SEARCH

SEARCH BY CITATION

Keywords:

  • haematopoietic stem cell transplantation;
  • children;
  • cytomegalovirus;
  • immune recovery;
  • relapse risk

Summary

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Funding
  8. Authorship contributions
  9. Conflicts of interest
  10. References
  11. Supporting Information

The interplay between immune recovery, cytomegalovirus (CMV)-reactivation, CMV-driven immunity and graft-versus-leukaemia effect (GVL) was analysed in 108 children (median age: 8 years) who underwent haematopoietic-stem cell transplantation (HSCT) for acute leukaemia. Follow-up was 2 years unless death or relapse occurred. CMV-polymerase chain reaction (PCR) was programmed weekly until month +3 post-HSCT. Immunomonitoring consisted of sequential lymphocyte subset enumerations and analyses of T-cell proliferative and γ-interferon responses to CMV and to adenovirus. In the 108 recipients, the 2-year relapse rate (RR) was 25% (median time to onset 4·5 months; range: 24 d–17 months). CMV reactivation occurrence was 31% (median time to onset 26 d). Donor/recipient CMV serostatus did not influence RR. Among the 89 recipients disease-free after day +120, i) early CMV-reactivation before day +30 was more frequent (P = 0·01) in the relapse recipient group opposed to the non-relapse group. ii) CD8+/CD28 and CD4+CD45RA T-cell expansions induced by CMV did not influence RR, iii) Recovery of anti-CMV and also anti-adenovirus immunity and of naïve CD4+ T-cells was faster in the non-relapse group (P = 0·008; 0·009 and 0·002 respectively). In contrast to adult acute myeloid leukaemia, CMV reactivation was associated with increased RR in this paediatric series. Accelerated overall immune recovery rather than CMV-driven immunity had a favourable impact on RR.

Allogeneic haematopoietic stem-cell transplantation (HSCT) is a potentially curative therapy for patients classified at high risk of leukaemia relapse. Myeloablative HSCT results in leukaemia eradication not only by the direct cytotoxic effect of the intensive conditioning regimen but also through the immune recognition of malignant cells by donor lymphocytes, referred to as the Graft-Versus-Leukaemia (GVL) effect (Horowitz et al, 1990).

Although significant strides have been made in reducing overall transplant-related mortality over the past 2 decades, relapse of disease still remains the most common cause of failure in patients undergoing HSCT (Gooley et al, 2010).

Many attempts have been made to enhance GVL by manipulating the donor transplant or by adding infusion of selected donor cells after transplantation (Rager & Porter, 2011). However, further elucidation of the interplay between immune recovery after HSCT and GVL remains necessary to broaden our understanding of leukaemic cell control as a guide for these new therapeutic approaches and to identify biomarkers of prognostic value for their indication. The primary objective of this study aimed to refine the immune parameters associated with the risk of relapse in paediatric HSCT.

An effect of cytomegalovirus (CMV) infection on the relapse rate (RR) was first suggested by Lonnqvist et al (1986). Several groups subsequently made similar observations in adults (Elmaagacli et al, 2011; Green et al, 2013; Ito et al, 2013). Furthermore, an association between donor or recipient CMV serostatus and decreased risk of relapse was reported in some paediatric settings (Behrendt et al, 2009). As far as we know, the potential protective effect of CMV reactivation has not been analysed in children so far. Furthermore, no study selectively investigated the impact of CMV-driven immunity on the risk of relapse. Here, we therefore evaluated immune parameters associated with RR in the context of early CMV reactivation and of immunity triggered by CMV in a paediatric series.

All consecutive children ongoing alloHSCT for high risk of leukaemia relapse were enrolled at the time of transplantation from April 2006 to December 2010 in a single centre to analyse the impact on relapse (i) of the CMV serostatus of any donor-recipient pairs (ii) of early CMV reactivation (from HSCT to month +3) in the recipients after HSCT and (iii) of both overall immune recovery and of CMV-driven immune recovery during the first 6 months post-HSCT.

Materials and methods

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Funding
  8. Authorship contributions
  9. Conflicts of interest
  10. References
  11. Supporting Information

Study population

A total of 124 consecutive patients who underwent alloHSCT after a myeloablative regimen between April 2006 and December 2010 for acute leukaemia in our centre were initially considered. The medical history of patients was prospectively collected and retrospectively analysed. The two parents or guardians of these paediatric patients gave written informed consent to all aspects of the transplantation procedure and to the use of medical records for research. Inclusion criteria were primary acute myeloid leukaemia (AML), acute lymphoid leukaemia (ALL) or biphenotypic acute leukaemia. Patients who had undergone previous alloHSCT (n = 4) and those who received haploidentical HSCT from a related donor (n = 3) were excluded for homogeneity purpose. In total, 117 patients were enrolled.

This study was approved by the local ethic committee of Robert Debré hospital, Paris-France.

Transplant procedures

The myeloablative preparative regimens in ALL patients consisted of 120 cGy hexafractionated total body irradiation (TBI) from a cobalt-60 (60Co) source and either cyclophosphamide (60 mg/kg/d for 2 d) or etoposide 60 mg/kg/d for 1 d. Myeloablative preparative regimens did not include TBI in children <2 years old and/or in children with AML. In these situations, IV Busulfan (total-dose, 16 mg/kg with patient body weight adaptation according to published data [de Berranger et al, 2014]) was combined with cyclophosphamide (200 mg/kg total dose) in most instances.

Acute graft-versus-host disease (aGVHD) prophylaxis consisted of blood level-adjusted intravenous ciclosporin A (CSA) alone in patients transplanted for ALL from a sibling donor. Patients transplanted for AML from a sibling donor received CSA and short-course methotrexate. Patients transplanted from an unrelated donor received CSA, short-course methotrexate and anti-thymocyte globulins (ATG). Finally, patients transplanted with cord-blood received CSA, methylprednisolone and ATG. Diagnosis and grading of GVHD was performed according to standard procedures (de Masson et al, 2012). First-line therapy of grades II-IV aGVHD consisted of intravenous methylprednisolone (2 mg/kg bodyweight) over 5 to 7 d, which was gradually tapered at 3-d intervals in responding patients thereafter. In patients with steroid-refractory aGVHD, additional immunosuppressive drugs were added according to the discretion of the treating physician.

Early CMV reactivation diagnosis and preemptive treatment

Weekly monitoring for CMV viraemia was performed using a qualitative polymerase-chain reaction (PCR) and, when positive, was quantified using a quantitative PCR followed by initiation of pre-emptive therapy in cases of viral load ≥1000 DNA copies/ml according to a standardized protocol (Matthes-Martin et al, 2012).

Treatment consisted of 2 weeks induction therapy using ganciclovir (5 mg/kg intravenously, twice daily), followed by 2 weeks of maintenance therapy with the same drug given once daily, 5 d a week; foscarnet was used instead in 10 patients (90 mg/kg twice daily for 14 d). CMV prophylaxis was not used. Standard HSCT procedures included acyclovir (paediatric dose 250 mg/m2 IV every 12 h) from day −2 as prophylaxis against varicella zoster and herpes simplex viruses.

Immunological investigations

Lymphocyte subsets analysis

Equipment used for data acquisition and analysis was supplied by Beckman-Coulter (FC500, Villepinte, France). Absolute values of lymphocyte subsets (T, TCD4, TCD8, B and natural killer [NK]) were determined using commercialized kits (CD45 tetra chrome kit, Beckman-Coulter) according to manufacturer recommendations. In addition, direct immunofluorescence staining was performed on whole blood using a combination of directly labelled monoclonal antibodies (Mab). CD4-PeCy5/CD45RA-PE/CD62L-FITC, CD3-FITC/CD8-PeCy5/CD28-PE, CD3-ECD/CD4-PeCy7/CD62L-FITC/CD45RA-PE and CD3-ECD/CD8-PeCy7/CD62L-FITC/CD45RA-PE/CD28-PeCy5 (BD biosciences, Le Pont de Claix, France; Beckman Coulter Company, Marseille, France). In brief, 100 μl of whole blood were added to a premixed solution of Mab at the appropriate dilution. Red blood cells were lysed with immunoprep (Beckman-Coulter). The CXP-analysis software (Beckman-Coulter) was used for analysis following gating according to small lymphocyte characteristics and then to high density CD3 and CD4 or CD8 expression. Absolute counts of CD4 and CD8 subsets were obtained by multiplying the percentages of the considered subset with CD4 or CD8 absolute counts.

T-cell proliferation assays

T-cell proliferation assays were performed as previously described (Guerin et al, 2010)). In brief, peripheral blood mononuclear cells (PBMCs) were isolated by density-gradient separation with UNI-SEPmaxi+ tubes (Novamed, AbCys, Paris, France) and resuspended at a cell-density of 106/ml in RPMI 1640 medium containing glutamine (1%) and supplemented with 10% pooled human sera. Cells were plated at 105 per well in triplicate. Predetermined optimal concentrations (Guerin et al, 2010; Pedron et al, 2011) of CMV or Adenovirus (AdV) antigens were added. Unstimulated cells were used as negative controls. Cells stimulated by CD3 Mab (Janssen-Cillag, Neuss, Germany) were used as positive controls. Antigen-stimulated cells were cultured for 6 d, after which [3H]-Thymidine at 29·6 kBq/ml (GE Healthcare, Buckinghamshire, UK) was added for 18 h. Results are given as counts per minute (cpm) and stimulation index (SI, i.e., cpm in stimulated cultures divided by cpm in unstimulated cultures). Based on studies comparing results in immune (healthy adults) versus non-immune (<2-year-old HSCT donors) individuals, a SI ≥4 was considered to be a positive response in CMV or AdV-stimulated cultures.

γ-interferon (IFNγ) -secreting cells

IFNγ secreting cells were enumerated in a flow cytometry-intracellular cytokine-staining assay, as previously described (Pedron et al, 2011). PBMCs were adjusted to 106/ml in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum (Gibco, Paisley, Scotland). The culture medium also contained 10 iu/ml recombinant interleukin 2 (Roche Diagnostics, Meylan, France), 2 μg/ml of CD28 Mab and 2 μg/ml of CD49d Mab (from Becton-Dickinson or Beckman-Coulter). The secretion inhibitor Brefeldin A (10 μg/ml) (Sigma, Saint-Quentin-Fallavier, France) was incorporated 1 h later. Unstimulated cells were used as negative controls. Stimulation by 0·2 μg/ml of CD3 Mab (Janssen-Cillag) was used as a positive control. Titrated amounts of CMV or AdV antigen were used for CMV- and AdV-specific stimulations. Cultures were incubated at 37°C in a humidified 5% CO2 atmosphere for 18 h. After fixation and permeabilization (IntraStain Kit, Dako, Cytomation, Glostrup, Denmark), CD3-FITC (Beckman-Coulter), CD4-PerCP, and IFNγ-PE (BD Biosciences) were incubated for 20 min at room temperature.

After washing, cells were analysed on FC500 or Navios instrument (Beckman-Coulter) using CXP analysis software (Beckman-Coulter). Files were gated on small CD3+/CD4+ lymphocytes.

The percentage of specific cells secreting IFNγ was calculated as the percentage of cells secreting IFNγ in stimulated cultures minus the percentage of unstimulated cells secreting the cytokine (background). The absolute counts of IFNγ-secreting CD4 cells were determined by multiplying the percentage with the CD4 counts obtained using CD3/CD4/CD8 (CD45 tetra chrome kit, according to the manufacturer's recommendations).

Statistics and methods

Patient characteristics were compared using Wilcoxon rank-sum tests and Fisher's exact tests.

All time-to-event outcomes were counted from the date of transplant to the date of event. Death was considered as a competing risk in analyses of GVHD and CMV reactivation. Relapse and Non-Related Mortality (NRM) were considered to be mutually competing risks.

Two-year relapse-free survival (RFS) was evaluated at month +24 post-HSCT. For competing risks analyses, cumulative incidence functions were estimated using usual methodology (Kalbfleisch & Prentice,1980).

Factors associated with relapse were analysed using Cox proportional hazards models (Prentice et al, 1978). The proportional hazards assumption was checked by examination of Schoenfeld residuals and Grambsch and Therneau's lack-of-fit test (Grambsch & Therneau, 1994). Chronic GVHD (cGVHD) was considered as a time-dependent covariate in the models.

All tests were two-sided and P < 0·05 were considered to indicate a statistically significant association. Analyses were performed using the r statistical software version 2.15.0 (R Development Core Team, 2011).

Results

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Funding
  8. Authorship contributions
  9. Conflicts of interest
  10. References
  11. Supporting Information

Patient characteristics and HSCT outcome

Nine of the 117 recipients were excluded from analysis because of early treatment failure unrelated to relapse; six failed to engraft, three had early lethal infections. The main characteristics of the 108 remaining recipients are shown in Table 1.

Table 1. Patient and graft characteristics of the 108 recipients
CharacteristicsAll patientsCMV serostatus
D+/−/R+D+/RD/R
  1. CMV, cytomegalovirus; D, donor; R, recipient; ALL, Acute Lymphoid Leukaemia; AML, Acute Myeloid or biphenotypic Leukaemia (i.e. lymphoid and myeloid leukaemia); BM, Bone Marrow; CBU, Cord Blood Unit; PB, Peripheral Blood.

Patients (n)108581139
Age, median (range), years8 (0–18)9 (0–18)8 (1–13)7 (1–16)
Male gender, n (%)61 (56)36 (62)6 (55)19 (49)
Female donor to male recipient, n (%)19 (18)11 (19)3 (27)5 (13)
Leukaemia characteristics
ALL/AML, n/n63/4534/248/321/18
Advanced disease stage, n (%)22 (20)12 (21)2 (18)8 (21)
Very high risk, n (%)64 (61)31 (54)8 (80)25 (66)
Donor, n (%)
Genoidentical44 (41)19 (33)6 (55)19 (49)
Unrelated64 (59)39 (67)5 (45)20 (51)
Stem cell source, n (%)
BM80 (74)43 (74)10 (91)27 (69)
CBU17 (16)8 (14)0 (0)9 (23)
PB11 (10)7 (12)1 (9)3 (8)
Conditioning regimen
Anti-thymoglobulin, n (%)41 (38)28 (48)3 (27)10 (26)
Total body irradiation, n (%)61 (56)35 (60)5 (45)21 (54)

All of the 108 eligible patients were in complete remission (CR) at the time of transplantation. All received a conventional myeloablative conditioning regimen. The graft source was bone marrow (BM) in the majority of patients (74%). Most donors (84%) were human leucocyte antigen (HLA)-matched (6/6) siblings (41%) or HLA-matched (≥9/10) unrelated donors (43%). Cord Blood Units (CBU) were ≥4/6 HLA-matched (16%). 74/108 subjects were relapse-free at +2 years and 1/108 was lost to follow-up at 14 months post-HSCT. The distribution of Donor/Recipient (D/R) CMV serostatus was 24 D+/R+, 34 D/R+, 39 D/R and 11 D+/R. Of note, one recipient, who was initially serologically classified as D/R, showed early viraemia (at day 2 post-alloHSCT). Given that this recipient had a positive proliferative response to CMV before alloHSCT (cpm; 17·386; index: 33), CMV serology was considered to be a false negative and the patient was definitely included in the R+ group. Cumulative incidence of aGVHD ≥ grade 2 was 68%. Among such cases, median time to onset was 18 d (range: 8–102). The incidence of cGVHD was 20% (limited 8%; extensive: 12%). Thirty-four patients had positive CMV-DNAemia during follow-up, with a cumulative incidence of 31% at 120 d. Median time to onset was 26 d (range: 1–73). R+ serostatus was a major factor contributing to the occurrence of CMV-DNAemia (P < 0·0001) whether the CMV serostatus of the donor was D+ or D (11/24 D+/R+ and 22/34 D/R+ had CMV-DNAemia positivity). In all cases, the first CMV-PCR positive result occurred in relapse-free patients; at this time, patients received CSA (n = 33) combined with methotrexate (n = 19) and corticosteroid (1 mg/kg, n = 6) as GVHD prophylaxis. In this study using preemptive treatment there was no CMV disease and infection resolved with treatment in all patients but one, in whom CMV-PCR positivity persisted until death from pseudomonas aeruginosa septicaemia with no evidence of CMV disease. Among the 108 recipients, a haematological leukaemic relapse was documented in 27, at a median of 6 months (range: 24 d to 17 months) after transplantation, corresponding to a 25% post-transplant leukaemic relapse incidence at 2 years.

In addition, one female recipient developed leukaemia of donor origin (carrying the Y chromosome from the male donor) more than 3 years after HSCT.

The main cause of death in non-relapse recipients within the first 2 years post-HSCT was severe infections (4%) associated in most instances with GVHD ≥ grade 2.

CMV serostatus was not associated with the risk of relapse in the 108 patients

Factors associated with RR in the 108 recipients were first tested in a univariate analysis. Results are shown in Table 2. RR was similar in patients with advanced disease i.e. presenting with AML in at least second complete remission (CR2) or ALL in at least third complete remission (CR3) and in patients with AML in first complete remission (CR1) or ALL in CR2 or less. Similar relapse incidences were also observed after HSCT from HLA-identical related donors, from ≥9/10 HLA-identical donors or from ≥4/6 HLA-identical CBU. In contrast, very high risk (VHR) status i.e. central nervous system (CNS) involvement at initial diagnosis and/or leucocyte counts >100 × 109/l at initial diagnosis and/or ALL with either KMT2A/MLL-involving translocations, RUNX1 overexpression, t(9;22) poor responders, t(6;9) at initial diagnosis or minimal residual disease positivity at transplant and/or AML with either WT1 overexpression, BCR/ABL1 transcript, t(9;22), t(9;11), KMT2A- or RUNX1-involving translocations, FLT3-ITD or a complex caryotype (≥3 abnormalities) at initial diagnosis was clearly associated with higher RR. Finally, cGVHD was associated with a GVL effect or, at least, less relapses.

Table 2. Factors associated with relapse in the 108 recipients
Variable n CIF (95%CI)HR (95%CI) P
  1. CIF, cumulative incidence frequency; 95%CI, 95% confidence interval; ALL, Acute Lymphoid Leukaemia; AML, Acute Myeloid or biphenotypic Leukaemia (i.e. lymphoid and myeloid leukaemia); CR, Complete Remission; HSCT, haematopoietic stem cell transplantation; MRD, 6/6 HLA-matched Related Donor; URD, ≥9/10 HLA-matched Unrelated Donor; CBU, ≥4/6 HLA-matched Cord Blood Unit; CMV, cytomegalovirus; GVHD, graft-versus-host disease.

  2. a

    No estimation of the cumulative incidence for time-dependent variables.

  3. Bold values indicate significant P values (P < 0·05).

Age (years)
<21136% (10–64)1 
2–96326% (16–37)0·59 (0·20–1·75)0·34
≥103418% (7–33)0·48 (0·14–1·63)0·24
Disease stage
ALL CR1-2, AML CR18625% (16–34)1 
ALL CR >2, AML CR ≥22223% (8–42)0·95 (0·36–2·51)0·92
Very high risk
No417% (2–18)1 
Yes6437% (25–48)4·55 (1·57–13·2) 0·005
HSCT group
MRD4425% (13–39)1 
URD4722% (11–35)0·81 (0·35–1·88)0·62
CBU1729% (10–52)1·17 (0·41–3·32)0·77
CMV serostatus
D+/−R+5824% (14–36)1 
D+/R1127% (6–55)0·95 (0·28–3·30)0·94
D/R3924% (12–38)0·86 (0·38–1·96)0·72
Anti-thymoglobulin
No6720% (11–30)1 
Yes4132% (18–47)1·78 (0·84–3·80)0·13
Total body irradiation
No4726% (14–39)1 
Yes6123% (13–34)0·78 (0·37–1·66)0·52
Acute GVHD 2–4
No40 a 1 
Yes68 a 0·77 (0·36–1·64)0·49
Chronic GVHD
No87 a 1 
Yes21 a 0·13 (0·017–0·95) 0·044
Overall CMV reactivation
No87 a 1 
Yes34 a 1·59 (0·58–4·28)0·31

Both univariate (Table 2) and multivariate analysis did not associate donor/recipient CMV serostatus (D/R versus D+/−/R+) with relapse (P = 0·72 and 0·91 respectively, in multivariate analysis). Note, however that an increased rather than decreased RR tended to be associated with CMV reactivation occurrence during the first 3 months post-HSCT.

Influence of CMV reactivation on overall immune recovery and immune parameters associated with relapse at 2 years in the 108 recipients

Lymphocyte repopulation following HSCT was determined by analysing absolute numbers of B, T-CD4 both naive (CD45RA+/62L+) and memory (CD45RA), T-CD8 (both CD28+ and CD28) and NK cells. The kinetics of lymphocyte subset recovery was influenced by CMV reactivation (Fig 1A). Within the first 6 months post-HSCT, CD8+ T-cell (not shown) and primarily CD28CD8+ T-cell (Fig 1) levels were higher in recipients reactivating CMV than in the non-reactivating group. Also, CD45RACD4+ T-cell and NK cell levels were higher in the recipients with CMV reactivation.

image

Figure 1. Kinetics of lymphocyte subset recovery according to CMV reactivation and relapse status in the 108 recipients. Median values of CD28 CD8+ T-cells, CD45 RA+ CD62L+ CD4+ T-cells (naïve CD4 T-cells) CD45 RA CD4+ T-cells (memory CD4 T-cells) and NK cells according to cytomegalovirus (CMV) reactivation (+) or not (−) (panel A) and to relapse (+) or not (−) (panel B) in the 108 recipients at month +1 (M1), +3 (M3) and +6 (M6) post-haematopoietic stem cell transplantation are shown. Error bars indicate standard deviation. Unpaired student's t test results are indicated when significant (P < 0·05).

Download figure to PowerPoint

Though full immune surveillance was not feasible in the recipients with early relapse, repopulation of lymphocyte subsets in the 108 recipients according to relapse is also shown in Fig 1B. An association between early recovery of naïve T cells and GVL was suggested.

Immune recovery of CMV-specific immunity according to RR in the 108 recipients

Data on proliferative responses to CMV during the first year post-HSCT were available in 99/108 recipients. 36/99 showed positive responses, with 30 (83%) developing immunity within the first 120 d post-HSCT.

Data on IFNγ-responses were available in 68/108 recipients. Although the proliferation assay and the IFNγ assay analysed distinct functions, the results from proliferative and cytokine assays were highly correlated (k = 0·81, P < 0·0001). Among the 100 patients investigated for immunity, the cumulative incidence of recipients who developed immunity to CMV within the first year post-HSCT was 42% (95% confidence interval [CI] 32–52) (range: 14 d to 8 months), of whom the great majority (37; 73%) had developed immunity at month +3 (±1). As expected, the development of CMV-specific immunity was strongly associated with the seropositivity of the recipient before HSCT (P < 0·0001). The kinetics of immune recovery of immunity to CMV is shown in Figure S1.

As relapse and death within the first 3 months post-HSCT were competing factors for detection of anti-CMV immunity, recovery of anti-CMV immunity according to RR was not analysed in the 108 recipients but in a more restricted group (see below).

Relapse incidence according to early viraemia and/or immune recovery in the 89 recipients alive and relapse-free after day 120

Considering the kinetics of immune recovery, a minimum of several months were needed after HSCT to accumulate a meaningful number of naïve CD4+ T-cells (Fig 1) and also to detect specific T-cell responses to AdV (not shown). Correlates between immune recovery and the risk of relapse were therefore only analysed in those recipients who were alive and relapse-free beyond month +3 post-HSCT (±30 d to allow for a 1-month tolerance in immunological investigations). Among the 89 recipients with these characteristics and who were evaluated for immunity during this time, leukaemia relapse was observed in 13 (15%), with a median time to HSCT of 11 months (range: 4–17 months).

Recovery of immunity to CMV at month 3 (±1) was observed in 37 (41%) recipients. CMV-DNAemia positivity was associated with the development of CMV-specific immunity in 22 out of 37. Fifteen recipients developed immunity to CMV while the virus remained undetectable in the blood, an observation that is commonly interpreted as occult reactivation. Finally, five recipients had CMV-DNAemia positivity but no immunity to the virus, which was interpreted as delayed immune recovery.

Our results confirmed that repopulation of naïve CD4 T-cells is accelerated in the subgroup without relapse (P = 0·04) (Fig 2B). Higher levels of NK cells (Fig 2) and of proliferative responses to CMV (P = 0·008), but also to AdV, were also observed in recipients without relapse compared to recipients with relapse (Fig 3).

image

Figure 2. Kinetics of lymphocyte subset recovery in the 89 recipients alive and relapse-free after day +120 according to CMV reactivation and relapse status. Median values of CD28 CD8+ T-cells, CD45 RA+ CD62L+ CD4+ T-cells (naïve CD4 T-cells), CD45 RA CD4+ T-cells (Memory CD4 T-cells) and NK cells according to cytomegalovirus (CMV) reactivation (+) or not (−) (panel A) and to relapse (+) or not (−) (panel B) in the 89 recipients alive and relapse-free after day +120, at month +1 (M1), +3 (M3) and +6 (M6) post-haematopoietic stem cell transplantation are shown. Error bars indicate standard deviation. Unpaired student's t test results are indicated when significant (P < 0·05).

Download figure to PowerPoint

image

Figure 3. Kinetics of T-cell proliferative responses to CMV and AdV according to relapse in the 89 recipients. Results are expressed as mean stimulation index (SI) at months +1 (M1), +3 (±1; M3) and +6 (M6) according to the relapse status at 2 years in recipients alive without relapse beyond day +120. Error bars indicate standard deviation. Unpaired student's and Mann–Whitney t test when significant (P < 0·05) are shown. CMV: cytomegalovirus; AdV: adenovirus.

Download figure to PowerPoint

Note also that CD8+CD28 and CD4+CD45RA T-cell levels were higher in CMV-reactivating opposed to the non-reactivating subgroups but similar in the subgroups with and without relapse.

Following characterization of the 89 patients as responders to CMV or to AdV, based on SI ≥4 in proliferation assays and adjustment for VHR status, multivariate analyses showed that immune response to both CMV and AdV were independently associated with GVL (P = 0·03 and 0·01, respectively). In deep contrast, early CMV reactivation (before day 30) was associated with an increased risk of relapse in univariate analysis (P = 0·01) that nearly reached significance in multivariate analysis (P = 0·06). Furthermore, multivariate analyses showed the lowest risk of relapse in recipients with early (before day 120) immunity to CMV but no CMV-reactivation at that time, and the highest risk in recipients with early (before day 120) CMV-reactivation but no immunity to CMV at that time. This observation further supports different risks of leukaemic relapse according to immunity to CMV and CMV-reactivation. Data from multivariate analyses are shown in Table 3.

Table 3. Multivariate analysis in the 89 recipients
VariableHR (95%CI) P
  1. HR, hazard ratio; 95%CI, 95% confidence interval; CMV, cytomegalovirus; AdV, adenovirus; RI+/−, recipients with (+) and without (−) immunity to CMV at month 3 (±1).

  2. V+/−: recipients with (+) and without (−) CMV reactivation during CMV-PCR surveillance (The first 3 months post-HSCT).

Very high risk
No1
Yes4·02 (1·07–15·1)0·039
CMV reactivation
V+1
V6·03 (4·21–13·56)0·06
CMV immunity
RI+1
RI0·32 (0·18–0·65)0·03
AdV immunity
RI+1
RI0·28 (0·09–0·36)0·01
CMV immunity/reactivation
RI/V+1
RI/V0·09 (0·017–0·54)0·008
RI+/V0·06 (0·005–0·69)0·024
RI+/V+0·19 (0·033–1·07)0·06

Discussion

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Funding
  8. Authorship contributions
  9. Conflicts of interest
  10. References
  11. Supporting Information

In this study, using a reasonable sized cohort of 108 paediatric recipients, cGVHD and VHR status were two factors associated with RR, as expected from previous studies (Ringden & Horowitz, 1989; Armand et al, 2012; Locatelli et al, 2012; Hochberg et al, 2013). As also expected from previous reports in the literature, our analysis ruled out a major role of the stem-cell source (MRD, unrelated donor or unrelated CBU) (Eapen & Wagner, 2010; Zhang et al, 2012), ATG or TBI usage (Wang et al, 2011) and aGVHD (Ringden & Horowitz, 1989) on RR. More surprisingly, disease stage had no impact on RR. In agreement with the interpretation reported by Hochberg et al (2013), the most likely explanation for this later observation is that VHR cases in CR1 were referred for HSCT while those at lower risk were probably not referred until they had relapsed.

This study analysed for the first time the impact of early CMV reactivation on GVL in children. Univariate analysis showed that CMV reactivation occurrence before day +30 was associated with increased RR in the 89 recipients alive without relapse after day +120, and this approached significance in multivariate analysis. Elmaagacli et al (2011) first suggested a clear-cut beneficial effect of early CMV viraemia on the risk of leukaemic relapse in an homogeneous population of adult AML patients grafted with genoidentical HSCT (Elmaagacli et al, 2011). More recently, Green et al (2013) observed a much less striking beneficial effect in a more heterogeneous population of AML patients. Furthermore, in contrast with Ito et al (2013), they did not found any association between CMV reactivation and RR in chronic myelogenous leukaemia patients. Finally, an effect of CMV reactivation on reducing relapse is not obvious in ALL (Green et al, 2013). Although we did not have sufficient statistical power to examine differences between disease subgroups, it is worth noting that a trend toward increased RR was observed in both ALL- and AML-reactivating recipients from this study (not shown). Further studies are required to determine whether the discrepant results observed within paediatric versus adult leukaemia relate to patient age and/or leukaemia nature. Further studies should also explore the prognostic value of early CMV viraemia in homogeneous subsets of leukaemia in children, though it remains difficult to address this issue because of paediatric leukaemia heterogeneity.

Of note, the increased RR was associated with CMV reactivation in the 89 recipients alive and relapse-free after day 120 post-HSCT but not in the total cohort. The involvement in this apparent controversy of a confounding factor is a plausible explanation. Indeed, the rate of VHR status among recipients who relapsed within the first 3 months post-HSCT was 12/13 (92%) but 7/14 (50%) among the recipients who relapsed beyond month +3 (P = 0·03).

An association between decreased risk of relapse and recovery of CMV immunity in both univariate and multivariate analysis contrasted with the increased risk of relapse associated with reactivation in univariate analysis that nearly reached significance in multivariate analysis. This apparent controversy arises as the result of dissociation between immune response and CMV reactivation in some of the recipients. It is generally accepted that subclinical levels of CMV replication that are not detected by PCR in blood (occult reactivations) allows a certain degree of antigen exposure to immune system, thus favouring anti-CMV T-cell reconstitution (Lilleri et al, 2009; Abate et al, 2012; Merindol et al, 2012). In contrast, elevated CMV-DNA copies/ml may anticipate, by several weeks or months, the appearance of a detectable CMV-specific response in immunocompromised individuals. As shown in Figure S2, the recovery of the proliferative response to CMV was accelerated in recipients with occult reactivations (RI+/V) as opposed to recipients with CMV-PCR positivity (V+). This observation supports that the optimal early recovery of T-cell response to CMV impeded virus expansion. We therefore propose that the absence of detectable CMV-DNAemia but recovery of immunity to the virus must be regarded as an indicator of efficient immune reconstitution, defining a group of patients with a low risk of clinical infections and a decreased risk of relapse. At variance to, an earlier study that analysed RR according to immunity to CMV in paediatric recipients (Nakamura et al, 2004), impaired recovery of anti-CMV immunity in recipients with persistent CMV reactivation (RI/V+) was associated with an increased risk of relapse. Nakamura et al (2004) suggested that immunity to CMV did not have a direct effect on GVL but was a surrogate factor for the immunocompetence of the recipient after HSCT. Our results, showing an association between GVL and early thymopoiesis as well as improved immune recovery of anti-AdV in addition to anti-CMV immunity give supportive evidences for this hypothesis.

That CMV-driven immunity had no impact on GVL in our paediatric series is further supported by the observation that increased accumulation of CD8+CD28 and CD45RACD4+ T-cells in CMV reactivating recipients had no influence on RR.

The expansion of CD28CD8+ T-cells by CMV has been documented in both immunocompetent and immunocompromised individuals, though reports on this phenomenon in HSCT are scarce. As far as we know, only one study previously assessed the impact of CD8+CD28 T-cell levels on relapse (Yakoub-Agha et al, 2009); the authors reported a beneficial effect on GVL of CD8+CD28 T-cell expansion. Differences in recipient age and/or leukaemia nature may explain the discrepancies between this study and ours.

Expansion of memory CD4+ T-cells by CMV is a less well-known phenomenon. As far as we know, only one study suggested that bystander-secreted differentiation-inducing factors during CMV infection might induce differentiation of CD4+ T-cells of different specificities (Fletcher et al, 2005). Whatever the mechanism involved, again this CMV-induced change does not appear to have an impact on GVL as far as paediatric leukaemia is concerned. Increased levels of NK cells were also observed in CMV-reactivating recipients. An impact of CMV on the NK cell compartment has been recently documented (Kreutzman et al, 2011). Also, NK cells are increasingly implicated as important mediators of GVL. This however appears to apply mostly to adult myeloid diseases and in the setting of haploidentical transplant (Ruggeri et al, 2002).

Finally, in both univariate and multivariate analysis, CMV serostatus of either donor or recipient at HSCT was not associated with RR. As far as we know, only two studies using pre-emptive antiviral chemotherapy have previously examined the effect of donor and recipient serostatus on paediatric alloHSCT outcome. An association between recipient CMV seropositivity and increased non-relapse mortality (P = 0·05) was noted in the former (Behrendt et al, 2009), but superior relapse-free survival when recipients and/or donors were CMV seropositive in the latter (Bordon et al, 2008). The reasons for these discrepancies are not clear. One explanation may be differences between the studies in the involvement of confounding factors for relapse, as appropriate statistical methodology was not used in these previous studies. Further studies using large series and appropriate statistical methods are required to definitely solve this point.

To conclude, despite a clear impact of CMV reactivation on the profile of immune recovery, an influence of CMV-driven immunity on GVL was not obvious in this paediatric series. In line with previous studies, a favourable impact on the risk of relapse of early thymopoiesis and development of T-cell functional activities was however observed (Parkman et al, 2006; Williams & Gress, 2008). We propose that the striking association of CMV reactivations with increased RR in this study may be regarded as an indicator of suboptimal immune recovery resulting in increased risk of viraemia and decreased GVL. Finally, changes by CMV reactivation in immune recovery observed in this study invite further exploration as to whether expansion by CMV reactivation of memory CD4 T-cells and/or CD8+CD28 T-cells and/or discrete subsets of NK cells play a role in adult settings where early CMV reactivation is associated with less relapse.

Acknowledgements

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Funding
  8. Authorship contributions
  9. Conflicts of interest
  10. References
  11. Supporting Information

We acknowledge the technical expertise of Guylaine Boiry, Priscilia Egremonte, Elodie Richez and Judith Tholle. We acknowledge the patient nurses and staff of the department of haematology without whom this research would not have been possible. Finally, we greatly thank Céline Neto for preparing the manuscript.

Funding

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Funding
  8. Authorship contributions
  9. Conflicts of interest
  10. References
  11. Supporting Information

The work was partly supported by the ‘Agence de la Biomédecine’ and by the ‘Assistance Publique-Hôpitaux de Paris’.

Authorship contributions

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Funding
  8. Authorship contributions
  9. Conflicts of interest
  10. References
  11. Supporting Information

Mohamed Jeljeli: analysed the results of immunological investigations, contributed to statistics analysis and to writing the manuscript; Valérie Guérin-El Khourouj: designed the immunological investigations and analysed the results; Raphael Porcher: performed the statistics; Sandrine Leveillé: recorded prospectively the clinical characteristics of the patients; Mony Fahd, Karima Yakouben and Marie Ouachée-Chardin: took care of the patients; Jerome LeGoff: performed the virological part of the study; Debora Jorge Cordeiro: contributed to result recording and analysis and contributed to writing the manuscript; Beatrice Pédron: analysed the HLA compatibility between donors and patients and contributed to the immunological part of the study; Andre Baruchel: contributed to the design of the study; Jean-Hugues Dalle and Ghislaine Sterkers: designed the study, analysed the results and wrote the manuscript.

References

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Funding
  8. Authorship contributions
  9. Conflicts of interest
  10. References
  11. Supporting Information
  • Abate, D., Cesaro, S., Cofano, S., Fiscon, M., Saldan, A., Varotto, S., Mengoli, C., Pillon, M., Calore, E., Biasolo, M.A., Cusinato, R., Barzon, L., Messina, C., Carli, M. & Palu, G. (2012) Diagnostic utility of human cytomegalovirus-specific T-cell response monitoring in predicting viremia in pediatric allogeneic stem-cell transplant patients. Transplantation, 93, 536542.
  • Armand, P., Gibson, C.J., Cutler, C., Ho, V.T., Koreth, J., Alyea, E.P., Ritz, J., Sorror, M.L., Lee, S.J., Deeg, H.J., Storer, B.E., Appelbaum, F.R., Antin, J.H., Soiffer, R.J. & Kim, H.T. (2012) A disease risk index for patients undergoing allogeneic stem cell transplantation. Blood, 120, 905913.
  • Behrendt, C.E., Rosenthal, J., Bolotin, E., Nakamura, R., Zaia, J. & Forman, S.J. (2009) Donor and recipient CMV serostatus and outcome of pediatric allogeneic HSCT for acute leukemia in the era of CMV-preemptive therapy. Biology of Blood and Marrow Transplantation, 15, 5460.
  • de Berranger, E., Cousien, A., Petit, A., Peffault de Latour, R., Galambrun, C., Bertrand, Y., Salmon, A., Rialland, F., Rohrlich, P.S., Vannier, J.P., Lutz, P., Yakouben, K., Duhamel, A., Bruno, B., Michel, G. & Dalle, J.H. (2014) Impact on long-term OS of conditioning regimen in allogeneic BMT for children with AML in first CR: TBI+CY versus BU+CY: a report from the Societe Francaise de Greffe de Moelle et de Therapie Cellulaire. Bone Marrow Transplantation, 49, 382388.
  • Bordon, V., Bravo, S., Van Renterghem, L., de Moerloose, B., Benoit, Y., Laureys, G. & Dhooge, C. (2008) Surveillance of cytomegalovirus (CMV) DNAemia in pediatric allogeneic stem cell transplantation: incidence and outcome of CMV infection and disease. Transplant Infectious Disease, 10, 1923.
  • Development Core Team, R. (2011) A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria.
  • Eapen, M. & Wagner, J.E. (2010) Transplant outcomes in acute leukemia. I. Seminars in Hematology, 47, 4650.
  • Elmaagacli, A.H., Steckel, N.K., Koldehoff, M., Hegerfeldt, Y., Trenschel, R., Ditschkowski, M., Christoph, S., Gromke, T., Kordelas, L., Ottinger, H.D., Ross, R.S., Horn, P.A., Schnittger, S. & Beelen, D.W. (2011) Early human cytomegalovirus replication after transplantation is associated with a decreased relapse risk: evidence for a putative virus-versus-leukemia effect in acute myeloid leukemia patients. Blood, 118, 14021412.
  • Fletcher, J.M., Vukmanovic-Stejic, M., Dunne, P.J., Birch, K.E., Cook, J.E., Jackson, S.E., Salmon, M., Rustin, M.H. & Akbar, A.N. (2005) Cytomegalovirus-specific CD4+ T cells in healthy carriers are continuously driven to replicative exhaustion. Journal of Immunology, 175, 82188225.
  • Gooley, T.A., Chien, J.W., Pergam, S.A., Hingorani, S., Sorror, M.L., Boeckh, M., Martin, P.J., Sandmaier, B.M., Marr, K.A., Appelbaum, F.R., Storb, R. & McDonald, G.B. (2010) Reduced mortality after allogeneic hematopoietic-cell transplantation. The New Journal England of Medicine, 363, 20912101.
  • Grambsch, P.M. & Therneau, T.M. (1994) Proportional hazards tests and diagnostics based on weighted residuals. Biometrika, 81, 515526.
  • Green, M.L., Leisenring, W.M., Xie, H., Walter, R.B., Mielcarek, M., Sandmaier, B.M., Riddell, S.R. & Boeckh, M. (2013) CMV reactivation after allogeneic HCT and relapse risk: evidence for early protection in acute myeloid leukemia. Blood, 122, 13161324.
  • Guerin, V., Dalle, J.H., Pedron, B., Ouachee-Chardin, M., Yakouben, K., Baruchel, A. & Sterkers, G. (2010) Cellular immune parameters associated with spontaneous control of CMV in children who underwent transplantation. Bone Marrow Transplantation, 45, 442449.
  • Hochberg, J., Khaled, S., Forman, S.J. & Cairo, M.S. (2013) Criteria for and outcomes of allogeneic haematopoietic stem cell transplant in children, adolescents and young adults with acute lymphoblastic leukaemia in first complete remission. British Journal of Haematology, 161, 2742.
  • Horowitz, M.M., Gale, R.P., Sondel, P.M., Goldman, J.M., Kersey, J., Kolb, H.J., Rimm, A.A., Ringden, O., Rozman, C. & Speck, B. (1990) Graft-versus-leukemia reactions after bone marrow transplantation. Blood, 75, 555562.
  • Ito, S., Pophali, P., Co, W., Koklanaris, E.K., Superata, J., Fahle, G.A., Childs, R., Battiwalla, M. & Barrett, A.J. (2013) CMV reactivation is associated with a lower incidence of relapse after allo-SCT for CML. Bone Marrow Transplantation, 48, 13131316.
  • Kalbfleisch, J.D. & Prentice, R.L. (1980) The Statistical Analysis of Failure Time Data. John Wiley & Sons, New York, NY.
  • Kreutzman, A., Ladell, K., Koechel, C., Gostick, E., Ekblom, M., Stenke, L., Melo, T., Einsele, H., Porkka, K., Price, D.A., Mustjoki, S. & Seggewiss, R. (2011) Expansion of highly differentiated CD8+ T-cells or NK-cells in patients treated with dasatinib is associated with cytomegalovirus reactivation. Leukemia, 25, 15871597.
  • Lilleri, D., Gerna, G., Fornara, C., Chiesa, A., Comolli, G., Zecca, M. & Locatelli, F. (2009) Human cytomegalovirus-specific T cell reconstitution in young patients receiving T cell-depleted, allogeneic hematopoietic stem cell transplantation. The Journal of Infectious Diseases, 199, 829836.
  • Locatelli, F., Schrappe, M., Bernardo, M.E. & Rutella, S. (2012) How I treat relapsed childhood acute lymphoblastic leukemia. Blood, 120, 28072816.
  • Lonnqvist, B., Ringden, O., Ljungman, P., Wahren, B. & Gahrton, G. (1986) Reduced risk of recurrent leukaemia in bone marrow transplant recipients after cytomegalovirus infection. British Journal of Haematology, 63, 671679.
  • de Masson, A., Bouaziz, J.D., Peffault de Latour, R., Wittnebel, S., Ribaud, P., Rubio, M.T., Micol, J.B., Suarez, F., Nguyen, S., Dalle, J.H., Yakouben, K., Robin, M., Xhaard, A., Ades, L., Bourhis, J.H., Rybojad, M., Bagot, M. & Socie, G. (2012) Limited efficacy and tolerance of imatinib mesylate in steroid-refractory sclerodermatous chronic GVHD. Blood, 120, 50895090.
  • Matthes-Martin, S., Feuchtinger, T., Shaw, P.J., Engelhard, D., Hirsch, H.H., Cordonnier, C. & Ljungman, P. (2012) European guidelines for diagnosis and treatment of adenovirus infection in leukemia and stem cell transplantation: summary of ECIL-4 (2011). Transplant Infectious Disease, 14, 555563.
  • Merindol, N., Salem Fourati, I., Brito, R.M., Grenier, A.J., Charrier, E., Cordeiro, P., Caty, M., Mezziani, S., Malette, B., Duval, M., Alfieri, C., Ovetchkine, P., Le Deist, F. & Soudeyns, H. (2012) Reconstitution of protective immune responses against cytomegalovirus and varicella zoster virus does not require disease development in pediatric recipients of umbilical cord blood transplantation. Journal of Immunology, 189, 50165028.
  • Nakamura, R., Battiwalla, M., Solomon, S., Follmann, D., Chakrabarti, S., Cortez, K., Hensel, N., Childs, R. & Barrett, A.J. (2004) Persisting posttransplantation cytomegalovirus antigenemia correlates with poor lymphocyte proliferation to cytomegalovirus antigen and predicts for increased late relapse and treatment failure. Biology of Blood and Marrow Transplantation, 10, 4957.
  • Parkman, R., Cohen, G., Carter, S.L., Weinberg, K.I., Masinsin, B., Guinan, E., Kurtzberg, J., Wagner, J.E. & Kernan, N.A. (2006) Successful immune reconstitution decreases leukemic relapse and improves survival in recipients of unrelated cord blood transplantation. Biology of Blood and Marrow Transplantation, 12, 919927.
  • Pedron, B., Guerin, V., Cordeiro, D.J., Masmoudi, S., Dalle, J.H. & Sterkers, G. (2011) Development of cytomegalovirus and adenovirus-specific memory CD4 T-cell functions from birth to adulthood. Pediatric Research, 69, 106111.
  • Prentice, R.L., Kalbfleisch, J.D., Peterson, A.V. Jr, Flournoy, N., Farewell, V.T. & Breslow, N.E. (1978) The analysis of failure times in the presence of competing risks. Biometrics, 34, 541554.
  • Rager, A. & Porter, D.L. (2011) Cellular therapy following allogeneic stem-cell transplantation. Therapeutic Advances in Hematology, 2, 409428.
  • Ringden, O. & Horowitz, M.M. (1989) Graft-versus-leukemia reactions in humans. The Advisory Committee of the International Bone Marrow Transplant Registry. Transplantation Proceedings, 21, 29892992.
  • Ruggeri, L., Capanni, M., Urbani, E., Perruccio, K., Shlomchik, W.D., Tosti, A., Posati, S., Rogaia, D., Frassoni, F., Aversa, F., Martelli, M.F. & Velardi, A. (2002) Effectiveness of donor natural killer cell alloreactivity in mismatched hematopoietic transplants. Science, 295, 20972100.
  • Wang, Y., Liu, D.H., Xu, L.P., Liu, K.Y., Chen, H., Chen, Y.H., Han, W., Shi, H.X. & Huang, X.J. (2011) Superior graft-versus-leukemia effect associated with transplantation of haploidentical compared with HLA-identical sibling donor grafts for high-risk acute leukemia: an historic comparison. Biology of Blood and Marrow Transplantation, 17, 821830.
  • Williams, K.M. & Gress, R.E. (2008) Immune reconstitution and implications for immunotherapy following haematopoietic stem cell transplantation. Best Practice & Research Clinical Haematology, 21, 579596.
  • Yakoub-Agha, I., Saule, P., Magro, L., Cracco, P., Duhamel, A., Coiteux, V., Bruno, B., Dufosse, F., Jouet, J.P., Dessaint, J.P. & Labalette, M. (2009) Immune reconstitution following myeloablative allogeneic hematopoietic stem cell transplantation: the impact of expanding CD28negative CD8+ T cells on relapse. Biology of Blood and Marrow Transplantation, 15, 496504.
  • Zhang, M.J., Davies, S.M., Camitta, B.M., Logan, B., Tiedemann, K., Eapen, M. & Thiel, E.L. (2012) Comparison of outcomes after HLA-matched sibling and unrelated donor transplantation for children with high-risk acute lymphoblastic leukemia. Biology of Blood and Marrow Transplantation, 18, 12041210.

Supporting Information

  1. Top of page
  2. Summary
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Funding
  8. Authorship contributions
  9. Conflicts of interest
  10. References
  11. Supporting Information
FilenameFormatSizeDescription
bjh12875-sup-0001-FigS1.docWord document109KFig S1. Kinetics of immunity to CMV development.
bjh12875-sup-0002-FigS2.docWord document29KFig S2. Accelerated immune recovery in occult opposed to documented reactivations.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.