Impact of Toll-like receptor-4 and tumour necrosis factor gene polymorphisms in patients with hidradenitis suppurativa


  • Funding sources
    This study was funded in part by a donation from Abbott Hellas SA.

  • Conflicts of interest
    None declared.

Athina Savva.


Background  Recent evidence has suggested that deranged immune responses play a role in the pathogenesis of hidradenitis suppurativa (HS).

Objectives  To investigate the role of single nucleotide polymorphisms (SNPs) of the tumour necrosis factor (TNF) and Toll-like receptor 4 (TLR4) genes in the physical course of HS; these genes encode for proteins implicated in the immune response of the host.

Methods  DNA was isolated from 190 patients with HS and 84 healthy controls. SNPs at the promoter regions −376G/A, −238G/A and −308G/A of the TNF gene and the Asp299Gly and Thr399Ile SNPs of the TLR4 gene were determined by polymerase chain reaction (PCR) and digestion of the PCR product by restriction enzymes; after electrophoresis on 2·0% agarose gel, products were visualized on under ultraviolet radiation.

Results  The presence of the −238 TNF gene polymorphism was associated with a predisposition to HS (P = 0·027). Susceptibility to the disease was strongly correlated with the presence of AGG/GGA/AGA/GAA TNF haplotypes in 32 (17%) patients compared with two (2%) controls (< 0·001, odds ratio 8·30, 95% confidence interval 1·94–35·52). The frequency of HS exacerbations and disease severity were greater in patients carrying any of the GAG/AGG/GGA/AGA/GAA haplotypes of the TNF gene. Thirty-two patients were given TNF antagonists. Nineteen of these patients were carriers of the GGG haplotype of the TNF gene, whereas 13 were carriers of other haplotypes; favourable responses as evidenced by the Sartorius score were registered in 15 (79%) and five (38%, = 0·025), respectively. Carriage of the TLR4 gene alleles was not associated with any disease parameter.

Conclusions  A significant role of SNPs at the promoter region of the TNF gene is indicated for susceptibility to HS and for response to TNF antagonists.

Hidradenitis suppurativa (HS) is a chronic, devastating, inflammatory skin disorder affecting areas rich in apocrine glands. Painful lesions such as deep-seated nodules, boils, abscesses and sinus tracts appear most commonly on the axillae, and the inguinal and/or the anogenital areas. The lesions progressively become swollen and rupture with the release of pus.1 This process recurs, leading to permanent sinus tract and scar formation. The onset of HS occurs after puberty and indiscriminately affects the global population at a reported point-prevalence rate between 1% and 4%.2,3 HS has a considerable impact on the quality of life and mental health of patients, deeply affecting functioning experiences in all domains, with the presence of fatigue, depression, social stigmatization and isolation.4 The Dermatology Life Quality Index score for HS is 8·9 and is greater than for any other skin disorder.5

The exact pathophysiology of HS remains unknown; there are many hypotheses.6 Accumulating evidence over the past years suggests that deranged immune responses play a crucial role in the pathogenesis of the disease. Several studies supporting this thesis have shown that: (i) circulating monocytes are downregulated for cytokine production;7 (ii) HS seldom coexists with Crohn disease which is an autoimmune disorder;8 (iii) concentrations of proinflammatory cytokines in biopsies from HS lesions and perilesional skin are elevated;9 and (iv) many patients benefit from treatment with tumour necrosis factor (TNF) antagonists. The effectiveness of TNF antagonists has been demonstrated in a variety of randomized and nonrandomized clinical trials but the number of patients enrolled in these studies is limited.10–13

The innate immune response to offending bacteria is initiated when well-preserved bacterial structures known as pathogen-associated molecular patterns sensitize pathogen recognition receptors (PRRs) of circulating monocytes and tissue macrophages. The best studied PRRs are Toll-like receptors (TLRs), which are transmembrane receptors. TLR4 has been well studied; it binds bacterial lipopolysaccharide (LPS) and endogenous danger-associated molecular patterns such as heat-shock protein 70 and high-mobility group box-1. These interactions lead to a series of intracellular activations ending in the gene expression and production of proinflammatory cytokines including TNF-α, interleukin (IL)-1β and IL-6, and of anti-inflammatory cytokines such as IL-10.14 Data of our group denote that circulating monocytes of patients with HS demonstrate downregulated responses for the production of TNF-α after stimulation with LPS.7

However, the integrity of the pathway of cytokine production after stimulation of TLR4 may be altered when patients bear single nucleotide polymorphisms (SNPs) encoding protein molecules participating in the cascade of stimulation of monocytes.15 The most broadly studied SNPs are those of the TLR4 and TNF genes. More precisely, regarding the TNF gene, three SNPs at the promoter region have been broadly described: guanine (G) to adenine (A) transitions at the −376 position (−376 TNF G/A, rs1800750); at the −308 position (−308 TNF G/A, rs1800629); and at the −238 position (−238 TNF G/A, rs361525).16 For the TLR4 gene, two SNPs have been well characterized: an adenine to guanine transition causing substitution of aspartic acid by glycine at amino acid position 299 (Asp299Gly, rs4986790) and a cytidine to thymidine transition causing substitution of threonine by isoleucine at amino acid position 399 (Thr399Ile, rs4986791).17 Existing published studies have investigated the importance of these two SNPs for the predisposition of the host to infectious diseases.18–21

Indirect evidence suggests a need to investigate the impact of these SNPs in HS. This evidence concerns the known genetic predisposition for HS,6 the hyporesponsiveness of circulating monocytes after TLR4 stimulation7 and the favourable responses after treatment with TNF antagonists.10–13 The present study aimed to investigate the impact of SNPs of the TLR4 and TNF genes in the natural course of HS.

Patients and methods

Patient population

Patients enrolled in the study were followed up at the outpatient Department of Immunology of Infectious Diseases of ATTIKON University Hospital during the period January 2005 to January 2012. The study was approved by the ethics committee of the hospital and written informed consent was provided by all the patients and healthy controls.

Inclusion criteria for the study were caucasian origin and a diagnosis of HS. Diagnosis of HS was based on the following criteria: (i) onset early after puberty; (ii) presence of subcutaneous nodules in areas of skin rich in apocrine glands; and (iii) a compatible history of recurrent drainage of pus from the affected areas.5,22 Exclusion criteria were: infections of human immunodeficiency virus (HIV)-1, hepatitis B virus (HBV) or hepatitis C virus (HCV); coexisting rheumatoid arthritis, seronegative arthritis, Crohn disease or psoriasis; and diabetes mellitus types 1 and 2. Clinical characteristics of patients were recorded comprising demographics, frequency of exacerbations, involved areas and Hurley stage of severity. Lesions were graded according to the clinical system of Hurley, which comprises three stages based on the presence and extent of cicatrization and sinuses: stage I, mainly characterized by single or multiple abscess formation, to stage III where multiple interconnected tracts and abscesses are present in the involved areas.22 A healthy control group included subjects of caucasian origin without HS and without any of the above exclusion criteria.

Thirty-two patients with Hurley stage III disease were treated with TNF antagonists; they received either infliximab or etanercept. Administration of anti-TNF agents for 10 of these patients was performed as part of a prospective nonrandomized, unblinded, clinical trial10 and for another 22 patients as compassionate use; the latter is a procedure taking place after providing written informed consent from the patients and followed by Institutional Review Board approval and approval by the National Organization of Medicines, as dictated by Greek legislation. Patients were administered treatment after a thorough work-up, provided that they had a negative skin tuberculin test; a normal chest X-ray; normal liver biochemistry; negative serology for HIV, for HBV and for HCV; and negative family history of any demyelinating disorder. Disease severity was assessed before the start of therapy and during the course of therapy using the score proposed by Sartorius et al.23 as follows: (i) anatomical region involved (left and/or right axilla, groin, gluteal, inframammary or other region – 3 points per region involved); (ii) number and scores of lesions (points per lesions of each region involved – fistulas 4, nodules 2, scars 1, folliculitis 1); (iii) the longest distance between two relevant lesions in each region, or size if there was only one lesion (< 5 cm, two points; 5–10 cm, 4; > 10 cm, 8); (iv) clear separation of lesions from adjacent normal skin (yes, 0; no, 6). Response to treatment was assessed by a > 30% decrease of the Sartorius score compared with baseline after at least 6 months of treatment.

Blood collection and genotyping

A sample of whole blood (3 mL) for genotyping analysis was collected from patients with HS and healthy controls in ethylenediamine tetraacetic acid (EDTA) and stored at −70 °C until processed. Genomic DNA was extracted from all individuals using the Purigen Blood Core Kit C (Qiagen, Valencia, CA, U.S.A.) according to the manufacturer’s instructions. The genotypic analysis was performed in a blinded fashion for clinical data, with analysts being unaware of the patients’ names.

Typing for the three SNPs of the TNF gene and the two SNPs of the TLR4 gene was performed using polymerase chain reaction (PCR) on a Sensoquest thermal LabCycler Gradient (SensoQuest, Göttingen, Germany) using 50 ng of genomic DNA at a final volume of 27 μL, with 50 mmol L−1 of MgCl2, 20 mmol L−1 of dNTPs and 1 mmol L−1 of Taq polymerase (all New England Biolabs, Ipswich, MA, U.S.A.). Targeted forward and reverse primers of each studied SNP, conditions of PCR and conditions of digestion of PCR products are shown in Table 1. All digested PCR products were electrophoresed on 2% agarose gel and visualized using ultraviolet radiation after ethidium bromide staining, with the application of positive controls in each run. Haplotype calculations were done using UNPHASED software (

Table 1. Conditions of polymerase chain reaction (PCR) for analysis of single nucleotide polymorphisms (SNPs)
SNPPrimers (forward/reverse)PCR conditionsPCR digestion
Denaturation: 5 min at 95 °C, 40 amplification cycles (annealing step 60 s at 95 °C, polymerization step 60 s at 52 °C, elongation step 60 s at 72 °C) NcoI 37 °C for 1 h
Denaturation: 4 min at 95 °C, 35 amplification cycles (annealing step of 30 s at 95 °C, polymerization step 30 s at 56 °C, elongation step 30 s at 72 °C) HinfI 37 °C for 1 h
rs18007505′-CCT CAG GAC TCA ACA CAG C-3′
Denaturation: 10 min at 95 °C, 35 amplification cycles (annealing step 60 s at 95 °C, polymerization step 60 s at 58 °C, elongation step 60 s at 72 °C) HpaI 37 °C for 4 h
Denaturation: 10 min at 95 °C, 35 amplification cycles (annealing step 60 s at 95 °C, polymerization step 60 s at 60 °C, elongation step 60 s at 72 °C) NcoI 37 °C for 4 h
rs3615255′-CAG ACC ACA GAC CTG GTC-3′
Denaturation: 10 min at 95 °C, 35 amplification cycles (annealing step 60 s at 95 °C, polymerization step 60 s at 58 °C, elongation step 60 s at 72 °C) BamHI 37 °C for 4 h

Statistical analysis

Comparisons between patients with HS and controls were done using the χ2-test. Odds ratio (OR) and 95% confidence intervals (CI) were estimated using Mantel–Haenszel statistics. The association between patients’ disease severity and carriage of SNPs was calculated using the χ2-test. Comparisons of demographic characteristics were done using the Mann–Whitney U-test for quantitative parameters and by the χ2-test for qualitative parameters. < 0·05 was considered significant.


A total of 190 caucasian patients with HS and 84 age-, sex- and ethnicity-matched healthy controls were enrolled in the study. The clinical characteristics of the patients are shown in Table 2.

Table 2. Clinical characteristics of patients (= 190) with hidradenitis suppurativa enrolled in the study
  1. aNumber of patients is given; many patients presented with involvement in multiple areas.

 Age (years), mean ± SD39·1 ± 11·2
 Male/female, n76/114
 Age at onset (years), mean ± SD26·4 ± 9·6
 Family history, n (%)29 (15·2)
Involved body areasa
 Gluteal area71

The distribution of each of the studied SNP alleles is shown in Table 3. Alleles were in Hardy–Weinberg equilibrium. From the five studied SNPs only the frequency of the −238 G/A SNP at the promoter region of the TNF gene was significantly higher in patients than in healthy controls (P = 0·002). Only haplotypes GGG and GAG of TNF occurred at major frequencies (i.e. > 5%) in the healthy controls (Table 4). The other haplotypes were found at a greater frequency in patients. More precisely, only two (2·4%) controls were carriers of any of the AGG/GGA/AGA/GAA TNF haplotypes compared with 32 (16·8%) patients (< 0·001, OR 8·30, 95% CI 1·94–35·52). Similar findings were not shown for any of TLR4 haplotypes (Table 4).

Table 3. Distribution of genotypes for the single nucleotide polymorphisms (SNPs) of the promoter region of the tumour necrosis factor (TNF) and the Toll-like receptor 4 (TLR4) genes among 190 patients with hidradenitis suppurativa and 84 healthy controls
SNPGenotype/alleleControls, n (%)Patients with HS, n (%) P-value for HW, controls P-value for HW, patients P-value
  1. HW, Hardy–Weinberg equilibrium.

TNF gene
 rs1800750GG83 (98·8)179 (94·2)0·9620·6810·146
GA1 (1·2)11 (5·8)
AA0 (0)0 (0)
G allele167 (99·4)369 (97·1)0·117
A allele1 (0·6)11 (2·9)
 rs1800629GG69 (82·1)155 (81·6)0·1710·5500·397
GA13 (15·5)34 (17·9)
AA2 (2·4)1 (0·5)
G allele151 (89·9)344 (90·5)0·876
A allele17 (10·1)36 (9·5)
 rs361525GG82 (97·6)163 (85·8)0·9120·2910·004
GA2 (2·4)27 (14·2)
AA0 (0)0 (0)
G allele166 (98·8)353 (92·9)0·003
A allele2 (1·2)27 (7·1)
TLR4 gene
 rs4986790AA74 (88·1)174 (91·6)0·5610·5440·377
AG10 (11·9)16 (8·4)
GG0 (0)0 (0)
A allele158 (94·0)364 (95·8)0·388
G allele10 (6·0)16 (4·2)
 rs4986791CC77 (91·7)175 (92·1)0·6900·2320·773
CT7 (8·3)14 (7·4)
TT0 (0)1 (0·5)
C allele161 (95·8)364 (95·8)0·981
T allele7 (4·2)16 (4·2)
Table 4. Distribution of haplotypes of TNF and of TLR4 genes in 190 patients with hidradenitis suppurativa and 84 controls
 Control, n (%)Patients, n (%) P-value
TNF haplotypes
 GGG67 (79·8)126 (66·3)0·037
 GAG15 (17·8)32 (16·8)
 AGG0 (0)6 (3·2)
 GGA1 (1·2)18 (9·5)
 AGA1 (1·2)5 (2·6)
 GAA0 (0)3 (1·6)
TLR4 haplotypes
 AC74 (88·1)171 (90·0)0·507
 GT6 (7·1)11 (5·8)
 GC4 (4·8)5 (2·6)
 AT0 (0)3 (1·6)

The impact of carriage of any of the studied gene SNPs on the natural course of HS was also investigated. To this end, the role of the studied SNPs was analysed with emphasis on the frequency of HS exacerbations, disease severity and response to treatment with TNF antagonists.

The frequency of HS exacerbations was 5·1 ± 1·24 per month (mean ± SE) in patients carrying any of the GAG/AGG/GGA/AGA/GAA haplotypes of TNF, and 2·50 ± 0·52 per month for patients carrying only the GGG haplotype (= 0·027). In patients carrying any of GT/GC/AT haplotypes of TLR4, the frequency of HS exacerbations was 4·87 ± 2·46 per month, and 3·22 ± 0·55 per month in patients carrying only the AC wild-type haplotype (= 0·522).

Seventy-two (37·9%) patients were categorized into Hurley stage I of severity, 72 (37·9%) into Hurley stage II of severity and 46 (24·2%) patients into Hurley stage III of severity. The associations between carriage of any of the studied SNPs and disease severity are shown in Table 5. It was apparent that carriage of the −308 G/A SNP of the TNF gene was associated with more severe disease. More precisely, it was found that among 126 carriers of the GGG haplotype of TNF, 24 (19·0%) were sufferers of Hurley III disease; this was the case for 22 among 64 (34·4%) patients carrying other TNF haplotypes (= 0·021, OR 2·22, 95% CI 1·13–4·39).

Table 5. Frequency of carriage of single nucleotide polymorphisms (SNPs) of tumour necrosis factor (TNF) and the Toll-like receptor 4 (TLR4) genes in association with disease severity
SNPGenotypeHurley I, n (%)Hurley II, n (%)Hurley III, n (%) P-value
 rs1800750GG66 (34·7)69 (36·3)44 (23·2)0·502
GA6 (3·1)3 (1·6)2 (1·1)
AA0 (0)0 (0)0 (0)
 rs1800629GG62 (32·7)63 (33·1)30 (15·8)0·028
GA10 (5·3)9 (4·7)15 (7·9)
AA0 (0)0 (0)1 (0·5)
 rs361525GG65 (34·2)59 (31·1)39 (20·5)0·350
GA7 (3·7)13 (6·8)7 (3·7)
AA0 (0)0 (0)0 (0)
 rs4986790AA70 (36·8)64 (33·7)40 (21·1)0·085
AG2 (1·1)8 (4·2)6 (3·1)
AA0 (0)0 (0)0 (0)
 rs4986791CC69 (36·3)65 (34·2)41 (21·6)0·452
CT3 (1·6)6 (3·2)5 (2·6)
TT0 (0)1 (0·5)0 (0)

Finally, it was assessed whether carriage of any of the studied SNPs may be related with favourable response to treatment with TNF antagonists. This involved the group of 32 patients who were treated with infliximab or etanercept. Treatment with infliximab consisted of single infusions of 5 mg kg−1 at weeks 0, 2, 6, 10 and 14; if required, infusions were repeated once every month. For patients treated with etanercept, the drug was administered subcutaneously at a dose of 50 mg once weekly for 12–24 weeks. Improvement was judged according to the Sartorius score; any decrease from the baseline by 30% by the end of treatment was considered a satisfactory response, as proposed elsewhere.7 From the 32 patients given TNF antagonists, 19 carried only the GGG haplotype of TNF, whereas 13 were carriers of other haplotypes; favourable responses were recorded in 15 (78·9%) patients and five (38·4%, = 0·025) patients, respectively. The OR for lack of response to treatment with TNF antagonists for carriers of at least one SNP allele of the TNF gene was 2·67 (95% CI 1·13–6·29). Concerning the TLR4 gene, 30 patients were carriers of the wild-type haplotype and only two were carriers of other haplotypes; 18 (60%) patients and two patients had favourable responses with TNF antagonists (100%, = 0·516), respectively (Fig. 1).

Figure 1.

 Percentage changes in Sartorius score after treatment with anti-tumour necrosis factor (TNF) agents in correlation with haplotypes of the TNF and the Toll-like receptor-4 (TLR4) genes.


To our knowledge, this is the first time that the presence of SNP alleles of protein molecules related to immune responses has been studied in patients with HS. From the two studied genes, only SNPs of the promoter region of the TNF gene are related both with susceptibility to HS and with the natural course of the disease. More precisely, carriage of the −238 SNP of the TNF gene is related to susceptibility to HS. Carriage of SNPs of the TNF gene is a moderator of the natural course of the disease; it is related to more frequent exacerbations and to more severe disease. Although no other study has been conducted on the linkage of susceptibility to HS with SNPs of the TNF gene, other studies have investigated the association of these SNPs with predisposition to other skin disorders. Three studies have been conducted in patients with psoriasis. One study was performed in a northern Polish population; findings indicated a predisposition to the disease in carriers of the −238 SNP allele of the TNF gene.25 Two other studies explored the role of the −238 and of the −308 SNP alleles of the TNF gene. They concluded that carriage of these SNPs is not an important genetic risk factor but it may be related to clinical manifestation of the disease.26,27

The probable association between Crohn disease and HS based on published data was not verified by our results. Carriage of polymorphisms of the TNF gene was not implicated in the pathogenesis of Crohn disease in numerous studies. However, the same region on chromosome 6p where the TNF gene is located also contains the human leukocyte antigen genes that have been strongly associated to Crohn disease, suggesting that the strong linkage disequilibrium across this gene-dense region confuses the designation of the SNPs involved.28 In contrast, two independent meta-analyses confirmed a strong association of TLR4 polymorphisms and mainly Asp299Gly alone29,30 or in genetic interaction with NOD2/CARD15,31 an intracellular PRR, with increased susceptibility to Crohn disease.

The efficacy of TNF antagonists in HS remains to be explored. Available evidence comes from studies with a limited number of patients; some single-arm studies and few blinded and randomized studies are available. Results are contradictory; some studies indicate significant improvement after treatment with etanercept,11,32 adalimumab33 or infliximab11,13 whereas others show poor efficacy.12,34–36 However, it should be underscored that even in studies where treatment with TNF antagonists was superior to placebo, many of the treated patients failed to respond. Presented results suggest that carriage of at least one of SNP allele of the TNF gene is strongly related to unfavourable response to anti-TNF therapy, explaining, at least in part, failure of treatment. Some studies conducted in patients suffering from rheumatoid arthritis, Crohn disease and psoriasis investigated the impact of carriage of SNP alleles of the TNF gene on response to treatment. Carriage of the −308 SNP allele of the TNF gene has been associated with better response to anti-TNF therapy in patients with rheumatoid arthritis;37 no association was registered for patients with Crohn disease treated with infliximab between carriage of any TNF gene polymorphisms and response to treatment.38 Another study in a cohort of Greek patients with psoriasis investigated the significance of one SNP of the TNF gene not included in the present study. Results disclosed a positive response to anti-TNF treatment in patients who were carriers of the polymorphism.39 The role of these SNPs in HS should be considered in light of the excess amounts of TNF-α in HS lesions compared with the levels found in lesions from patients with psoriasis9 and of the subsequent downregulation of cytokine concentrations in HS skin biopsies after administration of adalimumab.40

The present study failed to identify any impact of TLR4 SNPs on the susceptibility to HS or on disease severity. For many years, HS was known as acne inversa because acne most commonly affects the face but not the rest of the body, whereas HS behaves in the opposite way. SNPs of TLR4 do not induce any susceptibility for acne vulgaris whereas they seem to affect its course. In a recent study by our group, the role of the two SNPs of the TLR4 gene was investigated in 191 patients. The frequency of SNP carriage was similar between patients and healthy controls. However, carriage of SNP alleles of the TLR4 gene protected against the development of acne conglobata, which is the most severe form of acne.41 The lack of impact of TLR4 SNPs on the phenotype of HS in contrast to the impact on the phenotype of acne vulgaris adds further to the scientific rationale that acne and HS are different disease processes.

In conclusion, the presented results indicate a major role for SNPs at the promoter region of the TNF gene for susceptibility to and for the natural course of HS in a cohort of caucasian patients; carriage of SNPs is linked with disease severity and unfavourable responses to TNF antagonists. There is no doubt that further studies are warranted to confirm these findings in larger patient populations, especially in the case of patients treated with anti-TNF agents, and in other ethnicities. However, these findings introduce the need for pharmacogenomics as a marker of response to treatment.