SOX2 Immunity and Tissue Resident Memory in Children and Young Adults with Glioma

Patients and blood and tumor samples

Peripheral blood (n = 14) and paired tumor tissue (n = 4) were obtained intraoperatively from pediatric and young adult glioma patients undergoing surgical resection at Yale-New Haven Hospital. Samples were collected after obtaining informed consent under a Yale University Institutional Review Board approved protocol. Immediately after resection, tumor samples were placed in sterile RPMI-1640 with l-glutamine (Corning) with 1% penicillin/streptomycin. All patients included in this study had a confirmed glial tumor. Blood samples were obtained from patients with a range of pathological grades and undergoing a variety of treatments including surgical resection alone, active or recent chemotherapy and/or radiation therapy. No patients included in the study had received any corticosteroids within the last 2 weeks prior to blood sample collection. Peripheral blood mononuclear cells (PBMCs) were obtained by density gradient centrifugation process using Ficoll-Paque Plus (GE Health Care Life Sciences). Tumor samples were processed as previously described[25]. Briefly, freshly resected tumor specimens were minced with a razor under sterile conditions, followed by enzymatic digestion (RPMI-1640 with 1-glutamine [Corning], 1 mg/ml collagenase IV [Sigma-Aldrich], 1 U/ml DNAse [Qiagen], and 1% penicillin/streptomycin) for 30 min at 37°C. A single cell suspension was then obtained by passing the sample through a 70- μm cell strainer.

Cell culture

Fresh tumor samples were processed as described above. For establishment of primary adherent cell culture, cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) (ThermoFisher) supplemented with 10% FBS (Gibco) and 1% penicillin/streptomycin. The cells were seeded in 25-cm2 culture flasks and maintained at 37°C with 5% CO2 and culture medium was changed every 2-3 days.

Primary tumor cells and CHLA-01-MED cells (ATCC) were resuspended in neurosphere medium (NM), composed of DMEM/F-12 supplemented with N2 (Gibco), B-27 (ThermoFisher), 20ng/ml EGF (ThermoFisher), 20ng/ml FGF (ThermoFisher), 20ng/ml LIF (ThermoFisher), and 1% penicillin/streptomycin. Viable cells were seeded in a 25-cm2 culture flasks at a maximum concentration of 5x104/ml and maintained at 37°C with 5% CO2 and culture medium was changed every 2-3 days.

Immunophenotyping by mass cytometry

Fresh patient PBMCs and tumor single-cell suspensions were stained at the same time, as previously described[25]. PBMCs from healthy children (seen in our clinic for family history of mild bleeding disorder) were used as an additional control. Cells were suspended at up to 2 million/ml in 1× PBS for viability staining by Cell-ID Cisplatin (final concentration of 5 μM; Fluidigm Sciences). Cells were mixed well and incubated for 5 min at room temperature. The staining was quenched with MaxPar Cell staining buffer and washed twice before proceeding to the usual procedure of surface and intracellular staining, as per manufacturer’s protocol. A 37-antibody staining panel was used with 31 surface markers and 6 intracellular markers (Supplementary Table 1). Between 1x106 and 2x106 PBMCs and tumor cells were incubated in a volume of 100 μl cell staining buffer with Abs in a polystyrene tube for 30 min at room temperature. After staining, cells were washed twice with buffer before fixing with BD Cytofix fixation buffer (100 μl/million cells) and permeabilizing with BD Perm/Wash buffer. Fixed and permeabilized cells were stained with the intracellular cocktail for 30 min at room temperature. Cells were washed twice with buffer and suspended in 1ml intercalation solution containing MaxPar Intercalator-Ir in MaxPar Fix and Perm buffer at final concentration of 125 nM. Cells were left overnight in the intercalator solution, washed with staining buffer, and finally suspended at 106cells/ml in MaxPar water supplemented with 10% EQ 4-element calibration beads (Fluidigm) before acquiring on CyTOF 2 instrument (DVS; Fluidigm Sciences). To facilitate quantitative comparisons between data acquired on different days, single-cell data was normalized using beads. At the completion of data acquisition, files were concatenated into a single FCS file and the normalization beads were removed using the CyTOF II built-in software. All data were analyzed using Cytobank. An unsupervised cluster analysis tool (viSNE) was used to visualize the data in two dimensions based on the t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm[26].

Immunohistochemistry

Paraffin sections from pediatric brain tumor specimens were subjected to antigen retrieval at low pH with citrate buffer. Slides were then stained with monoclonal mouse anti-human SOX2 antibodies (1.25μg/ml, R&D, clone #245610). Tumor tissue was called positive for SOX2 if any nuclear staining was detected with the SOX2 antibody. SOX2-positive tumor cell nuclei were quantified in five randomly selected 40X optical fields.

Peptide libraries

Overlapping peptides libraries spanning the entire length of SOX2 were synthesized as previously described[27]. The SOX2 library consisted of 86 peptides divided into 4 sub-mixes (Supplementary Table 2). M1 peptides cover SOX2 residues 1–89, M2 residues 79–171, M3 residues 161–246 and M4 residues 236–321. A pool of peptides derived from cytomegalovirus, Epstein–Barr virus and influenza virus (CEF; Anaspec Inc.) and Candida albicans (Greer Laboratories Inc.) were used as a positive control.

Detection of antigen-specific T-cells

The presence of SOX2, CEF and Candida albicans reactive T-cells was detected based on antigen-dependent cytokine production and proliferation, as previously described[27,28]. Briefly, PBMCs were cultured either with media alone (control) or together with CEF peptides (5 μg/mL per peptide), Candida albicans (10 μg/ml) or SOX2 peptide pools (5 μg/mL per peptide) in 5% PHS, in 96-well round bottom plates (2.5 × 105 cells/well). PHA was used as a positive control. After 48 hrs., culture supernatants were harvested and examined for the presence of chemokine (C-X-C motif) ligand 10 (CXCL10, also known as IP10) using a Luminex assay, as per manufacturer’s instructions (Millipore, MA). The samples were collected on Luminex 100 instrument and analyzed using the xPONENT software (Luminex Corporation). Values ≥2-fold over the negative control were deemed positive, based on the analysis of inter- and intra-assays variation, as previously described[27]. In this assay the antigen-induced secretion of CXCL10 serves as a downstream marker of T-cell reactivity and depends on the presence of CD3+ T-cells as well as on the induction of IFN-γ, as previously described[27].

Antigen-dependent proliferation assays

The presence of antigen-dependent T-cell proliferation was examined using a CFSE dilution assay, as previously described[28]. PBMCs were labeled with 0.5 μM CFSE (Molecular Probes) and cultured with 1 μg/mL anti-CD28 and anti-CD49d antibodies (BD Biosciences), alone or in the presence SOX2 peptide mixes (5 μg/mL per peptide), CEF peptides (5 μg/mL) peptides, candida albicans (10 μg/ml) or PHA. Five days later, PBMCs were stained with anti-CD3, anti-CD4 and anti-CD8 antibodies (all BD Phamingen). T-cell proliferation was analyzed on a FACSCalibur cytofluorometer (Becton Dickinson). Flow cytometry data were analyzed using the FlowJo software.

siRNA transfection

On-TargetPlus SmartPool siRNAs for SOX2 (Cat. L-011778) and non-targeting pool (Cat. D-001810) were purchased from Dharmacon (Boulder, CO, USA). Neurospheres were dissociated with TrypLE (ThermoFisher) to make a single-cell suspension and then resuspended in Opti-MEM without phenol red (ThermoFisher) at 2.5x106/100ul along with 20ug of siRNA. Cells were then transferred to a 0.4cm electroporation cuvette (Bio-Rad, Cat. 165-2088) and pulsed with 500mV x 500msec with an ECM 830 electroporator (BTX-Harvard Apparatus). Cells were cultured in duplicate in NM in 6-well plates and harvested at 72-hrs.

Statistical analysis

A 2-tailed Student’s t-test was used to determine statistical significance and a p < 0.05 was considered statistically significant.

SOX2 immunohistochemistry in pediatric brain tumors.

Prior studies have characterized the expression of SOX2 in adult glioma[18,29,30]. In order to examine the expression of SOX2 in pediatric glioma, we analyzed 27 pediatric tumor samples using immunohistochemistry (Table 1). SOX2 expression was detected in tumor cells but not in the surrounding normal tissue in all juvenile pilocytic astrocytoma (JPA), diffuse astrocytoma, anaplastic astrocytoma and glioblastoma, and in 60% of oligodendrogliomas (Fig. 1a). SOX2 staining was nuclear and its intensity appeared to increase in higher grade lesions (Table 2). RNAi-mediated inhibition of SOX2 in primary JPA cells as well as a pediatric anaplastic astrocytoma cell line (CHLA-01) led to decrease in cell growth (Fig. 1b), consistent with prior studies[18,31]. Together, these data show that SOX2 can be a useful marker for histopathologic analysis of pediatric glial tumors, particularly to assess infiltration of normal tissue by tumor cells.

Detection of SOX2 T-cell response in pediatric and young adults with glioma.

We have previously developed an overlapping peptide library spanning the entire SOX2 protein that allowed us to evaluate naturally occurring T-cells against this antigen in patients with several adult human tumors[27,32,33]. We utilized this peptide library and antigen-dependent cytokine secretion (Fig. 2a) and proliferation assays (Fig. 2b) to detect anti-SOX2 T-cells in pediatric and young adults with glioma (Table 3, Supplementary Table 3). The presence of anti-SOX2 cytokine producing T-cells was detected in 5/14 (36%) of patients studied. There was no difference in the anti-SOX2 T-cell reactivity between patients with LGG and HGG (3/8 vs. 2/6 patients, respectively) (Fig. 2c). The region of the SOX2 protein that was most commonly recognized in patients with SOX2 immunity included amino-acid residues 79–171 (Fig. 2d). T-cell response to control antigens (CEF, a mix of peptides from viral antigens CMV, EBV, influenza; or Candida albicans) or polyclonal mitogen (PHA) were similar between patients with and without a SOX2 T-cell response indicating that the absence of SOX2 immunity was not due to global immune paresis (Fig. 2e). There were also no significant differences in tumor histology, age or type of treatment between patients with and without SOX2 T-cell reactivity (Supplementary Table 4). In two patients, we assessed T-cell reactivity to SOX2 in paired blood and freshly resected tumor tissue. Tumors in both patients expressed SOX2 (data not shown). No reactivity to SOX2 was detected in either blood or tumor tissue in one patient. Interestingly, in the second patient, SOX2 specific T-cells were only detected in the tumor, but not blood (Fig. 2f). These findings demonstrate that SOX2 is immunogenic in pediatric glioma patients. These findings also suggest that SOX2 specific T-cells can enter the tumor tissue and SOX2 T-cell reactivity may be detected in tumors even when it is not detectable in circulating T-cells.

Characterization of pediatric glioma immune cell infiltration

In order to better understand the immune microenvironment in pediatric glioma, we utilized single cell mass cytometry (CyTOF) to characterize immune cells from paired blood and freshly resected pediatric glial tumors in four patients (Table 4). We also obtained blood from healthy children as a control. The peripheral blood from patients with glioma had similar composition of immune cells (Fig. 3a–c) and expression of immune checkpoints as blood from healthy children (Fig. 3e). We found that the tumor tissue is enriched in myeloid cells (CD45+CD11b+), and these cells represent the dominant component of CD45+ cells infiltrating these tumors (Fig. 3a). Tumor-infiltrating cells of myeloid origin include both microglia (CD45dimCD11b+) and inflammatory monocytes/macrophages (CD45hiCD11b+), as previously described[34]. T-cells within tumor consisted of both CD4+ as well as CD8+ T-cells and there was no significant difference in the proportion of CD4+ or CD8+ T-cells between blood and tumor tissue (Fig. 3b). In contrast to peripheral blood, most of the tumor-infiltrating CD4+ and CD8+ T-cells were CD45RO+ memory cells (Fig. 3c), with a T_RM phenotype (CD45RO+CCR7-CD69+). About half of the CD8+ T_RM cells expressed CD103 (Fig. 3d).

Next, we examined the expression of several co-stimulatory and co-inhibitory molecules including PD-1, PD-L1, TIM3, BTLA, CD200, OX40, 4-1BB and TIGIT on peripheral blood of healthy children as well as paired peripheral blood and tumor infiltrating T-cells from children with glial tumors. Both CD4+ and CD8+ T-cells in the tumor had a higher proportion of PD-1+ T-cells (Fig. 3e) as well as a higher expression of PD-1 (Fig. 3f) compared to circulating T-cells. In contrast, the expression of other immune checkpoints studied were comparable between blood and tumor (Fig. 3e and f). Immune checkpoints are known to be expressed on only a subset of tumor-infiltrating T-cells. In glioma tissue, we found that expression of PD-1 is specifically enriched on the T_RM subset of tumor-infiltrating CD4+ and CD8+ T-cells (Fig. 3g–3h). Importantly, several of the PD-1+ T-cells also co-express other inhibitory checkpoints, including TIGIT and PD-L1 (Fig. 3i).

In addition to T-cells, we also compared the proportion and phenotype of innate CD56+NK cells in blood from healthy children as well as paired blood and tumor tissue from children with glioma. While the proportion of NK cells is comparable between blood and tumor tissue (Fig. 3a), tumor-infiltrating NK cells express significantly less granzyme and CD16 than their circulating counterparts (Fig. 3j). The loss of granzyme appears to be specific to tumor-associated NK cells, since CD8+ T-cells infiltrating these tumors retain granzyme expression (Fig. 3k). Our data show that the characteristics of T-cells and innate NK cells infiltrating the tumor tissue are distinct from their circulating counterparts.