Skip directly to content

Looking Where the Light Is

Garbage In, Garbage Out; Could It Be We're Asking Good Questions Of Bad Systems?

A Case of Turncoat Tropism -- Or, Living and Dying With "NSI"

August 1994

Why after more than a decade of AIDS research do we still know so very little about how HIV runs its devastating course through the human body? Besides the glaring need for a good animal model of AIDS, the answer may lie, in part, to imprecise science. For years, AIDS researchers have concentrated on studies of laboratory strains of HIV, using cell lines or "abnormally activated cells" which have generated an enormous body of knowledge, much of which may be biologically irrelevant.

As researchers realize that in vitro models of HIV infection may not correlate with the disease process in vivo, the best scientists are turning towards investigations of primary isolates of virus -- using cells and tissues derived from patients -- or examining retroviral infections in animals, notably SIV macaques, FIV in cats or HIV itself in SCID-hu mice.

In order to illustrate the need for increased emphasis on in vivo examinations of HIV infection, TAGline asked immunologist Mike McCune (who developed, along with Jerry Zack, the severe combined immunodeficient mouse with transplanted human fetal thymus and liver tissue -- called SCID-hu Thy/Liv mouse model) to highlight some of the discrepancies between what scientists observe in the test tube and what is happening in the body.

Mike mentioned that HIV cultured from a lab worker accidentally infected with HIV-IIIB (Gallo's notorious lab strain), a predominantly T-cell tropic strain of HIV in vitro, expressed a monocytotropic phenotype in vivo. That is, although preferentially infecting and reproducing in T-cells when studied in culture, the IIIB strain was found to betray its laboratory predilection once introduced into a living, breathing human host. With shameless infidelity, once having established itself within the cells and tissues, the virus ditched the commonplace T-cells and headed for the more exotic monocytes and macrophages. Mike commented that this incident "suggests that the isolates which are selected for growth in vitro may differ in significant ways from those that grow in vivo." As case in point, a study published last year (in the SCID-hu mouse) reported that these so-called "monocytotropic" strains of HIV actually create much greater T-cell depletion than the T-cell tropic strains.

Our description of viral strains based on in vitro phenotypes, such as cell tropism (types of cells which the virus preferentially infects), may be defining characteristics which are only relevant in culture. Mike also pointed out that early on in the epidemic it was observed in vitro that HIV-infected T-cells formed syncytia (giant, multi-nucleated, "clumped together" cells) and which led to the hypothesis that this form of direct cytopathicity was responsible for CD4+ T-cell loss in vivo. He than went on to say that "it is also clear that many of the cells that are dying during the course of HIV disease in vivo are not dying in the form of syncytia and, in fact, are in many cases not infected at all. Thus many of the 'cytopathic' effects of HIV appear to be mediated indirectly and/or are not simply through the formation of end-product giant cells." Thus, while we have characterized some strains of HIV as syncytia-inducing (SI), this phenomenon may only be happening in the test tube as well. Approximately half of those who succumb to AIDS do so with no detection of SI variants.

Different strains of HIV, SIV and FIV also exhibit varying behavior depending on whether they are observed in vitro or in animal models. SIV, for example, grows well in vitro without the nef gene, but is only pathogenic in rhesus macaques with nef. Further, there are isolates of FIV that are infectious in vitro but not in vivo. And finally, only primary isolates of HIV, but not isolates that have been tissue-culture adapted for long periods of time (for example strains IIIB, MN, RF) are infectious in the human organs of the SCID-hu mouse.

In general, Mike notes that "the events observed in tissue culture represent but a small subset of those that can occur in the human body. In vivo, most cells are resting; in vitro, the only cells that can be observed are those that are forced to grow continually. In vivo, cells exist with spatially organized compartments in tight cell-to-cell contact; in vitro, only dilute suspensions of cells can survive for long. These major differences in observational inputs tend to skew the conclusions from in vitro studies and to focus the effects of HIV on activated T-cells. The simple, hard truth is that HIV is a disease that affects many more cells than simply the CD4+ T-cell and many more systems that the peripheral blood and the peripheral lymphoid system."

"Virtually nothing is known about the events that occur in vivo from the time of successful HIV-1 contact to the demise of the patient with AIDS," Mike writes in an issue of Cell, "correspondingly, there is little that can be done to reverse the course of disease." But even as more and more scientists look to the use of clinical samples or animal models for their work, the National Institutes of Health (NIH) appears ill-prepared to support this kind of work. Complaints first surfaced last year from scientists who had difficulty obtaining clinical samples from the Multicenter AIDS Cohort Study ("MACS") or had problems getting access to monkeys from the Regional Primate Research Centers (RPRCs). While the NIH initially denied there were problems, the Division of AIDS (DAIDS) soon after changed its procedure for procuring clinical samples from its epidemiological cohort studies. In July, the National Task Force on AIDS Drug Development recommended that the NIH convene a workshop "to address the most appropriate use of and access to available animal models." If in vivo explorations of HIV pathogenesis are to be facilitated, NIH policy changes can come none too soon.