The transfer of adaptive immunity to CMV during hematopoietic stem cell transplantation is dependent on the specificity and phenotype of CMV-specific T cells in the donor

Abstract

The successful reconstitution of adaptive immunity to human cytomegalovirus (CMV) in hematopoietic stem cell transplantation (HSCT) recipients is central to the reduction of viral reactivation-related morbidity and mortality. Here, we characterized the magnitude, specificity, phenotype, function, and clonotypic composition of CMV-specific T-cell responses in 18 donor-recipient pairs both before and after HSCT. The principal findings were: (1) the specificity of CMV-specific T-cell responses in the recipient after HSCT mirrors that in the donor; (2) the maintenance of these targeting patterns reflects the transfer of epitope-specific T-cell clonotypes from donor to recipient; (3) less differentiated CD27(+)CD57(-) CMV-specific memory T cells are more likely to persist in the recipient after HSCT compared with more terminally differentiated CD27(-) CD57(+) CMV-specific memory T cells; (4) the presence of greater numbers of less differentiated CD8(+) CMV-specific T cells in the donor appears to confer protection against viral reactivation in the recipient after HSCT; and (5) CMV-specific T cells acquire a more differentiated phenotype and a restricted functional profile after HSCT. Overall, these findings define the immunologic factors that influence the successful adoptive transfer of antigen-specific T-cell immunity during HSCT, which enables the identification of recipients at particular risk of CMV reactivation after HSCT.

Figure 1

Total CMV-specific CD4⁺ and CD8⁺ T-cell responses in donors, recipients before HSCT, and recipients after HSCT. CD4⁺ (top panel) and CD8⁺ (bottom panel) T-cell responses are shown in donors (green), recipients before HSCT (red), and recipients after HSCT (yellow). After stimulation for 6 hours, the frequency of CMV-specific memory T cells was calculated using all possible Boolean combinations (where at least a single response was present) of cytokines (MIP-1β, IL-2, TNF-α, and IFN-γ) and CD107a expression for each ORF, and the total frequency was determined by the summation of responses to the 19 ORFs for both CD4⁺ and CD8⁺ subsets (after background subtraction). *Patients who received a DLI.

Figure 2

Contribution of ORF-specific T-cell response to the total CMV-specific T-cell response in the donor and recipient after HSCT. (A) Targeting patterns of CMV-specific memory CD4⁺ (left panels) and CD8⁺ (right panels) T-cell responses in representative donor-recipient pairs. The summation of responses measured by 5 functional outputs is shown for each ORF (CD4⁺ and CD8⁺) tested and for whole CMV lysate (CD4⁺ only); donor (green), recipient before HSCT (red), and recipient after HSCT (yellow). Response breadth (ie, the number of targeted ORFs) and dominance patterns (ie, the ORF-specific responses representing the majority of the total response) showed substantial variability. In some cases, pp65 and IE-1 were preferentially targeted (eg, ID 1 and ID 3 CD8⁺ T-cell responses); in other cases, the predominant responses were directed against different ORFs (eg, ID 7 CD8⁺ T-cell responses). In some cases, the CD4⁺ T-cell responses to whole CMV lysate were less than the summation of the ORFs, which may reflect the negative effect of freeze/thawing in lysate processing for peptide recognition. *Patients who received a DLI. (B) Contribution of each ORF-specific response to the total CD4⁺ (left panel) and CD8⁺ (right panel) CMV-specific T-cell response plotted for donors against recipients after HSCT. In each case, the ORF-specific response was normalized such that the contribution of each targeted ORF (in percentage) to the total CMV-specific T-cell response was calculated. Only the normalized data are shown.

Figure 3

Phenotypic analysis of donor CMV ORF-specific memory T-cell responses in relation to CMV reactivations and magnitude of the corresponding responses in the recipient after HSCT. (A) For donor ORF-specific T-cell populations that were less differentiated (CD27⁺CD45RO⁺CD57⁻), there was a greater probability of persistence and expansion (increased fold change) in the recipient after HSCT. For donor ORF-specific T-cell populations that were more differentiated (CD27⁻ memory; CD27⁻CD57⁺ memory), there was a greater probability of contraction (decreased fold change) in the recipient after HSCT. (B) In donors with a higher frequency of CD8⁺ CMV-specific CD27⁺ memory T-cell responses, fewer CMV reactivations were observed in the recipient after HSCT. Conversely, in donors with a higher frequency of more differentiated CD8⁺ CMV-specific memory T-cell responses (CD27⁻), a greater number of CMV reactivations were observed in the recipient after HSCT. No correlations between the phenotype of donor CMV-specific CD4⁺ T cells and the incidence of CMV reactivation in the recipient after HSCT were observed. The patients who reactivated CMV are shown in Table 1.

Figure 4

Changes in phenotype and cytokine production of CMV ORF-specific T cells in the donor and recipient after HSCT. (A) Phenotype of CMV ORF-specific responses in donors and recipients after HSCT. Matched ORF-specific responses to CMV acquired a more differentiated phenotype after HSCT, with significant increases in both CD27⁻ memory and CD57⁺ memory T-cell populations in both the CD4⁺ (left panels) and CD8⁺ (right panels) subsets; similarly, significant decreases were observed in the corresponding CD27⁺ memory T-cell populations. (B) Analysis of the contribution of each cytokine and CD107a to the increment in CMV-specific T-cell responses after HSCT is shown in the upper panels. Each of the 5 independent functions was analyzed for each ORF-specific T-cell response after HSCT and compared with the corresponding response in the donor. For CD4⁺ T-cell responses, IFN-γ secretion increased in the recipient after HSCT (P = .024); for CD8⁺ T-cell responses, an increment in CD107a mobilization and the production of IFN-γ, MIP-1β, and TNF-α was observed in the recipient after HSCT. When T-cell responses were categorized on the basis of the number of functions elicited on CMV antigen encounter (bottom panels), bifunctional responses increased within both the CD4⁺ and CD8⁺ subsets after HSCT (P = .017 and P = .003, respectively). The median is depicted in red. Only frequencies ≥ 0.01 are shown for graphical representation. D indicates donor; and R, recipient after HSCT. *P < .05, **P < .01 (Mann-Whitney U test).