HIV dementia: why might some viruses cause more problems than others?

Researchers are beginning to identify specific differences in HIV that change the risk of neurological complications, including some that may help in predicting the emergence of HIV dementia and that could lead to the identification of therapies that can protect against the emergence of HIV dementia, scientists heard last month at two international meetings on HIV-related neurological problems.

Although many researchers are unconvinced that there is a lower rate of HIV-related neurological disorders among some HIV subtypes, such as HIV-1C, than was seen in people with HIV-1B before highly active antiretroviral therapy (HAART), the idea has energised research into possible viral determinants of dementia — and whether those viral determinants are more or less common in certain subtypes.

“The notion that there are differences in neuropathogenesis really means that the virus must have evolved substantial and consistent differences and that those differences confer a biological phenotype that in turn affects whether a person has cognitive impairment, dementia, neuropathy and so forth. That kind of diversification usually requires that the subtypes be evolving in isolation for some time,” said Dr Ellis.

But HIV mutates rapidly, frequently altering its proteins to evade the immune response and to increase its efficiency in different hosts. As a result, it has become a rather diverse virus. In fact, the diversity of HIV within one infected individual “is greater than is seen globally with the influenza virus,” said Dr Ellis, while the global diversity of HIV “is many orders of magnitude larger than the diversity that one sees in influenza virus.”

However, despite the diversity of HIV within an infected individual, the virus that is transmitted from one individual to another is relatively conserved (retaining most, though not all, of the same genetic material as the virus which infected the previous host).

By examination of the gag and env genes of HIV isolates from infected people, as well as the closely related simian immunodeficiency virus (SIV) in monkeys, scientists have been able to note the degree of genetic relationships between viruses, classify them by groups and subtype and calculate when one group diverged from one another.

There are two separate types of HIV. HIV-1 and HIV-2 (found primarily in West Africa) are derived from different forms of SIV: HIV-1 most closely related to a type of SIV infecting chimpanzees (SIV_cpz), and HIV-2 to an SIV found in the sooty mangabey monkey (SIV_sm). These genetic relationships indicate that HIV-1 and HIV-2 were introduced separately into humans. Likewise, the three groups of HIV-1 (group M and the very rare groups O and N) were also probably introduced into humans upon separate occasions.

HIV-1 group M, which is responsible for the HIV pandemic, is believed to have crossed into humans sometime between 1915 and 1941, and all of its current subtypes or clades (A, B, C, D, F, G, H, J and K, plus the many CRFs of these clades) are believed to have evolved since that time from that one transmission event.

Without treatment, each subtype can cause AIDS. However, some data suggest there are differences in the course of disease between the different subtypes. Once in humans, it is well known that HIV-2 is much less virulent than HIV-1 with a dramatically slower rate of clinical progression. Meanwhile, recent studies suggest that HIV-1D, may be more virulent than HIV-1A.

One study from Uganda has reported that time to AIDS or death was more rapid in people infected with HIV-1D than those with HIV-1A, and another recently reported similar findings among women sex workers in Kenya.

“But if HIV clades influence neuropathogenesis, one has to postulate some kind of mechanism,” said Dr Ellis, “and many of the viral genes are believed to impact on various pathways of neuropathogenesis.”

In particular, Dr Ellis noted that laboratory studies have reported that several HIV products (notably tat and env (gp120, gp41) but also vpr, rev, and nef) can have adverse effects on cellular functioning in the CNS or affect neuro-tropic factor signalling.

However, it is important to remember that what happens in the test tube or even animal model can never reflect the complexity what is happening in the human brain. The relative contribution of any of these possible factors to neurotoxicity has never been conclusively demonstrated. In addition, despite years of research, no one is really certain whether the neurological problems seen in HIV-1B disease, such as HAD and MCMD, are on the spectrum of the same disorder or actually separate conditions triggered by different viral or immune processes.

It’s telling that in the animal model most commonly used for HIV neurological research, rhesus macaques are infected with SIV_sm.

According to Dr Ken Williams of Harvard University (who presented animal data at the conference showing that you can use drugs that don’t even penetrate the brain and still prevent neurologic damage in macaques) despite the differences between HIV-1 and SIV_sm, the rate of neurological damage in untreated macaque is around 30% in infected macaques, strikingly similar to the rates of HAD reported in industrialised countries with HIV-1B before HAART.

This suggests that if any subtype of HIV, such as HIV-1C, is less neurotoxic, it must have evolved in such a way that distinguishes it from HIV-1B and from what is seen in SIV.

Vinayaka Prasad of the Albert Einstein College of Medicine in New York and his colleagues believe they may have identified a smoking gun in the discovery that the tat protein of most HIV-1C is distinct from the tat protein generally found in other HIV and SIV subtypes.

Of all the HIV/SIV proteins, tat has most often been implicated in the pathogenesis of HAD. The recruitment of monocytes across the blood brain barrier is believed to be a crucial step in the pathogenesis of HAD. Tat is believed to facilitate this because it has been shown to have chemotactic effects, which means that it can act like a chemokine, directly recruiting immune cells to cross the blood brain barrier. It also increases the production of other monocyte chemokines such as MCP-1. (Weiss JM et al. J. Immunol. 163: 2953-2259).

In an already published report, nine out of ten HIV-1C isolates studied by Vinayaka Prasad and his colleague Dr Udaykumar Ranga, of the Jawaharalal Nehru Center for Advanced Scientific Research in Bangalore, India have an altered tat protein. The tat contains a specific amino acid mutation (a serine rather than a cysteine), which makes the protein lose its chemotactic properties. Upon mutating the serine back to a cysteine, monocyte chemotaxis is restored (Ranga 2004).

In a laboratory model for the blood brain barrier using viral isolates, Dr Ranga has found that clade B infected macrophages induce 2-3 times more monocyte migration than Indian clade C infected macrophages.

At the conference, Prasad presented new findings of studies using the SCID mouse model for HIV encephalitis. The SCID mouse HIV encephalitis model mimics some of the key features of HAD, with cognitive dysfunction, loss of neuronal integrity, gliosis (neuroglial cell proliferation), and infiltration of murine monocytes to the site of HIV infection. Plus, it also allows for the comparison of whole virus in hosts with the same genetic background — thus allowing researchers to directly compare two subtypes.

In this particular experiment, SCID mice were given an intracranial injection of monocytes that had been infected with either subtype B or subtype C or no infection (control). After five days, twelve days of behavioural studies were conducted upon the mice using a complex radial/water maze to assess cognitive and memory impairment (Prasad compared it to remembering where you had parked your car each day). Then magnetic resonance spectroscopy was performed to quantify N-acetyl aspartate levels (NAA is a measure of neuronal and axonal injury) and the mice brains were examined to detect monocyte infiltration and neuronal apoptosis.

So far, the histopathology studies showed that the mice’s brains contained similar amounts of HIV whether they were injected with HIV-1B or 1C (assessments of neuronal integrity, gliosis and infiltrates have yet to be completed). But despite having similar viral loads, the HIV-1B infected mice performed worse on the maze test. Essentially, these animals were less able to handle a higher working memory load. The clade C infected animals on the other hand, had a performance in between the clade B infected animals and the controls.

“Memory load effects showed that infection with both subtypes of HIV-1 impaired the ability to remember numerous items of information with the greatest deficits seen in clade B infections,” said Prasad.

However, they are open to other factors playing a role as well. While the early studies from India suggest that the numbers of HAD may be low, it does occur. And Dr Ranga has isolated viruses from a patient from Bangalore infected with HIV-1C with a severe case of HAD. Interestingly, the virus contained several unusual amino acid substitutions in the envelope glycoprotein that altered cellular tropism (the range of cells it could infect) and was able to form syncytia (clumps of virus or viral particles and cells) in MT-2 cells.

This is interesting because one of the possible molecular mechanisms that had previously been proposed for the increased aggressiveness of HIV-1D is its greater tendency to evolve changes in its envelope protein that allow binding to both CCR5 and CXCR4 receptors, thus allowing subtype D to fuse to a wider range of CD4 cell types.

This change is often seen in people with AIDS with subtype B as well (although it is not necessary for clinical progression), so this finding may simply be another effect rather than the root cause of subtype D’s greater pathogenecity. Nevertheless, once increased tropism is present, CD4 cell counts are more rapidly depleted (possibly because of cell death associated with syncytia).

Dr Ranga said that such a shift to CXCR4 tropism hasn’t really been seen with subtype C virus before. The patient had been infected for about 10 years, had no major opportunistic infections or symptoms aside from dementia, but has unfortunately since died. However, identification of such isolates from one individual does not necessarily mean that this particular shift was responsible for his neurological deterioration.

But another recently published study using from samples from the HIV Neurobehavioral Research Center at the University of California, San Diego patients has reported that a certain polymorphism in gp120, specifically in the hypervariable region of the V3 loop “was a lot more common in individuals who had very severe degrees of cognitive impairment than those who were either normal or had milder degrees of cognitive impairment,” said Dr Ronald Ellis (Pillai). “Interestingly, this particular variant could be present either in CSF or in the plasma.”

Unpublished data from the same researchers found that there was significant variability across clades in the frequency of this particular polymorphism. It was found in 10-11% of the isolates from clade B and D, <1-6% of clade A and in 5% of those with clade C.

Dr Dana Gabuzda of the Dana Farber Cancer Institute presented similar findings, identifying two different HIV envelope glycoprotein variants that enhance HIV’s entry and replication in macrophages and which are associated with dementia in patients.

The first env variant, N283 was found in a high frequency of brains of 43 people with HAD/HIVE (39%; n=330 sequences) but was rare in the brains of 24 people with HIV without HAD/HIVE (8%; n=151 (p<0.001). It appeared to increase the affinity of gp120 for CD4, and increased viral replication in macrophages and microglia. The second variant, D386, was much more commonly found in the blood and lymphoid tissue, rather than the brain, but was more associated with HAD nonetheless. It was found from thirteen HAD patients (26%, n=185) compared with nine non-HAD patients (7%, n=99). This variant is characterised by the loss of a glycan (sugar) that lets it bind to and infect macrophages more readily but not microglia or PBMC.

“This suggests that therapeutics targeting Env-CD4 interactions may be beneficial for preventing CNS infection and neurologic injury in HIV-infected patients,” she noted.

Dr Gabuzda has since analysed viral isolates from different clades from the Los Alamos database. The variant N283 and D386 (no glycan) was found in 49% and 17% of Clade A envelopes (n=59), 7% and 18% of Clade B envelopes (n=176), 0% and 8% of Clade C envelopes (n=122) and in 10% and 8% of Clade D respectively. However, she noted that these were simply samples without any functional data (any record of neurological symptoms in the source patient).

Part one - Are some subtypes of HIV more likely to cause neuroAIDS than others?

Part three - HIV dementia: are viral or host factors driving differences between subtypes? (full references to all articles in the series can be found in part three).