma cell lines (GSC0514: 1.02 ± 0.33, Fig. 4A p = 0.005; J3T-Bg: 1.02 ± 0.33, Fig. 4B p = 0.03; GSC1110: 0.55 ± 0.11, Fig. 4C p = 0.008; and G06A: 1.02 ± 0.33, Fig. 4D p = 0.04).CCL2 mRNA expression was not altered in any glioma cell line following co-culture with CD4 + helper T cells (Fig. 4A, B, C, D).

Figures
Figures

Figure 1 High
Figure 1

Figure 3 The
Figure 3