28 CHAPTER 2 Results Do various clones dominate the T cell response at different locations within one single inflamed joint? In large joints, ST biopsy specimens can be obtained from multiple anatomic locations. For example, in the knee, biopsy specimens can be taken from either the SP or IP region. It is unknown whether the T cell repertoires are different at these distinctive locations. To investigate this, we included seven RA patients (see Table 1 for characteristics) with inflamed knee joints and took biopsy specimens from both the SP and IP regions from both inflamed joints (in total, n = 14) and analyzed the TCRβ repertoires. For comparison, we used TCR repertoires from the paired PB samples. We used 26,866 quality-filtered, randomly selected TCRβ sequences per sample. The number of TCRβ clones found in the SP and IP regions (mean 6679 [SD 2612] versus 7869 [SD 2550], respectively) and the number of HECs (mean 8.0 [SD 3.8] versus 7.6 [SD 3.8]) were comparable (Fig. 1A, 1B). The HECs accounted for 29.9% (mean, SD 23.4) and 18.2% (SD 12.8) of the total repertoire, respectively (Fig. 1C). PB showed a similar number of TCRβ clones (9982, SD 5898) with a mean of 6.9 HECs (SD 6.2) that accounted for 40.7% (SD 32.4) of the total repertoire (Fig. 1A–C). Subsequently, using PB as control, we analyzed to what extent the 25 most expanded (top 25) TCRβ clones in the SP and IP regions showed overlap. Of the 25 most expanded TCRβ clones in the SP region, 54.3% (mean, SD 17.7) were also present among the 25 most expanded TCRβ clones in the IP region. This overlap was significantly lower when comparing the IP samples with PB (mean 20.0%, SD 8.0, p < 0.0001; Fig. 1F), leading us to believe that dominant TCRβ clones from the ST are hardly present in the PB. Comparable results were found if we restricted the analysis to the top 10 or top 100 TCRβ clones (Supplemental Fig. 1). Of the top 25 SP TCRβ clones, 86% could be retrieved among the top 1000 of the IP TCRβ clones (Supplemental Fig. 2), showing that the vast majority of the expanded SP TCRβ clones is also present in the IP region. If we do not focus on the most dominant clones, but instead look at the total measured TCRβ repertoire, we used the Chao-modified Sørensen index to test for similarity between the different regions. For the measured TCRβ repertoires in SP and IP regions, we thus observed a score of 0.49 (mean, SD 0.11), which is significantly higher than that for the comparison of ST and PB (mean 0.15, SD 0.07; p < 0.0001) (Fig. 1G). For these analyses, we only included the patients from whom we had a complete dataset, and therefore paired analyses were possible.
RkJQdWJsaXNoZXIy MTk4NDMw