RETROVIRUS PIONEER REJECTS
No Retroviruses Can Cause AIDS, Virus Isolation
Expert Etienne de Harven Concludes
By Paul Phillpot
Reappraising AIDS Nov./Dec. 1998
Physician Etienne de Harven was a successful cancer researcher and
retroviral expert when he read in 1981 about the first four cases of what
would later be called AIDS. He watched as public hysteria over the
possibility of an infectious cause of AIDS transformed into a palpable
demand that scientists find one.
When National Institutes of Health official Robert Gallo answered the
demand in 1984 by proposing a retroviral cause, de Harven's concern turned
to disappointment. He feels that the near universal acceptance by scientists
and physicians of Gallo's hypothesis occurred only because they did not
properly study it.
Although concerned and disappointed, this did not surprise de Harven. He
had in the 1970s watched Gallo and other retrovirologists undermine the
scientific standards of their profession to make it seem that retroviruses
could cause a non-infectious disease, cancer, in humans. That enabled
retrovirologists to become the principal beneficiaries of federal cancer
It also set the stage for their eventual takeover of the emerging AIDS
research budget, which was growing rapidly due to public fear of a
In essays published in the spring issue of Continuum and this issue of RA,
de Harven makes plain his objection to the popular view of AIDS as a
virus-induced condition. He states that retroviruses, including HIV, lack a
capacity to cause immune deficiency, and endorses UC-Berkeley
retrovirologist Peter Duesberg's assertion that such factors as narcotics
and HIV drugs like AZT provide much better explanations for AIDS. He also
articulates support for Australian biophysicist Eleni Papadopulos-Eleopulos'
argument that HIV's existence has never been properly established, and that
all the HIV tests are invalid.
De Harven's public alliance with the AIDS Reappraisal movement represents a
major endorsement due to his sterling credentials. In 1953, he graduated
prima cum laude with his MD from the University of Brussels, Belgium. As a
research fellow, first at the Institut Gustave Roussy, in Villejuif, France,
and immediately afterwards at the Sloan Kettering Institute in New York, he
rapidly made two contributions to viral research. In 1958, he and
Sloan-Kettering's Charlotte Friend published the first electron micrographs
(pictures taken with an electron microscope) of the Friend leukemia virus
(FLV), a retrovirus just discovered by Friend in murine (mouse) leukemia
cells. In 1960, he demonstrated with electron microscopy that the assembly
of such retroviral particles occurs on the host cell's outer membrane. In
describing the steps leading to the release of retroviruses from host cells,
he coined the term "budding," which has become a staple in the vocabulary of
freshmen microbiology students.
In 1962, he joined the staff of the Sloan-Kettering Institute as its chief
of Ultrastructural Research Section, in a joint appointment as Professor of
Biology at Cornell. His laboratory became an international center for
ultrastructural studies of retroviruses. In 1981, he moved to Canada, to
lead the Electron Microscope Laboratory at the University of Toronto
Pathology Department, while serving as a staff pathologist at Toronto
General Hospital. He retired from both posts in 1993, and transferred to
southern France, where he continues his affiliation with the University of
Toronto as emeritus professor of pathology.
De Harven spent most of his research career characterizing and isolating
murine retroviruses, using filters and ultracentrifuges to purify the
particles, and electron microscopy to monitor the level of success of
purification and to study viral morphology. His extensive publication record
led him to serve as associate editor of Virology and Cancer Research, and as
editorial board member of Scanning Microscopy, Submicroscopic Cytology, and
the Journal of Electron Microscopy Technique.
De Harven doubted the infectious HIV-AIDS model from its inception in 1984,
when Gallo presented it in four papers published together in the May 4 issue
of Science. Among the evidence supporting his conclusion, Gallo claimed to
have isolated a new retrovirus, later called HIV, from 34% of the AIDS
patients that he examined. This seemed to be a very week correlation, since
66% of the patients lacked enough virus to isolate.
As for the patients in whom Gallo did claim to obtain retroviral isolates,
the evidence appeared very weak as well. Although retroviruses infect immune
cells, they don't kill them, so de Harven viewed them as unlikely candidates
for causing AIDS. Was Gallo proposing that his new retrovirus killed host
cells when replicating? No. In the cultures where Gallo claimed replication
of the retrovirus take place, hosts cells survive indefinitely.
Also, Gallo attempted to isolate his retrovirus only from supernatants
(cell-free culture fluid) of stimulated cell cultures, rather than from
fresh patient's plasma. Why was this? Could it be that there wasn't enough
virus in the plasma to isolate it in the first place? After all, culturing
causes viral populations to increase beyond the levels found in vivo.
"If there was enough virus to isolate directly from the plasma, as the
current concept of alleged 'viral load' suggests," de Harven says, "Gallo
could have done so very easily, and it would have made his fabulous story
much more convincing."
Then there were the replication data Gallo used to demonstrate that these
samples — which he called retroviral isolates — indeed contained viruses.
Gallo added the primary "isolates" (the ones made from supernatants of
cultures to which he had added patient tissue) to virgin cultures, which he
then spiked with powerful stimulants. He showed that from the resulting
secondary supernatants he could obtain a continuous supply of "retroviral
isolates" identical to the originals, indicating that something in them had
the capacity to replicate, a fundamental feature of viruses.
But since stimulants can induce culture cells to release all sorts of
material and particles, and since both the primary and secondary "isolates"
resulted only from stimulated cultures, de Harven wondered if these
"isolates" were merely artifacts of culture stimulation.
There were other reasons to doubt that these samples were viral isolates. A
retroviral isolate should contain a single species of RNA molecule, and it
should code for the proteins in the sample.
Although Gallo showed that his "isolates" contained RNA molecules and
several proteins, including the enzyme reverse transcriptase, he didn't show
that all the RNA molecules were identical, or that any one of them coded for
the proteins. Instead, he concluded that the molecules were viral simply
because some of the proteins reacted to antibodies in the fresh (uncultured)
plasma of 88% of his AIDS patients, and because he assumed — incorrectly, de
Harven says — that reverse transcriptase is an exclusive feature of
Gallo's electron micrography data particularly intrigued de Harven, as this
involved his specialty. He says that a convincing and proper demonstration
of the existence of a retrovirus should contain two sorts of electron
micrographs. One should show retrovirus-looking objects of identical size
and appearance in uncultured patient tissue samples; the other should show
isolates of those objects, meaning samples that exclusively contain objects
having identical size and appearance to each other and the objects observed
in the tissue. Instead, Gallo presented micrographs of stimulated cultures
to which he had added AIDS patient immune cells or plasma, but no
micrographs of his "isolates." The micrographs of the cultures showed cells
along with spherical objects nearby of various sizes which Gallo claimed
where representatives of his new retroviral species.
To this day, de Harven says he can find no studies that have improved upon
Gallo's electron microscopy evidence. No one has claimed to have produced
electron micrographs of retrovirus-looking objects in the uncultured plasma
of a single AIDS patient, or to have obtained "HIV isolates" from uncultured
plasma. De Harven calls this "a totally unacceptable contradiction with the
alleged PCR measurements of 'viral load' in these patients."
Only in 1997 did researchers — two teams, in fact — finally publish
electron micrographs of so-called "HIV isolates" (Virology 230:125 and 134)
made from stimulated cultures derived from AIDS patients' blood. As de
Harven suspected, the published pictures showed mostly non-viral material,
i.e. a heterogeneous mixture of debris plus, he says, "particles that
resembled cellular fragments but had no resemblance to retroviral
The authors labeled some of these particles HIV and stressed that control
cultures — cultures unexposed to AIDS patients' blood — produced no such
particles. "But those 'control' cultures had not been stimulated and,
therefore, are questionable," de Harven says.
The Good Old Days
De Harven says when he began his career in 1953, claims of viral isolation
and causation were much more convincing, and had been so since the birth of
virology, in 1892.
That year a Russian bacteriologist discovered he could induce mosaic
disease in healthy tobacco plants by injecting them with fluid from sick
plants. Since the fluid didn't contain any of the diseased tissue, he
concluded that there was some exogenous causative agent. He showed that the
agent was microbial rather than molecular, since microbes can sustain their
populations within an organism via replication, whereas exogenous molecules
can do so only via regular replenishment from an outside source. He also
established that this microbe was invisible with the light microscope and
smaller than any known bacteria or fungi — the only known microbes at that
time — by reproducing the results after passing the fluid through special
filters with ultramicroscopic pore sizes (a process called ultrafiltration).
In 1898, a Dutch scientist found that this filterable infectious agent was
an unknown type of microbe, since it remained active even when subjected to
challenges that would destroy bacteria and fungi. He called it a virus.
Using similar techniques based on ultrafiltration, scientists found viral
causes for other diseases, including many animal and even human diseases. In
1956, Friend demonstrated the existence of a mouse leukemia (immune cell
cancer) virus by inducing the disease in adult Swiss mice with injections of
ultrafiltered tissue extracts from leukemic laboratory mice.
Scientists learned to approximate the size of viruses by using filters of
different and relatively well-calibrated pore diameters. This enabled Friend
to demonstrate that her virus, like all other viruses then shown to cause
cancer in lab animals, probably had a diameter close to 100 nanometers (nm).
Such techniques also enabled scientists to purify viral samples according
to size, to remove entities that are much larger or smaller than their
virus, in attempts to produce viral isolates. De Harven and Friend followed
this standard procedure to obtain purified samples of their retrovirus.
The next step of the standard procedure called for them to produce
micrographs of their purified samples, to see how close they had come to
ideal isolation, and to document the size and morphology of the objects in
their sample. Their micrographs showed very little else but spherical
objects of the same size (100 nm) and morphology. The objects looked exactly
like those in electron micrographs of the leukemic mouse tissue from which
they'd abstracted their filtrates. The objects also closely resembled the
viruses shown to induce cancer in other laboratory animals.
In the 1970s, purification by ultrafiltration gave way to purification by
density. This process involves placing a sample on a sucrose gel that has a
graduated density, meaning that from top-to-bottom the gel's density
increases from light to heavy. By spinning the gel in a high-speed
centrifuge — a process called ultrafugation — the sample's contents settle
at their characteristic densities, forming bands that constitute
De Harven and Friend showed that Friend's virus, like all those associated
with cancers in lab animals, banded at 1.16 grams per millimeter (gm/ml).
"Collecting the 1.16 band is a very efficient method of concentrating
retroviruses," de Harven explains. "Retroviruses concentrated by this method
look similar, under the electron microscope, to those concentrated by the
old ultrafiltration method. The degree of purity of these retroviral
concentrates depends on the amount of cellular microvesicles and debris
contaminating the initial sample which may very well sediment at the same
density of 1.16 gm/ml."
Other methods came along to help scientists to study viruses.
Growing cells in cultures and infecting them with viral isolates permitted
virologists to distinguish two general groups of viruses. When some viruses,
like those associated with polio, influenza, and herpes, replicate inside
culture cells, the cells quickly die. Other viruses, including those
associated with cancer, lack this cytolytic capacity, and replicate without
killing their hosts. The viruses associated with cancerous lab animals,
including Friend's, actually make the culture cells become cancerous, a
process called transformation or immortalization, adding support to the
contention that they cause cancer in vivo.
Virologists also developed a way to study the proteins and gene molecules —
RNA or DNA — of viruses. By using chemicals that break viruses apart, and a
process called electrophoresis that separates molecules according to
molecular weight, scientists can turn a 1.16 gm/ml retroviral isolate band
into several new bands, each one representing an isolate of a different
molecular species that composes the virus.
Logically, this process applied to a retroviral isolate should produce a
single band with RNA molecules in it, and those molecules should be
identical to each other and code for the proteins that form the other bands.
Even before the advent of this technology, Friend showed that her virus,
like the others associated with cancer, contained RNA, which is why these
viruses were called "RNA Tumor Viruses."
In 1970, scientists found that the protein forming one of the
electrophoresis bands from retroviral isolates was an enzyme capable of
reverse-transcribing RNA molecules into DNA molecules. They named this
enzyme reverse transcriptase, and gave RNA tumor viruses the new name
That same year Duesberg showed that these viruses caused cancer because
their RNA molecules contained a special extra onc-gene. This contrasted with
the cyctolytic viruses, which other scientists found contained a gene that
coded for a cytolytic enzyme responsible for destroying the host cell's
De Harven says that the collective lessons of retrovirology by 1970 led to
an important fundamental conclusions: a 1.16 gm/ml band qualified as a
retroviral isolate only after electron microscopy showed that it looked like
one, and unstimulated cultures showed that it behaved like one. This is
because on many occasions virologists obtained 1.16 gm/ml bands that flunked
the electron microscopy test, including bands that contained nothing at all
that even vaguely resembled viruses. And some 1.16 gm/ml bands that did
resemble retroviral isolates failed to demonstrate any replicative capacity
at all, and therefore qualified as isolates of non-viral objects that
In other words, everything that looks like a retrovirus is not necessarily
a retrovirus, and everything that looks like a retroviral isolate is not
necessarily a retroviral isolate.
De Harven says this conclusion holds even for 1.16 gm/ml bands that contain
reverse transcriptase, which virologists have "found not to be unique for
retroviruses, but instead a normal constituent of many cells."
No retroviral isolates from humans
The retrovirus-cancer link demonstrated among special inbred lab animals
caused scientists to wonder if retroviruses might explain some forms of
This idea initially interested de Harven. But he and some other
retrovirologists began to doubt that retroviruses could cause human cancer.
For one thing, the retrovirus-cancer link in animals was reserved for
special cases of inbred mice, chickens, and cats, and even then the link was
far from perfect. Among these peculiar laboratory animals, scientists could
usually isolate retroviruses from the subjects that had cancer, but
sometimes they could not. And sometimes they could isolate retroviruses even
in the absence of cancer. Furthermore, these retroviral isolates would not
induce cancer when injected into wild mice and chickens. Nor could anybody
isolate retroviruses from any wild animals, or even other laboratory
animals, besides special inbred cats.
Nonetheless, the entire profession in 1955 set its sights on isolating
retroviruses from humans suffering from various types of cancers. By 1970,
the examination of many cancerous human tissues resulted in some electron
micrographs that showed — mixed in with the cells and other material — a
few particles "with a vague resemblance to retroviruses," but no micrographs
of isolates. "Unfortunately," de Harven says, "the term 'virus-like
particles' kept being described in the literature, wrongly perpetuating the
notion that actual retroviruses were occasionally associated with some human
Typical retroviruses — objects of identical size and appearance, isolates
of which contain a single RNA species that codes for the proteins in the
sample — were never conclusively demonstrated for humans. "This was in sharp
contrast to the highly reproducible demonstration, by electron microscopy,
of typical retroviruses in a variety of mouse and chicken leukemias and
Gallo lowers the standards
Nonetheless, prior to HIV's official birth in 1984, Gallo published the only
three claims for isolation from human subjects of retroviruses, each one of
which he said caused a different form of human leukemia. His claims flunked
the strict criteria established previously by scientists demonstrating the
existence of retroviruses in lab animals, and causally linking those
retroviruses to cancer.
Gallo's "evidence" resembled what he'd later use to demonstrate HIV's
existence and causal role in AIDS: the presumption that all 1.16 gm/ml bands
that contain reverse transcriptase are "retroviral isolates"; the
presumption that anything that vaguely resembles a retrovirus is a
retrovirus; the use of stimulated cultures; micrographs of those cultures,
but not of the density-purified bands; the failure to obtain them from
uncultured plasma; the failure to use unstimulated cultures; the failure to
obtain these bands from most of the patients examined; the failure to
demonstrate that these bands affect virgin cultures in a way that explains
the disease (rather than transforming the culture cells, Gallo's 1.16 gm/ml
bands required that the culture cells already be cancerous); the failure to
demonstrate that his bands contained a single RNA species and that it coded
for the proteins in the band; the failure to demonstrate that the RNA
contained a gene — in this case, an onc-gene — that could explain the
"Why no micrographs of the 1.16 gm/ml bands themselves?" de Harven asks.
"Most probably, Gallo's electron microscopy results from his 'isolates' were
negative and swiftly ignored."
Yet a large segment of the biomedical profession, eager to consolidate the
hypothetical role of human retroviruses in human diseases, accepted Gallo's
new standards. Though one of his first claims was eventually discarded, the
pump was primed for the retroviral hunt among AIDS patients.
"The faith in retroviruses as pathogens assumed quasi-religious
proportions," de Harven laments. "Since electron microscopy could not
demonstrate viruses in the 1.16 bands from human subjects, we forgot about
microscopy and started relying on 'markers.' Microscopy is time-consuming
and skill-demanding. Who has time for that? Not when research funding was
getting difficult and when major pharmaceutical corporations were starting
to finance 'crash programs' for speedy answers.
"When retroviruses are legion, molecular markers provide a useful approach
to quantification probably better than direct particle counting under the
electron microscope (which I always found difficult). But without isolates,
the use of markers is methodological nonsense. 'Markers' of what? We know
that all of the so-called 'HIV markers' are totally non-specific." *