The evolution and diversity of TNF block haplotypes in European, Asian and Australian Aboriginal populations

A TNF Block Genotype may Influence CMV Retinitis in HIV Patients without Affecting Systemic Viral Replication

BACKGROUND

A conserved TNF block haplotype marked by the minor alleles of rs1800629 (TNFA-308*A) and rs9281523 [BAT1(intron 10)*C] has been linked with several immunopathological conditions and with rapid progression of HIV disease. Reported associations with cytomegalovirus (CMV) retinitis in HIV patients before or during early antiretroviral therapy (ART) may therefore reflect greater replication of CMV in advanced HIV disease or an immunopathological response to CMV in the retina.

OBJECTIVE

As all Indonesian HIV patients display high levels of CMV replication, we evaluated whether TNF block genotypes alter markers of their burden of CMV and/or associate with retinitis.

METHODS

We assessed 79 consecutive HIV patients beginning ART, 25 HIV patients with a history of CMV-retinitis and 63 healthy adults. HIV RNA, CD4 T-cell counts, CMV-reactive antibody and CMV DNA were measured and alleles of TNFA-308, BAT1(intron 10) and TNFA-1031 (rs1799964) were determined.

RESULTS

TNFA-308 and BAT1(intron 10) were in complete linkage disequilibrium. Patients carrying minor alleles at both loci had higher levels of CMV-reactive antibody after one month on ART (p=0.01), but not at other time points spanning 1 year on ART. 50% of patients had detectable CMV DNA before ART, irrespective of TNF block genotypes. However, the TNFA-308*A/- BAT1(intron 10)*C haplotype was more common in CMV-retinitis patients than other patients or healthy controls (p<0.01).

CONCLUSION

The TNFA-308*A/BAT1(intron 10)*C haplotype appears to affect CMV-induced pathology rather than CMV replication.

TNF-Block Genotypes Influence Susceptibility to HIV-Associated Sensory Neuropathy in Indonesians and South Africans

HIV-associated sensory neuropathy (HIV-SN) is a disabling complication of HIV disease and antiretroviral therapies (ART). Since stavudine was removed from recommended treatment schedules, the prevalence of HIV-SN has declined and associated risk factors have changed. With stavudine, rs1799964*C (TNF-1031) associated with HIV-SN in Caucasians and Indonesians but not in South Africans. Here, we investigate associations between HIV-SN and rs1799964*C and 12 other polymorphisms spanning TNF and seven neighboring genes (the TNF-block) in Indonesians (n = 202; 34/168 cases) and South Africans (n = 75; 29/75 cases) treated without stavudine. Haplotypes were derived using fastPHASE and haplotype networks built with PopART. There were no associations with rs1799964*C in either population. However, rs9281523*C in intron 10 of BAT1 (alternatively DDX39B) independently associated with HIV-SN in Indonesians after correcting for lower CD4 T-cell counts and >500 copies of HIV RNA/mL (model p = 0.0011, Pseudo R² = 0.09). rs4947324*T (between NFKBIL1 and LTA) independently associated with reduced risk of HIV-SN and shared haplotype 1 (containing no minor alleles) associated with increased risk of HIV-SN after correcting for greater body weight, a history of tuberculosis and nadir CD4 T-cell counts (model: p = 0.0003, Pseudo R² = 0.22). These results confirm TNF-block genotypes influence susceptibility of HIV-SN. However, critical genotypes differ between ethnicities and with stavudine use.

Polymorphisms in inflammation-related genes and the risk of primary varicose veins in ethnic Russians

Inflammation was shown to be activated in varicose veins, although its role in the development of vein wall transformation remains inconclusive. We aimed to investigate the influence of 13 inflammation-related single nucleotide polymorphisms (SNPs) TNF rs1800629 and rs3093661, IL1A rs1800587, IL1RN rs4251961, IL6 rs1800795 and rs1800796, IFNG rs2430561, IL10 rs1800896, TGFB1 rs1800469, HIF1A rs11549465, NFKB1 rs28362491, and rs4648068 on the risk of primary varicose veins (PVVs) in ethnic Russians. We genotyped 709 patients with PVVs and 278 individuals without a history of chronic venous disease and performed a single SNP and a haplotype analysis. Several associations with P < 0.05 were revealed in our study. Variant allele HIF1A rs11549465 T, TNF rs3093661 A, and NFKB1 rs28362491 ATTG deletion showed the reverse association with PVV risk, and allele IL6 rs1800795 C was associated with the increased risk of the studied pathology. Haplotype analysis revealed associations of TNF haplotypes rs3093661 A-rs1800629 G and IL6 rs1800795 C-rs1800796 G with the decreased and the increased risk of PVVs, correspondingly. However, all the observed associations failed to reach statistical significance after the correction for multiple testing, which was set at a level of 10^-3 due to many tests performed. Our study therefore provides evidence that investigated polymorphisms do not play a major role in susceptibility to PVVs.

Genetics of HIV-associated sensory neuropathy and related pain in Africans

Despite the use of safer antiretroviral medications, the rate of HIV-associated sensory neuropathy (HIV-SN), the most common neurological complication of HIV, remains high. This condition is often painful and has a negative effect on quality of life. Up to 90% of those with HIV-SN experience pain for which there is no effective analgesic treatment. Genetic factors are implicated, but there is a lack of a comprehensive body of research for African populations. This knowledge gap is even more pertinent as Africans are most affected by HIV. However, recent studies performed in Southern African populations have identified genes displaying potential as genetic markers for HIV-SN and HIV-SN-associated pain in Africans. Here, we review the published studies to describe current knowledge of genetic risk factors for this disease in Africa.

TNF Block Gene Variants Associate With Pain Intensity in Black Southern Africans With HIV-associated Sensory Neuropathy

OBJECTIVES

HIV-associated sensory neuropathy (HIV-SN) is a common neurological complication of HIV infection, and it is often painful. Tumor necrosis factor (TNF)-α is implicated in neuropathic pain, but associations between neuropathic pain and polymorphisms in the TNFA gene have not been identified. The "TNF block" is a region of high linkage disequilibrium within the central major histocompatability complex that contains several genes involved in the regulation of inflammation, including TNFA. Polymorphisms in the block have been associated with an altered risk of HIV-SN, but no investigations into whether this region is associated with the painful symptoms of neuropathy have been undertaken. Therefore, we investigated whether polymorphisms in the TNF block are associated with pain intensity in black Southern Africans with HIV-SN.

METHODS

Single-nucleotide polymorphisms (SNPs) defining TNF block haplotypes and African-specific tagSNPs were genotyped in samples from 150 black Southern Africans with HIV-SN.

RESULTS

One SNP allele, rs28445017*A, was significantly associated with an increased pain intensity after correction for age, sex, and the CD4 T-cell count. A common 3-SNP haplotype containing rs28445017*G remained associated with a reduced pain intensity after correction for covariates and multiple comparisons.

DISCUSSION

We identified a novel genetic association between polymorphisms in the TNF block and the pain intensity in black Southern Africans with HIV-SN. Our study implicates rs28445017 in painful HIV-SN, although its precise role and whether it may be causative is unclear. rs28445017 was not associated with the risk for HIV-SN as such, highlighting potential differences between the pathophysiology of the neuropathy and the painful features of the neuropathy.

Role of TNF block genetic variants in HIV-associated sensory neuropathy in black Southern Africans

HIV-associated sensory neuropathy (HIV-SN) is a common neurological complication of HIV infection. The TNF block is a region within the central MHC that contains many immunoregulatory genes. Polymorphisms and haplotypes of the TNF block have been associated with increased risk of HIV-SN in Asians and whites. Here we investigated genetic associations with HIV-SN in 342 black Southern Africans (190 cases and 152 neuropathy-free controls) using single nucleotide polymorphisms (SNPs) spanning the TNF block and a set of haplotypes defined by 31 SNPs in Asian and white populations (denoted FVa). We included population-appropriate tagSNPs derived from an African population (Yoruban, YRI, HapMap) and derived extended haplotypes comprising 61 SNPs (denoted FVa_ext b). We found no association between HIV-SN and carriage of two SNPs (TNF-1031/rs1799964*C and BAT1 (intron10)/rs9281523*C) associated with HIV-SN in whites and Asians. Additionally, a haplotype containing TNF-1031/rs1799964*C associated with increased risk of HIV-SN in Asians, but was not present in this African population. However, alleles of seven SNPs associated with reduced risk of HIV-SN (corrected for age, height and multiple comparisons). These were rs11796*A, rs3130059*G, rs2071594*C, NFKBIL1-62/rs2071592*A, rs2071591*A, LTA+252/rs909253*G, rs1041981*C. One haplotype (FV18_ext1), not containing these alleles, was associated with increased risk of HIV-SN after correction for age, height and multiple comparisons. Our results confirm the involvement of genes in the TNF block in altering risk for HIV-SN, but genotypes critical in this African population differed from those affecting HIV-SN in whites and Asians. These differences support the need for genetic association studies in diverse populations.

Tumor necrosis factor haplotype diversity in Mestizo and native populations of Mexico

The so-called tumor necrosis factor (TNF) block includes the TNFA, lymphotoxin alpha and beta (LTA and LTB) genes with single-nucleotide polymorphisms (SNP) and microsatellites with an allele frequency that exhibits interpopulation variability. To date, no reports have included both SNPs and microsatellites at the TNF block to study Mestizo or Amerindian populations from Mexico. In this study, samples of five Mexican Mestizo populations (Durango, Guadalajara, Monterrey, Puebla, and Tierra Blanca) and four native-Mexican populations (North Lacandonians, South Lacandonians, Tepehuanos, and Yaquis) were genotyped for two SNPs (LTA+252A>G and TNFA-308G>A) and four microsatellites (TNFa, d, e, and f), to analyze the genetic substructure of the Mexican population. Allele and haplotype frequencies, linkage disequilibrium (LD), and interpopulation genetic relationships were calculated. There was significant LD along almost all of the TNF block but the lowest D' values were observed for the TNFf-TNFd pair. Mestizos showed higher allele and haplotype diversity than did natives. The genetic differentiation level was reduced among Mestizos; however, a slightly, but significant genetic substructure was observed between northern and southern Mexican Mestizos. Among the Amerindian populations, the genetic differentiation level was significantly elevated, particularly in both North and South Lacandonians. Furthermore, among Southern Lacandonians, inhabitants of Lacanja town were the most differentiated from all the Mexicans analyzed. The data presented here will serve as a reference for further population and epidemiological studies including these TNF polymorphisms in the Mexican population.

Tumor necrosis factor-alpha polymorphism at position -238 in preeclampsia

Background:

Preeclampsia is the most common serious disorder during pregnancy and studies show several immune-related processes in its pathophysiology. The role of cytokines and their expression remains controversial in this field. One of the cytokines of interest in recent studies has been TNF-α, which has been shown to have a higher level in maternal plasma of preeclamptic women.

Objectives:

This study was designed to evaluate the role of TNF-α polymorphism at position -238 in the risk of developing preeclampsia during pregnancy.

Patients and Methods:

One hundred fifty three preeclamptic cases and 140 healthy pregnant women were retrieved from two major hospitals of Mashhad, Iran. Methods a case-control study were designed. Anyone with a history of inflammatory disease, hypertension, or chronic kidney disease was excluded. DNA was extracted from peripheral blood leukocytes. Both groups were genotyped for the polymorphism of the TNF-α gene at position -238 by the RFLP method with Ava II enzyme. Allele and genotype frequencies were compared using one-way ANOVA and the Fisher’s exact test.

Results:

There were significant differences between the two groups in TNF-α genotype at position -238 (P < 0.001). In the preeclamptic group, the frequency of the AA genotype was higher (P < 0.001) and the frequency of the GG genotype was lower (P < 0.001). The overall prevalence of the A allele at position -238 was higher in preeclamptic cases (P < 0.001).

Conclusion:

In this study group, TNF-α -238 polymorphism was shown to be different in preeclamptic and non-preeclamptic pregnant women. The AA genotype and the A allele may carry an increased risk for developing of preeclampsia.

Diversity of extended HLA-DRB1 haplotypes in the Finnish population

The Major Histocompatibility Complex (MHC, 6p21) codes for traditional HLA and other host response related genes. The polymorphic HLA-DRB1 gene in MHC Class II has been associated with several complex diseases. In this study we focus on MHC haplotype structures in the Finnish population. We explore the variability of extended HLA-DRB1 haplotypes in relation to the other traditional HLA genes and a selected group of MHC class III genes. A total of 150 healthy Finnish individuals were included in the study. Subjects were genotyped for HLA alleles (HLA-A, -B, -DRB1, -DQB1, and -DPB1). The polymorphism of TNF, LTA, C4, BTNL2 and HLA-DRA genes was studied with 74 SNPs (single nucleotide polymorphism). The C4A and C4B gene copy numbers and a 2-bp silencing insertion at exon 29 in C4A gene were analysed with quantitative genomic realtime-PCR. The allele frequencies for each locus were calculated and haplotypes were constructed using both the traditional HLA alleles and SNP blocks. The most frequent Finnish A∼B∼DR -haplotype, uncommon in elsewhere in Europe, was A*03∼B*35∼DRB1*01∶01. The second most common haplotype was a common European ancestral haplotype AH 8.1 (A*01∼B*08∼DRB1*03∶01). Extended haplotypes containing HLA-B, TNF block, C4 and HLA-DPB1 strongly increased the number of HLA-DRB1 haplotypes showing variability in the extended HLA-DRB1 haplotype structures. On the contrary, BTNL2 block and HLA-DQB1 were more conserved showing linkage with the HLA-DRB1 alleles. We show that the use of HLA-DRB1 haplotypes rather than single HLA-DRB1 alleles is advantageous when studying the polymorphisms and LD patters of the MHC region. For disease association studies the HLA-DRB1 haplotypes with various MHC markers allows us to cluster haplotypes with functionally important gene variants such as C4 deficiency and cytokines TNF and LTA, and provides hypotheses for further assessment. Our study corroborates the importance of studying population-specific MHC haplotypes.

Genetic polymorphism of tumor necrosis factor promoter region and susceptibility to develop Hodgkin lymphoma in a Mexican population

Hodgkin lymphoma (HL) is a rare neoplasm of the lymphatic system, in which inflammation and allelic variants in cytokines have been proposed as etiological factors. Epstein-Barr virus infection is often associated as a risk factor in HL and since cytokines are involved in the humoral response to viral infection. Our aim was to study the association between single nucleotide polymorphisms (SNPs) located in the tumor necrosis factor (TNF) gene (- 376G> A, - 238G> A and 581G> A) in a sample of Mexican patients (56 cases) and their susceptibility to develop HL, comparing these SNPs among healthy individuals (127 controls). Frequencies for TNF - 238G> A and TNF 581G> A showed no significant differences between cases and controls. However, the proportion of cases with the GA genotype of - 376 SNP showed a significant difference as compared to controls, odds ratio = 4.41 (95% confidence interval: 1.21-16.6), p = 0.02. We found that in this group of patients from Mexico the SNP - 376G> A in TNF shows an association with higher risk for HL.

IL10 and TNF variants and risk of non-Hodgkin lymphoma among three Asian populations

Genetic variation in immune-related genes, such as IL10 and TNF, have been associated with the development of non-Hodgkin lymphoma (NHL) in Caucasian populations. To test the hypothesis that IL10 and TNF polymorphisms may be associated with NHL risk in Asian populations, we genotyped 20 single nucleotide polymorphisms (SNPs) within the IL10 and TNF/LTA loci in three independent case-control studies (2635 cases and 4234 controls). IL10 rs1800871, rs1800872, and rs1800896 were genotyped in all three studies, while 5 of the remaining SNPs were genotyped in two studies, and 12 in a single study. IL10 rs1800896 was associated with B cell lymphoma [per-allele odds ratio (OR) = 1.25, 95 % confidence interval (CI) 1.08-1.45; p trend = 0.003], specifically diffuse large B cell lymphoma (DLBCL) (per-allele OR = 1.29, 95 % CI 1.08-1.53; p trend = 0.004), as well as T cell lymphoma (per-allele OR = 1.44, 95 % CI 1.13-1.82; p trend = 0.003). TNF rs1800629, which was genotyped in only two of our studies, was also associated with B cell lymphoma (per-allele OR = 0.77, 95 % CI 0.64-0.91; p trend = 0.003), specifically DLBCL (per-allele OR = 0.69, 95 % CI 0.55-0.86; p trend = 0.001). Our findings suggest that genetic variation in IL10 and TNF may also play a role in lymphomagenesis in Asian populations.

TNF haplotypes in a Southern African population resemble those seen in Caucasians and Asians

Haplotypes spanning the tumor necrosis factor (TNF) gene block in the central major histocompatibility complex were defined in a Southern African population using 31 single-nucleotide polymorphisms. Twenty haplotypes accounted for 91.8% of the cohort. The haplotypes matched those described previously in Caucasian and Asian populations, supporting the hypothesis that TNF block haplotypes are ancient and highly conserved. They are presented here as a tool for disease-association studies.

TNFA promoter alleles--frequencies and linkage with classical HLA genes in a South German Caucasian population

The tumor necrosis factor alpha (TNFA) promoter region exhibits several polymorphisms, which have been hypothesized to influence gene expression, thereby associating positively or negatively with inflammatory conditions. Many studies have focused on single nucleotide polymorphisms (SNPs) taking not into account additive or inverse effects between different SNPs. We typed 1,021 healthy Caucasian volunteer stem cell donors for their TNFA promoter as well as their HLA-A,-C,-B,-DRB1 loci. Using statistical methods, we reconstructed TNFA promoter alleles and analyzed their frequency and linkage with HLA genes. We show that the number of TNFA promoter alleles frequent enough to be analyzed in clinical studies is limited and that a strong linkage with classical HLA genes is present, especially for the extended HLA-haplotype HLA-A*01:01/HLA-C*07:01/HLA-B*08:01/TNFA promoter allele 3/HLA*DRB1*03:01. Taking into account SNP frequency information, it is possible to quite accurately deduce TNFA promoter alleles by generic Sanger sequencing, obviating the need for elaborating allele-specific sequencing. This information may enable investigators to consider the complete TNFA regulatory region in a phase-separated manner in contrast to previous approaches examining only one or few isolated SNPs.

Genetic predisposition to respiratory infection and sepsis

Genetic variations, in part, determine individual susceptibility to sepsis and pneumonia. Advances in genetic sequence analysis as well as high throughput platform analysis of gene expression has allowed for a better understanding of immunopathogenesis during sepsis. Differences in genes can also modulate immune and inflammatory response during sepsis thereby translating to differences in clinical outcomes. An increasing number of candidate genes have been implicated to play a role in sepsis susceptibility, most of which are controversial with few exceptions. This does not refute the significance of genetic polymorphisms in sepsis, but rather highlights the difficulties and pitfalls related to genetic association studies. These difficulties include differences in study design such as heterogeneous patient cohorts and differences in pathogenic organisms, linkage disequilibrium, and lack of power for detailed haplotype analysis or examination of gene-gene interactions. There is extensive diversity in the pathways of inflammation and immune response during sepsis making it even harder to prove the functional and clinical significance of one single genetic polymorphism which could be easily masqueraded or compensated by other upstream or downstream events of the pathway involved. The majority of studies have analysed candidate genes in isolation from other possible polymorphisms. It is likely that susceptibility to sepsis is the result of polymorphisms from multiple genes rather than one single mutation. Future studies should aim for multi-centered collaborative approach looking at genome wide association or gene profiling to provide a more complete appraisal of the key genetic players in determining genetic susceptibility to sepsis. This review paper will summarise the prominent candidate gene polymorphisms with known functional changes or those with haplotype data. In addition, a summary of the expanding research in the field of epigenetics and post-sepsis immunosuppression will be discussed.

The major histocompatibility complex: a paradigm for studies of the human genome

The major histocompatibility complex (MHC) on chromosome 6 is one of the most intensively studied regions of the human genome and has many features which make it unique. It is the source of much research interest because of its role in autoimmune and infectious disease susceptibility, and of diagnostic interest because of its role in transplantation and rejection. It is the most gene-dense and SNP-rich region of the genome, with large number of complex haplotypes and other features which must be taken into account when analysing the MHC in the laboratory. This article provides a brief overview of the MHC highlighting some of the issues that must be considered when developing new methods and assays.

An assessment of the association between childhood asthma and HLA DRB1*03 using extended haplotype analysis

Ancestral haplotypes between human leukocyte antigen (HLA) class I and class II alleles are well-recognized in the literature. We previously published a positive association between the class II HLA allele DRB1*03 and the subsequent development of asthma in a retrospective cohort of 383 children. To refine this association, we investigated whether DRB1*03-specific haplotypes extending across the HLA are associated with asthma incidence. We found evidence of strong HLA DRB1*03-dependent linkage disequilibrium across the region, but no association between DRB1*03 ancestral haplotypes and childhood asthma. We did, however, observe a trend toward a positive association between HLA DRB1*03 and asthma by adding non-ancestral DRB1*03 positive haplotypes. Our results suggest that the role of the HLA DRB1*03 in asthma susceptibility is independent of ancestral-haplotype-mediated linkage disequilibrium.

Characterization of tumor necrosis factor-α block haplotypes associated with susceptibility to chronic venous leg ulcers in Caucasian patients

Polymorphisms in the central major histocompatibility complex (MHC) are associated with several immunopathologic and inflammatory diseases, including chronic venous leg ulcers (CVLU). Because of strong linkage disequilibrium, identification of loci affecting disease susceptibility must be based on comparisons between haplotypes. Here we examine the association of conserved tumor necrosis factor (TNF) block haplotypes with CVLU susceptibility. A total of 171 Caucasian patients with CVLU were compared with 173 age-/gender-matched controls, excluding individuals with type 1 diabetes or rheumatoid arthritis. A total of 194 healthy subjects formed a separate population-based control group. Samples were typed for 38 tumor necrosis factor (TNF) block single nucleotide polymorphisms (SNP), human leukocyte antigen (HLA)-B and HLA-DRB1 alleles. TNF haplotypes were derived using the PHASE algorithm and assigned numbers (FVx) defined previously. The patients and matched controls shared 16 TNF block haplotypes. The patients had increased carriage of FV16 and alleles of the 8.1 and 60.3 MHC ancestral haplotypes (AH). CVLU risk is modulated by alleles within FV16 (e.g., TNF-308A and BAT1intron10 C insertion) or near FV16 in the 8.1AH. CVLU risk may also be mediated by unidentified alleles (not in FV22) marked by HLA-B40 and HLA-DR13. FV16 appears to be the best MHC and TNF block marker of susceptibility. After disease onset, an individual's TNF block haplotype does not modulate CVLU severity.

Can immunogenetics illuminate the diverse manifestations of respiratory infections?

Improved technologies for high-throughput genotyping and the establishment of well-defined cohorts prompted hope that polymorphisms would be discovered that define a patients' risk of respiratory disease or aid in diagnosis. Genetic pitfalls encountered in this quest include genotyping errors, ethnic differences and linkage dysequilibrium. Differences in the definition of the disease phenotype also create discrepancies, so immunogenetic testing has not yet reached the clinic. However, associations between a polymorphism and a disease phenotype place the gene or one in linkage dysequilibrium on the path to the disease. Here we review studies of immune-related genes that are illuminating the immunopathogenesis of community-acquired pneumonia and mycobacterial infections.

Tumor necrosis factor-alpha gene polymorphisms are associated with severity of acute graft-versus-host disease following matched unrelated donor bone marrow transplantation in children: a Pediatric Blood and Marrow Transplant Consortium study

Tumor necrosis factor (TNF)-alpha plays a significant role in conditioning related toxicities and the development of acute graft-versus-host disease (aGVHD). TNF-alpha gene polymorphisms are associated with rejection after organ transplantation and aGVHD in matched related donor blood and marrow transplantation (BMT) recipients. Few studies have been published on unrelated donor BMT in the pediatric age group. In this study, we examined the relationship between specific polymorphisms in TNF pathway genes and the occurrence and severity of aGVHD. Recipient single-nucleotide polymorphisms (SNPs) in TNF-alpha and TNF receptor superfamily members 1A (TNFRSF1A) and 1B (TNFRSF1B) were investigated. In a multi-institutional Pediatric Blood and Marrow Transplant Consortium trial, a total of 180 pediatric patients (mean age, 11.0 years) were prospectively evaluated for clinical outcomes after matched unrelated donor BMT. All patients received myeloablative conditioning and two-drug GVHD prophylaxis with cyclosporine or tacrolimus, with methotrexate in the majority of patients. TNF-alpha genotypes were not correlated with the overall incidence of aGVHD. Significant associations were seen between TNF-alpha variant alleles and the severity of aGVHD (grade II-IV and grade III-IV), especially when analyzed in whites only (n = 165). Grade II-IV aGVHD was correlated with recipient -857T allele (hazard ratio [HR], 0.47; P = .04), -238A allele (HR, 1.76; P = .002), and d3/d3 genotype (HR, 0.64; P = .03). Severe (grade III-IV) aGVHD was associated with TNF-alpha -1031C allele (HR, 2.38; P = .03), -863A allele (HR, 3.18; P = .003), and d4/d4 genotype (HR, 2.82; P = .01). After adjusting for clinical factors, the association of -1031C, -863A, -238A, and d4/d4 genotypes with severity of aGVHD remained statistically significant. No correlation between selected SNPs in TNFRSF1A or TNFRSF1B and the incidence or severity of aGVHD was found. Our findings indicate clinically important relationships between genetic polymorphisms in TNF-alpha and the severity of aGVHD in this cohort. Improved understanding of this relationship may allow for a risk-adjusted approach to GVHD prevention in pediatric BMT.