Does HIV cause AIDS?

Paul J Travers p_travers at
Fri Jun 3 10:39:40 EST 1994

In article <R.Burge.148.000CC7D3 at>,
R.Burge at (Richard Burge) wrote:

> Maybe HIV doesn't cause AIDS. Crazy notion - or is it?
> I was at a lecture a couple of days ago by Kary Mullis, 1993 Nobel 
> Chemistry Laureate (inventor of PCR).

I was at the same lecture, and I wasn't impressed; I left with the
overwhelming impression that the man is either a fool or a charlatan.  His
sole argument against the proposition that HIV causes AIDS revolves around
his observation that he couldn't find a paper explicitly stating that HIV
causes AIDS (and since there are so many hungry post-docs out there, if it
could be published it would have been published and the post-doc would no
longer be hungry).  Now there may not be such a paper (and I wouldn't be
surprised if there weren't) but the failure to prove a hypothesis is not
quite the same thing as proving it false.  A key question here is "What
constitutes proof?", short of infecting someone with a purified preparation
of the virus and standing back to see what happens.  I note in passing that
Dr Mullis, for all his claims that HIV is harmless, would not volunteer to
be the guinea pig.  Now, I think that the body of evidence consistent with
the hypothesis that HIV causes AIDS is pretty impressive, and I know of no
evidence to the contrary.  If anyone wants to present any such evidence I
would be very interested.

  Now onto the Mullis Hypothesis.  In may ways this is a hybrid of the
Jerne Idiotypic Network Hypothesis (but substituting retroviruses for
idiotypes) with Rolf Zinkernagel's ideas on "immunological exhaustion". 
Mullis suggests that we are awash with retroviruses and that most of these
are harmlessly integrated.  Some of these retroviruses will be
T-lymphotropic and on occasion, one such retrovirus will be activated when
its host T cell is activated and proliferates.  This offers a new
immunological stimulus, activating further T cells, which may activate
further retroviruses and so on, ad infinitem, or ad nauseam, whichever
comes the sooner.  As you can see, if each retrovirus is unique to an
individual T cell clone then it is equivalent to an idiotype, and if it is
shared between a number of clones then it is the equivalent of a regulatory
idiotype.  Now, we know that a proliferating network of anti-idiotypic
responses neither overwhelms nor exhausts the imumune system (and I believe
that Jerne got HIS Nobel prize for suggesting one mechanism why it should
not), and I can think of no reason why a proliferating network of
anti-retroviral responses shouldnot behave in exactly the same way.

  You may argue that a retrovirus presents more epitopes to the immune
system than an idiotype does, but I don't think this is relevant; the
Mullis hypothesis does not suggest that all T cells harbour retroviruses so
that if, say, a retrovirus could encode 50 discrete epitopes (remember,
this is T cells), then at least one in 50 cells have to be infected to give
the same antigenic challenge as the antiidiotypic response (lots of
handwaving going on at this point). 

		It's also hard to imagine a model for immunological exhaustion by
retroviruses that is specific only for the CD4 subset of T cells - CD8
cells can also respond to retroviral antigens and can be exhausted.  Maybe
I'm just not thinking clearly enough but I do not understand how the Mullis
model could lead to the subset specific depletion seen in AIDS.

  OK, that's the theoretical aspects out of the way, now what about the
practical.  Firstly, Mullis proposes that HIV is essentially a passenger,
that it is one of a myriad of retroviruses that infect us all and is only
found in AIDS patients because (a) they are immunosuppressed and as a
result carry lots of viruses and (b) there is a test for it and not for any
of the other retroviruses.  Well, Mullis was hot on asking for evidence in
his talk, I'd like to know the evidence for his statement that there are
hundreds of retroviruses infecting us.  I don't know of any evidence that
says that AIDS patients carry  a very extensive viral load; certainly, not
every person with AIDS has CMV, one of the viruses originally postulated to
cause AIDS.  I would doubt that there are general infections of flu, polio,
mumps, measles, vaccinia, rabies, HTLV-I and II (if you want retroviruses,
but who said that immunosuppression favoured only retroviruses).  In fact,
isn't it the case that the ony virus consistently isolated from AIDS
patients is HIV?

  Secondly, the "retroviral catastrophe" theory is not dependent on HIV per
se.  So, unless HIV is in fact extremely common (and the frequency of HIV+,
at least at the antibody level must be known by the bloodbanks and is, I
believe, relatively low, at least in the UK), you would expect to find
individuals who had been exposed to the myriad other viruses but NOT to
HIV, but who would also show the symptoms of AIDS.  Where are these
individuals?  By the same token, since we only eliminate HIV+ blood from
the blood supply (well, hepatitis too), and not any of these other
retroviruses (for which we have no means to test), unless there were a
surprising correlation between the presence of other retroviruses and the
presence of HIV, we should still see transfusion related AIDS cases, all of
which would be HIV(-)ve.  Where are these cases?

		Now, as for the transmission of these retroviruses;  Mullis suggests that
the appearance of AIDS in habitues of the bathhouses was a direct
consequence of the numbers of contacts made by these individuals.  Not
necessarily sexual contacts, but simply a result of a local high population
density in a hot and humid environment being conducive to the (probably
respiratory) spread of infectious agents.  Well, I don't know how familiar
Kary Mullis is with the London Underground in the rush hour, but to me it's
a Dantean vision of Hell.  One is pressed into a small space with hundreds
of strangers, all coughing and sneezing, half of whom forgot to wash that
morning while the other half put on too much aftershave.  If you're lucky,
the person whose face is inches from yours will not have eaten garlic the
previous evening.  Unfortunately, a large proportion of the population of
Greater London travel under such conditions 5 days a week.  If there really
were a variety of retroviruses out there, with respiratory spread, London
would be a charnel house of AIDS victims.  And this is not the case.  (I
understand the public transport systems in Tokyo are even more crowded,
anybody got any figures for the incidence of AIDS in Tokyo?)  

  Let's face it, this retrovirus catastrophe theory just doesn't make
sense. Nice try, but no cigar. 

  Finally (for those of you who have persevered so far, it's almost over),
let's address the question of the change in latency.  Well, the simplest
explanation I can think of for a change in the latency of AIDS goes like
this.  In the beginning, people with AIDS were identified because they got
sick - they were relatively late in the disease progression and the average
time from diagnosis to death was short.  Later on, after HIV was identified
as a probable cause,  patients were identified as being HIV+, but at an
early point in the disease progression, and these individuals have a longer
mean time from diagnosis to death.  So the statistical increase in latency
may be a consequence of earlier diagnosis, rather than a change in the
disease process itself.  It's worthwhile noting in passing that if HIV were
irrelevant to AIDS then there would be no correlation between HIV
positivity and mortality from AIDS, and no effect of early HIV diagnosis on
mean time to death in AIDS patients.  Mullis may make a big play of the
change in latency, but I think it argues against him.

	OK, sorry for the length of this and the extent of the ramblings, but I
could not let Mullis' statements go unchallenged.  I think he (and the
Sunday Times, who sponsored the event) has behaved irresponsibly, and the
cachet of a Nobel prize gives his views greater weight with the
non-scientist, and often poorly scientfically-informed public.  I'll be
happy to enter into discussion on any of the points raised, particularly
with anyone more informed than I.

  All views expressed are mine;  I have no idea what the views of the ICRF
or Birkbeck College are in these issues.

                               Paul Travers

Paul J Travers                          phone : +44-(0)71-631-6262 (office)
ICRF Structural Biology Unit             "       "       "    6268 (lab)
Birkbeck College                          fax : +44-(0)71-436-8918
Malet Street
London WC1E 7HX                         email :  p_travers at
England                                    or :  paul at

More information about the Immuno mailing list