In lung transplant recipients (LuTRs), human cytomegalovirus (HCMV) DNAemia may be associated with HCMV disease and reduced survival of the allograft. Because T cells are essential for controlling HCMV replication, we investigated in this prospective study whether the kinetics of plasma HCMV DNA loads in LuTRs are associated with HCMV-specific CD8+ T cell responses, which were longitudinally assessed using a standardized assay. Sixty-seven LuTRs were monitored during the first year posttransplantation, with a mean of 17 HCMV DNA PCR quantifications and 11.5 CD8+ T cell tests performed per patient. HCMV-specific CD8+ T cell responses displayed variable kinetics in different patients, differed significantly before the onset of HCMV DNAemia in LuTRs who subsequently experienced episodes of DNAemia with high (>1000 copies/mL) and low plasma DNA levels (p = 0.0046, Fisher's exact test), and were absent before HCMV disease. In HCMV-seropositive LuTRs, high-level DNAemia requiring preemptive therapy occurred more frequently when HCMV-specific CD8+ T cell responses fluctuated, were detected only after HCMV DNA detection, or remained undetectable (p = 0.0392, Fisher's exact test). Thus, our data indicate that HCMV-specific CD8+ T cells influence the magnitude of HCMV DNAemia episodes, and we propose that a standardized measurement of CD8+ T cell immunity might contribute to monitoring the immune status of LuTRs posttransplantation.
Infection with human cytomegalovirus (HCMV) may manifest in lung transplant recipients (LuTRs) as HCMV syndrome or tissue-invasive HCMV disease (1–5). In HCMV-seropositive recipients (R+), HCMV replication may occur, due to either reactivation of latent virus during immunosuppressive treatment or reinfection with a donor-transmitted HCMV strain (1,2,5,6). In contrast, seronegative recipients (R–) acquire HCMV primary infection most frequently after transfer of an allograft from a serospositive donor (D+) (1,2,5,6). Such D+/R– LuTRs are at particular risk for HCMV disease, probably due to their lack of immunity, with the organ of the donor acting as viral reservoir (1).
The level of HCMV DNAemia has been shown to be predictive for the development of HCMV disease during post–lung-transplant surveillance and therefore serves as an indicator for preemptive antiviral therapy (7–9). HCMV DNAemia may also trigger chronic inflammatory processes that have been associated with bronchiolitis obliterans syndrome, and there is evidence that multiple and prolonged episodes of HCMV DNAemia significantly reduce long-term survival rates (10–13). Although measurement of plasma HCMV DNA levels is a routine tool in the monitoring of LuTRs, the functional mechanisms that lead to hematogenous HCMV dissemination are not entirely understood.
The HCMV-specific cellular immune response has been identified as an essential host factor for controlling HCMV infection (14,15). In particular, CD8+ cytotoxic T cells have a key function because they specifically recognize and kill HCMV-infected cells, with CD4+ T helper cells providing necessary stimulatory signals. Although R– LuTRs must first generate a specific T cell response when they acquire a primary HCMV infection, R+ patients display an already primed HCMV-specific T cell immunity, which might be impaired by the high-dose immunosuppressive treatment administered after lung transplantation (LuTX; Refs. 16–19). The quality of the specific T cell response in primary HCMV infection and the kinetics with which responses recover in R+ LuTRs posttransplant determine the number of subsequent reactivation episodes and the patients' susceptibility for HCMV disease (20–24).
Currently, the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company, Darmstadt, Germany) is the only commercially available and standardized test kit for monitoring the response of HCMV-specific CD8+ cytotoxic T cells (25). This assay has been optimized for the clinical setting and is used to quantify interferon-γ (IFNG), which is produced by specific CD8+ T cells after ex vivo stimulation with various HCMV T cell epitopes (25,26). A previous study in which the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) was used demonstrated that solid-organ recipients in whom a significant HCMV-specific IFNG response could be detected at the end of at least 3 months of antiviral prophylaxis had a lower incidence of HCMV disease during the later posttransplant follow-up (22).
However, to date, the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) has not been applied to longitudinally investigate in LuTRs whether the dynamics of HCMV-specific IFNG responses are associated with HCMV DNAemia. Therefore, in this prospective study, we have analyzed individual kinetics and intraindividual changes of plasma HCMV DNA loads and HCMV-specific IFNG responses in a large cohort of LuTRs at serial time points during the first year posttransplant.
Materials and Methods
In this prospective study, a total of 68 LuTRs (35 females, 33 males) with a mean age of 53 years (range: 21–68) were included. The most common indication for LuTX in the patient cohort was chronic obstructive pulmonary disease (46 patients), followed by cystic fibrosis (7 patients) and α1 antitrypsin deficiency (5 patients). All patients received a lung transplant (65 double and 3 single lung transplants) at the Medical University of Vienna between October 2007 and October 2009. Induction therapy with antithymocyte globulin or alemtuzumab was administered to patients with cystic fibrosis and idiopathic pulmonary hypertension (two patients). Postoperatively all patients received a triple-drug immunosuppressive treatment, including a calcineurin inhibitor (cyclosporine or tacrolimus), an antimetabolite (mycophenolic acid) and corticosteroids (prednisolone). Target trough levels during the first year posttransplant were 250–350 ng/mL for cyclosporine and 10–15 ng/mL for tacrolimus. In addition, all patients received antiviral prophylaxis, which consisted of hyperimmune globulin (1 mL/kg bodyweight) at days 0, 7, 14 and 21 after LuTX, ganciclovir administered intravenously for 3 weeks, and valganciclovir administered orally (450–900 mg, twice daily, depending on the patient's weight and renal function) for a duration determined by the patient's HCMV serostatus. Patients with low and intermediate risk of HCMV disease (D–/R–, D+/R+ and D–/R+ constellation) received oral prophylaxis for a mean of 96 days (range: 41–198), whereas high-risk D+/R– patients were treated for a mean of 318 days (range: 211–452). The clinical surveillance schedule comprised weekly visits to the outpatient clinic with collection of blood samples for 2 months posttransplant and monthly thereafter, with additional samples taken if clinically indicated. When plasma HCMV loads exceeded 1000 copies/mL during follow-up, preemptive valganciclovir therapy was initiated. All patients gave written consent for participation, the protocol was approved by the institutional review board, and the study was conducted in accordance with the declaration of Helsinki and the guidelines of the local ethics committee (EC protocol 165/2007).
Quantification of HCMV DNA
Quantitative assessment of the HCMV DNA load in plasma samples was performed by PCR using a Cobas Amplicor HCMV Monitor-Test Kit on a COBAS Amplicor Analyzer (Roche Molecular Systems, Branchburg, NJ, USA).
QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company)
HCMV-specific CD8+ T cell responses were assessed using the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) as described by the manufacturer and as reported previously (19,22,25,26). From each of the patients, a heparinized whole-blood sample was obtained at each visit to the outpatient clinic, from which aliquots of 1 mL were placed into three tubes. The first tube contained a mix of 21 peptides, each with a length of 8–13 amino acid residues, derived from HLA-restricted CD8+ T cell epitopes of various HCMV proteins, including phosphoprotein (pp) 65, pp 50, glycoprotein B, immediate early antigen (IE)-1 and IE-2, as described in detail elsewhere; the second tube contained the mitogen phytohemagglutinin and served as a positive control and the third tube was the negative control and contained sterile phosphate-buffered saline (19,22,25,26). The tubes were then shaken vigorously and incubated for 18–24 h at 37°C. After incubation, supernatants were recovered, and IFNG levels (IU/mL) were measured using an enzyme-linked immunosorbent assay according to the manufacturer's instructions. A response in the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) was considered positive if the IFNG response to HCMV peptides exceeded the negative control value at least by 0.2 IU/mL. Only when the mitogen response was higher than 0.5 IU/mL was the test result interpreted as determinate. In episodes of HCMV DNAemia whole-blood samples for the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) were obtained before preemptive antiviral treatment was initiated and the result of the assay was not communicated to the physician.
Fisher' s exact test was used to evaluate whether HCMV-specific IFNG responses in preceding samples were associated with subsequent high and low plasma HCMV DNA levels (low-level episodes that occurred under antiviral prophylaxis were excluded) and could be linked to the occurrence of high-level HCMV DNAemia in R+ LuTRs. Receiver operator characteristics curve (ROC) analysis was used to evaluate the performance of the preceding specific IFNG response on prediction of subsequent HCMV DNAemia. A p value of <0.05 was considered statistically significant. GraphPad Prism version 5.0 software was used for the statistical analysis.
Prospective recruitment of patients and acquisition of serial samples
Sixty-one of the 68 LuTRs enrolled for this study completed a follow-up of 1-year posttransplant. Seven LuTRs died during the first year after LuTX but could be followed for at least 7 months. One patient who displayed an HLA haplotype that did not include epitopes covered by the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) exhibited no specific IFNG responses and was excluded from all further analyses. The remaining 67 LuTRs showed the following HCMV serostatus: D–/R+, n = 11; D+/R+, n = 28; D+/R–, n = 17; D–/R–, n = 11. From these 67 patients, 1121 plasma samples with a mean of 17 samples per patient (range: 8–43) were collected during the follow-up at a mean interval of 25.6 days (range: 2–95), and HCMV DNA loads were measured. In parallel, a total of 758 heparinized whole-blood samples with a mean of 11.5 serial samples per patient (range: 5–23) were acquired prospectively at a mean interval of 31.7 days (range: 4–111) and were tested for HCMV-specific CD8+ T cell responses using the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company). Of these, 598 samples (78.9%) gave a determinate result. Of the indeterminate results, which were all due to a low mitogen response, 33.9% were observed during the first, 33.2% during the second and 13.2% during the third month posttransplant, when immunosuppression was highest dosed and antiviral prophylaxis was still administered.
HCMV-seropositive LuTRs: HCMV-specific IFNG responses and absence of HCMV DNAemia
First, we analyzed whether the absence of HCMV DNAemia in HCMV-seropositive LuTRs was associated with detectable HCMV-specific IFNG responses. In 7 of 39 (17.9%) R+ LuTRs, HCMV DNA remained undetectable in all plasma samples collected during the first year posttransplant (Figure 1). Five of these patients (71.4%) showed HCMV-specific CD8+ T cell immunity, which was detected using the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) early during the administration of antiviral prophylaxis and remained stable during the further follow-up, in which also the blood levels of calcineurin inhibitors were stable (Figures 1 and 2A). For one patient, the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) tested positive after antiviral prophylaxis was ceased, and for another individual, IFNG responses remained permanently negative.
HCMV-seropositive LuTRs: HCMV-specific IFNG responses and high-level HCMV DNAemia
Next, we investigated HCMV-specific CD8+ T cell immunity in R+ LuTRs before episodes of high-level HCMV DNAemia. Thirteen of 39 (33.3%) R+ LuTRs developed HCMV DNAemia, with HCMV DNA loads exceeding 1000 copies/mL (Figure 1), with a median plasma HCMV DNA load of 11 100 copies/mL (range: 1690–1 170 000). In all of these patients, preemptive antiviral therapy was started after detection of viral DNA and sampling for CD8+ T cell analysis.
In 9 of these 13 (69.2%) patients, HCMV-specific IFNG responses were undetectable at the last time point before HCMV DNA was first detected in the patient's plasma. The mean interval between the performance of the corresponding QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) and the occurrence of HCMV DNAemia was 33.1 days (range, 18–44). Before high-level HCMV DNAemia, IFNG responses were absent in the patients, remaining undetectable during the entire follow-up (n = 2; Figures 1 and S1A), developing only after HCMV DNA detection (n = 5; Figures 1 and 2B) or showing fluctuating kinetics and decreasing levels before HCMV DNAemia due to an increase of immunosuppression (n = 2; Figures 1 and 2C).
In contrast, in 4 of 13 (30.8%) R+ LuTRs (3 D+/R+, 1 D–/R+), high-level HCMV DNAemia did develop despite the fact that HCMV-specific IFNG responses were detectable before and at the same time point when HCMV DNA was detected in the patient's plasma (Figure S1B). In all these patients, the blood levels of calcineurin inhibitors increased over the target concentration within 8 weeks before the detection of HCMV DNA.
HCMV-seropositive LuTRs: HCMV-specific IFNG responses and low-level HCMV DNAemia
HCMV DNAemia with viral DNA levels lower than 1000 copies/mL occurred in 19 out of 39 (48.7%) R+ LuTRs, and we further analyzed whether HCMV-specific IFNG responses could be detected before these episodes.
Before low-level HCMV DNAemia, IFNG responses to HCMV were observed in 12 out of 19 (63.2%) patients and could be detected early during antiviral prophylaxis (n = 6; Figure S1C) or afterwards but before the initial appearance of HCMV DNA (n = 6; Figure S1D). In all of these patients, the responses remained stable during the entire further follow-up.
In contrast, in 7 of 19 (36.8%) LuTRs, no HCMV-specific IFNG responses were detected before low-level HCMV DNA detection in plasma, either because of fluctuating dynamics (n = 3; Figure 1) or because they were completely absent until then (n = 4; Figure 1).
HCMV-seronegative LuTRs: HCMV-specific IFNG responses and HCMV DNA plasma loads during primary HCMV infection
Next, we investigated the association of plasma HCMV DNA loads and development of HCMV-specific IFNG responses in HCMV-seronegative LuTRs. In 11 of 28 R– LuTRs (39.3%; D+/R–: n = 10), HCMV DNAemia could be detected during the first year posttransplant, either during (n = 5) or after administration of antiviral prophylaxis (n = 6), with HCMV DNA levels exceeding 1000 copies/mL in nine of these patients (81.8%; D+/R–: n = 8). HCMV-specific IFNG responses were undetectable before all episodes of initial HCMV DNA detection (Figure 1).
Subsequent development of HCMV-specific IFNG responses could be monitored in two patients with high-level HCMV DNAemia (after 207 and 118 days, respectively; Figure S2A), as well as in both LuTRs with low-level HCMV DNAemia (simultaneously and after 75 days; Figure S2B).
HCMV DNA remained completely undetectable in the other 17 (60.7%) seronegative LuTRs (Figure 1), of which 16 never showed IFNG responses. However, one D+/R– patient developed stable HCMV-specific CD8+ T cell immunity although HCMV DNA was permanently undetectable in plasma (Figures 1 and S2C) and bronchoalveolar lavage fluid samples (BALF; data not shown).
Association of HCMV-specific IFNG responses, high-level HCMV DNAemia and occurrence of HCMV disease
We next analyzed statistically whether HCMV-specific CD8+ T cell immunity before episodes of HCMV DNAemia was associated with the level of plasma HCMV DNA in the subsequent episode (mean interval: 34.5 days, range: 10–55). We therefore investigated a total of 41 HCMV DNAemia episodes, of which 16 were preceded by a positive HCMV-specific IFNG response and 25 were not. As shown in Figure 3A, HCMV DNA loads exceeding 1000 copies/mL were detected in 72% (18/25) of the HCMV DNAemia episodes that were not preceded by a positive HCMV-specific IFNG response and in only 25% (4/16) of the episodes that were preceded by a positive response (Fisher's exact test, p = 0.0046; Figure 3A). A ROC analysis was performed for prediction of HCMV DNAemia based on the preceding IFNG response in all the patients (n = 67) and the area under the curve was 0.639 (p = 0.041; Figure 3B).
We then investigated whether distinctive kinetic patterns of HCMV-specific IFNG responses were associated with development of HCMV DNAemia with viral DNA levels exceeding 1000 copies/mL. For this purpose, we tested all 39 R+ LuTRs from our cohort and found that high-level HCMV DNAemia had developed in 53% (9/17) of the patients in whom HCMV-specific IFNG responses fluctuated and decreased, recovered only after viral DNA detection or remained permanently undetectable. In contrast, DNAemia was observed in only 18% (4/22) of patients in whom HCMV-specific CD8+ T cell immunity was detectable early during antiviral prophylaxis and/or before HCMV DNA detection and showed stable dynamics thereafter (Fisher's exact test, p = 0.0392; Figure 3C).
Finally, we investigated whether HCMV disease was associated with the absence of HCMV-specific IFNG responses. Four LuTRs from our cohort (3 D+/R–, 1 D+/R+) developed HCMV tissue-invasive disease (HCMV pneumonia; verified by histology) and in all of these cases, the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) performed before that had tested negative. In the three D+/R– LuTRs, responses were persistently absent (Figure 2D), whereas in the D+R+ LuTR, HCMV-specific CD8+ T cell immunity fluctuated and decreased before HCMV disease developed (Figure S2D).
In this prospective study, we did a longitudinal analysis of plasma HCMV DNA loads in LuTRs in relation to HCMV-specific CD8+ T cell responses determined using the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company). We provide evidence that CD8+ T cell immunity varies among these patients during the first year posttransplant and contributes to the control of HCMV replication.
High-level HCMV DNAemia is linked to the development of HCMV disease and can even trigger allograft rejection (7–13). Therefore, a deeper understanding is required of how the level of HCMV replication is affected by the immune system, in particular by cytotoxic CD8+ T cells, which are essential for containing HCMV (14–23).
One finding of our study was that HCMV-specific CD8+ T cell responses display different kinetics and have an impact on the level of HCMV DNAemia that is reached during a given episode. This was revealed by our observations that IFNG responses differed significantly between patients who subsequently developed high and low plasma HCMV DNA loads and that high-level HCMV DNAemia was more frequently preceded by a lack of responses. Consistent with these data, a recently published study of LuTRs with a D+/R– constellation demonstrated that development of an adequate CD8+ T cell response during HCMV primary infection was essential to be protected from relapsing high-level viremia (23).
It has also become evident from our data that the incidence of high-level HCMV DNAemia in R+ LuTRs might be linked to individual patterns of specific CD8+ T cell responses, because it was significantly higher in individuals who displayed fluctuations, delayed detection or complete absence of IFNG responses due to the immunosuppressive treatment. In contrast, in the majority of patients who showed early and stable responses during the entire follow-up, either HCMV DNAemia was not detected, given the sensitivity of the applied viral load assay and the investigated sample type, or the viral loads remained so low that they did not require preemptive treatment (27). This is consistent with earlier studies which showed that the posttransplant kinetics, and in particular the earliest time point at which HCMV-specific CD4+ and CD8+ T cell responses could be detected in solid-organ recipients determined whether subsequent HCMV replication could be efficiently controlled (24,28).
Furthermore, we found evidence that CD8+ T cells might influence whether HCMV disease develops after LuTX. This parallels previous data from Kumar et al., who showed that HCMV disease in solid-organ recipients may be predicted using the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company). The authors demonstrated that patients who showed a detectable IFNG response at cessation of antiviral prophylaxis displayed a lower incidence of HCMV disease during the subsequent follow-up (22). In our study, we show that the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) that was performed during the last visit to the outpatient clinic before HCMV disease developed, gave a negative result in every case. This finding supports the idea that HCMV disease might be linked to a lack of specific CD8+ T cell responses, although the number of disease cases occurring in our cohort was small. However, we demonstrate that serial measurements using the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) are required to allow an identification of short-term fluctuations of CD8+ T cell functionality that might be associated with HCMV disease in individual cases, as we have shown here.
In this study, we prospectively applied the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) in a large cohort comprised exclusively of LuTRs. After LuTX, HCMV replication is not always associated with HCMV DNAemia but can be restricted to the lung allograft, as indicated by HCMV detection solely in BALF. Westall et al., who also applied the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company), investigated HCMV-specific IFNG responses in relation to the presence of HCMV in BALF and found that in most R+ LuTRs with either no viral replication in the lung compartment or only low levels, CD8+ T cell immunity was stable or even increased. Our data now indicate that the appearance of HCMV in the blood might also be related to specific CD8+ T cells, which probably affect HCMV DNAemia by accumulating at the site of viral replication, releasing antiviral cytokines and killing infected cells, thereby partially eliminating the source from which the virus spreads into the bloodstream.
The earlier study by Westall et al., however, also showed that kinetics of HCMV-specific CD8+ T cell immunity varied among individual LuTRs and could not be used to predict local HCMV replication episodes at any given time point (19). Difficulties in predicting HCMV control by monitoring CD8+ T cells also arise from our data on viral DNA detection in the blood, although early detection and stable dynamics of HCMV-specific IFNG responses were observed for most of the patients in whom HCMV DNAemia was absent. In particular, we observed that, in a small fraction of four R+ LuTRs, high-level HCMC DNAemia occurred in spite of stable responses. In the control of HCMV replication, the importance of CD8+ T cell specificity, that is, the spectrum of viral proteins to which responses are directed, has been highlighted previously (21,23,28–31). Furthermore, it has been demonstrated that virus-specific CD4+ T helper cells are necessary for the generation and maintenance of CD8+ T cells (24,28,30–33). The QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) is designed to stimulate CD8+ T cells specifically with a mixture of multiple HCMV peptides and therefore does not allow the specific targeted epitope to be identified or CD4+ T cell functionality to be evaluated (25,34). We therefore presume that in the patients who developed high-level HCMV DNAemia despite the fact that CD8+ T cell responses were present, specific epitopes might have been targeted that were not necessarily associated with viral control. Relating to this, it has been shown recently that only responses directed against pp65 correlate with protection from viremia, whereas another study adversatively highlighted the importance of IE-1 (21,23). Also, a lack of CD4+ T cell help, which might have been caused by the high dosed immunosuppression, could provide a plausible explanation for the simultaneous detection of high-level HCMV DNAemia and specific CD8+ T cell responses, as it has been demonstrated recently that CD8+ T cells alone are not sufficient to control HCMV infection and that kidney transplant recipients with specific CD4+ T cell responses are protected from HCMV replication when the immunosuppression is stable (20,24,35). Because LuTRs may be infected by multiple HCMV strains, incomplete cross-protection by CD8+ T cells against a de novo donor-transmitted virus strain could also have been played a role, which would be strengthened by the fact that three of the four patients mentioned above displayed a D+/R+ constellation and were therefore infected with at least two different HCMV strains (36,37). Thus, it would be of future interest to investigate the epitope specificity of T cell responses in parallel to viral heterogeneity in individual hosts, as well as other immunological and virological factors that, together with CD8+ T cells, might influence HCMV replication in LuTRs.
In conclusion, we demonstrate that in LuTRs HCMV-specific IFNG responses display variable kinetics and have an impact on DNA levels during episodes of HCMV DNAemia and simultaneous HCMV disease development. We have also identified distinctive patterns of specific IFNG responses in R+ patients that were associated with high plasma HCMV DNA loads requiring preemptive antiviral treatment. Our data suggest that the QuantiFERON®-CMV assay (Cellestis GmbH, a QIAGEN company) can make an important contribution to monitoring the immune status of LuTRs during posttransplant surveillance.
The authors thank Barbara Dalmatiner for technical support and Bettina Jaradat for assistance in acquiring the clinical data.
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation, and no part of this study was carried out or funded by a commercial organization.