Vaccine efficacy of an attenuated but persistent Mycobacterium tuberculosis cysH mutant

It is estimated that two billion people worldwide are latently infected with Mycobacterium tuberculosis, the aetiological agent of tuberculosis (TB) (Dye et al., 2002b), and that 223 % of these latently infected persons will develop active disease at some point in their lifetime (Parrish et al., 1998). A new vaccine against TB is needed, especially in the context of the lengthy drug treatment requirement, the proliferation of multidrug-resistant strains and the burgeoning HIV/AIDS pandemic (Dye et al., 2002a). Currently, an attenuated strain of Mycobacterium bovis (bacille CalmetteGuérin; BCG) is used as a vaccine in most parts of the world. The BCG vaccine has a number of drawbacks, including the waning of protective immunity over time (Sambandamurthy & Jacobs, 2005), highly variable efficacy in adults (Fine, 1995) and poor safety in immunocompromised individuals (Armbruster et al., 1990; Besnard et al., 1993). Therefore, a new vaccine with improved efficacy and a greater margin of safety is urgently required.

Recently, several studies have described the vaccine potential of attenuated M. tuberculosis strains in the mouse model of tuberculosis. Most of these attenuated strains were gene-deleted mutants that were auxotrophic for various substrates and only some of these strains were able to generate protection equivalent to that of the BCG vaccine (Sambandamurthy & Jacobs, 2005). For example, the pantothenate (ΔpanCD), tryptophan (ΔtrpD) and proline (ΔproC) auxotrophic strains generated an overall protection comparable to that generated by BCG (Sambandamurthy et al., 2002; Smith et al., 2001). However, when mice were immunized once with either the purine (ΔpurC), leucine (ΔleuD) or lysine (ΔlysA) auxotrophic strains, they did not confer protection comparable to that conferred by BCG (Hondalus et al., 2000; Jackson et al., 1999; Pavelka et al., 2003). Here, we report an auxotrophic (for cysteine and methionine) M. tuberculosis strain (ΔcysH), with a limited replicating capacity in mouse tissues, that generated protection equivalent to or slightly better than that generated by the BCG vaccine. Attenuation of the ΔcysH mutant in the mouse model of tuberculosis has been described previously (Senaratne et al., 2006).

Bacterial strains and culture conditions. M. bovis BCGPasteur (BCG-P), ΔcysH and M. tuberculosis Erdman were grown in Middlebrook 7H9 broth containing 10 % ADC (Becton Dickinson), 0.2 % glycerol and 0.05 % Tween 80 or on Middlebrook 7H10 agar containing OADC (Becton Dickinson), 0.5 % glycerol and the antifungal agent cycloheximide (100 µg ml¹) (Sigma-Aldrich). For the growth of ΔcysH, all of the above media were supplemented with 2 mM methionine. Our previous experiments have shown that the minimum concentration of methionine necessary for the replication of ΔcysH is 2 µm. The presence or absence of the above amounts (2 mM) of methionine on 7H10 agar plates made no difference to the recovery of c.f.u. of M. tuberculosis Erdman. Construction and Southern blot confirmation of the ΔcysH mutant have been described previously (Senaratne et al., 2006).

Infection and vaccination of mice. Female C57BL/6 mice (Jackson Laboratories) were vaccinated subcutaneously with approximately 5x10⁵ c.f.u. ΔcysH or BCG-P suspended in 0.2 ml PBS containing 0.05 % Tween 80 (PBST). A control group of mice was mock-vaccinated with 0.2 ml PBST. Twelve weeks after vaccination, three mice from each of the vaccinated groups were sacrificed, and lungs, spleen and liver were harvested and homogenized in PBST. The number of tubercle bacilli in the corresponding organ homogenates was assessed by plating part of the homogenate on 7H11 agar (Difco) plates and c.f.u. were enumerated 21 days later. Within the limits of the level of detection (10 bacilli per organ), no bacteria were detected in the lungs or liver of either group of mice. However, on average, approximately 100 c.f.u. per spleen (SD <47) were detected in both groups of mice.

Sixteen weeks after vaccination, all of the mice (vaccinated or mock-vaccinated) were infected with M. tuberculosis Erdman via the inhalation route using the Inhalation Exposure System (Glas-Col). The inoculum doses in each group were assessed as described above by harvesting the right lung of three mice (from each group of mice) 24 h post-infection (p.i.). The dose of infection was approximately 1000 bacilli per lung for each infection. At 27 and 56 days p.i., the bacterial load of each organ (right lung, spleen and liver) was determined by c.f.u. enumeration as described above. Student's t-test was used to analyse the differences between c.f.u. counts obtained from different mouse groups. Differences were considered significant at the given P values.

Histopathology. Left lungs were fixed in formaldehyde and sectioned for histopathological analysis. Lung sections were either stained with haematoxylin and eosin (H&E) or stained for acid-fast bacilli (Histology Consultation Service, Everson, WA, USA). Lung sections from three mice per each group per time point were analysed by two veterinary pathologists at the University of California at Davis, USA.

Bacterial burden is significantly lower in ΔcysH- and BCG-vaccinated mice
After infection with M. tuberculosis Erdman, six mice were sacrificed from each infected group at two different time points (27 and 55 days) to analyse the bacterial burden in the mouse tissues. The number of c.f.u. recovered from all three organs of mock-vaccinated mice at both time points p.i. was significantly higher (P<0.02) than the number of c.f.u. recovered from ΔcysH- or BCG-vaccinated mouse organs (Fig. 1). The mean number of c.f.u. recovered from BCG-vaccinated mouse livers at day 55 p.i. was 5825 (±3412) and from the livers of ΔcysH-vaccinated mice was 4400 (±787) (P<0.026). However, all other c.f.u. recovery from ΔcysH-vaccinated and BCG-vaccinated mouse organs was not significantly different (P>0.075). In summary, as seen in Fig. 1, the c.f.u. counts in the lungs, spleens and livers of the vaccinated mice were reduced by more than tenfold compared with the c.f.u. counts in mock-vaccinated control mice at 27 days p.i.

(12K): Fig. 1. Vaccine efficacy of ΔcysH assessed by bacterial burden in mouse organs. Results are shown as the bacterial burden of the right lung (a), spleen (b) and liver (c) of ΔcysH-, BCG- or mock-vaccinated mice infected with M. tuberculosis Erdman 16 weeks after vaccination. Mice were infected with aerosolized bacteria (∼1000 bacilli); n=3 mice per group per time point.

Lung histopathology in ΔcysH- and BCG-vaccinated mice
In addition to the bacterial burden, we used another virulence parameter, lung histopathology, to assess the protective effect of ΔcysH. Two pathologists independently analysed M. tuberculosis Erdman-infected lung sections of vaccinated and mock-vaccinated mice. All mice developed a granulomatous interstitial pneumonia, which progressed from areas of interstitial expansion and well-demarcated nodules to coalescing nodules and diffuse lesions over time. At 27 and 55 days, both pathologists found that the extent and severity of the inflammation in the lung sections from ΔcysH-vaccinated mice was similar to or less than the inflammation in the BCG-vaccinated mice. In contrast, they found that the lesions in the mock-vaccinated mice progressed more quickly. At both 27 and 55 days, the inflammation was more extensive and severe than in either group of vaccinated mice (Fig. 2). A Fites acid-fast stain on the lung sections demonstrated very low numbers of acid-fast bacilli in the ΔcysH- and BCG-vaccinated groups at both 27 and 55 days, and at 27 days, the numbers of acid-fast bacilli in the mock-vaccinated groups were much higher than those in the ΔcysH- and BCG-vaccinated groups (Table 1).

(78K): Fig. 2. Vaccination of mice with ΔcysH decreases lung pathology. Lung sections were harvested at 55 days p.i. and stained with H&E. (a) All groups had a granulomatous interstitial pneumonia. At 55 days p.i., the mock-vaccinated group had lesions that were predominantly diffuse, whereas the lesions in the BCG- and ΔcysH-vaccinated groups were more contained, with nodules, coalescing nodules and rare diffuse areas of inflammation. A larger percentage of the parenchyma was involved in the mock-vaccinated group compared with the BCG- and ΔcysH-vaccinated groups. (b) The inflammatory infiltrate in all groups was predominantly foamy, with epithelioid macrophages intermingled with minimal numbers of neutrophils. There were small amounts of necrosis at 55 days p.i. and the mock-vaccinated group had greater neutrophil infiltration and necrosis compared with the BCG- and ΔcysH-vaccinated groups. All of the mice were infected as in Fig. 1. Photographs were taken by a veterinary pathologist and are representative of each group at the given time point. Bars, 300 µm (a); 15 µm (b).

Table 1. Summary of pathology report of mouse lung histology at 27 and 55 days p.i. Mouse lungs used for pathology analysis were from the same experiment as shown in Figs 1 and 2; values are the mean of measurements from three mice in each group. Four H&E-stained and two acid-fast stained sections from each mouse were analysed.

Thus vaccination with ΔcysH resulted in pulmonary lesions and a bacterial burden similar to or slightly less severe than those of BCG-vaccinated mice and markedly less severe than those of mock-vaccinated animals (Figs 1 and 2; Table 1). Therefore, we concluded that immunization with ΔcysH induced a protective response that was equivalent to that of the BCG vaccine. It is interesting to note that, in contrast to vaccinated mice, c.f.u. counts in the lungs and spleens of mock-vaccinated mice showed a considerable reduction in c.f.u. from 27 to 55 days p.i. (Fig. 1). This reduction in the number of bacteria may be due to the induction of a stronger immune reaction in the later phase of the infection in response to the large number of bacteria present in the early stage of the infection in mock-vaccinated mice. However, the pathological injury to lung tissues at 55 days p.i. was much higher in mock-vaccinated mice than in vaccinated mice (Fig. 2 and Table 1). The more pronounced pathological injury in mock-vaccinated mice may also be explained by the stronger immune reaction in the later stage of the infection in mock-vaccinated mice.

ΔcysH is defective in the production of 5'-adenosine phosphosulfate reductase, which is needed for the synthesis of cysteine and methionine (Senaratne et al., 2006). As reported previously, ΔcysH remains virulent in immunocompromised Rag1^/ mice (Senaratne et al., 2006); therefore, the use of the single deletion mutant ΔcysH as a vaccine candidate in its present form may not be warranted, due to its potential pathogenicity in immunocompromised individuals. However, the ability of ΔcysH to generate protection similar to that of BCG remains of immunological interest.

As mentioned above, several deletion mutant strains of M. tuberculosis that were auxotrophic for one substrate were tested previously as vaccine strains (Sambandamurthy & Jacobs, 2005); of all the auxotrophic strains evaluated, only panCD, proC and trpD mutants were able to confer protection similar to that of BCG (Sambandamurthy et al., 2002; Smith et al., 2001). It is interesting to note that the panCD, proC and trpD mutants had limited ability to replicate in mouse tissues (Sambandamurthy et al., 2002; Smith et al., 2001), whereas the replication of other auxotrophic mutants in mouse tissues was either negligible or nonexistent (Hondalus et al., 2000; Jackson et al., 1999; Pavelka et al., 2003). Our previously published data demonstrated that ΔcysH has the ability to replicate in mouse tissues (Senaratne et al., 2006), and in this study, we have shown that the level of protection provided by ΔcysH is similar to that of BCG. Therefore, the outcome of our ΔcysH vaccine experiment supports the hypothesis of Sambandamurthy & Jacobs (2005) that the capacity of live vaccine strains to generate long-term immunity in mice is related to their ability to undergo limited replication in vivo. Additionally, in a previous study, the persistence of several BCG substrains in mouse spleens was assessed to acquire information about their vaccine efficacy (Lugosi, 1992). In this study, we were able to detect both BCG and ΔcysH in mouse spleens 12 weeks after vaccination. In summary, it is possible that persistence of vaccine strains through limited replication in vivo generates better protective immunity against M. tuberculosis.

In contrast, a recent finding showed that ΔlysA ΔpanCD and ΔleuD ΔpanCD mutants auxotrophic for two substrates did not replicate in mouse tissues, yet provided protection similar to that of BCG as a vaccine (Sambandamurthy et al., 2005; Sampson et al., 2004). It is not clear why vaccination with the ΔlysA and ΔleuD single mutants failed to reduce tissue pathology and organ burden relative to BCG (Hondalus et al., 2000; Pavelka et al., 2003), whereas the ΔlysA ΔpanCD and ΔleuD ΔpanCD triple mutant strains, generated by creating two additional mutations in pantothenate biosynthesis (panCD) in the ΔlysA and ΔleuD single mutants, respectively, provided protection similar to that of BCG (Sambandamurthy et al., 2005; Sampson et al., 2004). One explanation may be that only certain types of M. tuberculosis mutants can induce a protective immune response. In this respect, the ΔcysH mutant falls into the category of ΔpanCD and suggests that deletion of lysA or leuD in ΔcysH may generate a potential vaccine strain similar to the ΔlysA ΔpanCD and ΔleuD ΔpanCD strains. In addition, combination of these mutations may improve the overall safety profile of the resultant mutant, providing potential for use in immunocompromised individuals and newborns. These systematic approaches to TB vaccination studies may contribute to the knowledge necessary to generate an efficient vaccine against M. tuberculosis.

We thank Kathleen Dunphy for a critical review of this manuscript. This work was supported by NIH-Al-51622.

Footnotes

†Present address: School of Chemistry and Bio21 Molecular Science and Biotechnology Institute, Victoria 3010, Australia.