Immunotherapy of tuberculosis with Mycobacterium leprae Hsp65 as a DNA vaccine triggers cross-reactive antibodies against mammalian Hsp60 but not pathological autoimmunity

Abstract

Despite substantial efforts in recent years toward the development of new vaccines and drugs against tuberculosis (TB), success has remained elusive. Immunotherapy of TB with mycobacterial Hsp65 as a DNA vaccine (DNA-hsp65) results in a reduction of systemic bacterial loads and lung tissue damage, but the high homology of Hsp65 with the mammalian protein raises concern that pathological autoimmune responses may also be triggered. We searched for autoimmune responses elicited by DNA-hsp65 immunotherapy in mice chronically infected with TB by evaluating the humoral immune response and comprehensive histopathology using stereology. Cross-reactive antibodies between mycobacterial and mammalian Hsp60/65 were detected; however, no signs of pathological autoimmunity were found up to 60 days after the end of the therapy.

Keywords: :

• tuberculosis
• autoimmune response
• heat-shock protein
• DNA vaccine
• immunotherapy

Tuberculosis (TB) remains a leading cause of death with more than 8 million cases every year and 1.4 million deaths worldwide, representing one of the three most impacting infectious diseases for human health.¹ One of the reasons for the limited success in the combat of TB is the lack of new drugs to shorten disease treatment, which typically lasts at least 6 mo. As a result, immunotherapy for TB has gained new attention because it may also provide an adjuvant effect for conventional chemotherapy by improving the immune response of the patient against the bacillus. In this regard, three candidates for TB immunotherapy have been discussed specifically in the literature: (1) M. vaccae, (2) a post-infection vaccine called RUTI made from detoxified, fragmented M. tuberculosis cells, and (3) a DNA vaccine encoding the heat-shock protein 65 of M. leprae (DNA-hsp65).^2-4 M. vaccae is the only immunotherapy that has been tested in TB patients and, despite high individual variability, it showed some improvement in treatment outcome.⁵

Considering the well-known pathological reaction triggered in individuals sensitized or infected with M. tuberculosis, known as the Koch phenomenon,⁶ it is important to evaluate the variety of autoimmune responses triggered by immunotherapeutic candidates based on mycobacterial products, in M. tuberculosis infected hosts. This is particularly important for Hsp65-based immunobiological agents, such as DNA-hsp65 vaccine, because of its high homology with the human Hsp60 and because of the reported association of this protein as a target antigen in some human autoimmune diseases, including arthritis, atherosclerosis, and diabetes.^7-9 For this reason, we have evaluated, in vitro and in vivo, autoimmune responses in mice that have been challenged with a high dose virulent M. tuberculosis strain and immunized with DNA-hsp65 for therapeutic purposes.

The experimental protocol (Ethics Committee approval 094/2009) was performed as recently described,¹⁰ where 8-wk-old BALB/c mice were challenged with 1 × 10⁵ H37Rv M. tuberculosis cells via intra-tracheal route. Mice started to receive immunotherapy with DNA-hsp65 at day 30 post-infection. A total of four doses of 100 μg of DNA each were administered at 10 d intervals (50 μg in each quadriceps). Control mice received saline or empty vector pVAX1 (Invitrogen). At days 10 and 60 after the last dose of DNA, equivalent to 70 and 120 d after infection, respectively, the mice were euthanized for sample collection. These time points are referred to as short and long follow-up, respectively. In this model we typically observe a significant decrease in M. tuberculosis colony forming units (CFU) counts in lungs, spleen, and liver of DNA-hsp65-treated mice, in parallel with a decrease in lung inflammation, indicating a beneficial effect of the therapy.^10,11 These immunotherapeutic effects of DNA-hsp65 are mainly associated with the induction of anti-Hsp65 specific T cells producing IFN-γ, as previously demonstrated.¹² Moreover, as shown for other HSP family members, it is possible that Hsp65 also has an adjuvant activity by assisting antigen processing/presentation because it was reported that Hsp65 possesses a proteolytic activity for several types of peptides, including peptides derived from itself.¹³ It was also shown that the abolition of this property, by site direct mutagenesis, interferes with the disease outcome of NZB/NZW F1 lupus-prone mice that were immunized with Hsp65, and as a consequence, the disease was accelerated.¹⁴

To evaluate the humoral immune response following DNA-hsp65 therapy of TB, we first determined the production of anti-Hsp65 antibodies by ELISA. The levels of IgG1 antibodies specific against the M. leprae Hsp65 recombinant protein (rHsp65) were detected in all of the experimental groups at short and long time points (Fig. 1A) but were significantly higher after DNA-hsp65 immunization 10 d after the end the therapy, indicating the success of the immunization capacity of the DNA-hsp65 vaccine to induce specific antibodies. Although the levels of these antibodies were maintained in DNA-hsp65-immunized mice, the saline and vector groups presented an increase of these cross-reactive antibodies, indicating the progression of the infection (Fig. 1A). Contrary to the findings for the IgG1 isotype, IgG2a anti-rHsp65 antibodies were highly produced in all of the infected groups (i.e., saline, vector, and DNA-hsp65) at both time points (Fig. 1B). This indicates that the M. tuberculosis infection itself is able to elicit the production of IgG2a antibodies that cross-react with the M. leprae rHsp65. In addition, the levels of IgG2a were considerable higher than those of IgG1; these observations are consistent with the previously reported predominant Th1 immune response to mycobacterial Hsp65.^11,15,16

Figure 1. Production of specific and cross-reacting antibodies against mycobacterial Hsp65, human Hsp60 and DNA after M. tuberculosis infection in mice untreated or treated with DNA-hsp65 immunotherapy. (A) anti-IgG1 and (B) IgG2a antibodies reactive to recombinant M. leprae Hsp65 protein. (C) Detection of antibodies that cross-react with human Hsp60. (D) anti-DNA antibodies. The mean value obtained in healthy non-infected animals in each measurement was subtracted from all the individuals in each group prior to performing statistical analysis. *P < 0.05 compared with all other groups at the same time point by one-way ANOVA and Tukey’s multiple comparisons post-test. F1, female F1 of NZB/NZW mice used as positive controls.

Because some degree of cross-reactivity was observed in infected mice recognizing the vaccine-encoded protein in all of the experimental groups, we next evaluated whether produced antibodies could also cross-react with the homolog human protein (Hsp60). In view of the high homology between the human and the murine Hsp60s (97%), this may be considered a valid way to evaluate induced autoreactivity in this context. Indeed, the infection alone also triggered the production of antibodies that recognize the human Hsp60 (Fig. 1C), indicating infection-induced autoreactivity. It is interesting to note that the presence of these antibodies was clearly detected in the long follow-up sample only, indicating that this was a very late phase feature of the infection itself and was not particularly increased by the vaccine because the antibody levels were not different among DNA-hsp65-treated and untreated infected animals. Nonetheless, immunization of healthy mice with the DNA-hsp65 (not challenged with M. tuberculosis) can induce the production of antibodies that cross react with the human Hsp60, as demonstrated before.¹⁷ However, this antibody production is not associated to tissue damage, supporting our interpretation that the DNA-hsp65 vaccine does not seem to be involved in triggering pathological autoimmune reactions. It is important to highlight that in this previous work, we observed the induction of anti-Hsp60 cross-reactive antibodies that last up to 210 d after the beginning of the immunization with DNA-hsp65, and in our present study, the evaluations went until 90 d from the first injection of DNA-hsp65. Considering these data, it is tempting to speculate that anti-Hsp-60 cross-reactive antibodies could last at least a similar time in animals infected with M. tuberculosis.

We also evaluated whether the immunization with plasmid DNA as a vector or in the form of DNA-hsp65 vaccine triggered the production of anti-DNA antibodies, which display significant relevance in pathological autoimmunity. A strong production of anti-DNA antibodies was observed in the positive control sample from NZB/NZW F1 lupus-prone mice but not in the experimental groups (Fig. 1D). Taken together, the evaluation of the humoral immune response showed that the M. tuberculosis infection in mice elicits antibodies that cross-react with both the M. leprae Hsp65 and the human Hsp60 but does not trigger anti-DNA antibodies. Moreover, the immunotherapy with DNA-hsp65 of animals infected with M. tuberculosis does not boost the production of anti-Hsp60 cross-reactive antibodies. Nevertheless, the characterization of the anti-Hsp60 antibodies regarding IgG1 or IgG2a subtypes could also provide a better idea regarding if these cross-reactive antibodies increase mainly because of the presence of M. tuberculosis or because of DNA-hsp65 immunization in the infected animals, a shortcoming of the study that needs further evaluation. However, the evaluation of the humoral immune response, per se, does not discriminate the potentially pathological autoimmunity, and the immunized groups—with either the plasmid alone or with the DNA-hsp65 vaccine—could eventually develop pathological autoimmunity and tissue aggression where Hsp60 is upregulated during inflammation. Thus, we performed an exhaustive search for signs of pathological autoimmunity following the immunization of infected mice with either the vector or the DNA-hsp65 vaccine. Two different laboratories of pathology performed a single-blinded evaluation of samples obtained from 18 different organs (listed below). Six to ten animals per experimental group were included in the analysis. In addition, to obtain a reliable evaluation that takes into account the entire volume of each organ, the preparation of the samples for the histopathological analysis involved the application of stereological principles that allow for the evaluation of the whole organ compared with random sections commonly obtained to perform histopathological analysis.^18,19 Therefore, the organs were sliced in two different manners, depending on whether they were isotropic or anisotropic, before the inclusion in paraffin. For isotropic organs, defined as those in which it is not possible to define the orientation of the section at the histological level (i.e., adrenal gland, articulation and bones, eyes, liver, lungs, lymph nodes, pancreas, skin, spleen, ovary, thymus, thyroid), ten sections from each organ were obtained at different levels at equal intervals, covering the entire organ, and stained for hematoxylin and eosin (HE). For anisotropic organs (i.e., brain, heart, kidneys, large intestine, small intestine, and skeletal muscle), the orientator methodology was applied to convert the organ to isotropic (Fig. 2), and then all of the resultant pieces of the organ were immersed together in paraffin, sliced, and stained, as described before. The alterations found in the histopathological evaluation were classified as minor or major, depending on the frequency and the size of the lesions observed. Minor alterations were found in essentially all organs, which included foci of congestion, hemorrhage, and foci of lymphoplasmacytic inflammatory infiltration (Table 1). Considering that these alterations were of minor intensity and were observed in all of the experimental groups, including in healthy control animals, they were not considered to be associated either with the infection or with the immunization with plasmid or DNA-hsp65 vaccine.

Figure 2. Preparation of isotropic and anisotropic organs for paraffin embedding and histological sectioning. Isotropic organs, such as the spleen (A), do not require special procedures and were fixed and embedded in paraffin after their extraction. Anisotropic organs, such as the kidney (B), were pre-sectioned following the orientator method.¹⁸ For this method, (1) the organ was cut at random; (2) resultant fragments were cut again with a perpendicular section to the first plane, and (3) step 2 was performed again with the fragments obtained in step 2. After step 3, fragments are considered uniformly isotropic. Then, all of the fragments were immersed in paraffin and sectioned as in A. Grid unit on graph paper equals to 0.5 cm².

Table 1. Frequency of minor lesions observed in the histopathology evaluation*

	Saline f/n**	Vector f/n	DNA-hsp65 f/n
Minor congestion	4/9	3/6	5/10
Minor infiltration	5/9	0/6	5/10
Hemorrhagic foci	1/9	1/6	1/10

* The Table includes the lesions observed at both 70 and 120 d in all organs, except for those lesions well-known to be due to M. tuberculosis infection (i.e., lungs, liver, spleen, thymus, and lymph nodes). **f/n = frequency of the lesion / number of animals evaluated.

In contrast, major alterations were observed in the lungs, lymph nodes, thymus, spleen, and liver of infected mice but not in healthy controls, indicating that the alterations were due to active TB disease (Fig. 3). With the exception of the lungs, the pathological findings consisted of slight lymphoid hyperplasia with mononuclear cell infiltration, with similar intensity among the experimental groups, which is consistent with commonly reported findings due to M. tuberculosis infection.^20,21 The alterations in the lungs presented the characteristics of TB pneumonia, consisting of a multifocal process of granulomatous aspect with parenchymatous, subpleural and peribronchovascular distribution, with the typical presence of foamy macrophages surrounded by lymphocytes and some neutrophils. The DNA-hsp65-treated group displayed a more focalized process than the more spread aspect observed in the saline and vector-injected animals, as described recently.¹⁰

Figure 3. Histopathological analysis of mice infected with M. tuberculosis and treated or not with DNA-hsp65 immunotherapy. Sections of lung, spleen, lymph node, thymus, and liver were obtained from not infected, saline, vector, and DNA-hsp65 groups. The lung samples presented an active inflammatory process with granuloma-like structures across the tissue in samples from saline and vector injected groups. These lesions were less prominent and smaller in DNA-hsp65-treated animals. In contrast, samples from all of the infected groups presented a slight lymphoid hyperplasia in spleen, lymph node, thymus, and liver. The figure shows the results at 120 d after infection. Similar findings were observed at day 70. Tissue was fixed with 10% buffered formalin, and 5 μm sections were stained with hematoxylin and eosin as described in the text. Magnification panels, 100 ×.

Considering that Hsp60 is a widely expressed molecule across the tissues and increases its expression in stress conditions,²² we could expect that DNA-hsp65 treated animals show signs of autoimmune attack in organs where tuberculosis infection is not common and/or worsen the lesions in the organs where M. tuberculosis is present, as is observed during the Koch phenomenon. In this manner, the stereological analysis, commonly used for meticulous search of lesions or transgene expression,^23-25 although not able by itself to discriminate between the lesions caused by the infection from those that could be triggered by the breakdown of tolerance to Hsp60/65, indicates that DNA-hsp65 treatment resulted in diminution of lung damage and did not trigger pathological autoimmunity.

Taken together, our present data on the humoral immune response in addition to the results from the stereologic histopathological evaluation of treated and untreated M. tuberculosis-infected mice indicate that, although cross-reactive antibodies to Hsp60 are elicited in these animals, no histological alteration compatible with autoimmune inflammation could be attributed to the vaccine or plasmid up to 60 d after the end of the therapy. These findings are in accordance with previous data from our group in a similar post-vaccination long-term follow-up evaluation (up to 210 d) in normal BALB/c mice, as mentioned before.¹⁷

Our data contrast with the observations of Taylor and colleagues,²⁶ who reported lung damage and no effect in bacterial counts in mice immunized with Hsp65. Other similar deleterious results were reported in the same study with a vaccine encoding Ag85, another well-studied antigen by different laboratories for a new anti-TB vaccine, some of which were evaluated in clinical trials.^27,28 Considering that several laboratories that have tested the efficacy of Hsp65 and Ag85-based vaccines in prophylactic and therapeutic schemes did observe a significant reduction in both lung inflammation and bacterial counts, it has been discussed that the harmful effects observed by Taylor et al.²⁶ could be due to endotoxin contamination of the vaccines used.²⁹ In addition, the downregulation of Th2 related genes and the reduction in the overall Th17 response in mice with active TB that received DNA-hsp65 therapy were also observed by our group, together with a higher lung tissue preservation.^10,30 These studies suggest that DNA-hsp65 immunization does not only act as a pro-Th1 molecule and a Th2/Th17 inhibitor but also triggers an immunomodulatory response, fine-tuning inflammation. It seems that DNA-hps65 immunization elicits a complex immune response that can be affected not only by the host’s genetic background but also by other ongoing immune responses, such as in infections and in pathological autoimmunity. Accordingly, systemic lupus erythematosus and autoimmune uveitis can be worsened by Hsp65 immunization,^14,31 but the immunization with vector DNA or DNA-hsp65 results in the reduction of joint inflammation and pancreatic islet infiltration in models of arthritis and diabetes,^32,33 indicating immunoregulatory properties of this immunobiological agent. Interestingly, regulatory T (Treg) cells seem to be part of this immunoregulatory effect because mice vaccinated with DNA-hsp65 prior to M. tuberculosis challenge presented a higher number of CD4⁺Foxp3⁺ T cells than mice immunized with plasmid alone or than those infected but not immunized.³⁴ Considering that the uncontrolled inflammation may culminate in important deleterious effects in an ongoing infection,^6,35,36 we favor the hypothesis that Treg cells may also have an important participation in the beneficial immune response induced by the DNA-hsp65 by controlling excessive inflammation. The role of Treg cells in the immunoregulatory effect of DNA-hsp65 immunization needs to be investigated further.

In conclusion, our results show that mice chronically infected with M. tuberculosis and treated with DNA-hsp65, which reduces bacterial loads systemically and inflammation in lungs, present in parallel the production of cross-reactive antibodies against Hsp60/65 but do not present pathological autoimmunity up to 60 d after the end of the therapy. Evaluations in animals treated with a combination of chemotherapy and immunotherapy at longer follow-up as well as in animals genetically predisposed to develop autoimmune diseases could be valuable to further investigate the potentially harmful effects of DNA-hsp65 immunotherapy of TB.

Abbreviations:
TB	=	tuberculosis
Hsp	=	heat-shock protein
DNA-hsp65	=	DNA vaccine encoding Hsp65

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

This research received support from Fundação de Amparo à Pesquisa do Estado de São Paulo, FAPESP (Grant number 2009/06793-7). We thank Dr Carlos A Mandarim-de-Lacerda for excellent discussion prior initiation of the study.

10.4161/hv.28249