SEARCH

SEARCH BY CITATION

Keywords:

  • Activating KIR genes;
  • centromere KIR;
  • CMV infection;
  • kidney transplantation;
  • telomere KIR

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Materials
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

Cytomegalovirus (CMV) infection is a common complication after organ transplantation. Previous studies have demonstrated that activating killer-cell immunoglobulin-like receptors (KIR) may reduce the rate of CMV infection. KIR genes can be divided into haplotype A (containing a fixed set of inhibitory receptors) and haplotype B (carrying additional activating KIR genes). The KIR locus is divided into a centromeric and a telomeric portion, both of which may carry A or B haplotype motifs. We studied a cohort of 339 kidney transplant recipients to elucidate which KIR genes protect from CMV infection. CMV infection occurred in 139 patients (41%). Possession of telomeric (hazard ratio 0.64, 95% confidence interval 0.44–0.94, p = 0.02) but not centromeric (HR 0.86, 95% CI 0.60–1.23, p = 0.41) B motifs was associated with statistically significant protection from CMV infection. Due to linkage disequilibrium, we were not able to identify a single protective gene within the telomeric B complex (which may contain the KIR2DS1, KIR3DS1, KIR2DL5A and KIR2DS5 genes). The presence of known or putative ligands to activating KIR did not significantly modify the influence of telomeric B group genes. We confirm that B haplotypes protect from CMV infection after kidney transplantation and show that this arises from telomeric B haplotype genes.


Abbreviations: 
CMV

cytomegalovirus

HR

hazard ratio

KIR

killer-cell immunoglobulin-like receptors

NK cell

natural killer cell

PCR

polymerase chain reaction

SOT

solid organ transplantation

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Materials
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

Cytomegalovirus (CMV) is the most common viral infection after hematopoietic cell and solid organ transplantation (SOT) (1). Recent evidence has pointed to a critical role of natural killer (NK) cells in the control of CMV infection in transplant recipients. Immunosuppressive drugs mandatory to avoid organ rejection inhibit the T-cell activity but interestingly, NK-cell functions do not appear to be strongly affected in the studies on SOT that have included in vivo data (2).

The effector mechanisms of NK cells are mediated by the balance of activating and inhibitory receptors and their ligands expressed by target cells. Several families of receptors have been described. Among them, the killer-cell immunoglobulin-like receptors (KIR) are of special interest in the context of CMV infection. The KIR family includes 15 genes divided into 9 inhibitory KIR and 6 activating KIR. Depending on an individual's KIR genotype, patients’ NK cells may express varying numbers of activating KIR receptors. Carrying more activating KIR genes has been demonstrated to significantly reduce the incidence of CMV infection after transplantation (3–6). KIR are not expressed exclusively on NK cells, but are also found on CMV-specific CD8+ T cells and gamma–delta T cells (7,8). Small patient numbers and a high linkage disequilibrium between the different activating KIR genes have so far precluded identification of a single KIR receptor responsible for protection from CMV infection.

KIR genes can be divided into haplotypes A and B according to the gene content. KIR haplotypes A have simple, fixed gene numbers and haplotypes B have a more variable gene content with additional activating KIR genes. The position of each KIR gene is fixed, and centromeric (Cen) or telomeric (Tel) motifs can be defined for each haplotype, i.e. Cen-A, Tel-A, Cen-B and Tel-B. In hematopoietic stem cell transplantation, recent data have demonstrated that both Cen-B and Tel-B motifs strongly contribute to relapse protection and improve survival (9). Pregnancy disorders have been shown to be less frequent in women that possess Tel-B (10). In CMV infection, the role of specific KIR genes or groups of KIR genes remains to be clarified. To address this question we combined two large cohorts of patients undergoing kidney transplantation and we analyzed the importance of KIR gene position in haplotypes A and B with regard to the risk of CMV infection.

Patients and Materials

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Materials
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

Patients

Three hundred and thirty-seven patients undergoing kidney transplantation at two centers (Geneva University Hospital, Geneva, Switzerland, n = 217; and Basel University Hospital, Basel, Switzerland, n = 122) between 2000 and 2008 were combined in this analysis. Induction and maintenance immunosuppressive regimens were as previously described (4,5) and are summarized in Table 1. CMV serological constellation was assessed by detecting anti-CMV IgG antibodies at time of transplant in donors and recipients. Starting in 2003, patients with high-risk constellation (donor positive, recipient negative) received prophylaxis with valgancyclovir for 60 (Geneva) or 120 (Basel) days, respectively. Patients were screened for CMV (by pp65 antigenemia or by quantitative polymerase chain reaction, PCR) regularly during follow-up (at least once/month). Assessment of CMV replication by pp65 antigenemia only was used in patients transplanted before 2002 (n = 42), whereas all patients transplanted after 2002 were tested by PCR. Testing for pp65 was performed using a commercial kit (CINAkit; Argene, Biosoft, Varilhes, France). Quantitative PCR for CMV was performed as previously described (Geneva: Ultrasensitive Cobas Amplicor CMV DNA Monitor, Roche Diagnostic Systems, Branchburg, NJ, USA (4); Basel: Real-time PCR using previously published primers purchased from Applied Biosystems, Rotkreuz, Switzerland) (11). Written informed consent was obtained from all study participants and the study was IRB approved at both centers.

Table 1.  Patient characteristics
Patient age at transplantation
 Median (range)55 ( 5–86)
Gender
 Male197 ( 58)
 Female142 ( 42)
Type of transplant (n, %)
 Living donor133 ( 39)
 Deceased donor206 ( 61)
CMV serology (n, %)
 Don neg/Rec neg52 ( 15)
 Don neg/Rec pos72 ( 21)
 Don pos/Rec neg67 ( 20)
 Don pos/Rec pos139 ( 41)
Antibody induction (n, %)
 none23 ( 7)
 Anti-CD25 mab273 ( 81)
 ATG43 ( 13)
Maintenance immunosuppression (n, %)
 Tacrolimus, MMF, prednisone146 ( 43)
 Cyclosporine, MMF, prednisone121 ( 36)
 Tacrolimus, MMF, papamycine46 ( 14)
 Tacrolimus, azathioprine, prednisone12 ( 4)
 Other14 ( 4)

HLA and KIR Genotyping

HLA-A and -B typing was performed by the serology technique (Biotest AG, Dreieich, Germany) and molecular biology (reverse sequence-specific oligonucleotide method, One Lambda, Canoga Park, CA, USA), and HLA-C typing was performed exclusively by molecular biology (One Lambda). KIR genotyping was performed using the KIR Genotyping SSP kit (Pel-Freez, Rogers, AK, USA) or using a reverse sequence-specific oligonucleotide method (OneLambda) according to the providers (4,5) Both systems detected the inhibitory receptors KIR2DL1, 2DL2, 2DL3, 2DL5, 3DL1, 3DL2, 3DL3, the activating receptors 2DL4, 2DS1, 2DS2, 2DS3, 2DS4, 2DS5, 3DS1 and the common variants of 2DL5, 2DS4 and 2DP1 + 3DP1 (not expressed).

Statistical Analysis

Cumulative incidence of CMV infection was compared in Cox models adjusted for CMV donor and recipient serostatus at time of transplant and for valgancyclovir prophylaxis. Valgancyclovir administration was included in the Cox models as a time-dependent covariate. KIR genotypes were grouped into AA if they contained only the canonical group A haplotype genes (KIR3DL3, KIR2DL3, KIR2DL1, KIR2DL4, KIR3DL1, KIR2DS4 and KIR3DL2). Any genotype containing additional KIR genes is referred to as a BX, as it contains at least one group B haplotype (12). BX genotypes can be further dichotomized according to the localization of any additional KIR genes: KIR2DS2, KIR2DL2, KIR2DL5B and KIR2DS3 are localized centromerically of KIR2DL4 and are referred to as Cen-B genes. KIR3DS1, KIR2DL5A, KIR2DS5 and KIR2DS1 are localized telomerically of KIR2DL4 and are referred to as Tel-B genes (13). Patients are classified as Cen-BX or Tel-BX if they carry at least one of the respective group B genes, whereas the remaining are referred to as Cen-AA and Tel-AA. The A and B haplotypes with the KIR gene contents of Tel and Cen regions are depicted in Figure 1. The HLA-C1, -C2 and Bw4 ligands of specific KIR genes are also indicated.

image

Figure 1. Schematic view of representative KIR gene haplotypes A and B. The framework KIR genes are in white, the activating KIR genes are in gray and the inhibitory KIR genes are in black. The HLA-C1, C2 and HLA-Bw4 cognate ligands of the inhibitory KIRs are depicted. «?» indicates a degree of uncertainty with regard to the HLA ligands of the activating KIR gene 2DS1 and 3DS1. KIR A haplotypes have a fixed number of KIR gene, KIR haplotypes B have variable gene content.

Download figure to PowerPoint

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Materials
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

CMV infection occurred in 139 of the 339 patients in this cohort (41%). In order to determine which activating KIR genes protect from CMV infection, genotypes were grouped into those containing Tel and Cen A and B haplotype motifs. One hundred patients (29%) carried AA genotype motifs exclusively in both the Cen and the Tel part of the KIR locus (Cen-AA/Tel-AA). Forty-seven patients (14%) carried a mixed Cen-AA/Tel-BX genotype. Eighty-one patients carried a mixed Cen-BX/Tel-AA genotype (24%). The remaining 111 patients carried B genotype motifs in both the Cen and the Tel part (Cen-BX/Tel-BX, 33%). Results are summarized in Table 2. The frequencies found are similar to those recently reported in a large cohort of healthy donors of hematopoietic stem cell grafts (9).

Table 2.  KIR genotype distribution
 Tel-AATel-BXTotal
Cen-AA100 ( 30%)47 ( 14%)147 ( 43%)
Cen-BX81 ( 24%)111 ( 33%)192 ( 57%)
Total181 ( 53%)158 ( 47%)339 ( 100%)

When analyzed separately, the presence of a Cen-B gene motif was less protective than the presence of a Tel-B motif (hazard ratio Cen-BX versus Cen-AA: 0.75, 95% confidence interval 0.54–1.06, p = 0.10; HR Tel-BX versus Tel-AA: 0.61, 95% CI 0.43–0.87, p = 0.006). Adjusted cumulative incidence curves for these analyses are shown in Figure 2. When the presence of both Cen-B and Tel-B motifs was analyzed as covariates in a single Cox model, the presence of a Tel-BX genotype retained statistical significance (HR 0.64, 95% CI 0.44–0.94, p = 0.02), whereas the presence of a Cen-BX genotype conferred only minimal protection (HR 0.86, 95% CI 0.60–1.23, p = 0.41, Table 3). These results suggest that the marginally protective effect of carrying Cen-B genes which was found in univariate analysis (Figure 2A) was largely due to the linkage disequilibrium of Cen-B genes with Tel-B genes.

image

Figure 2. Adjusted cumulative incidence of cytomegalovirus infection. Adjusted cumulative incidence of CMV infection after kidney transplantation in patients carrying at least on group B gene (BX) or only group A genes (AA) centromeric (left panel) or telomeric (right panel) of KIR2DL4.

Download figure to PowerPoint

Table 3.  Risk factors for CMV reactivation
 Hazard ratio95% CIp-Value
  1. *Coded as a time-dependent covariate.

CMV serostatus
 D+/R+1.00--
 D−/R−0.060.02–0.25<0.001 
 D−/R+0.970.63–1.500.90
 D+/R−1.851.14–3.010.01
 Valgancyclovir*0.200.07–0.560.002
Antibody induction
 Anti-CD25-Mab versus none1.330.52–3.390.55
 ATG versus none1.200.52–2.780.67
 Treatment for rejection*1.220.80–1.850.36
KIR genotype
 Cen-BX versus Cen-AA0.860.60–1.230.41
 Tel-BX versus Tel-AA0.640.44–0.940.02

Recent evidence from mouse strains has demonstrated features of adaptive immunity associated with NK cells (14). This raises the question whether antiviral effects of KIR genes are only found in R+ individuals with previous exposure to CMV or also in R− individuals. Interestingly, we found protective effects in both populations (R+: HR 0.65, 95% CI 0.43–0.98, p = 0.04; R− HR 0.45, 95% CI 0.22–0.88, p = 0.02). These data would indicate that the protective effect associated with Tel-B genes does not rely on previous exposure to CMV which might have induced the formation of memory NK cells. Tissue invasive CMV disease was recorded in six patients (colitis in five and pneumonitis in one). Of these six patients, five carried a Cen-A/Tel-A genotype, whereas one patient carried a Cen-A/Tel-B genotype.

We next analyzed genes within the Tel-B locus for their contribution to protection from CMV infection. Due to substantial linkage disequilibrium, hazard ratios were in a similar range for all Tel-B genes. In descending order, protective effects were greatest for KIR2DS1 (HR present versus absent 0.73, 95% CI 0.51–1.04, p = 0.09), and KIR3DS1 (HR 0.75, 95% CI 0.52–1.07, p = 0.12), followed by KIR2DS5 (HR 0.85, 95% CI 0.59–1.22, p = 0.38) and KIR2DL5A (HR 0.85, 95% CI 0.57–1.27, p = 0.43).

KIR2DS1 has recently been shown to be the only activating KIR receptor binding HLA class I antigens (namely those with the ‘C2’ specificity, e.g. HLA-Cw2, -Cw4, -Cw5, -Cw6) in vitro.(15) KIR3DS1 on the other hand—while not binding HLA in vitro—has been shown to interact with HLA-B antigens carrying the Bw4 specificity in patients with chronic HIV infection (16). An HLA-C2 ligand was detected in 223 of the 339 patients (66%). The hazard ratio for CMV infection of patients carrying the KIR2DS1 gene was comparable in HLA-C2 positive (HR 0.71, 95% CI 0.46–1.10, p = 0.14) and HLA-C2 negative patients (HR 0.79, 95% CI 0.42–1.50, p = 0.48). Two hundred fifteen patients carried the Bw4 epitope (64%), the putative ligand for KIR3DS1. Hazard ratios regarding the protective effect of KIR3DS1 were lower in Bw4+ compared to Bw4− patients; however this difference did not reach statistical significance (0.70, 95% CI 0.45–1.09, p = 0.11 in Bw4+ patients versus 0.88, 95% CI 0.47–1.66, p = 0.70 in Bw4− patients, interaction p-value comparing the effect = 0.69). Collectively, these data suggest that if indeed KIR2DS1 and/or KIR3DS1 are directly responsible for protection from CMV infection, then this protective effect may not be mediated via binding to HLA class I ligands.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Materials
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

Previous analyses have shown that carrying activating KIR genes protects patients from CMV infection after transplantation of solid organs or hematopoietic stem cells. Microarray analysis in kidney transplant patients experiencing primary CMV infection revealed that KIR2DS2, KIR2DS4, KIR3DL2 are upregulated on CMV-specific CD8+ T cells during the peak of the CMV infection, but returned to baseline levels in the latency phase (8). Using flow cytometry, KIR2DL2/S2 and KIR 2DL1/S1 have been found to be increased on T cells of CMV-infected healthy patients (7). The impossibility to discriminate by flow cytometry between homologous KIR2DL and KIR2DS molecules, due to the crossreactivity of available mAbs, limits the interpretation of the phenotypic data. In kidney transplants in particular, an inverse correlation between number of activating KIR genes and the probability of CMV infection was described in two independent cohorts (4,5). Limited patient numbers and strong linkage disequilibrium within the KIR locus precluded any further analysis attempting to identify a single KIR gene responsible for this effect. In the current study we attempted to better define the relationship between KIR genes and CMV infection by combining patients from two large cohorts.

Similar to recent studies in patients with acute leukemia (9) and women with reproductive failure (10) we classified group B KIR genes according to their localization with respect to KIR2DL4 into those with Cen and Tel localization. Patients may carry B genes in none, either or both the Cen and Tel region. Using this approach, we aimed to reduce the number of potential candidate genes.

Our analysis confirms that group B KIR haplotypes protect from CMV infection in patients after kidney transplantation and further maps the relevant locus to the Tel part of the KIR B haplotype, which may contain the genes coding for KIR2DS1, KIR3DS1, KIR2DS5 and KIR2DL5A. Due to high linkage disequilibrium between these genes, we were unable to ascribe the protective effect of the Tel-B gene motif to one single locus. Synergistic effects of multiple Tel-B KIR genes can also not be excluded. By analogy with an antiviral T-cell immune response in which multiple peptides/antigens stimulate several T-cell clones (17), the presence of multiple activating KIR genes of the Tel-B motif could be more efficient than a single one.

We were also unable to detect a significant modulation of the protective effect of Tel-B KIR genes by the presence or absence of HLA class I KIR ligands. One possible explanation is that the ligands recognized by Tel-B KIR genes are not HLA molecules. Indeed, recent data have demonstrated that—in addition to their function as receptors for self HLA—three-domain KIR (among them KIR3DS1) act as receptors for viral CpG oligodexynucleotides (18) which might explain the lack of association with HLA found in this study. Only functional studies will ultimately be able to discern which gene(s) might be involved in the protective effect of group B haplotypes and what ligands they recognize. Of particular interest is the fact that similar degrees of relative protection were noted both for CMV-naive (IgG-negative) patients as well as for individuals which had undergone primary CMV infection prior to kidney transplantation. Recent studies using mouse models have unraveled features of adaptive immunity in murine NK cells, among them the evidence of a memory NK-cell population against murine CMV (14). Our results indicate that the KIR-mediated protection from human CMV infection described in this study does not rely on previous exposure to CMV.

We believe that our data are of interest with regard to the prophylactic/preemptive strategies of CMV prevention after SOT. The cost of CMV prophylaxis is high and side effects of antiviral treatment are not unusual (19). Therefore, it is of great interest to know how we can restrain the therapy to the patients who really need it. The role of genetic factors in identifying patients more prone to CMV infection is therefore of great interest.

Our data confirm the protective effect of group B KIR haplotypes from CMV infection and maps the locus of resistance to the Tel part of the KIR gene cluster containing the genes coding for KIR2DS1, KIR3DS1, KIR2DS5 and KIRDL5A. Knowing the KIR haplotype motifs of solid organ transplant recipients could help to predict the risk of CMV infection and better stratify the patients for antiviral prophylaxis.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Materials
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

M.S. was supported by the Swiss National Science Foundation (grant PP00P3_128461/1) and by the University Hospital Basel (VFWAWF-Fonds). C.H. was supported by the Swiss National Science Foundation (grant PP00B-114850). The project was support by a grant of the Department of Medicine, Geneva University Hospital.

Disclosure

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Materials
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Materials
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References