Inflammatory Cytokine Profile in Individuals with Inherited Chromosomally Integrated Human Herpesvirus 6.

Abstract

Acute graft-versus-host-disease (aGVHD) is a major complication following hematopoietic cell transplantations (HCT). We have shown that HCT recipients in which either the donor or patient had inherited chromosomally-integrated human herpesvirus 6 (iciHHV-6) have a higher incidence of developing more severe aGVHD. Previous studies established that increased proinflammatory cytokines are associated with increased risk for aGVHD and non-relapse mortality post-HCT. We hypothesized that HCT recipients with donor or recipient iciHHV-6 (iciHHV-6^pos HCT cases) will have higher cytokine levels compared to HCT recipients without iciHHV-6 (iciHHV-6^neg HCT controls). We identified 64 iciHHV-6^pos HCT cases with plasma from days 7, 14, and/or 21 post-HCT and before aGVHD onset in patients who developed aGVHD. We identified 64 iciHHV-6^neg HCT controls matched for aGVHD risk factors. We also identified 28 donors with iciHHV-6 and 56 matched donors without iciHHV-6. We measured plasma cytokine concentrations for IL-6, ST2, TIM3, TNFα, TNFRp55, and CRP. We used Mann-Whitney tests and repeated measures models to compare cytokine levels. iciHHV-6^pos HCT cases had higher CRP levels on day 7 and day 21 and higher TNFRp55 levels on day 14 and day 21 compared to iciHHV-6^neg HCT controls. These findings were recapitulated in a repeated measures model. The differences were most evident among patients who subsequently developed aGVHD grades 2-4. Additionally, iciHHV-6^pos HCT cases had earlier-onset aGVHD (median, 20 versus 27 days post-HCT; p=0.02). There were no differences in cytokine levels among healthy donors with or without iciHHV-6. This study demonstrates that HCT recipients with iciHHV-6 have higher proinflammatory cytokines that may be associated with increased risk for aGVHD.

INTRODUCTION

Human Herpesvirus 6 (HHV-6) is a Roseolovirus that comprises two distinct species, HHV-6A and HHV-6B¹. While HHV-6B is known to cause roseola infantum (exanthema subitum or sixth disease), a generally mild illness characterized by fever and rash², the pathology of HHV-6A is largely unknown. Most children are infected by HHV-6B by the age of 3 years, and the seropositivity in the general population reaches up to 95%³. The virus can infect a variety of cells including CD4⁺ T-lymphocytes³. HHV-6B reactivation is seen in 30% - 50% of allogeneic hematopoietic cell transplant (HCT) recipients and is the most common cause for viral encephalitis in this setting, as well as a risk factor for the development of GVHD⁴. HHV-6B reactivation is also associated with the development of acute graft-versus-host disease (aGVHD), cytopenias, cytomegalovirus (CMV) reactivation, and an increased overall mortality⁵.

Unlike other herpesviruses, HHV-6 species establishes latency through integration of the HHV-6 genome into telomeric regions⁶. Integration into germ cells is possible and leads to offspring that carry the viral genome in every nucleated cell of the body⁷. This condition is termed inherited chromosomally integrated HHV-6 (iciHHV-6) and occurs in approximately 1% of the human population⁸. It has been shown that the virus is able to reactivate from this state in immunosuppressed patients^9-11, but the consequences of this condition are largely unknown.

Acute graft-versus-host-disease (aGVHD) is a common complication after HCT that contributes to substantial morbidity and mortality^12-14. The pathological mechanisms are still not fully understood. We previously performed a large-scale screening study in which we identified 87 HCTs in which either the donor, recipient, or both were confirmed to have iciHHV-6. These patients had significantly increased risk for aGVHD grades 2-4 and CMV reactivation compared to unaffected HCT recipients¹⁵. Several pro-inflammatory cytokines, including T-cell immunoglobulin and mucin-domain containing-3 (TIM3), interleukin 6 (IL-6), tumor necrosis factor alpha (TNFα), soluble TNF receptor 1 (TNFRp55), suppression of tumorigenicity 2 (ST2), and C-reactive protein (CRP) have been identified as early biomarkers associated with aGVHD severity and increased risk of non-relapse mortality and death in HCT recipients^16-19. CRP, TNFα, IL-6, and ST2 have also been associated with increased risk for heart diseases in non-immunocompromised populations^20-22. Interestingly, a large population-based study found an association between iciHHV-6 and angina pectoris²³. To explore a potential pathophysiological pathway through which iciHHV-6 may increase the risk for aGVHD and other outcomes such as heart disease, we tested plasma samples for TIM3, IL-6, TNFα, TNFRp55, ST2, and CRP at multiple post-HCT time-points in a cohort of HCT recipients and healthy stem cell donors with and without iciHHV-6.

METHODS

The protocol was approved by the Fred Hutchinson Cancer Research Center (Fred Hutch) Institutional Review Board.

Individuals and specimens

We previously performed a large-scale study in which we screened 4,319 HCT donor-recipient pairs from 1992-2013 for iciHHV-6¹⁵. Briefly, we used a specimen pooling strategy²⁴ to test genomic DNA of individuals through quantitative polymerase chain reaction and confirmed the results in a subset using digital droplet PCR^25,26. Using this approach, we identified 87 individuals (recipients and/or their donors) with iciHHV-6.

In this manuscript, we report the results of two cohort studies. Cohort study 1 consists of HCT recipients in which either the donor, recipient, or both were iciHHV-6^pos (iciHHV-6^pos HCT cases). Controls were selected from HCT recipients in which neither donor nor recipient were iciHHV-6^pos (iciHHV-6^neg HCT controls) and were matched 1:1 with the iciHHV-6^pos HCT cases for transplant year (before 2000 vs. on and after 2000), conditioning regimen (myeloablative vs. non-myeloablative), HLA-matching (matched vs. mismatched), and recipient-donor relation (related vs. unrelated). From these patients, we identified individuals for whom we had ≥1 plasma samples from days 7, 14, and 21 (+/− 3d) post-HCT and before aGVHD onset, as relevant; samples were stored at −80°C in the Fred Hutch Infectious Disease Sciences biorepository. In order to avoid biasing the cohort with healthier individuals by excluding patients who died early after HCT, patients were considered for matching if they met all criteria and had at least 1 available sample.

Cohort study 2 consists of 28 stem cell donors with iciHHV-6 (iciHHV-6^pos donor cases) and 56 donors without iciHHV-6 (iciHHV-6^neg donor controls) matched for age (+/− 4 years), sex, and race. We tested one plasma sample per individual in Cohort study 2.

Clinical data

Clinical data were abstracted from medical records and research databases.

Cytokine assay

Plasma samples were batch tested for TIM3, TNFRp55, ST2, IL-6, TNFα, and CRP levels using the Luminex multiplex assay (Luminex, Austin, Texas), as per the company’s specifications. Briefly, samples and standards were incubated with Luminex microbeads (one unique bead population per cytokine) coated with cytokine-specific antibodies. Beads were washed and incubated with biotinylated cytokine antibodies, washed again and then incubated with phycoerythrin-streptavidin conjugate. Samples and standards were analyzed for fluorescence intensity using a Luminex 200 instrument. The measured intensity was proportional to the cytokine concentration. Standard curves were generated for each cytokine from the standard read out and sample concentrations calculated from these curves. The limit of detection for each assay is detailed in Table S1. Testing was performed by technicians blinded to individual iciHHV-6 status and clinical data.

Statistical analysis

Matching criteria for each time-point were compared using Fisher’s exact test. Median values for each cytokine were compared at day 7, 14, and 21 time-points between iciHHV-6^pos HCT cases and iciHHV-6^neg HCT controls within each cohort using Mann-Whitney tests. The data were log₁₀-transformed for presentation of the results. If a value was below the limit of detection (LOD), it was assigned a value of the LOD/2. We performed exploratory analyses in which we stratified these comparisons by aGVHD severity or the iciHHV-6 source (donor, recipient, or both). Fisher’s exact tests were performed using OpenEpi²⁷. Competing risk analysis using Gray’s test was performed and cumulative incidence curves were created using R²⁸ and the Cumulative Incidence Curves in R package²⁹. Mann-Whitney tests were performed and box plot figures were created using Graphpad Prism version 8.0 for Windows, GraphPad Software, La Jolla California USA, www.graphpad.com. To evaluate comparisons between cases and controls across all time points, we also fit linear models with generalized estimating equations (GEE) and robust variances to account for repeated measures across time. Models evaluated the impact of day of testing and case versus control status along with interaction terms between day of testing and case versus control status to determine whether cytokines had different slopes across time. Similarly, aGVHD severity was tested as a main effect and interaction with case versus control status. Based on our a priori power calculations for Cohort 1, the minimum estimated effect sizes for the comparison of mean cytokine values between iciHHV-6^pos cases and iciHHV-6^neg controls that we would be powered to detect at each time point (based on an anticipated 77 cases with 1:1 matching) ranged from 0.47-0.51 of a standard deviation difference in means with α=0.05 and 80% power. For Cohort 2, the minimum estimated effect size that we would be powered to detect (based on an anticipated 30 cases with 1:2 matching) was 0.63 of a standard deviation difference in means. Because we did not adjust for multiple comparisons, P values between 0.05 and 0.01 should be interpreted with caution.

RESULTS

HCT Recipient Characteristics: Cohort Study 1

Cohort 1 consisted of 64 iciHHV-6^pos HCT cases for which either patient (n=36), donor (n=19), or both (n=9) were iciHHV-6^pos (HHV-6A, n=19; HHV-6B, n=45). These cases were matched with 64 iciHHV-6^neg HCT controls in whom neither donor nor recipient had iciHHV-6. Fifty-eight patients had plasma available at all three time-points, 30 patients had plasma at two time-points, and 40 patients had plasma at 1 time-point. All included samples were obtained before aGVHD onset. The number of patients who developed aGVHD prior to the day 14 and day 21 timepoints are detailed in Table S2; no data were included from these post-aGVHD onset timepoints. Among patients that subsequently developed aGVHD, the median time between the last tested sample and aGVHD onset was 4 days (interquartile range [IQR], 2-7 days) in iciHHV-6^pos HCT cases and 6 days (IQR, 4-14 days) in iciHHV-6^neg HCT controls (Table S3).

The demographics and clinical characteristics of Cohort 1 can be found in Table 1 and separated by time-points in Table S4. The matching criteria were similar between iciHHV-6^pos HCT cases and iciHHV-6^neg HCT controls at each individual time-point, suggesting that variable sample availability did not skew matching on baseline risk factors for aGVHD. The GVHD prophylaxis regimens were comparable between both groups, although this was not used as a matching criterion. IciHHV-6^pos HCT cases were more likely to be male, of Caucasian race, to have an advanced underlying disease, and to be CMV seropositive (donor and/or recipient) compared to the iciHHV-6^neg HCT controls (Table 1). CMV viremia and bacteremia during the study time period were infrequent and equally distributed between cases and controls (Table S5).

Table 1.

Overall demographics and clinical characteristics of iciHHV-6^pos HCT cases (donor, recipient, or both with iciHHV-6) and iciHHV-6^neg HCT controls (neither donor nor recipient with iciHHV-6).

	iciHHV-6pos HCT Cases	iciHHV-6neg HCT Controls
Patients (available samples)	64 (112)	64 (162)
Age median (IQR)	44.4 (21-68)	43.4 (12-67)
Sex of patient
Female	22 (34)	27 (42)
Male	42 (66)	37 (58)
Race of patient
Caucasian	62 (97)	45 (70)
Other	2 (3)	19 (30)
CMV serostatus
CMV positive	49 (77)	46 (72)
CMV negative	15 (23)	18 (28)
Patient-donor relation*
related	31 (48)	31 (48)
unrelated	33 (52)	33 (52)
HLA-match*a
HLA matched	53 (83)	53 (83)
HLA mismatched	11 (17)	11 (17)
Underlying disease severityb
More advanced	27 (21)	22 (34)
Less advanced	37 (58)	42 (66)
Underlying disease
Acute leukemia	26 (41)	19 (30)
Chronic leukemia	21 (33)	23 (36)
Any lymphoma	5 (8)	4 (6)
Otherc	12 (19)	18 (28)
Conditioning regimen*d
Myeloablative	43 (67)	43 (67)
Nonmyeloablative or reduced intensity	21 (33)	21 (33)
GVHD prophylaxis regimen
Cyclosporine plus methotrexate	32 (50)	31 (48)
Cyclosporine plus mycophenolatee	9 (5)	7 (11)
Tacrolimus plus methotrexate	11 (17)	14 (22)
Otherf	12 (19)	12 (19)
Year of treatment*
Before 2000	27 (42)	27 (42)
On and after 2000	37 (58)	37 (58)

Data are presented as absolute values (percent) unless otherwise indicated. iciHHV-6^pos HCT cases and iciHHV-6 ^neg HCT controls were matched for criteria that are starred*.

HCT – hematopoietic cell transplant; iciHHV-6 – inherited chromosomally integrated human herpesvirus 6; IQR – Interquartile range; CMV – cytomegalovirus; HLA – human leukocyte antigen; GVHD – graft-versus-host disease

^aHLA-match indicates 10/10 allele match or antigen match.

^bMore advanced underlying disease refers to diagnoses other than acute myeloid leukemia, acute lymphoblastic leukemia, or lymphoma in first remission, chronic myeloid leukemia in chronic phase, and refractory anemia without excess blasts.

^cOther underlying diseases include: aplastic anemia, erythroleukemia, Chediak-Higashi syndrome, myelofibrosis, myelodysplastic syndrome, agnogenic myeloid metaplasia, multiple myeloma, paroxysmal nocturnal hemoglobulinuria, refractory anemia, refractory anemia with excess blasts, Waldenstrom’s macroglobulinemia.

^dMyeloablative regimens include: any regimen containing ≥800 cGy TBI, any regimen containing carmustine/etoposide/cytarabine/melphalan (BEAM), or any regimen containing bulsulfan/cyclophosphamide with or without antithymocyte globulin.

^eMycophenolate comprises mycophenolate motefil or mycophenolate sodium.

^fOther regimens consisted of different combinations of the listed medications. Each alternative combination contained up to three iciHHV-6^pos cases, iciHHV-6^neg controls, or both. Four patients also received post-HCT cyclophosphamide (2 iciHHV-6^pos cases and 2 iciHHV-6^neg controls).

Acute GVHD grades 2-4 occurred earlier after HCT involving iciHHV-6

The cumulative incidence of aGVHD grades 2-4 was similar in this cohort of iciHHV-6^pos HCT cases (78.1%) and iciHHV-6^neg HCT controls (73.4%; Figure 1). However, the median day of onset of aGVHD grades 2-4 after HCT occurred 1 week earlier in iciHHV-6^pos HCT cases compared to iciHHV-6^neg HCT controls (20 versus 27 days post-HCT, p=0.02; Mann-Whitney

Figure 1.

Cumulative incidence probability estimates for the onset of aGVHD grades 2-4 in iciHHV-6^pos HCT cases (HCT cases, dashed lines) and iciHHV-6^neg HCT controls (HCT controls, solid lines) within 100 days post-HCT. Non-GVHD related death was treated as a competing risk. P = 0.1 (Gray’s test).

Proinflammatory cytokines were higher after HCT involving iciHHV-6

A total of 382 post-HCT plasma samples were tested for six cytokines (TIM3, TNFRp55, ST2, IL-6, TNFα, and CRP). CRP plasma concentrations were higher in iciHHV-6^pos HCT cases compared to iciHHV-6^neg HCT controls on day 7 (7.67 versus 7.41 log₁₀pg/mL, p=0.02; Figure 2A) and day 21 (7.17 versus 6.86 log₁₀pg/mL, p=0.06; Figure 2A). For all CRP results, the levels were above the upper limit of normal (10 mg/L) in 90 of 112 (80.4%) samples from iciHHV-6^pos HCT cases compared to 95 of 162 (58.6%) samples from iciHHV-6^neg HCT controls. TNFRp55 plasma concentrations were also higher in iciHHV-6^pos HCT cases compared to iciHHV-6^neg HCT controls on day 14 (3.68 versus 3.62 log₁₀pg/mL, p=0.07) and day 21 (3.76 versus 3.66 log₁₀pg/mL, p=0.005; Figure 2B). No differences were detectable in plasma concentrations of IL-6, TIM3, ST2, and TNFα for all time-points (Figure 2C-F). However, the majority of IL-6, ST2, and TNFα concentration values were below the LOD (62%, 75%, and 98% respectively; Table S1), which limited analyses for IL-6 and ST2 and precluded analyses for TNFα. Numerical results are summarized in Table S6.

Figure 2.

Box-and-Whisker plots displaying median and interquartile ranges of measured cytokines (A-F, as labeled) in iciHHV-6^pos HCT cases (HCT cases, grey boxes) and iciHHV-6^neg HCT controls (HCT controls, white boxes) at days 7 (n_case=41 and n_control=64), 14 (n_case=44 and n_control=53), and 21 (n_case=27 and n_control=45) (+/− 3 days) post-HCT. All values are presented as log₁₀ pg/mL. Exact P values are displayed and were calculated using the Mann-Whitney test. The dashed line represents the limit of detection divided by 2.

We next stratified our comparisons for CRP and TNFRp55 by subsequent aGVHD severity. Across all time-points and both cytokines, we generally observed higher cytokines in iciHHV-6^pos HCT cases compared to iciHHV-6^neg HCT controls in individuals with aGVHD grades 0-1 and those with aGVHD grades 2-4 (Figure 3). The most apparent differences were in day 21 samples among patients who subsequently developed aGVHD grades 2-4 (TNFRp55, 3.78 versus 3.66 log₁₀pg/mL; p=0.002; CRP, 7.48 versus 6.92 log₁₀pg/mL; p=0.02).

Figure 3.

Box-and-Whisker plots displaying median and interquartile ranges of CRP plasma levels in the first row and TNFRp55 plasma levels in the second row. Separate plots were made for each time-point (day 7, 14, and 21 (+/− 3 days) post-HCT) and are stratified by aGVHD severity (grades 0-1 vs. grades 2-4) and iciHHV-6 status (iciHHV-6^pos HCT cases in grey boxes and iciHHV-6^neg HCT controls in white boxes). All values are presented as log₁₀ pg/mL. Exact P values are displayed and were calculated using the Mann-Whitney test.

To incorporate data from all time points per patient, we compared mean differences in log₁₀ cytokine levels per day (for CRP and TNFRp55) between iciHHV-6^pos HCT cases and iciHHV-6^neg HCT controls using a repeated measures GEE model. In this analysis, iciHHV-6^pos HCT cases had overall higher mean CRP and TNRFp55 results compared to iciHHV-6^neg HCT controls (Table 2). For CRP, the mean value decreased over time, whereas the mean value increased over time for TNFRp55. There were no significant differences in the slopes of mean cytokine values over time between iciHHV-6^pos HCT cases and iciHHV-6^neg HCT controls (interaction P values 0.94 for CRP, 0.07 for TNFRp55). We next evaluated whether the findings differed by subsequent acute GVHD severity (grades 2-4 versus 0-1) and found no statistical evidence that the mean cytokine differences between iciHHV-6^pos HCT cases and iciHHV-6^neg HCT controls were affected by subsequent aGVHD severity (data not shown). Inclusion of aGVHD in the model did not affect the observed association between iciHHV-6 status and cytokine levels. Furthermore, there was no interaction between the iciHHV-6 status and subsequent aGVHD grade (i.e. the effect of iciHHV-6 status did not differ depending on subsequent aGVHD grade; interaction P value 0.94 for CRP, 0.96 for TNFRp55).

Table 2:

Comparisons of mean log₁₀ cytokine levels (CRP and TNFRp55) between iciHHV-6^pos HCT cases and iciHHV-6^neg HCT controls using repeated measures GEE models.

CRP
Variable	Estimate	95% CI	P-value
Day (change of 1 day)	−0.03	(−0.04, −0.01)	<0.001
iciHHV-6 case vs. controla	0.33	(0.12, 0.54)	0.002
Intercept	7.51	(7.28, 7.74)	<0.001
TNFRp55
Variable	Estimate	95% CI	P-value
Day (change of 1 day)	0.007	(0.003, 0.01)	<0.001
iciHHV-6 case vs. controla	0.12	(0.03, 0.22)	0.009
Intercept	3.48	(3.41,3.55)	<0.001

iciHHV-6 indicates inherited chromosomally integrated HHV-6; CI, confidence interval.

^aThe estimate demonstrates the mean log₁₀ difference in cytokine concentration per day for samples tested up to day 21 (+/− 3 days) post-HCT.

The iciHHV-6 source (donor versus recipient) and species (A versus B) were not associated with plasma concentrations of proinflammatory cytokines

We explored a potential influence of the source (donor, recipient, or both) of iciHHV-6 on the cytokine concentrations. Cytokine plasma concentrations for each timepoint were compared between the three possible combinations of donor and recipient iciHHV-6 status (donor and recipient iciHHV-6^pos; donor iciHHV-6^pos and recipient iciHHV-6^neg; donor iciHHV-6^neg and recipient iciHHV-6^pos) with iciHHV-6^neg HCT controls. There was no clear effect of the iciHHV-6 source on the magnitude of differences in cytokine levels between these subgroups of iciHHV-6^pos HCT cases compared to the iciHHV-6^neg HCT controls (data not shown).

We also explored whether there was a difference in CRP and TNFRp55 plasma concentrations when iciHHV-6^pos HCT cases were stratified by HHV-6 species (A [n=19] versus B [n=45]). This demonstrated no differences between the groups (Table S7).

Stem Cell Donor Characteristics: Cohort Study 2

Cohort 2 consisted of 28 donor cases with iciHHV-6 and 56 donor controls without iciHHV-6. The donor cases had a median age of 49 years, 43% were female, and all were of Caucasian race. The donor controls had a median age of 43 years, 43% were female, and 97% were of Caucasian race.

Proinflammatory cytokines were not elevated in healthy donors with iciHHV-6

A total of 84 samples from Cohort 2 were measured for six cytokines. We did not observe any differences between iciHHV-6^pos donor cases and iciHHV-6^neg donor controls for IL-6, ST2, TIM3, TNFRp55, and CRP (Figure 4). Additionally, all donor cytokine levels were lower than day 7 post-HCT cytokine levels with differences between median cytokine concentrations ranging from 0.2 to 1.3 log₁₀pg/mL and p values of 0.14-0.96 (excluding TNFα; Figure 4). Comparable to Cohort 1, the majority of IL-6, ST2, and TNFα concentrations were below the LOD (Table S1).

Figure 4.

Box-and-Whisker plots displaying median and interquartile range of IL-6, ST2, TIM3, TNFRp55 and CRP plasma concentrations in iciHHV-6^pos donor cases (donor cases, white boxes), iciHHV-6^neg donor controls (donor controls, light grey boxes), and iciHHV-6^pos HCT cases and iciHHV-6^neg HCT controls (combined) using the day 7 (+/− days) post-HCT time-point (HCT cases and controls, dark grey boxes). All values are presented as log₁₀ pg/mL. Exact P values are displayed and were calculated using the Mann-Whitney test. There were no differences in cytokine levels between iciHHV-6^pos donor cases and iciHHV-6^neg donor controls for any cytokine. However, median differences in combined donor versus combined HCT recipient cytokine concentrations ranged from 0.2 to 1.3 log₁₀ pg/mL (excluding ST2 given most values were below the limit of detection [LOD]). TNFα levels are not depicted given that 94% of all results were below the LOD.

DISCUSSION

We performed two cohort studies of individuals with (cases) and without (controls) iciHHV-6 consisting of 64 iciHHV-6^pos HCT cases and 64 matched iciHHV-6^neg HCT controls (Cohort Study 1), as well as 28 iciHHV-6^pos stem cell donor cases and 56 iciHHV-6^neg matched stem cell donor controls (Cohort Study 2). We tested plasma samples for 6 cytokines over time to test the hypothesis that individuals with iciHHV-6 have higher median proinflammatory cytokine concentrations than individuals without iciHHV-6. We demonstrated that iciHHV-6^pos HCT cases had higher concentrations of CRP and TNFRp55 after HCT, as well as an earlier onset of aGVHD grades 2-4, compared to iciHHV-6^neg HCT controls. However, there were no differences in these cytokine levels in healthy donor cases and controls.

Acute GVHD involves a variety of cell types, such as monocytes, macrophages, T cells, and natural killer (NK)-cells, which express a range of proteins that mediate allo-reactivity and apoptosis of targeted host cells¹².Due to the inflammatory nature of aGVHD, the cytokine profile became a focus of research to predict aGVHD severity and onset. A number of proinflammatory cytokines have been relatively consistently identified as early biomarkers for subsequent aGVHD, increased risk of non-relapse mortality post-HCT, and increased risk for heart disease in non-immunocompromised populations^16-22. Cytokines are released during pre-HCT conditioning due to tissue damage by radiation and chemotherapy, which stimulate antigen presenting cells (APCs)¹². After receiving donor cells, patient APCs may activate donor T cells that exacerbate tissue damage by releasing further proinflammatory cytokines.

In addition to treatment-related triggers of cytokines and aGVHD, reactivation of herpesviruses such as HHV-6^4,30-32 and CMV^33,34 have been associated with aGVHD. A meta-analysis by Phan et al. found that HHV-6B reactivation is an independent risk factor for the subsequent development of aGVHD grades 2-4⁴. We previously found that HCT recipients in whom either the donor or recipient harbored iciHHV-6 had an increased risk for aGVHD grades 2-4¹⁵. Importantly, iciHHV-6 is not a dead-end for viral replication, and reactivation of the integrated HHV-6 strain has been demonstrated in vivo and in vitro^6,7,9. Tissue culture experiments showed an induction of viral gene transcription after exposure to histone deacetylase inhibitors^6,35. Subsequently infected naive T cells developed cytopathic effects. Strenger et al. found viral transcripts of lytic and latent genes in PBMCs of iciHHV-6^pos individuals¹⁰. Phan et al. recently reviewed a potential mechanism by which HHV-6B reactivation could cause or exacerbate aGVHD³⁶. They hypothesize that HHV-6B reactivation following conditioning chemotherapy causes depletion of CD4⁺ T cells, including a regulatory T cell (Treg) population, resulting in an uncontrolled expansion of allo-reactive T cells that promote aGVHD. Given that individuals with iciHHV-6 contain the viral genome in all nucleated cells of the body, it is biologically plausible that viral gene expression or reactivation early after conditioning chemotherapy may contribute to donor allo-reactive T cells in HCT recipients who have or receive donor hematopoietic cells with iciHHV-6, and this may influence the development of GVHD.

In this study, we demonstrated that iciHHV-6^pos HCT cases had higher plasma concentrations of CRP on days 7 and 21 post-HCT, as well as higher TNFRp55 concentrations on days 14 and 21 post-HCT, compared to iciHHV-6^neg HCT controls. These findings were recapitulated in longitudinal repeated measures models demonstrating that mean log₁₀ concentrations of CRP and TNFRp55 were elevated in iciHHV-6^pos HCT cases compared to iciHHV-6^neg HCT controls. After stratification by aGVHD severity, we found that the increased levels of CRP and TNFRp55 in iciHHV-6^pos HCT cases compared to controls were present at all time-points, but the largest differences were found in day 21 samples among individuals who subsequently developed aGVHD grades 2-4. We did not observe any differences when analyses were stratified by iciHHV-6 source and species.

Elevated CRP levels post-HCT have been shown to serve as a prognostic marker for aGVHD^16,37, and an early rise of CRP was associated with a higher number of major complications and transplant-related mortality³⁸. The role of TNFRp55 in the development of aGVHD is less clear with some studies demonstrating an increase of TNFRp55^16,39,40 and others not¹⁷. The absolute median differences spanning all time-points ranged between 0.04-0.1 log₁₀ pg/mL for TNFRp55 and 0.26-0.32 log₁₀ pg/mL for CRP, which is consistent with the magnitude of differences for these cytokine in other studies of HCT recipients who did and did not develop aGVHD^17,38. Interestingly, the largest differences in cytokine levels between iciHHV-6^pos HCT cases and iciHHV-6^neg HCT controls were at the day 21 time-point, which coincides with the median time of reactivation of HHV-6B after allogeneic HCT^41-43. Furthermore, we demonstrated that iciHHV-6^pos HCT cases developed aGVHD a median of seven days earlier than iciHHV-6^neg HCT controls. Although there are data to suggest that earlier-onset aGVHD is associated with higher non-relapse mortality in specific patients based on donor relation and HLA-matching, the differences in time of onset were larger than seen in our study, and the clinical significance of a 7 day difference in aGVHD-onset is unclear⁴⁴. Together with our findings of an increase in certain plasma cytokine levels, these data support a potential role of iciHHV-6 in as a mediator of aGVHD and post-HCT morbidity.

In the healthy stem cell donor cohort, there were no differences in cytokine concentrations between iciHHV-6^pos donor cases and matched iciHHV-6^neg donor controls. This could suggest that these individuals maintain control of viral reactivation in the absence of immunosuppressive conditions or treatments.

The strength of this study is the unique and large cohort of immunocompromised and healthy individuals with iciHHV-6, along with accompanying plasma samples at multiple timepoints. It is important to note that the results do not account for multiple comparisons. Many individuals had undetectable levels of IL-6, ST2, and TNFα (Table S1) using contemporary assays, which limited analyses for these cytokines. Although many of these samples had been in long-term storage at −80°C, there was no evidence of degradation of the samples, as values below the LOD were found in similar proportions of samples in early and later time periods (data not shown). We did not have samples that would allow for identifying HHV-6 gene expression or reactivation from the integrated viral strain in individuals with iciHHV-6, as this cannot be determined by standard PCR and is important to consider for future studies⁸. We focused this study on 6 cytokines that have consistently been associated with subsequent aGVHD and with heart disease in non-immunocompromised hosts^{16-19,38-40,45}, but future studies could be expanded to a broader panel of cytokines such as REG3α¹⁷, elafin¹⁷, sIL-2R³⁹, or sCD8³⁹. Statistical approaches utilizing receiver operating characteristic curves as suggested by Paczesny¹⁸ could allow for simultaneous analysis of multiple cytokines that might have a combined effect on the development of aGVHD. We note that the time between the last tested sample and onset of aGVHD was longer for iciHHV-6^neg controls compared to iciHHV-6^pos controls, which may have affected our results, but the difference was small. The overall incidence of grade 2-4 acute GVHD is higher at our center than reported at other centers, which is likely due to the high sensitivity of our approach for the diagnosis of upper gut GVHD using endoscopy⁴⁶. Finally, it is important to point out that although this study did not demonstrate a significantly higher incidence of aGVHD among iciHHV-6^pos HCT cases, we previously demonstrated this finding in a larger cohort study¹⁵, and the current study was not powered for or designed to test this question. Although other clinical complications have been associated with elevated proinflammatory cytokines, this study was not designed to evaluate clinical outcomes.

In conclusion, our study shows that HCT recipients in whom the donor or recipient have iciHHV-6 have higher proinflammatory cytokines than unaffected HCT recipients, as well as earlier onset of aGVHD grades 2-4. These findings support additional study into the pathological mechanisms of the role of iciHHV-6 in the development of aGVHD and a potential role for antiviral prophylaxis after HCT involving donors or recipients with iciHHV-6.

Highlights.

• HCT recipients in whom the donor or recipient had inherited chromosomally integrated HHV-6 (iciHHV-6^pos) have higher plasma CRP and TNFRp55 cytokine concentrations.

• IciHHV-6^pos HCT recipients had earlier-onset graft-versus-host disease by a median of 7 days.

• There were no differences in cytokine levels among healthy donors with or without iciHHV-6.