Response to: “Some considerations about γδ and CD8+ T-cell responses in blood after gluten challenge in treated celiac disease”

We appreciate the interest for our recent paper¹ by Núñez & Fernández-Bañares.² While acknowledging the main finding of our study that in celiac disease (CeD) patients there is sharing of TCRγδ and TCRαβ T-cell clonotypes between gut cells and CD103⁺CD38⁺ blood cells appearing in blood 6 days after oral gluten challenge, they are concerned that we found an increase in CD103⁺CD38⁺ blood T cells on day 6 in only a subset of patients, unlike previous reports.^3,4 Núñez & Fernández-Bañares raise three possible explanations for this lower degree of T-cell responsiveness in our study; (a) error arising from death of cells in cryopreserved samples (b) unclear criteria for scoring of positive response and (c) a low number of cells analyzed in each T-cell subset.

The protocols we employed in our study were chiefly based on those published by Han et al.³ As for (a) Han et al. analyzed cryopreserved cells. Yet, we were concerned about cryopreservation being a confounder. Therefore, as part of our study we performed comparative analysis of fresh or cryopreserved samples looking at the frequencies of CD103⁺ CD38⁺ CD8⁺αβ T cells and CD103⁺CD38⁺ γδ T cells in peripheral blood mononuclear cells (PBMCs) obtained from day-6 samples from six orally challenged CeD patients. In our hands, there were no obvious differences between fresh and cryopreserved samples (Fig. 1a). Likewise, in the hands of Núñez & Fernández-Bañares, the difference between fresh and frozen samples did not reach significance (p = 0.19). Further to the issue of cells analyzed, while our analysis¹ and that of Han et al.³ were performed on isolated PBMCs, Núñez & Fernández-Bañares analyzed whole fresh blood without pre-isolation of blood cells.² This factor could also contribute to different results of the studies.

Fig. 1: Analysis of CD103⁺ CD38⁺ T cells from blood.

a The effect of cryopreservation on the frequency of CD103⁺ CD38⁺ CD8⁺ αβ and CD103⁺ CD38⁺ γδ T cells. Parallel fresh and cryopreserved PBMC samples of day-6 samples from 6 challenged patients were compared. b Depiction of number of CD8⁺ αβ and γδ T cells analyzed from blood of treated CeD patients (n = 12) being part of 3-day gluten challenges as reported in¹. Patients challenged with cookie or bread are depicted with circles and triangles, respectively. The p value was calculated using a two-tailed paired Wilcoxon test.

For (b) we used criteria for a positive response as defined by Han et al.³. We agree that better criteria for positivity can be defined. However, the scope of our study was not to define the rate of responsiveness; rather we aimed to study clonal relatedness over time and in different body compartments.

As for (c) we agree that it is crucial to analyze sufficient numbers of cells when investigating rare cell populations. This is especially so for γδ T cells as they constitute approximately 5% of CD3⁺ T cells of which only ~1% express CD103. As our main focus was to investigate clonal relatedness of blood-derived and intraepithelial CD8⁺αβ and γδ T cells, we enriched γδ T cells with beads and index-sorted cells from a 14-day challenge protocol to ensure optimal sequencing coverage.¹ However, this came at the cost of recording high numbers of γδ T cells. To rule out an issue of insufficient cell sampling, in a 3-day challenge of 12 patients, we made sure to assess the frequency of CD103⁺ CD38⁺ CD8αβ and γδ T cells analyzing a minimum number of cells (>20,000 γδ T cells). The numbers of CD103⁺CD38⁺ CD8⁺αβ and γδ T cells analyzed for each patient in this series of experiments are shown in Fig. 1b. We thus believe that too few analyzed cells is unlikely to be an explanation for our findings.

Reassuringly, there seems to be consensus on the fact 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 celiac disease patients. The possible discord relates to the degree of the response. We will argue that there could be real differences between studies, and we would like to mention two factors that we believe affect T-cell responsiveness to oral gluten challenge. One factor is difference in how quiescent the disease is among subjects included in the various studies. As discussed by Sarna et al.⁵, and as recently supported by tissue proteomics analysis⁶, it is conceivable that in subjects with an existing low degree of gluten-induced immunity at baseline respond more vigorously to the gluten challenge. Another factor is that the degree of response is affected by amount of gluten or type of gluten vehicle used for the challenge. A 14-day gluten challenge study found more blood α4β7⁺ CD8⁺ T cells as well as HLA-DQ2.5:gluten peptide tetramer⁺ CD4⁺ T cells on day 6 when the gluten challenge was done with 10 g/day compared to 3 g/day⁷. In our study we used different gluten vehicles.¹ Possibly bread challenge gave better T-cell responses than challenge with müslibar or cookie, yet the small number of patients in each group prevent us from concluding whether there indeed was a difference between the vehicles used. As oral gluten challenge will be conducted in future studies of CeD pathogenesis, including in studies of CD103⁺ CD38⁺ CD8⁺ αβ and γδ T cells, we will advocate for standardization of the amount and type of gluten antigen used in such studies.