A very large number of studies have been performed investigating the importance of TNF-alpha in arthritis, especially RA. TNF-alpha has also attracted interest in recent years for its possible role in skeletal muscle damage. Increased protein degradation, as well as decreased body weight and food consumption, was demonstrated when TNF-alpha was administered to rats via a catheter into the external jugular vein [4]. Crush injury in mice leads to elevated TNF-alpha levels in skeletal muscle tissue [5] and there is an increase in TNF-alpha serum levels in response to repetitive strain injuries [6]. However, the experiments that have been performed have sometimes yielded apparently conflicting results. For example, an experiment in which TNF-alpha was administered to mice via an osmotic pump led to accumulation of inflammatory cells in skeletal muscles but no signs of atrophy or injury [7]. Furthermore, the results of studies on TNF receptor knockout and TNF-alpha antibody-neutralized mice indicate that TNF-alpha can actually be involved in the recovery of muscle function after traumatic muscle injury [8]. Therefore, it might be that the role of TNF-alpha in muscle injury varies with the type, severity, and stage of the injury [9]. In humans, TNF-alpha is known to be intimately involved in cachexia [10], a complex condition characterised by progressive muscle loss that affects up to 13% of patients with RA [11]. Nevertheless, a recent study showed acute elevation of TNF-alpha not to affect markers of systemic or skeletal muscle turnover in healthy humans [12].

Remarkably little data is available on the role of TNF-alpha in situations where there is a pronounced infiltration of inflammatory cells in the muscle tissue, that is, myositis. From studies in tissue other than muscle, it is known that macrophages and other immunoactive cells such as monocytes, mast cells, and neutrophils are responsible for TNF-alpha production [13–16]. Data addressing a possible TNF-alpha production by inflammatory cells in myositis comes almost entirely from studies of patients affected by a group of diseases known as “idiopathic inflammatory myopathies” (inflammatory myopathies) [17, 18]. These autoimmune diseases include mainly the subgroups inflammatory myopathic polymyositis, dermatomyositis, and inclusion body myositis [19]. In these conditions, inflammatory cell-related TNF-alpha expression is localised predominantly to macrophages [18]. TNF-alpha is also expressed in the inflammatory cells in crush-injured and transplanted muscle autografts in mice [7]. Finally, blockade of TNF-alpha in the dystrophic (mdx) mouse, which is the most frequently used model of Duchenne's muscular dystrophy, reduces TNF-mediated adverse responses to exercise-induced muscle damage [20, 21]. However, without further information, it is difficult to reach conclusions on the importance of TNF-alpha and the possible usefulness of TNF-blocking in muscle disorders, including in myositis [22]. Furthermore, it should be stressed that the majority of information on the TNF system for skeletal muscle tissue has come from studies of cultured myoblasts (e.g., [23]). Animal models are needed to advance our understanding of the disease mechanisms of TNF-alpha that are involved in myositis.

The tissue specimens were cut into 10μm thick fresh cryosections using a cryostat (with a knife washed in 70% EtOH in DEPC [diethylpyrocarbonate]-H₂O) and mounted onto Super Frost Plus slides (nr.041200, Menzel Gläser). The protocol that thereafter followed was that previously used in our laboratory for detection of mRNA for other substances (e.g., [34–36]).

An alkaline phosphatase-(AP) labelled anti-DIG antibody (Roche, Germany, 11 093 274 910) was used for detection. The sections were finally mounted in Pertex mounting medium.

The corresponding sense DIG-hyperlabeled ssDNA probe was used as a negative control. As a positive control probe, a β-actin antisense probe (GD5000-OP) was used, comparisons being made with sense β-actin probe (GeneDetect, New Zealand).

A unique model for the production of myositis in rabbit musculature (the soleus muscle) was utilized. The main finding was that cells in the inflammatory infiltrates in the myositis muscles were found to express TNF-alpha at both at the mRNA and protein levels. Colocalization between TNF-alpha and CD68 was noted for these cells. Expression of TNF-alpha in macrophages has previously been noted in other situations (e.g., [13, 16]), including inflammatory myopathies [17]. In contrast, in our recent studies using the current myositis model, expression of the vesicular glutamate transporter VGluT2 was noted in white blood cells in the inflammatory infiltrates other than macrophages [24]. A further main finding was that the muscle fibers and blood vessel walls in areas showing inflammatory infiltration exhibited TNF-alpha mRNA and that fibroblasts also were seen to exhibit TNF-alpha mRNA.

From a methodological point of view, it was clear that muscle fibers, blood vessel walls, and fibroblasts exhibited TNF-alpha mRNA but that no reactivity (muscle fibers, blood vessel walls) or very weak reactivity (fibroblasts) was noted at the protein level. The production level in these locations is therefore likely to be low, which precluded clear detection with our immunohistochemical methods. It is also possible that our in situ hybridization method detects very small quantities of TNF-alpha mRNA. Nevertheless, it has previously been shown that TNF-alpha can be expressed not only in inflammatory cells but also in injured muscle fibers and fibroblasts in response to muscle injury (crush-injury) [7] as well as in muscle fibers and cells in the connective tissue in inflammatory myopathies [17, 18].

The patterns of morphologic appearances of the inflammatory infiltrates and other morphologic changes seen resembled the appearances that can be seen in the muscle tissue in inflammatory myopathies [38]. Nevertheless, preliminary analysis using ELISA detecting anti-Jo-1 antibodies, which are known to correlate with disease activity for patients with inflammatory myopathy [39], does not lend proof to the theory that the myositis in our model is autoimmune in origin (unpublished observations). However, further studies on this aspect are warranted.

As noted previously, TNF-alpha immunoreactions in inflammatory cells invading muscles affected by myositis have previously only been documented in biopsies from patients with inflammatory myopathies [17, 18] and in muscle of mice in response to crush-injury [7]. Thus, in combination with this previous work, our results imply that TNF-alpha is intimately involved in the inflammatory process in myositis. Indeed, it has been suggested that TNF-alpha may have a role in the pathogenesis of the myositis in the inflammatory myopathies [40, 41] and that a marked inflammatory response involving TNF-alpha may be directly responsible for damaging muscle fibres in myopathic conditions [42].

Whether or not the TNF-alpha produced by the cells of the inflammatory infiltrates is entirely responsible for pro-inflammatory and damaging effects remains open to speculation. It is well known that TNF-alpha administration can have pro-inflammatory and detrimental effects, for example, leading to various catabolic changes as seen in studies on cultured skeletal muscle cells [4, 43, 44]. However, there is a marked discrepancy in the literature regarding the effect of TNF-alpha on the musculature. Some studies on myoblast cell culture show that TNF-alpha administration does not have catabolic effects (e.g., [45]), and other studies documenting accumulations of inflammatory cells in skeletal muscle in response to TNF-alpha administration [4] show no decrease in skeletal muscle proteins and no signs of muscle atrophy or injury. Perhaps these discrepancies reflect a dual role of TNF-alpha where in some circumstances inflammatory cell derived TNF-alpha can play a protective role [46] and also be involved in the recovery of muscle function after traumatic injury [9] and in muscle regeneration [47]. The discrepancies may also reflect the fact that different methods have been used in the studies that have been performed. The results in preliminary studies on inflammatory myopathies suggest that TNF blocking might be useful [48], but it is also emphasized that further studies are needed in order to clarify if this type of treatment is indeed useful [22]. Results of in vitro studies suggest that targeting TNF-alpha might be worthwhile in myositis [49, 50] and studies on dystrophic mdx mice subjected to wheel exercise indicate that TNF blockade can reduce myofiber necrosis [20, 21]. The use of anti-TNF treatment in studies on a rat model of repetitive reaching and grasping leads to an improvement in grip strength and attenuated task-induced increases in inflammatory cytokines, including TNF-alpha [51].

Although the use of other animal models have shown inflammation in muscle tissue in response to various forms of exercise [50, 52], the myositis model used in the current experiment is clearly distinguishable from these models. Thus, in contrast to these models, it leads to a marked presence of inflammatory infiltrates in the muscle tissue, that is, a morphology resembling that seen in inflammatory myopathies. In fact, no experimental myositis model exists which resembles the one used here and in which a marked presence of inflammatory infiltrates becomes present in the muscle tissue. Those for which such an infiltration has been demonstrated are the model of crush-injury described above [7] and models designed to help understand the mechanisms of inflammatory myopathies that occur in man. In these latter cases, myositis is induced by various infectious agents [53, 54], immunization with muscle components, for example, myosin [55, 56], and intraperitoneal injections with lipopolysaccharide [57]. The TNF system has not been examined in any of these myositis models replicating the inflammatory myopathies seen in man. Interestingly, there is evidence indicating a relationship between the inflammatory myopathies and another condition, muscular dystrophy, in the form of complex interactions between immunological and nonimmunological features of the diseases [58].