The Recollections of an Electron Microscopist

By Etienne de Harven

Reappraising AIDS Nov./Dec. 1998

For an electron microscopist who spent most of his research career studying retroviruses associated with murine leukemias and who followed closely their hypothetical impact on the field of human oncology, it was predictable that contemporary AIDS research was completely on the wrong track. The following "Recollections" are presented to explain why.

The importance of electron microscopy in the emergence of modern cell biology, between 1945 and 1965, is unanimously recognized. Unquestionably, the relationships between cell structures and cell functions would never have been elucidated without the high resolving power of the electron microscope (EM). What is less generally recognized, however, is the role virus research had in the study of cell ultrastructure. Historically, when Rüdenberg filed the patent for the electron microscope in 1931 (1), his motivation stemmed from the hope of visualizing the polio virus! And during the WW2 years, when electron microscopes started to be accessible to biologists, attempts to visualize "virus particles" associated with cancer cells of laboratory animals received first priority. Albert Claude, working at the Rockefeller Institute, succeeded in demonstrating the Rous sarcoma virus in chicken fibroblasts (2). And a few years later, Keith Porter et al. had a similar success in imaging the "milk factor" in murine mammary adenocarcinomas cells (3). The viral etiology of Rous sarcoma in chickens and of mammary tumors in mice were well established by microbiological ultrafiltration experiments, years before the EM pictures were published. Still, the direct observation of virus particles in these experimental tumors gave an enormous (today, we would perhaps say excessive!) impetus to virus research in oncology.

Viral etiology of several murine and avian malignancies had been clearly demonstrated by ultrafiltration experiments which permitted the calculation of the approximate diameter of the virus particles. Electron microscopists knew, therefore, what size particle to look for, frequently around 100 nm. This facilitated the initial identification of oncogenic viruses by EM, although it was later observed that many microvesicles or particulate components of normal cells fall into a similar size range.

The discovery by Charlotte Friend, at the Sloan Kettering Institute in New York, of a murine erythroleukemia transmissible by cell-free filtrates illustrates well the research approaches around 1955. Moreover, since I started working in Dr. Friend's laboratory at that time, the rationale of our research program is clearly in my mind. For EM, we gave priority to two types of sample: 1) a variety of tissues from leukemic Swiss mice (spleen, lymph nodes, thymus and bone marrow), and 2) ultracentrifugation pellets obtained from leukemic tissue cell-free filtrates which readily transmitted the disease to adult Swiss and/or DBA/2 mice. We knew from ultrafiltration experiments that the activity was absent when using filters with an average pore size diameter smaller than 200 nm. Classic theories on ultrafiltration indicated that the infective particles were therefore probably in the 100 nm range. The study of thin sections from plastic embedded leukemic tissue frequently revealed the presence of particles of approximately that diameter, closely associated with a variety of cells. The particles were bound by a single membrane and had a centrally located electron-dense core or nucleoid. They had a characteristic structure and strikingly constant diameters. To our knowledge, such particles did not resemble any known components of normal cells. However, they did resemble particles previously identified by others in several filterable experimental tumors and classified by W. Bernhard as "type C" particles (4). More importantly, we observed identical particles in ultracentrifugation pellets prepared from cell-free filtrates capable of transmitting the disease to susceptible mice. On the basis of these data, we hypothesized that these particles indeed represented the oncogenic virus etiologically related to the Friend erythroleukemia (5). Somewhat surprisingly, the virus was also found in close association with cells not apparently related to the leukemic process, such as bone marrow megakaryocytes, for example. These early EM studies also indicated that all electron- dense particles averaging 100 nm in diameter are not viruses and that a rigorous ultrastructural characterization was essential for adequate differentiation between viruses and "virus-like particles."

Fortunately, our EM studies of Friend leukemia soon added an important feature to the structural characterization of oncogenic RNA viruses. It appeared that virus assembly is a cell surface phenomenon, the surface membrane of the infected cell directly contributing the future viral envelope by a multiple step mechanism for which we coined the word "budding" (6). Viruses are released into intercellular spaces by such a budding process. EM identification of viruses of that group in other experimental malignancies became therefore more rigorous, observation of "budding particles" being required. This probably helped in eliminating thousands of "virus-like particles" observed in human malignancies and with which over-enthusiastic electron microscopists attempted to contaminate the literature! In addition, recognition of budding particles on cell surfaces clearly allowed us to identify infected single cells, and to observe that such cells are perfectly viable, with absolutely no ultrastructural evidence of any cytolytic effect of viral infection. Moreover, typical viruses were frequently observed in cells undergoing mitotic division (7).

Since obviously human experimentation was out of the question, the eventual observation in human cancer cells of particles resembling those well characterized in experimental tumors would have been of great interest, although by itself not conclusive. Around 1960, many laboratories around the world, using ever improving EM methodologies, were focusing on that target. At that time, well before the emergence of molecular biology, EM was the best we had to identify viruses in cell samples. The crucial role of EM was well recognized at the Cold Spring Harbor conference of 1962, when Lwoff, Horne and Tournier proposed classifying all viruses primarily on the basis of their morphological characteristics demonstrated by electron microscopy (8).

Continuing our studies on the Friend leukemia virus (FLV), and encouraged by Dr. J. Beard from Duke University who had considerable experience with avian leukosis, we oriented our effort toward the demonstration, by EM, of viremia in leukemic mice. The most efficient initial step in purifying avian leukosis viruses was to start not from leukemic tissues but from the blood plasma of leukemic chickens. We wondered if a similar situation could also prevail in leukemic mice. This was important for us because the efficiency of our early virus purification method from homogenates of leukemic tissues such as spleen or lymph nodes was poor. We developed, therefore, a very simple purification procedure based on two steps of Millipore filtration. A diluted plasma sample, around 10 ml (i.e. bleeding about 25 leukemic mice), was first clarified by aspiration filtration through 0.65 um filter; filtrate #1 was then further filtered through a 0.22 um membrane; filtrate #2 was then centrifuged at 30,000g for 120 min. The resulting pellet was extremely small...almost invisible, still worth embedding for EM! Thin sections of such pellets revealed the presence of a most impressive population of typical, well preserved virus particles, packed together, and with very little contaminating debris (9). This was our approach to the demonstration of viremia in 1965...

Meanwhile, many EM cancer research centers (those lead by Dr. W. Bernhard, at Villejuif, France, Dr. A.J. Dalton, at NCI, Bethesda Md, Dr. L. Dmochowski, at MD Anderson, Houston, TX, and us at Sloan Kettering, New York, NY), were spending a considerable amount of time in attempts to demonstrate virus particles associated with human cancer cells. "Virus-like particles" were occasionally reported but convinced nobody. Typical viruses were never conclusively demonstrated. This was in sharp contrast with the highly reproducible demonstration, by EM, of viruses in a variety of murine and avian leukemias and tumors. Very few papers were published to report on these negative findings in human cancers and leukemias. However, Haguenau, in 1959 (10), reported on the difficulty of identifying any typical virus particles in a large series of human mammary carcinomas. Bernhard and Leplus, in 1964 (11) in an EM survey of cases of Hodgkin's disease, lymphosarcomes, lymphoid leukemias and metastatic diseases failed to recognize virus particles associated with these malignant conditions. At Sloan Kettering in New York, I decided, in 1965, to stop surveying cases of leukemias and lymphomas by EM for the presence of viruses in view of our entirely negative results. This was reported at a conference on Methodological Approaches to the Study of Leukemias held at the Wistar Institute, in 1965 (12).

Of mice and men

Publication of these negative findings failed to discourage fanatical virus hunters! An explanation for these negative results had to be found somewhere! Perhaps the technique of EM by the thin section method was not the best approach? (although it worked perfectly for mice!). Preparing thin sections was time-consuming and skill- demanding! Who had time for that, when research funding was getting difficult, and when major pharmaceutical corporations were starting to finance "crash programs" for speedy answers?... Why don't we try the negative staining method? It is very easy and very fast! And, after all, it gave beautiful results with unenveloped viruses like adenovirus and polyoma. The results were absolute disaster because fragile RNA tumor viruses (not yet called retroviruses...) are badly distorted by air-drying during the negative staining procedures; they appear as particles with a long tail! Unfortunately many cell debris and vesicular fragments, when air-dried for negative staining, form similar "tailed" structures. Interpreting "tailed" particles as RNA tumor viruses was therefore a bonanza for virus hunters! Still, we had demonstrated that "tailed" virions were preparation artifacts which can be avoided by proper control of the osmolality and by osmium fixation prior to negative staining (13), or by critical-point drying (14). The chaos created by reports on "tailed" particles damaged the credibility of EM in the search for cancer related viruses. Cow's milk and human milk were being screened for "tailed" particles and Sol Spiegelman was eloquent on the possible risks of breast feeding...

A major discovery, which had nothing to do with EM, completely reoriented ideas on how RNA tumor viruses might work, the discovery by Temin and Baltimore, in 1970, of reverse transcriptase, RT. How RNA tumor viruses could induce genomic alterations of the cells they infect suddenly became comprehensible. Moreover, these viruses remained candidates as possible oncogens because they were known NOT to be cytolytic. RNA tumor viruses were given a new name, retroviruses, and the study of their eventual role in causing human cancer soon received levels of federal funding, after Nixon's "War Against Cancer Act," which were frighteningly in excess of what could have been expected from a frequently recurring, perhaps interesting but so far totally unproven hypothesis...


The profile of the research effort changed very drastically after the discovery of RT in 1970. Somehow, most of the research methods which had dominated the field of viral oncology from 1950 until 1970 were suddenly substituted by the most exclusive fashion of molecular biology. I observed this evolution almost as an outsider, since, in my view, electron microscopy was no longer a primary contributing method to tackle any further the hypothetical relationships between retroviruses and human cancer.

The 1970-1980 years were dominated by a series of ideas which would never have withstood scientific scrutiny 10 or 20 years before. For example:

1. It became acceptable to postulate that when viruses cannot be seen by EM in cancer cells, biochemical or immunological methods supposedly identifying viral "markers" were enough to demonstrate viral infection of the cells under scrutiny. Such markers can be an enzyme (RT), an antigen, various proteins, or some RNA sequences. Never seeing the viral particles was conveniently explained by the integration of the viral genome into the chromosomes of the alleged infected cells. To accommodate such interpretations implied complete oblivion of all we knew from previous research on cancer of experimental animals. Admittedly, in these models EM was only showing terminal steps of viral reduplication, initial phases being a series of molecular events which escape ultrastructural recognition. Still, in all the classical models such as murine and avian leukoses, visible terminal steps of viral replication (i.e. "budding") were always observed and regarded as essential for the spreading of infection from cell to cell.

2. Another shortcut with disastrous consequences has been the naive notion that any material banding at 1.16gm/ml represented retroviruses! Sure enough, true retroviruses band around that density. But this does not mean that whatever material bands at 1.16gm/ml is retroviral in nature! In the 1960s I was frequently asked to look at such 1.16gm/ml bands by biochemists: "Look at this, it forms a sharp band, it should be pure viruses!" Ultracentrifugation pellets obtained from such "sharp bands" showed, in thin sections for EM, an extreme variety of microvesicles and proteinaceous debris, but no retroviruses! Still, this approach has been (and still is!) used to identify virus "markers!" How sad it is to think that a simple EM control of such "bands" (which takes about two days, and costs a few hundred dollars, but has never been done before 1997) could have prevented these highly misleading interpretations of "markers" on which large budgets have been simply wasted...

3. Collecting viruses from the supernatant of cultures of virus infected cells raises other questions. We all remember the discovery, by Epstein (15) in 1964 of the EB virus in cultures derived from African Burkitt's lymphomas. This was an EM finding and the virus was immediately and properly classified as a member of the herpes group. To identify this DNA virus in cultured cells it was necessary to look into partially degenerating cells because, most obviously, the virus had a marked cytolytic effect. By complete contrast, retrovirus-carrying cells maintain excellent viability and released viruses can be easily recovered in the culture supernatant without the need to apply any lymphokine or growth factors stimulation to the cultures.

4. As far as scientific policy is concerned, research on potentially oncogenic viruses was dominated by the retrovirus hypothesis. Federal funding took the same direction, amplified by the incredibly naive idea that success was primarily a matter of money! Unusually large levels of federal support resulted in the creation of a retrovirus research establishment. Large numbers of research jobs were created in this venture. The intellectual freedom to think along other avenues of cancer research was rapidly dwindling, especially when major pharmaceutical companies started to offer tantalizing contracts to support polarized retrovirus research... The top priority was to demonstrate, at any cost, that retroviruses had something to do with human cancer, an hypothesis, however, which didn't receive the slightest support throughout the 1970s. Such a misdirected research effort would have been relatively inconsequential as long as public health was not involved. Unfortunately, the emergence of acquired immunodeficiency syndromes (AIDS) in 1981 gave the retrovirus establishment an opportunity to transform what could have been only an academic flop into a public health tragedy.


What happened after 1981 is so well known to all the readers of RA that I hesitate to elaborate any further. The events which have lead to today's crisis have been reviewed and analyzed most convincingly by Peter Duesberg (16). I must say that I read Duesberg's book with great attention but basically with no surprise, because the way research had been conducted on retroviruses in the 1970s had so dangerously set the stage for "Impure Science" (17)...

Soon after the first cases of "Gay related immune deficiency" were described by Michael Gottlieb it was obvious for all observers that Gallo and his associates were going to jump on the new syndrome as a Godsent opportunity to attempt to justify the lavish federal budgets they had consumed on retroviruses over the past 10 years. In 1980, the scientific community was getting more and more concerned about the absence of results in "The War against Cancer" based on retrovirus hunting. The minor episode of HTLV 1 was not enough, by far, to calm the fears of grossly misdirected federal research funds. The fact that the syndrome, soon tactically renamed "AIDS", had nothing to do with cancer was apparently of little embarrassment for Gallo. Frequent association with Kaposi sarcoma helped to blur the difference in the eyes of the public.

Dominated by the media, by special pressure groups and by the interests of several pharmaceutical companies, the AIDS establishment efforts to control the disease lost contact with open-minded, peer-reviewed medical science since the unproven HIV/AIDS hypothesis received 100% of the research funds while all other hypotheses were ignored. The general public and the medical community were made to believe that the presence of circulating antibodies is diagnostic of this disease, that Koch's postulates were outdated, that 90% of all cases of an infectious disease can be observed in males, that viremia can be measured by PCR enhancement of RNA fragments even when viral particles are not demonstrable, etc., etc...

Most conveniently, it was totally forgotten that heroin addicts were known for many decades to expose themselves to immuno- deficiencies, that nitrite inhalants have multiple toxic effects, that the extreme toxicity of AZT was known for over 20 years, that known retroviruses never have any cytolytic effects, etc., etc...

And to ensure that the AIDS establishment could profitably continue to flourish, research on any dissenting (i.e. non-HIV) hypothesis was carefully prevented by tight control of research funding and by the extreme difficulty of publishing anywhere any dissenting views... In the late 1980s, I was considering adding to my research program in Toronto more EM observations on samples from AIDS patients. Unfortunately, by that time the media and the CDC had so perfectly orchestrated the panic of a plague-like epidemic that I was quickly made to understand that my assistants would all transfer out of the lab if I had insisted to activate such a program... The HIV seropositivity test was still at that time regarded as providing reliable diagnostic data. Since then, Papadopulos and the Australian team have demonstrated that this is very far from the truth...(18).

Since my retirement in France, I take any opportunity to speak out, as openly as I can, along the lines of this little note. I am proud to be part of the "Group for the Re-appraising of the HIV-AIDS hypothesis" and I sincerely hope that the combined activities of "The Group" will soon contribute to a complete re-appraisal of the etiology of AIDS for the best interest of patient's care, and for a revival of scientific integrity in medical science!

Copyright 1998 by Prof. Etienne de Harven, MD; "Le Mas Pitou," 2879 Route de Grasse, 06530 Saint Cézaire sur Siagne, FRANCE. <>


1. Rüdenberg R (1932). Elektronenmikroskop (Electron microscope). Naturwissenschaften 20, 522.

2. Claude, A (1947-1948). Studies on cells: morphology, chemical constitution, and distribution of biochemical functions. The Harvey Lectures, Series XLIII, pp 121-164.

3. Prter KR & Thompson HP (1948). A particulate body associated with epithelial cells cultured from mammary carcinoma of mice of a milk factor strain. J. Exp. Med.,88:15-85.

4. Bernhard W (1960). The detection and study of tumor viruses with the electron microscope. Cancer Res. 20:712-727.

5. De Harven E & Friend C. (1958). Electron microscope study of a cell-free induced leukemia of the mouse: a preliminary report. J. Biophys. Biochem. Cytol. 4:151-156.

6. De Harven E & Friend C (1960). Further electron microscope studies of a mouse leukemia induced by cell-free filtrates. J. Biophys. Biochem. Cytol. 7:747-752.

7. De Harven E. (1962). Ultrastructural studies on three different types of mouse leukemia; a review. In "Tumors induced by viruses"pp. 183-206, Academic Press, Inc. New York.

8. Lwoff A, Horne R & Tournier, P (1962). Cold Spring Harbor Symposium on Quantitative Biology 27:51.

9. Friend C & de Harven, E (1965). A new method for purifying a murine leukemia virus. Fed. Proc. 24, N° 2. And: de Harven E (1965). Viremia in Friend murine leukemia: the electron microscope approach to the problem. Pathologie-Biologie 13 (3-4):125-134.

10. Haguenau F (1959). Le cancer du sein chez la femme. Etude comparative au microscope électronique et au microscope optique. Bull. Assoc. Franç. Etude du Cancer, 46:177-211.

11. Bernhard W & Leplus R (1964). In "Fine structure of the normal and malignant human lymph node". Pergamon Press, ed., Oxford.

12. De Harven E (1965). Remarks on Viruse, Leukemia and Electron Microscopy. In:"Methodological Approaches to the study of leukemias". Defendi, V. edit.; The Wistar Institute Press, Philadelphia, publ., pp. 147-156.

13. De Harven E & Friend C (1964). Structure of virus particles partially purified from the blood of leukemic mice. Virology 23:119-124.

14. De Harven E, Beju D, Evenson DP et al. (1973). Structure of critical point dried oncornaviruses. Virology 55:535-540.

15. Epstein MA, Achong BG & Barr YM (1964). Virus particles in cultured lymphoblasts from Burkitt's Lymphoma. Lancet 1:702-703.

16. Duesberg P (1966). "Inventing the AIDS Virus", Regnery Publishing, Inc., Washington DC.

17. Epstein S (1996). "Impure Science; AIDS, Activism, and the Politics of Knowledge". University of California Press, publ., Berkeley CA.

18. Papadopulos-Eleopulos E, Turner VF & Papadimitriou JM (1993). Is a positive Western blot proof of HIV infection? Bio/Technology 11:696-702.