Sequential anti-cytomegalovirus response monitoring may allow prediction of cytomegalovirus reactivation after allogeneic stem cell transplantation

Abstract

Background: Reconstitution of cytomegalovirus-specific CD3(+)CD8(+) T cells (CMV-CTLs) after allogeneic hematopoietic stem cell transplantation (HSCT) is necessary to bring cytomegalovirus (CMV) reactivation under control. However, the parameters determining protective CMV-CTL reconstitution remain unclear to date.

Design and methods: In a prospective tri-center study, CMV-CTL reconstitution was analyzed in the peripheral blood from 278 patients during the year following HSCT using 7 commercially available tetrameric HLA-CMV epitope complexes. All patients included could be monitored with at least CMV-specific tetramer.

Results: CMV-CTL reconstitution was detected in 198 patients (71%) after allogeneic HSCT. Most importantly, reconstitution with 1 CMV-CTL per µl blood between day +50 and day +75 post-HSCT discriminated between patients with and without CMV reactivation in the R+/D+ patient group, independent of the CMV-epitope recognized. In addition, CMV-CTLs expanded more daramtaically in patients experiencing only one CMV-reactivation than those without or those with multiple CMV reactivations. Monitoring using at least 2 tetramers was possible in 63% (n = 176) of the patients. The combinations of particular HLA molecules influenced the numbers of CMV-CTLs detected. The highest CMV-CTL count obtained for an individual tetramer also changed over time in 11% of these patients (n = 19) resulting in higher levels of HLA-B*0801 (IE-1) recognizing CMV-CTLs in 14 patients.

Conclusions: Our results indicate that 1 CMV-CTL per µl blood between day +50 to +75 marks the beginning of an immune response against CMV in the R+/D+ group. Detection of CMV-CTL expansion thereafter indicates successful resolution of the CMV reactivation. Thus, sequential monitoring of CMV-CTL reconstitution can be used to predict patients at risk for recurrent CMV reactivation.

Figure 1. Gating strategy for detection of tetramer-stained CMV-CTLs.

The gating hierarchy and tetramer positive cells in a patient expressing 3 of the HLA molecules represented in our tetramer set are shown. The lymphocytes were gated in the FS/SS quadrant (upper left) followed by the selection of CD3⁺CD8⁺-positive cells (lower left). Within the CD3⁺CD8⁺-positive population, the percentage of cells that bind each tetramer is determined in individual samples (200 µl whole blood each). Staining results using the negative control tetramer, HLA-A*0101 CMV tetramer, HLA-B*0702 CMV tetramer and HLA-B*0801 CMV tetramer are shown. The number of CMV-CTLs per µl blood is indicated in each tetramer plot.

Figure 2. CMV-CTL levels vary depending on the HLA-epitope tetramer used for detection and the occurrence of CMV-reactivation.

(A) Median circulating CMV-CTL levels in our patient cohort are shown for six different HLA-restricted CMV epitopes monitored by commercially available HLA-epitope tetramers. Data are shown for all patients with a detectable CD3+/CD8+ T cell response (CMV-CTLs >0 per µl blood) during the first year following HSCT, and summarized as mean values of CMV-CTLs calculated for each patient (filled squares). P-values were calculated using the Kruskal-Wallis test followed by Dunn's multiple comparison. Patients not mounting a response (0 CMV-CTL per µl blood) or with cell counts below the detection limit are denoted below the dashed line. (B) Median CMV-CTL levels for each HLA-restricted CMV epitope are shown for each patient who experienced (empty squares, CMV-R) or did not experience (filled squares, no CMV-R) CMV reactivations. The median CMV-CTL levels from the entire cohort are shown in the box above each group. No results are shown for the HLA-A*1101 tetramer, since only one patient was monitored. * p≤0.05, ** p≤0.01, *** p≤0.001.

Figure 3. Specific HLA combinations influence the CMV-CTL levels.

(A) Median CMV-CTL levels in patients expressing only and monitored with HLA-A*0201 (filled circles) or HLA-B*0702 (filled squares) tetramers, respectively, were calculated and compared to the levels in patients expressing both alleles. The level of HLA-A0201 CMV-CTL (white circle) was significantly lower in patients expressing HLA-A0201 and –B0702 compared to those only expressing the A0201 allele (filled circles). No such difference was found for HLA-B0702-binding CTL levels between patients who only express the B0702 allele (filled square) or both A0201 and B0702 alleles (white squares). (B) The median numbers of CMV-CTLs detected by each tetramer were compared in all patients monitored with at least 2 different tetramers (n = 176/278). The analysis for one patient is shown as an example. The number of CMV-CTLs detected by each tetramer are plotted over time. CMV reactivation occurred on day +41 (spike in the pp65 line), triggering CMV-CTL expansion. HLA-B*0801 detected the highest numbers of CMV-CTLs until day +100 when a switch occurred. * p≤0.05.

Figure 4. CMV-CTL monitoring enables early discrimination between R+/D+ patients in whom CMV reactivation did or did not occur.

(A) A threshold of 1 CMV-CTL per µl blood was applied to discriminate between patients with and without CMV reactivation in the R+/D+ group. P-values from log-rank testing are indicated for days +50 (*) and +75 (+). (B) CMV-CTL reconstitution after HSCT is shown for one patient in the R+/D+ group. CMV-CTL numbers per µl blood (left y-axis, empty squares) were plotted against the time in days after HSCT. The right y-axis (filled circles) shows the number of pp65-positive cells/400,000 leukocytes as the means of detecting CMV reactivation. Positive PCR results for CMV are indicated at the top of each graph (half filled triangle). CMV reactivation occurred on day +39 (spike in filled circle line) in this patient, after which CMV-CTL expansion occurred (rise in empty square line). (C) Reconstitution of CMV-CTLs within the R+D− group using a threshold of 1 CMV-CTL per µl blood. Until day +100 no significant reconstitution occurred. (D) Prolonged CMV immune reconstitution (empty squares) and necessary prolonged antiviral therapy (striped bar) in one patient from the R+/D− group.

Figure 5. Kinetics of CMV-CTL expansion after CMV reactivation.

(A) CMV-CTL levels around day +30 (the interval from day +15 to +45 was assessed, filled symbols) and day +60 (the interval from day +46 to +99 was assessed, filled symbols) in patients who experienced (circles) or did not experience (triangles) a CMV reactivation prior to day +100 in the R+/D+ group. Significant differences between groups were assessed by t-test with Welch's correction, and included the 34 data points not depicted because they lie outside axis limits. (B) Patients in the R+/D+ group who experienced no (triangles), a single (circles) or multiple (diamonds) CMV reactivations before day +100 were compared (t-test with Welch's correction) using the difference between the slopes of the lines created by the CMV-CTL levels during the interval from day +15 to +45 and the interval from day +46 to +99. The dotted line indicates no change in CMV-CTL level between the measurements in both intervals. ** p≤0.01.