Introduction

Since October 2001, when Bacillus anthracis was released in the United States as an act of bioterrorism, there has been a greater interest in determining if there are risk factors for inhalation anthrax infection. Exposure to Bacillus anthracis spores does not cause infection in all exposed individuals [1]. Epidemiologic studies of individuals infected by inhalation anthrax have suggested that a weakened immune system might increase susceptibility to infection by Bacillus anthracis [2]. Some of the infected individuals had a history of chronic pulmonary disease, including asthma, sarcoidosis, and tuberculosis [2–4]. Studies in mice have demonstrated a genetic basis for anthrax sensitivity [5, 6]. For example, macrophages from C3H mice are 100,000 times more sensitive to the Bacillus anthracis toxin than macrophages from A/J mice [6]. The current study examines whether there are genetic polymorphisms in humans associated with increased susceptibility to lung disease. Identification of genes associated with an increased risk of lung disease might identify individuals who might also be of increased susceptibility to inhalation anthrax infection.

The N AD(P)H ox idases (NOX) are a family of enzymes that are essential in host defense against microbial infection, as reviewed by Quinn and Gauss [7]. The central enzyme of the NAD(P)H oxidase is a flavin and heme-containing protein, the most well known being the phagocytic gp91phox (CYBB, NOX2) protein. gp91phox, and a number of related proteins including DUOX1 and DUOX2, are transmembrane proteins which transport electrons and generate reactive oxygen species (ROS) at the expense of NADH or NADPH. The activity of the oxidases are highly regulated by accessory proteins, including p22phox (CYBA), p47phox (NOXO1, NCF1), p67phox (NOXA2, NCF2), and p40phox (NCF4). Chronic Granulomatous Disease (CGD), associated with severe, recurrent, and chronic non-specific bacterial and fungal infections, is most commonly caused by mutations in p47phox, gp91phox, p67phox, and p22phox that severely compromise the respiratory burst activity of neutrophils.

Görlach et al were the first to identify the presence of at least one pseudogene copy of the p47phox (NCF1) gene on chromosome 7q11.23 [8]. By construction of a detailed physical map of this region Hockenhull et al determined that there were one normal wildtype copy and two pseudogene copies of NCF1 per chromosome [9]. Heyworth et al elegantly demonstrated that in some individuals, one of the pseudogene copies of NCF1, possibly by recombination or gene conversion, has reverted to the normal wildtype GTGT sequence (i.e. pseudowildtype) [10]. Thus, individuals with this low frequency polymorphism of NCF1, have 2 "wildtype" copies and one pseudogene copy per chromosome [10]. Therefore, individuals (with 2 chromosomes) can have a NCF1 pseudogene: wt copy ratio of either 2:1, 1:1 or 1:2. Although two groups have examined the association of the minor 1:1 and 1:2 alleles with inflammatory bowel disease, the conclusions were in conflict primarily due to differences in allele frequencies of the control population and sample size [11, 12]. Other polymorphisms in p47phox, p67phox and gp91phox, have not been shown to be associated with human disease other than CGD. Recently p47phox has been shown by positional cloning to regulate the severity of arthritis in rats [13]. The H72Y polymorphisms in p22phox (CYBA), associated with reduced respiratory burst in isolated human neutrophils [14], but has yet to be shown to be clearly associated with a disease phenotype [15–17]. DUOX1 and DUOX2, which are expressed in lung epithelium, regulates H2O2 [18–20] and acid [21] production in the airway but have not been shown to be associated with lung disease. Mutations in DUOX2 have been shown to be associated with mild hypothyroidism [22–24].

TLR2 is the receptor for peptidoglycans, lipoteichoic acid, lipoarabinomannan, mycolylarabinogalactan, and zymosan. Anthrax infection is thought to be partially mediated through the TLR2 pathway since TLR2 deficient mice are resistant to infection by the Sterne strain of Bacillus anthracis and HEK293 cells expressing TLR2, but not TLR4, are able to signal in response to exposure to heat-inactivated Bacillus anthracis [25]. Inactivation and killing of the tuberculosis mycobacterium is also mediated through TLR2 since macrophages from Tlr2-deficient mice or human macrophages blocked by anti-TLR2 antibodies failed to kill the bacteria [26]. Tlr9 and Tlr2 double knockout mice display a more pronounced susceptibility to infection by tuberculosis than single gene knockout mice [27]. The TLR2 polymorphism R753Q [28] and the R677W polymorphism in humans [29–31] have been shown to be associated with increase risk for tuberculosis infection. The R753Q polymorphism was not associated with a generalized increased risk of infection, e.g. individuals with R753Q were less responsive to infection by Borrelia burgdorferi, which causes Lyme Disease [32] and R753Q was not associated with increased susceptibility to Staphylococcus aureus infection [33].

Alpha-1-anti-trypsin (AAT) deficiency has been associated with increased susceptibility to lung disease, particularly emphysema [34, 35]. Although more than 70 variants have been described, only a few are associated with reduced AAT protein expression and/or reduced activity [35]. Several studies have suggested that simple heterozygosity for mutant alleles of AAT may predispose individuals to chronic obstructive lung disease [35–37]. The Z allele (E366K), which occurs at an allele frequency of 0.01–0.02 in people of European origin, is the most common allele associated with an increased risk of environmentally induced emphysema [34, 38–40]. Homozygous individuals of the AAT S allele (E288V) are not at risk for emphysema but compound heterozygotes of the Z and S allele or a null allele are of increased risk [39, 41]. Carriers of the AAT S and Z alleles are over-represented in individuals with lung cancer [42]

In this study, we attempted to determine whether normal nonsynonymous genetic variations identified by the Genbank SNP database or previously described in the literature to be present in the normal population in the genes for p47phox (NCF1), p67phox (NCF2), p40phox (NCF4), gp91phox (CYBB), p22phox (CYBA), DUOX1, DUOX2, TLR2, TLR9 and alpha-1 anti-trypsin (AAT) are associated with an increased susceptibility to tuberculosis, sarcoidosis, recurrent pneumonia, and atypical mycobacterial infection.

Materials and methods

Study participants

Anonymized blood samples from control individuals of European, non-Hispanic origin (n = 95) were obtained from Kaiser Permanente [43] or from The Scripps Research Institute GCRC blood drawing program. From a group of 31,247 participants in a Kaiser Permanente study of European, non-Hispanic origin [43], all individuals that had a documented medical history with hospitalization for lung diseases: atypical mycobacterial infection (n = 1), repeated episodes of pneumonia (n = 5), sarcoidosis (n = 46), and tuberculosis (n = 43), were selected and will be referred to as the lung disease group (n = 95). The participants in the Kaiser Permanente study were members of Kaiser Permanente attending a Health Appraisal Clinic and were not selected for underlying acute or chronic disease. All human samples were obtained with written consent. Approvals for the protocols involving the use of human individuals were obtained from the institutional review boards of The Scripps Research Institute and Kaiser Permanente.

p47phox/NCF1 pseudogene: wildtype ratio

Amplification of the region of p47phox exon 2 with the wildtype GTGT sequence and the pseudogene delGT sequence were amplified using primers p47phox/NCF1 Ex2F GCTTCCTCCAGTGGGTAGTGGGATC and p47phox/NCF 161R GGAACTCGTAGATCTCGGTGAAGC and 32P-labeled p47phox/NCF1 Ex2F primer under standard PCR conditions for 25 cycles. The 32P-labeled amplified DNA products were separated on a 10% acrylamide/urea/TBE sequencing gel. Autoradiography was used to visualize the wildtype and pseudogene amplified products, which differ by 2 nucleotides in length.

Genotyping of single nucleotide polymorphisms (SNPs) by allele specific oligomer hybridization (ASOH)

For the genes of this study, non-synonymous SNPs identified in Genbank's SNP database and/or non-synonynous SNPs associated with lung disease were investigated. Amplification of DNA regions encompassing the SNPs were amplified using the primers listed in Table 1. ASOH was performed using standard hybridization conditions [44] using 32P radiolabeled probes and washing temperatures described in Table 1. Genotyping was determined following visualization of the hybridized probe by autoradiography.

Table 1 Primer List. List of primers used for DNA amplification and ASOH.
Full size table

Statistics

The Fisher's Exact test was performed with GraphPad InStat using the raw data entered into a 2 × 2 contingency table. Power calculations were performed to give the probability of finding the differences between the gene frequencies as statistically significant, given the sample size.

Results

We examined 95 individuals of European, non-Hispanic origin with documented medical history with hospitalization for lung disease (46 with sarcoidosis, 43 with tuberculosis, five with recurrent pneumonia, and one with atypical mycobacterial infection) and 95 control individuals of European, non-Hispanic origin for differences in allele frequencies in genes involved in innate immunity.

P47phox/(NCF1)

Examination of the pseudogene: wt copy ratio of control versus lung disease individuals demonstrated no statistically significant difference in the frequencies of the pseudogene: wt ratios in the lung disease group as compared to the control group (Table 2).

Table 2 Pseudogene versus gene ratio. p47phox/NCF1 pseudogene: wt gene ratio in lung disease and control individuals. The data are presented as number of individuals with the indicated pseudogene:wt ratio and the number within parentheses indicates the calculated frequency.
Full size table

p67phox (NCF2), p40phox (NCF4), p22phox (CYBA), gp91phox (CYBB), DUOX1, DUOX2

SNPs in the p67phox (NCF2), p40phox (NCF4), p22phox (CYBA) and gp91phox (CYBB), DUOX1 and DUOX2 genes were examined. Some SNPs did not occur at a high enough frequency to be detected in our samples. None of the allele frequencies differed significantly between the lung disease and the control groups (Table 3).

Table 3 Summary of SNP Analyses. SNP analyses of candidate genes in lung disease versus control groups. Numbering of SNPs start from the ATG initiator methionine of the cDNA. Data are presented as number of alleles identified divided by total number of alleles examined. Numbers within parentheses are the calculated allele frequencies. Power calculations were performed using number of subjects.
Full size table

TLR2, TLR9, AAT

TLR2, TLR9, and AAT genes were examined. Again, many SNPs did not occur at high enough frequency to be observed. Most of the allele frequencies did not differ between the lung disease and control groups. The TLR2 polymorphism R753Q, associated with tuberculosis, was not shown to be different between the control or lung disease group. The TLR2 R677W polymorphism, also associated with tuberculosis, was not observed in either group. The R863Q SNP in TLR9 was absent from the lung disease group indicating that this polymorphism was not associated with increased lung disease. The AAT S (Glu288Val) and Z (E366K) alleles, associated with chronic obstructive lung disease, were examined and there was no difference in allele frequencies between the control and lung disease groups (Table 3).

Discussion

Since only a subset of individuals exposed to Bacillus anthracis spores develop pulmonary disease, the most life-threatening form of anthrax infection, it would be important to identify factors that lead to susceptibility to this type of infection. This might make it possible to identify those individuals who are at greatest risk and to provide them with the most aggressive treatment at the outset of infection. The ability to thus triage individuals in the case of a bioterrorism attack would be valuable. Moreover, understanding genetic susceptibility could lead to better management of individuals with pulmonary anthrax infection.

The genetic influences on resistance to infection are very strong. Indeed, genetic influences on resistance to infection appear to be greater than genetic influences on cancer or cardiovascular disease [45]. In the past few decades a considerable number of polymorphisms have been shown to cause infectious disease susceptibility in mice [6] and in humans [28, 31, 46]. Because infections caused by Bacillus anthracis are rare it was impossible to examine candidate polymorphisms in patients who actually developed pulmonary anthrax. Instead, it was necessary to use surrogate infections such as unusual mycobacterial infections, recurrent pneumonia, and tuberculosis or examine lung diseases such as sarcoidosis, which has been reported in cases of inhalation anthrax, for this study. The "lung disease group" in this study represented all the individuals with documented hospitalization for lung disease from a collection of 31,247 individuals of European, non-Hispanic origin unselected for any particular acute or chronic health problem. Candidate genes were chosen on the basis of their role in immunity against chronic infection as well as their participation in the innate immune response. This is a reasonable approach, since defects in the immune system generally increases susceptibility not to a single organism, but rather to multiple organisms that share some features in the pathogenesis of the disease that they produce.

Our analyses of genes of the NAD(P)H oxidase, p47 (NCF1), p67phox (NCF2), p40phox (NCF4), p22phox (CYBA), and gp91phox (CYBB), as well as other genes involved in innate immunity such as DUOX1 and 2, TLR2, TLR9 and AAT demonstrated that there were no differences between the control and lung disease group comprised of primarily sarcoidosis and tuberculosis individuals. There may, of course, be many other polymorphisms that affect susceptibility to Bacillus anthracis. Although the genes that we chose seemed to be reasonable candidates; there are many additional genes encoding products that could be important in effecting the course of anthrax in humans. For example, it has been suggested that susceptibility to Bacillus anthracis might involve myD88 [25]. Furthermore, susceptibility to infection by tuberculosis may be altered by variations in the vitamin D receptor gene [47]. Similarly, sarcoidosis has been shown to be associated with particular alleles in BTNL2 [48, 49], IL18 [50], and IFNa [51], and SLC11A1 [52].