Toll like receptor-3 ligand poly-ICLC promotes the efficacy of peripheral vaccinations with tumor antigen-derived peptide epitopes in murine CNS tumor models

C57BL/6 (H-2^b) and ovalbumin (OVA)-specific T-cell receptor (TCR) transgenic OT-1 mice (C57BL/6-background) were purchased from Taconic (Germantown, NY). Pmel-1 mice (The Jackson laboratory, Bar Harbor, ME) are transgenic for a human(h)gp100_25–33 – specific TCR, which cross-reacts with mouse(m) gp100_25–33. Animals were handled in the Animal Facility at the University of Pittsburgh per an Institutional Animal Care and Use Committee-approved protocol.

The mouse (H-2^b) GL261 glioma cell line was kindly provided by Dr. Robert Prins (University of California, Los Angeles, CA), with the OVA-transfected B16 (M05) melanoma cell line kindly provided by Dr. Louis D. Falo, Jr. (University of Pittsburgh, PA). These cell lines were maintained in mouse complete medium [RPMI 1640 supplemented with 10 % heat-inactivated fetal bovine serum, 100 units/mL penicillin, 100 μg/mL streptomycin, and 10 mmol/L L-glutamine (all reagents from Life Technologies, Inc., Grand Island, NY)] in a humidified incubator in 5% CO₂ at 37°C.

Preparation of i.c. tumor-bearing mice was performed as previously described [1, 21, 22]. Briefly, 5 × 10³ M05 or 5 × 10⁴ GL261 cells in 2 μl PBS were stereotactically injected through an entry site at the bregma 2 mm to the right of the sagittal suture and 3 mm below the surface of the skull of anesthetized mice using a stereotactic frame. The animals received s.c. vaccinations with corresponding peptides emulsified in incomplete Freund Adjuvant (IFA) (Difco Laboratories, Detroit, Michigan, MI) and i.m. administrations with poly-ICLC (50 μg/injection; Oncovir Inc, Washington, DC) on indicated days. Animals were monitored daily after treatment for the manifestation of any pathologic signs. In some experiments, symptom-free survival was monitored as the primary endpoint, and in other experiments, treated mice were sacrificed on indicated days to evaluate immunological endpoints, such as brain infiltrating lymphocytes (BILs).

All peptides, including H-2K^b-restricted OVA_257–264 (SIINFEKL), and mTRP2_180–188 (SVYDFFVWL), H-2D^b-restricted hgp100_25–33(KVPRNQDWL) and mEphA2_671–679 (FSHHNIIRL), I-A^b-binding HBV core_128–140 (TPPAYRPPNAPIL), were synthesized in the University of Pittsburgh Peptide Synthesis Facility, and were > 95% pure as indicated by analytical high-performance liquid chromatography and mass spectrometric analysis. Peptides were dissolved in phosphate-buffered saline (PBS)/10% dimethyl sulfoxide (DMSO) at a concentration of 2 mg/ml and stored at -20°C until use. Phycoerythrin (PE)-conjugated- H-2K^b/TRP2_180–188 tetramer was produced by the National Institute of Allergy and Infectious Disease (NIAID) tetramer facility at the Emory University Vaccine Center (Atlanta, GA). PE-H-2K^b/OVA_257–264 tetramer was purchased from Beckman Coulter Inc. (Fullerton, CA).

BILs were isolated using the methods described previously [1, 23]. In brief, mice were sacrificed by CO₂ asphyxia, then perfused through the left cardiac ventricle with PBS. Brains were mechanically minced and cells from each brain were resuspended in 70% Percoll (Sigma, Saint Louis, MO), overlayed with 37% and 30% Percoll, then centrifuged for 20 min at 500 × g. Enriched BIL populations were recovered at the 70%-37% Percoll interface [19, 24]. For in vivo inhibition of α₄-integrin, mice received i.p. injections with anti-CD49d monoclonal antibodies (mAbs) (150 μg/mouse for R1-2 and 9C10; BD PharMingen, San Diego, CA) or 300 μg/mouse isotype rat IgG_2bK, (clone A95-1; BD PharMingen, San Diego, CA) at the indicated time points.

TriColor (TC)-anti-CD3, fluorescein isothiocyanate (FITC)-anti-CD8α, PE-anti-CD8α, TC-anti-CD8, PE-anti-H-2K^b, FITC-anti-CD49d (α4-integrin), FITC-anti-TCRvβ13, purified anti-CD49d (clones R1-2 and 9C10) mAbs were purchased from BD PharMingen (San Diego, CA). PE-anti-TLR3 was purchased from eBioscience. IFN-β-neutralizing mAb (MIB-5E9.1) and the corresponding, functional grade, purified Armenian hamster IgG (eBio299Arm) were obtained from BioLegend and eBioscience, respectively. Single-cell suspensions of splenocytes (SPCs), lymph node (LN) cells and BILs were stained with 10 μg/ml PE-labeled MHC-peptide-specific tetramers in PBS containing 1% BSA for 15 min at room temperature, then washed once and stained with TC-anti-CD3 or FITC-anti-CD8 mAbs. For intracellular IFN-γ staining, cells were surface-stained with PE-anti-CD8 mAb, washed, fixed, and then permeabilized with Cytofix/Cytoperm buffer (BD PharMingen) before staining with anti-IFN-γ mAb (BD PharMingen). Flow-cytometric analyses were performed using Coulter EPICS cytometer for lymphocyte-gated cell populations (Beckman Coulter, Fullerton, CA).

C57BL/6 mice were immunized s.c. with the OVA (100 μg) and the helper HBV Core_128–140 (150 μg) peptides emulsified in IFA and i.m. poly-ICLC (50 μg/ml) on days 0 and 7. On day 14, erythrocyte-depleted SPC and lymph node cells from naïve C57BL/6 mice were either: 1) pulsed with 10^-6 M OVA-peptide, incubated at 37°C for 1 hour, and labeled with a high concentration of Carboxyfluorescein diacetate, succinimidyl ester (CFDA SE; CFSE) (2.5 μM) (CFSE^high cells); or 2) pulsed with no peptide and labeled with a low concentration of CFSE (0.25 μM) (CFSE^low cells). Then, equal numbers of cells from each populationwere mixed and i.v. infused into immunized mice (1 × 10⁷ cells in 200 μl PBS/mouse). At 4 hrs after i.v. infusion, SPC were evaluated for the ratio of CFSE^low to CFSE^high cells by flow cytometry. To calculate specific lysis, the following formula was used: ratio = (percentage CFSE^low/percentage CFSE^high). Percentage specific lysis = [1 – (ratio unprimed/ratio primed) × 100] [25].

C57BL/6 mice were immunized s.c. with synthetic, GL261-derived CTL epitopes (50 μg TRP-2_180–188 and hGP100_25–33, 100 μg mEphA2_671–679) and the helper peptide (150 μg HBV Core_128–140) emulsified in IFA (GAA-vaccine) in combination with i.m. poly-ICLC on days 0 and 7. SPCs were harvested on day 14, and in vitro re-stimulated for 5 days with IL-2 (20 IU/ml) and each of (5 μg/ml) three GAA peptides before SPCs were tested for their lytic activity against GL261 glioma cells and control, GAA-negative EL-4 cells in a 4-hr standard ⁵¹Cr-release assay as previously described [22].

Single cell-suspended BILs, SPCs and LN cells were restimulated with OVA_257–264 peptide (5 μg/ml) in the presence of rhIL-2 (20 U/ml) for indicated time. Culture supernatants were assessed for mIFN-γ using a specific ELISA kit (BD PharMingen, San Diego). IFN-inducible protein (IP)-10 production levels from tumor cells were determined by mIP-10 ELISA kit (R&D, Minneapolis, MN).

Perfusion-fixed brains were obtained from mice treated with GAA-vaccine and poly-ICLC on day 90 after the tumor inoculation. The brains were then embedded in optimal cutting temperature compound (4583; Sakura Finetek), frozen at -80°C, and thin coronal sections [20 μm for Luxol Fast Blue (LFB) and 5 μm for Hematoxylin & Eosin (H&E)] were made using a cryostat. Sections were stained with H&E to evaluate the overall infiltration of mononuclear immune cells, or with LFB to evaluate the demyelination.

Survival data were compared using Log rank test. Comparative T-cell responses were analyzed by one way analysis of variance for comparing means of three or more variables, ANOVA).

GL261 glioma cells (H-2^b) express well-characterized CTL epitopes, such as the H-2D^b-restricted mgp100_25–33 and the H-2K^b-restricted TRP-2_180–188 peptides [26]. In addition, we recently identified a novel T cell epitope H-2D^b-restricted mEphA2_671–679, derived from the receptor tyrosine kinase mEphA2, which is overexpressed in human [27] and murine gliomas, including GL261 [28]. We hereby define these three antigens (mgp100, TRP-2 and EphA2) as the glioma-associated antigens (GAAs) in the GL261 glioma cells. As demonstrated in Fig. 1, immuno-staining revealed a high-level expression of EphA2 in the GL261 tumor; whereas the normal brain did not demonstrate significant staining. These results confirm the Western Blot analyses reported in our previously published study [28]. In the case of the M05 melanoma model system, the OVA_257–264 peptide represents an additional "tumor" epitope [1].

We sought to determine whether poly-ICLC administration could enhance the induction of specific CTL response to peptide epitopes employed in vaccines. In vivo CTL assays against OVA_257–264-peptide demonstrated that concurrent i.m. poly-ICLC and OVA peptide-based vaccine administration yielded remarkably enhanced induction of specific CTL responses in both SPCs and draining LN cell populations (Fig. 2A). Furthermore, i.m. poly-ICLC administration augmented anti-GL261 CTL responses induced by vaccinations with 3 endogenous GAAs, mEphA2_671–679, hgp100_25–33and TRP-2_180–188 in addition to the HBVcore_128–140 T-helper epitope (GAA-vaccines) (Fig. 2Ba). EL-4 thymoma cells that do not express these GAAs were used as negative control target cells, and were poorly recognized (< 10% lysis) in all groups (Fig. 2Bb), supporting the specificity of the vaccine-induced CTL responses. Poly-ICLC administration also enhanced the SPC-production of IFN-γ following vaccinations with mEphA2_671–679 (Fig. 2C).

To determine whether poly-ICLC administration enhanced the infiltration of brain tumors by vaccine-induced OVA_257–264- specific CTLs, C57BL/6 mice bearing intracranial (i.c.) M05 tumor received vaccinations using OVA peptide with or without poly-ICLC for two cycles prior to isolation of BILs. The percentage (Fig. 3A) and the absolute numbers of CD3⁺/OVA tetramer⁺ BILs per mouse (Fig. 3B) were then analyzed by flow cytometry. As displayed in Figs. 3Aa and 3Ba, there were relatively few OVA-tetramer/CD3-double positive BILs in mice that had received mock-vaccination or poly-ICLC alone. Even though the percentage of OVA-tetramer/CD3-double positive cells appeared to be slightly elevated as a result of poly-ICLC treatment alone (Fig. 3Aa), the actual number of these cells was still low, due to the low total number of BIL obtained from this group (Fig. 3Ba). Although OVA-vaccines appeared to have increased the total number of OVA-specific T-cells moderately, due to the increase of total CD3⁺ lymphocytes in this group compared to groups treated with poly-ICLC alone or mock-treated mice (Fig. 3Ba), the numbers of OVA-tetramer/CD3-double positive cells remained less than 1000/mouse in these three groups in 2 independent experiments. In contrast, the addition of poly-ICLC to the OVA-vaccine regimen remarkably increased tumor-infiltration by OVA-specific responder T cells (3,900/mouse).

To improve the sensitivity and reliability of flow-cytometry based enumeration of BILs, in parallel experiments, i.c. M05 bearing mice received i.v. transfer of naïve OT-1 mice-derived SPCs and LN cells prior to the first immunization. As demonstrated in Figs. 3Ab and 3Bb, the combination therapy with OVA-vaccines and i.m. poly-ICLC administrations resulted in a remarkable increase in the percentage (Fig. 3Ab), and in the absolute numbers (2.06 × 10⁵/mouse) (Fig. 3Bb) of OVA-specific CD3⁺ BILs per mouse compared to other groups including mice receiving OVA-vaccines alone (2.05 × 10⁴/mouse).

To evaluate whether the increased numbers of OVA-reactive T-cells were also evident systemically, we also harvested lymphocytes from the spleen, draining inguinal LNs and cervical LNs, and evaluated these populations for their frequencies of OVA-reactive CD8⁺ T cells. The percentage of OVA_257–264-reactive CD3⁺ T cells among total lymphocyte gated populations ranged from 0.54 – 1.05% and 0.34–0.65% in mice treated with both OVA-vaccines plus poly-ICLC, vs. mice treated with OVA-vaccines alone, respectively (data not shown). Hence, poly-ICLC administration in combination with OVA-vaccination preferentially promoted the infiltration of vaccine-induced OVA-specific T-cells into brain tumor sites.

To assess the functional status of BILs, freshly-isolated BILs were evaluated for expression of IFN-γ following a brief (6 hrs) in vitro re-stimulation with the OVA_257–264-peptide and low-dose rhIL-2 (Fig. 3Ca). BILs obtained from mice receiving both OVA-vaccines and poly-ICLC produced high levels IFN-γ, whereas IFN-γ levels were undetectable in all other groups. SPCs and LN cells obtained from mice receiving the combination therapy produced higher levels of IFN-γ than those obtained from other groups following 96 hr in vitro stimulation (Fig. 3Cb–d) (P < 0.001). These results suggest that poly-ICLC delivery enhances Type-1 (i.e. IFN-γ expressing) function of vaccine-stimulated T cells, especially within the i.c. tumor-microenvironment.

Efficient CNS-tumor homing of Ag-specific T-cells activated by Ag-specific vaccines and poly-ICLC administration led us to determine whether poly-ICLC administration induces qualitative changes in the phenotype of vaccine-activated OVA-reactive T cells particularly in the context of homing receptors involved in brain tumor-tropism. Based on a recent study by Calzascia et al. demonstrating that up-regulation of VLA-4 expression on Ag-specific CTLs confers efficient CNS-tumor tropism [19], we evaluated expression of CD49d (Integrin α4 chain), a subunit for the VLA-4, on BILs (Fig. 4A) and SPCs (Fig. 4B) obtained from mice receiving OVA-vaccines and/or i.m. poly-ICLC administrations. As depicted in Fig. 4A, in vivo administration of poly-ICLC remarkably increased the number of α₄-integrin (CD49d)⁺ OVA-tetramer binding CD8⁺ cells in BILs. Although the percentage was lower compared to BILs, the combination regimen also increased the numbers of α₄-integrin (CD49d)⁺ OVA-tetramer binding cells in SPCs (Fig. 4B). These cells also expressed β₁-integrin (CD29), but did not express detectable α₄β₇ integrin heterodimers (data not shown), indicating that α₄-integrin (CD49d) is solely expressed in VLA-4 complexes.

We next performed parallel experiments using i.c. GL261-bearing mice treated by adoptively transfer of naïve Pmel-1 derived CD8⁺ cells followed by vaccination with gp100 peptides and poly-ICLC administration. We observed results similar to the M05 model, with increased i.c. glioma-infiltration by α₄-integrin/TCR-Vβ13 double-positive cells as a result of this treatment regimen (Fig. 4C).

VLA-4 allows activated T-cells to adhere to vascular cell-adhesion molecule (VCAM)-1⁺ endothelial cells [17, 18]. Therefore, these data suggest the possibility that VLA-4 may play a significant role in the CNS-tumor homing of CTLs induced by vaccine plus poly-ICLC co-treatment. To address this hypothesis, we evaluated the effects of specific mAb-mediated blockade of the VLA-4/VCAM-1 interaction (Fig. 4D). Before these in vivo experiments, we confirmed that each of two anti-α₄-integrin mAbs (Clones R1-2 and 9C10), and especially the combination of these two mAbs, effectively blocked the binding of activated T-cells to plastic plates coated with VCAM-1 (Sasaki et al. submitted). C57BL/6 mice bearing day 10 i.c. M05 tumors received i.p. injections of anti-α₄-integrin mAbs (R1-2, and 9C10), or control isotype mAb (rat IgG_2bK, clone, A95-1) 2 hrs prior to i.v. adoptive transfer of 5 × 10⁶ naïve OT-1 mouse-derived T-cells, which was immediately followed by OVA-vaccination and poly-ICLC administration. On day 13, i.p. mAb injections and vaccinations were repeated. BILs were harvested on day 16, and evaluated by flow cytometry. As depicted in Fig. 4D, the functional blockade of α₄-integrin by specific mAbs diminished the CNS-tumor homing of OVA-specific T-cells. Treatment with mAbs did not affect the peripheral expansion of OVA-reactive T-cells, as the numbers of OVA-tetramer reactive T-cells in lymphoid organs were not significantly altered by mAb treatments (Fig. 4E). These data suggest that the mAb-mediated blockade of α₄-integrin inhibits the tumor-homing of these effector cells, but not their induction or expansion. Even though efficient tumor-homing by Ag-specific CD8⁺ T cells most likely results from the orchestrated cooperation of multiple receptor-ligand combinations [29], these results indicate that poly-ICLC induced VLA-4 expression on vaccine-induced CTL dominantly influences their capacity to infiltrate into CNS tumors.

To evaluate whether the enhanced, systemic and local Ag-specific T-cell responses stimulated by the combination therapy would translate into a more clinically-relevant model, C57BL/6 mice were pre-immunized with GAA-vaccines on days -14 and -7, with or without poly-ICLC administration, before they received i.c. injection of 5 × 10⁴ GL261 cells in the right hemisphere on day 0. As depicted in Fig. 5A, all mice receiving mock treatments died by day 47. Treatment with poly-ICLC alone had no therapeutic effect when compared to the mock vaccine group (P > 0.05). Although GAA-vaccination alone resulted in long-term (> 90 days) survival in 3 of 10 mice, this level of protection did not reach statistical significance when compared to the control, mock treatment group (p = 0.0521). In contrast, addition of poly-ICLC to the GAA-vaccine protocol improved survival, with 9 of 15 animals still alive on day 90, a statistically significant benefit when compared with the control (P = 0.003) group.

To further determine whether i.c. tumor-bearing mice immunized with GAA-vaccines and poly-ICLC exhibit long-term anti-tumor memory immune response, survivors in experiments in Fig. 5A were re-challenged with 5 × 10⁴ GL261 cells in the contralateral hemisphere of the brain on day 90 after the initial tumor inoculation. As a control group, non-immunized age-matched mice also received the same i.c. GL261 cell-inoculations. On day 7 after tumor re-challenge, mice were sacrificed, and BILs were harvested. As shown in Fig. 5B, flow-cytometric analyses of BILs revealed an elevated percentage of H-2K^b/TRP-2_180–188 tetramer reactive CD8⁺ T cells in the vaccinated, long-term survivors when compared to BILs obtained from control non-immunized GL261-bearing mice. Furthermore, mice received the combinational regimen exhibited approximately 3-fold higher numbers of CD8⁺ BILs when compared to mice treated with GAA-vaccines alone (Fig. 5C). These results strongly suggest that the GAA-vaccines in combination with poly-ICLC, can induce effective, long-term memory responses that are protective against i.c.GL261 progression.

To determine whether the combination therapy with GAA-vaccines and poly-ICLC administration induces demyelination and/or autoimmune encephalitis, we evaluated brain sections obtained from treated mice by LFB staining and H&E staining using standard protocols [30]. Fig. 6 depicts representative stained sections derived from mice treated with the combinational regimen. Highly-myelinated structures, such as corpus callosum and internal capsule, were densely stained with LFB. There was no evidence of demyelination or abnormal immune cell infiltration throughout the brain. We also examined control mice treated with GAA-vaccines alone or with mock-vaccines, and found no evidence of autoimmunity in these animals (data not shown).