Variant genotypes of FcγRIIIA influence the development of Kaposi's sarcoma in HIV-infected men

Kaposi's sarcoma (KS) is the most common malignant condition associated with human immunodeficiency virus-1 (HIV) infection.1,2 Before the onset of the HIV epidemic, KS was a rare disease confined to specific risk groups: elderly men of Mediterranean or Ashkenazi Jewish background, Sub-Saharan Africans, and immuno-suppressed patients.2-6 Since the initial description of KS, there has been strong evidence to suggest that an infectious agent contributes to the pathogenesis of this disease. Recently, a KS-associated herpesvirus (KSHV), later designated as human herpesvirus-8 (HHV-8), has been determined to be the likely causative microbial agent critical in the pathogenesis of KS.7-10However, this virus appears to be necessary but not sufficient for development of KS.

Current insights into the pathogenesis of KS in HIV-infected individuals indicate that in its early stage, KS is an angio-proliferative inflammatory condition induced by infection with HHV-8.11,12 In the late stages of KS, lesions may develop a malignant phenotype and there is some evidence that they can resemble a true sarcoma as defined by monoclonality.13 The immunopathogenesis of KS includes the activation of lymphocytes, resulting in increased inflammatory cytokine production, activation of endothelial cells, and production of virally encoded angiogenic factors by HHV-8-infected cells.7-10,14-18 It is postulated that an increase in pro-inflammatory cytokine production, notably IFN-γ, TNF-α, IL-1, and IL-6 as well as HIV-1 tat protein, promotes activation and growth of endothelial cells, expression of adhesion molecules and integrins, and release of angiogenic molecules.14-19 Hence, a significant disruption of the extra- and intracellular signaling pathways by cytokines and other molecules of innate immunity appears to be a hallmark of KS in HIV-infected individuals.

The low-affinity Fc gamma receptors (FcγR) for IgG couple humoral and cellular immunity by mediating interaction between antibodies and effector cells, such as professional phagocytes (ie, neutrophils, monocytes, and macrophages) and natural killer (NK) cells.20 Surface FcγR on effector cells can direct phagocytosis, antibody-dependent cell-mediated cytotoxicity (ADCC), and activation of cytokine pathways.21,22 The structural heterogeneity of the genes that encode the FcγRs reflects the functional diversity of these receptors. The genes that encode the low-affinity FcγR map to a region of chromosome 1q22 are characterized by a high degree of sequence homology.23,24 

Polymorphic forms of the low affinity receptors,FcγRIIA, FcγRIIIA, andFcγRIIIB, have been described and are currently the focus of efforts to identify heritable risk factors for a range of diseases.25-31 Variant alleles have been described that have a high frequency (greater than 25%) in the general population. These alleles exhibit differing functional characteristics both in vitro and in clinical association studies. For example, a polymorphism in FcγRIIA at amino acid 131 results in altered binding affinity to IgG2.27 The FcγRIIIa-158 V isoform, which lies in the extracellular domain and affects ligand binding, has a higher affinity for IgG1, IgG3, and IgG4 than FcγRIIIa -158 F.28,29 Interestingly, FcγRIIIa is highly expressed on NK cells and monocytes. Neutrophils from NA2 homozygotes ofFcγRIIIB bind human IgG3 less avidly than those from NA1 homozygotes and exhibit a lower level of phagocytosis of erythrocytes sensitized with IgG1 and IgG3 anti-Rhesus D monoclonal antibody.21 Recently, it was shown that variant genotypes of these 3 FcγRs were randomly distributed in the populations studied, suggesting that each of the 3 FcγRs, namely,FcγRIIA, FcγRIIIA, andFcγRIIIB, provides a distinct set of functions in host immune responses, despite the high degree of homology among their sequences.32 

Clinical association studies analyzing these variant genotypes have shown a correlation between FcγR genotype and disease susceptibility or outcome, particularly in cohorts with underlying defects in immune function. Specifically, FcγRIIA andFcγRIIIB genotypes have been associated with granulomatous and rheumatologic complications of chronic granulomatous disease, renal complications of both systemic lupus erythematosus and Wegener's granulomatosis, as well as infection with encapsulated bacteria in HIV-infected and other immunocompromised populations.33-36 Although variant alleles of the FcγRs have not been associated with the occurrence or progression of HIV infection, clinical association studies investigating other genes, in particular variant alleles of the chemokine receptor family, have demonstrated a strong influence on infection and progression of HIV disease.37-44 

The above mentioned observations led us to consider the possibility that variant genotypes of low affinity FcγRs might influence susceptibility to KS. To investigate this possibility, we chose to genotype the 3 candidate FcγRs loci,FcγRIIA, FcγRIIIA, andFcγRIIIB, in a cohort of HIV patients with or without KS.33 

The first cohort tested, population I, consisted of 119 deceased white males enrolled on NCI protocols, who had acquired HIV through sex with other men. None of the patients received highly active antiretroviral therapy (HAART) and each died before 1996. The second cohort, population II, consisted of 131 HIV-infected patients (61 patients with KS, 70 patients without KS). The demographic and clinical profile of the patients in the confirmatory cohort, population II, was comparable to the first cohort in that it consisted of men with or without KS who had acquired HIV through sex with other men and were enrolled in NIAID protocols that did not include HAART. The 2 groups did not differ in age or CD4 count at time of death. All but 4 samples were collected from deceased patients. The activity of this protocol was determined to be exempt from the need for Institutional Review Board approval by the Office of Human Subject Research, National Institutes of Health, Bethesda, MD. Four patients still alive with confirmed KS provided informed consent for enrollment in a prospective study approved by the NIAID Institutional Review Board and were included. All but 4 patients were white North Americans; 2 controls and 2 with KS were African Americans.

Genomic DNA was extracted from banked cryopreserved lymphocyte pellets using the Puregene DNA Extraction Kit (Gentra Systems, Minneapolis, MN).

FcγRIIA polymorphisms were tested by allele-specific restriction digest according to a previously described method.45 

The 158V/F-FcγRIIIA polymorphism was detected by an allele-specific oligohybridization after a nested polymerize chain reaction (PCR) amplification of genomic DNA.28,29 A gene-specific 1.2-kilobase (kb) DNA fragment was amplified by PCR using the following primers, ATATTTACAGAATGGCACAGG and GACTTGGTACCCAGGTTGAA, and conditions, 5 minutes at 95°C, 35 cycles of 1 minute at 95°C, 1 minute at 56°C, and 1 minute at 72°C with a final extension of 8 minutes at 72°C.28,32 One microliter was used as a template for a nested PCR reaction with the primer pair, TCATCATAATTCTGACTTCT and CTTGAGTGATGGTGATGTTC, and conditions of 30 cycles of 1 minute each at 95°C, 62°C, and 72°C. [γ-32P]-ATP–labeled oligonucleotide probes corresponding to the F or V allele, GCAGGGGGCTTTTTGGGAGTAAA or GCAGGGGGCTTGTTGGGAGTAAA, were hybridized and washed in 6 × SSPE/1% SDS at room temperature, 42°C, and at 70.5°C for T allele and at 72.5°C for G allele. Autoradiography was then performed. Selected samples were confirmed by direct sequence analysis in duplicate; the 3′ oligonucleotide was used as the sequence primer with the Thermo Sequenase–radiolabeled terminator cycle sequencing kit (Amersham Life Sciences, Cleveland, OH) at 35 cycles and an annealing temperature of 55°C. Sequence analysis confirmed the presence of a C at nucleotide 531, present inFcγRIIIA but not inFcγRIIIB (which has a T). Two groups have characterized the polymorphism at base pair 559, a nonconservative T to G substitution, resulting in a change of phenylalanine (F) to valine (V) but designated the amino acid position differently, ie, 15829 or 176.28 Interestingly, a tri-allelic polymorphism of FcγRIIIA at nucleotide 230, 48L/H/R, which is strongly linked to 158V/F, does not appear to confer a significant biologic difference.28 

The FcγRIIIB-NA1/NA2 polymorphism was determined by an allele-specific PCR, using a modification of a protocol by Hessner et al.46 

A commercially available immunofluorescence assay kit for measuring antibodies against lytically expressed HHV-8 antigens (Advanced Biotechnologies, Columbia, MD) was used for detection of HHV-8 antibodies according to the manufacturer's instructions.47,48 Sera were tested at a 1:40 dilution and coded slides were scored by 3 independent investigators.

An initial analysis was performed on population I to explore a possible association between 1 or more loci and development of KS using χ² analysis (3 × 2 tables with 2 degrees of freedom). The data are presented without formal correction on the premise that candidate genes were chosen on the basis of previous in vitro data or association studies suggesting the functional significance of the polymorphisms.33 Because the second cohort, population II, was tested to confirm the findings for population I, a χ² analysis (3 × 2 tables with 2 degrees of freedom) of the second population is presented without correction. The effect of each FcγRs genotype on KS associated with HIV was analyzed by χ² analysis (2 × 2 with 1 degree of freedom). The association between KS and theFcγRIIIA locus in the combined population was tested using the Cochran-Mantel-Haenszel method of stratified analysis.49,50 

An exploratory analysis of the candidateFcγRs loci, FcγRIIA,FcγRIIIA, and FcγRIIIB, was performed on stored samples from population I (Table1). In this cohort, 58 patients had KS by time of death and 61 did not. Serologic evidence of HHV-8 infection by immunofluorescence assay on the last available serum sample conformed to published data: 39 of 46 KS patients (84.7%) were seropositive, compared with 21 of 55 patients (38.1%) without KS.51,52In an exploratory analysis in this cohort, population I (Table 1), there was a significant difference observed between individuals with and without KS with respect to 1 locus, FcγRIIIA(P =  .0035). Further analysis indicated (Table1) that the FcγRIIIA FF genotype was underrepresented in individuals with KS; only 21.8% of patients with KS had this genotype, compared with 48.3% of patients without KS (P = .003). In contrast, the heterozygous VF genotype was associated with KS; 65.5% of patients with KS had this genotype, compared with 35% of patients without KS (P = .0011). An association was not observed between KS and either the FcγRIIA orFcγRIIIB variant genotypes.

To confirm these results, a second population, population II, was subsequently genotyped (Table 1). The HHV-8 serologic status of the second population was determined on the last available banked serum. Of the patients tested, 47 of 55 KS patients (85.5%) were HHV-8 seropositive in contrast to 23 of 67 patients without KS (34.3%) who were seropositive, a distribution that did not differ statistically from the previous cohort or the published literature.51,52The analysis of population II (P = .018), as well as the analysis of the combined populations (ie, populations I and II; P = .00028) provided additional evidence (Table 1) that variant genotypes for FcγRIIIAwere strongly associated with KS. In population II, the FF genotype was associated with failure to develop KS (P = .0065); however, the association between the VF genotype and KS was only suggestive of a trend (P = .10).

Because the 2 study populations were comparable (ie, males with late stage HIV infection, CD4 counts below 200/μL, and identical rates for HHV-8 seropositivity), a stratified analysis combining both populations was performed using the Cochran-Mantel-Haenszel test.49,50This analysis indicated a strong association between theFcγRIIIA locus and KS (P = .00028) in the combined populations. Furthermore, analysis of the combined cohort of 240 genotyped patients (112 with KS and 128 without KS) demonstrated that the FF homozygous genotype was associated with a decreased likelihood of KS during HIV infection (P = .000061), whereas the VF heterozygous genotype was strongly associated with development of KS (P = .00063).

The distribution of genotypes did not differ significantly between the population without KS and a healthy control population of white North Americans from a recently published report at any of the 3 FcγR loci studied.32 Specifically, the distribution ofFcγRIIIA genotypes in the healthy control population was as follows: 91 (50%) FF, 71 (40%) VF, and 19 (10%) VV and did not differ from the combined group without KS (P = .77; χ² = 0.53).

We also examined whether FcγRIIIA might be involved in host response against HHV-8. An exploratory analysis of the combined populations with respect to HHV-8 serologic status andFcγRIIIA genotype, not taking into account KS status, does suggest an association (P = .0071) (Table 2). It appears that the FF genotype is protective against infection with HHV-8 that induces serologic response (P = .0017), whereas the VF genotype could be a risk factor for this form of HHV-8 infection (P = .023). Because HHV-8 infection appears to be an essential factor for the development of KS, an analysis restricted to subjects with detectable HHV-8 antibodies was performed for the combined population (n = 130; 86 with KS and 44 without KS). The significance of the 158 V/F polymorphism ofFcγRIIIA and predisposition to KS in individuals was still evident in those who were seropositive for HHV-8 (P = .036). The FF genotype was protective; 17 of 86 patients with KS in contrast to 18 of 44 without KS had this genotype (P = .01). On the other hand, the VF genotype was marginal in its association with development of KS in individuals who were seropositive for HHV-8 (55 of 86 with KS in comparison to 21 of 44 without KS; P = .076).

Our study of FcγR genotypes in 2 separate cohorts of HIV-infected men suggests that the homozygous FF genotype ofFcγRIIIA acts to partially protect HIV-infected individuals from developing KS. Conversely, the presence of at least 1 V allele is associated with KS. These findings suggest a possible role forFcγRIIIA in the pathogenesis of HIV-associated KS. Infection with HHV-8 could predispose individuals at risk to develop KS with the outcome determined in part by the genotype ofFcγRIIIA 158V/F. Certainly, other cofactors are critical, some of which could be influenced by low-affinity FcγR activity; these include HIV-associated disturbances in cytokine regulation, hormonal balance, angiogenic and transforming growth factors, and altered monocyte or NK activity.

Both in vitro and in vivo evidence are consistent with the hypothesis that cytokine balance plays a critical role in the pathogenesis of KS.11,12,15 Pro-inflammatory cytokines, IL-1β, IL-6, TNF-α, and oncostatin M are potent growth factors for KS spindle cells and are also found in excess within lesions, whereas the addition of the interleukin-1 receptor antagonist (IL-1RN) to KS spindle cells impairs growth in vitro.17,53-55 In vitro studies have shown that the VV genotype of FcγRIIIA results in an increase in NK cell activation, induction of apoptosis, and increased binding affinity for IgG1 and IgG3 compared with FF homozygotes.28,29 It will be of interest to study whether differential binding of IgG to FcγRIIIa results in alterations in downstream events, such as release of cytokines or chemokines which in turn could influence pathways critical for development of KS. We suggest that FF homozygotes might be protected from developing KS because of a less vigorous inflammatory response.

The small number of VV individuals precluded our ability to determine whether this genotype carries a moderately increased risk compared with VF heterozygotes. VF heterozygotes might be at a disadvantage because of the combination of the underlying defect in NK cells associated with HIV infection and the presence of a V allele that could more significantly modulate inflammatory pathways. The correlation of a deleterious outcome with heterozygosity for a variant form of a molecule of innate immunity has been reported previously56; tuberculosis in West Africa has been associated with heterozygous genotypes of several linked variants of the NRAMP-1 gene.

It is of interest that variant genotypes ofFcγRIIIA are associated with KS in HIV because FcγRIIIa is the predominant FcγR expressed on NK cells where it is closely associated with either the TCR-ξ or FcεRI chain.57 FcγRIIIa is a critical receptor on CD3-negative NK cells that act as effectors for ADCC and spontaneous lysis of sensitized viral-infected or malignant cells. In addition, on activation, NK cells secrete cytokines that influence pro- and anti-inflammatory pathways that are already deranged in HIV-associated KS, such as interferon-γ.58 Also, the absence of NK cells has been associated with increased severity of herpesvirus infections.59 Several case reports have suggested that the expression of variant FcγRIIIA on NK cells contributes to more severe illness during infection with herpesviruses but this limited study does not take into account the frequency of the variant alleles in the general population.60 

In the course of this study, it was determined that the FF genotype might be important in influencing not only the development of HIV-associated KS but also the outcome of infection with HHV-8. Further research will be required to determine whether genotypic differences in the binding of IgG, whether by NK cells or by other phagocytic cells, such as macrophages or monocytes, could alter response to HHV-8 infection, particularly in the presence of HIV infection. We infer from our analysis of FcγRIIIA genotypes and HHV-8 seropositivity that the reduced occurrence of seropositive HHV-8 infection associated with the FF genotype could be partially responsible for underrepresentation of the FF genotype in KS. These preliminary observations provide a possible link connecting KS andFcγRIIIA genotype, particularly because HHV-8 infection is an essential cofactor for KS. In this regard, the possible role of FcγRIIIa in modulating infection with HHV-8 is novel and bears further investigation, both for validation and elucidation of events critical for KS pathogenesis.

In this context, our results raise several interesting possibilities for understanding the contribution of FcγRIIIa to the pathogenesis of KS. Each of these possibilities reflects an antibody-dependent, HHV-8–specific role for FcγRIIIa on NK cells and professional phagocytes. Differences in genotype have been shown to alter IgG binding that could influence cytokine levels, the state of effector cell activation, or viral load. Our genetic epidemiologic study provides preliminary evidence for further investigation of these issues, based on a strong association between theFcγRIIIA locus and development of KS in HIV-infected men. It is also possible that the observed association is due to a different gene in linkage dysequilibrium withFcγRIIIA, which instead might directly play a role in host response to viral infection. Because the complexity of disturbances in immune regulation during HIV infection is sufficiently complicated, it is likely that a number of genes are involved in determining host susceptibility and response to HHV-8 infection as well as the development of KS.61 Nonetheless, our study has identified an immunologically interesting region of chromosome 1q22 as potentially important in the development of KS with theFcγRIIIA gene as the leading candidate responsible for this association.

We would like to thank Renée Chen and John O'Mara for their technical assistance.