Remarks on Methods for Retroviral Isolation

By Etienne de Harven
Continuum Spring 1998

Dr. Etienne de Harven is emeritus Professor of Pathology, University of Toronto. He worked in electron microscopy (EM) primarily on the ultrastructure of retroviruses throughout his professional career of 25 years at the Sloan Kettering Institute in New York and 13 years at the University of Toronto. In 1956 he was the first to report on the EM of the Friend virus in murine (mouse) leukemia, and in 1960, to coin the word "budding" to describe steps of virus assembly on cell surfaces. He will deliver a speech at the 12th World AIDS Conference in Geneva (June 28-July 3) at the session "HIV-testing: Open Questions about Specificity".

The most impressive developments of molecular genetics over the past 20 years do not make Robert Koch's postulates obsolete. The first of these postulates indicates that to be considered as pathogenic, a microorganism should be isolated in every single case of the disease. Still, according to E. Papadopulos et all and S. Lanka (2) isolation of HIV from fresh plasma of AIDS patients has never been achieved under any circumstances. Moreover, and most surprisingly, the "efficiency" of current antiviral therapeutic protocols (AZT tri-therapy) is being measured by determining "viral load" in the plasma of treated patients. "Viral load" implies viremia I.e. the presence of circulating viral particles in the peripheral blood. The virus incriminated being allegedly a retrovirus, this would have been the time to remember that the morphology of such viruses in several animal experimental tumors and leukemias had been extensively characterised by transmission electron microscopy (EM) over the past 40 years, the viral particles having a characteristic ultrastructure and a diameter ranging between 100 and 120 um. Some of them had been studied by methods of high resolution transmission electron microscopy.(3) In the 1960s, transmission electron microscopy was by far the best available method to identify viruses within or around diseased cells. Consequently, many cancer research centers all around the world, started to compete for the best equipment and training in EM, aiming at the demonstration in human malignancies of viruses similar to those which had just been recognized as significantly associated with tumors and leukemias of several laboratory animals. This approach to cancer research appeared highly justified when Lwoff, Horne and Tournier proposed to classify all viruses primarily on the basis of their morphological features demonstrated by electron microscopy.(4) Identification of viruses by EM in leukemic animal tissues became unambiguous when steps in virus assembly, i.e. the 'budding' of complete virions from the surface of the infected cells, were described.(5) In spite of considerable efforts, the search for similar, typical viruses in human malignancies remained entirely negative. Pleomorphic membranous microvesicles, approaching viral size, and frequently described in the literature on human malignancies as "virus-like particles" were without any pathogenic significance. As stated in 1965, typical RNA tumor viruses have never been observed in association with human neoplasia.(6)

Concentrations of retroviruses from murine and avian leukemic tissue homogenates were reproducibly achieved permitting titration of infectivity into receptive laboratory animals. This was not, however, an easy approach to the problem of virus purification, large amounts of microvesicles and cell debris being usually present. As far as virus purification was concerned, it soon became evident that when viremia is present, blood plasma was far better than tissue homogenates for efficient virus isolation and purification.

In the case of RNA tumor viruses, now called retroviruses, the demonstration of viremia in the blood plasma of experimental leukemic animals (chickens and mice) was published more than 35 years ago. A most efficient purification method including ultrafiltration and ultracentrifugation of a 1/1 dilution of plasma in heparinized Ringer's solution, allowed me to demonstrate packed retroviruses by transmission electron microscopy (7) in thin sections of pellets obtained by high speed centrifugation of the purified virus, quite clearly establishing that the amount of contaminating cell debris was remarkably small, a conclusion which could never have been reached by using the negative staining EM method. Using this simple ultrafiltration procedure, virions were never exposed to hypertonic shock. However, sedimentation in sucrose density gradients, at the density of 1.16 gm/ml, soon became the most popular method for retrovirus purification.(8) Interestingly, it was very well known by electron microscopists in the 1960s, that sharp bands sedimenting at the density of 1.16 frequently contained large amounts of microvesicles and cell debris of non-viral nature. These debris just happened to sediment in sucrose gradients at a density very similar to that of retroviruses clearly indicating that finding a "sharp band" at the density of 1.16 gm/ml was of little significance and was certainly far from any demonstration of retroviruses isolation.

But this conclusion was based on EM findings, and around 1970 the faith in retroviral oncology was assuming quasi-religious proportions! If EM cannot demonstrate viruses in the 1.16 bands, let us forget about EM and rely on other "markers"!

When around 1980, R. Gallo and his followers attempted to demonstrate that certain retroviruses can be suspected of representing; human pathogens, to the best of my bibliographical recollection, electron microscopy was never used to demonstrate directly viremia in the studied patients. Why? Most probably, EM results were negative and swiftly ignored! But over-enthusiastic retrovirologists continued to rely on the identification of so-called "viral markers", attempting to salvage their hypothesis.

When retrovirus particles are legion, the study of molecular markers can be useful, and provide an approach to quantification probably better than direct particle counting under the EM (which I always found very difficult). But when, using EM, retrovirus particles are absent relying exclusively on 'markers' is a methodological nonsense. 'Markers' of what?

Nevertheless, for the past ten years, HIV research and clinical therapeutic trials have been primarily based on the study of several HIV "markers".

First the antibody. Elisa, then Western Blot tests were hastily developed (at sizable financial profit eagerly split between the Pasteur Institute and the US). "Seropositivity" became synonymous with the disease itself, plunging an entire generation into behavioral panic, and exposing hundreds of thousands of people to 'preventive' antiviral AZT therapy which actually hastened the appearance of severe or lethal immunodeficiency syndrome. Appropriate controls were apparently never carried out or were never published. Still, back in 1993 it became clear that the so-called HIV antibody tests badly lacked specificity, (9) cross-reactivity being observed with patients suffering from a long list of pathological conditions including malaria, leprosy, auto-immune diseases and many more.

Secondly, 'viral proteins'. Several proteins have been identified as 'HIV markers', most frequently because they were identified in a variety of 1.16 bands. The case of the p24 "viral" antigen is a significant example and its lack of viral specificity has been well documented.(10)

Third, reverse transcription. If reverse transcriptase activity were a unique feature of retroviruses, it could have been an interesting molecular marker. Unfortunately, it has been shown that reverse transcriptase is found in the uninfected cells of yeasts, insects and mammals (11) and "has nothing to do with retroviruses as such" as well referenced in a recent report from S. Lanka. Moreover, K. Mullis himself does not support the use - to amplify and quantify the "HIV genome" - which is being made of the PCR methodology he developed, which is the current method of "measuring the viral load" in AIDS patients.

More disturbing is the fact that some 'markers' are searched for in the 1.16 gradient sedimenting material which is the density where intact virions are expected to be found, but not their molecular fragments. If lysed retrovirus particles released molecular markers, the 1.16 samples should at least initially allow investigators to demonstrate virus particles by EM. They don't. however after 15 years of most intensive HIV research, two independent groups finally decided to explore by electron microscopy the ultrastructural features of the material sedimenting at the 1.16 density. Working on "HIV-1 infected T-cell" cultures supernatants, both groups found that it contains primarily cellular debris and cell membrane vesicles which could definitely not be identified with HIV particles and rare "virus-like" particles.(12, 13) Still, this is the type of sample in which "viral markers" are currently identified and used to measure the effects of anti-viral drugs in current clinical trials.

In conclusion, and after extensive reviewing of the current AIDS research literature, the following statement appears inescapable: neither electron microscopy nor molecular markers have so far permitted a scientifically sound demonstration of retrovirus isolation directly from AIDS patients. This conclusion fully confiens the recent reports published in Continuum by E. Papadopulos and by S. Lanka.

Harvey Bialy, editor of the journal Bio/Technology has stated that (14) "A powerful hypothesis has to explain and predict. What kind of scientist continues to support a hypothesis that fails to explain and fails to predict?" The HIV/AIDS hypothesis fails to explain the considerable drop of T4 circulating lymphocytes in AIDS patients. It predicted a dramatic AIDS epidemic which was never observed (unless we accent the CDC's most surprising redefinition of AIDS as including some 31 "AIDS defining illnesses"!).

Obviously, the HIV/AIDS hypothesis has to be scientifically reappraised.(15) And, most urgently, the funding for AIDS research should no longer be restricted to laboratories working on an hypothesis which has never been proven. *

References

Author's address:
Dr Etienne de Harven
le Mas Pitou,
2879 Route de Grasse
06530 Saint Cezaire sur Siagne, France