INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Elevated serum angiotensin-converting enzyme (ACE) levels are found in approximately 60% of patients with sarcoidosis (1). Although initially thought to be a useful tool in monitoring the clinical course of the disease, controversy now surrounds the role of ACE levels in sarcoidosis (2)

The gene-encoding ACE (or dipeptidyl carboxy peptidase 1: DCP1) is located on chromosome 17q35 and consists of 26 exons spread over approximately 24 kb. A 250-bp deletion/insertion polymorphism exists in intron 16 of the ACE gene and the deletion variant is associated with higher serum levels of the enzyme. Thus, serum ACE levels correspond to ACE insertion/deletion (I/D) genotypes in the order: II < ID < DD. This I/D polymorphism is reported to account for 47% of the phenotypic variation in serum ACE levels (3).

The discovery of the I/D polymorphism with its consequent effects on serum ACE levels has encouraged several studies in patients with sarcoidosis to determine whether this variation influences susceptibility to disease or disease outcome (Table 1). Two main conclusions can be drawn from these studies: (1) Sarcoidosis populations differ substantially in their I/D frequencies (For example, in the Japanese, the I allele is more common than in whites); (2) Controversy persists as to whether there is an association between the ACE I/D polymorphism and sarcoidosis disease severity (3, 9, 12, 13).

Although some studies did not find an association between the ACE I/D polymorphism and sarcoidosis disease severity (3, 9), a study of Finnish patients (13) revealed an association between the deletion variant (D) and disease progression. In that study disease progression was defined as persistent chest radiographic changes and/or extrapulmonary disease at 5 yr from presentation. The association was identified only after the ID and DD genotypes were grouped together. In the African-American population (12) the reverse was the case: those with the insertion variant (I) had a threefold greater risk of chest radiographic progression after 2 yr, after grouping the ID and DD genotypes together. This association was not found in a US white population (12).

In a murine model of radiation-induced pulmonary fibrosis (4) the extent of fibrosis was inversely correlated with the intrinsic lung activity of both plasminogen activator (PLA) and ACE, supporting the choice of the ACE gene as a strong candidate gene for the development of fibrosis in sarcoidosis. However, previous studies concerning the clinical significance of the ACE I/D polymorphism rendered contradictory findings. Our study, therefore, aimed to clarify whether the ACE I/D polymorphism is associated with: (1) sarcoidosis disease susceptibility; (2) pulmonary disease severity, particularly fibrosis at 2 and at 4 yr from presentation; (3) pulmonary disease progression at 2 and at 4 yr from presentation. We examined two distinct European populations in order to detect whether the effects of the polymorphisms, if any, are independent of geographic and ethnic bias.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

UK cases (n = 184) and Czech cases (n = 56) were recruited consecutively for the outpatient clinics at the Royal Brompton Hospital, London (tertiary referral center), and the Olomouc Hospital (secondary center). Informed consent for genetic analysis was obtained prior to phlebotomy. All patients had (1) clinical features of pulmonary sarcoidosis, (2) diagnosis confirmed on transbronchial, Kveim, or lymphnode biopsy, and (3) received the appropriate treatment for care of the disease. Age at presentation, sex, and treatment history of both the UK and Czech sarcoidosis populations studied are outlined in Table 2. Current smoking history, defined as the smoking history at the time of presentation and throughout a 2-yr follow up period, is also recorded.

The healthy control UK population data used for comparison were published previously (14) and comprised unmatched white patients (n = 386) of UK origin recruited from three centers: London, Cardiff, and Belfast. The Czech control population consisted of unrelated, age, sex, and ethnically matched white blood donors (n = 179) from a region close to Olomouc.

Although ethnic uniformity (defined as "uniformity of people with a common cultural heritage that sets them apart from others": MeSH browser, Pubmed, National Library of Medicine) was thought to prevail within each group studied, it was assumed that the UK and Czech populations are ethnically distinct from each other. Project approval was granted by the local ethics committee at each of the centers.

Evaluating Sarcoidosis Pulmonary Disease Severity

Radiography. The chest radiographic series for each patient was examined and compared to determine disease coursewhen possible these comprised radiographs on presentation to hospital, followed by radiographs from 2 and 4 yr later, and the final/most recent radiograph. Each patient's radiographs were analyzed for disease severity by a respiratory physician using standard radiographic staging for sarcoidosis. Briefly, this staging comprises five broad stages: Stage 0, normal; Stage I, bilateral hilar lymphadenopathy (BHL); Stage II, bilateral hilar lymphadenopathy and parenchymal infiltration; Stage III, parenchymal infiltration with early features of fibrosis; Stage IV, irreversible fibrosis with loss of lung volume. Those who improved were defined as those patients whose radiologic appearances had changed from a higher to a lower radiologic stage, and those who progressed were defined as those who had changed from a lower to a higher stage at these two time intervals. Radiologic appearances were otherwise classified as unchanged. As a quality control measure, another respiratory physician carried out a blind assessment of the chest radiographs. Comparison of radiograph staging from the two independent assessments showed that < 5% of the results differed.

Presentation chest radiographic data were available only on 129 of the 184 UK patients with sarcoidosis studied. The rest had been previously returned to their appropriate referring hospitals. Chest radiographs, taken at least 2 yr after initial presentation, were available on 118 of the UK patients. Of this group, a total of 101 had radiologic follow-up at 4 yr. The difference in patient numbers at presentation compared with 4 yr onwards is accounted for by the presence of patients in whom a diagnosis had been recently made in this study. These patients fulfilled the diagnostic criteria for study inclusion, but because their diagnosis dates were more recent, they had not been assessed by radiography for the preferred 4-yr period.

Although a complete radiologic time series was not available for the Czech patients, chest radiographs taken 2 yr from presentation were available on 56 patients. Pulmonary severity for both UK and Czech populations was therefore assessed on radiologic appearances at 2 yr after presentation.

In the absence of a complete radiologic series for the Czech population, ACE genotype and progression of chest radiographic appearances (from presentation to 4 yr) were analyzed only in the UK sarcoid population (n = 101). Patients with Stage IV disease at presentation (n = 18), by definition, cannot improve or worsen radiologically and were therefore omitted from analysis.

Pulmonary function testing. All patients who had pulmonary function performed in the same laboratory at presentation and at the 4-yr follow-up were included for analysis. Ninety-five patients fulfilled the inclusion criteria for assessment of pulmonary function. The following parameters were recorded: FEV₁ and FVC (PK Morgan Instruments, Kent, England). Gas transfer measurements (DL_CO) were performed by using the carbon monoxide diffusing lung capacity single-breath technique and were adjusted for alveolar volume (6200 Autobox DL; Sensormedics, Yorba Linda, CA); the results were corrected for hemoglobin. Severity data were expressed as percent-predicted values. Disease progression data at 4 yr were expressed as the percentage change from presentation value.

Identification of the Polymorphism and Primer Sequences

Polymerase chain reaction with sequence-specific primers (PCR-SSP) was used to determine the ACE I/D genotype of each person studied. The conditions were identical to those described previously (15). The forward and reverse primer sequences consisted of 5'-CTGGAGACCACTCCCATCCTTT-3' and 5'-GATGTGGCCATCACATTCGT, respectively. From the 184 UK patients initially identified, 180 were successfully genotyped (98%); all 56 Czech sarcoid patients were typed.

Statistical Analysis

The MS-DOS computer package EpiInfo6 (available from the Centers for Disease Control and Prevention, USA via the internet http://www.cdc.gov/epo/epi/epiinfo.htm) was used to estimate the statistical power to detect significant differences between study groups given the group sizes and estimated/observed frequency of the factor being compared. The study was designed to predict an odds ratio of 2.0 with 80% confidence intervals (CI) and a power of 95%.

Statistical analysis of the ACE frequency and radiologic data for the two populations studied were calculated using chi-square contingency table analysis with the appropriate number of degrees of freedom. Analysis of pulmonary function data was performed using the Mann-Whitney U test. A p value < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ACE I/D Polymorphism and Sarcoidosis Susceptibility in UK and Czech Populations

The results of the ACE genotyping for the UK and Czech populations are given in Table 3. No significant difference was seen in the distribution of ACE I/D genotypes between patients and controls for allele, genotype, and phenotype frequencies.

Sarcoidosis Pulmonary Disease Severity

Radiology. For the 101 UK patients with a complete 4-yr chest radiographic series, the proportion of patients in each of the four severity categories varied according to the time points (Figure 1). At presentation, the proportions comprised 0% normal or Stage 0, 22.8% Stage I, 33.7% Stage II, 20.8% Stage III, and 22.7% Stage IV. At the 2-yr, 4-yr, and most recent/current assessment points, the proportions of patients in Stages II and III were found to have fallen dramatically, whereas the proportions in Stages I and IV had risen.

View larger version (28K): [in this window] [in a new window]   Figure 1.   UK pulmonary sarcoidosis severity distribution at 4 yr from presentation (n = 101).

The patients included in the Czech group were found to represent all stages of disease severity, from those patients who had a normal chest radiograph appearance, to those who had features of permanent fibrotic changes. The stratification of patients according to disease severity in the UK and Czech populations is given in Table 4. No association was identified between ACE I/D genotype, allele or phenotype frequency, and each radiographic severity group in either sarcoid population.

The distribution of the severity groups (normal/Stage 0 = 1.8%, Stage I = 50%, Stage II = 33.9%, Stage III = 14.3%, and Stage IV = 0%) in the Czech population differed from that of the UK population (normal/Stage 0 = 11.9%, Stage I = 21.2%, Stage II = 30.5%, Stage III = 11.9%, and Stage IV = 24.6%) for the same 2-yr time interval. The clinical significance of this is difficult to ascertain.

Pulmonary function testing. The median percentages of predicted values (range) at presentation for the three pulmonary function parameters were as follows: FEV₁ = 91 (26 to 136)%; FVC = 98 (33 to 146)%; DL_CO = 77 (20 to 145)%. When analyzed for ACE I/D genotype, the mean percentages of predicted values (± SDs) of FEV₁, FVC, and DL_CO (Figure 2) at presentation were remarkably similar for all three subgroups: II (FEV₁ = 85.7 ± 23.7%, FVC = 96.5 ± 22.7%, DL_CO = 70.4 ± 18.7), ID (FEV₁ = 83.1 ± 23.0%, FVC = 96.1 ± 15.2, DL_CO = 74.6 ± 14.9) and DD (FEV₁ = 87.5 ± 22.1, FVC = 96.0 ± 21.4%, DL_CO = 75.3 ± 23.4). Likewise, no difference was demonstrated between phenotype and allele frequency data and disease severity defined on pulmonary function testing.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Comparison of ACE I/D genotype and DLCO (% pred) at presentation in a sarcoidosis population.

ACE I/D Polymorphism and Pulmonary Fibrosis

Analysis of ACE I/D genotype in relation to pulmonary fibrosis did not reveal any significant findings for either the UK (p = 0.9) or Czech (p = 0.7) populations.

ACE I/D Polymorphism and Pulmonary Disease Progression

Radiology. Thirty-eight patients (46%) demonstrated radiologic improvement, 26 (31%) progression, and 19 (23%) no change. An association was not identified between genotype frequency and radiographic progression at 2 yr or at 4 yr (Figure 3) from presentation. Similarly, no association was detected for phenotype or allele (Table 5) frequency. The genotypes were then grouped in a similar method (II + ID versus DD and ID + DD versus II) to the previous studies (12, 13). This approach revealed an association between carriage of the I allele (II + ID versus DD) and disease progression (p = 0.02).

View larger version (17K): [in this window] [in a new window]   Figure 3.   ACE I/D genotype and radiologic progression (from presentation to 4 yr) in a UK sarcoidosis population (n = 83).

From the radiologic data it was seen that, over time, the patients' distribution tended to taper toward the two extremes, i.e., normal (Stage 0) or severe (Stage IV) disease (Figure 1). Because the largest changes in distribution were seen in patients with Stages II and III, these were analyzed more closely to determine if there was an association between ACE I/D genotype and progression from Stage II/III to Stages I or IV. Again no association was detected.

Pulmonary function testing. No significant differences and no marginal trends in FEV₁, FVC, or DL_CO (Figure 4) over a 4-yr period were detected when comparison was made for the three ACE I/D genotype groups (Table 6). Similarly, no associations were identified between pulmonary function testing and ACE I/D phenotype or allele frequency. In contrast to radiologic progression, genotype grouping did not yield any differences in pulmonary disease progression defined on lung function testing.

View larger version (12K): [in this window] [in a new window]   Figure 4.   Comparison of ACE I/D genotype and change in DLCO (% pred) at 4 yr from presentation in a sarcoidosis population.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study was designed to attempt to clarify the relationship between the ACE I/D polymorphism and the severity and progression of lung disease in sarcoidosis. In order to remove any bias of local geographic or ethnic differences, UK white and Slavic (Czech) white patient cohorts were studied. We found no association between genotype or allele frequencies and disease severity and progression in either of these populations.

The large number of previous association studies examining the ACE I/D polymorphism in this disease reflects the interest in the role of this gene in sarcoidosis. The ACE I/D polymorphism has been shown to affect serum ACE levels (3). However, controversy surrounds not only the role of serum ACE levels in sarcoidosis but also the ACE I/D polymorphism in sarcoidosis susceptibility and disease progression. This study aimed to clarify this role and to look particularly for a relationship between the polymorphism and pulmonary fibrosis with which serum ACE levels are associated. The results from this study do not support a role for the ACE I/D polymorphism in sarcoidosis susceptibility in the United Kingdom and Czech white populations. This concurs with the findings of Maliarik and colleagues (12) in the United States white populaton but differs from the African-American population, which showed an increased risk for sarcoidosis associated with ID heterozygotes (RR = 1.3; 95% CI = 0.72 to 2.36) and DD homozygotes (RR = 3.17; 95% CI = 1.5 to 6.71). Furthermore, disease progression was paradoxically associated with the II variant in the latter population. The difference in findings between the white and African-American populations suggests that the role of the I/D polymorphism is population-dependent. Consistent with this is that a racial difference in the relationship between serum ACE levels and the I/D polymorphism has been reported (16).

This study did not demonstrate any association between ACE I/D genotype or allele frequency and pulmonary disease severity or progression in two distinct European white sarcoid populations. These findings concur with those of Papadopoulos and colleagues (17) who found no association between isolated cases of sarcoidosis and disease severity despite observing a significant increase in DD genotype in patients with coexisting autoimmune disease and sarcoidosis and chest radiographic Stage III. Genotype grouping analysis (II + ID versus DD and DD + ID versus II) in this study revealed an association between carriage of the insertion allele (II + ID) and radiographic progression. However, physiologic assessment of disease progression yielded no association. These findings differ from those of the Finnish study of white individuals in which an association between disease progression and the deletion variant was found. Although the United States African-American study revealed a similar association between the insertion variant and radiologic progression, the use of radiologic change as the sole criterion for progression of sarcoidosis is suspect. Furthermore, as the functional role of the homozygote and heterozygote states of the ACE I/D polymorphism remain unknown, clear conclusions cannot be derived from genotype grouping. Analysis of allele frequency, as in this study, may be more informative.

No association was demonstrated between the ACE I/D polymorphism and pulmonary fibrosis in the two sarcoid populations we studied. This is the first study to look for an association with lung fibrosis specifically. Approximately 20% of patients with sarcoidosis develop fibrosis (18). Although data from the Czech population in this study were in keeping with this number (14%), the United Kingdom population contained a significantly higher proportion of subjects with fibrosis (47%). The main reason for this is the patient population from which the United Kingdom group was recruited. These patients were derived from a specialist tertiary respiratory hospital to which asymptomatic patients are less likely to be referred. Furthermore, as those with more severe disease are followed at more frequent intervals and for a longer period, the likelihood of these patients being recruited to this study was higher.

One criticism of this study is that the control United Kingdom data were unmatched for age and sex, which could account for the absence of any association between ACE I/D genotype and sarcoidosis susceptibility and pulmonary severity or progression. However, as there was no significant difference in ACE I/D genotype distribution between this unmatched control United Kingdom population and control data from previously published United Kingdom studies (6, 7), we believe our comparison to be acceptable. Furthermore, as the genotype frequency in both United Kingdom and Czech patients are similar to the patient data from the two United Kingdom studies of whites mentioned as well as patient data from the aforementioned United States white group (12), this suggests that the results in this study are not biased by either ethnic or geographic population similarities or variations.

In summary, we aimed to clarify the role of the ACE I/D polymorphism in sarcoidosis susceptibility and specifically pulmonary severity, fibrosis, and progression. We conclude that the I/D polymorphism has no role in sarcoidosis susceptibility in European white individuals and that it is not a regulatory variant in this disease. We believe that an additional functional variant, possibly in linkage with the I/D polymorphism, exists within the ACE gene. This variant may have a regulatory region that contains a response element to either the causal agent of sarcoidosis or to a mediator that this agent releases. This theory is supported by the work of Rigat and colleagues (19) and Tiret and coworkers (20) who demonstrated that the I/D polymorphism accounted for only 47 and 28% of the phenotypic variance of serum ACE, respectively. A recent sequence variation analysis of the complete dipeptidyl carboxypeptidase-I (angiotensin-I-converting enzyme) DCP1 gene and its flanking regions revealed 78 polymorphisms within the ACE gene that can be grouped into three main haplotype branches (21, 22). Of these 78 polymorphisms, five resulted in amino acid changes. Further genetic studies are necessary to identify which of these polymorphisms provides the functional basis of the ACE gene. This knowledge will allow us to determine the exact contribution of this gene in sarcoidosis.