Nonmyelin-Specific T Cells Accelerate Development of Central Nervous System APC and Increase Susceptibility to Experimental Autoimmune Encephalomyelitis¹ Free

Multiple sclerosis (MS)³ is a debilitating, paralytic disease characterized by inflammation, demyelination, and other harmful pathologic processes in the CNS (1). Studies in experimental autoimmune encephalomyelitis (EAE) have demonstrated that myelin-specific T cells cause autoimmune CNS inflammation and paralysis in laboratory rodents (2, 3, 4, 5). Immune status studies have suggested that myelin-specific T cells also cause the clinical paralysis of MS (6, 7). However, under a variety of conditions, myelin-specific rodent T cells do not induce EAE (8, 9, 10) and many healthy humans possess circulating myelin-specific T cells (11, 12, 13). Furthermore, histocompatible APC in the CNS are required for myelin-specific T cells to induce disease (14, 15). Therefore, the presence and activity of myelin-specific T cells are necessary but not sufficient to cause CNS inflammation. In addition to myelin-specific T cells, disease-potentiating characteristics such as the presence and activity of CNS APC control susceptibility to CNS inflammation.

Using the interspecies bone marrow cell (BMC) and T cell cotransfer method in EAE-susceptible SCID mice, it was shown previously that nonmyelin-specific T cells promoted the development of functional, bone marrow (BM)-derived APC in the CNS (16). Since CNS APC are required for EAE (14, 15), these results suggested that previous immune responses to nonmyelin Ags might be capable of augmenting susceptibility to CNS inflammation induced subsequently by myelin-specific T cells. For this report, the Lewis rat-SCID mouse xenogeneic cell transfer model of EAE was used to evaluate the influence of nonmyelin specific T cells on the severity and course of disease induced by myelin-specific T cells. In the presence of chicken OVA-specific T cells, formation of BM-derived CNS APC was enhanced and mice were much more susceptible to disease induction compared with the absence of OVA-specific T cells. A higher disease incidence, an earlier day of onset, a longer duration, a higher peak severity, and a higher cumulative disease index (CDI) each indicated a highly significant enhancement of disease susceptibility associated with nonmyelin-specific T cells. Furthermore, the number of myelin-specific encephalitogenic T cells required to induce EAE was greatly reduced in the presence of nonmyelin-specific T cells.

Six- to 8-wk-old female Lewis rats were obtained from Harlan Sprague Dawley (Indianapolis, IN). Inbred female SCID mice (8–11 wk old) were obtained from the SCID mouse breeding facility within the Veterinary Medical Unit at the Veterans Affairs Medical Center. Animal care was in accordance with institutional guidelines.

Eight rats per line were each immunized by s.c. injection with a synthetic peptide of basic protein (BP) (BP87–99), chicken OVA (100 μg Ag/rat), or no Ag in CFA containing 100 μg of heat-inactivated Mycobacterium tuberculosis strain H37RA (Difco, Detroit, MI). Nine days after immunization, draining lymph nodes (LN) were collected and a single-cell suspension of LN cells was prepared. LN cells were stimulated in vitro (8 × 10⁶ cells/ml) for 3 days with Ag (BP87–99 or chicken OVA; 200 μg/ml) depending on the desired specificity of the line. Stimulated cells were expanded in rIL-2-containing medium for 5–7 days. This was followed by restimulation of the T cells with Ag in the presence of Lewis rat irradiated thymocytes for 3 days. Cells were selected and maintained through alternate cycles of stimulation with Ag and expansion in IL-2. Ag specificity was verified for each cell line during Ag stimulation using an in vitro thymidine uptake proliferation assay, and the disease-inducing potential of the T cell lines was confirmed by adoptive transfer of activated T cells (10–20 × 10⁶ cells/rat, i.p. injection) into Lewis rats (data not shown). T cell lines were evaluated for expression of the indicated phenotype markers by fluorescent mAb staining and flow cytometry.

All experiments involving rat BMC utilized T cell-depleted BMC. Two or three adult female Lewis rats were used as BM donors. The femurs and tibias were dissected free of surrounding tissue and collected in ice-cold RPMI 1640 medium. After cutting off the ends of each bone, BMC were flushed out with ice-cold RPMI 1640 using a 1-ml syringe. BMC were pelleted by centrifugation, resuspended in 10 ml of ice-cold RPMI 1640, counted, and then incubated with a 1/160 dilution (final) of mouse anti-rat CD3 mAb (BD PharMingen, San Diego, CA) for 30 min on ice. Following this, cells were washed and incubated with goat anti-rat IgG Ab-coated magnetic beads (Miltenyi Biotec, Auburn, CA). Cells were then washed and run through a magnetized column (Miltenyi Biotec) that had been pretreated with PBS/4% BSA for 30 min. The column was washed five to six times with elution medium (PBS/0.5% BSA/2 mM EDTA) to elute all unbound, CD3-negative cells. The column was then demagnetized by removal from the magnet and CD3-positive cells were eluted from the column. The total yield obtained from the column depletion procedure was 80- 95%. Undepleted, T cell-depleted, and T cell-enriched BM fractions were stained with mAbs specific for rat T cells to verify the purity of the respective populations and to evaluate the effectiveness of the magnetic sort. After T cell depletion, it was not possible to detect T cells in the T cell-depleted BMC fraction by fluorescent mAb staining and flow cytometry using Abs specific for rat CD3. BMC were evaluated for expression of the indicated phenotype markers by fluorescent mAb staining.

The indicated number of Lewis rat T cell-depleted BMC were injected (0.2 ml/mouse) by i.v. injection into anesthetized, irradiated (300–350 Gy) C.B-17/scid mice. BMC were injected either alone or within 5 min of Lewis rat OVA-specific T cells (0.2 ml/mouse, i.p.). Seven days after injection of BMC, mice received an injection of BP-specific, encephalitogenic rat T cells (0.2 ml/mouse, i.p.). Groups of mice were followed daily and compared for the presence of clinical neurologic deficit (paralysis).

Mice were evaluated daily for disease and scored as follows: 0, normal; 1, limp tail or mild hind limb weakness; 2, limp tail and moderate hind limb weakness or mild ataxia; 3, moderately severe hind limb weakness; 4, limp tail and severe hind limb weakness or moderate ataxia; 5, paraplegia with no more than moderate forelimb weakness; or 6, limp tail and paraplegia with severe forelimb weakness or severe ataxia.

Spinal cord cells from SCID mice were evaluated for the presence of rat cells using rat-specific fluorescent mAbs (BD PharMingen) specific for rat myelomonocytic cells (CD11b/c, clone OX-42, noncross-reactive with mouse cells), T cells (CD3, clone G4.18), and rat class II MHC (RT-1B, clone OX-6). A single-cell suspension of spinal cord cells was isolated from euthanized donor mice. The cells from a single spinal cord were resuspended in 5 ml of 80% Percoll (Pharmacia, Peapeck, NJ) and overlaid with 5 ml of 40% Percoll in a 15-ml centrifuge tube. Cells were centrifuged for 30 min at 1600 rpm and cells from the 80/40 interface were collected. Cells were washed, counted, and aliquoted for mAb staining. Cells were incubated with fluorescent mAbs (1 × 10⁵ cells/0.2 ml/tube) for 20 min on ice. Reactivity with specific mAbs was evaluated by flow cytometry using a BD Biosciences FACScan (Mountain View, CA). Dot plots of fluorescence intensity for FITC- and PE-labeled cells were evaluated using CellQuest software (BD Biosciences) to identify and quantitate distinct populations of cells. Isotype-matched (control) fluorescent Abs were used at the same concentration as each specific Ab for every experiment. Percentage of true positive staining was determined by subtracting the percentage of the isotype control staining (background) from the percent staining of each specific mAb.

RT-PCR was used with rat-specific cytokine primer pairs to detect cytokine-specific mRNA in cultured, OVA-specific or BP-specific rat T cells. mRNA was isolated from an in vitro stimulation culture containing 10 × 10⁶ Lewis rat T cells. Frozen pelleted cells were treated with 1 ml of TRIzol reagent (phenol/guanidine isothiocyanate; Molecular Research Center, Cincinnati, OH,). After 5 min at room temperature, 0.2 ml of chloroform was added. After 15 min, the samples were centrifuged at 12,000 rpm for 15 min at 4°C. The aqueous phase was removed to a new tube and 0.5 ml of isopropanol was added. After mixing, mRNA was stored for 10 min at room temperature. This was centrifuged at 12,000 rpm for 8 min at 4°C. The supernatant was carefully aspirated away and the remaining pellet was mixed with 1 ml of 70% (v/v) ethyl alcohol and spun at 7500 rpm for 5 min at 4°C. After aspirating away the ethyl alcohol, the pellet was air dried and resuspended in 10 μl of diethyl pyrocarbonate-treated water. This was transferred to a 0.65-ml tube and 2 μl of random primers (Life Technologies, Rockville, MD) was added. This was heated to 70°C for 10 min and then quick chilled on ice for 3 min. To this was added 4 μl of first-strand buffer (Life Technologies), 2 μl of 0.1 M DTT, 1 μl of 10 mM dNTP (nucleotides), and 1 μl of reverse transcriptase (Superscript II; Life Technologies). The mixture was incubated for 1 h at 42°C. The reaction was terminated by incubation at 70°C for 15 min. This 20-μl volume of cDNA was aliquoted and stored at −80°C for later use in the PCR. For the PCR, 0.5 μl of cDNA was added to each PCR tube containing 2.5 μl of 10× PCR buffer (Life Technologies), 0.5 μl of 50 mM MgCl₂, 0.5 μl of 10 mM dNTP, 0.5 μl of 20 mM specific 5′ primer, 0.5 μl of 20 mM specific 3′ primer, 2.5 μl of Taq polymerase enzyme (Life Technologies), and 17.5 μl of diethyl pyrocarbonate-treated water. Samples were amplified on a thermocycler (PerkinElmer 9600, PerkinElmer, Norwalk, CT) for 30 or 35 cycles (depending on the primer): 94°C, 45 s melting; 57°C, 45 s annealing; and 72°C, 90 s elongation. From each tube, 10 μl of the amplified DNA was added to each lane of a 1.5% agarose gel. PCR products were run for ∼1 h at constant voltage (96 V). A Polaroid photograph was taken of each gel using a UV illuminator. PCR primers were specific for rat sequences and were designed by Dr. T. Keller (Oregon Health and Science University Molecular Biology Core Service, Portland, OR) to be noncross-reactive with mouse sequences. Detection of specific PCR products of the correct predicted molecular size was dependent on the addition of rat cDNA.

Groups of SCID mice were prepared by transferring by injection T-depleted BMC alone or BMC with OVA-specific T cells (Fig. 1). Migration of BM-derived CD11b/c⁺ rat cells in the mouse CNS was evaluated for each group 7–11 days following BMC transfer in the presence and absence of OVA-specific rat T cells (Fig. 2). On days 8–11, the percentage of CD11b/c⁺ BM-derived rat cells was greatly elevated in the spinal cord of individual mice that received BMC injected along with OVA T cells. The proportion of BM-derived cells in the CNS was ∼5- to 10-fold greater during this time period in the animals that received OVA-specific T cells.

Groups of mice that had received BMC alone or BMC with OVA-specific T cells were compared for susceptibility to EAE following i.p. injection of BP-specific T cells 7 days after injection of BMC with or without OVA-specific T cells (Fig. 3). Two, 8, or 15 × 10⁶ myelin-specific T cells (Fig. 3) induced a more severe paralytic disease course in recipients of BMC and OVA-specific T cells compared with recipients of BMC alone (Fig. 2). Disease severity in the two groups was also evaluated by comparing disease incidence, day of onset, disease duration (days), mean maximal (peak) severity, and CDI (Table I and Fig. 4). Groups of animals that received BMC and OVA-specific T cells exhibited a higher disease incidence, an earlier day of onset, a longer disease duration, a higher mean maximal disease severity, and a higher CDI compared with groups of mice that received BMC without OVA-specific T cells (Table I and Fig. 3). The number of encephalitogenic, BP-specific T cells required to induce a particular level of disease was much less in recipients of BMC plus OVA-specific T cells compared with recipients of BMC alone (Fig. 4). Spinal cord inflammation was more severe in recipients of BMC and OVA-specific T cells compared with recipients of BMC alone (Fig. 5).

The nonmyelin (OVA)-specific and myelin (BP)-specific T cells were compared to assess the likelihood that nonmyelin-specific and myelin-specific T cells might differ in their potential to augment susceptibility to disease. OVA-specific and BP-specific T cells appeared to be very similar with respect to expression of CD4 (Fig. 6,A) and CD49d (Fig. 6,B). The OVA- and BP-specific T cell lines expressed similar cytokines at detectable levels (Fig. 6,D). Although the PCR products for monocyte chemoattractant protein- 1, stromal-derived factor-1β, IL-10, and TNF-β appeared to be somewhat greater from the OVA-specific T cells, these quantitative differences are not significant since the assay was not performed in a quantitative fashion. Neither cell line expressed detectable IL-4 mRNA when assayed by RT-PCR. MHC class II expression was difficult to assess based on its low level of expression and therefore appeared to vary only slightly between the two T cell lines, with a higher percentage of the BP-specific line expressing MHC class II (RT-1B, Fig. 6,C). OVA-selected and BP87–99-selected rat T cell lines were specific for their respective selecting stimulus and did not cross-react (Fig. 7). The elevated susceptibility to disease in BMC plus OVA-specific T cell recipients depended on BMC and BP-specific T cells and was not merely due to the animals having more T cells since mice that received 40 × 10⁶ OVA-specific T cells and 40 × 10⁶ BP-specific T cells failed to develop EAE (Table II).

The results reported here provide experimental evidence of a positive association between three conditions: 1) the presence and activity of nonmyelin-specific T cells in the CNS before disease; 2) an increase in the number of BM-derived cells in the CNS; and 3) a heightened susceptibility to paralytic CNS inflammation induced indirectly by myelin-specific T cells. Thus, nonmyelin-specific T cells appear to elicit a heightened susceptibility to CNS inflammation by promoting an elevation in the number and function of BM-derived APC in the CNS. The results also establish a crucial, prepathogenic role for circulating BM-derived APC precursors in disease susceptibility induced by nonmyelin-specific T cells.

Enhanced susceptibility to CNS inflammation mediated by nonmyelin-specific T cells may be relevant to the etiology of CNS inflammation in MS. Conditions that cause the onset of new disease and that determine the progression of clinical disease in MS are not adequately understood (17). Certain crucial immune system characteristics are thought to be important, including the number, frequency, phenotype, specificity, location, and activity of effector and regulatory myelin-specific and nonspecific T cells (18). The results reported here reveal a distinct, disease-enhancing functional capability of nonmyelin-specific T cells during periods preceding disease onset. Thus, immune responses to nonmyelin Ags (during infections or following other common antigenic challenges) could exert chronic and/or acute influences over susceptibility to T cell-induced inflammation by promoting a shift in conditions within the CNS whereby the pathogenic functions of myelin-specific T cells are more easily triggered. Although the molecular signals which trigger this distinct activity in nonmyelin-specific T cells have not been defined, it seems reasonable and likely that these are simply the same signals provided by the in vitro conditions used to generate the nonmyelin-specific and myelin-specific T cell lines and are the same signals as are present during Ag activation in the peripheral LN (19).

Because of the apparent in vitro phenotypic similarities between the OVA- and BP-specific T cells and because of the previously demonstrated ability of BP-specific T cells to recruit CD11b⁺ BM-derived cells to the CNS in the same time frame as nonmyelin-specific T cells (16), it seems most likely that nonmyelin-specific and myelin-specific T cells are similar with regard to their potential to recruit circulating BM-derived myelomonocytic lineage cells to the CNS before Ag recognition. Thus, the molecular recognition of specific myelin Ag and MHC is not required for enhanced formation of BM-derived CNS APC and enhanced susceptibility to CNS inflammation mediated by either nonmyelin-specific or myelin-specific T cells. This raises the fascinating possibility that, in addition to myelin-specific T cells, certain nonmyelin-specific T cells present in the peripheral lymphoid tissues and CNS may participate in the disease process even in the absence of stimulation by specific neural Ag. Such a shared, early role for myelin-specific and nonmyelin-specific T cells could involve direct or indirect, local, or systemic effects on APC development and/or blood-brain barrier permeability. We are currently pursuing these possibilities to define more precisely the cellular and molecular interactions responsible for altering susceptibility to disease.

Migration of nonmyelin-specific T cells into the CNS has been recognized for decades in multiple rodent models (20, 21). Enhancement of disease severity by nonmyelin-specific T cells has been observed previously in immunocompetent rodents (22) and nonmyelin-specific T cells in the CNS have been proposed to play a pathologic role in EAE and MS (23). However, since the experiments here were conducted in SCID mice in the presence of only a very limited, defined T cell compartment, the results leave open the possibility that nonmyelin-specific T cells might be subject to regulation or other influences in immunocompetent animals and humans. As examples of this, induction of an immune response to nonmyelin Ag protected immunocompetent SJL mice from disease (24, 25) and conditions or treatments which induced suppressor T cells inhibited EAE in immunocompetent rodent strains (26, 27). We are currently conducting experiments to evaluate these possibilities in this model of EAE.

The current results should also be considered in the context of previous clinical epidemiological studies since there has been considerable interest in the reported associations among infections, immunizations, and/or other environmental exposures and MS (28). Although a specific environmental agent which is causative in MS has not been identified (29, 30), several have been suggested and experimental studies in EAE have provided compelling evidence supporting the relevance of various infectious or environmental agents as etiologic triggers for the pathogenic, neural Ag-specific immune reactions which lead to CNS inflammation (31). The results presented here demonstrate a cellular mechanism whereby immune challenge by various environmental agents (non-neural Ags) might induce nonpathologic changes within the CNS which enhance susceptibility to CNS inflammation mediated subsequently by neural Ag-specific T cells.

The results show that formation of BM-derived CNS APC and susceptibility to EAE are reduced when nonmyelin-specific T cells are absent compared with when they are present. This suggests that treatments which decrease the number and activity of nonmyelin-specific T cells may provide clinical benefit by reducing the severity of inflammation and paralysis caused by myelin-specific T cells. Furthermore, the results provide a rationale for understanding how acute changes in immune status following environmental exposures might precipitate an encephalitogenic event in MS (30). For these reasons, it will be important to consider the number, quality, location, and function of nonmyelin-specific T cells as crucial susceptibility factors in CNS inflammation.

This work was conducted at the Veterinary Medical Unit and Research Service, U.S. Department of Veterans Affairs Medical Center (Portland, OR). Founder mice for the C.B-17 SCID mouse breeding colony were kindly provided by M. Bosma. Histological specimens were prepared by Carolyn Gendron (Pathology Laboratory, Oregon Cancer Institute) and photographed by Michael Moody (Medical Media, Portland Veterans Affairs Medical Center).