Heat shock protein 70 (hsp70) as a major target of the antibody response in patients with pulmonary cryptococcosis

Abstract

Cryptococcus neoformans causes infection in individuals with defective T cell function, such as AIDS, as well as without underlying disease. It has been suggested that humoral as well as cellular immunity might play an important role in the immune response to C. neoformans infection. We have recently shown, using immunoblotting, that the 70-kD hsp family of C. neoformans was the major target molecule of the humoral response in murine pulmonary cryptococcosis. In this study we also used immunoblotting to define the antibody responses in the sera of 24 patients with pulmonary cryptococcosis: 21 proven and three suspected diagnoses. Anti-C. neoformans hsp70 antibody was detected in 16 of 24 (66.7%) patients with pulmonary cryptococcosis. Fourteen of 17 (82.3%) patients with high antigen titres (≥ 1:8) and two of seven (28.6%) patients with low titres (≤ 1:4) had detectable levels of anti-hsp70 antibody. Sera from patients positive for anti-hsp70 antibody showed high titres in the Eiken latex agglutination test for the detection of serum cryptococcal antigen. Our results indicate that the 70-kD hsp family from C. neoformans appears to be a major target molecule of the humoral response, not only in murine pulmonary cryptococcosis, but also in human patients with pulmonary cryptococcosis.

INTRODUCTION

Hsp or stress proteins are synthesized by all cells in response to various types of environmental stress and probably function as molecular chaperones in normal physiological processes. In recent years it has been observed that members of the hsp family, including hsp70 from several pathogenic microbes, are antigenic. Moreover, hsp of an invading microbe are often immunodominant targets of cellular and humoral immune responses [1]. This is particularly true for members of the hsp70 family, which are among the most immunogenic proteins of pathogenic microorganisms [2]. For instance, hsp70s are thought to play an important role in infections caused by Plasmodium [3], Trypanosoma [4], Schistosoma [5], Leishmania [6], and Mycobacterium [7] species. In the field of fungal pathogens, a member of the hsp70 family of Histoplasma capsulatum was recognized as a target of cell-mediated, protective immunity [8] and the hsp70 family of Candida albicans appears to be expressed as both B cell and T cell immunogens in human subjects, with a possible important role in infections caused by C. albicans [9]. Further, the microbial hsp are the antigens of choice for vaccines in various infectious diseases. Several studies have suggested the potential usefulness of microbial hsp as a candidate for the design of subunit vaccines, e.g. hsp90 in candidiasis [10], hsp60 in histoplasmosis [11], and hsp70 in schistosomiasis [12].

The fungus Cryptococcus neoformans is primarily a pathogen for individuals with impaired cell-mediated immunity. Infection with C. neoformans is a major problem in HIV-infected individuals; in New York City, 6–8% of patients with AIDS develop cryptococcal meningitis [13]. There is also an increased prevalence of cryptococcosis in patients receiving therapeutic doses of corticosteroids, patients with lymphoreticular malignancies (especially Hodgkin's disease), those with a renal transplantation, and with sarcoidosis (even in the absence of corticosteroid therapy). Diabetes mellitus has also been cited as a predisposing factor for cryptococcosis. However, patients with cryptococcosis but without any underlying disease may also sometimes report at health clinics.

Several studies have shown the importance of T lymphocytes, CD4⁺ and CD8⁺, in mediating pulmonary clearance of C. neoformans in mice [14]. B cell-deficient mice are not at increased risk of Cryptococcal infection [15]. However, the presence of antibodies, acting as potent opsonins [16] required for natural killer cell [17] and leucocyte [18] anti-fungal activity in vitro suggests that humoral immunity plays an important role in this infection. Recently, there has been renewed interest in antibody immunity to glucuronoxylomannan (GXM) for prevention [19] and treatment [20] of human infection. To date, some studies of C. neoformans protein have been described [21–23]. However, little is known about the proteins secreted or released by C. neoformans, despite evidence that they elicit important immune responses.

We recently analysed the serological responses of mice with pulmonary cryptococcosis [24]. Briefly, Western blotting analysis showed that experimentally induced pulmonary cryptococcosis in (BALB/c × DBA/2)F₁mice was associated with the appearance of serum antibodies to a 77-kD protein derived from C. neoformans as well as to 18-, 22-, 25-, 36- and 94-kD proteins. The immunodominant 77-kD band also reacted with antibodies against hsp70 family members. Further, we purified a 77-kD antigen from C. neoformans cell extracts. N-terminal amino acid sequencing of the 77-kD antigen confirmed that it was a member of the hsp70 protein family. These results showed that the 70-kD hsp family from C. neoformans was the major target molecule of the humoral response in murine pulmonary cryptococcosis. To our knowledge, this was the first report of hsp from C. neoformans.

In the present study, we examined the antibody response in the sera of 21 patients with proven pulmonary cryptococcosis and three patients with suspected disease, and compared these responses with those in patients with other pulmonary diseases and deep mycoses.

PATIENTS AND METHODS

Subjects

Patients with pulmonary cryptococcosis

Serum samples were collected from 21 patients with pulmonary cryptococcosis admitted to the Nagasaki University Hospital and affiliated hospitals during a 3-year period from 1993 to 1996. The study protocol was approved by the ethics committee of our institution and informed consent was obtained from all participants in this study before sample collection. The clinical characteristics of these patients are summarized in Table 1. The diagnosis of cryptococcosis was established by correlating the clinical features and histopathological examination of transbronchial lung (TBLB) or percutaneous lung aspiration biopsies in 20 patients, and by sputum culture in one. Biopsy specimens were stained by Gomori methenamine silver and mucicarmine stains. Sputum, percutaneous lung aspirates and cerebrospinal fluids (CSF) were cultured on Sabouraud dextrose agar plates. CSF was examined at least once from all 21 patients at the time of positive diagnosis, and was repeated if signs or symptoms of meningitis subsequently developed. Meningitis was diagnosed by the growth of C. neoformans from CSF. All patients were negative for HIV infection. Sera were stored at −80°C until they were tested.

Table 1.

Clinical and laboratory data of patients with pulmonary cryptococcosis. Serum cryptococcal antigen titre on initial presentation

Patients with suspected cryptococcosis

Three patients had multiple nodular shadows on the chest roentgenograms. The clinical characteristics of these three patients are summarized in Table 1 (cases 11, 12 and 17). Although several bronchoscopic studies were performed, culture and histopathological studies of TBLB, bronchoalveolar lavage specimens and sputum failed to detect C. neoformans. However, sera of these three patients were positive by Eiken latex agglutination test (Eiken, Tokyo, Japan). They were therefore suspected as cases of pulmonary cryptococcosis. Administration of anti-fungal agents for several months resulted in clinical improvement in these patients.

Control patients and healthy volunteers

Serum samples were obtained from 11 patients with pulmonary tuberculosis, 10 patients with aspergillosis (five aspergilloma, five semi-invasive aspergillosis), and 10 patients with systemic candidiasis with Cand-Tec assay (Ramco Labs Inc., Houston, TX) titre ×4 to ×16. Sera were collected from 16 healthy subjects (six males and 10 females; age range 21–67 years) with no evidence of pulmonary cryptococcosis.

Antigen

Cryptococcus neoformans-derived hsp70 and its fragment 43-kD antigen as well as hsp60 were purified by ATP agarose chromatography as described previously (Fig. 1a) [24]. Briefly, a modified protein extraction procedure was used based on the method described by Hazen & Cutler [25]. A serotype A encapsulated strain of C. neoformans, YC-11, was grown on Sabouraud dextrose agar plates for 48 h at 30°C, after which the plates were transferred to 42°C for 30 min, and then to 30°C for 3 h. Cryptococcus neoformans cells heat-stressed in this way were then centrifuged and washed. The cell pellet was suspended in a protein extraction buffer (10 mm Tris–HCl pH 7.5, 6 mmβ-mercaptoethanol (β-ME), 10 mm Mg acetate, 60 mm NH₄Cl, 1.0 μg/ml pepstatin A and 1.0 mm PMSF) and transferred to a 50-ml thick-walled tube (Becton Dickinson, Frankline Lake, NJ). Glass beads were added to the cell suspension. The cells were vortexed while cooling and then sonicated in an ultrasonic cleaner (Yamato, Tokyo, Japan). Specimens were cleared, and the supernatant filtered (Sterivex-GV 0.22 μm Filter unit; Millipore, Bedford, MA) and the filtrate was precipitated with 80% ammonium sulphate. In the next step, a modified procedure for purification of hsp70 was used based on the method described by Welch & Feramisco [26]. Briefly, the protein sample was dialysed against 10 mm Tris acetate pH 7.5, 10 mm NaCl, 0.1 mm EDTA and a 100 000 g supernatant obtained. This sample was resolved on a fast performance liquid chromatography (FPLC) system (Mono Q; Pharmacia Inc., Piscataway, NJ) equilibrated with 10 mm Tris acetate pH 7.5, 10 mm NaCl, and 0.1 mm EDTA. The proteins were eluted in a linear gradient of a 20–500 mm NaCl and collected as 1.0 ml/min fractions. Hsp70-containing fractions were detected by immunoblotting. The ATP eluate was concentrated by centrifugation over a Centricon 10 membrane (cut-off mol wt approx. 10 000; Amicon Corp., Danvers, MA)

Fig. 1.

(a) SDS–PAGE analysis of the ATP binding protein from Cryptococcus neoformans whole cell extract. The antigens were electrophoresed on a 10% gel and then stained with silver. Positions of molecular mass markers are shown on the left. (b) Western blot analysis of ATP binding protein including hsp70 from C. neoformans with sera from patients with pulmonary cryptococcosis. The antigens were electrophoresed on a 10% gel, transferred to PVDF membrane, and incubated with a 1:50 dilution of sera from patients with pulmonary cryptococcosis or 1:250 dilution of sera from infected mice. Lanes A–U, sera from patients with pulmonary cryptococcosis; lane X, sera from infected mice (post-infection day 23). The positions of protein molecular weight markers, with respective sizes, are indicated on the left.

Western blotting

Protein samples were denatured and reduced by heating for 5 min at 100°C in 62.5 mm Tris–HCl pH 6.8 containing 640 mmβ-ME, 10% (v/v) glycerol, 2.0% (w/v) SDS, and 0.05% (w/v) bromophenol blue. The solubilized antigen preparations were electrophoresed on SDS–10% PAGE gels (Fig. 1a). After electrophoresis, the proteins were electroblotted onto polyvinylidine difluoride (PVDF) membrane (ATTO, Tokyo, Japan). A molecular mass standard (Pharmacia Biotech, Uppsala, Sweden) containing phosphorylase b (94.0 kD), albumin (67.0 kD), ovalbumin (43.0 kD), carbonic anhydrase (30.0 kD) and trypsin inhibitor (20.1 kD) was electrophoresed in parallel with the test samples. The PVDF membrane was washed three times for 20 min in 0.01 m PBS with 0.1% Tween 20 (Sigma, St Louis, MO) (PBS–T) and then once for 20 min in PBS, and incubated in PBS with 1% H₂O₂ for 2 h at room temperature. This was followed by incubation in PBS–T with 2% normal goat serum (Dako, Tokyo, Japan) (G–PBS–T) and 2% skim milk for 2 h at room temperature to block the unoccupied protein binding sites. After washing with PBS–T three times for 20 min, the membrane was incubated with patient sera diluted 1:50 in G–PBS–T overnight at 4°C with gentle agitation. After three washes with PBS–T, the strips were incubated for 30 min with peroxidase-coupled affinity-purified goat anti-human IgG (H+L) (BioRad, Hercules, CA) second antibody diluted 1:5000 in PBS–T. The strips were rinsed three times for 20 min in PBS–T before the addition of enhanced chemiluminescence (ECL) Western blotting detection reagents (Amersham, Aylesbury, UK) as the peroxidase substrate, and then exposed for 3–5 s to autoradiography film (Hyperfilm-ECL; Amersham).

Eiken latex agglutination test

Sera (300 μl) were first treated with pronase at 56°C for 30 min, after which the reaction was stopped by heating at 100°C for 5 min. Neat samples (75 μl) and samples diluted with 0.1% bovine serum albumin (BSA) containing 0.2% glycine-buffered saline pH 8.4 (75 μl) were mixed with anti-cryptococcal antibody-coated latex (25 μl) and mixed by shaker (125 ± 25 rev/min) for 10 min. A positive control (50 ng/ml of soluble capsular polysaccharide, isolated from the supernatant of the cultured C. neoformans type A) and a negative control (sample diluent) were used in each assay. When a sample was considered positive, the titre was determined by serial dilution.

Statistical analysis

The χ² test was used to determine the significance of differences in proportions between groups. P < 0.05 was considered statistically significant.

RESULTS

Cryptococcal antigen detection test

Of the patients with proven cryptococcosis, the Eiken latex agglutination test detected cryptococcal antigen at titres of ≥ 1:1 in 19 of 21 (90.5%), and at titres of ≥ 1:8 in 14 of 21 (66.7%) patients (Table 1). Three patients were suspected of having pulmonary cryptococcosis on the basis of a high antigen titre by the Eiken Latex test; > ×32, ×32, and ×16, respectively (Table 2). No antigen was demonstrated in the sera of control subjects and healthy volunteers.

Table 2.

Numbers of cases with positive anti-Cryptococcus neoformans hsp70 antibody

Immunoblotting

Pulmonary cryptococcosis (proven and suspected)

We examined by Western blotting the immunoreactivity of serum samples from patients with cryptococcosis. ATP binding protein, including hsp70 (77 kD) from C. neoformans, was obtained by affinity chromatography on ATP-agarose as described previously [23]. Proteins (66 kD and 43 kD) were coincidentally purified with hsp70 (Fig. 1a). The 66- and 43-kD proteins were identified by N-terminal amino acid sequencing as members of hsp60 and a fragment of hsp70, respectively [23]. ATP eluate was subjected to SDS–PAGE and electroblotted onto a PVDF membrane. Anti-C. neoformans hsp70 antibody was detected in 16 of 24 (66.7%) patients with pulmonary cryptococcosis (including those with suspected disease (Fig. 1b; Table 1). When titres of ≥ 1:8 were taken as high titre in the Eiken test, anti-hsp70 antibody was detected in 14 of 17 (82.3%) patients. In contrast, two of seven (28.6%) patients with titres of ≤ 1:4 had detectable anti-hsp70 antibody (Table 1). Sera from three of four patients who had received long-term corticosteroid and/or immunosuppressive therapy were unreactive to hsp70 from C. neoformans. The 66- and 43-kD proteins were not reactive with sera from patients with cryptococcosis or sera from C. neoformans-infected mice (Fig. 1b).

Control groups and healthy volunteers

In contrast to the findings in patients with confirmed pulmonary cryptococcosis, anti-hsp70 antibody was detected in the sera of only two of 10 (20.0%; P < 0.05) patients with aspergillosis, one of 10 (10.0%; P < 0.05) patients with candidiasis, and two of 15 (13.3%; P < 0.05) healthy controls. Interestingly, the sera of four out of 11 (36.3%) patients with pulmonary tuberculosis tested positive to the anti-hsp70 antibody (Table 2).

DISCUSSION

Following our recent description that the hsp70 family derived from C. neoformans was the major target molecule for the humoral response in murine pulmonary cryptococcosis [24], we extended these results in the present study by investigating whether sera from cryptococcosis patients would react with hsp70 from C. neoformans. Hsp70 from C. neoformans was used as the antigen source, and sera from patients with pulmonary cryptococcosis (including three with suspected disease), other pulmonary infectious diseases, systemic candidiasis, as well as normal individuals were analysed for antibody to C. neoformans hsp70. Anti-C. neoformans hsp70 antibody was detected in 16 of 24 (66.7%) patients with proven or suspected pulmonary cryptococcosis. In three patients with suspected cryptococcosis, C. neoformans was not demonstrable by histopathological examination or culture. However, serum antigen titres in the Eiken test were ×16, ×32 and ×32, respectively (Table 1) and the patients responded clinically to administration of anti-fungal agents.

The immunocompetent patients, including those with cryptococcal meningitis (case 4), received proper treatment, and they did not relapse. None of the patients with underlying disease that received anti-fungal agents for 3–9 months simultaneously receiving treatment for underlying disease, suffered a relapsed. No patients with a severe immunosuppressive condition such as AIDS were included in our study. Only one patient (case 1) did not properly receive anti-fungal agents, and died of aggravation of underlying disease, Because of the above statement, we can not estimate that the presence of anti-hsp70 antibody has clinical value as prognostic factor (e.g. signifying a likelihood of a favourable outcome or relapse rate). Using sera from patients before and after treatment, we made a comparison with intensities of 70-kD band to demonstrate the amount of antibodies against hsp70 on Western blotting film. However, the difference was not quantitatively obvious.

The sensitivity and specificity of the Eiken latex agglutination test for serum cryptococcal antigen are high [27] and the test is useful in patients with pulmonary cryptococcal disease [28]. The sensitivity and specificity of the Eiken latex test in our laboratory were examined using 201 serum samples from 26 patients with cryptococcosis and 175 patients without cryptococcosis. When a cut-off titre level of ≥ 1:8 was considered positive, the sensitivity was 80.8% (21/26) and specificity was 98.9% (173/175). When the cut-off value of ≥ 1:1 was considered positive, the sensitivity and specificity were 96.2% (25/26) and 88.6% (155/175), respectively. A titre of ≥ 1:8 is considered to be strong evidence of active cryptococcosis, and, due to the high sensitivity, a titre of ≥ 1:1 is suitable for screening. High antigen titre and anti-C. neoformans hsp70 antibody were detected in three patients with suspected disease, suggesting that they had cryptococcal infection. Patients with sera positive for anti-C. neoformans hsp70 antibody were significantly more frequently associated with high titre (≥ 1:8, 82.3%) than with low titre (≤ 1:4, 28.6%) in the Eiken test. Four patients who received corticosteroid and/or immunosuppressive agents for several years were seropositive in the Eiken test. The sera of three of these four patients were not reactive with C. neoformans hsp70. Prolonged corticosteroid and immunosuppressive therapy may have limited the production of antibodies in these cases.

Hsp are highly conserved proteins during evolution, with important biological functions in protein biogenesis. Therefore, most pathogen-derived antibodies to hsp70 are thought to cross-react with other pathogen-derived hsp70. The fact that N-terminal amino acid sequences of hsp70 from various pathogens such as Saccharomyces cerevisiae, Haemophilus influenzae and Escherichia coli are quite similar [24] supports this notion. Healthy individuals appear frequently to express antibodies to hsp70. For example, Arora and co-workers [29] reported that 12 of 20 (60.0%) healthy individuals exhibited antibodies against the recombinant hsp70 of L. donovani. Kindas-Mugge et al. [30] demonstrated that IgG antibodies to hsp70 from a fibrosarcoma cell were found in 11 of 47 (23.4%) healthy subjects. In the present study, we found that sera from patients with pulmonary tuberculosis, aspergillosis, candidiasis and also healthy volunteers cross-reacted with C. neoformans hsp70. Thus, it is possible that cross-reactivity exists but is not ubiquitous across all cases.

The C-terminal region of hsp70 is the major target for the humoral immune response and thus may contain a species-specific B cell epitope(s) [31]. In contrast to the highly conserved N-terminal portion, diverse amino acid sequences have been observed in C-terminal fragments of hsp70 [32]. The 43-kD antigen, N-terminal fragment of hsp70, did not react with sera from any patient with cryptococcosis (Fig. 1b). These observations suggest that antibodies to hsp70 in patients with pulmonary cryptococcosis are mostly directed to the C-terminal portion of hsp70 in which a C. neoformans-specific amino acid sequence may be located. In fact, C. neoformans hsp70 appears to have a unique amino acid sequence, as some commercially available antibodies to all mammalian hsp70 do not cross-react with hsp70 from C. neoformans. For example, anti-hsp70 MoAb (SPA 820; human HSC70 and HSP70; Stress Gen Biotechnologies Corp., Victoria, BC, Canada) and anti-hsp70 MoAb (MA3-007; Affinity Bioreagents, Neshanic Station, NJ) cross-reacted with a broad range of species from yeast to human but not with C. neoformans (data not shown). Because of the cross-reactivity to conserved portion, whole sequence of hsp70 from C. neoformans should not be adopted as a diagnostic test, however. As stated above, these data suggest that the specific epitope of hsp70 from C. neoformans is prevalent and is involved with the immune reaction in cryptococcosis.

At present, the DNA structure of hsp70 from C. neoformans is under investigation in our laboratory, The sequencing of hsp70 from C. neoformans will reveal the epitope recognized by each antibody and following experiments will show the transition of antibodies in cryptococcosis patients quantitatively. Elsaghier et al. reported anti-mycobacterial hsp70 antibodies were reacted in 55% of 20 patients with lung mycobacterioses [33]. In our data, sera from 36.3% of 11 patients with pulmonary tuberculosis reacted to hsp70 from C. neoformans. Although the frequencies of both are approximate, the epitopes recognized by sera should be different between patients of Elsaghier et al. and ours.

Collectively, our results indicate that the 70-kD hsp family of C. neoformans appears to be a major target molecule of the humoral response, both in murine and human pulmonary cryptococcosis. Although C. neoformans hsp70 proteins have been recognized as immunodominant antigens in cryptococcosis, the epitope of hsp70 has not yet been determined. Recent studies have demonstrated that immune responses against microbial hsp mediate protection against infection [3,11,34–36]. Based on these studies, the possibility of using the 77-kD antigen, with its unique epitope structure, in designing a specific vaccine against C. neoformans infection should be examined.

Finally, our recent studies show that hsp70 from C. neoformans elicited a Th1 immune response to spleen cells in murine pulmonary cryptococcosis (data not shown). These data suggest that hsp70 from C. neoformans should be a target molecule of humoral as well as cellular immune response. Antibody to hsp70 from C. neoformans might be IgG subclass related to Th1 immune response (e.g. IgG3 or IgG2a in mice). Further detailed studies are now required to establish fully the immunological events involved in these protection mechanisms.