INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The incidence of Mycobacterium avium-intracellulare (MAI) pulmonary disease in human immunodeficiency virus (HIV)- negative patients has been increasing over the last 20 years (1). Although this rising incidence is primarily due to an aging patient population and the increased use of immunosuppressive therapy, a substantial proportion of patients, accounting for 24-59% of non-HIV-related cases, have no predisposing risk factors (2, 3). These patients are predominantly caucasian, older, nonsmoking women with no preexisting lung disease and no known immune deficiency. MAI pulmonary disease in this group has proven difficult to treat, resulting in significant morbidity and even mortality (2).

It is unclear why these previously healthy patients become infected by an opportunistic pathogen. A number of hypotheses have been proposed in the literature to explain the mechanism of MAI pulmonary disease in this patient population. These theories have focused on increased host susceptibility due to altered anatomical defense mechanisms, such as an unidentified systemic connective tissue disorder (4), or voluntary cough suppression causing bronchiectasis in dependent lobes, predisposing to MAI colonization (5). However, disease susceptibility could also be explained by an immune deficiencey caused by alterations on a genetic rather than anatomic level. The resulting subtle immune defect would be more consistent with the unexplained observations in this populationa preponderance of Caucasian females, disease occurring later in life, and MAI as the sole pathogen.

In support of the concept that a previously unrecognized host immunodeficiency could account for susceptibility to MAI pulmonary disease, two mutations in the interferon  receptor 1 (IFN-R1) gene have been reported in patients with atypical mycobacterial disease (6, 7). The first, found in four children with disseminated disease, was a point mutation at nucleotide 395. In addition, a deletion at nucleotide 131 was found in a fifth child who developed disseminated disease after a Mycobacterium bovis BCG vaccine. Both produced premature stop codons, resulting in a truncated IFN-R1 protein.

In mice, resistance to intracellular pathogens such as atypical mycobacteria, Salmonella, and Leishmania has been shown to be controlled by the natural resistance-associated macrophage protein (nramp) gene (8). nramp encodes a putative transmembrane protein that mediates macrophage response to infection in the early stages. A single point mutation in the nramp gene, which results in a glycine-to-aspartic acid substitution (G105D), was shown to correlate with overwhelming M. bovis infection in mice (9). NRAMP1, the human genomic counterpart, has been sequenced and shares significant homology with murine nramp (10). In fact, the region of human NRAMP1 corresponding to that in nramp containing the murine susceptibility mutation is completely conserved between the two species (10). In addition, several polymorphisms have been identified in human NRAMP1, two of which predict amino acid changes in the cytoplasmic tail of the putative protein (11), and a third that is found in the promoter region of the gene (12). Extrapolating from the key role nramp plays in regulating murine resistance to infection by intracellular organisms, and the high degree of homology between mouse nramp and human NRAMP1, investigators have searched small populations of patients with Mycobacterium tuberculosis (11), leishmaniasis (12), and familial disseminated atypical mycobacteria (13) for the existence of a disease-associated NRAMP1 polymorphism, without success.

We felt that an alteration in either a single or multiple gene(s) would explain the apparent increased immune susceptibility to nontuberculous mycobacteria in our patient population. We further hypothesized that MAI pulmonary disease in previously healthy women might be explained by a defect in either their NRAMP1 or IFN-R1 gene, resulting in increased susceptibility to disease. The human NRAMP1 gene was analyzed in these patients with MAI pulmonary disease, in an attempt to identify a mutation corresponding to that found in susceptible mice, as well as to identify possible disease-associated NRAMP1 polymorphisms. The patients were also screened for the two known IFN-R1 mutations that have been already associated with human nontuberculous mycobacterial disease.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Recruitment

Eligible patients were identified from the Stanford University Hospital pulmonary clinic (Stanford, CA) by chart review. The diagnosis of MAI pulmonary disease was established according to the criteria published by the American Thoracic Society for diagnosis of disease caused by nontuberculous mycobacteria (14). Inclusion criteria were symptomatic disease with typical radiological findings (bronchiectasis or multiple small nodules) as well as chart documentation of either (1) multiple positive sputum cultures, (2) positive bronchoalveolar lavage culture, or (3) positive transbronchial biopsy culture. Exclusion criteria were a history of immunosuppressive therapy or an underlying immunocompromised state. Patients were contacted during a regular scheduled clinic visit or by telephone. Controls were four volunteers in our laboratory, one female, none of whom had pulmonary disease. The study was approved by the Administration Panel on Human Subjects in Medical Research and written consent was obtained from all patients and controls. Genomic DNA was extracted from 15 ml of whole blood using a blood and cell culture DNA kit from Qiagen (Chatsworth, CA). Genomic DNA that had been previously extracted (P.J.O., Genetics Department, Stanford University, Stanford, CA) from 18 additional controls (17 female, all caucasian, without any known pulmonary diseases) was used for additional analysis of the polymorphism in the NRAMP1 promoter region.

Polymerase Chain Reaction Amplifications

The region of NRAMP1 corresponding to the region in murine nramp containing the G105D mutation was amplified by polymerase chain reaction (PCR) using sense (5' GTA GGA GTT AGA GAC CCC TGG ACC AGG CTG G 3') and antisense (5' CGT AGT TAT CGA GGA AGA GGA AGA AGA AGG TG 3') primers. Each 50-µL reaction included 200 ng of genomic DNA, 500 ng of each primer, and a 200 µM final concentration of each dNTP. Two units of Taq DNA polymerase were added after initial denaturization at 94° C for 10 min, followed by 35 cycles at 94° C for 40 s, 61° C for 40 s, 72° C for 40 s, and a final extension at 72° C for 10 min. The PCR products were fractionated on 1.5% agarose gels in 1× TAE (0.04 M Tris-acetate, 0.001 M EDTA [pH 8.0]) buffer and visualized by ethidium bromide staining. The NRAMP1 promoter region was amplified using sense (5' CCA ACG AGG GGT CTT GGA ACT CCA 3') and antisense (5' GTC AAT ACC CCA TGA CCA CAC CCC 3') primers. The 3' region of NRAMP1 containing two other polymorphisms was amplified using sense (5' GCA TCT CCC CAA TTC ATG GT 3') and antisense (5' AAC TGT CCC ACT CTA TCC TG 3') primers and an annealing temperature of 59° C. The IFN-R1 region containing nucleotide 395 was amplified using sense (5' GTT AAA GCC AGG GTT GGA CA 3') and antisense (5' CAT CTC GGC ATA CAG CAA ATT CTT GT 3') primers and an annealing temperature of 56° C. The IFN-R1 region containing nucleotide 131 was amplified using sense (5' GCC TAC ACC AAC TAA TGT TA 3') and antisense (5' ATA GTT CTT TAC CTC TAC GG 3') primers and an annealing temperature of 58° C.

Denaturing High-Performance Liquid Chromatography

Denaturing high-performance liquid chromatography (HPLC) is capable of resolving heteroduplex from homoduplex DNAs, thereby revealing the presence of point mutations arising from allelic variation (15). The PCR products containing the regions containing the mouse nramp mutation, the human NRAMP1 promoter polymorphism, and the IFN-R1 deletion at nucleotide 131 were analyzed using denaturing HPLC as previously described to determine the presence of allelic variations.

Restriction Fragment Length Polymorphism Analyses

The 3' region of NRAMP1 was analyzed for the presence of polymorphisms using restriction fragment length polymorphism (RFLP) and specific restriction enzymes AvaII and FokI. Restriction enzyme Tsp45I was used to search for the presence of mutation at nucleotide 395 in IFN-R1. Ten nanograms of PCR products was digested in each reaction as recommended by the supplier (New England BioLabs, Beverly, MA). Restriction digests were fractionated on 4% NuSieve (FMC BioProducts, Rockland, ME) agarose gels in 1× TAE and visualized by ethidium bromide staining.

Microsatellite Sizing

Polymorphisms identified by denaturing HPLC in the NRAMP1 promoter region were further characterized by microsatellite sizing. The PCR products were amplified using primers of identical sequence, except that the sense primer was end labeled with fluorescein, and AmpliTaq Gold polymerase (Perkin-Elmer, Foster City, CA) was substituted to minimize slippage caused by multiple base repeat sequences during amplification. The product was then analyzed for exact nucleotide product size by denaturing acrylamide gel electrophoresis.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Clinical Characteristics

The first eight patients meeting eligibility criteria who were able to come to clinic for venipuncture were enrolled. Their characteristics are summarized in Table 1. Of the eight patients, seven were female, and seven were caucasian. The mean age at time of enrollment was 66.8 yr (range, 52-77 yr). The majority of these patients had no known significant prior pulmonary disease. The most common presenting symptom was chronic productive cough. Six of the eight patients had bronchiectasis, and five had small nodular disease defined radiographically on chest radiograph and/or chest computed tomography. Diagnostic testing had to proceed to brochoscopy in four of the patients; the other four had sputums persistently positive for MAI.

Human NRAMP1 Genomic DNA Analyses

We identified four independent alleles, located in three separate regions of human NRAMP1, which were felt to be potential candidates to modulate NRAMP1 expression and/or function. To explore possible mutational or allelic variation at these sites, genomic DNA was isolated from patients and controls and used for PCR amplification of these three regions.

The first NRAMP1 region investigated correlated to the segment in murine nramp containing the previously identified susceptibility mutation (9). PCR primers were designed to amplify a 128-bp product that would include the human sequence homologous to the mouse mutation site. These PCR products were analyzed by denaturing HPLC, comparing patients among themselves as well as against controls. The 128-bp NRAMP1 PCR products from all eight patients and four controls were homogeneous, with no evidence of allelic variation. In addition, a subset of the 128-bp PCR products of the patients was sequenced directly, and was subcloned into pCRII vectors and sequenced (data not shown). All sequences were wild type. Thus, there was no evidence of any point mutation in the region of human NRAMP1 homologous to the mouse nramp G105D mutation.

The 3' region of NRAMP1, including exon 15, which encodes a portion of the cytoplasmic tail of the putative protein product, was the second region analyzed. A polymorphism in exon 15 has been shown to occur from a single base mutation resulting in an amino acid substitution (D543N) (11). A second polymorphism consisting of a deletion of four nucleotides is found in the 3' untranslated region (UTR), a region of the gene that may regulate translation (11). Both of these polymorphisms were felt to have a potential role in modulating NRAMP1. The 3' region of NRAMP1 was PCR amplified. A search for the polymorphisms was made by restriction enzyme digests with AvaII and FokI, respectively, as previously described (11). AvaII RFLP digest of the eight patients and four controls resulted in identical fragment lengths, thus ruling out differing alleles in exon 15. FokI RFLP digest of all 12 samples also revealed no difference between patients or controls, ruling out a different allele in the 3' UTR in our patients.

A polymorphism previously identified in the NRAMP1 promoter region was the third region investigated (12). PCR amplification of genomic DNA from eight patients and four controls produced a 108 to 122-bp product (depending on the polymorphism) containing this promoter region. Denaturing HPLC comparison of these 12 PCR samples suggested the existence of an allele in 3 of the patients that was not found in the other 5 patients or in any of the controls. As the polymorphism consisted of a varying number of (gt) repeats, the PCR reaction was modified to minimize any role of slippage during amplification. Reamplified and fluorescein-labeled PCR products were analyzed by microsatellite sizing, which corroborated the previous finding of a different allele in the same three patients. To determine if this difference was statistically significant, genomic DNA from an additional 18 controls was also PCR amplified using the modified technique and analyzed by microsatellite sizing. Ten of the 18 controls also revealed the existence of allelism. Thus, although a polymorphism is expressed in 3 of our 8 patients, it is also found in 10 of 22 controls who have no pulmonary disease, and thus did not correlate with disease susceptibility.

IFN-R1 Genomic DNA Analyses

Two published polymorphisms in IFN-R1 have been associated with disseminated atypical mycobacterial disease. We searched the genomic DNA of our patients for the presence of both of these mutations.

The first analysis focused on the polymorphism at IFN-R1 nucleotide 395, consisting of a point mutation of C to A. Primers had been previously designed to PCR amplify a 70-bp product that is cleaved by restriction enzyme Tsp45I if it is wild type, but that loses the restriction enzyme site if there is a mutation (6). Tsp45I digest of the PCR 70-bp products of all eight patients and four controls resulted in all 12 PCR products being cleaved, indicating the absence of a mutation at nucleotide 395.

The second IFN-R1 polymorphism consisted of a C deletion at nucleotide 131 (7). This region of interest was PCR amplified in all eight patients and four controls. The 12 PCR products were analyzed by denaturing HPLC. They showed complete homogeneity, ruling out the existence of a deletion at nucleotide 131.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This is the first investigation that seeks a genetic basis for an immune deficiency correlating to MAI pulmonary disease in elderly women.

Retrospective reviews of MAI pulmonary disease have identified a common patient phenotype of a thin, postmenopausal woman, often with anatomic abnormalities such as mitral valve prolapse or pectus excavatum. Our series at Stanford, which includes more than 20 such patients, substantiates these observations. Previous reviews have suggested that it is the phenotype itself that leads to disease susceptibility, such as in "Lady Windermere syndrome" (5). We proposed that this phenotype, rather than being the cause of disease, might instead be an indicator of a specific genetic immunodeficiency that is independent of the phenotype. This explanation would address several observations regarding this patient population that cannot be resolved by prior theories. First, these women appeared entirely immunocompetent until their presentation with MAI pulmonary disease, and even then, had no apparent increased susceptibility to other infectious agents. Second, they did not acquire disease until later in life. Furthermore, although they mount a granulomatous reaction against the infection, the response is insufficient to eradicate the opportunistic pathogen. Thus, morbidity is a consequence of the indolent granulomatous inflammation leading to bronchiectasis.

It has been shown that in mice, susceptibility to Mycobacterium bovis is controlled by an autosomal recessive gene. The nramp protein product is predicted to be a membrane-bound transporter whose pleiotropic effects control the early phase of resistance to mycobacterial growth in macrophages (16). The expression of human NRAMP1 mRNA is highest in macrophages as well as in lungs and spleen, two immune effector organs (10).

In humans, a genetic basis for increased susceptibility to atypical mycobacteria has been suggested by a report of familial childhood disease (17). In six children who developed disseminated atypical mycobacterial disease, three of the four from Malta were related and the remaining two Greek Cypriots were siblings. It was demonstrated that tumor necrosis factor  production in response to stimulation by endotoxin or IFN- was diminished in the children as well as in their parents to a lesser degree. In addition, the production of IFN- in response to mycobacterial antigens and recall antigens was decreased in both children and parents. As in our pulmonary population, the immune deficiency was specific for atypical mycobacteria; over years of follow-up these children showed no inadequate response to other pathogens except for Salmonella. It was originally felt that these children represented the human equivalent to the mice that were homozygous for the nramp susceptibility mutation. However a subsequent study has ruled out any NRAMP1 mutation in these patients (13).

Further investigations revealed a mutation in the IFN-R1 gene in the four Maltese children resulting in the absence of the receptor on peripheral blood mononuclear cells (6). A second, independent mutation in the same gene has also been reported in a Tunisian child with disseminated disease following BCG vaccination (7). These children, although in many ways different from the population of elderly, female patients with localized disease, add plausibility to our theory that susceptibility to MAI pulmonary disease may be based on the genomic level, whether due to changes in a single gene or a consequence of interactions among multiple genes.

Given the key role nramp plays in conferring resistance to atypical mycobacteria in mice, we first focused our search on an alteration in human NRAMP1 to explain MAI pulmonary disease. However, we found no mutation in human NRAMP1 correlating to the susceptibility mutation in murine nramp. In addition, there was no correlation between disease and any of the three of the four published NRAMP1 polymorphisms. The fourth polymorphism encodes an alanine-to-valine substitution in exon 9, which was felt not to be a significant amino acid change and was thus not investigated (11). Therefore, we found no evidence for allelism in NRAMP1 to correlate with MAI pulmonary disease in older, HIV-negative women.

Although a mutation in the NRAMP1 gene is probably not the cause of disease in these patients, disease susceptibility may still be controlled by this gene. Although the putative NRAMP1 protein may be qualitatively normal, its quantitative expression may be regulated in some fashion and thus diminish the host immune response. This modulation may occur as a result of the action of another gene(s), or possibly as a result of the changing hormonal milieu in these elderly, generally postmenopausal, patients. However, further investigations of NRAMP1 beyond gene expression and regulation must await isolation of the putative transmembrane protein.

The IFN-R1 gene appears to be a more promising candidate for further studies to explain MAI pulmonary disease susceptibility. In addition to the mutations previously discussed correlating to human familial disease, there have been reports of clinical improvement following IFN- treatment of refractory disseminated atypical mycobacterial disease (18). Although our study did not find evidence of mutation in this gene to correlate with disease, this was by no means an exhaustive analysis of the IFN-R1 gene. The fact that two separate mutations have already been found to exist and correlate with disseminated disease leads to the intriguing possibility of other, perhaps less severe, polymorphisms that may lead to more localized disease. To our knowledge, a complete search for polymorphisms has not been done for the human IFN-R1 gene. Once these polymorphisms have been identified, they should be searched for in these elderly women with MAI pulmonary disease. Meanwhile, a more complete analysis of IFN-R1 expression should be investigated in this patient population.

Although we did not find a genetic mutation in NRAMP1 or IFN-R1 to correlate with MAI pulmonary disease in otherwise healthy women, we still believe that disease susceptibility is a consequence of a subtle, as yet to be identified immune deficiency. This theory is supported by the fact that an autosomal recessive gene controls disease in a mouse model, and that mutations in the IFN-R1 gene correlate with disease in children. Future investigations should focus on elucidating the etiology of disease in this population, as this would potentially lead to new, and possibly superior, treatment options.