CMV-specific cellular therapy for acute myeloid leukemia?

To the editor:

We read with interest the recent paper by Elmaagacli et al documenting a striking association between early CMV replication after allogeneic transplantation and a subsequent reduction in relapse risk in patients with acute myeloid leukemia (AML).¹  While the precise mechanism(s) underpinning the correlation remain obscure, the authors hypothesize that CMV-specific T cells with cross-reactive T-cell receptors, which were stimulated to expand by viral antigenemia, might be responsible. They further suggest that “controlled” induction of viral replication might be a future strategy to reduce relapse rates in these patients.

We use T cell–depleted conditioning regimens in patients with AML: an ablative regimen (120mg/kg cyclophosphamide, 14.4Gy TBI, plus 90mg/m² fludarabine with unrelated donors) along with in vitro alemtuzumab (20 mg mixed with the graft) in younger patients without significant comorbidities; a reduced intensity regimen (RIC; 150mg/m² fludarabine, 140mg/m² melphalan) with in vivo alemtuzumab in older patients or those with significant comorbidities. While alemtuzumab-containing regimens are associated with very high levels of CMV replication, the delayed T-cell reconstitution that characterizes their use would potentially eliminate the beneficial effects predicted by Elmaagacli et al, in much the same way as they suggest after CD34-selected grafts.¹  However, during this period we have also performed a series of cellular immunotherapy studies, infusing CMV-specific T cell lines or directly-selected CMV-specific T cells.^2-4  These have generally been infused early after transplantation, at which time the profound lymphopenia combines with viral antigen exposure to reproducibly induce massive expansions of CMV-specific T cells. This setting would therefore provide perhaps an ideal one in which to test their hypothesis on the protective effect of CMV-specific T cells. One hundred patients received transplants between January 2001 and January 2011, and survived more than 50 days after transplant without relapse. None had refractory disease. All gave written informed consent for the transplant and for collection of the data presented here.

The cumulative incidence of relapse did not differ significantly according to donor source or conditioning intensity (Figure 1A-B). Disease status had the greatest impact (probably explaining the lack of positive impact of unrelated donors because transplants performed in CR1 more commonly used sibling donors), although not reaching statistical significance (Figure 1C). CMV replication was detected by quantitative PCR in 4/7 intermediate-risk patients (seronegative recipient/seropositive donor), 49/51 high-risk patients (seropositive recipient) and 0/42 low-risk patients (seronegative recipient/donor). The group with viral replication (PCR⁺) were well matched with those without (PCR⁻); no statistically significant differences in terms of age (median 45 [17-67] years versus 38 [15-66] years), conditioning (18/53 versus 15/47 RIC), donor source (27/53 versus 30/47 unrelated), or status at transplant (33/53 versus 29/47 CR1). The 5-year cumulative incidence of relapse was 22% in PCR⁺ versus 28% in PCR⁻ patients (P = .4658, Figure 1D).

Nineteen patients received CMV-specific adoptive cellular therapy a median of 32 days after transplant. There were no significant differences in age, conditioning, or number in CR1 (12/19 versus 50/81), although this group included a higher number with related donors because of the inclusion criteria of the cellular immunotherapy studies (12/19 versus 31/81, P = .0707). We previously demonstrated the rapid kinetics of recovery of durable CMV-specific immunity in these patients. The 5-year cumulative incidence of relapse was 25%, versus 25% in those not receiving cellular immunotherapy (P = .7827, Figure 1E). While we appreciate that this is a retrospective study and that the number of patients receiving cellular immunotherapy is relatively low, we find no evidence for a protective effect of CMV-specific T cells. It is perhaps unlikely that a trial will be initiated to address this question directly, although 2 prospective national randomized studies of CMV cellular immunotherapy in the United Kingdom may provide further insights if sufficient patients with AML are enrolled.