Circulating granulysin levels in healthcare workers and latent tuberculosis infection estimated using interferon-gamma release assays

According to global estimation, latent Mycobacterium tuberculosis (MTB) infection is present in approximately one third of the human population (~2 billion people) [1]. Latent infection follows prolonged survival of MTB despite containment by host defense mechanisms [2] and contributes significantly to the risk of overt disease following reactivation of MTB under conditions of compromised immunity [3], which presents a major obstacle to achieving global tuberculosis (TB) control.

Interferon-gamma release assays (IGRA) are diagnostic of TB infection and operate on the principle that MTB-specific antigens provoke immune responses in whole blood after the establishment of infection [4]. IGRA offer more specific detection of latent TB infection (LTBI) than tuberculin skin test in many circumstances [5]. Accordingly, QuantiFERON-TB Gold In-tube enzyme-linked immunosorbent assay (ELISA)-based IGRA are recommended in multiple current guidelines, including those of the United States Center for Disease Control and Prevention for testing LTBI among people who are at risk of TB infection (i.e., healthcare workers) [6].

As an effector molecule, the saposin family protein granulysin (GNLY) is involved in protective immunity, and is released from natural killer (NK) cells, NKT cells, γδ T cells, and cytotoxic T lymphocytes (CTLs). This molecule directly eliminates extracellular MTB and intracellular bacteria in the presence of perforin by causing osmotic shock and inducing apoptosis [7]. Recently Walch et al. [8] suggested that GNLY initially penetrates infected cells and delivers bactericidal granzymes to intracytoplasmic bacteria. Hence, in the presence of GNLY, granzymes eliminate bacteria independently of host cell death.

Precursor and mature GNLY proteins of 15 and 9 kDa, respectively, have been identified, and mature GNLY is found with perforin and granzymes in cytotoxic granules. Moreover, a previous report has shown that the 9-kDa mature protein is mainly secreted from activated CTLs, whereas the 15-kDa precursor protein is found in plasma under physiological conditions [9]. Although the precursor protein may have more potent chemotactic and inflammatory activities, its physiological function remains poorly understood [10].

Circulating GNLY levels may be a useful indicator of overall host cellular immunity [11]. Moreover, GNLY expression has been identified as a marker for outcomes in cancer and organ transplantation patients and is an important mediator of damage in skin diseases [10]. In TB, plasma GNLY levels were significantly lower in patients with active disease than in healthy individuals and increased after successful anti-TB treatment [12]. In an animal study, GNLY expression was associated with protection after vaccination against TB in cattle [13]. However, it is not clear whether GNLY levels are associated with human LTBI. In the present study, we investigated the relationship between serum GNLY concentrations and LTBI status as diagnosed using IGRA, determined GNLY mRNA expression in blood cells, and analyzed GNLY genetic polymorphisms.

In the present study, we investigated GNLY in LTBI as identified using IGRA among healthcare workers in Hanoi, Vietnam. Serum GNLY concentrations in the IGRA-positive group were significantly lower than those in the IGRA-negative group. Moreover, serum GNLY concentrations were significantly positively correlated with GNLY mRNA expression levels in blood and were dependent on the number of C alleles of the rs11127 SNP in exon 4 of GNLY, whereas GNLY mRNA expression levels were not dependent on the number of C alleles of the rs11127 SNP in exon 4 of GNLY. Of the three parameters (SNP genotype, mRNA expression, and concentrations of GNLY in blood), only circulating levels of GNLY protein were significantly negatively associated with IGRA-positive status.

GNLY, a major antimicrobial molecule, has been shown to preferentially bind and disrupt cholesterol-poor membranes of microbes [10]. However, no homolog has been found in rodents, and investigations using human samples are required to elucidate the significance of GNLY in vivo, as shown in the present study. In a recent study [8], bactericidal mechanisms involving cooperation of GLNY and granzymes have been clarified. Specifically, GLNY delivers granzymes to intracellular bacteria leading to oxidative damage through cleavage of bacterial antioxidants. Thus, GNLY shows both cytolytic and bactericidal activities with other effector molecules such as granzymes and perforin, although GNLY acts independently at micro-molar concentrations and under hypotonic or acidic conditions [18, 19].

Since CTLs play crucial roles in the containment of MTB in TB granulomas [2, 19], GNLY appears to be an important effector molecule in human TB infection, alone and in cooperation with granzymes and perforin. In a previous study in which the same ELISA system was used [11], GNLY concentrations of 244 presumably TB-unexposed healthy Japanese individuals were 3.7 ± 3.2 ng/ml (mean ± standard deviation), and these values were even higher than those in IGRA-negative Vietnamese individuals presumably resistant to TB infection in our study. It suggests that TB exposure may decrease serum GNLY levels, and that LTBI state may further suppress the level by some unknown mechanism, though effects of ethnicity and other background such as comorbidity and nutritional status should be taken into consideration and it is difficult to draw any conclusions from direct comparison between different studies. Moreover, previous studies of active TB disease have shown that GNLY levels in the blood vary with clinical stage [12, 20]. Although serum GNLY concentrations were lower in the LTBI group than in the non-LTBI group, this GNLY concentration (median, interquartile range; 2.144, 1.703–2.824 ng/ml) was higher than that shown previously in newly diagnosed active TB (median ± standard error, 1.511 ± 0.287 ng/ml) and relapsed TB (1.458 ± 0.329 ng/ml) patients in our collaborative study using the same ELISA system [20]. Potential interpretations of low GNLY levels in TB infection are as follows. First, the present data may reflect sequestration of GNLY-producing cells: Lymphocytes with high GNLY expression accumulate at the sites of latent infection. At these sites, continuous cross-talk between the host immune system and the pathogen [21] may facilitate GNLY turnover, and accumulation of these GNLY-producing lymphocytes potentially results in a relative decrease in circulating GNLY in infected individuals. Moreover, it is well known that MTB-specific CTLs accumulate in areas surrounding infected epithelioid cells of TB granulomas [19, 22] and may also occur in individuals with LTBI. This condition may differ from that of chronic TB, which is characterized by persistent inflammation and active clinical manifestations, and is associated with diminished GNLY in granuloma-associated CTLs, resulting in impaired CTL activities [23]. Second, protein levels of GNLY were negatively associated with interferon gamma release in the present study, whereas GNLY expression levels and GNLY polymorphisms were not, suggesting that GNLY may lack stability in TB infection, and its rates of metabolism or degradation may be affected by chronic inflammation and immune reactions in TB. Finally, during the very early stages of TB infection, NK and NKT cells may destroy pathogens and infected cells [24], and as innate and intermediate immune cells, they express both GNLY and granzymes. Hence, limited expression of GNLY in these cells may lead to failure to eliminate MTB and it may cause IGRA-positive status. However, although frequencies of the C allele of the rs11127 GNLY polymorphism may lead to low serum GNLY concentrations, these inherent genotypes were not associated with IGRA-positive status, indicating that low GNLY concentrations may be a consequence rather than a cause of LTBI.

Sputum smear-positive cases of TB have long been targeted in TB control programs [25] and preventive therapy for LTBI in the general population has not been feasible in most countries with high TB burdens, including in Vietnam. However, identification of subgroups of individuals who are at high risk of developing active disease may provide a useful and cost-effective strategy to control TB [26]. Thus, future investigations are required to explore the roles of GNLY and other antimicrobial molecules as biomarkers, for example, in prospective cohort studies with follow up of blood levels during the first two years after infection.

Our study has some limitations. We could not recruit individuals with no TB exposure. We compared our results with those obtained from previous studies and discussed possible interpretations of our findings. Because previous studies of granzymes [8] and perforin [7] reveal that GNLY works in cooperation with these molecules to eliminate MTB, it may be of further value to investigate the co-localization of these effector molecules using flow cytometry or cell population analyzers to offer greater certainty of this assertion. Moreover, the 15-kDa precursor form of GNLY is better known as a potent chemotactic factor during inflammation than as a bactericidal molecule. Since the present sandwich ELISA system captures both forms of GNLY, further studies are required to determine which form of GNLY is predominant in blood from patients with LTBI and active TB. Lastly, the SNP rs11127 (C/T) of exon 4 leads to an amino acid substitution (Ile104Thr) in the GNLY protein and may accelerate consumption or enhance stability of GNLY. Since GNLY genotype was associated with GNLY concentration, but not with mRNA expression level, we can assume that GNLY polymorphism may have affected GNLY concentrations at the protein level; and GNLY may have been expressed in the way partly independent of GNLY genotype. A genetic variation that directly affects GNLY concentrations was not clearly determined in this study because of strong linkage disequilibrium around the GNLY locus. Nevertheless, the SNP rs11127 has been reported to have an important role in clearance of hepatitis B virus [27, 28], and further study on GNLY genetic polymorphism is necessary whether it may contribute to optimize granulysin vaccines against TB [29, 30].

Serum GNLY concentrations in IGRA-positive participants were significantly lower than in those with IGRA-negative status. Therefore, changes in circulating GNLY, presumably at the protein level, may be involved in the immunological condition of LTBI.