|Home | About | Journals | Submit | Contact Us | Français|
Experimental stroke induces a biphasic effect on the immune response that involves early activation of peripheral leukocytes followed by severe immunodepression and atrophy of spleen and thymus. In tandem, the developing infarct is exacerbated by influx of numerous inflammatory cell types, including T and B lymphocytes. These features of stroke prompted our use of Recombinant T Cell Receptor Ligands (RTL), partial MHC class II molecules covalently bound to myelin peptides. We tested the hypothesis that RTL would improve ischemic outcome in brain without exacerbating defects in peripheral immune system function.
Four daily doses of RTL were administered subcutaneously to C57BL/6 mice after middle cerebral artery occlusion (MCAO), and lesion size and cellular composition were assessed in brain, and cell numbers were assessed in spleen and thymus.
Treatment with RTL551 (I-Ab molecule linked to MOG-35−55 peptide) reduced cortical and total stroke lesion size by ~50%, inhibited the accumulation of inflammatory cells, particularly macrophages/activated microglial cells and dendritic cells, and mitigated splenic atrophy. Treatment with RTL1000 (HLA-DR2 moiety linked to human MOG-35−55 peptide) similarly reduced the stroke lesion size in HLA-DR2 transgenic mice. In contrast, control RTL with a non-neuroantigen peptide or a mismatched MHC class II moiety had no effect on stroke lesion size.
These data are the first to demonstrate successful treatment of experimental stroke using a neuroantigen specific immunomodulatory agent administered after ischemia, suggesting therapeutic potential in human stroke.
The role of inflammatory factors and their contribution to tissue injury in stroke has been well established. Following dense cerebral ischemia, there is rapid accumulation of inflammatory cells associated with breakdown of the blood-brain barrier (BBB) (1-3). These infiltrating cells, including neutrophils, macrophages, and lymphocytes (4), exacerbate the evolving ischemic injury. Although the importance of lymphocytes and adaptive immunity is not well-defined, mice lacking T and B cells sustain smaller infarcts than normal mice (5,6). Furthermore, cerebral ischemia engages the systemic immune system, first by massive activation, then by immunosuppression marked by spleen and thymus atrophy and loss of immunocompetent cells (7-11). Accordingly, successful immunotherapy after ischemia would need to reduce inflammatory damage and concomitantly support survival of peripheral immunity.
Recombinant T cell receptor (TCR) Ligands (RTL) are partial MHC II molecules comprised of covalently linked β1 and α1 chains tethered to antigenic peptides. These unique molecules were designed as minimal ligands for the TCR of peptide-specific T cells. Unlike four-domain MHC II molecules that can induce T cell activation and apoptosis (12), RTL are partial agonists that deviate autoreactive T cells to become non-pathogenic (13,14). We previously demonstrated that RTL could prevent and/or reverse clinical signs of experimental autoimmune encephalomyelitis (EAE) (15-17), collagen-induced arthritis (18) and experimental autoimmune uveitis (19). Recent studies with RTL551 (I-Ab molecule covalently tethered to mouse myelin oligodendrocyte glycoprotein - mMOG-35−55 peptide), showed dramatic reversal of MOG-35−55-induced EAE, reduction of pathogenic CNS cells, down-regulation of endothelial cell adhesion molecules (ICAM and VCAM) and broad reduction of CNS chemokines/receptors (20).
One prevailing hypothesis is that cerebral ischemia breaches the BBB, thereby exposing CNS antigens and facilitating development of anti-CNS responses (21). Recent reports suggest that induction of T cell tolerance to CNS antigens by mucosal administration of myelin basic protein (MBP) or mMOG-35−55 peptide can reduce infarct size and CNS autoimmunity and improve clinical outcome in mice with experimental stroke (22-25). Based on these observations and RTL effects on EAE, we hypothesized that treatment of experimental stroke with RTL551 would reduce infarction volume through a neuroantigen-specific mechanism.
The study was conducted in accordance with National Institute of Health guidelines for the use of experimental animals, and the protocols were approved by the Institutional Care and Use Committee. All experiments utilized age matched, sexually mature (20−25g) male mice (C57BL/6, Charles River; or HLA-DRB1*1502 mice obtained from Dr. Chella David, described previously (26,27)). The mice were bred and housed at the Portland Oregon Veterans Affairs Medical Center according to institutional guidelines.
RTL molecules consist of the α1 and β1 domains of MHC II molecules expressed as a single polypeptide with or without antigenic amino terminal extensions (28,29): RTL551 = I-Ab linked to mouse MOG-35−55 peptide (MEVGWYRSPFSRVVHLYRNGK); RTL553 = recombinant I-Ab linked to I-Ea-52−68 (3Kp) peptide (ASFEAQKAKANKAVD (30,31)); RTL342M = HLA-DR2 linked to mouse MOG-35−55 (mMOG35−55) peptide; RTL1000 = HLA-DR2 linked to human MOG-35−55 (hMOG35−55) peptide (MEVGWYRPPFSRVVHLYRNGK). RTL were constructed de novo or by sequential site-directed mutagenesis of previous constructs (20). Protein purification was as previously described (28,29) with a 30 to 40mg yield of purified protein per liter of bacterial cell culture.
Mice were randomized to treatment with RTL 551 0.1 ml (1mg/ml, subcutaneous injection at onset of reperfusion followed by doses at 24, 48 and 72h of reperfusion for a total of 4 treatments), RTL553, RTL342M, RTL1000 or vehicle (Tris-HCl, pH=8.5). To determine if RTL 551 altered arterial blood pressure or blood gases, separate cohorts of mice were dosed with RTL 551 (n=3) or vehicle (n=3), then continuously monitored for two hours. Arterial blood gases, osmolarity and electrolytes were also measured in these cohorts at two hours post-treatment.
Transient MCAO was performed in isofluorane-anesthetized male mice using 60min of intraluminal filament occlusion as described (6,7). Temporalis muscle temperature was maintained at 35.5−37.5°C throughout MCAO surgery with warming lights. Adequacy of occlusion was confirmed by intra-ischemic laser-Doppler flowmetry (Moor Instruments Ltd, Oxford, England). Sham-operated mice experienced identical surgical procedures, exclusive of vascular opening and filament insertion. Tissue was harvested under deep anesthesia at 96h of reperfusion. Cerebral infarct size was determined by standard triphenyltetrazolium chloride (TTC, Sigma Aldrich) staining and digitally quantified by Sigma Scan Pro5 software (Jandel, San Real, CA, USA), as described (7). In additional experiments, mice were saline-perfused via the ascending aorta to remove non-parenchymal cells. Brain was divided into the ischemic right hemisphere and non-ischemic left hemisphere, then dissociated mechanically through a 150-mesh screen. CNS mononuclear cells were isolated by Percoll gradient centrifugation as described (32).
Spleen and thymus were isolated from vehicle- and RTL-treated sham and stroke mice. Single cell suspensions were prepared by passing the tissue through a 100μm nylon mesh screen. Cells were washed using RPMI and red cells were lysed using 1X red cell lysis buffer (eBioscience, San Diego, CA). The cells were washed twice, counted and resuspended in stimulation medium containing 10% FBS. For real-time polymerase chain reaction, splenocytes were pelleted, snap frozen and stored at −80C until tested. For FACS staining, cells were washed with staining medium (1X PBS, 0.5% BSA, 0.02% sodium azide) before adding specific antibodies. Cardiac blood was collected in 3mg/ml EDTA, and the cells were pelleted after lysis of red cells, washed, counted and resuspended in staining medium for FACS staining.
Four-color (FITC, PE, PI, allophycocyanin) fluorescence flow cytometric analyses were performed to determine the phenotypes of brain and blood mononuclear cells. Cells were washed with staining medium and stained with a combination of the following mAbs: CD3e (145−2C11), CD4 (L3T4), CD8 (Ly-2), CD11b (M1/70), CD11c (HL-3) and CD19 (1D3) for 20min on ice. After incubation with mAb, cells were acquired with a FACSCalibur (BD Biosciences). Forward and side scatter parameters were chosen to identify lymphocytes. Dead cells were gated out using propidium iodide discrimination. Data were analyzed using FCS express software (De Novo). For each experiment, cells were stained with appropriate isotype control Abs to establish background staining and to set quadrants before calculating the percentage of positive cells.
Statistical differences between parameters assessed in vehicle and RTL groups were determined by Student's t-test. Values of p≤0.05 were considered significant. Data are presented as means ± SD.
Treatment with RTL 551 did not alter arterial blood pressure over a two hour observation window, and there were no differences at two hours as compared to vehicle-treated mice (Veh) (RTL 80±6 mmHg; Veh 75±8, n=3 per group). Similarly, there were no differences in arterial blood pH (RTL 7.33±0.03; Veh 7.34±0.02), pCO2 (RTL 49±5 mmHg; Veh 41±5) or pO2 (RTL 160±17mmHg; Veh 178±4). Plasma electrolytes, glucose, and osmolarity were also not different in RTL vs. vehicle treated-groups (data not shown). Four daily treatments with RTL551 treatment resulted in ~50% reduction in infarction volume in the cortex (49.1±12.6, vehicle vs. 25.6±18.4, RTL-treated mice, p=0.009) but not in the striatum (Fig. 1A). Total infarction in RTL551-treated mice was 26.5±12.4 vs. 39.7±11.0 in vehicle-treated controls (p=0.04). These data demonstrate that RTL551 can successfully reduce the cortical and total stroke lesion volumes when administered after induction of MCAO.
To evaluate peptide specificity, C57BL/6 mice were treated with RTL553, which has the same MHC moiety as RTL551 (I-Ab) but is linked to a non-neuroantigen peptide (I-Ea-52−68). As is shown in Fig. 1B, four daily treatments of MCAO mice with RTL553 had no effect on infarction volume in the cortex or striatum (39.8±23.4 in cortex and 93.1±9.0 in striatum). To address the contribution of RTL MHC II domains to therapy, MCAO mice were treated with RTL342M, which has the same mMOG-35−55 peptide as RTL551 but a different MHC II moiety (HLA-DR2 vs. I-Ab). Similar to RTL553, treatment of MCAO with RTL342M failed to reduce infarction volume (Fig 1C). These data demonstrate that successful RTL treatment of stroke requires a neuroantigen peptide (mMOG-35−55) and the autologous MHC moiety for the treated mice.
We further evaluated the therapeutic activity of humanized RTL1000, containing the HLA-DR2 moiety covalently linked to hMOG-35−55 peptide, on MCAO in DR2 mice. As shown in Fig. 1D, four treatments of RTL1000 significantly reduced cortical infarction volume (27.7±18.2, vehicle vs. 5.3±5.2, RTL1000-treated mice). These data demonstrate that RTL1000, containing a neuroantigen peptide and a matched MHC II moiety, can significantly reduce stroke lesion size in the cortex of a different mouse strain, further documenting the RTL treatment parameters established above.
We hypothesized that RTL551 treatment would reduce inflammatory cell infiltration into brain after stroke. In fact at 96h post MCAO, RTL551 treatment reduced recovery of total brain mononuclear cells by 64% (from 220,000 to 80,000/mouse) and viable leukocytes by 52% (from 39,600 to 19,200/mouse) from the ischemic ipsilateral hemisphere, with much lower cell counts (~60,000 total brain cells and ~10,000 viable leukocytes) recovered from the unaffected contralateral hemisphere from both vehicle- and RTL-treated mice. Using four-color flow cytometry, we found comparable percentages of most cell types in ischemic and unaffected hemispheres from vehicle- and RTL-treated mice. However, RTL551 treatment markedly down-regulated CD45hiCD11bhi cells that represent macrophages and activated microglia in the ischemic right hemisphere of the brain (44% in vehicle-treated mice vs. 15% in RTL551-treated mice, Fig. 2A). Considering absolute numbers of recovered viable leukocytes, the reduction of CD45hiCD11bhi cells after RTL551 treatment (83%) was even more dramatic (from 17,424 to 2,880/mouse, Fig. 2B and Table 1). RTL551 treatment also caused a similar reduction (81%) in dendritic cells (DC) and less pronounced reductions in absolute numbers of the other cell populations in ischemic brain (−35 to −59%, Table 1). In contrast, there were much lower baseline levels of these populations in the left non-ischemic side of the brain (1.4−26%) that were relatively less affected by RTL551 treatment (−33 to +137%,Table 1).
To evaluate the effects of RTL551 treatment on stroke-induced atrophy, cell numbers in spleen and thymus were counted in post-ischemic vehicle- and RTL-treated mice and compared to their respective sham-operated controls. As expected, MCAO induced a significant decrease in spleen and thymus cell numbers in the vehicle-treated group (Fig. 3). Interestingly, viable cell counts were significantly increased in spleens of RTL551-treated vs. vehicle-treated mice (4.6±3.5, vehicle- vs. 17.1±17.3, RTL-treated mice, p=0.02) 96h after reperfusion. In contrast, treatment with RTL553 or RTL342M did not improve spleen cell counts (data not shown). The drastic reduction in thymic cell counts induced by MCAO was not altered by either RTL551 (Fig. 3) or RTL553 (not shown) vs. vehicle treatment.
Given the partial restoration of spleen cell numbers in RTL551-treated mice, we tracked survival of specific splenic cell types. As shown in Table 2, RTL treatment did not affect the percentages of T cells, B cells, macrophages or DC in the spleen, although DC were decreased in vehicle-treated MCAO- vs. sham-operated mice. As is shown in Table 3, we confirmed our previous finding (8) that MCAO increased circulating CD11b+ macrophages in vehicle-treated MCAO vs. sham-operated mice and further noted that RTL treatment limited this increase (+21% in vehicle-treated MCAO mice vs. +11% in RTL551-treated MCAO mice).
Our study demonstrates for the first time that treatment with RTL after onset of MCAO reduces cortical and total infarct size, inhibits infiltrating inflammatory cells, particularly activated macrophages/microglial cells and DC, into post-ischemic brain and partially preserves spleen cell numbers that are typically ablated after MCAO. This result was specific to RTL551 treatment in C57BL/6 male mice and verified using a ‘humanized’ RTL1000 construct to treat MCAO in HLA-DR2-Tg mice. The results clearly show that the therapeutic activity of RTL requires a neuroantigen peptide (mouse or human MOG-35−55) tethered to an MHC moiety that closely matches the class II of the treated mouse strain (I-Ab for C57BL/6 mice and HLA-DR2 for DR2 Tg mice). In contrast, treatment of C57BL/6 mice with RTL553 comprised of I-Ab coupled to I-Ea-52−68 (a non-neuroantigen peptide) and RTL342M comprised of HLA-DR2 (non-matched MHC class II) coupled to mMOG-35−55 peptide did not have therapeutic effects.
Beyond the RTL-bound MOG-35−55 peptides that are known to induce inflammation in C57BL/6 and DR2 mouse strains (20,27), it is unknown what other neuroantigen peptides might also be effective in treating MCAO. It seems likely that the key features for therapeutic activity in stroke are antigenicity in combination with self MHC II molecules and expression in infracted areas of the brain. This important issue awaits additional study. The need for a matched MHC moiety in the treating RTL suggests that therapeutic activity may require direct ligation of neuroantigen-reactive T cells by the intact RTL molecule rather than internalization and re-presentation of the RTL-bound MOG peptide by host APC.
The influx of inflammatory cells into brain after infarct was evaluated in detail in a previous study (4). Although T cells, B cells, neutrophils and macrophages were all present, the predominant infiltrating population observed at 48, 72 and 96h after occlusion was the macrophage/activated microglial cell population that can be discerned by FACS as CD45hiCD11bhi. This population in our study comprised ~45% of the viable leukocyte population at 96h post-occlusion, in agreement with the Stevens study. Concurrent with the reduction in the infarct size, RTL551 treatment reduced all of the infiltrating cell types in the right brain hemisphere, but particularly inhibited (by >80%) the infiltration of CD45hiCD11bhi macrophages/activated microglial cells and CD11c DC. This result in stroke is quite similar to the previously-described effect of RTLs in blocking cellular infiltration into the CNS during experimental multiple sclerosis (20). This decrease in emigrating cells from the periphery may also help to explain RTL-induced preservation of spleen cell numbers during the second immunosuppressive stage of stroke.
The major unresolved question regarding the ability of RTL to inhibit lesion size after MCAO is establishing the conceptual framework for how it might work. The developing lesion in experimental stroke is propagated in part by infiltration of inflammatory cells, including T cells, B cells, neutrophils and activated macrophages. Our previous studies demonstrated reduced cortical stroke lesion size in T and B cell deficient SCID mice (6), and studies by others implicated T cells but not B cells or neutrophils as necessary inflammatory contributors (5). Moreover, mucosal induction of T cell tolerance to myelin antigens or transfer of tolerized T cells to naïve mice undergoing MCAO resulted in reduced infarct size and number of IFN-γ producing cells and an increase in cells secreting IL-10 and TGF-β (21-23, 25).
Our working hypothesis is that normal brain antigens such as MOG are released in higher quantities in the stroke lesion and presented to the immune infiltrating cells in an inflammatory environment, probably after ingestion of damaged brain tissue by phagocytic cells such as microglia, macrophages and dendritic cells. The rapid brain-to-spleen signaling that increases general activation of peripheral immune cells within 6h of reperfusion (7) might allow activated T cells of many different specificities to infiltrate the stroke lesion site. Those T cells that are specific for MOG or other brain antigens might then be triggered to recruit other inflammatory cells into the lesion and thereby increase damage to the surrounding tissues. This scenario represents to some degree a native ability of T cells to recognize and respond to self antigens, not dissimilar in concept to the natural IgM antibodies specific for nonmuscle myosin heavy chain type II A and C implicated in reperfusion injury in skeletal and intestinal reperfusion injury described by Zhang et al. (33).
It seems likely that RTL treatment induces T cell tolerance to MOG-35−55 peptide in vivo through a different mechanism but with the same outcome as that induced by mucosal administration of MBP (22) or MOG-35−55 peptide (23). We established previously that RTL specifically targets myelin specific T cells and profoundly changes their functional properties from pro-inflammatory to anti-inflammatory cells that secrete IL-10, IL-13, and TGF-β3 (16). As such, the RTL-‘tolerized’ MOG-specific T cells could inhibit entry of other T cells into the CNS, prevent release of pro-inflammatory cytokines at the stroke lesion site and reduce infiltration of activated macrophages/microglial cells. This explanation for RTL inhibition of stroke is supported our previous observation that mice pre-treated with RTL prior to immunization with neuroantigens are profoundly protected against subsequent induction of EAE (34). Moreover, by short-circuiting early brain-to-spleen activation, RTL modulation of MOG-specific T cells may have contributed to the observed partial preservation of spleen cell numbers.
In summary, we here demonstrated therapeutic activity of myelin antigen specific immunomodulatory agents (recombinant TCR ligands, RTL551 and RTL1000) that were effective when administered after induction of MCAO. RTL therapy greatly reduced influx of inflammatory cells, particularly macrophages and DC, into the ischemic hemisphere and partially prevented splenic atrophy that accompanies the downstream immunosuppressive phase of stroke. Future studies are underway to determine how long after MCAO the RTL can be given and still exert therapeutic benefit. RTL1000, a humanized RTL comprised of MOG-35−55 peptide covalently linked to an HLA-DR2 moiety, is currently in clinical trials in MS, and potentially might be useful for treatment of human stroke patients.
The authors thank Ms. Eva Niehaus for preparing the manuscript. This work was supported by US Public Health Service NIH grants NS33668, NRO3521, NS49210, AI43960, and NS47661; National Multiple Sclerosis Society grant RG379-A-4; The Nancy Davis Center Without Walls; and the Biomedical Laboratory R&D Service, Department of Veterans’ Affairs.
CONFLICT OF INTEREST DISCLOSURE
Drs. Offner, Vandenbark, and OHSU have a significant financial interest in Artielle ImmunoTherapeutics, Inc., a company that may have a commercial interest in the result of this research and technology. This potential conflict of interest has been reviewed and managed by the OHSU and Portland VAMC Conflict of Interest in Research Committees.