CDR3 Size Spectratyping and Sequencing of Spectratype-Derived TCR of Spinal Cord T Cells in Autoimmune Encephalomyelitis¹

Experimental autoimmune encephalomyelitis (EAE)³ is inducible in susceptible strains of rats and mice by immunization with neuroantigens such as myelin basic protein (MBP) or by adoptive transfer of MBP-specific T cells into naive animals. Recent analysis indicates that EAE-inducing T cells bear CD4 molecules and use a limited number of α- and β-chains of the TCR (1). Furthermore, the complementarity-determining region 3 (CDR3) of TCR of long term-cultured encephalitogenic T cell clones is rather short, and some amino acid residues are conservatively preserved (2, 3). In the rat, encephalitogenic T cell clones established from guinea pig MBP-immunized animals mainly use Vβ8.2 (4). Moreover, we (5) and others (6) have demonstrated by immunohistochemical and flow cytometric studies that Vβ8.2⁺ T cells infiltrate the central nervous system (CNS) at the early stage of EAE and become a predominant population throughout the course of EAE. However, more detailed analysis by these methods is difficult because a complete set of mAb specific for rat TCR β-chain family is not available at present. Conventional Vβ analysis using PCR products of TCR also has limitations. As clearly demonstrated by Karin et al. (7), diverse Vβ gene transcripts were detected in the CNS with full blown EAE even though Vβ8.2⁺ T cells are still a predominant population.

The determination of CDR3 size by spectratyping is a powerful tool with which to analyze the T cell repertoire under normal and pathologic conditions (8). First, this method provides a profile of the T cell repertoire without any need for in vitro T cell expansion. Second, oligoclonal expansion of T cells that bear a particular TCR phenotype in mixtures of lymphocytes is readily detectable as a dense band on a gel. Thus, we applied this method to analysis of the repertoire of spinal cord T cells to determine whether T cells in the target organ show characteristics similar to long term-cultured encephalitogenic T cells. For this purpose, cDNA prepared from spinal cord tissue with inflammatory lesions at various stages of actively induced EAE was amplified and analyzed by CDR3 size spectratyping. Furthermore, cDNA of TCR showing oligoclonal expansion was extracted from bands on a gel, cloned, and sequenced. It was revealed that oligoclonal expansion of Vβ8.2 with a short CDR3 occurred throughout the course of EAE and that the majority of Vβ8.2 spectratype-derived cDNA clones had the same nucleotide sequence as that of in vitro-established encephalitogenic T cells.

Lewis rats were purchased from Seiwa (Fukuoka, Japan) and used at 8 to 12 wk of age.

Active EAE was induced in Lewis rats as described previously (9). Each rat was injected in the hind footpads on both sides with an emulsion containing 100 μg of guinea pig MBP in CFA (M. tuberculosis H37Ra, 5 mg/ml). The clinical stage of EAE was divided into four (grade 1, floppy tail; grade 2, mild paraparesis; grade 3, severe paraparesis; grade 4, tetraparesis or moribund condition) (10). At different times, rats were killed under ether anesthesia, and several segments of the lumbar spinal cord were snap-frozen in OCT compound. Twenty sections 20 μm thick were cut in a cryostat and stored at −80°C until use. In this study, the early, peak, and recovery stages of EAE generally refer to days 10 to 11 (grade 1), days 13 to 14 (grade 3 or 4), and day 18 to 20 (grade 0), respectively.

RNA was extracted from frozen sections using RNAzol B (Biotecx Lab, Houston, TX). cDNA was then synthesized by reverse transcription with SuperScript Preamplification System (Life Technologies, Gaithersburg, MD) and amplified in a thermal cycler (Perkin Elmer, Norwalk, CT) using Amplitaq Gold (Perkin Elmer) and primer pairs for TCR. Cycling conditions for PCR and nested PCR were as follows: 95°C for 10 min for denaturation and hot start; 55°C for 1 min for annealing; and 72°C for 1 min for extension followed by 40 cycles of 95°C for 1 min, 55°C for 1 min, and 72°C for 1 min. Primers for Vβ1–20 were the same as those used in the previous study (5). Two types of Cβ primers, Cβ outer (5′-TGTTTGTCTGCGATCTCTGC-3′) and Cβ inner (5′-TCTGCTTCTGATGGCTCA-3′), were used in this study. They were labeled with Cy-5 or rhodamine or remained unlabeled.

CDR3 size spectratyping was performed as described previously (11) with a few modifications. cDNA was amplified with Vβ-specific and rhodamine-labeled Cβ outer primers, and undiluted or diluted PCR products were added to an equal volume of formamide/dye loading buffer and heated at 94°C for 2 min. Two microliters of the samples were applied to a 6% acrylamide sequencing gel. Gels were run at 30 W for 3 h and 30 min at 50°C. Then, the fluorescence-labeled DNA profile on the gel was directly recorded using an FMBIO fluorescence image analyzer (Hitachi, Yokohama, Japan). Spectratypes revealed by this analysis usually consisted of 5 to 7 bands. We designated each band as band I, II, or III in order of molecular size; e.g., the band representing the smallest Vβ8.2 PCR products was Vβ8.2 band I.

cDNA in PCR products or isolated from bands (in most cases, band I) on the acrylamide gel was reamplified with Vβ and unlabeled Cβ inner primers. Then, PCR products were ligated into pT-Adv vector and cloned using the AdvanTAge PCR Cloning Kit (Clontech Laboratories, Palo Alto, CA) according to the manufacturer’s instruction. The plasmid DNA was then sequenced using Cy5-labeled Cβ inner primer and an Autoread Sequencing Kit on an ALFexpress DNA sequencer (Pharmacia Biotech, Tokyo, Japan). CDR3 length is defined as a region starting from an amino acid residue after the CASS sequence of most Vβ segments and ending before the GXG box in the Jβ region as described previously (12). Thus, Vβ-CASSDSSYEQYFGPG-Jβ would count as eight amino acids.

We performed CDR3 size spectratyping of TCR at the early, peak, and recovery stages of EAE. At least three rats were individually examined at each time point, and essentially the same results were obtained. Representative results are shown in Figures 1, 2 and 3. It is known that encephalitogenic, but not nonencephalitogenic, T cells have a short CDR3 in their TCR (2, 3). Therefore, we first checked the profile of Vβ1–20 spectratypes at the early stage of EAE (Fig. 1). Although dense bands indicating oligoclonal expansion of TCR were recognized in Vβ2, Vβ4, Vβ8.2, Vβ9, Vβ14, Vβ15, Vβ17, and Vβ19, only Vβ8.2 and Vβ17 showed expansion of TCR with the shortest CDR3 (indicated by an arrow and an arrowhead, respectively). At the peak stage of EAE, the oligoclonal expansion seen at the early stage became less pronounced, and each Vβ generally had five to seven spectratypes with one or two dense bands in their middle portion (Fig. 2). Vβ8.2 was the only exception. Spectratypes consisting of Vβ8.2 were rather rare compared with other Vβs, and the band representing the shortest CDR3 was very dense (arrow in Fig. 2). With regard to Vβ17, the spectratype profile seen at the early stage did not persist and at the peak stage exhibited a pattern typical of other Vβs except Vβ8.2 (in Fig. 2, arrowhead). We wished to establish whether there are clones that expand at the recovery stage; such clones would suggest the presence of regulatory T cells in the target organ. However, no newly expanded clones (except Vβ8.2) were detected at this stage, and the spectratype pattern was essentially the same as that at the peak stage (Fig. 3). In Figure 4, spectratypes of Vβ8.2 and Vβ17 at the early, peak, and recovery stages of EAE are depicted together. The spectratype pattern of Vβ8.2 remained unchanged throughout the course of EAE, and band I always showed oligoclonal expansion. In sharp contrast, Vβ17 spectratypes showed oligoclonal expansion only at the early stage and diversified later on (Fig. 4).

On the basis of the findings obtained by CDR3 size spectratyping, we decided to determine the nucleotide sequence of the CDR3 of Vβ8.2 and Vβ17 spectratypes. For this purpose, cDNA was extracted from Bands I of Vβ8.2 and Vβ17 on a polyacrylamide gel and reamplified by nested PCR. The cDNA was then cloned, and the nucleotide and amino acid sequences of the CDR3 of each clone were determined. As shown in Table IA, a large number of clones possessed the DSSYEQYF sequence which was frequently seen in in vitro-established encephalitogenic T cell clones (2, 3) at the early stage of EAE (83% of sample 1518 and 63% of sample 1780, average 71%). The average percentage of clones that possessed this sequence was much higher than that of randomly sequenced spinal cord-derived clones. In the latter case, clones that had the DSSYEQYF sequence accounted for 34% (13). More interestingly, all the clones without DSSYEQYE possessed Asp-Ser or (X)-Ser at positions 1 and 2 which are also reported as common motifs of encephalitogenic T cells (2). The T cell repertoire of Vβ8.2 band I-derived clones at the peak stage of EAE was essentially the same as that at the early stage (Table IB). The predominance of the DSSYEQYF sequence in spectratype-derived cDNA clones persisted at the recovery stage, although it became less obvious (Table IC). TCR that did not have the common motif (DVWETTQYF) was the second largest population. With regard to Vβ17, the predominance of a certain sequence was not observed (Table II). However, all the sequenced clones possessed His and Gly at positions 1 and 2.

Recent studies have demonstrated that the TCR of encephalitogenic T cells in the rat has several unique characteristics. First, using in vitro-established T cell clones, it was demonstrated that encephalitogenic T cells mainly use Vβ8.2 (4). Later, the predominance of Vβ8.2⁺ T cells was confirmed in situ in the CNS of Lewis rats with EAE by immunohistochemistry (5, 14). However, Sun et al. claimed that TCR β-chain usage of encephalitogenic T cells is not restricted to Vβ8.2 (15, 16). Indeed, transfer of MBP-reactive Vβ10⁺ T cells induced EAE indistinguishable from that induced by Vβ8.2⁺ T cells (17). Recently, it was clearly demonstrated that preferential usage of Vβ8.2 by encephalitogenic T cells is shaped in the intact Lewis thymus (18). Therefore, the non-Vβ8.2 T cells involved in the induction of EAE seem to make up only minor populations. Second, the TCR of encephalitogenic T cells has an unique CDR3. Gold et al. (2) and Zhang et al. (3) first described that the TCR of encephalitogenic Vβ8.2⁺ T cells has a short CDR3, suggesting that these T cells are ontogenically old and generated before terminal deoxynucleotidyl transferase appears, i.e., before day 4 after birth (19, 20). Furthermore, positions 1 and 2 were highly preserved and were Asp-Ser or (X)-Ser in most cases (3, 17). We (13) and others (21) reported that these common motifs were also frequently found in the CDR3 of TCR isolated from the spinal cord, but not from the lymph node.

Although these studies, i.e., random sequencing of CDR3 using T cell clones and lines or T cells isolated from the target organ, provided useful information, they have limitations. In the present study, we used CDR3 size spectratyping to detect oligoclonal expansion of certain Vβ transcripts. As shown in Figures 1, 2 and 3, screening of a complete set of Vβ family transcripts could be easily achieved using uncultured samples, thereby avoiding the bias that can be produced from culture. CDR3 size spectratyping performed at the early, peak, and recovery stages of EAE revealed several intriguing findings. First, oligoclonal expansion of Vβ8.2 with a short CDR3 persisted throughout the course of EAE. This finding is, at least in part, attributable to the fact that encephalitogenic T cells always recognize the immunodominant epitope of MBP (68–88 sequence) in MBP-induced EAE in the rat and that there is no inter- or intramolecular determinant spread (22). Even under such conditions, it was difficult by routine RT-PCR analysis to find Vβ8.2 dominance especially at the peak or later stage of the disease (7). Second, oligoclonal expansion similar to Vβ8.2 was seen in Vβ17 transcripts at the early stage. Since T cells reacting with MBP89–101, an encephalitogenic peptide sequence in both mice and rats (23, 22), bear Vβ17 in SJL/J mice (24), it is possible that Vβ17⁺ T cells in Lewis rats are also encephalitogenic. Last, no Vβ other than Vβ8.2 showed oligoclonal expansion at the recovery stage. This finding suggests that regulatory T cells that expand at the later stage are not involved in the recovery from EAE. Alternatively, regulatory T cells may exist in the expanded Vβ8.2 population. In addition, it is possible that regulatory T cells may rely more on restricted Vα expression.

Determination of nucleotide sequences of Vβ8.2 spectratype-derived TCR confirmed the findings reported by Buenafe et al. (21) that the Asp-Ser CDR motif is frequently present in spinal cord-derived cDNA clones after random sequencing. Moreover, we demonstrated that the majority of Vβ8.2 clones extracted from an expanded spectratype (60–70% depending on the stage of EAE) have the DSSYEQYF sequence in the CDR3, and the rest except one found at the recovery stage (DVWETQYF) have either Asp-Ser or (X)-Ser motif (Table I). Percentages of these sequences are much higher that those obtained by random sequencing because most TCR clones with a longer CDR3 (bands II and III according to the nomenclature in this study) do not have the CDR3 common motif (our unpublished observation). The CDR3 profile of Vβ8.2 at the recovery stage was slightly different from those seen at the early and peak stages. The frequency of the DSSYEQYF sequence decreased slightly (60%), and TCR clone without Asp-Ser or (X)-Ser motif constituted the second largest population. As mentioned in the previous paragraph, T cells that bear this sequence might have a regulatory function.

In the present study, we analyzed a complete set of TCR β-chain family of spinal cord T cells of Lewis rats with active EAE by CDR3 size spectratyping and then determined the nucleotide sequence of the spectratype-derived CDR3 of interest. Since the TCR of encephalitogenic T cells has been extensively investigated and well characterized mainly using long term-cultured T cell clones, we were able to detect one definite and several possible TCR spectratypes in the target organ associated with EAE development according to the criteria of TCR of encephalitogenic T cells. Conversely, it is highly possible to identify autoimmune disease-associated TCR spectratype(s) based on characteristics of spectratypes. We are currently investigating rat experimental autoimmune carditis (EAC) by the strategy used in this study. EAC is inducible by immunization with cardiac myosin (25) and transferable to naive animals by sensitized T cells (26), but the exact nature of carditis-inducing T cells is unknown. We screened Vβ1–20 spectratypes of heart-infiltrating T cells and found several candidate spectratypes that expanded clonally and thus may be associated with the development of EAC (manuscript in preparation). Likewise, CDR3 size spectratyping is a powerful tool for the screening of autoimmune-associated TCR, and the strategy, including the methods shown in this study, can be applied for all T cell-mediated autoimmune diseases.