Impact of Cytomegalovirus Reactivation and Natural Killer Reconstitution on Outcomes after Allogeneic Hematopoietic Stem Cell Transplantation: A Single-Center Analysis

Transplant outcomes in patients with cytomegalovirus reactivation

Journal Pre-proof

Impact of Cytomegalovirus Reactivation and Natural Killer Reconstitution on Outcomes after Allogeneic Hematopoietic Stem Cell Transplantation: A Single Center Analysis Taiki Ando , Taisei Suzuki , Yasufumi Ishiyama , Satoshi Koyama , Takayoshi Tachibana , Masatsugu Tanaka , Heiwa Kanamori , Hideaki Nakajima PII: DOI: Reference:

S1083-8791(19)30639-1 https://doi.org/10.1016/j.bbmt.2019.09.028 YBBMT 55739

To appear in:

Biology of Blood and Marrow Transplantation

Received date: Accepted date:

5 July 2019 24 September 2019

Please cite this article as: Taiki Ando , Taisei Suzuki , Yasufumi Ishiyama , Satoshi Koyama , Takayoshi Tachibana , Masatsugu Tanaka , Heiwa Kanamori , Hideaki Nakajima , Impact of Cytomegalovirus Reactivation and Natural Killer Reconstitution on Outcomes after Allogeneic Hematopoietic Stem Cell Transplantation: A Single Center Analysis, Biology of Blood and Marrow Transplantation (2019), doi: https://doi.org/10.1016/j.bbmt.2019.09.028

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc. on behalf of the American Society for Transplantation and Cellular Therapy

Highlights ・ Post-transplant cytomegalovirus (CMV) reactivation is associated with non-relapse mortality and overall survival. ・ Natural killer (NK) cell subset significantly contributed to a lower cumulative incidence of relapse. ・ Natural killer cell reconstitution improved post-transplant outcomes, especially in patients with CMV reactivation.

ORIGINAL ARTICLE

Impact of Cytomegalovirus Reactivation and Natural Killer Reconstitution on Outcomes after Allogeneic Hematopoietic Stem Cell Transplantation: A Single Center Analysis

Taiki Ando1*, Taisei Suzuki1, Yasufumi Ishiyama1, Satoshi Koyama2, Takayoshi Tachibana1, Masatsugu Tanaka1, Heiwa Kanamori1, and Hideaki Nakajima2

1

Department of Hematology, Kanagawa Cancer Center, Yokohama, Japan;

2

Department of Hematology and Clinical Immunology, Yokohama City University School of

Medicine, Japan;

Short title: Transplant outcomes in patients with cytomegalovirus reactivation

Correspondence: Dr. Taiki Ando Department of Hematology, Kanagawa Cancer Center 2-3-2 Nakao, Asahi-ku, Yokohama 241-8515, Japan Phone: 81-45-520-2222 Fax: 81-45-520-2202, E-mail: [email protected] Financial disclosure statement: There are no conflicts of interests to report.

Abstract Cytomegalovirus (CMV) reactivation and natural killer (NK) cell reconstitution are well recognized immunological events occurring after allogeneic stem cell transplantation (allo-SCT). We aimed to study the outcome of CMV reactivation and NK cell reconstitution in patients with hematological malignancies after allo-SCT. We retrospectively studied 246 adult patients (152 males, 94 females, median age of 51 years, 18–69 years) who underwent allo-SCT for hematological malignancies at the Kanagawa Cancer Center. CMV reactivation was defined as initiation of pre-emptive CMV therapy following pp65 antigenemia surveillance. All patients’ lymphocyte subsets were monitored by flow cytometry at day 180, 365, and 730 days posttransplant. The median follow-up period was 3.2 years (0.8–9.6 years). CMV reactivation occurred in 141 patients (57%) at a median of 45 days (15–93 days). In patients without CMV reactivation (CMVR-) vs. those with CMV reactivation (CMVR+), five-year OS, NRM, and CIR were 79 vs. 55% (P < 0.001), 3 vs. 16% (P = 0.012), and 28 vs. 38% (P = 0.09), respectively. CD8+ T cell and CD3-CD56+ NK cell subset were higher in CMVR+ at day 100 posttransplant. Multivariate analysis showed that adverse factors for OS were represented by no remission, CMV reactivation, and lower CD16+CD57- NK cell counts. Overall, a higher NK cell subset significantly contributed to a lower CIR. Among subgroups of CMVR+, CD16+CD57- NK cells represented a favorable factor for OS, NRM, and CIR. CMV reactivation was an adverse event after allo-SCT. NK cell reconstitution may contribute to improved outcomes, especially in CMVR+ subgroups.

Keywords: Cytomegalovirus reactivation, natural killer cell reconstitution, non-relapse mortality, allogeneic stem cell transplantation

Introduction Allogeneic hematopoietic stem cell transplantation (allo-SCT) is a curative therapy for hematological malignancies via the graft-versus-tumor effects. Immunity is impaired during the first month. A complete cell count recovery takes years. Specifically, following an allo-SCT, immune cell subsets gain regular immune reconstitution (IR) over different timelines [1]. The donor-derived IR’s regular function is affected by various factors (e.g., conditioning regimen, graft type, stem cell dose, graft-versus-host disease (GVHD) prophylaxis, GVHD itself, and presence of infection). Successful IR after allo-SCT is associated with superior outcomes [2, 3]. Cytomegalovirus (CMV) reactivation is a known allo-SCT complication. In patients with acute myeloid leukemia (AML), CMV reactivation has indirect effects, such as a lower incidence of disease relapse [4, 5]. Flow cytometry (FCM) enables identification of lymphocyte subsets and their maturation as IR of T, B and natural killer (NK) cells. Several studies have demonstrated that rapid lymphocyte repopulation with T, B, and NK cells, identified by FCM, reduces the incidence of infection, GVHD, and disease relapse [6-9]. This study aims to clarify immunological specifications, following CMV reactivation, and its clinical impacts on post-transplant

outcomes. By using FCM for IR 100 days post-transplant, according to CMV reactivation, this study might enable the immunological prediction of allo-SCT outcomes.

Patients and methods 1. Patients and transplant This study included adult patients (≥18 years old) with hematological malignancies. Patients underwent their first allo-HSCT between April 2009 and December 2017 at the Kanagawa Cancer Center. Clinical data were collected from their medical charts. Exclusion criteria were death or graft failure in fewer than 100 days from allo-SCT. Myeloablative conditioning (MAC) regimens were defined as follows: cyclophosphamide (120 mg/kg) and total body irradiation (TBI) >10 Gy, or busulfan at >8 mg/kg (oral) or >6.4 mg/kg (intravenous). Reduced-intensity conditioning (RIC) regimens were either fludarabine 125 mg/m2 plus melphalan 80 (or 140) mg/m2 with 2 or 4Gy TBI or busulfan at 3.2 mg/kg (intravenous). GVHD prophylaxis mainly consisted of tacrolimus (TAC) or cyclosporine (CSP) and a short course of methotrexate (MTX). Three individuals experienced post-transplant cycrophosphamide (PTCY). Additionally, 15 patients received antithymocyte globulin (ATG) combination prophylaxis. Diagnosis and grading of acute and chronic GVHD was performed as previously described [10, 11]. The graft’s sources included donor bone marrow (BM), peripheral blood stem cell (PBSC), and single umbilical cord blood (UCB).

These sources were not modified by in or ex vivo T-cell depletion. Regardless of graft sources, all patients routinely received levofloxacin, fluconazole, and aciclovir as antibacterial, antifungal, and antiviral prophylaxis, respectively. Additionally, all patients received lenograstim (10 μg/kg/day), from day 5 until recovery, with an absolute neutrophil count of 0.5×109 cells/L for three consecutive days. Engraftment was measured by donor/recipient chimerism, through bone marrow aspiration, at 28 days. Complete donor chimerism was defined as ≥95% donor-derived leukocytes in BM samples. Measurements were performed by fluorescent in situ hybridization, with specific probes for the sex chromosomes, or by multiplex short tandem repeat polymerase chain reaction. As later described, lymphocyte subsets by FCM and immunoglobulin G (IgG) were measured on days 100, 180, 365, and 730 post-transplant. The median values of cell counts were used for IR measurements at each time points as high (greater than or equal to the median) vs low (less than the median). An early IR was defined when

patients experienced a higher cell counts on days 100 from allo-SCT.

2. CMV reactivation and preemptive therapy Patients received weekly pp65 antigenemia surveillance, starting at the time of neutrophil engraftment following allo-SCT. CMV prophylaxis was not used in any patient. CMV reactivation was defined as the start of CMV preemptive therapy. Intravenous foscarnet (FCV), ganciclovir (GCV), or oral Valganciclovir were administered as CMV preemptive

therapy. The latter was generally initiated when at least two CMV pp65 antigen–positive cells per 50,000 white blood cells were detected, or a systemic steroid therapy (corticosteroid ≥0.5 mg/kg) was induced. CMV disease was diagnosed per previously reported recommendations [12].

3. Flow cytometry analysis Subsets of peripheral blood (PB) samples were analyzed by FCM at 100, 180, 365, and 730 days following allo-SCT. To ensure adequate cell evaluation, at least 10,000 lymphocytes were measured. Viable lymphocytes were gated using FSC/SSC and 7AAD staining. Briefly, two color reagent panels of fluorochrome-conjugated monoclonal antibodies, specific for the representative surface antigen (CD4*CD8, CD20*CD2, CD3*CD56, and CD16*CD57), were used. The absolute lymphocyte cell (ALC) counts of cell subsets were calculated by assessing the percentage of the presented subsets as determined by FCM. Subsets were acquired on a Gallios cytometer. Data were then analyzed with Kaluza software (Beckman Coulter, Brea, CA, USA).

4. Statistical analysis To compare the baseline characteristics and cell counts of lymphocyte subsets, Fisher’s exact test and the Wilcoxon rank-sum test were used for univariate analysis, and a multiple logistic

regression model for multivariate analysis. Quantitative measures of IR were summarized using median (range) and compared between groups using Wilcoxon rank-sum tests at each time point. Overall survival (OS) and disease-free survival (DFS) was recorded from the post-transplant day until the day of death or disease progression. Thus, both events including an event of chronic GVHD were for chronic GVHD, relapse-free survival (GRFS) [13]. The Kaplan–Meier technique was used to analyze the probability of OS and DFS [14]. Landmark analyses of survival and disease-free survival based on CMV reactivation and NK cell IR at the landmark time were conducted using Kaplan Meier estimates and compared using the log-rank test. The log-rank test was used for univariate analysis and the Cox proportional hazard model for multivariate analysis for OS [15]. Cumulative incidences of relapse (CIR), non-relapse mortality (NRM), and GVHD were evaluated using the Fine and Gray model for univariate and multivariate analyses. Differences between means were considered statistically significant at two-sided P < 0.05. Statistical analysis was performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan). The latter is a graphical user interface for R (the R Foundation for Statistical Computing, Vienna, Austria) [16].

Results 1. Immune reconstitution according to CMV reactivation At the Kanagawa Cancer Center, 246 adult patients with hematological malignancies

underwent allo-SCT and had available IR data. All patients had a PB lymphocyte analysis, by FCM, of at least 10,000 cell counts at 100, 180, 365, 730 days after allo-SCT. In patients without CMV reactivation (CMVR-) vs. those with CMV reactivation (CMVR+), an identification of lymphocyte subsets was summarized for ALC, CD4+ T cell, CD8+ T cell, CD20+ B cell, CD3-CD56+ NK cell and CD16+CD57- NK cell counts (/µl) (Supplemental Figure 1). Significant differences of median cell counts were observed in CD8+ T cell and CD3-CD56+ NK at 100 day from allo-SCT. Timely changed IR were observed in presented subsets according to CMV reactivation. IR differences had gradually decreased among groups after allo-SCT (Supplemental Figure 2). CD16+CD57- NK cell is defined as a terminal matured in NK cell subsets. Its activation occurs by both producing interferon-γ and cytotoxic activity, especially enriched in CB graft. There were no significant correlations between the presence of CMV reactivation and CD16+CD57- NK cell reconstitution. Using a variance inflation factor, a positive correlation was observed between both CD3-CD56+ NK cell and CD16+CD57+ NK cell counts with the experience of CMV reactivation. In the present study, both subsets were confounding factors. Of note, our data included 112 patients (45.5%) with CB as a graft source. Therefore, we focused on CD16+CD57- NK cells as a comparison with patient characteristics and survival outcomes combined with CMV reactivation.

2. Patient characteristics The patients’ characteristics, classified in terms of CMV reactivation and NK cell reconstitution, are described in Table 1.CMV-R: CMV reactivation following allogeneic stem cell transplantation, CMV serostatus: R for recipient, D for donor. NK cell reconstitution: CD16+CD57- NK cell subset was focused, IR: immune reconstitution. The median age was 51 years (range: 18-69). At the time of transplantation, diagnosis was AML (n=131, 53.4%), acute lymphoblastic leukemia (n=61, 24.8%), myelodysplastic syndrome (n=32, 13%), chronic myeloid leukemia (n=7, 2.8%), myeloproliferative neoplasms (n=7, 2.8%), and other hematological malignancy (n=8, 3.2%). For GVHD prophylaxis, 189 patients received TAC base and 57 patients was CSP, respectively. Disease risk were standard risk in 166 patients and high risk in 80 patients. MAC and RIC were employed in 99 (40.2%) and 147 patients (59.8%), respectively. BM, PBSC, and CB was selected as graft sources in 98 (39.8%), 36 (14.6%), and 112 patients (45.5%), respectively. According to the serostatus of recipient (R) versus donor (D), i.e., R+/D+, R+/D-, R-/D+, and R-D-, the cumulative incidence of CMV reactivation was 60, 67, 33, and 15%, respectively (P < 0.001). CMV reactivation was experienced in 141 patients (57.3%) on days 100 from allo-SCT, whereas early NK cell IR was in 131 patients (53.3%), respectively. There was no significant correlation with the presence of CMV reactivation and NK cell IR (P = 0.702). Statistical correlation with CMV reactivation was observed in higher ages (≥50), CMV serology (recipient seropositivity), conditioning regimens

(RIC), graft sources (CB), and acute GVHD (grade 2-4), respectively (each P < 0.05). By multivariate analysis, a relationship was found for CMV reactivation with both age [odds ratio (OR) 2.66 (95% CI 1.22–6.52), P = 0.018], CMV seropositivity [OR 5.45 (2.46–12.3), P < 0.001], and acute GVHD [OR 3.14 (95% CI 1.61-8.65), P < 0.001], respectively. Additionally, patient characteristics differed according to CD 16+CD57- NK cell IR. There was a statistical relationship with early NK cell IR in disease risk (standard risk), graft source (CB), and acute GVHD (grade 2-4) each with P < 0.05. Multivariate analysis showed a significant correlation between NK cell IR and both disease risk [OR 2.47 (1.72–8.31), P = 0.0021] and graft source [OR 1.99 (1.39–4.79), P = 0.015].

3. Survival outcomes The median follow-up duration was 3.2 years (range: 0.8–9.6 years). The cumulative incidence of neutrophil engraftment was achieved in all patients with a median of 19 days (range: 13–44). For all patients, the five-year OS, cumulative NRM, and CIR rates were 66, 10, and 33%, respectively. The cumulative incidence of grade II-IV acute GVHD and chronic GVHD affected 98 (39.8%) and 80 (32.5%) patients, respectively. Our study results show that both CMV reactivation and early NK cell IR were significant events for post-transplant outcomes. Outcomes in CMVR- vs. CMVR+ patients, for five-year OS, NRM, and CIR were 78.8 vs. 55.1% (P < 0.001), 3.0 vs. 16.2% (P = 0.012), and 27.8 vs. 38% (P = 0.091),

respectively (Figure 1A). Thus, the sub group analyses based on CMV reactivation were shown in Supplemental Table. Of note, outcomes in the groups High vs. Low CD16+CD57- NK cell counts groups, for five-year OS, NRM, and CIR were 75.0 vs. 35.7% (P < 0.001), 9.4 vs. 28.9% (P < 0.001), and 16.2 vs. 52.4% (P = 0.0007), respectively (Figure 1B). Notably, for all patients, five-year OS curves were significantly different based on a combination with CMV reactivation and NK cell IR (Figure 2). Of 141 CMVR+ patients, 78 had higher NK cells and 63 had lower NK cells. Outcomes in terms of OS for these two groups, NRM and CIR, were 69.1 vs. 26.7% (P < 0.001), 11.4 vs. 21.8% (P = 0.045), and 26.4 vs. 48.8% (P = 0.011), respectively. Likewise, outcomes for five-year cumulative incidence of chronic GVHD and chronic GRFS were 35.9 vs. 46.5% (P = 0.188) and 21.3 vs. 67.5% (P < 0.001), respectively (Figure 3). Multivariate analysis’s results are shown in Table 2. Univariate analysis for OS demonstrated that significant variables included graft sources, disease risk, CMV reactivation, and NK cell (CD16+CD57-). Multivariate analysis found high disease risk (P = 0.006), CMVR+ (P = 0.019) and low NK cell (CD16+CD57-) (P < 0.001) as adverse predictors. Univariate analysis for NRM demonstrated that significant variables were CMV recipient seropositivity, graft sources, CMV reactivation, and NK cell (CD16+CD57-). Subsequently, the multivariate analysis demonstrated a significant difference in patients with CMVR+ for NRM (P = 0.011).

Finally, univariate analysis for CIR demonstrated that significant values were CMV recipient seropositivity, disease risk, and NK cell (CD16+CD57-). Multivariate analysis then revealed two adverse predictors, including high disease risk (P < 0.001) and low NK cell (CD16+CD57-) (P = 0.031). Causes of death are shown in Table 3. In summary, the probability of idiopathic interstitial pneumonia and maladaptive immunity were significantly greater in patients with CMVR+ (P = 0.012 and P = 0.065, respectively). One patient had confirmed CMV pneumonia at autopsy. Two of three patients with CMVR+ who had died by infection were being treated for GVHD. One of two patients with CMVR+ who experienced hemorrhage was being treated for pneumonia at the time of death.

Discussion

In the present retrospective study for 246 patients with hematological malignancy who received allo-SCT, we demonstrate that post-transplant CMV reactivation was an adverse predictor for NRM and OS. Thus, early CD16+CD57- NK cell IR had a lower CIR and better OS. The presence of CMV reactivation and delayed NK cell IR showed dismal outcomes following allo-SCT. Based on these findings, and given our observations on causes of death, we believe that unbalanced immunity (both maladaptive and deficient) is key for higher NRM in patients with CMV reactivation.

Several studies have reported that CMV reactivation’s frequency was associated with acute GVHD [5, 17-19], severe bacterial infection [19-21], and abnormal balanced immunity, especially in patients using corticosteroids [20]. In line with these findings, our data showed that CMV reactivation’s early onset and intervention of corticosteroids could result in fatal NRM events during the late phase following allo-SCT. Current studies, systematic review and phase 3 double blind trial data, suggested treatment option for CMV prophylaxis after allo-SCT. Antiviral prophylaxis with placebo demonstrated overall effectiveness in reducing CMV reactivation (OR, 0.51; 95% CI, 0.42-0.62) [21]. Whereas, letermovir prophylaxis resulted in a lower risk of CMV infection than placebo as well [22]. Thus, multiple large numbers patients were analyzed by previous studies [23-25]. CMV reactivation was conducted as predictor for higher NRM and linked to complex immune effects after all-SCT. Comparing these data with our findings, CMV prophylaxis might promise the lower incidence of NRM and better OS in patients following allo-SCT. In terms of disease progression, previous studies have reported that CMVR+ was associated with a reduced risk of relapse in patients with AML and/or CML [5, 17, 23], and hematological malignancy [26]. In this study, a statistical tendency for a higher CIR in patients with CMVR+ was observed. Of note, disease recurrence was the principal cause of death. A multivariate analysis showed that the CMV reactivation’s frequency was related to a higher age (≥50) and CMV serostatus, especially in recipient seropositivity. Such results are based on both relatively

old patients and various hematological malignancies. These characteristics may have resulted in a higher disease recurrence. NK cells are cytotoxic lymphocytes which play important roles in early phase of immunity after allo-SCT. Their roles have been well documented as a major components of the innate immunity and modulators of the adaptive immunity. Following a statistical screening, we isolated finally-differentiated NK cell for analysis. Our multivariate analysis results showed a significant impact of High CD16+CD57- NK cell IR in suppressing disease relapse. Of note, high NK cell IR had a dual contribution for chronic GVHD and disease recurrence in patients with CMV reactivation. Similarly, earlier studies reported that an immunocompromised status in the form of NK cell dysfunction might set the stage of CMV reactivation and subsequent mortality in the late phase [20, 27-29]. Therefore, the immunological environment should be evaluated for IR prior to allo-SCT. Based on patient profiles of NK cell IR, we observed that disease risk (high risk) and graft source (BM and PBSC) were significantly correlated with delayed NK cell recovery. Some studies reported similar results [6, 30]. Additionally, the number of NK cells in the graft sources had an impact on disease relapse and acute GVHD [31-32]. In our study, we demonstrated that, eventually, both disease burden and less/lack of NK cells in grafts might adversely influence proper NK cell IR. For instance, dual status of CMV seropositivity and high disease risk at allo-SCT might provoke a donor derived maladaptive immunity in the late

phase after allo-SCT. Taking into account both features of NK cell IR and CMV reactivation prior to allo-SCT, CMV prophylaxis was highly expected to induce proper immunity and superior outcomes following all-SCT. The present study had some limitations. Firstly, due to the retrospective nature of the study design, available data were heterogeneous and incomplete. Second, the monitoring of IR data using PB sample was performed after 100 day from allo-SCT, not within 100 days nor prior the allo-SCT. To understand further kinetics and impacts of IR, it is important to acknowledge the prior phase and earliest phase of these data in allo-SCT. Third, comparison of patients with CMV reactivation prophylaxis represented an important issue. Of note, in our data, none of the patients had a CMV prophylaxis agent. Further prospective studies are needed to compare patients with/without CMV prophylaxis and evaluate their clinical outcomes. Finally, a significant amount of clinical information was lacking, including: severe infection (Bacterial/fungus/viral including CMV infection), late onset of CMV reactivation/recurrence, use of corticosteroids, and acute GVHD details. The limited data indicated corticosteroid use and a MAC regimen were associated with lower median values of NK cell subsets (data not shown). The present study revealed particular clinical outcomes in patients with CMV reactivation after allo-SCT, focusing on immune reconstitution, especially on terminal matured CD16+CD57- NK cells.

In conclusion, our study results demonstrate that CMV reactivation represented an adverse event for post-transplant outcomes, especially in NRM. Furthermore, NK cell reconstitution may contribute to improved transplantation outcomes, especially in subgroups of CMVR+ patients.

Acknowledgement: This presented study was integrated assets by all patients with hematological malignancy and physician’s effort of the clinical front line. The authors thank all patients that participated in this study; and staffs who gathered fresh blood samples for the purely saved as huge data. The data have been preserved within the Kanagawa Cancer Center. For long-term survival observation, this study patients were followed-up main by Yokohama Cooperative Study Group for Hematology (YACHT). Contribution: T.A. and H.K. contributed conception of the paper; T.A. designed and data collected, analyzed the data and wrote the manuscript; T.T., MT., and HK. developed analysis and supervised the manuscript; All authors contributed to manuscript revision, read and approved the submitted version.

Conflict of interests: The authors report no conflict of interest.

References 1. Copelan E.A. Hematopoietic stem-cell transplantation. N. Engl. J. Med. 2006; 354: 1813-1826. 2. Seggewiss R, Einsele H. Immune reconstitution after allogeneic transplantation and expanding options for immunomodulation: an update. Blood. 2010; 115(19): 3861-3868. 3. Ogonek J, Kralj Juric M, Ghimire S, et al. Immune reconstitution after allogeneic hematopoietic stem cell transplantation. Front Immunol. 2016; 7: 507. 4. Nicolle H. R. Litjens, Lotte van der Wagen, Jurgen Kuball and Jaap Kwekkeboom. Potential beneficial effects of cytomegalovirus infection after transplantation. Front Immunol. 2018; 9: 389. 5. Takenaka K, Nishida T, Asano-Mori Y, et al. Cytomegalovirus reactivation after allogeneic hematopoietic stem cell transplantation is associated with a reduced risk of relapse in patients with acute myeloid leukemia who survived to day 100 after transplantation: The Japan Society for Hematopoietic Cell Transplantation Transplantation-related Complication Working Group. Biol. Blood Marrow Transplant. 2015; 21: 2008-2016. 6. Bartelink IH, Belitser SV, Knibbe CA, et al. Immune reconstitution kinetics as an early predictor for mortality using various hematopoietic stem cell sources in children. Biol. Blood Marrow Transplant. 2013; 19(2): 305-313.

7. Tian DM, Wang Y, Zhang XH, et al. Rapid recovery of CD3+CD8+ T cells on day 90 predicts superior survival after unmanipulated haploidentical blood and marrow transplantation. PLoS One. 2016; 11(6): e0156777. 8. Storek J, Wells D, Dawson MA, Storer B, Maloney DG. Factors influencing B lymphopoiesis after allogeneic hematopoietic cell transplantation. Blood. 2001; 98(2): 489-491. 9. Minculescu L, Marquart HV, Friis LS, et al. Early natural killer cell reconstitution predicts overall survival in T cell replete allogeneic hematopoietic stem cell transplantation. Biol. Blood Marrow Transplant. 2016; 22(12): 2187-2193. 10. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus conference on acute GVHD grading. Bone Marrow Transplant. 1995;15(6): 825-828. 11. Filipovich AH, Weisdorf D, Pavletic S, et al. National Institutes of Health consensus development project on criteria for clinical trials in chronic graftversus-host disease: I. Diagnosis and staging working group report. Biol. Blood Marrow Transplant. 2005; 11(12): 945-956. 12. Ljungman P, Griffiths P, Paya C, et al. Definitions of cytomegalovirus infection and disease in transplant recipients. Clin. Infect. Dis. 2002; 34: 1094-1097.

13. Holtan SG, DeFor TE, Lazaryan A, et al. Composite end point of graft-versus-host disease-free, relapse-free survival after allogeneic hematopoietic cell transplantation. Blood. 2015; 125(8): 1333-8. 14. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. Am. Stat. Assoc. J. 1958; 53: 457-481. 15. Cox DR. Regression models and life tables. J R Stat Soc Series B Stat Methodol. 1972; 34: 187-220. 16. Kanda Y. Investigation of the freely-available easy-to-use software “EZR” (Easy R) for medical statistics. Bone Marrow Transplant. 2013; 48: 452-458. 17. Margaret L. Green, Wendy M. Leisenring, Hu Xie, et al. CMV reactivation after allogeneic HCT and relapse risk: evidence for early protection in acute myeloid leukemia. Blood. 2013; 122: 1316-1324. 18. Kheav VD, Busson M, Scieux C, et al. Favorable impact of natural killer cell reconstitution on chronic graft-versus-host disease and cytomegalovirus reactivation after allogeneic hematopoietic stem cell transplantation. Haemtologica. 2014; 99(12): 1860-1867. 19. van der Heiden P, Marijt E, Falkenburg F, Jedema I. Control of cytomegalovirus viremia after allogeneic stem cell transplantation: a review on CMV-specific T cell reconstitution. Biol Blood Marrow Transplant. 2018; 24:1776-1782.

20. Blyth E, Withers B, Clancy L, Gottlieb D. CMV-specific immune reconstitution following allogeneic stem cell transplantation. Virulence. 2016; 7(8): 967-980. 21. Chen K, Cheng MP, Hammond SP, et al. Antiviral prophylaxis for cytomegalovirus infection in allogeneic hematopoietic cell transplantation. Blood Adv. 2018 28; 2(16): 2159-2175. 22. Marty FM, Ljungman P, Chemaly RF, et al. Letermovir Prophylaxis for Cytomegalovirus in Hematopoietic-Cell Transplantation. N Engl J Med. 2017; 377(25): 2433-2444. 23. Ljungman P, Brand R, Einsele H, et al. Donor CMV serologic status and outcome of CMV-seropositive recipients after unrelated donor stem cell transplantation: an EBMT megafile analysis. Blood. 2003 ; 102(13): 4255-4260. 24. Ramanathan M, Teira P, Battiwalla M, et al. Impact of early CMV reactivation in cord blood stem cell recipients in the current era. Bone Marrow Transplantation 2016; 51(8):1113-1120. 25. Stern L, Withers B, Avdic S, et al. Human Cytomegalovirus Latency and Reactivation in Allogeneic Hematopoietic Stem Cell Transplant Recipients. Front Microbiol. 2019; 10: 1186. 26. Cichocki F, Cooley S, Davis Z, et al. CD56dimCD57+NKG2C+ NK cell expansion is associated with reduced leukemia relapse after reduced intensity HCT. Leukemia. 2016 ; 30(2): 456-463.

27. Cook CH. Cytomegalovirus reactivation in "immunocompetent" patients: a call for scientific prophylaxis. J. Infect. Dis. 2007; 196: 1273-1275. 28. Ito S, P Pophali, wu Co, E K Koklanaris, et al. CMV reactivation is associated with a lower incidence of relapse after allo-SCT for CML. Bone Marrow Transplant. 2013; 48: 1313-1316. 29. Zdziarski P. CMV-specific immune response-new patients, new insight: central role of specific igg during infancy and long-lasting immune deficiency after allogenic stem cell transplantation. Int. J. Mol. Sci. 2019; 20(2): E271. 30. Keating AK, Langenhorst J, Wagner JE, et al. The influence of stem cell source on transplant outcomes for pediatric patients with acute myeloid leukemia. Blood Adv. 2019; 3(7): 1118-1128. 31. Tanaka M, Kobayashi S, Numata A, Tachibana T, Takasaki H, Kanamori H. The impact of the dose of natural killer cells in the graft on severe acute graft-versus-host disease after unrelated bone marrow transplantation. Leuk. Res. 2012; 36(6): 699-703. 32. Maggs L, Kinsella F, Chan YT, et al. The number of CD56 dim NK cells in the graft has a major impact on risk of disease relapse following allo-HSCT. Blood Adv. 2017; 1(19): 1589-1597.

Tables: Table 1. Patient characteristics, based on the presence of CMV reactivation and NK cell reconstitution. Table 1

CMV reactiv ation Charact eristics Age, years

NK cell reconsti tution

Tota CMVR l -

CM VR+

medi an

51 (18-

(rang

69)

44 (18-69)

P

earl y IR

dela yed IR

P

52 <0.001 (18-6

49 (18-

47 (18-

0.0 76

8)

69)

67)

62

59

69

56

77

45

54

37

128

104

13

9

e)

Sex

< 50

121 (49. 2)

68

53

≥ 50

125 (50. 8)

37

88

male

152 (61.

62

90

<0.001

0.507

0.6 09

0.7 85

8)

PS

fema le

94 (38. 2)

43

51

0-1

222 (90. 2)

96

126

≥2

22 (8.9)

9

13

unkn own

2 (0.9)

CMV

R+/D 75

serostatu s

+

（

R+/D 124

0.895

2 30

45

0.9 21

2 <0.001

36

39

(30. 5)

0.0 59

41

83

75

49

recipient/

-

(50. 4)

R-/D

19

+

(7.7)

R-/D -

27 (11. 0)

unkn

1

own

(0.4)

donor）

Disease

13

6

8

11

23

4

11

16

1

AML 131 (53. 4)

55

76

ALL

70

61

29

32

42

19

MDS 32 (13. 0)

10

22

10

22

CML 7

5

2

5

2

MPN 7 (2.8)

1

6

2

5

ML

3

3

2

4

2

0

0

2

77

89

98

68

33

47

53

46

61(2

0.165

0.2 52

4.8)

(2.8)

6 (2.4)

PMF

2 (0.8)

Disease risk

Stan dard

166 (67.

0.1

0.0 093

5) High

80 (33. 5)

28

52

Conditio

MA

99

53

46

ning

C

(40. 2)

RIC

147 (59. 8)

52

95

BM

98

45

53

Graft

0.006

0.0 63

0.011

78

69

41

57

<0.

source

GVHD prophyla xis

(39. 8) PBS

36

C

(14. 6)

CB

001 22

14

13

23

112 (45. 5)

38

74

77

35

CSP base

57 (23. 2)

44

13

26

31

TAC base

189 (76.

61

128

105

84

grade 148 0-1 (60. 2%)

84

64

88

60

grade 98

21

77

43

55

0.159

0.0 53

8) Acute GVHD

2-4

<0.001

0.0 19

(39. 8%)

CMVR: CMV reactivation following allogeneic stem cell transplantation, CMV serostatus: R for recipient, D for donor. NK cell reconstitution: CD16+CD57- NK cell subset was focused, IR: immune reconstitution. PS: performance status (ECOG). AML: acute myeloid leukemia, ALL: acute lymphoblastic leukemia, MDS: myelodysplastic syndrome. CML: chronic myeloid leukemia, MPN: myeloproliferative neoplasm, ML: malignant lymphoma. PMF: primary myelofibrosis, GVHD: graft-versus-host disease, CSP: cyclosporine, TAC: tacrolimus. MAC: myeloablative conditioning, RIC: reduced intensity conditioning, BM: bone marrow, PBSC: peripheral blood stem cell, CB: cord blood.

Table 2. Prognostic factor for OS, NRM, and CIR. OS

N

Multivaria

P

te analysis

NR

CI

M

R Multivaria

P

te analysis

Multivaria

P

te analysis

CMV recipient seropositivit y No

46

1

0.17

1

7 Yes

Unknow

0.09 2

19

0.48

1.75

9

(0.17-1.38

(0.78-9.02

)

)

1

n Graft sources BM/PBS

13

1

0.585

1

C

4

CB

11

0.85

0.39

2

(0.46-1.54

(0.11-1.40

)

)

0.15 1

Disease Risk Standard

16

1

0.006

1

6

＜ 0.00 1

Hiｇｈ

80

2.07

2.24

(1.28-4.39

(1.50-4.83

)

)

CMV reactivation CMVR-

10

1

0.019

1

5 CMVR+

NK cell

0.01

1

1

0.31 9

14

1.88

3.25

1.17

1

(1.21-3.66

(1.42-6.95

(0.79-2.22

)

)

)

(CD16+CD5 7 -) High

12

1

6 Low

<0.00

1

1

0.10

1

3

0.03 1

12

2.16

2.21

1.92

0

(1.45-4.18

(0.72-6.85

(0.06-3.84

)

)

)

OS: overall survival, NRM: nonrelapse mortality, CIR: cumulative incidence of relapse. BM: bone marrow, PBSC: peripheral blood stem cell, CB: cord blood, CMVR: cytomegalovirus reactivation

Table 3. Causes of death CMV reactivation

Disease progression

N=246

CMVR(n=105)

CMVR+ P (n=141)

56 (22.8%)

17 (16.2%)

39 (27.7%)

0.045

Unbalanced Immunity

19 (7.7%)

3 (2.9%)

16 (11.3%)

0.015

16 (6.5%)

3 (2.9%)

13

0.065

NRM

Maladaptive immunity

(9.2%)

GVHD

6 (2.4%)

3 (2.9%)

3 (2.1%) 0.702

BO

2 (0.8%)

0

2 (1.4%) 0.509

OP

8 (3.2%)

0

8 (5.7%) 0.012

Infection

3 (1.2%)

0

3 (2.1%) 0.263

Secondary cancer

2 (0.8%)

1 (0.95%)

1 (0.7%) 0.993

TMA

1 (0.4%)

0

1 (0.7%) 0.961

VOD

1 (0.4%)

0

1 (0.7%) 0.961

2 (0.8%)

0

2 (1.4%) 0.509

Deficient Immunity

Treatment completion

Hemomorrhage

One patient had confirmed CMV pneumonia in pathological autopsy. Two of three patients with CMVR+ who had died by infection had been receiving treatment for GVHD. One of two patients with CMVR+ who experienced hemorrhage was being treated for pneumonia at the time of death.

Figure captions Figure 1. Transplant outcomes according to A, CMV reactivation, and B, CD16+CD57- NK cell reconstitution. The survival rates were at five years after allo-SCT. There was no significant correlation with the presence of CMV reactivation and CD16+CD57- NK cell IR on days 100 from allo-SCT, especially in median cell counts and basement characteristics. The overall survival benefit was found to be almost 20% higher in patients without CMV reactivation. Patients who had CMV reactivation experienced five times higher NRM after allo-SCT. The indirect effect of CMV reactivation, such as disease suppression, was not observed. A significant difference in survival outcomes, according to CD16+CD57- NK cell reconstitution, was observed for all patients. Figure 2. Dual impacts of CMV reactivation CD16+CD57- NK cell IR were found after allo-SCT. CD16+CD57- NK cell subset significantly contributed to a lower cumulative incidence of relapse. NK cell IR might have improved post-transplant outcomes, especially in patients with CMV reactivation. Figure 3. Survival curves according to CD16+CD57- NK cell reconstitution in patients with CMV reactivation. A) OS, B) NRM, C) CIR, D) chronic GVHD, and E) chronic GRFS. Outcomes were shown in subgroup patients with CMV reactivation. Outcomes without CMV reactivation were provided in Supplemental Table. A late onset of survival events was adversely observed in patients with delayed CD16+CD57- NK cell reconstitution. For delayed

NK cell IR, only 20% of patients with CMV reactivation experienced free events from chronic GVHD and disease recurrence (P < 0.001). These findings explain the fact that a lack of balanced immune reconstitution might induce both disease relapse events and maladaptive immunity as GVHD related events as well. Timely NK cell reconstitution was required for patients with CMV reactivation for late phase of survival events. Thus, CMV prophylaxis might be required for long term, event-free survival.