Relative Contribution of Naïve and Memory T Cells to Alloreactivity in Hematopoietic Cell Transplantation

Introduction: Graft-versus-host disease (GvHD) is a major impediment to the cure of blood disorders by hematopoietic cell transplantation (HCT). GvHD is mediated by alloreactive T cells recognizing histocompatibility antigen (HAg) mismatches between patient and donor. Naïve T cells are thought to be the main mediators of alloreactive responses since, theoretically, memory T cells would have never been exposed to and selected by alloantigens, except in multiparous women or transfused individuals. Accordingly, clinical trials using naïve T cell-depleted allografts are being conducted with the aim to reduce GvHD after human leukocyte antigen (HLA)-matched HCT. However, several groups have shown that memory T cells can also mediate alloreactive responses, in particular against mismatched HLA. We hypothesized that the relative importance of naïve vs. memory T cell alloreactivity depends on the matching status of the patient-donor pair. Specifically, we reasoned that naïve-depletion strategies will be most efficient in HLA-identical sibling HCT, where minor (m)HAg presented by self-HLA are the only targets of T cell alloreactivity, but less so in HLA-matched unrelated HCT, where HLA-DPB1 mismatches (mmDPB1) are frequent and potentially recognized through molecular mimicry by both naïve and memory T cells.

Methods: In order to model T cell alloreactivity to mHAg and to major HLA mismatches post HCT, we used a quantitative in vitro assay based on co-culture of responder and stimulating cells. Naïve (CD45RA⁺CD45RO^-) and memory (CD45RA^-CD45RO⁺) CD4⁺ T cells were enriched from peripheral blood mononuclear cells from healthy individuals using microbead technology to >95% purity and used as responders. Irradiated transduced HeLa cells engineered to express single HLA-DP antigens and the necessary machinery for HLA class II antigen presentation were used to stimulate CD4⁺ T cells. HeLa transductants expressing the autologous (i.e. DP-matched, response restricted to mHAg) or an allogeneic (mmDPB1) DP antigen were used to challenge naïve and memory CD4⁺ cells from each responder. After 14 days of culture, T cells were restimulated overnight and the levels of T cell response were quantified by cell surface expression of the activation marker CD137.

Results: In 36 independent T cell cultures from 8 different individuals, the overall levels of alloreactivity against mHAg were significantly lower than those against mmDPB1 (mean 50.3% vs 20.7%, p<0.0001) (Figure 1A). Consistent with current concepts, alloreactivity to mHAg was significantly higher in the naïve than in the memory subset (mean 27.7% vs 10.5%, p=0.015) (Figure 1B). This was most evident in 5/8 responders (mean 38.4% vs 13.3%, p=0.016), in particular in females under 40 years of age. In 3 of the 8 responders, mHAg alloreactivity was generally low and not significantly different between the naïve and the memory subsets (mean 10.3% vs 12.9%, p=0.73). In contrast, alloreactivity against mmDPB1 was evenly distributed between the naïve and the memory subset (mean 52.1% vs 48.5%, p=0.62) in all responders, independent of age, sex or cytomegalovirus serostatus of the responder (Figure 1C). Interestingly, naïve DPB1*04:01-restricted mHAg alloreactive CD4⁺ T cells were able to cross-recognize the structurally similar (i.e. permissive) DPB1*04:02 (mean 43.3%) but not the dissimilar (i.e. non-permissive) DPB1*09:01 (mean 14.1%) (Figure 1D). Moreover, when purified CD4⁺ cells from self-DPB1*04:01 homozygous donors were challenged with DPB1*04:02 or DPB1*09:01, naïve CD4⁺ T cells were the main source of alloreactive responses against the permissive mmDPB1 (mean 25.0% vs 7.4% for naïve and memory cells, respectively), while both memory (mean 50.0%) and naïve (mean 46.0%) CD4⁺ cells elicited strong alloresponses against the non-permissive mmDPB1.

Conclusion: Our data provide the first direct experimental evidence that alloreactivity against mmDPB1 is stronger than against mHAg, and importantly that it is mediated equally by naïve and memory CD4⁺ T cells while the mHAg response is mediated mainly by the naïve subset. However, our data also suggests that some mmDPB1 involving structurally (and hence functionally) similar alleles (in general permissive) might behave similarly to DPB1 matches. These observations should be taken into account in clinical trials aimed at improving the outcome of unrelated HCT by selective depletion of naïve T cells.

View largeDownload slide

View largeDownload slide

Close modal