Introduction

Abstract Introduction Materials and Methods Results Discussion References

Fibronectin (FN) is a multifunctional heterodimer glycoprotein composed of two similar polypeptide chains. It is involved in regulating interactions between the cell and extracellular matrix proteins, serving multiple functions in tissue repair and fibrosis. FN is encoded by a gene of over 75 kb in length, located on chromosome 2q34-36 and consisting of at least 50 exons with an average length of 147 base pairs, alternating with introns of variable lengths (1- 5). Six restriction fragment length polymorphisms (RFLPs) of the fibronectin gene have been identified: HindIII, HaeIII b, TaqI a, TaqI b, MspI, and HaeIII a (Figure 1) (2, 6).

View larger version (22K): [in this window] [in a new window]   Figure 1.   Localization of the FN RFLPs along the FN gene. Dashed lines indicate approximate intron positions. Bold lettering denotes the RFLPs examined in this study.

There are different forms of FN produced by the variable splicing of a single gene. Alternative splicing at three different domains can produce up to 20 FN subunit isoforms (5, 11). The extra domain I (ED I), which can be excluded from or included in FN, is present in cellular fibronectin (produced by fibroblasts and other mesenchymal cells) and absent from plasma fibronectin (secreted by hepatocytes). ED I is overexpressed in wound healing, epithelial fibrosis, and vascular intimal proliferation (5). The extra domain II (ED II) is not present in plasma fibronectin and its function is unknown (5). Five different forms of FN mRNA can be generated by alternative splicing of the type III connecting segment (III CS) (5). FN III CS splice variants are involved in cellular inflammation, and it has been suggested that they influence chemotaxis of fibroblasts and bronchial epithelial cells, as well as directing macrophages and lymphocytes to the airway epithelium (12, 14).

FN acts in the lung as a chemoattractant and adhesive substrate for fibroblasts, providing these cells with the structure for proliferation. Activated fibroblasts, in turn, produce cellular FN and organize FN matrix assembly on the cell surface via FN receptors, inducing further migration, attachment, and proliferation of fibroblasts. These interactions between fibroblasts and FN are thought to play an important role in the development of lung fibrosis (4, 12).

Systemic sclerosis (SSc) is a generalized connective tissue disorder, characterized by thickening and fibrosis of the skin (scleroderma), Raynaud's phenomenon, widespread damage of small arteries and microvessels, musculoskeletal manifestations, and distinctive forms of visceral involvement (15). Fibrosing alveolitis (FASSc) occurs in roughly 55% of patients in collected series. Prognosis is difficult to evaluate because of the impact of the disease in other organs. However, once lung disease has reduced physiological measures by 50%, survival is reduced by 50% and morbidity increases (16, 17). Indeed, lung fibrosis is the major cause of death in this disorder, and therefore the ability to predict the development of pulmonary fibrosis is important (18, 19). This prediction will involve the identification of genetic and environmental risk factors that will individually and collectively determine a complex disease risk assessment. This will vastly improve our surveillance procedures and the timing of commencement of treatment.

FN is released in increased amounts by alveolar macrophages from patients with SSc and active alveolitis. This overexpression of FN in the epithelial lining fluid of the lower respiratory tract is likely involved in the pathogenesis of FASSc, contributing, with other growth factors, to the recruitment of fibroblasts to the lung and their adhesion and proliferation at sites of lung repair (20).

There is evidence to support the concept of a genetic predisposition to SSc and pulmonary fibrosis. Analysis of the major histocompatibility complex (MHC) class II loci has demonstrated that DR3/DR52a is associated with pulmonary fibrosis in SSc patients. Among the autoantibodies associated with disease subsets, antitopoisomerase autoantibodies (anti-Scl-70) are correlated with lung fibrosis. The presence of DR3/DR52a and/or antitopoisomerase autoantibodies gives a relative risk of developing pulmonary fibrosis of 16.7; the two risk factors show some, but not complete, overlap (18).

Such genetic association explains some, but not all, of the genetic susceptibility to lung fibrosis in SSc. Other polymorphisms are likely to play a role. Of the candidate genes known to be involved in pathogenesis, polymorphisms of proteins encoding genes involved in the development of fibrosis might be likely candidates. In this study, we set out to determine whether specific polymorphisms in the fibronectin gene confer germline genetic susceptibility to the development of pulmonary fibrosis in SSc patients.

	Materials and Methods

Abstract Introduction Materials and Methods Results Discussion References

Patients

FN polymorphisms were investigated in 161 patients with SSc (29 men and 132 women; 94 subjects with FASSc and 67 subjects with SSc without FA [NFASSc]) and in 253 control subjects.

The median age of the SSc patients was 57 yr, range 30 to 75 yr. The median disease duration was 5 yr. All were Caucasians from the United Kingdom. No selection process was used for inclusion in this study other than willingness to take part in the study and full disease characterization. All patients fulfilled the American Rheumatology Association preliminary criteria for SSc (21). On the basis of the extent of skin involvement, patients were classified as having either diffuse cutaneous SSc (dcSSc) or limited cutaneous SSc (lcSSc). Patients with truncal and acral skin involvement were classified as dcSSc and patients with skin involvement limited to hands, face, feet, and forearms were classified as lcSSc (15, 22). Fibrosing alveolitis in SSc patients was diagnosed by the following criteria: no other known causes of fibrosis, bilateral persisting radiographic shadows, and restrictive pulmonary deficit and/or reduced gas transfer measurements (25). The presence of FASSc was confirmed by high-resolution computerized tomography. Interspaced 3-mm sections were obtained from the lung apices to the lung bases at 10-mm intervals and reconstructed with a high resolution "bone" algorithm (26). The diagnosis of FASSc was confirmed by lung biopsy in 34 patients of the fibrotic group (n = 94). Written patient consent was obtained for all subjects, and project authorization was given by the Ethics Committee of the Royal Brompton National Heart and Lung Hospital.

Control subjects were healthy Caucasians from the United Kingdom with no previous history of fibrosing lung disease. The median age of the control group was 43 yr, range 18 to 64 yr; 53% were men.

DNA Extraction

Peripheral venous blood (20 ml) was collected into 5% Na₂ ethylenediaminetetraacetic acid and stored at 20°C, pending genomic DNA extraction using a modified high-salt extraction technique (27, 28).

Polymerase Chain Reaction

To identify specific polymorphisms in the FN gene, previously published primers were used in polymerase chain reactions (PCR) to amplify the regions containing four RFLPs within the human fibronectin gene (HaeIII b, MspI, HindIII, TaqI b) (2). HaeIII b and TaqI b were selected for study because of their proximity to the splicing region. Each PCR amplification was performed with 10 ng of genomic DNA in a total volume of 25 µl containing 6.25 pmol of each primer, 2.5 µl of PCR reaction buffer (10× concentrated) (Boehringer Mannheim, Lewes, East Sussex, UK), 200 µM of each deoxynicotinamide triphosphate (Pharmacia Biotech, St. Albans, Herts, UK), and 0.75 U of Taq DNA polymerase (Boehringer Mannheim).

The FN HaeIII b site was amplified with the primers F3 (5'-AGC TCT ATT CCA CCT TAC AAC ACC G-3') and F4 (5'-CTC CCA GGA GAC TGT GAG CAC-3') and the following cycle conditions: 30 s at 94°C, 30 s at 56°C, and 30 s at 72°C (30 cycles). The FN MspI site was amplified using the primers F7 (5'-GCC TGG TAC AGA ATA TGT AGT G-3') and F8 (5'-TGC CAT TAA GAG CAA CGA TGC-3') and the following cycle conditions: 1 min at 94°C, 1 min at 56°C, and 2 min at 72°C (30 cycles). The FN HindIII polymorphic site was amplified with the primers F1 (5'-CAG ATA AAT CAA CAG TGG GAG C-3') and F2 (5'-TGC GAT GGT ACA GCT TAT TCT C-3') and the following cycle conditions: 1 min at 94°C, 1 min at 56°C, and 2 min at 72°C (30 cycles). The FN TaqI b polymorphic site was amplified using the primers F5 (5'-TGC ATT AGC GTT ATG GCC ATG-3') and F6 (5'-GTT TGT TGT GTC AGT GTA GTA-3') and the following cycle conditions: 30 s at 94°C, 30 s at 55°C, and 30 s at 72°C (30 cycles). The final extension was maintained for 10 min. All PCR amplification was automated with a thermal cycler (Crocodile III; Oncor Appligene, Chester-Le-Street, Durham, UK).

Digestion

After performing a PCR for each polymorphic site (HaeIII b, MspI, HindIII, TaqI b), 25 µl of each product were digested to completion with HaeIII, MspI, HindIII, and TaqI (all Pharmacia Biotech), respectively, under conditions recommended by the supplier.

Gel Electrophoresis

The digested product was electrophoresed on a 1.5% agarose gel containing 0.1% wt/vol ethidium bromide in Tris-acetate (TAE) for 60 min at 100 V, and a photograph was taken. The digested PCR products were compared with the known cleavage patterns and each sample was assigned a genotype accordingly (2).

Four random PCR products from each RFLP site were purified using the Wizard PCR Preps DNA Purification System (Promega, Southampton, UK), and then manually sequenced using Sequenase V 2.0 (USB, Amersham, Bucks, UK). Using the Basic Local Alignment Search Tool sequence database program (BLAST; available online at http://www.hgmp.ac.uk), we confirmed that the amplified region had a 100% homology with the human FN gene (29).

Data Handling and Statistical Analysis

In this association, study correlations were made between disease subtype, pulmonary fibrosis, and the different genotypes. For analytical purposes, patients were divided into those with and without fibrosing alveolitis: FA was defined by an abnormal computed tomography scan, confirmed by biopsy in 35 patients, with lung function tests with vital capacity less than 70% predicted. The data sets in this study were compared using a standard 2 × 2 chi-squared analysis using SIGTEST, a computer-based program that uses a Woolf-Haldane correction in cases of small numbers. This program has a facility for the calculation of the chi-squared statistic, the significance value, and the relative risk. Each group of subjects (i.e., the SSc group as a whole, and the subgroups FASSc and NFASSc) was compared with the control group, and the significance values after correction (P_corr) for the number of analyses made are stated in the tables. A P_corr value of < 0.05 was considered statistically significant. Given the size of our patient population, a difference in genotype frequency of 10% would produce a significance value of < 0.05. The control population was tested for conformity to the Hardy-Weinberg equilibrium using a 2 × 2 chi-squared test between observed and expected numbers.

	Results

Abstract Introduction Materials and Methods Results Discussion References

Genotypes and Allele Frequencies

To determine whether FN gene polymorphisms differ between patients with SSc and control subjects, four RFLPs (HaeIII b, MspI, HindIII, and TaqI b) of the FN gene were analyzed in 414 individuals (161 patients and 253 control subjects). For each restriction enzyme, three genotypes were identified corresponding to the presence of the cutting site on neither (represented as AA for HaeIII b, CC for MspI, EE for HindIII, and GG for TaqI b), one (represented as AB, CD, EF, and GH, respectively), or both chromosomes (represented as BB, DD, FF, and HH, respectively) (Figure 2).

View larger version (77K): [in this window] [in a new window]   Figure 2.   HaeIII (a), MspI (b), HindIII (c), and TaqI (d) digestions of the PCR product showing three different genotypes. The size of each fragment is indicated on the right of the gel. A 100-bp DNA ladder is shown in the left lane.

Genotype and allele frequencies for each RFLP are shown in Tables 1234 for control subjects and patients with SSc. The distribution of the different genotypes and alleles was consistent with the Hardy-Weinberg equilibrium.

HaeIII

Comparison of the scleroderma patients with control subjects for the FN HaeIII genotypes showed a significant difference in the heterozygote (AB) populations. Although this genotype was found in 61% of the patients, it was present in only 46% of the control subjects (P_corr = 0.008) with a reciprocal decrease in the frequency of the homozygote cutter (BB) genotype (19% patients, 34% controls; P_corr = 0.003).

When subgroup analysis was performed, we found that the FA group was responsible for this difference. For the FN RFLP HaeIII, the distribution of genotypes AA, AB, and BB in the control subjects was 20%, 46%, and 34%, respectively, and for the FASSc patients the distribution was 21%, 62%, and 17%, respectively. The allele frequencies in the control group were 0.42 for allele A and 0.58 for allele B, and in the FASSc group 0.52 for allele A and 0.48 for allele B (Table 1). By using chi-square analysis of the FN HaeIII RFLP pattern, we found a significant association (P_corr = 0.022) between patients with FASSc and the genotype AB and a reciprocal negative association (P_corr = 0.006) between patients with FASSc and the genotype BB. Similar but not significant trends were observed in the NFASSc population.

When patients were categorized according to scleroderma subtype (dcSSc and lcSSc), differences were again observed for the heterozygote genotype. Sixty-five percent of dcSSc patients had the AB genotype (P_corr = 0.034, compared with controls) and 16% had the BB genotype (P_corr = 0.028). Similar findings emerged from comparisons of the lcSSc group with controls: 60% had the AB genotype (P_corr = 0.026) and 20% had the BB genotype (P_corr = 0.016).

MspI

Comparisons between SSc patients and control subjects for the FN MspI genotypes did not show any significant difference. The distribution of genotypes CC, CD, and DD in the control subjects was 5%, 54%, and 41%, respectively, whereas in the patient group it was 6%, 62%, and 32%.

When subgroup analysis was performed, comparisons of the FASSc patients with control subjects showed a significant decrease in the homozygote (DD) FASSc patients. This genotype was present in only 28% of FASSc patients, whereas it was present in 41% of control subjects (P_corr = 0.038). The allele frequencies in the control subjects were 0.32 for allele C and 0.68 for allele D, and in the FASSc patients 0.39 for allele C and 0.61 for allele D (Table 2).

On the basis of scleroderma subtype (dcSSc and lcSSc), no significant differences were observed between dcSSc or lcSSc and the control population for the FN RFLP MspI. Six percent of dcSSc patients had the CC genotype, 67% the CD genotype, and 27% the DD genotype, whereas 6% of lcSSc patients had the CC genotype, 63% the CD genotype, and 31% the DD genotype.

HindIII and TaqI

When comparisons were made between SSc patients and controls for the HindIII and TaqI restriction enzyme cutting sites, no differences were seen in genotype or allelic frequencies in whole-group or disease subtype analysis (Tables 3 and 4). For both enzymes, the majority of patients exhibited homozygote cutting. (Genotype frequencies for HindIII were SSc 60%, control subjects 55%; TaqI SSc 82%, control subjects 80%. Allelic frequencies for HindIII cutting were SSc 0.78, controls 0.75; TaqI cutting SSc 0.91, control subjects 0.89.)

Genotype Associations

To determine whether associations of HaeIII and MspI genotype frequencies were better determinants of FA than single-enzyme RFLP analysis, we assessed the relative risk of developing FA when both heterozygote (AB and CD) genotypes were present in the same individual.

A highly significant association was found. Forty-five percent of the FASSc patients were heterozygous for both restriction enzymes compared with 29% of control subjects. This resulted in an odds ratio of 3.08 (P = 0.0059). The coassociation of HaeIII and MspI heterozygosity confers an almost twofold excess risk of developing FA in the context of SSc, with an etiological fraction of 22.27% (Table 5).

	Discussion

Abstract Introduction Materials and Methods Results Discussion References

This study has shown a strong association between FN gene polymorphisms and fibrosing alveolitis. Primary analysis has shown a significant difference when we compared scleroderma as a whole with control subjects. Disease subset analysis of those with lung disease compared with those without fibrosis showed that the difference between subjects with scleroderma and control subjects is determined by those with lung fibrosis. First, we found that the genotype frequency of FN HaeIII AB was significantly higher in patients with scleroderma who have fibrosing alveolitis than in the normal population, whereas no significant association was seen in the NFASSc group. Second, the combination of genotypes AB (FN HaeIII RFLP) and CD (FN MspI RFLP) carries an increased relative risk of almost twofold of developing fibrosing alveolitis. Third, allele A (FN HaeIII RFLP) was more frequent in the fibrotic group. This study is therefore consistent with the concept that FN polymorphisms may contribute to the development of fibrosing alveolitis in systemic sclerosis. Our patient population was representative of the demography of systemic sclerosis in terms of male/female ratio, age range, and prevalence of FASSc. To ensure that we did not confound our study by ethnic mixing, the populations studied were Caucasian only.

The fundamental pathogenetic mechanism of lung injury in SSc remains unknown. A role is attributed to alveolar macrophages, present in increased numbers in bronchoalveolar lavages from SSc patients. Activated alveolar macrophages that produce growth factors or fail to release inhibitory signals may influence fibroblast proliferation and connective tissue-matrix synthesis. One of the growth factors released by the macrophage is FN. It has been demonstrated that FN is overexpressed in the respiratory tract of patients with FASSc and can contribute to the recruitment and attachment of the fibroblasts in scleroderma lung disease (30). This is consistent with the studies of Rennard and colleagues (31), who previously showed that macrophages from patients with idiopathic pulmonary fibrosis produced FN at a rate 20 times higher than normal macrophages, suggesting that FN plays an important role in the pathogenesis of fibrosis. These studies indicated that FN was a good candidate gene for polymorphism studies and our study has demonstrated an increased relative risk to patients with HaeIII and MspI heterozygosity.

To our knowledge there are no studies pointing to associations between polymorphisms of the FN gene and fibrosing alveolitis or other diseases. The relevance of fibronectin polymorphisms in scleroderma is, however, supported by a study of fibroblast DNA obtained from the skin of sclerotic lesions of Japanese patients (32). Two point mutations were identified in the region adjacent to the cell-attachment tetrapeptide DNA sequence in the cell-binding domain (exon 7) of the fibronectin gene. This study supports the concept that a mutant FN might cause aberrant cell interaction in the pathogenesis of skin sclerosis, although the site of these mutations is different from those reported in this study.

In conclusion, we have shown that FN polymorphisms are present in human disease and may be important factors in the development of fibrosing alveolitis in SSc. At present, the FN RFLPs have not been associated with functional changes, and further work is required to evaluate the functional consequences of the polymorphisms described in this study.