Altered Thymic Function during Interferon Therapy in HCV-Infected Patients

Interferon alpha (IFNα) therapy, despite good efficacy in curing HCV infection, leads to major side effects, in particular inducement of a strong peripheral T-cell lymphocytopenia. We here analyze the early consequences of IFNα therapy on both thymic function and peripheral T-cell homeostasis in patients in the acute or chronic phase of HCV-infection as well as in HIV/HCV co-infected patients. The evolution of T-cell subsets and T-cell homeostasis were estimated by flow cytometry while thymic function was measured through quantification of T-cell receptor excision circles (TREC) and estimation of intrathymic precursor T-cell proliferation during the first four months following the initiation of IFNα therapy. Beginning with the first month of therapy, a profound lymphocytopenia was observed for all T-cell subsets, including naïve T-cells and recent thymic emigrants (RTE), associated with inhibition of intrathymic precursor T-cell proliferation. Interleukin (IL)-7 plasma concentration rapidly dropped while lymphocytopenia progressed. This was neither a consequence of higher consumption of the cytokine nor due to its neutralization by soluble CD127. Decrease in IL-7 plasma concentration under IFNα therapy correlated with the decline in HCV viral load, thymic activity and RTE concentration in blood. These data demonstrate that IFNα-based therapy rapidly impacts on thymopoiesis and, consequently, perturbs T-cell homeostasis. Such a side effect might be detrimental for the continuation of IFNα therapy and may lead to an increased level of infectious risk, in particular in HIV/HCV co-infected patients. Altogether, this study suggests the therapeutic potential of IL-7 in the maintenance of peripheral T-cell homeostasis in IFNα-treated patients.


Introduction
The hepatitis C virus (HCV) causes persistent infection in approximately two thirds of cases leading to chronic liver disease, liver failure, and, eventually, hepatocellular carcinoma in a substantial proportion of infected individuals. The most common therapy for chronic hepatitis C consists of pegylated interferon-a (IFNa) and ribavirin administration which results in viral clearance in 43-46% (genotype 1) to 80%, (genotype 3) of treated patients [1]. Interferon will continue to be a major component of new direct acting antivirals for treatment of HCV [2].
IFNa is produced in large amounts during the acute phase of many viral infections [3,4,5,6]. Direct activation of interferonstimulated genes enhances naïve T-cell survival through increased Bcl-2 and reduced Bax activation [7] and contributes to clonal expansion of antigen-specific T-cells [8]. Recent data suggest that early therapeutic intervention with pegylated IFNa rescues polyfunctional memory T-cells expressing high levels of the IL-7 receptor alpha chain (CD127) and Bcl-2, allowing a higher rate of sustained viral response [9]. However, despite good efficacy, IFNa-based therapies lead to sustained anemia, thrombocytopenia, neutropenia and lymphocytopenia [10,11,12,13,14]. Moreover, pegylated IFNa therapy enhances the risk of infection in older HCV-infected patients and HIV-infected individuals, independently from neutropenia [15,16,17].
The mechanisms of action of IFNa include inhibition of different hematopoietic growth factors [18,19], possibly affecting lymphoid differentiation at an early stage [20], and modifications in cell homing [12,21,22]. The mechanisms involved in IFNa therapy-associated leukocyte depletion remain poorly understood.
Others and we have documented a strong reduction in the ability of HIV-infected patients to sustain thymic production as a direct consequence of a drop in intrathymic precursor T-cell proliferation [23,24,25]. Similar thymic impact was also seen during early SIV-infection in the rhesus macaque model [26]. The capacity of the thymus to produce recent thymic emigrants (RTEs) is, in large part, dependent on thymocyte proliferation [27]. Indeed, extensive thymocyte proliferation occurs between T-cell receptor beta (TCRB) and alpha (TCRA) chain rearrangements. The extent of this proliferation directly correlates with thymic export [28]. The extent of cell proliferation in the thymus can be measured in patients through estimating, in peripheral blood cells, the ratio (sj/bTREC ratio) between the frequency of signal joint T-cell receptor excision circles (sjTREC), produced during the excision of the TCRd locus prior to TCRa chain rearrangement, and that of DbJbTREC T-cell receptor excision circles (TRECs) produced during TCRBD to TCRBJ rearrangement [29]. These by-products of TCR rearrangement processes are generated by the circularization of the chromosomal DNA excised during TCR rearrangements respectively occurring at the DN3 (DbJbTREC) and DP (sjTREC) stages of differentiation. Due to the fact that TRECs do not replicate during mitosis, increased proliferation between DN3 and DP leads to the reduction of DbJbTREC frequency in RTEs as compared to sjTREC frequency and consequently to an increase of the sj/bTREC ratio [23]. The correlation between initial plasma IFNa levels and the speed of thymic dysfunction observed during HIV primary infection suggested that IFNa, produced as part of the innate immune response to infection, participates in the impairment of thymopoiesis. However, no direct evidence of the relationship between IFNa production and thymic dysfunction was provided by these studies. In contrast, Arizcorreta and colleagues showed that IFNa and ribavirin therapy induces a substantial reduction of circulating sjTRECs, in HIV/HCV co-infected patients, accompanied by sustained naïve CD4 + T-cell defect, suggesting thymic dysfunction [10]. Similarly, in the SIV-infected rhesus macaque model, we showed that IFNa therapy induced a strong decrease of circulating RTE numbers as defined either by sjTREC frequency and numbers or by CD31 hi expression on naïve T-cells [30]. Interestingly, in these animals, recombinant interleukin (IL)-7 therapy more than abrogated the deleterious effects of IFNa therapy [30]. IL-7 is a key cytokine implicated at various levels of thymocytes differentiation. It allows cell survival during the rearrangement processes, and is implicated in the extensive thymocyte proliferation, in particular in intermediate single positive (ISP) and early DP cells [31,32,33,34]. This proliferation participates in the development of naïve T-cell diversity [35]. While up regulated by IFNa [36,37], the cyclin-dependent kinase inhibitor P27/Kip1 is negatively regulated by IL-7 [38], allowing ISP and early DP thymocytes to proliferate. Moreover, IFNa also inhibit IL-7 dependent proliferation through down modulation of the common c chain, that participates, together with CD127 to the IL-7 receptor [39]. We here investigated the early impact of IFNa therapy on thymic function and naïve T-cell homeostasis in both HCV-infected and HIV/HCV co-infected patients who started IFNa therapy.
Within CD4+ naïve T-cells, RTEs can be identified by their higher expression of the platelet endothelial cell adhesion molecule-1 (PCAM-1 or CD31) [40]. While the number of RTEs was similar in HCV-infected patients at study entry and healthy individuals ( These data demonstrate that, as early as one month following treatment initiation, IFNa induces stronger alterations of naïve Tcell subsets, and more specifically in the RTE compartment than in any other T-cell subset, suggesting a specific effect on thymopoiesis. We thus analyzed the evolution of intrathymic precursor T-cell proliferation, peripheral T-cell cycling, IL-7 plasma concentration and IL-7 receptor alpha chain (CD127) expression, different factors affecting naive T-cell homeostasis.

IFNa therapy affects thymic function
Despite differences between the 3 groups at study entry, RTE cycling rate, as estimated through measurement of Ki-67 expression, did not change significantly during the follow-up period ( Figure 3A). These data demonstrate that the observed changes in sjTREC frequencies were not a consequence of variations of RTE proliferation during IFNa therapy but more probably due to reduced thymic production.
We thus estimated thymic output through quantification of the sj/bTREC ratio in all groups of patients ( Figure 3B). The sj/ bTREC ratio estimates the extent of thymocyte proliferation between TCRB rearrangement and the excision of the T-cell receptor delta (TCRD) locus [23]. This parameter directly reflects the extent of thymic production and, contrarily to sjTREC values, is independent from peripheral RTE proliferation or survival capacity [28]. The sj/bTREC ratio was already low in HIVinfected patients (p,0.005 as compared to healthy control donors; Figure 3B bottom left panel) and did not evolve further under IFNa therapy in co-infected patients ( Figure 3B, bottom right panel). In contrast, acutely HCV-infected patients demonstrated higher than normal sj/bTREC ratio at baseline (p,0.05 as compared to aged matched healthy controls), showed a significant reduction in sj/bTREC ratio at M1 (p = 0.014) and M2 (p = 0.001; Figure 3B, top panel). Finally, a similar decline in the sj/bTREC ratio was observed during IFNa therapy in chronically HCV-infected patients (p,0.02 at M1, M2 and M3; Figure 3B, central panel).

Reduction of IL-7 plasma levels under IFNa treatment
Precursor T-cell proliferation in the thymus is, at least in part, dependent upon IL-7. We thus quantified plasma IL-7 concentration in all groups of patients. At study entry, HCV-and HIV/ HCV-infected patients presented with elevated plasma IL-7 (median = 10.3 pg/mL, range (6.7-12.9) in acutely HCV-infected patients; 8.3 pg/mL (6.3-10.5) in chronic HCV-infected patients and 7.15 pg/mL (4.3-13.5) in co-infected subjects), as compared to that observed in healthy control individuals (p,0.001 for any patients' group; Figure 4A). Surprisingly, while lymphocytopenia established, IL-7 plasma concentrations significantly decreased in both groups of HCV-infected patients (30,54,18 and 29% decrease at M1 to M4 in acute infection, p,0.05; 25, 46, 26 and 16% decrease at M1 to M4 in chronic infection, p,0.05; Figure 4B left and central panels). In contrast, IL-7 plasma levels did not significantly evolve in co-infected individuals during the first month of IFNa therapy ( Figure 4B right panel). Only patients with the highest IL-7 plasma levels showed a reduction in the concentration of this cytokine.
Decreased plasma IL-7 concentrations could be a consequence of reduced IL-7 production, increased consumption by T-cells or sequestration by soluble IL-7 receptor (sCD127). In both HCVinfected and HIV/HCV co-infected patients, neither sCD127

Evolution of Thymic function parallels IL-7 plasma levels
Considering the variations in all the parameters we used to evaluate thymic function, we then sought to evaluate the impact of changes in IL-7 plasma levels on de novo production from the thymus and on the number of both sjTREC and circulating CD4+ RTEs.
Variations in plasma IL-7 levels also correlated with changes in the proportions (D%Ki-67+ in CD4+RTEs; r = 0.806, p = 0.0002; Figure 5D, left panel) and numbers (DKi-67+RTEs; r = 0.706, p = 0.002; Figure 5E, left panel) of cycling RTEs in acute and chronic HCV infected patients and with D%Ki-67+RTE counts in co-infected patients (r = 0.709, p = 0.022; Figure 5E, right panel). Overall, IL-7 concentration was associated with reduced thymopoiesis and RTE proliferation, lower consequently leading to limited circulating RTE and naïve T-cell counts. These data strongly suggest that changes in IL-7 plasma levels during IFNa therapy directly impact the homeostasis of RTEs.

Discussion
We herein demonstrated that IFNa-based therapy leads to major lymphocytopenia in naïve T-cell compartments, in particular in the RTE subset. Several mechanisms could be implicated in the establishment of such a lymphocytopenia [41]. Among these, enhanced apoptosis [42,43], cell sequestration in lymphoid or non-lymphoid organs [12,21,22] and regulation of peripheral T-cell homeostasis [20]. In our study, no major change in cell survival (Bcl-2 expression) or T-cell activation (CD25 and CD69 expression) was observed during the follow-up period (data not shown). Moreover, we did not observe any significant modification in Ki-67 expression in any T-cell subset during the first month of therapy (data not shown and Figure 3). Finally, IFNa-induced T-cell homing, although rapid and massive, is only a transient process [22] suggesting that this mechanism marginally contributes to the observed long lasting lymphocytopenia.
Interestingly, both sjTREC quantification (sjTREC/mL) and intrathymic precursor T-cell proliferation (sj/bTREC ratio) were affected very early on after initiation of therapy ( Figures 2B and  3B). While sjTREC frequency and concentration in peripheral blood can be affected by modifications of parameters that impact on peripheral T-cell homeostasis (cycling, survival/apoptosis, homing), the sj/bTREC ratio is a marker of the intrathymic proliferation history of RTEs. Indeed, this parameter is generated by cell proliferation that occurs between TCRb chain rearrangement and the excision of TCRd locus. Further cell cycling after TCRa chain rearrangement does not modify the sj/bTREC ratio as both type of TRECs are similarly diluted upon cell proliferation. Accordingly, while exported to the periphery, the sj/bTREC ratio of mature T-cells cannot be modified. Therefore, while the observed decrease in sjTREC concentration (figure 2) can be a consequence of modifications of circulating T-cell homeostasis, the decline of the sj/bTREC ratio observed during the first months of IFNa therapy (figure 3) defines changes in thymocyte proliferation, thus in thymic output [28]. Acutely infected patients demonstrated a higher sj/bTREC ratio at baseline than patients in the chronic phase. However, this group was younger (Median = 31.5 (26-47)) versus Median = 53.5 (37-61)) than the chronic group (p,0.01; data not shown) and demonstrated normal sj/bTREC ratio for their age. Similar evolution of thymic function and circulating T-cell subsets were observed in both groups of HCV-infected patients, irrespective of the development stage of HCV pathology. The lack of effect of IFNa therapy in HIV/HCV co-infected patients might be due to the fact that, as expected for chronically HIV-infected individuals, these patients already had a low thymic function at study entry. The impairment of thymopoiesis in HCV-infected patients under IFNa therapy is reminiscent of that observed during the acute phase of HIV-1 infection [23] which suggested that long term production of IFNa, as part of the anti-HIV innate immune response, may play a role in the observed thymic defect. The correlation between decline in IL-7 plasma levels under IFNa therapy and both thymic dysfunction and reduced T-cell counts, in particular in the naïve and RTE compartments ( Figures 5A and  5B), confirms this hypothesis. Finally, in a recent study, we showed that IFNa treatment leads to decreased sjTREC frequency as well as reduced naïve T-cell and RTE counts in SIV-infected rhesus macaques [30]. Such an effect was accompanied by a 30-40% decrease in IL-7 plasma levels in these animals and could be counteracted by injection of recombinant simian IL-7 [30]. One could expect that such an effect of type I IFNs is not restricted to HIV-infection as many viral infections induce IFNa responses and cause transient lymphocytopenia in the infected hosts [3,4,5,6].
Moreover, the IFNa-induced reduction of thymic function and its probable consequences on naïve T-cell diversity may contribute to the higher infectious risk associated with IFNa therapy, in particular observed in older patients [15,16,44]. There are multiple sources for circulating IL-7 during viral infections including lymphoid organs, epithelial cells and recently the liver was identified as a major source of IL-7. Moreover, increased plasma IL-7 levels can also be observed during viral infection in non-lymphopenic individuals ( [33] and unpublished data), suggesting a role in the development of immune responses. Indeed, this cytokine participates to T-cell homing in various lymphoid and non-lymphoid tissues through stimulation of local chemokine productions [45]. Increased IL-7 plasma levels in lymphopenic individuals is likely due to reduced consumption [46] yet augmented production to counteract lymphopenia cannot be excluded [33]. The recent identification of the liver as an IL-7 producing tissue upon TLR stimulation [47] makes it tempting to speculate that HCV-infection can also, through TLR activation, stimulate IL-7 production by the liver. Indeed, non-lymphopenic HCV-infected patients demonstrate similar IL-7 plasma levels than lymphopenic HIV-infected individuals [33,48] suggesting that most of the IL-7 production in untreated HCV-infected patients was not linked to circulating T-cell counts. The reduction of IL-7 plasma levels while lymphocytopenia establishes under IFNa therapy, the absence of a correlation between IL-7 plasma levels and CD127 expression and the concomitance of decreases in IL-7 plasma levels and HCV viral load under therapy suggest that viremia might be driving IL-7 production before initiation of therapy. Our data suggest that, before initiation of IFNa therapy, actively replicating HCV leads to the overproduction of IL-7. Subsequent reduction of IL-7 production upon initiation of therapy probably reflects the elimination of IL-7 producing HCV-infected hepatocytes. This sudden reduction of IL-7 plasma levels may lead to diminished thymopoiesis. The fact that IL-7 plasma levels did not reach normal levels when HCV became undetectable may suggest that, after the initial decline that follows the drop in viremia, IL-7 plasma levels were regulated, as in HIVinfected patients [33] and in IFNa-treated SIV-infected rhesus macaques [30], as a consequence of lymphocytopenia through either reduced consumption or increased production in lymphoid organs [49]. Future studies with a longer follow-up period, in particular after the end of IFNa therapy and recovery from lymphocytopenia are required to further elucidate this point.
We herein demonstrated that a substantial reduction in thymic export was observed in HCV-infected patients, during the first months of IFNa therapy. This effect directly paralleled IFNainduced lymphocytopenia and decreased IL-7 plasma levels, initially high in HCV-infected patients. These data suggest that IL-7 production by the liver, a consequence of active HCV replication, was reduced while patients controlled HCV viremia. Restricted IL-7 plasma levels might, in association with the antiproliferative effect of IFNa, limit T-cell production in the thymus. Our study highlights the therapeutic potential of IL-7 as a complement to the standard IFNa based treatment to help HCVinfected patients to sustain normal circulating T-cell counts, and restore the diversity of the peripheral T-cell repertoire through its central thymopoietic effect. Restoring the breadth and intensity of T-cell control over the HCV virus might be immediately beneficial for the HIV/HCV co-infected population and offer new promising avenues for chronic HCV in the context of massive drop of HCV viral load after short term treatment with new antiviral compounds that will continue to be administered in combination with IFNa [50].

Patients characteristics
Sixteen HCV-infected patients (C-1 to C-16) and ten HIV/ HCV co-infected patients (I-1 to I-10) naïve to IFNa therapy were enrolled in this study. A summary of the virological and immunological status of patients at baseline is shown in table 1. All the HIV/HCV co-infected patients but one were under HAART with undetectable viremia (,40 HIV copies/mL). Chronically infected patients (C-9 to C-16 and I-1 to I-10) initiated pegylated IFNa/ribavirin treatment (IFNa-2a: Pegasys, 180 mg weekly, Ribavirin: Copegus, 800 mg to 1000 mg daily) and were followed over a 4 months period. Patients included in the acute phase of HCV infection (C-1 to C-8) were treated with pegylated IFNa (IFNa-2a: Pegasys, Roche, 180 mg weekly) [51,52]. Blood samples were taken monthly on EDTA. Two milliliters of total blood were 2-fold diluted in FCS/20%DMSO frozen at 280uC and conserved in liquid nitrogen. These total blood samples were subsequently used for flow cytometry analyses. Plasma was separated from the remaining eight milliliters and mononuclear cells were purified on Ficoll Hypaque (Eurobio, Courtaboeuf, France) and frozen for further analyses. Patients from the HCV mono-infection group were followed at the Centre de Recherche du CHUM, Hôpital Saint Luc, Montreal, QC, Canada and its collaborators as previously described [9,53]. Patients from the HIV-HCV groups were followed at the Hôpital Henri Mondor, Creteil, France. Clinical protocols conformed to Figure 5. Variations in IL-7 plasma levels correlate with evolution of RTE production. Correlations. between variations in IL-7 plasma levels (DIL-7) and either variations in (A) total (CD4 + + CD8 + ) naïve T-cell counts (Dnaïve T-cell counts), (B) RTE defined as CD31 hi naïve CD4 + T-cells (DRTE CD4 counts), (C) the sj/bTREC ratio (Dsj/bTREC ratio), (D) the frequency of Ki-67 + cells in the RTE CD4 + T-cell subset (D%Ki-67 + in CD4 + RTEs) or (E) the number of circulating Ki-67 + CD4 + RTEs (DKi-67 + RTE counts) between study entry and month 1 of therapy were calculated for acutely (black symbols) and chronically (white symbols) HCV-infected patients (left panels) and HIV/HCV co-infected patients (right panels). Correlation coefficients (Spearman's r) and the associated probabilities (p) are shown. doi:10.1371/journal.pone.0034326.g005 ethical guidelines of the authors' institutions and the US Department of Health and Human Services' human experimentation guidelines. This study was approved by both the Ethical committee of Centre Hospitalier de l'Université de Montreal (CHUM) and the ethical committee of Hôpital Henri Mondor, Créteil, France. Samples were obtained with the written subjects' informed consent.
Immunophenotyping and flow cytometry analysis FACS analyses were performed on cryopreserved samples. After thawing blood cells were incubated for 15 minutes at 4uC with conjugated monoclonal antibodies (mAbs). For intracellular labeling, cells were permeabilized with the Cytofix/Cytoperm Kit (Becton Dickinson) before incubation with specific mAbs according to the manufacturer's instructions. Samples were then washed, fixed in 2% paraformaldehyde phosphate-buffered saline (PBS/PFA 2%) and acquired using a Cyan cytofluorometer (Dako) and analyzed with FlowJo 8.7 software.

IL-7 plasma quantification
IL-7 was quantified in the plasma using the IL-7 Quantikine HS kit according to the manufacturer's instructions (R&D Systems Europe). Plasma soluble-CD127 quantification Soluble plasma IL-7 receptor (sCD127) quantification was performed as previously described [54].

TREC quantifications
Parallel quantification of the sjTREC and the 13 DJbTRECs, together with CD3c gene (used as a housekeeping gene) was performed for each sample using LightCyclerTM technology (Roche Diagnostics) with a technique adapted from [29]. Intrathymic precursor T-cell proliferation was evaluated through calculation of the sj/bTREC ratio as described [23].

Statistical analysis
Statistical analyses (Spearmans rank correlations and Wilcoxon matched -paired signed-rank tests) were performed using the Stata/IC 10.0 (Stata corporation, College Station, Tx U.S.A.). Due to the exploratory nature of the study there was no correction for multiple comparisons, and calculated p values are reported herein.