We have read the recent publication by Risnes et al. ‘Circulating CD103+ γδ and CD8+ T cells are clonally shared with tissue-resident intraepithelial lymphocytes in celiac disease’ in Mucosal Immunology.1 The authors make a great contribution showing that the gut-homing γδ and CD8+ T cells found in the blood of coeliac disease (CeD) patients after gluten re-exposure displayed a T-cell repertoire partially overlapping with intraepithelial lymphocytes in the gut. However, the authors also claim that, in contrast to previous reports,2,3 they could detect the increase of blood CD38+ CD103+ γδ and CD8+ αβ T cells on day 6 after a 3-day gluten challenge only in a subset of CeD patients, while the gluten-induced CD4+ T-cell response was detected in most patients. We find several possible reasons that could explain the apparent discrepancy.

First, it could be due to a methodological issue. We worked with fresh blood instead of cryopreserved material, as Risnes et al. did. We compared the results obtained on day 6 after gluten challenge from ten of our CeD patients using those two starting materials (Fig. 1). Although non-significant differences were found, it can be observed that the percentage of the CD8+ T-cell subset studied was always higher using fresh blood.

Fig. 1: Comparison between the percentages of CD103+ β7+ CD38+ CD8+ T cells in relation to the total number of CD8+ T cells observed in celiac disease patients on day 6 of a 3-day gluten challenge using frozen PBMCs or fresh blood.
figure 1

Lines connect results from individual patients. Centerlines in the box show the median; box limits indicate the 25 and 75th percentiles (IQR); whiskers extend to the highest and lowest observation excluding outliers. P value was calculated using a two-tailed paired Wilcoxon test. Y scale was converted to binary logarithm because data cover a wide range of values.

Full size image

Second, the authors do not explain what they have considered as a γδ or a CD8+ T-cell response to the gluten challenge. In our experience, these T cells are very unusually observed in non-CeD patients and very small increases in the percentage of the studied cells at day 6 are indicative of CeD. Comparison with a non-CeD control group is required to establish the cut-off point, considering an increase in percentage, defining a positive T-cell response. Therefore, it remains the possibility that the response exists in more patients than those claimed by Risnes et al. in their paper, as suggested by the high correlation they observed between γδ or CD8+ T-cell responses and the response characterised by HLA-DQ2.5 gluten tetramer CD4+ T cells (r = 0.84 and r = 0.73 between tetramer+ CD4+ T cells vs. CD8+ and vs. γδ T cells on day 6, respectively).

Finally, an additional difference might lie in the total number of cells analysed in each T-cell subset. γδ T cells are very scarce in blood and the low numbers present in many patients preclude to detect the gluten-induced γδ T-cell response. We already alerted about this issue in our report,3 which can also involve CD8+ T cells in some individuals. As already stated in the original work of Han et al.2 the extent of the CD103+ β7+ CD38+ γδ as well as CD8+ T-cell responses varies greatly between patients.

In summary, it seems clear that a 3-day oral gluten challenge induces the activation and concomitant presence in blood of activated gut-homing γδ, CD8+ and CD4+ T cells in treated CeD patients. More in-depth analyses are required before asserting specific comparisons between those cell types.