Secondary Logo

Share this article on:

Tumor necrosis factor region polymorphisms are associated with AIDS and with cytomegalovirus retinitis

Deghaide, Neifi HSa; Rodrigues, Maria de LVb; Castelli, Erick Ca; Mendes-Junior, Celso Tc; Figueiredo, José FCd; Donadi, Eduardo Aa

doi: 10.1097/QAD.0b013e32832e5591
BASIC SCIENCE

Background: The tumor necrosis factor (TNF) gene is located within the highly polymorphic major histocompatibility complex region, exhibiting the -308 GA promoter region polymorphism and six microsatellites (TNFa–f) spanning the region nearby the TNF locus.

Objective: In the present study, we evaluated the frequency of -308 GA and TNFa-e polymorphisms and respective haplotypes (in chromosomal sequence: TNFd-TNFe-308GA-TNFc-TNFa-TNFb), in 222 patients with AIDS, 52 of whom exhibited cytomegalovirus retinitis, and in 202 healthy HIV-negative individuals.

Method: TNF microsatellite and single nucleotide polymorphism typings were performed by PCR followed by polyacrylamide gel electrophoresis.

Results: The TNF-308A allele and the 4-3-G-2-7-1 haplotype were associated with susceptibility to AIDS, whereas the TNFb4 allele and the 3-3-G-1-11-4 haplotype were associated with protection against AIDS development. The TNFc2 allele and the 4-1-G-2-2-1 haplotype, which contains the TNFc2 allele, were associated with cytomegalovirus retinitis.

Conclusion: This study highlights that polymorphic sites spanning the region nearby the TNF locus are associated with AIDS per se and with cytomegalovirus retinitis in AIDS patients.

aDivision of Clinical Immunology, Department of Internal Medicine, School of Medicine of Ribeirão Preto, University of São Paulo, Brazil

bDepartment of Ophthalmology, Otorhinolaryngology and Head and Neck Surgery, School of Medicine of Ribeirão Preto, University of São Paulo, Brazil

cDepartamento de Química, Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto, Universidade de São Paulo, 14040-901, Ribeirão Preto-SP, Brazil

dDivision of Infectious Diseases, Department of Internal Medicine, School of Medicine of Ribeirão Preto, University of São Paulo, Ribeirão Preto, Brazil.

Received 22 December, 2008

Revised 4 May, 2009

Accepted 19 May, 2009

Correspondence to Dr Neifi H.S. Deghaide, Laboratório de Imunologia Molecular Bloco G 2°Andar, Divisão de Imunologia Clínica, Departamento de Clínica Médica, Hospital de Clínicas da Faculdade de Medicina de Ribeirão Preto-USP, Avenida Bandeirantes 3900, Ribeirão Preto 14048-900, São Paulo, Brazil. Tel: +55 16 3602 2566; fax: +55 16 36336695; e-mail: n.deghaide@bol.com.br

Back to Top | Article Outline

Introduction

The AIDS epidemic has launched investigations for predicting disease susceptibility and complication markers. Despite regional differences, the highest prevalence of opportunistic infections in AIDS patients on all continents has been related to Pneumocystis carinii pneumonia, cytomegalovirus retinitis (CMV-R) and Toxoplasma gondii encephalitis [1]. Although the introduction of highly active antiretroviral therapy has dramatically improved life expectancy of patients with AIDS, CMV-R is still being observed particularly in patients presenting TCD4+ cell counts lower than 50 cell/μl [2,3]. This condition requires special ophthalmologic attention because of its severity and morbidity. In Brazil, the most common eye infection in patients with AIDS is CMV-R, which produces marked deterioration of the retina with decreased visual acuity due to its localization in the central part of the retina [4,5]. Despite the efforts to control the development of fatal AIDS, resistance to available drugs has emerged [6], justifying the need to search for markers that predispose AIDS patients to developing CMV-R [7–9] and other complications. The immune response against HIV may contribute to the pathogenesis of AIDS and its complications, with emphasis on the role of cytokines, particularly tumor necrosis factor (TNF). As TNF has the ability to induce HIV expression in chronically-infected cells [10], increased levels of soluble TNF may contribute to disease progression [11,12]. In addition, TNF has been detected in the vitreous of patients with AIDS exhibiting CMV-R, contributing to destruction of the retina [13].

The pro-inflammatory cytokines TNF and lymphotoxins are encoded by genes localized in the major histocompatibility complex class III region, situated between human leukocyte antigen (HLA) class I (HLA-A, HLA-B, HLA-C) and class II (HLA-DR, HLA-DQ and HLA-DP) genes [14]. The TNF region has several polymorphic sites, including short tandem repeats (STRs) also known as microsatellites and single nucleotide polymorphisms (SNPs). Six microsatellites spanning the region nearby TNF and lymphotoxin genes have been described, named TNFa, TNFb, TNFc, TNFd, TNFe and TNFf [15,16], presenting different numbers of dinucleotide repeats characterizing different alleles. Two relevant SNPs have been described in the TNF promoter region: -238GA and -308GA. As many of these TNF microsatellites and SNPs have been associated with the magnitude of the cytokine production [17], they have been studied in several autoimmune diseases, in tumors and in parasite disorders. Given the role of TNF in the progression of AIDS and in the development of disease complications, in this study, we evaluated the TNFa-e microsatellite alleles and the TNF promoter region (SNP-308GA) in AIDS patients presenting or not presenting CMV-R.

Back to Top | Article Outline

Materials and methods

Patients

The study was conducted on 222 patients with AIDS diagnosed according to Centers for Disease Control and Prevention (CDC) criteria [18], of whom 52 exhibited CMV-R (mean age = 35.92 ± 9.58 years, 75% men) and 170 did not (mean age = 37.52 ± 7.38 years, 62.9% men). A group of healthy individuals consisting of 202 individuals from the same geographical region, presenting negative serology for HIV infection, aged 18–9 years (mean = 33.28 ± 8.31 years, 72.2% men), was also studied. Although the Brazilian population may be highly admixed, the ethnic composition of the control group did not differ significantly from the total HIV group (P = 0.9512 ± 0.0024). Demographic, epidemiological and laboratory features of patients and controls are shown in Table 1. The study protocol was approved by the local Research Ethics Committee (protocol # 4018/2007), and all patients provided their written informed consent to participate.

Table 1

Table 1

Back to Top | Article Outline

Tumor necrosis factor microsatellites

Peripheral blood was used for DNA extraction using a salting-out procedure. The primer sequences used to amplify TNF microsatellites were identical to those reported elsewhere [16]. All TNF microsatellite amplifications were performed using similar reagents and cycling conditions. The final volume of the reaction mixture for each sample was 25 μl, including 20 ng of genomic DNA, 0.3 mmol/l deoxynucleoside triphosphate (Pharmacia, Uppsala, Sweden), 0.3 mmol/l of each primer of a TNF microsatellite locus, 1 U DNA polymerase (Biotools, Palo Alto, California, USA), 1× amplification buffer (0.2 mol/l Tris-HCl, pH 8.5, 0.5 mol/l KCl) and 2.5 mmol/l of MgCl2.

The TNFa, TNFb, TNFd and TNFe loci were amplified in two steps. First, single fragments of amplified TNFa/b or TNFd/e loci were obtained using the primers described by Udalova et al. [16]. Second, the TNFa, TNFb, TNFd and TNFe fragments were produced using the first reaction product as template [16]. The first amplification for the TNFa/b and TNFd/e loci and the single amplification for the TNFc locus were carried out with an initial denaturation step at 94°C for 5 min, 28 cycles at 94°C for 1 min, 61°C for 1 min and 72°C for 1 min, with a final extension at 72°C for 10 min. The second amplification for the TNFa, b, d or e loci were carried out with an initial denaturation step at 94°C for 5 min, 32 cycles at 94°C for 1 min, 60°C for 1 min and 72°C for 1 min, with a final extension at 72°C for 10 min. Amplicons were separated by 12% denaturing polyacrylamide gels stained with silver. Homozygous or previously sequenced samples were used as controls. Figure 1 illustrates some TNF microsatellite allele typing.

Fig. 1

Fig. 1

Back to Top | Article Outline

Tumor necrosis factor single nucleotide polymorphism promoter region

The TNF -308G/A promoter region was amplified using two specific previously described primers: -308A (5′-ATAGGTTTTGAGGGGCATGA-3′), -308G (5′-ATAGGTTTTGAGGGGCATGG-3′) [19] and a generic primer (5′-CAGGGGAAAACTTCCTTGGT-3′), which produced 259 bp fragments [20]. The human hemoglobin locus (amplified with primers 5′-CGGTATTTGGAGGTCAGCAC-3′ and 5′-CCCACCACCAAGACCTACTT-3′), which produced a 548 bp fragment, was used as an internal control. The reaction was performed in duplicate in a final volume of 10 μl, including 20 ng of genomic DNA, 0.25 mmol/l deoxynucleoside triphosphate (Pharmacia) and 3 pmol of a specific primer and of a generic one, 2 pmol of the internal control primers, 0.75 U DNA polymerase (Invitrogen, Grand Island, New York, USA) and 1× buffer (20 mmol/l Tris-HCl, pH 8.4, 50 mmol/l KCl) and 1.5 mmol/l of MgCl2. After an initial denaturation step at 94°C for 5 min, samples were submitted to 32 cycles at 94°C for 45 s, 64°C for 45 s and 72°C for 1 min, with a final extension at 72°C for 5 min. Amplification products were submitted to electrophoresis using 10% polyacrylamide gels stained with silver (Fig. 1).

Back to Top | Article Outline

Statistical analysis

Allelic frequencies and observed heterozygosity (H O) were computed by the direct counting method. Adherence of genotypic proportions to expectations under Hardy–Weinberg equilibrium was tested by an exact test [21] employing the GENEPOP 3.4 software [22]. The GENEPOP 3.4 program was also used to perform the pair wise exact test of population differentiation [23] based on alleles and genotype frequencies. The expected heterozygosity values (H Sk) and their standard deviations were estimated by the ARLEQUIN program version 3.1 [24].

The presence of a significant association between microsatellites and SNP alleles was evaluated by a likelihood ratio test of linkage disequilibrium [25] using the ARLEQUIN 3.1 program [24]. Given the positive association but unknown gametic phase, the PHASE method, implemented by the PHASE v2 computer program [26], and the Expectation-Maximization algorithm implemented by the HelixTree package (HelixTree 6.0.1; Golden Helix, Inc., Bozeman, Montana, USA) were used to infer TNF haplotypes. In a conservative way, the estimated haplotypes for each sample were compared, and only samples with the same haplotypes inferred by both methods were used for further analysis. The HelixTree Package and the PHASE program inferred haplotypes with a mean probability of 0.9774 and 0.9616, respectively, and 377 out of 424 samples (88.9%) had the same haplotype determined by both methods.

The allele and genotype frequencies were compared using the two-sided Fisher's exact test implemented in the GraphPad InStat 3 software, which was also used to estimate the odds ratio (OR) and its 95% confidence interval (CI).

Back to Top | Article Outline

Results

The genotypes of all the loci evaluated in this study did fit the Hardy–Weinberg equilibrium expectations. An exact test of population differentiation based on allele and genotype frequencies, considering only the multiallele loci, revealed that the TNFb frequency distributions were significantly different between the total HIV group and controls (P = 0.0237 ± 0.0018 and 0.0236 ± 0.0017).

The frequencies of each TNF allele found in this study are presented in Table 2. In order to detect the potential influence of specific alleles, 2 × 2 contingency tables were created for each allele to detect possible differences in frequencies between groups using the Fisher exact test (Table 3). When AIDS patients presenting or not presenting CMV-R were compared, the TNFc1 allele was found to be associated with protection against CMV-R. Given that the TNFc locus is a biallelic marker, the TNFc2 allele was found to be associated with susceptibility to CMV-R (P = 0.0447, OR = 0.6157, 95% CI = 0.3924–0.9659). When the total HIV group was compared with the control group, the TNF-308A allele was found to be associated with susceptibility to HIV acquisition. Given that the SNP-308 G/A is also a biallelic marker, the -308G allele was found to be associated with protection against HIV acquisition (P = 0.0187, OR = 1.747, 95% CI = 1.109–2.752). Likewise, the TNFb4 allele was associated with protection against HIV acquisition (P = 0.0060, OR = 0.6703, 95% CI = 0.5042–0.8911).

Table 2

Table 2

Table 3

Table 3

Taking all samples together, a test of linkage disequilibrium revealed a high linkage disequilibrium among the loci evaluated. For almost all pairs of loci, the probability of linkage disequilibrium was 0.0000 ± 0.0000, except for the TNFc/-308 pair (P = 0.0165 ± 0.0010). The only pair that did not present significant linkage disequilibrium was TNFe/-308 (P = 0.2495 ± 0.0043). Given the positive association but unknown gametic phase, the PHASE and Expectation-Maximization algorithms were used to infer the TNF haplotypes. Considering the TNF haplotypes (as given in the chromosomal sequence, i.e. TNFd-TNFe-308G/A-TNFc-TNFa-TNFb), the 3-3-G-1-11-4 haplotype was associated with protection against AIDS (P = 0.0156, OR = 0.4999, 95% CI = 0.2887–0.8654), whereas the 4-3-G-2-7-1 haplotype was associated with susceptibility to AIDS (P = 0.0212, OR = 5.137, 95% CI = 1.117–23.612) (Table 3). When the CMV group was compared with the AIDS patients without CMV-R, the 4-1-G-2-2-1 haplotype was found to be associated with susceptibility to CMV-R (P = 0.0326, OR = 2.181, 95% CI = 1.107–4.295) (Table 3). A test of population differentiation based on the haplotype frequencies did not reveal any significant differences between groups.

Back to Top | Article Outline

Discussion

Several studies have reported the association of TNF microsatellite and promoter region polymorphisms with TNF production, although some of these results have been controversial. TNFa2 and TNFc2 alleles have been associated with high TNF production, whereas TNFa6 and TNFc1 alleles have been associated with low serum levels of TNF [27]. On the other hand, another study has associated the TNFa2, TNFa6 and TNFa10 alleles as low-producer alleles, whereas the TNFa4 and TNFa11 alleles were associated with high-serum TNF levels [28]. The TNFd3 allele has also been associated with in-vitro production of TNF by mononuclear peripheral blood cells [20]. No association of TNFb and TNFe alleles with serum TNF levels has been reported. The presence of an adenine at the −308 position in the promoter region has been associated with increased transcription of TNF as compared with the presence of a guanine in the same position [17].

The TNF locus is located in the HLA complex, which has been associated with susceptibility to or progression from HIV infection to AIDS. Patients possessing the HLA-B*35-Cw*04 haplotype progress more rapidly to AIDS after HIV infection, whereas HLA-B27 has a protective effect [29]. Rapidly progressive HLA haplotypes have also been reported by us to be associated with the development of CMV-R [30]. In addition, the TNFc2 microsatellite allele, which has been associated with high production of TNF [27], has also been associated with the progression from HIV infection to AIDS [31].

Considering the strong-linkage disequilibrium in the HLA complex, the evaluation of TNF polymorphic sites is of interest as this cytokine is highly involved in AIDS pathogenesis, and there are few studies evaluating these polymorphisms in patients with AIDS. In this study, the frequencies of TNF microsatellites and SNPs in the TNF promoter region were evaluated in order to search for markers associated with AIDS and with one of its major complication (i.e. CMV-R).

The TNFb4 allele conferred protection against AIDS, as its frequency was over-represented in healthy controls when compared with AIDS patients. Likewise, the TNFb4-associated TNFd3-TNFe3-308G-TNFc1-TNFa11-TNFb4 haplotype and the -308G allele conferred protection against AIDS (Table 3). On the other hand, the -308A allele and the TNFd4-TNFe3-308G-TNFc2-TNFa7-TNFb1 haplotype were associated with susceptibility to AIDS (Table 3). In both haplotypes described above, with opposite effects, the TNF low producer -308G allele was present; however, in the protection haplotype, there is an additional dose of the TNFc1 allele, which is associated with low TNF production. Therefore, it seems reasonable to hypothesize that different combinations of alleles may contribute to TNF production. However, soluble TNF levels were measured in 25 CMV-R patients, in 28 patients without CMV-R and in 31 controls (data not shown). No statistically significant differences were observed. As expected, normal individuals did not exhibit detectable levels of TNF. Among patients with AIDS, only four presented detectable plasma levels of soluble TNF, one with CMV-R and three without CMV-R.

Considering its influence regarding CMV-R acquisition or development, the biallelic TNFc locus was associated with susceptibility (TNFc2) or protection (TNFc1) against CMV-R. Additionally, the TNFd4-TNFe1-308G-TNFc2-TNFa2-TNFb1 haplotype was associated with susceptibility to CMV-R (Table 3). On this basis, we may conclude that the TNFc2 allele confers susceptibility to CMV-R rather than the TNFc1 allele confers protection. On the other hand, in a previous report, the TNFc2 allele was associated with a slow progression from HIV infection to AIDS in a cohort of 24 English patients [31]. According to our previous results, patients possessing HLA-A-B-C alleles associated with rapid progression to AIDS were also associated with CMV-R [30]. No significant differences were observed when the concomitance of B35/Cw4 in patients with AIDS presenting or not presenting TNF-c2 was evaluated (data not shown), indicating an independent role of TNF microsatellite alleles. With respect to the TNFd or TNFe locus, no significant association was observed in this study regarding susceptibility to or protection against the development of CMV-R or AIDS, and there is no information in the literature about the association of these alleles with the magnitude of TNF secretion.

Although the mechanisms related to the participation of TNF in the development of CMV-R have not been elucidated, this study indicates that the TNF region per se, or as a result of its high linkage disequilibrium with the HLA complex, is involved in susceptibility to AIDS and CMV-R development.

Back to Top | Article Outline

Acknowledgements

The authors acknowledge Dr Renata Toscano Simões and Dr Fernando Queiróz Cunha for technical assistance.

Back to Top | Article Outline

References

1. Belfort R Jr. The ophthalmologist and the global impact of the AIDS epidemic LV Edward Jackson Memorial Lecture. Am J Ophthalmol 2000; 129:1–8.
2. Goldberg DE, Wang H, Azen SP, Freeman WR. Long term visual outcome of patients with cytomegalovirus retinitis treated with highly active antiretroviral therapy. Br J Ophthalmol 2003; 87:853–855.
3. Jacobson MA, Stanley H, Holtzer C, Margolis TP, Cunningham ET. Natural history and outcome of new AIDS-related cytomegalovirus retinitis diagnosed in the era of highly active antiretroviral therapy. Clin Infect Dis 2000; 30:231–233.
4. Jabs DA, Green WR, Fox R, Polk BF, Bartlett JG. Ocular manifestations of acquired immune deficiency syndrome. Ophthalmology 1989; 96:1092–1099.
5. Jabs DA, Martin BK, Forman MS, Ricks MO. Cytomegalovirus (CMV) blood DNA load, CMV retinitis progression, and occurrence of resistant CMV in patients with CMV retinitis. J Infect Dis 2005; 192:640–649.
6. Kantor R, Shafer RW, Follansbee S, Taylor J, Shilane D, Hurley L, et al. Evolution of resistance to drugs in HIV-1-infected patients failing antiretroviral therapy. AIDS 2004; 18:1503–1511.
7. Price P, Keane NM, Stone SF, Cheong KY, French MA. MHC haplotypes affect the expression of opportunistic infections in HIV patients. Hum Immunol 2001; 62:157–164.
8. Veronese Rodrigues Mde L, de Castro Figueiredo JF, Deghaide NH, Romao E, Vieira de Souza N, Donadi EA. Frequency of HLA class 1 and 2 alleles in Brazilian patients with AIDS and cytomegalovirus retinitis. Acta Ophthalmol Scand 2003; 81:514–516.
9. Veronese Rodrigues Mde L, Demarco ALG, de Castro Figueiredo JF, Deghaide NHS, Romao E, Donadi EA. HLA-B35 antigen is associated to susceptibility to ocular infections in patients with AIDS. Ophthalmic Res 2004; 36:86.
10. Poli G, Kinter A, Justement JS, Kehrl JH, Bressler P, Stanley S, et al. Tumor necrosis factor α functions in an autocrine manner in the induction of human immunodeficiency virus expression. Proc Natl Acad Sci USA 1990; 87:782–785.
11. Godfried MH, van der Poll T, Weverling GJ, Mulder JW, Jansen J, van Deventer SJ, Sauerwein HP. Soluble receptors for tumor necrosis factor as predictors of progression to AIDS in asymptomatic human immunodeficiency virus type 1 infection. J Infect Dis 1994; 169:739–745.
12. Morlat P, Pereira E, Clayette P, Derreudre-Bosquet N, Ecobichon JL, Benveniste O, et al. Early evolution of plasma soluble TNF-alpha p75 receptor as a marker of progression in treated HIV-infected patients. AIDS Res Hum Retroviruses 2008; 24:1383–1389.
13. Mondino BJ, Sidikaro Y, Mayer FJ, Sumner HL. Inflammatory mediators in the vitreous humor of AIDS patients with retinitis. Invest Ophthalmol Vis Sci 1990; 31:798–804.
14. Hajeer AH, Hutchinson IV. TNF-alpha gene polymorphism: clinical and biological implications. Microsc Res Tech 2000; 50:216–228.
15. Hajeer AH, Hutchinson IV. Influence of TNFalpha gene polymorphisms on TNFalpha production and disease. Hum Immunol 2001; 62:1191–1199.
16. Udalova IA, Nedospasov SA, Webb GC, Chaplin DD, Turetskaya RL. Highly informative typing of the human TNF locus using six adjacent polymorphic markers. Genomics 1993; 16:180–186.
17. Kroeger KM, Carville KS, Abraham LJ. The -308 tumor necrosis factor-alpha promoter polymorphism effects transcription. Mol Immunol 1997; 34:391–399.
18. Prevention CfDCa. From the Centers for Disease Control and Prevention. 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. JAMA 1993; 269:729–730.
19. Marsh HP, Haldar NA, Bunce M, Marshall SE, le Monier K, Winsey SL, et al. Polymorphisms in tumour necrosis factor (TNF) are associated with risk of bladder cancer and grade of tumour at presentation. Br J Cancer 2003; 89:1096–1101.
20. Tuglular S, Berthoux P, Berthoux F. Polymorphisms of the tumour necrosis factor alpha gene at position -308 and TNFd microsatellite in primary IgA nephropathy. Nephrol Dial Transplant 2003; 18:724–731.
21. Guo SW, Thompson EA. Performing the exact test of Hardy–Weinberg proportion for multiple alleles. Biometrics 1992; 48:361–372.
22. Raymond M, Rousset F. GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. J Hered 1995; 86:248–249.
23. Raymond M, Rousset F. An exact test for population differentiation. Evolution 1995; 49:1280–1283.
24. Excoffier L, Laval G, Schneider S. Arlequin ver. 3.0: an integrated software package for population genetics data analysis. Evol Bioinform Online 2005; 1:47–50.
25. Excoffier L, Slatkin M. Incorporating genotypes of relatives into a test of linkage disequilibrium. Am J Hum Genet 1998; 62:171–180.
26. Stephens M, Smith NJ, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 2001; 68:978–989.
27. Pociot F, Briant L, Jongeneel CV, Molvig J, Worsae H, Abbal M, et al. Association of tumor necrosis factor (TNF) and class II major histocompatibility complex alleles with the secretion of TNF-alpha and TNF-beta by human mononuclear cells: a possible link to insulin-dependent diabetes mellitus. Eur J Immunol 1993; 23:224–231.
28. Derkx HHF, Bruin KF, Jongeneel CV, Waal LP, Brinkmam BMN, Verweij CL, et al. Familial differences in endotoxin-induced TNF release in whole blood and peripheral blood mononuclear cells in vitro: relationship to TNF gene polymorphism. J Endotoxin Res 1995; 2:19–25.
29. Dean M, Carrington M, O'Brien SJ. Balanced polymorphism selected by genetic versus infectious human disease. Annu Rev Genomics Hum Genet 2002; 3:263–292.
30. Fernandes AP, Goncalves MA, Zavanella RB, Figueiredo JF, Donadi EA, Rodrigues ML. HLA markers associated with progression to AIDS are also associated with susceptibility to cytomegalovirus retinitis. AIDS 2003; 17:2133–2136.
31. Khoo SH, Pepper L, Snowden N, Hajeer AH, Vallely P, Wilkins EG, et al. Tumour necrosis c2 microsatellite allele is associated with the rate of HIV disease progression factor. AIDS 1997; 11:423–428.
Keywords:

AIDS; cytomegalovirus; microsatellites; retinitis; single nucleotide polymorphisms; tumor necrosis factor

© 2009 Lippincott Williams & Wilkins, Inc.