Palsy of the rear limbs in Mycobacterium lepraemurium-infected mice results from bone damage and not from nerve involvement

Abstract

A small but relatively constant proportion (3–5%) of mice chronically infected with Mycobacterium lepraemurium (MLM) develops bilateral paralysis of the rear limbs. The aim of the study was to investigate whether or not the bilateral leg palsy results from nerve involvement. Direct bacterial nerve infection or acute/delayed inflammation might possibly affect the nerves. Therefore, palsied animals were investigated for the presence of: (a) histopathological changes in the leg tissues including nerves, bones and annexes, and (b) serum antibodies to M. lepraemurium and M. leprae lipids, including phenolic glycolipid I from M. leprae. Histopathological study of the palsied legs revealed that the paralysis was not the result of direct involvement of the limb nerves, as neither bacilli nor inflammatory cells were observed in the nerve branches studied. Antibodies to brain lipids and cardiolipin were not detected in the serum of the palsied animals, thus ruling out an immune response to self-lipids as the basis for the paralysis. Although high levels of antibodies to MLM lipids were detected in the serum of palsied animals they were not related to limb paralysis, as the nerves of the palsied legs showed no evidence of inflammatory damage. In fact, nerves showed no evidence of damage. Paralysis resulted from severe damage of the leg bones. Within the bones the bone marrow became replaced by extended bacilli-laden granulomas that frequently eroded the bone wall, altering the normal architecture of the bone and its annexes, namely muscle, tendons and connective tissue. Although this study rules out definitively the infectious or inflammatory damage of nerves in murine leprosy, it opens a new avenue of research into the factors that participate in the involvement or the sparing of nerves in human and murine leprosy, respectively.

Introduction

Skin and peripheral nerves infection by Mycobacterium leprae are the hallmarks of human leprosy. Contrary to human leprosy, nerve involvement is not a recognized feature of murine leprosy [1]. In spite of this, we have reported previously that a small but consistent percentage (3–5%) of mice infected with Mycobacterium lepraemurium (MLM) develops bilateral paralysis of the rear limbs 5–6 months after intraperitoneal inoculation of the bacterial cells, thus suggesting either an autoimmune nerve-system disease (triggered possibly by cross-reacting mycobacterial lipids) or the direct infection of nerves by the mycobacterium itself [2]. The first possibility was reasonably ruled out when it was demonstrated that the palsied animals lacked antibodies to self-lipid antigens (brain lipids) or self-lipid-related lipids (cardiolipin) while bearing high levels of anti-MLM antibodies [2]. Antibodies to MLM lipid antigens have been shown to be a reliable marker of infection [3], and although they reach very high levels in advanced disease they do not seem to participate in the pathology of murine leprosy. The present communication deals with the second possibility as the cause of the rear limb paralysis, i.e. the probable infectious or inflammatory involvement of nerves in our murine model of leprosy.

Materials and methods

To assess immunological changes occurring during the infection, 100 National Institute of Health (NIH) mice were each inoculated with 50 × 10⁶ MLM bacilli by the intraperitoneal route. Each week, during 6 months, four MLM-infected and four control mice were killed and used to investigate several parameters of the humoral- and cell-mediated immunity (the results of this study will be reported in a separate communication). Three of 100 infected mice (3%) developed bilateral palsy of the rear limbs (Fig. 1) and were investigated for the presence of bacilli or leprosy lesions in the nerve tissue (including brain, spinal cord and sciatic and dermal nerves) and the adjacent structures of the palsied legs. Specimens of nerve tissue-containing structures (whole brain, spinal cord and affected legs) were collected and fixed with 4% neutral formalin for a few days, then the bony specimens were subjected to decalcification according to Pearse [4] before being processed for paraffin sectioning. Three-micron-thick sections were then prepared and treated with the conventional stains: haematoxylin–eosin for general histology and Ziehl– Neelsen for acid-fast bacilli.

Fig. 1.

Two National Institute of Health (NIH) mice severely infected with Mycobacterium lepraemurium (MLM) showing bilateral paralysis of the rear limbs.

Antibodies to lipid antigens

Antigens

Whole lipids from MLM-infected mouse tissues and M. leprae-infected armadillo tissues were prepared according to Folch et al. [5], as reported previously [3]. Lipids extracted from M. leprae-infected tissue served as a source of phenolic glycolipid-I (PGL-I). PGL-I was purified to a single lipid spot (thin-layer chromatography on silica gel) by a combination of the methods described by Vemuri et al. [6] and Hunter et al. [7], as reported elsewhere [8]. Extraction of phospholipids and cerebrosides from the mouse brain has also been described [2].

Antibodies

The sera of 12 healthy mice and 12 MLM-infected animals (including the sera of the three palsied mice) were included in this study. Sera of the leprosy-infected mice were selected from many more mouse sera on the basis of their reactivity (the more reactive ones) with the homologous whole MLM-lipids in the enzyme-linked immunosorbent assay (ELISA) assay referred to below.

ELISA assay

The antibody detection system was a conventional ELISA test adapted for lipid antigens. The assay, described previously in detail [2,3,8], included coating the ELISA wells with 5 µg of lipids dissolved in ethanol, blocking the wells with defatted milk in phosphate buffered saline (DFM-PBS), with the addition of test sera diluted 1 : 100 in DFM-PBS, peroxidase-goat anti-whole mouse immunoglobulins at the appropriate dilution (1 : 1500) with detergent-free PBS washings between steps and the substrate-chromogenic mixture (hydrogen peroxide and o-phenylene diamine), arresting the reaction with sulphuric acid and reading the absorbency at 492 nm.

The lipid antigens used in this study included whole lipids extracted from M. lepraemurium-infected mouse tissue (MLM-L), whole lipids extracted from M. leprae-infected armadillo tissue (ML-L), phenolic glycolipid-I (PGL-I) purified from ML-L and phospholipid/cerebroside lipids extracted from a healthy mouse brain (BR-L).

Statistical analysis

The non-parametric Mann–Whitney U-test was used for statistical analysis.

Results

The following results correspond to the palsied animals which were killed by the fifth month of infection.

Histological alterations

Dermal nerves

No dermal nerve alteration was observed in the leg tissue of the three palsied animals. Nerves appeared intact, therefore free of bacilli, bacilli-containing granulomas and inflammatory infiltration of any kind (Fig. 2).

Fig. 2.

Dermal nerve (DN) of a palsied leg of a mouse infected with Mycobacterium lepraemurium. Note the healthy appearance of the nerve, e.g. the lack of foamy or bacilli-containing macrophages and the lack of inflammatory cell infiltrate of any type (haematoxylin–eosin, 12·5 × 40 × magnification).

Spinal cord

Neither bacilli nor lesions were observed in the dermal nerves or in the spinal cord and its branches of mice with advanced murine leprosy and bilateral paralysis of the rear limbs (Fig. 3). The lack of murine leprosy lesions or bacilli in the brain (not illustrated), spinal cord and their branches, sciatic (not illustrated) and dermal nerves, and the large amount of bacilli-laden granulomas outside the nerve tissue-containing structures such as the skin and annexes, was striking.

Fig. 3.

Transverse section of a lumbar vertebra of a mouse with advanced infection with Mycobacterium lepraemurium and bilateral paralysis of the rear limbs. The contrasting heavy infection of the vertebral body and the spared spinal cord (SC) are evident. The confluent bacilliferous granulomata (CBG) completely substitutes the bone marrow (BM) in the spongy vertebral body (bacilli are identified by the fuchsia colour they take in the Ziehl–Neelsen stain, 12·5 × 40 magnification).

Femur and tibia

The extensive damage of the leg bones, and bones in general, was surprising. Bone marrow of the palsied limbs contained numerous granulomas made up of bacilli-laden macrophages. Granulomas frequently invaded and eroded extended zones of the bone, including the periostium (Fig. 4), and this was also true for the tarsal and metatarsal bones and vertebral column at any level. However, no evidence was observed of acute (polymorph nuclear) or delayed (mononuclear) inflammatory reactions either in the nerves or in the bones of the palsied animals.

Fig. 4.

Longitudinal section of the femur of a palsied leg. The bone marrow tissue appears totally substituted by a confluent bacilliferous granulomata. Erosion (E) and hyperplasia (HP) of the bone periostium are common features of the palsied legs (Ziehl–Neelsen stain, 12·5 × 40 × magnification).

Bone-associated muscles

A frequent finding was the presence of bacilli-laden granulomas, some of them ‘leaking out’ from the eroded bones, interfering with the attachment of the muscle bundles to the bones (Fig. 5).

Fig. 5.

Tarsal section of a mouse with bilateral paralysis of the rear limbs showing a portion of a bone (BPO, bone periostium) and a muscular bundle (MB) whose trajectory is intervened by a bacilliferous granulomata (BG). (a) Haematoxylin–eosin stain (12·5 × 40 × magnification); (b) Ziehl–Neelsen stain for acid-fast bacilli (12·5 × 40 × magnification)

Bones in non-palsied mice

The leg bones of mice with similar times of infection (ca. 5 months), but without paralysis, showed a substantially lesser degree of involvement and although bacilli-laden macrophages were present, most of the bone marrow still had a healthy appearance. In these specimens, no erosion, hyperplasia or metaplasia of the bone periostium was observed (Fig. 6).

Fig. 6.

Transverse femur section of a non-palsied leg taken from a mouse with a 5-month infection with Mycobacterium lepraemurium (MLM). This should be compared with Fig. 4, from a palsied animal. Although granulomas with AFB (G) are found in the bone marrow (BM), they are not extended enough to completely replace the BM nor to injure the bone tissue (BP, bone periostium (PO)) itself (Ziehl–Neelsen stain, 10 × magnification).

Antibodies

Antibodies in MLM-infected versus healthy mice.

Compared to healthy animals, the MLM-infected mice produced antibodies to the lipids of MLM but not, at significant levels, to the other lipids tested, including the lipids of M. leprae (whole lipids and phenolic glycolipid-I) and the lipids extracted from a mouse brain (mainly phospholipids). The mean reactivity (OD_{492 nm}) of the sera to the whole lipids extracted from M. lepraemurium (MLM-L) was 0·662 ± 0·114 in the MLM-infected group (MLM-mice) and 0·016 ± 0·015 in the healthy group (control mice) (P < 0·0001). With the other lipids tested, the average reactivity of the sera of both groups, control and MLM-infected mice, was very low and comparable, with no meaningful statistical differences (P > 0·05 for most cases): ML-L (0·004 ± 0·006 versus 0·020 ± 0·014), PGL-I (0·009 ± 0·012 versus 0·043 ± 0·029) and BR-L (0·016 ± 0·019 versus 0·036 ± 0·016), respectively. These results are shown in Fig. 7.

Fig. 7.

Reactivity of the sera of 12 normal mice and 12 mice infected with Mycobacterium lepraemurium (MLM) with the lipid antigens MLM-L (whole lipids extracted from MLM), ML-L (whole lipids extracted from M. leprae-infected tissues), PGL-I (phenolic glycolipid-I of M. leprae) and BR-L (lipids extracted from a mouse's brain, mainly phospholipids).

Discussion

Despite several general resemblances between murine and human leprosy there are important differences between the two diseases, the most obvious being related to nerve involvement, frequent in human leprosy but absent in murine leprosy. Both illnesses are produced by acid-fast bacilli which are intracellular parasites of macrophages, both slowly evolve granulomatous diseases, the skin and internal organs are affected similarly in the two diseases and they show similar immunological alterations that account for the gradual loss of cell-mediated immunity to the infecting mycobacteria.

Nerve involvement has long been recognized as a characteristic of human leprosy, and it is still one of the two typical features with which a diagnosis can be made reliably, both clinically and histopathologically (the other is affected skin). Some investigators have considered the possible participation of antiphospholipid antibodies in neurological alterations of the disease as antibodies to phospholipids and to certain mycobacterial lipids, such as phenolic glycolipid-I (PGL-I), have been detected in the serum of people with lepromatous leprosy [8–14] and although most authors have found no association between the presence of antilipid antibodies and the neural pathology of the disease, others have suggested such an association [15,16]. Notwithstanding that neural damage in leprosy results, more frequently, from the direct infection of nerves, Mycobacterium leprae has a unique predilection for Schwann cells (SC), the glial cells of the peripheral nervous system. M. leprae invasion of SC leads eventually to the neurological damage that underlies the sensory motor loss and subsequent deformity and disability associated with this disease. Only recently have studies begun to elucidate the early events of M. leprae infection of SC on a molecular level and the host and bacterial factors that determine the neural predilection of this bacterium. It has been shown that α-2 laminin and α-dystroglycan on the Schwann cell axon serve as initial targets for M. leprae[17,18]. Although M. leprae-surface molecules that participate in the bacterial invasion of peripheral nerves are not well known, several superficial components, including phenolic glycolipid-I (PGL-I), a 21 kDa protein and a 19 kDa lipoprotein, have been regarded as possible candidates [19–21]. PGL-I and 19-kDa lipoprotein might interact not only with α-2 laminin and the α-dystroglycan receptor, they might also interact with Toll-like receptors present on the surface of SC. Interaction of TLR-2 on the surface of SC with M. leprae ligands may not only facilitate entry of the bacteria to the SC; it also might contribute to nerve injury in leprosy by triggering apoptosis of these cells [22].

The results of the present communication lead us to conclude that the bilateral paralysis of the rear limbs observed in 3–5% of mice with a long-lasting murine leprosy infection is not due to nerve involvement (Figs 2 and 3) but to invasive damage of the leg bones (Figs 3 and 4), that frequently affects the structure and function of the local tendons and muscles (Fig. 5). This result explains the bilateral paralysis of the rear limbs reported previously by our group [2] and confirms the lack of susceptibility of nerve tissue in murine leprosy [1]. Those animals with a similar duration of infection but with a less severe bone-marrow involvement do not show the bone or muscular alterations seen in palsied mice (Fig. 6).

The lack of nerve involvement in murine leprosy probably depends more on the structure of M. lepraemurium than on the characteristics of the murine SC themselves, as murine SC have been shown to be susceptible to infection by M. leprae in studies conducted both in vitro[23,24] and in vivo[25]. If PGL-I in M. leprae is a key determinant for the entry of M. leprae to SC, then the absence of PGL-I in M. lepraemurium might well account for the lack of nerve involvement in murine leprosy. The lack of reactivity of sera of M. lepraemurium-infected mice with PGL-I indicates indirectly the absence of this lipid in this bacterium (Fig. 7). Coating MLM with PGL-I or transforming MLM with the gene coding for the 21 kDa-protein of M. leprae will probably allow this bacterium to penetrate the mouse Schwann cells. This subject is currently under research in our laboratory.