Reversible ALS-like disorder in HIV infection

Patients and methods.

Results.

Discussion.

All our patients met the criteria for an ALS-like disorder18 combining both lower and upper motor neurons signs but with an unusually rapid and reversible course. We excluded inflammatory demyelinating polyneuropathy and multifocal neuropathy with conduction blocks that may also mimic MND,21 because the electrophysiologic features of these syndromes were absent in all our patients. Motor polyradiculopathy was also excluded on the basis of clinical and electrophysiologic criteria.19 A pure motor lumbosacral polyradiculopathy recently was reported early in the course of four patients with HIV-infection.22 Patients of the latter report22 were identified by their unique presentation combining symmetric severe isolated motor neuropathy of the lower limb with depressed or absent tendon reflexes in the legs. In the two patients who underwent spinal MRI, a contrast enhancement of the anterior spinal root was disclosed.22 These clinical and radiologic features are far different than those of our patients. CMV and herpes zoster infections also were ruled out in all our patients, because a progressive motor syndrome with both upper and motor neuron signs in a patient with CMV radiculitis23 and motor involvement in acute herpes zoster24 have been described. Cervical spondylitic myelopathy was also excluded by MRI. Finally, no other clear-cut causes of HIV-1 have been shown in our patients.

During a 13-year period, six patients with HIV-1 infection in a large series of 1700 were confirmed to have an ALS-like disorder, yielding a frequency of 3.5 per 1000, which exceeds the expected incidence of ALS in the general population (annual incidence varies from 0.4 to 1.76 per 100,000). Previous reports of HIV-1-associated MND (table 1) strongly suggest that this association is not coincidental but is etiologically related, and the autopsy-confirmed case8 clearly showed a loss of motor neurons in cervical and lumbar horns. Furthermore, the clinical course of one case14 and ours, with stabilization or even complete remission of neurologic symptoms after successful treatment of the HIV-1 infection, further strengthens the pathogenetic link between MND and the infection. Indeed, spontaneous resolution of MND is exceedingly rare and is not described with chronic HIV-1-related myelopathy. Moreover, the reappearance of ALS symptoms subsequent to the anti-HIV therapy escape reinforce the hypothesis of a causal link between HIV and ALS-like disorder, because long-term monotherapy or simple bitherapy is not optimal.

Like the majority of reported cases,7-14⇓⇓⇓⇓⇓⇓⇓ at ALS-mimic syndrome onset all our patients were immunodepressed, and markers25 of CNS HIV infective status (CSF p24 antigen and HIV viral load) were highly positive. It is therefore conceivable that the ALS-like disorder was probably driven by increased levels of viral replication in the CNS or the CSF compartments and improved with clearance of HIV-1 RNA from CSF. Some HIV-1-related CNS diseases are largely dependent on the magnitude of viral replication within the CNS and notably the neuronal loss.26 Supporting this hypothesis, the CSF HIV RNA of both patients controlled after neurologic improvement was undetectable. Moreover, we reported a myelo-meningo-polyneuropathy with high CSF HIV-1 RNA level that worsened on AART despite concomitant 2.85 logarithm reduction of plasma HIV-1 RNA level and only improved when CSF viral load decrease was obtained by anti-HIV-1 therapy strengthening.27 Also, five of our six patients were examined before the multitherapy era and the better control of the plasma HIV-1 viral load. Since that period we have no longer observed ALS-like disorder in treated patients who are HIV-1 seropositive. This may be explained by three potential mechanisms: 1) suppression of CNS viral replication may be caused by direct CNS penetration of antiretroviral drugs; 2) suppression of plasma viral load may reduce further CNS seeding and thereby be effective treatment; and 3) improvement of the patient’s immune system may decrease HIV-associated neurologic complications. In no case did we observe worsening of the neurologic feature with immune restoration.

The role of enteroviruses in the development of ALS still remains hypothetical. A recent study failed to detect the novel echovirus recently identified in the spinal cord of French patients with ALS.28 In a prospective study, opportunistic viruses and notably presumptive enteroviruses were identified by culture in the CSF of 5.6% of patients with HIV infection.29 No consistent correlation was yet observed between the detection of these viruses within the CSF and the presence or future development of neurologic disease. These results weaken the hypothesis that ALS is an opportunistic infection, but specific PCR has not been performed in our patients. Nonetheless, cranial nerve palsies and myelopathy were associated with CNS echovirus 6 infection in a patient with mild immunodepression and HIV-1.30

Except for Patient 2, who was completely cured with zidovudine monotherapy, therefore excluding sporadic ALS, all our patients were younger (mean age: 34 years) than the usual median age at onset of ALS (typically between 55 and 60 years).4,31⇓ That was similar for the other published HIV-1- or HIV-2-associated MND concerning patients ranging in age from 2610 to 40 years.14 Another particularity of our cases, and the majority of the other published cases, is the unusual rapidity of the ALS evolution, which was subacute in all patients, singularly progressing over a few weeks before anti-HIV therapy’s introduction or modification. Sporadic ALS evolves more slowly, over months or years, with median survival time of approximately 3.5 years from onset of symptoms.31 Moreover, early age at onset (younger than 50 years) is associated with longer survival, notably in cases of monomelic ALS.31 Two patients (1 and 4) developed rapidly progressive dementia, confirming widespread HIV-1 infection of the CNS and failure of the therapeutics, whereas dementia is seen in less than 5% of patients with sporadic ALS.31

Apart from opportunistic infections and lymphoma, vacuolar myelopathy characterized by spastic paraparesis, sensory ataxia, and urinary incontinence is probably the most common spinal involvement in HIV infection.32 It is primarily a disorder of the white matter with multiple vacuoles affecting the lateral and posterior funiculi of the thoracic spinal cord. The clinical and electrophysiologic features of HIV myelopathy are far different from those of our patients who had at onset asymmetric monomelic amyotrophy without sensory abnormalities and normal somatosensory-evoked potentials. Yet elements of ALS have been demonstrated in some cases of AIDS vacuolar myelopathies.33 Other types of HIV-1-associated myelopathies, predominantly affecting the anterior horn but without a clinical feature of ALS type, have been described. A 26-year-old man with HIV and negative HTLV-I serology developed a spastic paraparesis without sensory abnormalities as the first manifestation of HIV infection.34 Autopsy disclosed demyelination and vacuolar degeneration of the spinal cord’s anterior and lateral columns unusually respecting the posterior funiculum. Another unusual spinal cord degeneration, principally affecting the pyramidal tracts, also has been reported as a manifestation of HIV-1 infection.32

The MND viral theory has been revived in the last 10 years in the light of the following observations. First, in addition to HIV-associated MND,7-15⇓⇓⇓⇓⇓⇓⇓⇓ it was recognized that another retrovirus, HTLV-I, caused a myelopathy, and a number of these patients had lower motor neuron signs.35 One case of ALS was also associated with HTLV-I.17 In a series of 30 patients with sporadic ALS, 17 were demonstrated to have an antibody response to an unknown HTLV cross-reactive virus, and the tax-rex region of the HTLV genome was found in the peripheral blood cells of six of 15 patients tested.36 Several studies4,6,16⇓⇓ support the findings of retroviral antibodies in ALS sera. Second, it has long been noted that ALS could occur in association with neoplasia, especially lymphoma.37 Animal models of lymphotropic and neurotropic retroviral diseases suggest that lymphoid malignancies in patients with ALS could be merely a biologic marker of infection by a yet uncharacterized human retrovirus. Mice expressing transgenic human spuma retrovirus develop myoatrophy and encephalopathy, including affection of extramotor cortex as found in ALS.38 It has been demonstrated that the mouse retrovirus Cas-Br-E is associated with both leukemia and paralytic neurologic disease, comprising a vacuolar noninflammatory polio-encephalo-myelopathy in which spinal cord anterior horn cells are destroyed.37 Moreover, determinants of leukemia and neurologic disease have been mapped to a viral envelope gene and to the viral long terminal repeat region.39

If the sustained clinical response and recovery in our patients with anti-HIV therapies support the link between ALS-mimic syndrome and HIV infection, it conversely brings no evident clues about the exact mechanism by which HIV drives neurons to death. Direct infection of neurons is not yet suspected to play a major role in ALS. Neuronal loss is demonstrated in mice infected with a murine retrovirus,40 and HIV-like immunoreactivity of rare anterior horn cells has been shown,41 but ultrastructural studies did not identify HIV virions in these cells. In addition, there is no consensus with regard to persistent neuron infection by HIV. An indirect mechanism is more probable, as already suggested in vacuolar myelopathy and HTLV-1 myelopathy. One case11 associated with a monoclonal IgM anti-asialo GM1 is interesting in light of speculation that some forms of ALS may be immunologically mediated.4 HIV triggers such a robust immune response, a source of secretion of several cytokines in plasma and CSF, that AIDS dementia complex is said to be a cytokine disease driven by HIV.25 The toxicity of numerous cytokines has been proposed as a mechanism of neuronal death in ALS. Besides, observations conducted in vitro with neuronal cells indicate that HIV-1 glycoprotein envelope gp120 or Tat protein exert neurotoxic properties, possibly with the participation of cell-derived factors, by increasing the intracellular level of Ca²⁺ through a mechanism involving excessive stimulation of the excitatory amino acid receptors.25 Recent studies of patients with ALS have focused increasingly on altered glutamate metabolism in chronic neuronal loss and on high plasma and CSF concentrations of glutamate,42 leading to treatment of ALS with the glutamate release blocker riluzole. A common mechanism then appears operative and involves the activation of voltage-dependent Ca²⁺ channel and NMDA receptor-operated channels.25,42⇓

Because ours is a referral hospital and therefore represents a highly selected HIV-1-infected patient population with neurologic symptoms, we observed the largest published series and the longest follow-up of HIV-associated ALS-like disorder. This and other reports7-15⇓⇓⇓⇓⇓⇓⇓⇓ suggest that in some specific settings (especially in monomelic forms occurring in young individuals) a search for HIV-1, and more exceptionally HIV-2, infection is warranted because AART may result in improvement or stabilization of neurologic symptoms. Although no proof of a definitive relationship between HIV infection and ALS exists, the recognition of the simultaneous occurrence of these disorders may be instrumental in the eventual discovery of an etiologic mechanism that may be common to both with the development of molecular and virologic biology.