The Lymphocytic Infiltration in Calcific Aortic ... - Semantic Scholar

The Journal of Immunology

The Lymphocytic Infiltration in Calcific Aortic Stenosis Predominantly Consists of Clonally Expanded T Cells1 Henry D. Wu,* Mathew S. Maurer,* Richard A. Friedman,† Charles C. Marboe,‡ Elena M. Ruiz-Vazquez,§ Rajasekhar Ramakrishnan,¶ Allan Schwartz,* M. David Tilson,储 Allan S. Stewart,储 and Robert Winchester2‡§ Valve lesions in degenerative calcific aortic stenosis (CAS), a disorder affecting 3% of those older than 75 years, are infiltrated by T lymphocytes. We sought to determine whether the ␣␤ TCR repertoire of these valve-infiltrating lymphocytes exhibited features either of a polyclonal nonselective response to inflammation or contained expanded clones suggesting a more specific immune process. TCR ␤-chain CDR3-length distribution analysis using PCR primers specific for 23 V␤ families performed in eight individuals with CAS affecting tri- or bileaflet aortic valves revealed considerable oligoclonal T cell expansion. In five cases, ␤-chain nucleotide sequencing in five selected V␤ families showed that an average of 92% of the valve-infiltrating T cell repertoire consisted of expanded T cell clones, differing markedly in composition from the relatively more polyclonal peripheral CD8 or CD4 T cell subsets found even in this elderly population. Twenty-four of the valveinfiltrating T cell clones also had the same clone identified in blood, some of which were highly expanded. Interestingly, 22 of these 24 shared clones were CD8 in lineage (p ⴝ 1.5 ⴛ 10ⴚ12), suggesting a possible relationship to the expanded CD8ⴙCD28ⴚ T cell clones frequently present in the elderly. Additionally, the sequences of several TCR ␤-chain CDR3 regions were homologous to TCR ␤-chains identified previously in allograft arteriosclerosis. We infer that these findings are inconsistent with a nonselective secondary response of T cells to inflammation and instead suggest that clonally expanded ␣␤ T cells are implicated in mediating a component of the valvular injury responsible for CAS. The Journal of Immunology, 2007, 178: 5329 –5339.

C

alcific aortic stenosis (CAS)3 is a progressive disease characterized by calcified nodules in the body of the valve leaflets that restrict valvular motion and impede blood flow from the left ventricle. CAS is the most common reason for valve replacement surgery (1). Because CAS occurs in 3% of those older than 75 years (2, 3), from the time of its description over 100 years ago, CAS had been attributed to the degenerative effects of chronic hemodynamic stress and “wear and tear” on the valve (4), although why only a subset of individuals would respond to this prevalent stress with CAS is unknown. Olsson et al. (5) and Otto et al. (6) were among the first to demonstrate a variable, but often quite appreciable T lymphocyte infiltration in CAS. Further work showed that most T cells infil-

*Division of Cardiology, †Department of Biomedical Informatics, ‡Department of Pathology, and §Division of Autoimmune and Molecular Diseases, ¶Division of Biostatistics, and 储Department of Surgery, Columbia University College of Physicians and Surgeons, New York, NY 10032 Received for publication July 11, 2006. Accepted for publication January 15, 2007. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This study was supported by U.S. Public Health Service Grants R01 HL084599 and U19 AI046132, a Clinical Associate Physician Award from the National Institutes of Health (M01RR00645), and the Victoria and Esther Aboodi Award (to H.D.W.). M.S.M. was supported by a career development award from National Institute on Aging (K23-AG00966-A1A). 2 Address correspondence and reprint requests to Dr. Robert Winchester, Columbia University College of Physicians and Surgeons, 630 West 168th Street, New York, NY 10032. E-mail address: [email protected] 3 Abbreviations used in this paper: CAS, calcific aortic stenosis; VEGF, vascular endothelial growth factor; PB, peripheral blood; BLAST, basic local alignment search tool; p-MHC, peptide-MHC; HTLV, human T cell leukemia virus.

Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 www.jimmunol.org

trating the valve are concentrated in aggregates surrounding regions of calcification or arrayed about newly appearing blood vessels that express vascular endothelial growth factor (VEGF), VEGF receptors, ICAM-1, and VCAM-1 (7–9). The T cells exhibit considerable evidence of activation, including expression of CD25 and HLA-DR, whereas cells of the surrounding valvular mesenchymal tissues also strongly express HLA-DR, consistent with the release from activated lymphocytes of inflammatory mediators such as IFN-␥ (5, 10). The modified mesenchymal cells also express genes characteristic of osteoblasts, suggesting that calcification results from an active, regulated osteogenic process (9, 11). Because CAS is more prevalent in anatomically variant bileaflet aortic valves and occurs at younger ages, the finding of similar T lymphocyte infiltration in bicuspid valves emphasized the importance of this inflammation in CAS pathogenesis (7). CAS shares some risk factors with atherosclerosis and apolipoprotein E knockout mice develop CAS (2, 12, 13). However, as emphasized by Otto et al. (14), CAS and atherosclerosis are not synonymous, given that only half of the patients with CAS have concomitant coronary artery disease, whereas the large majority of those with atherosclerosis do not develop CAS (12). Additionally, bileaflet aortic valves are disproportionally affected with CAS, further suggesting that distinct factors account for the development of CAS (15). Moreover, the lack of efficacy of intensive lipid-lowering therapy by statins in halting the progression of CAS (16) suggests that processes other than those involved in atherosclerosis may be at play in CAS. Among the possibilities that could account for the conspicuous presence of T cells within CAS valve lesions is that the entrance of T cells is a secondary consequence of chemokines released by valvular injury or non-Ag-specific innate immune system activation of macrophages resulting in a polyclonal T cell infiltration

5330

CLONAL EXPANSIONS OF T CELLS IN AORTIC STENOSIS

Table I. Case information Case Designation

CAS01

CAS02

CAS03

CAS04

CAS05

CAS06

CAS07

CAS08

Age (years) Sex Ejection fraction (%) Aortic valve peak gradient (mm Hg) Aortic valve mean gradient (mm Hg) Aortic valve area (sq cm) Coronary artery disease (no. of vessel) Coronary bypass Hypertension Hyperlipidemia Diabetes Cigarette smoking Statin use Valve architecture

73 M 60 68 43 0.8 3 Yes Yes No No No No Trileaflet

81 F 50 110 67 0.57 0 No Yes Yes Yes No Yes Trileaflet

72 F 60 73 30 0.8 0 No No No No No No Trileaflet

84 M 60 105 60 0.6 2 Yes Yes No Yes No No Trileaflet

60 M 55 76 42 0.66 0 No No No Yes No No Trileaflet

41 M 65 49 31 1.1 0 No No No No No No Bileaflet

65 M N/A N/A N/A N/A 0 No No No No No No Trileaflet

76 F N/A N/A N/A N/A 0 No No No No Yes No Trileaflet

M, Male; F, female.

analogous to that of stable atherosclerotic plaques (17–20). Conversely, if the repertoire consisted of a small number of highly expanded T cell clones, it would favor the interpretation that some features of an adaptive immune response were occurring within the valve perhaps in response to stress-induced molecules, with the potential for the T cell activation and mediator release associated with clonal expansion to enhance valve injury. Because these two primary scenarios can be distinguished by TCR repertoire analysis, the goals of this study were to delineate the clonal composition of the infiltrate through structural features of the clonally specific ␤-chain TCR in the lymphocyte repertoire and determine the presence of expanded shared putative progenitorcirculating clones in either the CD4 or CD8 subsets of peripheral blood (PB).

Materials and Methods Study patients Residual aortic valve tissues and blood samples were obtained at surgery from eight individuals undergoing valve replacement for severe CAS (Table I). These specimens included seven tricuspid valves and one bicuspid aortic valve. No case had a history of rheumatic fever nor did they have significant aortic insufficiency or mitral valve pathology on echocardiography. Two had significant coronary atherosclerosis requiring coronary artery bypass. A third had hyperlipidemia without coronary atherosclerosis. A histologically normal-appearing valve was obtained during heart transplant for end-stage heart failure secondary to dilated cardiomyopathy. Histologic sectioning and immunochemical staining were performed as described previously (21). The Institutional Review Board of Columbia University approved the study. All participants gave informed consent.

Analysis of ␣␤ TCR repertoire ␤-chain CDR3 length distribution The clonal composition of the T cells infiltrating the stenotic aortic valve was analyzed using RNA extracted from separated PB CD4 and CD8 subpopulations and from the total sample of the aortic valve tissue as described previously (21–23). Briefly, the valve is minced, transferred to RNAStat-60 reagent, homogenized in a Tissuemizer (IKA Labortechnik), and after centrifugation out of calcific particles, RNA was isolated according to the manufacturer’s instructions (Tel-Test). The quality of the RNA was determined on an Agilent 2100 Bioanalyzer according to the manufacturer’s instructions, using a RNA microchip. cDNA was prepared using oligo(dT) priming. The repertoire composition was first determined in a lower resolution global assessment by ␤-chain CDR3 length distribution or spectratype analysis to characterize the degree of clonal expansions for each of the 23 TCR ␤-chain variable region families (BV or TRBV) in each CAS case, as described previously (21, 22). Briefly, primers specific for a particular variable region family (BV) and the constant region (CB) were used in PCR to amplify the TCR ␤-chains in each BV family. The sequence for all 23 primers for the 5⬘ TRBV families, the 3⬘ TRBC primer, and the

runoff primer are shown in the Supplement Table4 along with the size of the product for CDR3 length (length ⫽ 11). The amplified PCR product was fluorochrome-labeled in a runoff reaction using a labeled primer and the bands reflecting different CDR3 lengths, resolved on an automated analyzer, as described previously (21, 22). The presence of clonal expansions in the T cell compartments was determined by comparing the observed ␤-chain length distributions to a reference composite profile generated from a healthy control group of 140 samples, as described previously (21). A clonal expansion was considered to be present at a particular CDR3 length when the observed distribution in that BV family exceeded the 95th percentile of the expected frequency by 5% at that ␤-chain length in reference normals (21).

␤-Chain nucleotide sequence analysis In five trileaflet valves, the clonal composition of the repertoire was determined by nucleotide sequence in five BV families. These families were chosen for sequencing based on the presence of clonal expansions in the valve and one or more putatively shared oligoclonal expansions at the same CDR3 length in either or both the CD8 or CD4 subsets of PB. The cloning and sequencing of the same type of PCR product used in the length distribution technique involved a high throughput procedure as described previously (21). From 20 to 105 transcripts were successfully sequenced and translated from each of the three T cell compartments for each BV family after elimination of nonproductive rearrangements. The alternative IMGT (24) and Arden et al. (25) nomenclature, respectively, for variable (V region, TRBV, BV), diversity (D region, TRBD, BD), joining (J region, TRBJ, BJ), and CDR3 length delimiters (CAS. . . FGXG, C. . . FGXG) is included. The IMGT/VQUEST website was used to identify these regions and the location of the ␤-turn in the CDR3 region: http://imgt.cines.fr: 8104/IMGT vquest/share/textes/index.html (26). A clone was considered as expanded if two or more transcripts were identified, although inferring the degree of clonal expansion from the numbers of TCR transcripts identified in repertoire analysis is necessarily an approximation influenced by factors such as stage of clonal activation. Because the sequence analysis method determines the composition of the repertoire at the level of nucleotide sequence, rather than just ␤-chain length, it reveals the presence of clones that would otherwise not be apparent in the lower resolution spectratype method that only detects much more expanded clones in settings of lower polyclonal background. The sequences of all valve-infiltrating ␤-chain are available in GenBank, accession nos. EF415300 –EF415485, see supplemental table.

Sequence analysis bioinformatic methods The CDR3 amino acid motifs in the valve were examined for the presence of homology as described previously (21). Basically, the motifs were first examined for homology with sequences in the nr database using the Webbased National Center for Biotechnology Information protein-protein basic local alignment search tool (BLAST) resource (http://www.ncbi.nlm.nih. gov/BLAST) (27) and for similarities in TCR structure (26). The motifs were then examined at higher stringency for homology with sequences in the nr database using the GCG implementation of BLAST (27, 28), with the 4

The on-line version of this article contains supplemental material.


5331

PAM30 matrix, word size (size ⫽ 2) and gap cost existence (gap cost ⫽ 9, extension ⫽ 1), as is appropriate for very short sequences (29). Additional sequence comparisons were made using the GCG program “FindPatterns” using an identity criterion. To explore whether these homologies occurred by chance, the CDR3 regions of 163 CAS clones and a disease control of 116 T cell clones obtained from the kidneys of individuals with lupus nephritis were each compared using the Findpatterns (28) to the corresponding region of both 223 productively rearranged graft-arteriosclerosis clones (of 243 total) (Genpept, AAG15594-AAG15836; GenBank, AY006091-AY006333 (30)) and 99 atherosclerosis clones (18). Sequences were accepted as matched either by requiring 10 identical residues not counting the first 3 (e.g., CAS), or by a more lenient criterion in which 2 mismatches were accepted, as long as they involved residues in the same physicochemical similarity group according to the criterion of Miyata et al. (31). The number of matches expected by chance in the Findpatterns search is the probability of a single match, Pmatch, that equals the product of the frequencies of the amino acids (32) in the CDR3 region corrected for the number of CAS and graft clones, NC and Ng: Ematch ⫽ NC Ng Pmatch. Examination of the three-dimensional structure of CDR3␤ sequences that contain the YEQYFG CDR3 motif (33) was made using the molecular graphics program DeepView (Swiss-PDB Viewer) (34) to locate the potential contact residues.

Statistics The difference in degree of clonal expansion between distributions was assessed using the Hamming distance statistic that provides an overall measure of the extent that a given length probability distribution differs from a reference distribution, as described previously (22). The KolmogorovSmirnov test was performed to compare the difference between CDR3 length distribution expressed by T cells within the aortic valve, as determined by nucleotide sequencing, to length distributions observed within CD8 and CD4 peripheral subsets in the same CAS cases as described previously (21). The likelihood of the finding that T cell clones within the aortic valve originated from peripheral CD8 but not the CD4 subset was calculated using binomial probability.

Results TCR ␤-chain CDR3 length distribution analysis revealed multiple oligoclonal T cell expansions among lymphocytes infiltrating stenotic aortic valves Representative illustrations of the H&E and immunochemical staining using CD3, CD4, and CD8 mAbs of three CAS cases showing the composition of the lymphocytic infiltration in trileaflet and bileaflet aortic valves are shown in Fig. 1. These demonstrate the expected infiltration by CD4 and CD8 T cells around areas of neovascularization, calcific deposition, and in the aortic subendothelium of the valve. The distribution of TCR ␤-chain CDR3 lengths in seven tricuspid stenotic aortic valves was found to be highly restricted, with skewed length spectra indicating a considerable number of oligoclonal expansions among the infiltrating T cells, as illustrated by CAS01 (Fig. 2A), CAS05 (Fig. 2B), and CAS04 (Fig. 3B). There were oligoclonal expansions in most of the 23 BV families comprising this valve TCR repertoire, from 1 in BV15 up to 4 in BV5. This information is summarized for all 23 BV families studied in each sample, where the top section of Table II illustrates that the average number of oligoclonal expansions within the aortic valve for each sample is greater than in the PB CD8 or CD4 subsets, and middle section shows that these oligoclonal expansions also take up a greater proportion of the TCR repertoire within the aortic valve than they do in the paired PB CD8 or CD4 subsets, reflecting their greater degree of expansion. For example, in CAS01, 27 oligoclonal expansions were found in the heart valve compared with 22 in the CD8 subset, but the greater enlargement of the clones in the valve relative to the balance of the valve TCR repertoire was shown by the fact that 74% of the total valve repertoire in CAS01 (Table II, middle) consisted of oligoclonally expanded T cells, whereas in the CD8 subset the expanded clones only occupied 19% of the repertoire. Among all valves, an average of 57% of the

FIGURE 1. H&E and immunochemical staining using CD3, CD4, and CD8 mAbs of three representative CAS cases illustrating the composition of the lymphocytic infiltration in the aortic valve. Image set A, from CAS05, shows an aggregate of infiltrating lymphocytes in a trileaflet valve containing CD4 and CD8-staining T cells, with a preponderance of CD4 T cells. Image set B, from CAS06, shows a dense perivascular infiltrate of CD4 and CD8 T cells in a region of neovascularization in a bicuspid valve. The proportion of CD4 to CD8 T cells is nearly equal. Image set C shows a lymphoid aggregation in trileaflet CAS07 near a region of calcification. The number of CD4 T cells was slightly greater than the number of CD8 T cells.

␤-chain TCR valve repertoire across all 23 BV families (Table II, middle) was occupied by 37 T cell oligoclonal expansions (Table II, top). CAS05 (Fig. 2B) exhibited the most heterogeneous and polyclonal infiltration of T cells by the ␤-chain length distribution method, with 28% of the repertoire containing 31 clonal expansions (Table II, top and middle). Indicative of the selective entry of T cells into the tissue, no T cells were detected in some BV families, such as BV6S1 (BV6) in CAS01 (Fig. 2A) and BV25 in CAS05 (Fig. 2B). An average of 20% (range, 12– 44%) of BV families lacked expression only in the valve (data not illustrated). As expected in these older individuals, the CD8 subset exhibited frequent oligoclonal expansions designated by arrowheads. One case with a calcific bicuspid valve similarly revealed oligoclonal patterns (Fig. 3A) indistinguishable from those observed in the tricuspid valves (Figs. 2, A and B, and 3B). A normal valve had no detectable TCR ␤-chains in any BV family (data not illustrated).

5332


FIGURE 2. TCR ␤-chain CDR3 length distribution of two CAS cases, CAS01 and CAS05, comparing the T cell repertoire of the valve with that of blood CD8 and CD4 subsets. A, Illustrates the oligoclonal nature of T cells in trileaflet CAS valve in case CAS01 for six representative BV families. The height of each fluorescent peak is the relative proportion of the TCR repertoire occupied by TCR with a particular CDR3 length. For example, within BV15, a single expansion consisting of a uniform population of T cells with CDR3 length of 9 aas (Arden et al. (25)) was detected. This pattern differed markedly from the ␤-chain length distributions exhibited by paired peripheral CD8 and CD4 T cells, which were composed of multiple peaks distributed across a spectrum of CDR3 lengths. The CD4 subset length distribution was more Gaussian, whereas the CD8 subset exhibits several oligoclonal expansions (arrowheads). Certain BV families in PB contained oligoclonal expansions (arrows) of the same CDR3 length as those found expanded in the valve (arrows connected by lines), as illustrated by the oligoclonal expansion at CDR3 length (length ⫽ 9) for the valve and CD8. Similarly, the ␤-chain TCR repertoires in the other BV families also exhibited restricted and skewed distributions in contrast to patterns in the corresponding PB subsets. BV6A (6S1) illustrates an instance where no TCR expression was detected. B, TCR ␤-chain CDR3 length distribution illustrating that valve CAS05 contains a slightly more polyclonal infiltration in addition to oligoclonal T cells expansions. Six representative BV families are illustrated. The repertoire in BV1, -5, and -8 exhibited an increased proportion of apparently polyclonal T cells compared with the repertoires in the other valves. The ␤-chain TCR repertoires in the remaining BV families exhibited a more restricted and skewed distribution in contrast to patterns in the corresponding PB subsets. BV 25 illustrates an instance where no TCR expression was detected in the valve. The faint gray peaks are different size calibration markers.

The difference between the ␤-chain TCR length distributions of the stenotic valves and CD4 or CD8 subsets, or a weighted combination of both CD4 and CD8 approximating the ratio in normal

blood (PB), is further shown by comparing the proportion of the repertoire present at each CDR3 length between each paired sample, as measured by the Hamming distance statistic. Table


5333 Table II. Oligoclonal expansions occupy the major proportion of the T cell repertoire infiltrating the aortic valve as revealed by ␤-chain CDR3 length distribution in 23 BV families The Number of Oligoclonal Expansions within the Aortic Valve is Greater Than in the Peripheral Blood CD8 or CD4 Subsets Number of oligoclonal expansions Case

Valve

CD8

CD4

CAS01 CAS02 CAS03 CAS04 CAS05 Average

27 39 49 39 31 37

22 29 30 29 30 28

5 18 1 20 18 12

Oligoclonal Expansions Occupied a Greater Percentage of the TCR Repertoire within the Aortic Valve Than in the Peripheral Blood CD8 or CD4 Subsets Percentage of repertoire occupied by oligoclonal expansions Case

Valve

CD8

CD4


74 70 63 52 28 57

19 50 8 54 27 33

10 15 10 23 10 14

CDR3 Length Distributions within the Aortic Valve Differ from those in the Peripheral CD8 and CD4 Compartments Based on Hamming Distance Median hamming distance for all BV families (range) Case

FIGURE 3. A, TCR ␤-chain CDR3 length distribution analysis of five representative BV families illustrating the oligoclonal nature of T cells that infiltrate the aortic valve of a individual with CAS with a bileaflet valve. The oligoclonal T cell repertoire in the bicuspid valve is similar to that found upon ␤-chain length distribution analysis of trileaflet valves as shown in Figs. 2 and 3B. B, An additional example of ␤-chain length distributions of five BV families expressed by T cells infiltrating a trileaflet valve, CAS04, revealing a marked preponderance of oligoclonal expansions.

II, bottom, shows for each case the elevated Hamming distances observed across all 23 BV family length distribution histograms between those of the T cell repertoire infiltrating the valves and to those from PB. The average distance for all samples and BV families is 47 ( p ⬍ 0.01). Although the Hamming distance difference indicates the repertoires greatly differ between the valve and blood CD4 or CD8 subsets, a smaller number of oligoclonal expansions are found at the same CDR3 length and are designated by arrows joined by tie bars (Fig. 2, A and B). BV families containing these potentially shared clones with either CD4 or CD8 subsets from each of the five CAS cases were chosen for analysis by nucleotide sequencing. TCR ␤-chain nucleotide sequence analysis reveals the T cell repertoire within the aortic valve predominantly consist of expanded clones The 3,612 ␤-chain sequences remaining after exclusion of unsuccessful rearrangements were analyzed for BV, D, and J element usage, translated into amino acid sequences, and identical se-

CAS01 CAS02 CAS03 CAS04 CAS05

Valve vs CD8

Valve vs CD4

Valve vs PBⴱ

60 (23–100) 50 (13–100) 46 (15–90) 49 (17–88) 29 (10–87)

70 (31–100) 50 (16–100) 35 (12–84) 43 (24–83) 25 (11–53)

52 (25–91) 54 (20–84) 43 (25–84) 44 (21–83) 27 (9–70)

ⴱ, PB equals weighted average for CD4 and CD8 distributions, 70:30, approximating composition of whole blood.

quences were grouped together forming expanded clones (Supplement Table) and summarized as distribution histograms (Fig. 4). These histograms resemble the length distribution histograms in Figs. 1, A and B, and 2, but reveal the clonal composition at each CDR3 length, showing that the ␣␤ T cell valve repertoire was mainly composed of a small number of variably expanded T cell clones. For example, in Fig. 4A, the CAS01 valve repertoire consisted entirely (100%; Table III) of 21 expanded clones in the 5 sequenced BV families, with BV15 consisting of a single clone composed of 24 sequences. The other four cases (Fig. 4, B–E) similarly illustrate that the repertoire of valve-infiltrating T lymphocytes is composed of a limited number of T cell clones, many of which are expanded. An average of 92% of the valve TCR repertoire consisted of expanded T cell clones (Table III, bottom), with at least 97% of the repertoire in CAS01, -2, -3, and -4 composed of expanded clones. CAS05 was the most clonally diverse, with 27 expanded clones that occupied 66% of the repertoire, a considerably higher number than revealed by the length distribution method in these same BV families (Fig. 2B). The remaining 34% of the repertoire in CAS05 contained 40 unexpanded single sequence T cell clones. The presence of unexpanded clones (e.g., Fig. 4B) CAS02 (BV2 CDR3 lengths ⫽ 10, 11, and 12) and BV9 (length ⫽ 10), designated by boxes of unit height, provide an internal control for interpreting the relative degree of expansion in the

5334


FIGURE 4. Summary of T cell repertoire composition of the ␤-chain nucleotide sequence comparing the aortic valve and the peripheral CD8 and CD4 T cell compartments in each of the five CAS cases CAS01 through CAS05 (A–E). TCR transcripts with identical ␤-chain nucleotide sequence were grouped together and plotted according to their inferred amino acid sequence CDR3 length in histogram form, with the height of each column subdivision representing the number of identical sequences. The sequence analysis sometimes revealed separate clonal expansions not suspected by the length distribution method in Figs. 1 and 2. For example, A shows that the peak at CDR3 length (length ⫽ 9) in the CD8 blood subset is found to be composed of eight clones, whereas the valve contains only a single expanded clone that is identical in sequence to one of the CD8 blood clones, indicated by the filled rectangle and the tie line containing the CDR motif of the shared clone (CATSDFIADTQYF). The size of the shared expanded clone in blood varied considerably, with the largest expansions in blood of shared clones found in CAS04 and CAS05.

other clones. As shown in Fig. 4E, unexpanded clones predominate in BV1 and BV5. Fig. 4A reveals that the BV4 oligoclonal expansion in CAS01 at CDR3 length (length ⫽ 9) in Fig. 2A is actually composed of three clones, suggesting that these expanded T cell clones are not randomly distributed and may reflect selection by the same peptide Ag (23, 35). The CDR3 amino acid sequence motifs of some expanded clones with the same CDR3 length are homologous. For example, in CAS03 the BV8 family contains three clones at CDR3 length 9, of which two, CASSLGQKETQYF and CASSAGGTDTQYF, exhibit considerable homology (homologous residues underlined) and in CAS04, the

BV15 family contains the clonal expansions of CDR3 length 9, CA TQPVGTDTQYF, which differ by 3 aa residues from a second clone in the valve (CATSDLGTDTQYF) (Supplement Table). Furthermore, 27% of T cell expansions with the same CDR3 length use the same BJ segment. The T cell-infiltrative process in CAS is highly selective and generally does not resemble the repertoire in blood The TCR repertoires within the valve differed from those of the blood T cell subsets in terms of the greater degree of clonal expansion and the smaller proportion of the repertoire occupied


FIGURE 4. (continued)

5335

5336


Table III. ␤-Chain CDR3 nucleotide sequencing reveals that the TCR repertoire within the calcific aortic valve is composed of a small number of highly expanded T cell clones

Of the total of 158 T cell clones identified in the valve, 24 shared an identical ␤-chain CDR3 sequence at the nucleotide level with a clone in PB (Supplement Table) and depicted as amino acid motifs in the tie line linking these clones in Fig. 4. Strikingly, in 22 of 24 of the shared clones, the presumptive clonal precursor originated from the CD8 subset. Because this subset occupies ⬃30% of the peripheral T cell pool, the probability of this observed CD8 predominance being a chance event is p ⫽ 1.5 ⫻ 10⫺12.

siderable homology between the CDR3 of a valve clone from BV8 in CAS05 and a BV4 clone from CD8 subset in CAS04. The four underlined amino acid residues is the calculated location of the ␤-turn between the F and G strands of the ␤-pleated sheet and is near the likely region of p-MHC contact. There is a nonconservative amino acid difference not in the same physicochemical similarity group in this region (31). Set3 and Set4 show sequence homology between two different CD4 clones from blood and valve tissue clones. Set3 differs only in the J region 3⬘ to the ␤-turn, whereas Set4 share all but 2 aas. To explore the hypothesis that some sequences will be public motifs homologous to amino acid sequences found in previously identified sites of inflammation or other situations (35), the valve sequences were aligned to those in the nucleotide database using BLAST, and a program that estimates the position of the ␤-loop of the CDR3 region that is a contact site with p-MHC. The most significant homologies are summarized in Table IV. The CD4 T cell TCR sequence of Set4 from blood was found homologous to two clones identified in inflammatory sites of diseases unrelated to CAS, human T cell leukemia virus (HTLV) polymyositis, and Rasmussen encephalitis, suggesting that the valve clone of Set4 may be a bystander clone. Set5 shows the homology of a valve clone with an anti- HTLV-1 clone. The ␤-chains differ in length by 1 aa, but the putative turn is similar, containing one conservative amino acid difference (31). Set6 is a second example of a homology with a different HTLV-1-reactive clone that differs by two nonconservative changes. Set7 is a homology with an alloreactive HLA-B27specific clone that differs in sequence by one conservative amino acid in the ␤-turn. Set8 contains a valve clone homologous to both a clone from a multiple sclerosis lesion and one from graft arteriosclerosis. Sets9 through 13 each exhibit the highest or second highest homology with a CDR3 region of a TCR ␤-chain isolated from graft arteriosclerosis. Set9 exhibits the greatest homology, differing only by the conservative F-L substitution, whereas Set11 and Set12 have conservative changes in the ␤-turn between residues belonging to the same physicochemical similarity class (31). Examination of the three-dimensional structure of sequences containing the YEQYF motif shows that Y is the main contact between the CDR3␤ region and the peptide, and that in the sequences of Set10 the gap between the CAS and graft sequences due to differing CDR3 length is several amino acids removed from this contact region. In Set11 the three mismatches are between residues belonging to the same physicochemical similarity class (31). To explore whether these homologies occurred by chance, the CDR3 regions of 158 CAS clones and a disease control of 116 T cell clones obtained from the kidneys of individuals with lupus nephritis were each compared with the corresponding region of both 223 productively rearranged graft-arteriosclerosis clones in GenBank (30) and 99 atherosclerosis clones (18). Only the 10 aa match of Set10 was found, with no matches between CAS and atherosclerosis, or between the renal T cells and either atherosclerosis or graft arteriosclerosis. The number of matches expected by chance was 2 ⫻ 10⫺7, implying the observed match is unlikely to have occurred by chance.

Amino acid homology between structurally unrelated clones suggesting possible recognition of common peptide Ags

Discussion

Table IV, Set1, illustrates complete CDR3 amino acid sequence identity of two clones of CDR3 length (length ⫽ 7) from CAS05 that differ in nucleotide sequence and use different BV elements, suggesting that they may have been selected by a common peptide-MHC (p-MHC) structure. Set2 illustrates con-

This initial molecular characterization of the ␣␤ T lymphocyte repertoire in valve leaflets in CAS revealed the infiltrate is predominantly composed of a limited number of expanded T cell clones, and that a subset of these clones were also identified in blood, where they were predominantly highly expanded and CD8

The Number of Expanded T Cell Clones within the Aortic Valve in five Sequenced BV Families Is Smaller Than the Number in the Peripheral Blood CD8 or CD4 Subsets Number of expanded clones Case

Valve

CD8

CD4


21 25 16 20 27 22

38 26 23 38 24 30

54 35 35 25 24 35

The Proportion of TCR Repertoire Occupied by Expanded T Cell Clones Across Five Sequenced BV Families within the Aortic Valve Is Greater Than in Peripheral CD8 or CD4 Subsets Percentage of repertoire accounted for by expanded clones Case

Valve

CD8

CD4


100 97 97 98 66 92

50 90 67 89 65 72

32 63 44 63 32 47

by unexpanded T cell clones (Fig. 4 and Table III). In 16 of the 25 BV families sequenced across the five samples, the significance of the difference between the CDR3 length of PB and aortic valve repertoires was p ⬍ 0.0005. Unlike the valve repertoire, the T cell subsets in PB contained a preponderance of unexpanded T cell clones, and generally the CD8 T cell repertoires were composed of multiple expanded as well as unexpanded clones, whereas CD4 T cell repertoires were composed of predominantly unexpanded T cell clones. Sequencing revealed additional minimally expanded clones in blood not detected by the spectratype method (Fig. 1, A and B). Furthermore, sequence comparison of each clone in the valve vs blood revealed 86 of the 110 expanded aortic valve T cell clones were not found in blood, whereas 416 expanded PB clones were not detected in the valve. Expanded T cell clones within the aortic valve with putative progenitors in blood predominantly originate from the peripheral CD8 subset

3

4

8

1

CAS04 Valve 8_3_L_BV15_tissue_Cl.30 Set10 Graft arteriosclerosis AY006316

CAS04 Valve 10_4_L_BV15_tissue_Cl.26 Set11 Graft Arteriosclerosis AY006100

CAS01 Valve 12_8_R_BV5_tissue_Cl.55 Set12 Graft arteriosclerosis AY006281

CAS05 Valve 11_1_D_BV5_tissue_32 Set13 Graft arteriosclerosis AY006145

.

C A S S L G L A G V D E Q F F

C A S S L G T S A A G E L F F C A S S L - T S A A G E L F F

10

C A S S L G P G T G V K L F F

11 11

C A S S L G P G Q G V Y E Q Y F

C A S S R A S G S Y E Q Y F

10 12

C A T S R A G G P Y E Q Y F

C S A R T G T A S Y E Q Y F

10

10

C A - - T G T A S Y E Q Y F

C A S S F G L A G V D E Q F F

11

C A S S S - G G T D T Q Y F C A S S L T G G T Y E Q Y F

11

C A S S A - G G T D T Q Y F

C A S S P G T S Y E Q Y F

C A S S P G T P Y E Q Y F

C A S S Y P L A G V N E Q F F

C A S S F S L A G V N E Q F F

C A S S L - G R G Y N E Q F F

C A S S L Y G R S Y N E Q F F

cc ccc cc

ctcggt

cca

ttta

tgtgccagcagc.... tgtgccagcagttt...

5-4ⴱ01 28ⴱ01

ctc

gtccc

5-1ⴱ01

tgcgccagcagcttgg

gacc

cgggcc

11-2ⴱ01 tgtgccagcagc..... 5-5/6ⴱ01 tgtgccagcagcttgg

ag

24-1ⴱ01 tgtgccaccagt......

2ⴱ01

2ⴱ01

1ⴱ01

1ⴱ01

2ⴱ01

2ⴱ01

1ⴱ01

2ⴱ01

2ⴱ02

20-1ⴱ01 tgcagtgctaga..

aca

gc

1ⴱ01

tgcgccagcagcttgg

5-1ⴱ01

2ⴱ01 1ⴱ01

2ⴱ01

2ⴱ01

1ⴱ01

2ⴱ02

2ⴱ02

1ⴱ01

None

2ⴱ01 2ⴱ01 2ⴱ02

2ⴱ01

2ⴱ01

1ⴱ01

None

None

2ⴱ01

1ⴱ01

TRBD Name (IMGT)

24-1ⴱ01 tgtgccacc.........

tgtgccagcagctt..

5-6ⴱ01

tggc

c

28ⴱ01 7-7ⴱ01 tgtgccagcagt..... tgtgccagcag.....

12-3ⴱ01 tgtgccagcag......

tgtgccagcagc.....

cccc

taccccctcgct

7-9ⴱ01

tgtgccagcagt.....

6-1ⴱ01 12-3ⴱ01 tgtgccagcag......

tgtgccagcagctt..

5-4ⴱ01

ctc

tacggccgt

9ⴱ01? tgtgccagcagc....

5-5/6ⴱ01 tgtgccagcagcttg.

tgtgccagcagt..... tgtgccagcagt..... tgtgccagcagt.....

C A S S P G L A G G P T D T Q Y F 27ⴱ01 C A S S P G L A G G H G D T Q Y F 28ⴱ01 C A S S P G L A G G P N E - Q Y F 6-5ⴱ01

25-1ⴱ01 tgtgccagcagccaaga

gatggcgattggt

tct

tgtgccagcagccaa

29-1ⴱ01 tgcagcgttgaa 3-2ⴱ03

gag

12-3ⴱ01 tgtgccagc........

gcca

ct

N1b (N, P)

tttcgcc

9 10

8

tgtgccagcagcg...

V-Region

12-3ⴱ01 tgtgccagcagt.....

9ⴱ01

TRBV Namea (IMGT)

C S V S P G L A G G P T D T Q Y F 29-1ⴱ03 tgcagcg.....

C A S S Q D G D W S P L H F

C A S S Q D G D W S G A F F

C S V E S G G T I Y F

C A S E S G N T I Y F

C A S S A T G E Q Y F

C A S S A T G E Q Y F

CDR3 Amino Acid Sequence

9

9

9

11

11

10

11

13 13 12

13

10

10

7

7

7

7

CDR3 Length (Arden)a

..gactagcg......

gggactagcg......

gggacagggg..

gggacaggg...

......agcggg....

......agcgggggg.

gggacag.....

gggacag......

....ctagcggg....

....ctagcgggag..

......agcgggggg. ...acaggggg.

.......gcgggggg.

gggactag........

gggaca......

..........ggag..

...actagcgggag..

.....agggg..

gggactagcggggggg .ggactagcgggggg. gggactagcgggaggg

gggactagcggggggg

........cgggg...

....cagg....

......ggggg.

........cgggg..

...acaggg...

D-Region

cgg

cgg

tt

cgt

ag

cc

cat

ccagcta

agtgg

tagac

aacg

a

c

tg

t

a

cc tcacg ccta

cc

attgg

N2b (N, P)

J-Region

2-2ⴱ01 ......ccggggagctgtttttt

2-2ⴱ01 ......ccggggagctgtttttt

1-4ⴱ01 ...........aaactgtttttt

2-7ⴱ01 ......ctacgagcagtacttc

2-7ⴱ01 ...ctacgagcagtacttc

2-7ⴱ01 cctacgagcagtacttc

2-7ⴱ02 .cctacgagcagtacttc

2-7ⴱ02 ......cgagcagtacttc

2-1ⴱ01 ........atgagcagttcttc

2-1ⴱ01 ..........gagcagttcttc

2-3ⴱ01 ...acagatacgcagtatttt 2-7ⴱ01 ....tacgagcagtacttc

2-3ⴱ01 ...acagatacgcagtatttt

2-7ⴱ01 ...ctacgagcagtacttc

2-7ⴱ01 ..cctacgagcagtacttc

2-1ⴱ01 .......aatgagcagttcttc

2-1ⴱ01 ......caatgagcagttcttc

2-1ⴱ01 ....tacaatgagcagttcttc

2-1ⴱ01 .tcctacaatgagcagttcttc

2-3ⴱ01 ..cacagatacgcagtatttt 2-3ⴱ01 .gcacagatacgcagtatttt 2-7ⴱ01 .....acgagcagtacttc

2-3ⴱ01 ..cacagatacgcagtatttt

1-6ⴱ01 .........tcacccctccacttt

1-1ⴱ01 ..........agctttcttt

1-3ⴱ01 ........caccatatatttt

1-3ⴱ01 .tctggaaacaccatatatttt

2-7ⴱ01 .......gagcagtacttc

2-7ⴱ01 .......gagcagtacttc

TRBJ Name (IMGT)

b

The IMGT and Arden et al. (25) nomenclature for variable (V region, TRBV, BV), diversity (D region, TRBD, BD), joining (J region, TRBJ, BJ), and CDR3 length (CAS...FGXG, C...FGXG). N segment: nucleotides added in a template-independent or palindromic manner at the joining junctions. Presumed palindromic nucleotides are set apart by a space. The periods designate portions of the V, D, and J gene element genomic sequences excised during recombination. The four underlined amino acid residues is the calculated location of the ␤-turn between the F and G strands of the ␤-pleated sheet provided by the IMGT website, and is the likely region of p-MHC contact.

a

2

CAS05 Valve 11_2_D_BV5_tissue_20 Set9 Graft arteriosclerosis AY006292

2

1

4

10

CD8 L34734

Valve 9_2_D_BV8_tissue_Cl.12

CD8 AB044125

Valve 11_1_D_BV5_tissue_01

AB044104

11_4_R_BV5_tissue_Cl.29

1

13_1_R_BV14_CD4_Cl.42 AB076794 U55163

CAS03 Valve 9_10_E_BV8_tissue_21 Set8 Multiple sclerosis AB011247 Graft arteriosclerosis AY006318

CAS05 Set7 Alloreactive HLA-B27

CAS05 Set6 HTLV-reactive clone

CAS01 Set5 Anti-HTLV-1 tax clone

5

13_5_R_BV4_tissue_Cl.18

2

CD4 10_2_E_BV16_CD4_Cl.23

3

1

Valve 10_1_A_BV9_tissueCl.24

CD8 7_3_R_BV4_CD8_Cl.70

1

2

Valve 7_2_D_BV8_tissue_Cl.8

CD8 7_1_D_BV8_CD8_22

Clone Size

3

Representative Sequence Clone Name (Arden (25) BV)a

Valve 7_3_D_BV1_tissue_26

Site

CAS01 Valve Set4 CAS01 CD4 HTLV polymyositis Rasmussen encephalitis

CAS02 Set3 CAS03

CAS05 Set2 CAS01

CAS05 Set1 CAS05

Case

Table IV. Amino acid homology between different pairs of structurally distinct clones

The Journal of Immunology 5337

5338 in lineage. Taken together with prior observations on the phenotype and activation state of the infiltrating T cells, the expression of HLA-DR and evidence of osteoblast differentiation by the valvular mesenchymal cells, and the presence of new blood vessels in regions of lymphoid infiltration expressing VEGF, VEGF receptors, ICAM-1, and VCAM-1 (5–11), the present observations are not consistent with the hypothesis that the T cell infiltration in CAS is a nonspecific, innate immune response to inflammation. They rather suggest that clonally expanded ␣␤ T cells mediate at least some of the valvular injury, although the nature and magnitude of this contribution to the development of CAS remains unknown. The selectivity of the process underlying the presence of the expanded T cells infiltrating the valve is supported by the absence in tissue of T cells from certain BV families, the highly significant predominance of the CD8 subset among the clones shared with blood, the large proportion of clonal expansion in both blood CD4 and CD8 repertoires that were not identified in the valve and vice versa, the elevated proportion of expanded clones to unexpanded clones found in the valve, the structural homologies evident between CDR3 regions of unrelated clones as well as the frequent representation of the same CDR3 length in the valve. This selectivity suggests the operation of cognitive immune recognition events in the formation of the inflammatory infiltration. Several observations make unlikely the possibility that the clonal expansions artifactually result from PCR amplification of extremely small numbers of T cells in the valve. Real-time PCR for GAPDH and CB could be performed on valve samples CAS01 and CAS03, the two valve samples with the most oligoclonal T cell repertoires, and revealed, relative to the amount of GAPDH, that CAS01 contained 6.93% ␣␤TCRCB and CAS03 1.66% ␣␤TCRCB transcripts. Adjusted for the values obtained for purified CD4 T cells, this result provides an estimate that 1.7% of the cells in valve sample CAS01 and 1.6% of sample CAS03 are composed of infiltrating T cells, respectively. This estimate is consistent with the numbers of T cells demonstrable histologically within CAS valves in this study (Fig. 1) and in the literature (5–7, 9). Furthermore, using the same techniques, comparably small numbers of T cells obtained with needle biopsy of inflamed synovial tissues and analyses of fewer than 1,000 PB T cells yield a predominantly polyclonal repertoire (21, 23). The findings bear on the relationship between atherosclerosis and CAS. The composition of the TCR repertoire in CAS differs from that in the stable vascular atherosclerotic lesion, where the lymphocytes are mainly polyclonal and perhaps chiefly recruited by lipid-stimulated macrophages (17–20). Moreover, no shared T cell clones have been found between the stable atheromatous plaques and PB (18). The predominance of shared CD8 T cell clones also distinguishes CAS from the clonal expansions of T cells that have been implicated in adverse events occurring in unstable coronary and carotid plaques that were predominantly derived from the unusual subset of CD4⫹CD28null effector T cells (36 – 41). There is increasing evidence that some expanded T cells in an immune-driven inflammatory infiltrate do not recognize the peptide inciting the inflammation, a category designated as “bystander clones” (42). Their presence is considered to reflect the attraction, activation, and proliferation of memory/effector T cells induced by chemokines and immunoreactants found in the inflammatory milieu (42). The considerable sequence homology by BLAST exhibited by several TCR clones in CAS to clones from multiple inflammatory sites, such as multiple sclerosis lesions and HTLV-1 infection (Table IV), suggests that some CAS clones are likely bystander clones related to the non-Ag-specific component of in-

CLONAL EXPANSIONS OF T CELLS IN AORTIC STENOSIS flammation. However, the minor proportion of polyclonal, nonexpanded T cells compared with other inflammatory sites, such as atherosclerotic plaques of inflamed synovia (17–19, 21), suggests that inflammation-mediated T cell recruitment is not a highly prominent feature of CAS. Intriguingly, BLAST analysis showed that the disease entity with the most high-scoring T cell CDR3 motifs ␤-chain homologous to CAS clones in GenBank was cardiac allograft arteriosclerosis, a form of arteriosclerosis involving a specific cell-mediated immune response developing during cardiac allograft rejection and mediated by expanded CD8 lineage T cells clones (30). This sharing of clone sequences between unrelated individuals is an example of a public repertoire, and the frequent selection of conserved amino acid motifs is an example of a type 2 TCR repertoire bias (35). We speculate that this homology suggests that some clones in CAS and allograft arteriosclerosis recognize the same peptide. One explanation of how coronary arteries in the rejecting heart might express the same peptide as the aortic valve is that the allogeneic graft rejection response triggers a secondary activation of auto-, not alloreactive T cells that recognize vascular stress self-peptides, a possibility envisioned by Slachta et al. (30), and that these same molecules are induced in heart valves as a result of prolonged mechanical strain and elicit the T cell infiltration and clonal expansions in CAS. Perhaps binding and presentation of these peptides to T cells occurs with only certain allelic class I MHC molecules, accounting for the development of CAS in a small subset of all individuals that would be expected to express the stressinduced molecules. The finding that the large majority of T cell clones shared between blood and the valve were CD8 in lineage, together with occurrence of CAS in the elderly, suggests the hypothesis that the shared clones are related to the clonally expanded CD28-negative presenescent CD8 T cell clones in blood that normally increase in frequency with age and in some situations of chronic viral infection, autoimmunity, and malignancy (43– 49). The expression on these CD8 clones of NK family receptors that recognize molecules expressed by stressed cells provides a mechanism of costimulation that could lower the threshold for triggering clonal proliferation by still undefined self p-MHC (50). The evidence of extensive clonal expansion among the valveinfiltrating T cells in CAS gives further impetus to changing the paradigm of this disease from a passive, irreversible, degenerative process to one based, at least in part, on inflammation and injury mediated by elements of the adaptive immune response. It also suggests a rationale for studying whether therapy directed to the expanded activated T cell clones in CAS might retard the otherwise relentless progression of this disease.

Acknowledgment We thank Dr. Myron Weisfeldt for suggesting the collaborative approach to this problem.

Disclosures The authors have no financial conflict of interest.

References 1. Carabello, B. A., and F. A. Crawford, Jr. 1997. Valvular heart disease. N. Engl. J. Med. 337: 32– 41. 2. Stewart, B. F., D. Siscovick, B. K. Lind, J. M. Gardin, J. S. Gottdiener, V. E. Smith, D. W. Kitzman, and C. M. Otto. 1997. Clinical factors associated with calcific aortic valve disease: cardiovascular health study. J. Am. Coll. Cardiol. 29: 630 – 634. 3. Lindroos, M., M. Kupari, J. Heikkila, and R. Tilvis. 1993. Prevalence of aortic valve abnormalities in the elderly: an echocardiographic study of a random population sample. J. Am. Coll. Cardiol. 21: 1220 –1225. 4. Pomerance, A. 1972. Pathogenesis of aortic stenosis and its relation to age. Br. Heart J. 34: 569 –574.

The Journal of Immunology 5. Olsson, M., C. J. Dalsgaard, A. Haegerstrand, M. Rosenqvist, L. Ryden, and J. Nilsson. 1994. Accumulation of T lymphocytes and expression of interleukin-2 receptors in nonrheumatic stenotic aortic valves. J. Am. Coll. Cardiol. 23: 1162–1170. 6. Otto, C. M., J. Kuusisto, D. D. Reichenbach, A. M. Gown, and K. D. O’Brien. 1994. Characterization of the early lesion of ‘degenerative’ valvular aortic stenosis: histological and immunohistochemical studies. Circulation 90: 844 – 853. 7. Wallby, L., B. Janerot-Sjoberg, T. Steffensen, and M. Broqvist. 2002. T lymphocyte infiltration in non-rheumatic aortic stenosis: a comparative descriptive study between tricuspid and bicuspid aortic valves. Heart 88: 348 –351. 8. Soini, Y., T. Salo, and J. Satta. 2003. Angiogenesis is involved in the pathogenesis of nonrheumatic aortic valve stenosis. Hum. Pathol. 34: 756 –763. 9. Mazzone, A., M. C. Epistolato, R. De Caterina, S. Storti, S. Vittorini, S. Sbrana, J. Gianetti, S. Bevilacqua, M. Glauber, and A. Biagini. 2004. Neoangiogenesis, T-lymphocyte infiltration, and heat shock protein-60 are biological hallmarks of an immunomediated inflammatory process in end-stage calcified aortic valve stenosis. J. Am. Coll. Cardiol. 43: 1670 –1676. 10. Olsson, M., M. Rosenqvist, and J. Nilsson. 1994. Expression of HLA-DR antigen and smooth muscle cell differentiation markers by valvular fibroblasts in degenerative aortic stenosis. J. Am. Coll. Cardiol. 24: 1664 –1671. 11. Rajamannan, N. M., M. Subramaniam, D. Rickard, S. R. Stock, J. Donovan, M. Springett, T. Orszulak, D. A. Fullerton, A. J. Tajik, R. O. Bonow, and T. Spelsberg. 2003. Human aortic valve calcification is associated with an osteoblast phenotype. Circulation 107: 2181–2184. 12. Ortlepp, J. R., F. Schmitz, T. Bozoglu, P. Hanrath, and R. Hoffmann. 2003. Cardiovascular risk factors in patients with aortic stenosis predict prevalence of coronary artery disease but not of aortic stenosis: an angiographic pair matched case-control study. Heart 89: 1019 –1022. 13. Tanaka, K., M. Sata, D. Fukuda, Y. Suematsu, N. Motomura, S. Takamoto, Y. Hirata, and R. Nagai. 2005. Age-associated aortic stenosis in apolipoprotein E-deficient mice. J. Am. Coll. Cardiol. 46: 134 –141. 14. Otto, C. M., and K. D. O’Brien. 2001. Why is there discordance between calcific aortic stenosis and coronary artery disease? Heart 85: 601– 602. 15. Otto, C. M. 2002. Calcification of bicuspid aortic valves. Heart 88: 321–322. 16. Cowell, S. J., D. E. Newby, R. J. Prescott, P. Bloomfield, J. Reid, D. B. Northridge, and N. A. Boon. 2005. A randomized trial of intensive lipid-lowering therapy in calcific aortic stenosis. N. Engl. J. Med. 352: 2389 –2397. 17. Stemme, S., L. Rymo, and G. K. Hansson. 1991. Polyclonal origin of T lymphocytes in human atherosclerotic plaques. Lab. Invest. 65: 654 – 660. 18. Oksenberg, J. R., G. T. Stavri, M. C. Jeong, N. Garovoy, J. R. Salisbury, and J. D. Erusalimsky. 1997. Analysis of the T-cell receptor repertoire in human atherosclerosis. Cardiovasc. Res. 36: 256 –267. 19. Swanson, S. J., A. Rosenzweig, J. G. Seidman, and P. Libby. 1994. Diversity of T-cell antigen receptor V␤ gene utilization in advanced human atheroma. Arterioscler. Thromb. 14: 1210 –1214. 20. Li, Y., R. F. Schwabe, T. DeVries-Seimon, P. M. Yao, M. C. Gerbod-Giannone, A. R. Tall, R. J. Davis, R. Flavell, D. A. Brenner, and I. Tabas. 2005. Free cholesterol-loaded macrophages are an abundant source of tumor necrosis factor-␣ and interleukin-6: model of NF-␬B- and map kinase-dependent inflammation in advanced atherosclerosis. J. Biol. Chem. 280: 21763–21772. 21. Curran, S. A., O. M. FitzGerald, P. J. Costello, J. M. Selby, D. J. Kane, B. Bresnihan, and R. Winchester. 2004. Nucleotide sequencing of psoriatic arthritis tissue before and during methotrexate administration reveals a complex inflammatory T cell infiltrate with very few clones exhibiting features that suggest they drive the inflammatory process by recognizing autoantigens. J. Immunol. 172: 1935–1944. 22. Costello, P. J., R. J. Winchester, S. A. Curran, K. S. Peterson, D. J. Kane, B. Bresnihan, and O. M. FitzGerald. 2001. Psoriatic arthritis joint fluids are characterized by CD8 and CD4 T cell clonal expansions appear antigen driven. J. Immunol. 166: 2878 –2886. 23. Jabri, B., J. M. Selby, H. Negulescu, L. Lee, A. I. Roberts, A. Beavis, M. Lopez-Botet, E. C. Ebert, and R. J. Winchester. 2002. TCR specificity dictates CD94/NKG2A expression by human CTL. Immunity 17: 487. 24. Lefranc, M. P. 2003. IMGT, the international ImMunoGeneTics database. Nucleic Acids Res. 31: 307–310. 25. Arden, B., S. P. Clark, D. Kabelitz, and T. W. Mak. 1995. Human T-cell receptor variable gene segment families. Immunogenetics 42: 455–500. 26. Giudicelli, V., D. Chaume, and M. P. Lefranc. 2004. IMGT/V-QUEST, an integrated software program for immunoglobulin and T cell receptor V-J and V-D-J rearrangement analysis. Nucleic Acids Res. 32: W435–W440.

5339 27. Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25: 3389 –3402. 28. Accelrys. 2001. Program Manual for the Wisconsin Package, Version 10.2. Madison: Accelrys. 29. Altschul, S. F. 1991. Amino acid substitution matrices from an information theoretic perspective. J. Mol. Biol. 219: 555–565. 30. Slachta, C. A., V. Jeevanandam, B. Goldman, W. L. Lin, and C. D. Platsoucas. 2000. Coronary arteries from human cardiac allografts with chronic rejection contain oligoclonal T cells: persistence of identical clonally expanded TCR transcripts from the early post-transplantation period (endomyocardial biopsies) to chronic rejection (coronary arteries). J. Immunol. 165: 3469 –3483. 31. Miyata, T., S. Miyazawa, and T. Yasunaga. 1979. Two types of amino acid substitutions in protein evolution. J. Mol. Evol. 12: 219 –236. 32. McCaldon, P., and P. Argos. 1988. Oligopeptide biases in protein sequences and their use in predicting protein coding regions in nucleotide sequences. Proteins 4: 99 –122. 33. Maynard, J., K. Petersson, D. H. Wilson, E. J. Adams, S. E. Blondelle, M. J. Boulanger, D. B. Wilson, and K. C. Garcia. 2005. Structure of an autoimmune T cell receptor complexed with class II peptide-MHC: insights into MHC bias and antigen specificity. Immunity 22: 81–92. 34. Guex, N., and M. C. Peitsch. 1997. SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis 18: 2714 –2723. 35. Turner, S. J., P. C. Doherty, J. McCluskey, and J. Rossjohn. 2006. Structural determinants of T-cell receptor bias in immunity. Nat. Rev. 6: 883– 894. 36. Liuzzo, G., J. J. Goronzy, H. Yang, S. L. Kopecky, D. R. Holmes, R. L. Frye, and C. M. Weyand. 2000. Monoclonal T-cell proliferation and plaque instability in acute coronary syndromes. Circulation 101: 2883–2888. 37. Biasucci, L. M., G. Liuzzo, D. J. Angiolillo, G. Sperti, and A. Maseri. 2000. Inflammation and acute coronary syndromes. Herz. 25: 108 –112. 38. Caligiuri, G., G. Paulsson, A. Nicoletti, A. Maseri, and G. K. Hansson. 2000. Evidence for antigen-driven T-cell response in unstable angina. Circulation 102: 1114 –1119. 39. Wick, G., M. Knoflach, and Q. Xu. 2004. Autoimmune and inflammatory mechanisms in atherosclerosis. Annu. Rev. Immunol. 22: 361– 403. 40. Nadareishvili, Z. G., D. E. Koziol, B. Szekely, C. Ruetzler, R. LaBiche, R. McCarron, and T. J. DeGraba. 2001. Increased CD8⫹ T cells associated with Chlamydia pneumoniae in symptomatic carotid plaque. Stroke 32: 1966 –1972. 41. De Palma, R., F. Del Galdo, G. Abbate, M. Chiariello, R. Calabro, L. Forte, G. Cimmino, M. F. Papa, M. G. Russo, G. Ambrosio, et al. 2006. Patients with acute coronary syndrome show oligoclonal T-cell recruitment within unstable plaque: evidence for a local, intracoronary immunologic mechanism. Circulation 113: 640 – 646. 42. McNally, J. M., and R. M. Welsh. 2002. Bystander T cell activation and attrition. Curr. Top. Microbiol. Immunol. 263: 29 – 41. 43. Morley, J. K., F. M. Batliwalla, R. Hingorani, and P. K. Gregersen. 1995. Oligoclonal CD8⫹ T cells are preferentially expanded in the CD57⫹ subset. J. Immunol. 154: 6182– 6190. 44. Hamann, D., S. Kostense, K. C. Wolthers, S. A. Otto, P. A. Baars, F. Miedema, and R. A. van Lier. 1999. Evidence that human CD8⫹CD45RA⫹CD27⫺ cells are induced by antigen and evolve through extensive rounds of division. Int. Immunol. 11: 1027–1033. 45. Posnett, D. N., J. W. Edinger, J. S. Manavalan, C. Irwin, and G. Marodon. 1999. Differentiation of human CD8 T cells: implications for in vivo persistence of CD8⫹CD28⫺ cytotoxic effector clones. Int. Immunol. 11: 229 –241. 46. Posnett, D. N., R. Sinha, S. Kabak, and C. Russo. 1994. Clonal populations of T cells in normal elderly humans: the T cell equivalent to “benign monoclonal gammapathy.” J. Exp. Med. 179: 609 – 618. 47. Goronzy, J. J., and C. M. Weyand. 2003. Aging, autoimmunity and arthritis: T-cell senescence and contraction of T-cell repertoire diversity: catalysts of autoimmunity and chronic inflammation. Arthritis Res. Ther. 5: 225–234. 48. Effros, R. B., M. Dagarag, C. Spaulding, and J. Man. 2005. The role of CD8⫹ T-cell replicative senescence in human aging. Immunol. Rev. 205: 147–157. 49. Hadrup, S. R., J. Strindhall, T. Kollgaard, T. Seremet, B. Johansson, G. Pawelec, P. thor Straten, and A. Wikby. 2006. Longitudinal studies of clonally expanded CD8 T cells reveal a repertoire shrinkage predicting mortality and an increased number of dysfunctional cytomegalovirus-specific T cells in the very elderly. J. Immunol. 176: 2645–2653. 50. Raulet, D. H. 2003. Roles of the NKG2D immunoreceptor and its ligands. Nat. Rev. 3: 781–790.