The enigma of tuberculous vasculopathy - Is it time to review our dogmas?
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.190284
Source of Support: None, Conflict of Interest: None
Central nervous system (CNS) tuberculosis remains a formidable challenge to clinicians, largely due to lack of sensitive laboratory tests for diagnosis, and its devastating course with high a morbidity and mortality that complicates the course of the disease, despite the availability of curative antituberculous therapy. The prevalence of neurotuberculosis is second only to human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) in the developing world, with an estimated 8.7 million new cases worldwide (13% with HIV) that kills one person every 20 seconds.
The great diversity in the pathogenetic mechanisms of tuberculous meningitis (TBM) and its chronic course with diverse pathological lesions, results in a plethora of complications that include cerebral infarction, cranial nerve palsies and hydrocephalus. Tuberculous vasculitis has emerged as an important cause of stroke in the young in developing countries. Its occurrence correlates with the stage of the clinical disease, being more common in advanced stages. The predictors of stroke were investigated in an elegant study by Kalita and Misra. The stages of meningitis, hydrocephalus, degree of exudate, and hypertension were found to be predictors of stroke in TBM.
The “granuloma” – the pathologic hallmark of tuberculosis, that is central to the evolution of a successful cellular immune response, is a proverbial double edged sword, contributing to the clinical manifestations of active disease, as well as the neurological complications. Chemotherapeutic eradication of bacilli, however, is rendered difficult, owing to the sequestration of bacilli within its avascular core, shielded by a fibrous encapsulation developed as a healing phenomenon. Infection by M marinum transparent in zebra fish has permitted observations of dynamic changes in the granulomas over time. Intriguingly, the findings in this study challenge the long held belief that granuloma formation is critical to pathogen containment, and suggest instead that granulomas may indeed promote infection, raising the “friend–foe” hypothesis.
Incidence of strokes in TBM varies from 13–57%, depending on the mode of evaluation. At autopsy, 22–56% of patients have detectable infarctions. In survivors of TBM, it is the extent of vascular involvement and cerebral infarction that are the main determinants of the prognosis. Children with large, multiple and bilateral infarcts have worse motor and neuro-developmental outcome.
The central role of “tuberculous vasculopathy” in the pathogenesis of stroke and other neurological complications comes from the detailed autopsy study reported by Dastur et al., from India. A basal exudate of variable thickness typically located in the interpeduncular fossa [Figure 1], encircles the optic chiasma and anterior cerebral vessels, extending along the Sylvian fissures to entrap the internal carotid as well as the middle cerebral arteries and their perforating branches. Its extension posteriorly, covers the pontomesencephalic, medullary and cerebellar cisterns, the fibrinous tags causing blockage of the Foramina of Luschka More Details and Magendie with resultant hydrocephalus. The consistent distribution of tubercules along the pial vessels and the occasional restriction of tuberculous lesions to an arterial territory prompted the conclusion that infarcts originated from a primary vascular pathology.
Vascular involvement in TBM seen on cerebral angiography supported the autopsy findings. The characteristic angiographic triad of narrowing of supraclinoid portion of the internal carotid artery, a widely sweeping pericallosal artery, or outward bowing of the thalamostriate vein, and delayed circulation in the middle cerebral artery with scanty collaterals and early draining veins support vasculitic pathology involving the internal carotid system. The moyamoya pattern of angiographic changes in TBM with net like clusters of thin vessels in the region of basal ganglia and base of the brain reflects the development of collaterals.
Histomorphologic studies report a spectrum of cerebrovascular alterations involving the large calibre, muscular vessels the of circle of Willis and also the smaller calibre perforating vessels. Dastur et al., described periadventitial inflammation in muscular arteries traversing the fibrocaseous exudates in the subarachnoid space, with variable involvement of the vessel wall. In chronic stages, proliferative subintimal fibrosis and at times, epitheloid granulomas in the vessel wall were noted. Acute changes, fibrinoid necrosis, aneurysmal dilation and venous thrombosis were infrequent. Lammie and co-workers classified the spectrum of cerebrovascular pathology in TBM into three patterns as: (1) Infiltrative, (2) proliferative and (3) necrotizing vascular lesions. Their relative frequencies are dependant on the duration of TBM. The initial infiltrative lesions in the early stages evolve into a predominantly proliferative reaction after 2–3 weeks, as the exudates thicken and organize. Coexistence of acute and chronic vascular changes in vessels within the same exudate can occur, indicating recurrent infection. The influence of anti-tuberculous therapy on the vascular pathology and its resulting sequel has not been addressed. It is tempting to speculate that vascular alterations and infarction could also result as a consequence of immune reconstitution induced by antituberculous therapy (inducing a brisk immune response to degrading bacillary products similar to that occurring in reversal reactions in leprosy). Hemorrhagic lesions in TBM are rare, are more common in the presence of HIV (personal observation), attributable to venulitis, and less commonly to venous thrombosis. Dastur et al., by electron microscopic studies in the reactive zone around the tuberculoma, demonstrated the presence of osmiophilic altered basement membrane material or reticulin amidst the inflammatory cells and suggested that basement membrane protein related immune pathology may be responsible for perpetuating the vascular inflammatory response in TBM  reminiscent of the pathogenesis of giant cell arteritis, with elastic tissue engulfed within giant cells. This hypothesis has not been evaluated in subsequent studies.
Infarcts in TBM are multiple, bilateral and localised to the 'tuberculous zone' involving the caudate nucleus, anterior thalamus, anterior limb and genu of the internal capsule, in contrast to the posterior location of the ischemic thromboembolic infarcts. This localization has been the basis for the belief that infarcts are secondary to the involvement of the small calibre branches of medial striate, thalamotuberal and thalamostriate perforators embedded in the exudates and likely to be stretched by a coexistent hydrocephalus, rather than the main arterial branches. Large infarcts arising from the involvement of the proximal portion of the middle, anterior and posterior cerebral arteries as well as the supraclinoid portion of the internal carotid and basilar arteries are less common as documented by MRI, angiography and autopsy studies. In a recent autopsy based study from Chandigarh, macroscopic infarcts were observed to be localised to the middle cerebral artery territory secondary to proliferative lesions in large vessels. Microscopic infarctions were restricted mostly to the basilar artery distribution in the brainstem with necrotizing lesions. Infiltrative lesions were seen in all calibre vessels and did not necessarily cause infarction. Similar vascular changes are also noted in cerebral toxoplasmosis and cysticercosis.
Autopsy findings of cerebral vessels “strangled” within the exudates in basal cisterns, and prominent fibroproliferative and inflammatory vasculitis observed on microscopic examination, have perpetuated the widely held view that cerebral infarction in TBM results from vasculitis or luminal occlusion from intimal proliferation, or a combination of both these factors. However, autopsy studies have also commented on the curious dissociation between vascular lesions and absence of infarcts in the expected territory of the occluded vessels. Dastur et al., were “impressed by the absence of thrombosis as a factor producing arterial occlusion in TBM.” Even radiologically, it has been reported to be extremely uncommon to find an organizing thrombus in an appropriate vascular territory that matches the age of infarction. Arteritis also occurs in the absence of infarction.
These observations suggest that other hemorrheological factors such as vasospasm, coagulation factors, inflammatory cytokines, etc., may play a role in causation of infarcts in TBM. Shankar et al., were the first group to give importance to the finding in TBM, of vascular smooth muscle vacuolation occurring in both inflamed and uninflamed vessels; this finding is similar to that seen in meningeal vessles in traumatic subarachnoid hemorhage, in other forms of chronic meningitis, in renal vessels in experimental rat model of vasoconstriction by noradrenaline infusion, and in ischemic enteropathy. Smooth muscle vacuolation was demonstrated to represent vasospasm with accumulation of potassium in the vacuoles by electron microscopic histochemistry. The authors suggested that a similar pathogenetic mechanism operates in TBM, causing vasospasm and ischemic infarction. The process being dynamic and spasmodic, can explains the “reversibility” of stenosis demonstrated by Dalal et al., on cerebral angiograms, and the absence of luminal thrombosis seen at autopsy.
The role of autonomic regulation of cerebral circulation in generation of vasospasm has also been suggested, based on the findings in experimental feline models of subarachnoid haemorrhage, wherein supersensitivity of alpha adrenergic receptors to small quantities of norepinephrine and vasoactive hormones was seen in the basilar artery and pial vessels. Altered balance of vasoactive eicosanoids synthesised by brain tissue- thromboxane-2, a potent vasoconstrictor, and prostacyclin PGI2, a potent vasodilator, as occurs in subarachnoid haemorrhage, may lead to a dynamic ischemic injury.
Alterations in procoagulant, antithrombotic, fibrinolytic, platelet and vascular endothelial functions may contribute to an increased risk of thrombosis and infarction. A decrease in anticoagulants (Protein S), an increase of procoagulant factors (Factor VIII), and, raised levels of plasminogen activator inhibitor-1 (PAI-1) are documented to be more pronounced in stage III of TBM than in stage II.
The role of antigenic components of M tuberculosis in initiating and maintaining an immunological response was evaluated by Shankar et al., who demonstrated the localization of various Mycobacterial antigens on the endothelium and smooth muscles of arteries and arterioles sparing the internal elastic lamina. An interesting possibility of a synergistic role of bacterial infections has been proposed due to the shared bacterial wall protein of 65kDa antigen (heat-shock protein; HSP-65). Priming of the host by previous bacterial infection could induce an exaggerated immune response, cross reactive to 65kDa antigen of the mycobacterial cell wall. The role of inflammatory cytokines, tumor necrosis factor (TNF α), vascular endothelial growth factor (VEGF) and matrix metaloproteineases (MMPs) in damaging the blood brain barrier, attracting inflammatory cells, and the release of vasoactive autocoids have been proposed. Cerebral infarction has been associated with CSF interleukin (IL)8 and IL10 in patients with TBM. CSF IL6 is the only parameter that was independently associated with disease progression and the severity of disease.
Higher CSF VEGF and MMP-9 concentrations have been reported in TBM. A potent inducer of vascular permeability and angiogenesis, VEGF has been localized in the microvessels and perivascular cells in TBM. TNFα induces VEGF and its vasculotoxic effect has been attributed to its prothrombotic action and production of nitric oxide and free radicals. A TNFα inhibitor, thalidomide, has been shown to reduce mortality in experimental animals.
The resultant hemodynamic hypoperfusion may result from an interaction between vasospasm, hemorrheologic alterations, and neural control of smooth muscle, inflammatory mediators and thrombosis. Understanding the exact pathogenetic mechanism operating is critical to instituting an appropriate therapeutic modulation, to ameliorate the vascular complications, and improve prognosis. The use of corticosteroids with antituberculous therapy is believed to reduce mortality and morbidity by ameliorating the inflammatory effects and preventing hydrocephalus. Similarly, aspirin is also shown to reduce mortality in TBM, but the role of steroids and aspirin in preventing stroke in TBM needs evaluation.
Recent studies have suggested a link between tuberculosis, atherosclerosis and cardiovascular disease. The hazard ratio of ischemic stroke in a 3-year follow-up study was 1.52 times higher compared to patients without tuberculosis, even several years after the recovery from tuberculosis. A pro-atherogenic effect of antibody-mediated response to mycobacterial HSP 65, by cross reactivity with self-antigens in the blood vessels, has been demonstrated. Injection of mycobacterial HSP 65 protein induced arteriosclerosis in rabbits. In humans, a significant correlation between anti-HSP 65 antibodies and carotid atherosclerosis has been reported. Persistent immune activation in latent and active tuberculosis may also lead to cerebrovascular disease. Further research to investigate this link between tuberculosis and cerebrovascular disease is very relevant to our country, where both these diseases are in epidemic proportions.
Progress has been rather slow in reducing the impact of tuberculosis in parts of the world endemic for the disease. Current research is focused on improving the efficacy of antimicrobial treatment by developing adjunctive host-directed therapies to prevent the sequel. These aim at modification of granulomas, and other angiogenic, metabolic, age-related, and extracellular matrix modulators. Granuloma-targeted therapy identifies therapeutic targets that promote a protective response and limits tissue damage. The architecture of the granulomas with its hypoxic centre and the external fibrous and poorly vascular wall hinders permeation of the antituberculous drugs. Use of anti-VEGF bevacizumab in a rabbit model, was able to create more structurally permeative vasculature in the pulmonary granulomas, permitting the delivery of a low-molecular weight tracer. Imatinib, by inhibiting T cell population expansions, may help to downregulate the production of pro-inflammatory cytokines that push the balance of the granuloma towards destructive processes. Metformin, the biguanide class of anti-diabetic drug, has been shown to restrict the Mycobacterium tuberculosis growth, reduce tissue pathology and inflammation, enhance the host immune response, and decrease the overall disease severity in animal models. Vitamin D signaling pathways are also being explored, and promoting tissue repair to limit the unfavourable complications of Mycobacterium tuberculosis infection by therapeutic modulation of extracellular matrix (e.g., inhibition of TGF-β) is being carried out. Although these approaches appear to be attractive and promising, the heterogeneity of the disease process within and between individuals and the resulting disease complexity poses formidable challenges, as reviewed in the previous issues of Neurology India., Better understanding of the pathology and pathogenesis of the disease is essential for determining the factors that affect drug distribution into the site of infection to target the pathogen, and develop adjuvant host directed therapies to revolutionize antituberculous chemotherapy. Meanwhile, specific biomarkers in the cerebrospinal fluid that predict the development of neurological complications need to be sought so that these novel targeted adjuvant therapies can be individualized to combat this chronic debilitating disease and contribute to global disease control.
The author gratefully acknowledges her mentor, Prof.SK Shankar, Emeritus Professor, Department of Neuropathology, NIMHANS for the critical review of the write up.