INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Approximately 60 to 80% of nontuberculous mycobacterium in Japan is caused by Mycobacterium avium complex (MAC) composed of two distinct organisms, M. avium and M. intracellulare (1). Although the introduction of new macrolides has improved therapeutic results of pulmonary MAC infection, it is still considered intractable and sometimes progresses rapidly even in human immunodeficiency virus (HIV)-negative patients (2). MAC is environmentally ubiquitous, and thus, exists in sites such as water, soil, air, and even cigarettes (5). Three clinical syndromes are produced by MAC, including disseminated MAC, pulmonary MAC, and MAC lymphadenitis (8). Disseminated MAC develops in persons with advanced immunodeficiency conditions. Although specific risk factors for pulmonary MAC have not been identified, it is known that pulmonary MAC usually develops in persons with pulmonary lesions because of prior tuberculosis infection and heavy smoking (4). This may suggest that an impairment of mucociliary clearance might be an important predisposing factor. Aging is also suggested to be a risk factor. However, even in the nonsmoking younger generation, pulmonary MAC infection patients are occasionally observed who have neither past history of pulmonary disease nor underlying immunodeficiency. Recently, Jouanguy and coworkers (9, 10) and Newport and coworkers (11) have shown that mutation in the interferon-gamma receptor gene is associated with susceptibility to nontuberculous mycobacterial infection. Furthermore, impairment of mycobacterial immunity has been reported in human interleukin-12 receptor deficiency (12, 13). Major histocompatibility complex (MHC) molecules are known to play an important role in host-defense mechanisms. In this study, we investigated the human leukocyte-associated antigen (HLA) phenotypes since MHC molecules might play a role in the susceptibility to pulmonary MAC infection.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

HLA phenotypes were tested in 59 consecutive patients with pulmonary MAC infection (22 males, 37 females; 67.9 yr, range 40-88 yr) who were either treated or observed at Municipal Ida Hospital, City of Kawasaki from January 1996 to December 1997. These patients were from all over Japan. This study was approved by the committee of the hospital, and informed consent was obtained from all patients. Data from a healthy Japanese population (767 subjects), which were selected randomly and nationwide and reported at the Tenth Japan HLA Workshop (14), were used as a control. Therefore, we thought there would be no genetic difference between the MAC and the control groups.

Diagnosis of MAC infection was determined according to the criteria of the American Thoracic Society (4). In all patients, sputum or bronchial wash specimens were used to perform acid-fast bacilli (AFB) smear, culture, and polymerase chain reaction (PCR) detection of anti-acidic bacilli DNA (15). When Mycobacterium tuberculosis had been previously detected in sputum culture, history of tuberculosis (TB) in the MAC patient was regarded as positive. Main pulmonary lesions of patients with MAC were categorized into two groups (middle lobe and/or lingula [M/L] and apex groups) on the basis of blinded assessment of chest radiographic and computed tomographic (CT) scan findings as judged by three pulmonologists. Seven patients with MAC died of respiratory failure caused by lung destruction by MAC and concomitant bacterial infections. All pulmonary MAC patients had normal CD4-positive lymphocyte counts (more than 25% of total lymphocytes or more than 800/mm³) as estimated by fluorescent-activated cell sorter (FACS) analysis. HIV antibodies were not examined.

HLA phenotypes on T- and B-lymphocytes were estimated by the modified Terasaki-National Institutes of Health (NIH)-Standard method (16, 17). HLA-A, B, and DR loci were determined and the frequencies were compared between MAC and control groups.

Odds ratios (OR) were calculated according to Woolf's formula, and p values were defined by chi-square analysis. OR were indicated only when the p values are < 0.05 after correction for the number of antigens tested for each locus (HLA-A: 17; HLA-B: 38; HLA-DR: 11). A level of corrected p value (pc) < 0.05 was accepted as statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

HLA-A33 and HLA-DR6 antigen frequencies in patients with MAC were significantly increased compared with those in the control population (Table 1). Regarding haplotype, the frequency of A33-B44-DR6 in the pulmonary MAC patients was significantly increased compared with those of the control population (Table 1). Twenty-nine of 59 patients with MAC had TB histories. There were no significant differences in frequencies of either A33, DR6 antigen or haplotype A33-B44-DR6 between the TB history (+) and () groups (Table 2). In addition, there were no statistically significant differences in the above HLA frequencies between M/L and apex groups although the frequencies of HLA-A33 antigen and haplotype A33-B44-DR6 tended to be higher in patients with apex lesions (Table 2). There were no significant differences in any HLA phenotypes between patients who died from MAC progression and patients who are alive (Table 3). However, A33 antigen was not observed in patients who died.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our study demonstrated that frequencies of HLA-A33, DR6 antigens and haplotype A33-B44-DR6 in pulmonary MAC patients were significantly increased compared with those in the control population. This is the first report showing a significant linkage between HLA, whose genes are located on the short arm of chromosome 6, and pulmonary MAC infection. Our findings suggest that the development of pulmonary MAC infection is at least partly controlled by genetic factors.

In our MAC population, approximately 50% of the patients had TB histories, which is consistent with the notion that pulmonary lesions caused by TB may predispose the patient to subsequent MAC infection (18). However, there was no difference in the frequencies of HLA-A33, DR6 antigens and haplotype A33-B44-DR6 between MAC patients with and without TB histories. These observations imply that persons with these HLA antigens are at high risk for pulmonary MAC infection regardless of their TB history. Although the number of patients is relatively small in this study, we think that these HLA antigens could be useful markers for evaluating the risk of pulmonary MAC infection irrespective of TB history in a Japanese population.

Although bacterial infection is believed to be mainly affected by environmental factors and host immune conditions (19), several reports have shown that resistance to atypical mycobacteria and intracellular pathogens are controlled by a dominant autosomal gene, Bcg, in mice (20). The gene is thought to be identical to the Ity and Lsh loci, which control resistance to salmonella (24) and are located on the proximal part of mouse chromosome 1. Levin and coworkers (25) reported six children with familial disseminated atypical mycobacterial infection, suggesting that these patients are phenotypically similar to Lsh/Ity/Bcg susceptible mice. The mutation in the interferon-gamma receptor gene and interleukin-12 receptor deficiency have been recently reported to cause susceptibility to nontuberculous mycobacterial infection (9). These observations and our findings suggest that susceptibility to MAC infection is in part associated with genetic factors.

In this study, we found significant association between MAC infection and two HLA antigens and one haplotype. The relationship between either tuberculosis or leprosy and HLA has been reported previously (26). We speculated two mechanisms of the associations. First, the HLA-DR6 and/or A33 molecules may have a lower affinity than other HLA molecules for restricted MAC peptide, and thereby, T-lymphocyte recognition may be difficult to induce in the infected host. Second, a susceptible gene for MAC infection may be located near the MHC region. Further cellular immunity analyses are needed to elucidate the mechanisms that diminish the protective response to MAC infection.

As stated previously, both HLA-A33 and HLA-DR6 in pulmonary MAC patients were significantly increased compared with those of a control Japanese population. This may be related to the fact that A33-DR6 is a common haplotype in the Japanese population (14). However, we did not divide HLA-DR6 into HLA-DR13 and HLA-DR14 in this study because of the well-known discrepancy between serological and DNA methods in HLA typing. In our next study, DNA typing to identify not only the DR locus but also the DQ locus will be performed because Goldfeld and coworkers (29) have reported association of an HLA-DQ allele with clinical tuberculosis.

As shown in Tables 2 and 3, there were no differences in the frequencies of DR6 either between patients with M/L lesions and those with apex lesions, or between patients who died from MAC and those who are alive. In contrast, the frequency of A33 tended to be lower in patients with M/L lesion and in patients who died from progressive MAC. Although the progression of MAC infection can be affected by multiple clinical factors including immune status, nutrition, and pulmonary function, there is a group of patients with subacutely progressive course despite their normal immune responses (3). In terms of chest radiographic findings, pulmonary MAC patients can be categorized into at least two groups: fibrocavitary lesions at apex and interstitial nodular lesions at middle lobe or lingula (3). These observations indicate that further investigation will be necessary to elucidate the possible association of HLA with these clinical subtypes of MAC patients.

In summary, we suggest that the development of pulmonary MAC infection is associated with either HLA-DR6 and/or A33 antigen itself or possibly with a susceptibility gene located near the MHC region in a Japanese population.