We selected 40 allo-HSCT recipients from a local cohort (https://clinicaltrials.gov/ct2/show/NCT03642977), including 20 positive for HPgV-1 by rRT-PCR assay on all of their analysed timepoints and 20 negative for HPgV-1 by rRT-PCR assay on all of their analysed time points: among 20 HPgV-1-viremic patients, 20 had rRT-PCR assay performed at day 0/-7, day 30, 3, and 6 months, and 12 had rRT-PCR assay performed at 12 months. Among 20 HPgV-1-non viremic patients, 20 had rRT-PCR assay performed at day 0/-7, day 30, 3, and 6 months, and 15 had rRT-PCR assay performed at 12 months and one patient had rRT-PCR assay performed at 18 months. Patients were matched according to criteria that can influence the immune reconstitution, namely diagnosis, conditioning, graft source, T cell depletion, TBI, donor type and CMV status ( Table 1 ). 20 healthy donors from the Geneva University Hospitals blood transfusion center were also analyzed as a control group.

The study was approved by the cantonal ethic committee (CCER 2019-01153) and patients and healthy controls (HC) gave their written informed consent.

190 μl of plasma were spiked with 10 μl of standardized canine distemper virus of known concentration and extracted with the NucliSENS easyMAG (bioMérieux, Geneva, Switzerland) nucleic acid kit in a 25 ul elution volume, according to the manufacturer’s instruction. Extracted RNA was used for HPgV-1-specific rRT-PCR screening analysis, using a previously published assay (21). rRT-PCR assay reaction was performed using the QuantiTect Probe RT-PCR Kit (Qiagen, Valencia, CA. USA) on a StepOnePlus instrument (Applied Biosystems, Rotkreuz, Switzerland) under the following cycling conditions: 50°C for 30 min; 95°C for 15 min; 45 cycles of 15 s at 94°C; and 1 min at 55°C. Data were analysed with the StepOne software V.2 (Applied Biosystems). Analytical sensitivity was assessed with a plasmid-derived transcribed RNA including the target region (kindly provided by Professor JT Stapleton) and showed a lower limit of quantification (LLOQ) = 2.6E3 copies/mL.of plasma.

Peripheral blood mononuclear cells were isolated from EDTA anti-coagulated peripheral blood obtained at 1, 3, 6 and 12 months post-transplantation using Ficoll. Cells were cryopreserved and stored in liquid nitrogen until analysis. After thawing, cryopreserved PBMC were stained with monoclonal antibodies specific for the following antigens: CCR7 (FITC, clone 150503, R&D Systems), CD127 (BV786, clone HIL-7R-M21, BD Biosciences), CD16 (BV480 and PerCPCy5.5, clone 3G8, BD Biosciences), CD25 (APC and PECy7, clone M-A251, BD Biosciences), CD27 (APC-R700, clone M-T271, BD Biosciences), CD27 (BV786, clone L128, BD Biosciences), CD3 (BV711 and PerCPCy5.5, clone UCHT1, BD Biosciences), CD38 (BV605, clone HB7, BD Biosciences), CD4 (BV480, clone SK3, BD Biosciences), CD4 (BV650, clone L200, BD Biosciences), CD45RA (BUV395 and PE, clone HI100, BD Biosciences), CD56 (BV421, clone HCD56, Biolegend), CD57 (BV605, clone QA17A04, Biolegend), CD57 (PECF594, clone NK-1, BD Biosciences), CD8 (BUV737, clone SK1, BD Biosciences), CD8 (BV711, clone RPA-T8, BD Biosciences), HLADR (PECy7, clone L243, BD Biosciences) and PD1 (PE, clone EH12.2H7, Biolegend).

Intracellular staining for cytotoxic molecules and transcription factors was performed overnight at 4°C on fixed and permeabilized cells with FoxP3/transcription factor staining buffer set (e-Bioscience) using antibodies against Eomes (eFluor 660, clone WD1928, eBiosciences), Foxp3 (PECF594, clone 236A/E7, BD Biosciences), GranzymeB (AF700, clone GB11, BD Biosciences), Perforin (FITC, clone B-D48, Diaclone) and Tbet (PE, clone ebio4B10, eBiosciences).

Data were acquired on a BD LSRFortessa Cell Analyzer (BD Biosciences) and analyzed with FlowJo software (FlowJo LLC). Subsets of CD4 and CD8 were defined according to CD45RA, CD27 and CCR7 expression as follows: Naïve T cells (TN) CD45RA+/CD27+, central memory T cells (TCM) CD45RA-/CCR7+, effector memory T cells (TEM) CD45RA-/CCR7- and effector memory re-expressing CD45RA T cells (TEMRA) CD45RA+/CD27- ( Supplemental Figure 1 ). Subsets of NK cells were defined based on CD56 and CD16 expression as immature CD56 bright and mature CD56 dim NK cells as shown in Supplemental Figure 1 .

Stem cell factor (SCF), IL-7 and IL-15 were quantified in plasma samples using the the LEGENDplex™ Human Hematopoietic Stem Cell Panel (Biolegend). 50 ul of plasma from HSCT patients and healthy controls were diluted 1:1 and the assay was performed according to manufacturer’s instruction. All samples were tested in duplicate. The samples were acquired on a Attune NxT Flow cytometer (Invitrogen) and data were analyzed using the LEGENDplex™ Data Analysis Software.

X² or Fischer exact tests were used for categorical variables. Mann-Whitney test or Wilcoxon matched pairs signed rank test were used for continuous variables. A P value <.05 was considered statistically significant. Statistics were performed using R version 3.2.0 and R Studio version 1.3.1056.

We first assessed the potential impact of HPgV-1 replication on absolute numbers of CD4 and CD8 T cells as well of NK cells at different time-points after allo-HSCT. Patients were selected based on HPgV-1 replication and defined as HPgV-1 positive (HPgV-1+) or negative (HPgV-1-) based on HPgV-1 detection at all or none studied time-points during the first year after transplantation ( Figure 1A ). As expected, the majority of allo-HSCT recipients displayed persistent CD4 T cell lymphopenia up to 1-year post-transplantation while most of them reconstituted normal levels of CD8 T cells and NK cells by 3-months and 1-month post-allo-HSCT respectively ( Figure 1B ). We did not detect any significant differences between HPgV-1+ and HPgV-1- patients in terms of CD4 and CD8 T cell absolute numbers up to 1-year after allo-HSCT ( Figure 1B , upper and middle panel). In contrast, HPgV-1+ patients displayed significantly higher numbers of NK cells (median 267 (range 51-685) cells/ul) at 3 months after allo-HSCT compared with HPgV-1- patients (119 (26-725) cells/ul, p=0.02; Figure 1B , lower panel). Collectively, these data reveal minimal impact of HPgV-1 replication on absolute numbers of major lymphocytes subsets recovered after transplantation with only a slight but significant difference in absolute numbers of NK cells recovered at 3 months after allo-HSCT.

In addition of quantitative defects, allo-HSCT recipients display major qualitative abnormalities in lymphocyte subsets during immune-reconstitution, namely a skewing in T cell phenotype and repertoire (15) and an immature phenotype in NK cells (22, 23). To evaluate the potential impact of HPgV-1 replication on lymphocyte reconstitution, we next compared the immunophenotype of T and NK cells after transplant in HPgV-1+ and HPgV-1- recipients. Dimensionality reduction of FACS data using Uniform Manifold Approximation and Projection easily identified major T and NK cell subsets in both healthy controls and allo-HSCT recipients ( Figures 2A, B ). As predicted, allo-HSCT recipients displayed a significant increase in less mature CD56^bright NK cells and a significant decrease of naïve CD4 and CD8 T cells compared to healthy controls ( Figure 2B ). After stratification of allo-HSCT recipients based on HPgV-1, we detected higher proportions of less differentiated CD56^bright NK cells mirrored by lower percentages of mature CD56^dim NK cells in HPgV-1+ compared to HPgV-1- allo-HSCT recipients ( Figure 2C , left panels). Such difference persisted at each time point studied until 6 months after transplantation ( Figure 2C , left panels). We did not observe any differences in CD8 or CD4 T cell subset distribution (naïve, central memory, effector memory, TEMRA) between HPgV-1+ and HPgV-1- allo-HSCT recipients ( Figure 2C , middle and right panels).

To better assess the relationship between HPgV-1 viremia and the increase of less differentiated CD56^bright NK cells, we compared absolute numbers of CD56^bright and CD56^dim NK cells over the first year of transplantation in HPgV-1+ and HPgV-1- allo-HSCT recipients. As shown in Figure 3 , this analysis revealed a significant increase in CD56^bright NK cell numbers in HPgV-1+ compared to HPgV-1- negative patients at 3, 6 and 12 months after transplantation ( Figure 3 , upper panels) while we did not observe any differences in CD56^dim NK cell numbers between the two patient groups.

Collectively, these results indicate that HPgV-1 replication after allo-HSCT was associated with further imbalance of CD56^bright/CD56^dim NK cell distribution due to an increase in CD56^bright NK cells during the first year after HSCT in HPgV-1+.

NK cell subsets reconstitution after allo-HSCT is governed by the availability of homeostatic cytokines regulating their proliferation and survival, namely stem cell factor (SCF) and IL-7 for CD56^bright NK cells and IL-15 for CD56^dim NK cells. We hypothesized that HPgV-1 replication might influence NK cell subsets reconstitution by affecting the levels of homeostatic cytokines. To test this hypothesis, we measured the plasmatic levels of SCF, IL-7 and IL-15 at day 30 after allo-HSCT. No difference was observed in SCF levels between HPgV-1-viremic and non-viremic patients as well as between either patient group and healthy controls ( Figure 4 , left panel). Interestingly, HPgV-1-viremic allo-HSCT recipients displayed significantly higher IL-7 levels compared to both HPgV-1-non-viremic patients and healthy controls ( Figure 4 , middle panel). Both HPgV-1-viremic and HPgV-1-non-viremic allo-HSCT recipients displayed reduced levels of IL-15 compared with healthy controls while we did not detect any significant difference between the two patient groups ( Figure 4 , right panel). Analysis performed at later time points (3, 6 and 12 months) did not detect any significant difference in SCF, IL-7 and IL-15 levels between HPgV-1-viremic and HPgV-1-non-viremic allo-HSCT recipients (data not shown). These data show an association between HPgV-1 replication and IL-7 levels in plasma 30 days after allo-HSCT.

To gain further insights into the phenotypic abnormalities of NK cells during immune-reconstitution after allo-HSCT in the presence of HPgV-1 replication, we next measured the expression of additional markers reflecting NK cell differentiation at 6 months after allo-HSCT. According to their less differentiated phenotype, we observed a significant reduction in the proportion of NK cells expressing the surface Fcγ receptor CD16 in HPgV-1+ patients (58% (16–74)) compared with HPgV-1- individuals (70% (23-90); p=0.011 Figure 5A ). In addition, NK cells from HPgV-1+ allo-HSCT recipients displayed significantly reduced proportion (24% (3-76) vs 45% (8-85) in HPgV-1- patients; p=0.0045) of cells expressing the N-CAM family molecule CD57, believed to identify cells at final stages of peripheral NK cell maturation (24) ( Figure 5A ). Looking at expression of CD27, a molecule employed to identify less differentiated murine and human NK cells, we observed a trend not reaching statistical significance toward an increase in CD27+ NK cells in HPgV-1+ allo-HSCT recipients 18% (3-55) vs 12% (0.5-21) in HPgV-1- patients; p=0.078; Figure 5A ). Finally, we analysed the expression of PD-1, a well-established marker of T cell activation and exhaustion whose expression on NK cells is controversial. According to recent reports (25), with the only exception of one patient, we detected minimal PD-1 expression at NK cell surface of either HPgV-1+ or HPgV-1- allo-HSCT recipients ( Figure 5A ). When CD56^bright and CD56^dim NK cell subsets were analysed separately, the reduction in CD16 and CD56 in HPgV+ recipients was observed in the mature CD56dim NK cell subset while no difference was detected in CD56bright NK cells which, as expected, expressed only low level of these differentiation markers ( Figure 5B ). Collectively, this phenotypic analysis supports an association between HPgV-1 replication and an immature NK cell phenotype during reconstitution after allo-HSCT.

NK cell-mediated cytotoxicity plays a major role in anti-infectious and anti-tumor immunity after allo-HSCT. We hypothesized that the impaired NK cell phenotype observed in HPgV-1+ allo-HSCT recipients might be associated with a reduced capacity to produce cytotoxic molecules. At 6 months after allo-HSCT, we observed no differences in the proportions of perforin producing NK cells between HPgV-1+ and HPgV-1- patients ( Figure 6A , upper panels). Conversely, we observed a significant reduction in cells expressing the cytotoxic molecule granzyme B among NK cells recovered from HPgV-1+ allo-HSCT recipients (58% (22-78)) compared to cells from HPgV-1- patients (76% (48-97); p=0.0045; Figure 6A , lower panels). When CD56^bright and CD56^dim NK cell subsets were analysed separately, we did not observe any difference in perforin or granzyme B expression in NK cell subsets depending on HPgV-1 replication ( Figure 6B ), indicating that the reduced expression of these molecules detected at the global NK cell population level were mainly due to differences in the CD56^bright/CD56^dim NK cell ratios. Collectively, our analysis reveals that HPgV-1 replication was associated with an immature NK cell phenotype resulting in a decreased expression of the cytotoxic molecule granzyme B.