Oligoclonal T Cell Expansions in Atherosclerotic Lesions of Apolipoprotein E–Deficient Mice

Methods

Results

ApoE-KO mice were divided into 2 groups that were either fed a high-cholesterol diet with 1.25% cholesterol or a western diet containing 0.15% cholesterol. This leads to different progression rates: whereas mice fed a western diet for 10 weeks exhibited fatty streaks, those fed the high-cholesterol diet for the same length of time had developed fibrofatty plaques.^{33 34} The latter type of lesions were observed in mice maintained on the western diet for 18 weeks, whereas mice fed the high-cholesterol diet for 18 weeks exhibited large, advanced fibrofatty plaques (Figure 1⇓). The different diet treatments result in different immune effector mechanisms,¹⁶ and the present design permitted us to evaluate whether the immune detector mechanisms also varied depending on serum cholesterol levels or disease progression rate.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

CD4⁺ T cells in fibrofatty plaques from aortic root of apoE^−/− mice. Immunohistochemical staining of lesions from mice fed the high-cholesterol diet for 10 weeks (EH10) or 18 weeks (EH18) as well as mice fed the western diet for 18 weeks (EW18) is shown. The frames in the top panels are shown in the bottom panels at higher magnification. Magnification ×50 (top panels) and ×200 (bottom panels). The arrows point at sites in lesions that stain positive for CD4⁺ T cells. A calcification is marked with an asterisk.

An immunohistochemical analysis of lesions from the different groups of mice confirmed the abundance of CD4⁺ T cells in all types of lesions (Figure 1⇑).

Atherosclerotic lesions were microdissected from the aortic root of each mouse. The sample from each mouse (Table 1⇑) was treated individually (except samples 8 to 10; see Methods). The mRNA was isolated individually and used for RT-PCR analysis (see below). For comparison, mRNA from lymph node tissue was also analyzed. The TCR V region repertoire was analyzed by using 5′ primers specific for the different V domains in combination with a common 3′ primer from the C domain. This analysis provides information about the usage of different V domains by T cells in lesions. However, immune specificity is determined not only by V domain usage but also by nucleotide transferase activity associated with recombination events during T-cell differentiation. This activity adds or removes nucleotides at the various Vα-Jα, Vβ-Dβ, and Dβ-Jβ junctions in the maturing TCR genes and results in variations in the length of the mature TCRα and TCRβ genes. Together, V domain usage and junctional variability determine the conformation of the complementarity-determining region 3 (CDR3 region) of the TCR, which in turn determines the binding specificity of the antigen receptor.^{35 36} To assess junctional diversity as well as V domain usage, PCR was followed by fragment length analysis on high-resolution polyacrylamide gels in a DNA sequencer. The spectrum of size differences across the CDR3 region is known as the spectratype. Because the relative intensity of a given size peak is proportional to the amount of mRNA molecules in the starting material, an increase in height and area of a particular peak signals expansion of T-cell clones.¹⁷

When Vα and Vβ profiles were run on lymph nodes, each TCR type (except Vα9 and Vβ12) showed a gaussian distribution (Figures 2⇓ and 3⇓). Each peak within one V region profile differed by 3 bp, corresponding to peptide differences of 1 amino acid. Five to 7 different peaks were found for each TCR fragment, corresponding to a variation in length of 5 to 7 amino acids within the CDR3 region.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Spectratype profiles of the CDR3 length diversity in the Vα chain mRNA of TCR in atherosclerotic plaques. mRNA was extracted from plaques or lymph nodes (control tissue), and RT-PCR analysis of the CDR3 region in the TCR was performed as described in Methods. The PCR reaction was run with a fluorescently labeled common C-region primer and V-region–specific primers for each Vα. PCR products were analyzed on high-resolving PAGE in an automatic sequencer. The y axis shows fluorescence intensity; the x axis, CDR3 size. The boxed numbers show the length in base pairs of the PCR products in the corresponding peak. Each row contains all PCR products derived from one sample with use of the complete panel of Vα primers. All rows show the analysis from individual apoE^−/− mice (m2, m3, etc), except for m8 to 10, for which samples from 3 mice were pooled. Sample number (m2 to m13), time on diet in weeks (w), and type of diet (either high-cholesterol diet [HCD] or western diet [WD]) are indicated. Top row, Lymph nodes of apoE^−/− mice fed HCD for 10 weeks show gaussian distribution of CDR3 profiles corresponding to a polyclonal T-cell population. Subsequent rows, Plaques from apoE^−/− mice have strongly skewed repertoires for the Vα TCR, indicating monoclonal or oligoclonal T-cell populations.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Spectratype profiles of the CDR3 length diversity in the Vβ chain of the TCR in T cells infiltrating the atherosclerotic plaque. The analysis was performed as in Figure 2⇑ and on the same samples; eg, sample m2 is from the same mouse in both Vα and Vβ analyses. Top row, Lymph nodes of apoE^−/− mice fed HCD for 10 weeks show gaussian distribution of CDR3 profiles corresponding to a polyclonal T-cell population. Subsequent rows, Plaques of apoE^−/− mice exhibit strongly skewed repertoires for the Vβ TCR, indicating monoclonal or oligoclonal T-cell population.

The TCR analysis of plaque tissue gave a totally different result, as shown in Figures 2⇑ and 3⇑ and in Table 3⇓. First, only a limited set of V domains was represented in each of the 10 samples. The number of V domains expressed in each aorta varied between 1 and 8, with an average of 5.2 Vα and 5.6 Vβ domains. Second, the detected TCR fragments exhibited strongly skewed profiles and never showed the gaussian distribution found in lymph nodes. In several samples, only one peak was visible, implying a monotypic expansion of T cells (eg, Figure 3⇑, row m2, Vβ6).

Vβ6 was expressed as a dominant peak in all aortic samples (Figure 3⇑ and Table 3⇑). In 4 of them, a single peak of TCR Vβ6 was observed, indicating a monoclonal T-cell population within these plaques (eg, Figure 3⇑, rows m2 and m7). In the 6 remaining samples, 2 expanded peaks were identified (eg, Figure 3⇑, rows m5 and m6). The second most common Vβ domains, Vβ5.2 and Vβ16, were expressed in 5 of the 10 samples.

In the Vα analysis, the most frequent V domain was 34S, which was found in 9 of 10 samples (Table 3⇑). As for Vβ6, the profile was skewed, and one peak was dominant. The second most common Vα domain expressed was Vα9, which was also present as a single peak. However, it occurred as a single peak also in the lymph nodes. This transcript is probably derived from a rearranged pseudogene.²²

When comparing the mice fed a western diet with those maintained on the high-cholesterol diet, we found no differences in TCR profiles. In contrast, the stage of lesions appeared to influence the TCR patterns: more advanced lesions exhibited a more restricted TCR heterogeneity than did earlier stages of disease. This was particularly evident for Vβ domains; compare samples m2 to m4 in Figure 3⇑, which represent advanced fibrofatty lesions, with samples m11 to m13, which are derived from early fibrofatty plaques.

Discussion

A cellular immune response to a specific antigen induces proliferation of antigen-specific T cells, and this leads to expansion of T-cell clones carrying one unique TCR. Because T cells are important components of the atherosclerotic lesion at all stages and may recognize antigens present in the lesions, it is important to clarify whether clonal expansions of antigen-specific T cells occur in atherosclerosis.^{1 3} Previous studies of complicated human carotid lesions have failed to provide evidence for such expansions,^{13 14} but the situation may be entirely different in other phases of atherosclerosis. The development of murine models of the disease permits a dissection of all the phases of the disease with the use of clonotypic T-cell markers. The use of the recently developed spectratyping RT-PCR technique reveals the Vα and Vβ gene usage together with the unique length of the CDR3 region in each T cell and serves as a clonotypic indicator.¹⁷ Because the mouse models have defined genotypes and not the same degree of complexity in the major histocompatibility complex (MHC) gene cluster as the human patients show in their HLA, this will be reflected in the TCR usage. In particular, the V domain usage will be less diverse, because all the mice in the present study carry the same MHC genes. Finally, the sensitivity of spectratyping allows an analysis of the antigen-binding domain of the TCR at a level that was not possible before when only Southern blot analysis¹⁴ or standard PCR¹³ was available.

With this approach of spectratyping RT-PCR of the TCR, we now demonstrate that the T-cell population of atherosclerotic lesions is highly skewed compared with the gaussian distribution found in lymph nodes. Not only is the representation of V domains much more limited in the lesions, but the CDR3 variability is also much more restricted. Interestingly, TCR heterogeneity was more reduced in the more mature plaques compared with earlier ones. This could imply that a heterogeneous population of T cells is initially recruited to the forming lesion by antigenically nonspecific mechanisms, followed by a selective expansion of T cells carrying specific reactivity to local antigens. The comparison with very advanced human lesions suggests that such a selective expansion is followed by extended chronic inflammation, leading to adhesion and immigration of heterogeneous T cells and a dilution of the specific population. This possibility should be tested when the precise specificities of the TCR represented in the human lesions are known.

One TCR V domain was represented in all lesions: Vβ6. Only 1 or 2 Vβ6 peaks were found in most samples, and the dominant peaks were 179 to 182 bp and thus differed by only 1 codon. It seems likely that this expansion of TCR Vβ6 was caused by proliferation of antigen-specific T cells carrying this TCR. This expansion could theoretically arise either from an expanded population of activated Vβ6⁺ T cells in the blood or by local proliferation of Vβ6⁺ T cells in the plaque. Because no expansion of Vβ6⁺ T cells was observed in lymph nodes, it is less probable that a systemically expanded Vβ6 population was recruited to the lesion. Instead, the most likely explanation for the selective expansion of Vβ6 mRNA in lesions is that Vβ6⁺ T cells undergo local clonal proliferation in the lesion itself. This, in turn, suggests that plaque components are presented as antigens to local T cells. We have previously shown that CD4⁺ T cells of human lesions recognize oxidized LDL, and we have recently observed that murine T-cell hybridomas reactive with oxidation-induced LDL epitopes express Vβ6 (A. Nicoletti, G. Paulsson, G.K. Hansson, unpublished observation, 1999). Together, these results suggest that TCR Vβ6 is used in the recognition of oxidized LDL in atherosclerotic lesions, leading to clonal expansion of CD4⁺ T cells carrying this type of TCR.

In contrast to the dominance of Vβ6, no single Vα domain was expressed in any lesion sample. Although TCR Vα34S was present in 9 of 10 lesion samples, it was missing in 1 of 3 early fibrofatty plaques of the apoE-KO mice fed a western diet (Figure 2⇑, m7). Therefore, Vα expression was more heterogeneous than Vβ expression, possibly reflecting a promiscuous usage of Vα domains in the generation of antigen-specific TCR–recognizing plaque antigens. This interpretation is in line with recent observations that each T-cell epitope can be recognized by many different TCRs and that each clonotypic TCR can recognize many different epitopes.^{37 38}

In spite of the relative heterogeneity of Vα expression, the selective expression of T cells carrying Vβ6 suggests that such cells are important in the pathogenesis of atherosclerosis. Autoimmune diseases such as experimental autoimmune encephalomyelitis have been reversed by treatment with monoclonal antibodies specific for the TCR-Vβ type carried by disease-mediating T cells.³⁹ It will now be interesting to determine whether atherosclerosis can also be controlled by such reagents.