MHC Class II on Ccl19⁺ Reticular Cells Mitigates Graft-Versus-Host Disease Via Maintenance of Regulatory T Cells

Allogeneic T cell priming is considered as an essential event determining the outcome of allogeneic hematopoietic stem cell transplantation (allo-HCT), ideally triggering anti-leukemic responses (GvL effect) or, at worst, causing life-threatening acute graft-versus-host disease (aGvHD). During aGvHD initiation, alloreactive T cells are activated by host antigen presenting cells (APCs), rapidly expand and subsequently exert tissue damage. Recently, it was discovered that in absence of host hematopoietic APCs, aGvHD cannot be prevented, suggesting a crucial role of non-hematopoietic APCs for priming alloreactive T cells (Toubai et al., Blood 2012, Li et al., J Immunol. 2012, Koyama et al., Nat Med 2012). However, the exact location and identity of host non-hematopoietic APCs triggering alloreactive T cell responses remains controversial and needs to be proven in vivo. Fibroblastic reticular cells (FRCs) have shown to provide the crucial delta-like notch ligand to alloreactive T cells (Chung et al., JCI 2017) in aGvHD, therefore we investigated the role of FRCs MHC class II in aGvHD and their potential role as non-hematopoietic APCs in MHC class II dependent manner.

In vitro cultured FRCs cell line as well as FRCs from lethally irradiated mice upregulate MHCII and co-stimulatory molecules. Moreover, FACS sorted FRCs (CD45^-CD24^-CD31^-gp38⁺) were able to process DQ-OVA via MHC class II machinery, indicating that FRCs have the potential to activate CD4⁺ T cells.

Employing allo-HCT mouse models in combination with flow cytometry and advanced microscopy techniques, we explored early alloreactive T cells activation initially in a myeloablatively conditioned MHC major mismatch allo-HCT setting (FVB/NàC57Bl/6). We generated MHCII^ΔCcl19 mice with a Ccl19-intrinsic deletion of MHC class II on all Ccl19 expressing reticular lineage cells by crossing mice with floxed H2-Ab1 gene (H2-Ab1^fl) with a mouse expressing Cre recombinase under the control of the Ccl19 promoter (Ccl19^Cre). On day+3 after allo-HCT, CD4⁺ T cells activation (CD44 and CD25 expression) and proliferation (Ki67 expression and CFSE dilution) did not differ in the MHCII^ΔCcl19 mice from H2-Ab1^fl wildtype littermates. To further elucidate FRCs MHC class II in aGvHD milieu, we utilized iFABP-tOVA transgenic model in which OVA is expressed by intestinal epithelial cells as well as ectopically by FRCs of the lymphoid organs. OT-II cells transferred from Rag^Δ background mice failed to proliferate in the mLNs of lethally irradiated iFABP-tOVA, whereas excessive proliferation was observed in CD11c.DOG mice (where OVA is presented by CD11c-expressing cells). Taken together these results indicate that MHCII on FRCs does not play a role in direct antigen presentation and CD4⁺ T cell activation.

Next, we asked whether MHCII on FRCs influences alloreactivity of CD4⁺ T cells in the symptomatic phase of aGvHD. Indeed, in MHCII^ΔCcl19 mice, CD4⁺ T cells expressed higher levels of effector molecules: CD44 and CD127 as well as the proliferation marker Ki67 on day +30 of allo-HCT. Furthermore, the proportion of donor CD90.1⁺CD4⁺FoxP3⁺ regulatory T cells (Tregs) were reduced in MHCII^ΔCcl19 mice as compared to H2-Ab1^fl wild-type littermates. This led to overall poor survival of MHCII^ΔCcl19 mice by day+60 after allo-HCT. At this time point in MHCII^ΔCcl19 mice CD4⁺ T cells displayed higher levels of CD44, CD127 and Ki67 and down-regulated PD-1 and Lag3.

To further elucidate the effect of FRCs MHC class II on CD4⁺FoxP3⁺ donor Tregs, we transplanted CD90.1⁺CD4⁺CD25^hi Tregs, TCD BM from FVB mice along with naïve luc⁺ CD90.1⁺CD4⁺ T cells from FVB.L2G85 mice. Tregs protected against aGvHD in H2-Ab1^fl littermate controls whereas Tregs could not protect MHCII^ΔCcl19 recipients rendering them susceptible to aGvHD and to poor overall survival.

Conclusively, these results indicate for the first time that MHC class II on FRCs assists to maintain donor Tregs in the SLOs after allo-HCT. Conclusively, we propose a model in which FRCs promote T cell alloreactivity by providing notch ligands (Chung et al., JCI 2017) in the initiation phase and mitigate aGvHD by maintenance of Tregs via MHC class II in the aGvHD-effector phase.