Selective accumulation of virus-specific CD8⁺ T cells within the peripheral blood stem cell compartment

To the editor:

The absence of cellular immunity is central to the pathogenesis of herpesvirus-mediated diseases after allogeneic hemopoietic stem cell transplantation (HSCT).^1,2  For both bone marrow (BM)– and granulocyte-colony stimulating factor–mobilized peripheral blood stem cells (PBSCs) HSCT, donor-derived Epstein-Barr virus (EBV) and cytomegalovirus (CMV) peptide–specific CD8⁺ T cells clones undergo early expansion and persist long-term, with additional diversification arising from novel antigen-specific clones from donor-derived progenitors.³  Whether BM or PBSC is the superior source of antiviral CD8⁺ T cells is unclear. Given that PBSC has largely replaced BM as a source of stem cells for HSCT, it is unlikely that herpesvirus effector T-cell reconstitution will ever be compared prospectively. PBSC grafts contain 10 to 30 times more T cells than BM⁴  and a randomized study found proven viral infections were more frequent in BM than PBSC recipients,⁵  suggesting viral-specific T-cell immunity is enhanced in PBSC. Recently Moss showed in lung cancer patients that herpesvirus-specific BM-derived CD8⁺ T cells have unique homing properties relative to herpesvirus-specific CD8⁺ T cells present in unmobilized peripheral blood (PB).⁶  Immunodominant EBV-lytic peptide–specific CD8⁺ T cells were enriched in BM but were reduced for CMV peptide–specific CD8⁺ T cells relative to PB. EBV-latent peptide–specific CD8⁺ T cells were equivalent, which has relevance in the context of posttransplantation lymphoproliferative disorder for which impaired EBV-latent CD8⁺ T-cell immunity is a risk-factor.⁷  A comparison of herpesvirus-specific cellular immunity in PBSC versus PB has yet to be performed.

We assayed the PBSC of 16 patients and the PB of 26 age-matched healthy volunteers. Although PBSC was obtained in patients and compared with PB of healthy subjects, our previous data indicate that CMV/EBV effector T-cell immunity is not impaired in these patients.⁸  The study had ethics approval and was conducted in accordance with the Declaration of Helsinki. In line with Moss in BM versus PB,⁶  CD8⁺ T cells were significantly increased and naive CD8⁺ T cells were reduced in PBSC relative to PB. Otherwise, CD8⁺ T-cell subsets in PBSC relative to PB were strikingly different compared with BM (Table 1). Undifferentiated CD8⁺ T cells (coexpressing the costimulatory molecules CD27 and CD28) were reduced. In addition, CD8⁺ T cells expressing the lymphoid tissue homing molecule CD62L were lower which is consistent with recent mobilization of T cells from BM into the circulation. In keeping with this, we observed elevation of the chemokine receptors CXCR3 and CCR5 (markers of migratory effector T cells), but not CXCR6 (which facilitates egress into the lung, liver, and joints⁹ ), whereas in BM there is enrichment for CXCR3⁻CCR5⁺CXCR6⁺CD8⁺ T cells. As expected, the normal hierarchy of EBV lytic greater than latent T-cell responses seen in PB (P = .01) was preserved in PBSC (P = .03). Critically and in contrast to BM,⁶  in PBSC EBV-latent peptide–specific interferon-γ (IFN-γ) secreting CD3⁺CD8⁺ T cells were 20-fold lower than in PB (perhaps reflecting the lower EBV viral load in PBSC), whereas EBV-lytic and CMV peptide–specific IFN-γ CD8⁺ T cells were equivalent to PB. These results imply that within PBSC there is selective recruitment of CD8⁺ T-cell populations with a distinct functional and homing phenotype. In contrast to BM, EBV-latent but not CMV peptide–specific CD8⁺ T-cell immunity is impaired in PBSC relative to PB. The data have implications for HSCT and adoptive immunotherapy.