Linking CMV Serostatus to Episodes of CMV Reactivation Following Lung Transplantation by Measuring CMV-Specific CD8+ T-Cell Immunity


  • G. P. Westall,

    1. Heart and Lung Transplant Unit, Department of Allergy, Immunology and Respiratory Medicine, Alfred Hospital, and Monash University Medical School, Melbourne, Victoria, Australia
    Search for more papers by this author
  • N. A. Mifsud,

    1. Department of Medicine, Monash University, Alfred Hospital, Melbourne, Victoria, Australia
    Search for more papers by this author
  • T. Kotsimbos

    Corresponding author
    1. Heart and Lung Transplant Unit, Department of Allergy, Immunology and Respiratory Medicine, Alfred Hospital, and Monash University Medical School, Melbourne, Victoria, Australia
      * Corresponding author: Tom Kotsimbos,
    Search for more papers by this author

* Corresponding author: Tom Kotsimbos,


CMV-specific immunity was assessed in a longitudinal cohort of 39 lung transplant recipients (LTR) who were followed prospectively from the time of transplant using a novel assay. At the time of surveillance bronchoscopy, CMV-specific CD8+ T-cell responses were assessed in the peripheral blood, using the QuantiFERON®-CMV assay, which measures IFN-γ-secreting T cells following stimulation with CMV peptides. In total, 297 samples were collected from 39 LTR (CMV D+/R−, n = 8; D+/R+, n = 18; D−/R+, n = 6; D−/R−, n = 7). CMV-specific T-cell immunity was not detected in any of the CMV D−/R− LTR. In CMV seropositive LTR levels of CMV immunity were lowest early posttransplant and increased thereafter. While levels of CMV-specific immunity varied between LTR, measurements at any one time point did not predict episodes of CMV reactivation. In CMV mismatched (D+/R−) LTR, primary CMV immunity was not observed during the period of antiviral prophylaxis, but typically developed during episodes of CMV reactivation. Measuring CMV-specific CD8+ T-cell function with the QuantiFERON®-CMV assay provides insights into the interrelationship between CMV immunity and CMV reactivation in individual LTR. A better understanding of these dynamics may allow the opportunity to individualize antiviral prophylaxis in the future.


Following lung transplantation the use of immunosuppressive medication that specifically targets T-cell function is a major contributor to the increased risk of cytomegalovirus (CMV)-related disease in these patients (1). Studies on the precise role of CMV-specific T cells following lung transplantation are limited (2), despite the potentially high associated morbidity and mortality of CMV disease in this patient group.

The CMV-naive recipient receiving a lung allograft from a CMV seropositive donor is at highest risk of CMV-related complications because of the absence of a primed CMV-specific immune response (3). These patients often demonstrate CMV infection, particularly following cessation of routine antiviral prophylaxis (4), suggesting that at this time they may not yet have acquired protective immunity against CMV. The ability to measure and time the de novo acquisition of host CMV-specific CD8+ T-cell immunity may provide insights into why CMV reactivation is so commonly observed in these ‘high-risk’ lung transplant recipients (LTR).

CMV seropositive LTR are also at risk of CMV reactivation/disease. T-cell-targeted immunosuppression that controls alloreactivity also impairs host antiviral immunity potentially resulting in CMV reactivation from the latent state (5). Measuring longitudinal changes in host CMV-specific immunity from the time of transplant may provide insights into why some, but not all CMV seropositive LTR experience episodes of CMV reactivation/disease.

The QuantiFERON®-CMV assay (Cellestis Ltd., Melbourne, Australia) is an in vitro diagnostic test that uses HLA-restricted CMV epitopes to stimulate CD8+ T cells in whole blood (6). Detection of interferon-γ (IFN-γ) by enzyme-linked immunosorbent assay (ELISA) identifies patients with CMV-specific CD8+ T cells. In a cohort of longitudinally followed LTR, we correlated CMV-specific CD8+ T-cell immunity in the blood with episodes of CMV reactivation/infection in the lung allograft.

Materials and Methods

Lung transplant cohort

Patients undergoing lung transplantation at the Alfred Hospital between November 2004 and August 2006 were included into a longitudinal analysis of CMV-specific immunity. All LTR received a standard triple-therapy immunosuppression regimen, consisting of prednisolone, azathioprine and a calcineurin inhibitor (cyclosporine A or tacrolimus). Induction therapy with the IL-2 receptor blocker, basiliximab was given as a calcineurin-sparing agent to 24 patients (irrespective of CMV serostatus pretransplant), who were judged pretransplant to be at higher risk of developing renal dysfunction peritransplant.

Patients at-risk for CMV infection or reactivation (donor [D] and/or recipient [R] CMV seropositive) were given i.v. ganciclovir (5 mg/kg) for 2 weeks followed by oral valganciclovir (900 mg/day) for a further 18 weeks. Primary CMV mismatches (D+/R−) also received CMV hyperimmune gammaglobulin on days 1, 2, 3, 7, 14, 21, 28 and 35 posttransplantation. As per our previous studies (4), episodes of histologically proven CMV pneumonitis or high CMV loads (>46 000 copies/mL in the broncholaveolar lavage [BAL], or in the blood) were treated with i.v. ganciclovir (5 mg/kg twice daily for 2 weeks). Subclinical CMV reactivation (<46 000 copies/mL BAL) in the primary CMV mismatched LTR were treated with oral valganciclovir (900 mg/day for 2 weeks).

Surveillance bronchoscopy, with BAL and transbronchial biopsy sampling, was performed at 0.5, 1, 2, 3, 6, 9, 12 and 18 months posttransplantation, or if clinically indicated. Acute allograft rejection was diagnosed according to standard criteria (7). All patients gave written consent and the study was approved by the Alfred Hospital ethics committee.

Blood and BAL samples

At the time of bronchoscopy, peripheral whole blood samples (9 mL) were collected in Vacuette sodium heparin tubes (Greiner Bio-one, Kremsmunster, Austria) for the QuantiFERON®-CMV assay. BAL samples were collected and pooled as previously described (2).

QuantiFERON®-CMV assay

Peptides from HLA-restricted (HLA-A1, -A2, -A3, -A11, -A23, -A24, -A26 and HLA-B7, -B8, -B27, -B35, -B40, -B44, -B51, -B57, -B58, -B60) CD8+ T cell epitopes of various human CMV peptides derived from phosphoprotein (pp) 65 and pp 50, glycoprotein B (gB) and the immediate early 1 (IE-1) antigen were used (8). These epitopes with their known HLA specificities can be predicted to provide MHC coverage for greater than 95% of any given racial population ( Briefly, 1 mL aliquots of heparinized blood were incubated overnight (37°C) with either HLA-restricted CMV peptide epitopes (2 μg/mL of each of 21 peptides), sterile PBS (negative control) and phytohemagglutinin (PHA; positive control). The supernatant (plasma) was collected for IFN-γ quantification by ELISA. A test is considered positive if the IFN-γ response to CMV peptides is significantly above (>0.2 IU/mL IFN-γ) the negative control value. A positive response to CMV peptides, without a response to mitogen, is a valid result suggesting established cellular immunity. A low response to mitogen (<0.5 IU/mL) indicates an indeterminate result if the CMV peptide response is also negative. All LTR samples were blinded for both identity and clinical status to Cellestis Ltd. who performed the QuantiFERON®-CMV assay.

Quantification of CMV load

At the time of routine surveillance bronchoscopy, CMV load was measured in the BAL (COBAS Amplicor CMV monitor test, Roche Diagnostic Systems, NSW, Australia) as described elsewhere (8). CMV load in the BAL supernatant was reported as copies per milliliter (copies/mL). The lower and upper limits of the assay are 400 and 100 000 copies/mL, respectively. CMV load was only assessed in the blood if extrapulmonary CMV reactivation was suspected.

Statistical analysis

Numerical data were expressed as means and standard errors of the mean (SEM). Group comparisons were performed using the Spearman's rank correlation coefficient. Between-group analysis was made with the unpaired t-test. Statistical significance was defined as p < 0.05.


The relationship between antiviral prophylaxis, CMV reactivation and CMV-specific immunity, as measured by the QuantiFERON®-CMV assay, was evaluated in 39 LTR (mean age 47 years, range 19–65; 21 males; 35 double lung transplants) from the time of transplant (mean follow-up 16 months, range 6–28 months).

QuantiFERON®-CMV and mitogen responses

CMV-specific CD8+ T-cell immunity was assessed in the blood using the QuantiFERON®-CMV assay on a total of 297 posttransplant samples (mean 8 samples per patient; range 3–13). All samples were collected at the time of bronchoscopy. In this study, 94/297 (30.3%) of samples were classified as indeterminate, on the basis of a negative mitogen response. Mitogen responses were more often negative early posttransplant (<8 weeks)—the period of maximal immunosuppression—compared to later time points (2.4 ± 0.4 IU/mL vs. 5.9 ± 0.3 IU/mL, p < 0.0001).

CMV serostatus and infection/reactivation

The CMV serostatus of the LTR were as follows: D+/R+, n = 18; D+/R−, n = 8; D−/R+, n = 6; D−/R−, n = 7. While CMV reactivation in the lung allograft, as diagnosed by PCR, was common, no episodes of CMV pneumonitis (histological detection of CMV inclusions) were observed. The incidence of CMV reactivation was no higher in those LTR who received induction therapy with basiliximab (12/24 vs. 4/15, p = ns). Patterns of CMV infection/reactivation, as measured by PCR, differed between LTR and could be largely predicted by their individualized risk of CMV infection/reactivation, as determined by the D/R CMV serostatus at the time of transplant—that is highest in the CMV mismatched LTR. The 7 CMV D−/R− LTR did not experience CMV reactivation nor develop CMV immunity during the study.

Primary CMV infection—the detection of CMV DNA in the BAL of D+/R− LTR—was identified in 7/8 CMV mismatched patients, and in all cases was seen following cessation of antiviral prophylaxis (mean 31 weeks; range 27–39 weeks). CMV DNA became undetectable in all 7 LTR following the recommencement of antiviral therapy.

CMV reactivation in the lung allograft was seen in 11/24 CMV seropositive LTR, and in all cases occurred following cessation of antiviral prophylaxis (mean 38 weeks; range 28–52 weeks). In 6 patients CMV reactivation was transient, at low level (median 5755 copies/mL BAL supernatant; range 963–17 300) and did not require treatment. This contrasted with 5 LTR who developed significant CMV reactivation in the lung allograft (>100 000 copies/mL) requiring i.v. ganciclovir to clear the virus. Risk factors for viral recrudescence were identified in 4/5 CMV seropositive LTR who developed significant CMV reactivation: posttransplant hypogammaglobulinemia, n = 2; X-linked agammaglobulinemia, n = 1; recent augmentation of immunosuppression following an episode of A2 rejection, n = 1.

CMV reactivation was assessed in both the BAL and blood in a smaller cohort of 72 paired samples. Overall, in 55 paired samples CMV DNA was not detected in either the blood or the BAL, in 11 paired samples CMV was detected in the BAL but not the blood, in 6 paired samples CMV DNA was detected in both compartments and on no occasion was CMV detected in the blood in the absence of CMV in the BAL.

CMV immunity: CMV mismatched LTR

The temporal relationship between CMV infection in the lung allograft and the detection of primary CMV immunity in the CMV mismatched LTR is shown in Table 1. In CMV D+/R− LTR, viral immunity was first detected on average 45 weeks (range 17–113 weeks) posttransplant. Of the 7 CMV mismatched LTR who developed primary CMV infection, 2 developed CMV immunity contemporaneous to the detection of CMV, 3 experienced primary infection prior to the detection of CMV immunity and 2 LTR experienced CMV reactivation after the acquisition of CMV immunity.

Table 1.  Development of CMV reactivation and primary CMV immunity in CMV mismatched (D+/R−) LTR
Patient1Initial CMV reactivation (weeks posttransplant)Initial CMV load (copies/mL BAL)Time to positive QuantiFERON®-CMV assay (weeks posttransplant)
  1. 1All patients received antiviral prophylaxis until 20 weeks posttransplant.

127  519027
227 72 80038
328  923026
430100 000113 
531 18 70076
634  698034
739100 00033
8CMV DNA negative     017

CMV immunity: CMV seropositive LTR

The QuantiFERON®-CMV response was positive at some point posttransplant in all but one of the CMV seropositive LTR. The exception was a patient whose haplotype (HLA-A30,-; HLA-B13,-) included epitopes not covered by the pooled HLA-restricted CMV peptides that constitute the QuantiFERON®-CMV assay (6). Data from this LTR were not included in subsequent group analysis. Levels of CMV immunity were highly variable between individual CMV seropositive LTR (n = 23) at any given time point posttransplant. The absolute levels of CMV immunity at any one time, as measured by the QuantiFERON®-CMV assay, did not predict episodes of CMV reactivation, nor were the QuantiFERON®-CMV responses any lower in those CMV seropositive LTR who experienced CMV reactivation.

The longitudinal analysis of the CMV seropositive LTR who experienced either no or minimal CMV reactivation (n = 18) revealed two broad patterns in CMV immunity over time. In half of these ‘stable’ patients, the QuantiFERON®-CMV responses did not fluctuate with time from transplant (Figure 1A), a pattern that contrasted to that seen in the other 9 ‘stable’ CMV seropositive LTR who all demonstrated a progressive increase in QuantiFERON®-CMV responses with time from transplant (Figure 1B). Reflecting these trends, the QuantiFERON®-CMV responses in the CMV seropositive LTR were overall lower in the first 8 weeks posttransplant compared to their level beyond 8 weeks from transplant (3.2 ± 0.6 IU/mL vs. 4.8 ± 0.4 IU/mL; p < 0.05).

Figure 1.

CMV immunity in stable CMV seropositive LTR. CMV immunity, as measured with the QuantiFERON®-CMV assay, was assessed at multiple time points posttransplant in 18 CMV seropositive LTR who did not develop significant CMV reactivation (>46 000 copies/mL BAL). In 9 CMV seropositive LTR the QuantiFERON®-CMV responses did not change with time from transplant (Figure 1A—representative plots from 2 LTR), which contrasted with the other 9 CMV seropositive LTR in whom the QuantiFERON®-CMV responses tended to increase with time from transplant (Figure 1B—representative plots from 2 LTR).

The QuantiFERON®-CMV responses in the 5 CMV sero-positive LTR who developed significant CMV reactivation (>100 000 copies/mL BAL) were less uniform (Figure 2). In three cases (Figure 2A–C), we observed a striking decrease in the QuantiFERON®-CMV responses prior to the episode of CMV reactivation.

Figure 2.

CMV immunity in CMV seropositive LTR with significant CMV reactivation. QuantiFERON®-CMV responses plotted against time from transplant in 5 CMV-seropositive LTR (A–E) who developed significant CMV reactivation (>100 000 copies/mL BAL; depicted by solid arrow) following cessation of antiviral prophylaxis at 20 weeks posttransplant.


We have used the QuantiFERON®-CMV assay to delineate longitudinal patterns of CMV immunity in the first year posttransplant in a prospectively followed cohort of 39 LTR, both receiving and having ceased antiviral prophylaxis. In CMV mismatched LTR (D+/R−), the acquisition of de novo primary CMV immunity is delayed and may partly explain why CMV reactivation is so prevalent following cessation of antiviral prophylaxis. In these patients, this study is unable to answer whether antiviral prophylaxis is suppressing CMV replication to such low levels that the host is unable to prime a primary antiviral cellular immune response. If this is the case, then it would be likely that CMV infection in the mismatched LTR in particular would continue to be common following cessation of antiviral drugs, even if the duration of prophylaxis was extended to 1 year or longer posttransplant, as has been previously described (9).

In CMV seropositive LTR levels of CMV, immunity in the blood generally increase with time from transplant, but the absolute magnitude of the QuantiFERON®-CMV responses at any given point does not predict for episodes of CMV reactivation. In the CMV seropositive LTR, all of whom should have established cellular immunity to CMV pretransplant, this study provides a rationale for the continued use of empirical antiviral prophylaxis in this patient group. In the early posttransplant period (<8 weeks), levels of CMV immunity were significantly depressed compared with later time points; most likely a consequence of the higher net levels of immunosuppression early posttransplant. Individual levels of CMV immunity varied markedly between LTR, which is perhaps not unexpected given the considerable heterogeneity in levels of CMV immunity seen in immunocompetent normal subjects (10). Consequently, the QuantiFERON®-CMV result from an individual LTR at any one time point cannot be used as a diagnostic tool to diagnose CMV disease. We were, however, interested to see whether changing trends in CMV immunity over time could predict future episodes of CMV reactivation. Notably, in 3 CMV seropositive LTR who developed significant CMV reactivation in the lung allograft (BAL CMV load >100 000 copies/mL), there was a decrease in CMV immunity in the peripheral blood prior to the episode of CMV recrudescence in the lung allograft. Further studies will be required to verify the consistency of this observation, which if confirmed could provide the clinician with a predictive test of future CMV infection/disease, and potentially an opportunity to initiate preemptive treatment.

There are a number of studies in LTR using quantitative molecular diagnostic assays to detect CMV DNA that have demonstrated that CMV infection/reactivation is more common in the lung allograft than in the blood, and that high viral loads correlate with CMV disease and the histological detection of CMV inclusions in the allograft (11–13). Our own group has shown in LTR, CMV DNA is three times more likely to be detected in the lung allograft than in the blood, and that when CMV is detected in the blood it is always also present in the BAL (4). As such, it is now our clinical practice to only measure CMV load in the blood if we clinically suspect extrapulmonary CMV disease. With better antiviral prophylaxis, viral load testing and preemptive treatment strategies, the histological detection of CMV inclusions is not as common. In particular, it is widely accepted that a negative viral load test is an excellent negative predictor for CMV reactivation/disease, and it is our own experience that if antiviral treatment is targeted to a rising CMV load in the BAL then we do avoid clinical/histological CMV disease. It does, however, need to be appreciated that results and threshold values for treatment of CMV are likely to differ between laboratories that use different sampling methods and amplification protocols (14).

The QuantiFERON®-CMV assay has been developed to measure CMV immunity in the blood (6) and not the lung allograft. Given that CMV infection/reactivation is more likely in the lung allograft in LTR, it can be questioned whether CMV immunity in the blood, as measured with the QuantiFERON®-CMV assay, accurately reflects antiviral immunity in the lung allograft. Prior studies by our own group using MHC class I-restricted tetramers have partly addressed this issue. We have shown that the development of CMV immunity in the periphery does indeed parallel that in the lung allograft following lung transplantation, albeit at reduced levels (2). We have also demonstrated that CMV-specific CD8+ T cells in the BAL were no more differentiated than those in the blood, suggesting that activated CD8+ T cells were not being redistributed to and sequestered at the site of active infection. Similarly, McDyer and colleagues in a smaller study of 4 CMV-seroconverted D+/R− LTR demonstrated similar to slightly higher frequencies of CMV-specific CD8+ T cells in the BAL compared to the peripheral blood (15). It is therefore likely that QuantiFERON®-CMV responses in the blood can be used as a surrogate marker of CMV immunity in the lung allograft, although this will need to be formally analyzed in future studies.

A high percentage of results were indeterminate as defined by the mitogen response (positive control) being less than 0.5 IU/mL; most likely due to the T cell directed immunosuppression that all patients received. One limitation of the QuantiFERON®-CMV assay was exemplified by the 1 CMV seropositive LTR who had an unusual haplotype (HLA-A30,-; HLA-B13,-) that included antigens not covered by the pooled HLA-restricted CMV epitopes that constitute the QuantiFERON®-CMV assay. Prearmed with this information the test would be predicted to be negative. In the clinical setting, knowledge of the patient's haplotype will be essential in interpreting the results of the Quanti FERON®-CMV assay.

In summary, we have provided the first demonstration of the applicability of the QuantiFERON®-CMV assay in a prospectively followed cohort of LTR. In CMV mismatched LTR, the QuantiFERON®-CMV assay accurately tracks the development of de novo CMV immunity and further studies will be required to assess whether the assay could be used to tailor the length of antiviral prophylaxis in the individual LTR, thereby further reducing the likelihood of CMV infection in ‘at-risk’ patients. In the CMV seropositive LTR, lower levels of CMV immunity early posttransplant highlights the need to maintain antiviral prophylaxis during this period.


This study was supported by research grants received from Cellestis Ltd., Melbourne, Australia. Glen Westall is a recipient of a National Health and Medical Research Council (NHMRC) postgraduate research scholarship. Nicole Mifsud is supported by a NHMRC Peter Doherty postdoctoral fellowship.