ISIS Report 03/04/07
Beyond the HIV-Causes-AIDS Model
Dr. Mae-Wan Ho follows the trail of how a bad mathematical model
has misled AIDS policies with disastrous consequences, and recent attempts to
find a better model
A fully
referenced version of this article is posted on ISIS members’ website. Details
here
An electronic version of this report, or any other ISIS report, with full references,
can be sent to you via e-mail for a donation of £3.50. Please e-mail the title
of the report to: report@i-sis.org.uk
Alternatively, order you copy of ISIS Report Unraveling
AIDS and get a free CD of all new articles and others on AIDS from SiS
archives to bring you right up to date http://www.i-sis.org.uk/onlinestore/books.php#236
“A model lacking in predictive and explanatory power”
“More than 20 years into the AIDS era, it has become increasingly clear
that the current single-virus causation model is lacking in predictive and
explanatory power.” This is how Rebecca Culshaw, assistant professor of mathematics
at the University of Texas, Tyler, USA, begins her most recent paper published
in the winter 2006 issue of the Journal of American Physicians and Surgeons [1].
Culshaw has announced why she “quit HIV” in March 2006 [2] (On
Quitting HIV, this series), and this had lured me onto the fascinating trail
of how a bad mathematical model has misguided AIDS policies for so many years
worldwide, and more importantly, alerted me to recent attempts to find alternative,
more realistic models.
Perhaps quitting HIV is
not the same as quitting AIDS, as the world is desperately in need of a good
model, to save lives and end human suffering on a gigantic scale.
AIDS disease
is generally characterised by a decline in CD4+ T lymphocytes circulating
in the blood, which are responsible for cell-mediated immunity. As a result,
the patient becomes susceptible to opportunistic infections (those affecting
weakened immune systems) such as tuberculosis, pneumonia, meningitis, and
other diseases caused by parasites, bacteria and viruses that can enter and
multiply in the cells of the body.
But models that assume the
human immunodeficiency virus (HIV) plays a central role in disease progression
run into considerable difficulties. If the decline in CD4+ cells is due to
HIV killing the cells, then there should be a correlation between the ‘viral
load’, which estimates the amount of virus in the body, and the CD4+ cell
count. But that is not the case. CD4+ cell count is not a reliable indicator
of disease progression at all, nor for that matter is viral load [3] (Chapter
2, Unraveling AIDS,
ISIS Report), and they bear little relationship to each another. This has
been confirmed in a recent study on untreated HIV+ individuals [4].
Although higher viral loads
are associated with greater CD4+ cell decline, only a very small proportion
of CD4+ cell loss, about 4 – 6 percent, is influenced by viral load. The authors
reporting the new study called for future efforts [4] “to delineate the relative
contribution of other mechanisms.” In short, as Culshaw states [1]: “It has
been extremely difficult to construct a realistic theoretical model of immune
suppression that is entirely mediated by HIV.”
Why is it important to have a realistic model of the disease?
A realistic model not only can predict how the disease will progress, it can
also help in developing effective treatment and prevention. Since the discovery
of HIV, mathematical models have been constructed precisely for those purposes:
to determine the rates of progression to AIDS, to define optimal drug regimens
for therapy, to develop vaccines, and as a desperate last resort, microbicide
vaginal gels [5] (Concentrating
Exclusively on Sexual Transmission of HIV is Misplaced, this series). However,
the vast majority of the models lack predictive power because the mechanisms
of disease and the fundamental nature of the immune system are both poorly understood.
Meanwhile, the consequences of models based on a wrong hypothesis are all too
clear, as Culshaw has starkly stated [2].
The toxicity of HAART treatment is now widely accepted [3] (Unraveling AIDS, Chapter
7). The scandal of toxic drugs being tested on defenceless foster children in
New York City and mothers and babies in Uganda [6-8] (US Foster Children Used in AIDS Drugs
Tests; Guinea Pig Kids in AIDS
Drugs Trials ; NIH-Sponsored
AIDS Drugs Tests on Mothers and Babies ; SiS 27) was widely publicised
more recently in Harper’s Magazine [9]. This has reopened the acrimonious
debate [10] between AIDS ‘dissidents’ and the orthodox community of researchers
and activists led by Robert Gallo, the controversial co-discoverer of HIV. The
litany of vaccine failures has reached epic, controversial proportions [3] (Unraveling AIDS, Chapters
9-13); and the third large-scale clinical trials of anti-HIV gels has just been
terminated because it was not only ineffective, but actually increased the risk
of HIV infection [5]. There are many compelling reasons to confront the bad
model itself.
The Ho/Shaw model on why “hit hard hit early”
The model of HIV causes AIDS disease that has come to dominate global policies
on AIDS, from diagnosis to therapy and prevention, is barely 12 years old. It
was created in two high profile papers published in the 12 January 1995 issue
of the journal Nature [11, 12]. Two research teams, led respectively
by David Ho of the Aaron Diamond AIDS Research Centre NYU School of Medicine,
New York, and George Shaw of the University of Alabama at Birmingham, used experimental
antiretroviral drugs to follow how HIV viral load and CD4+ cell counts change
after drug administration. From the changes, they estimated the rates of viral
replication and elimination from the body as well as the rates at which CD4+
cells are killed and replaced by cell proliferation.
The results were astonishing;
they were touted as giving a radically new understanding of HIV infection,
one in which the immune system is in a constant battle with HIV from the moment
of initial infection. As the distinguished late mathematician Serge Lang,
a prominent AIDS dissident wrote [13]: “These papers largely provided the
justification for the new phase of protease inhibitor and cocktail treatments,
as well as for the expanded use of surrogate markers such as “viral load”
and CD4 counts for AIDS disease. Each of these represented a significant departure
in terms of HIV/AIDS diagnosis, maintenance, treatment, and epidemiological
reporting.”
In the Ho study [11], a
protease inhibitor code named ABT-538 was given at 600 to 1 200 mg per day
to 20 HIV+ individuals whose pre-treatment CD4+ lymphocyte counts ranged from
36 to 490 per mm3 and viral load, measured by a new quantitative
branch polymerase chain reaction from 15 – 554 x 103 virus particles
per ml.
Following treatment, every
patient had a rapid and dramatic decline in plasma viral load over the first
two weeks, between 11 and 275-fold reduction, with a mean of 66-fold, i.e.,
a 98.5 percent drop. The initial decline was assumed to be exponential, allowing
the half time of viral decay (time it takes for half of the virus particles
to disappear) to be estimated as 2.1 + 0.4 days. That showed HIV-replication
must be “highly productive”, the authors claimed; and the virus particles
were cleared as fast as they were produced. In other words, a steady state
standoff was established in the body, so that the viral load measured at any
time remained roughly the same. The estimated minimum production rate – the
same as the minimum clearance rate - averaged 0.68 + 0.13 x 109
virus particles per day, which is really quite modest, considering that each
infected cell can produce a hundred virus particles.
The paper was heavily criticised.
The estimates depended on the assumption that drug treatment does not affect
viral clearance, and that there was a pre-existing steady state between viral
production and viral clearance, regardless
of the amount of virus in circulation. Curiously, the estimated
viral clearance/production rate bore no relationship to the initial viral
load or to the CD4+ lymphocyte count, which was difficult to reconcile with
the idea that the virus was killing the CD4+ cells by invading the cells to
replicate and burst the cells. In that case, the more virus particles and
the more cells, the higher should be the production/clearance rate.
After ABT-538 treatment, CD4+ lymphocyte counts rose in each
of 18 patients that could be evaluated. Some increases were dramatic and others
quite modest. From the slope of the line depicting the rise in CD4+ lymphocyte
counts assuming an exponential increase, a doubling time of about 15 days
was estimated during the (assumed) pre-treatment steady state. The slopes
were inversely correlated with
baseline CD4+ cell counts, however, which too was difficult to explain. In
patients with lower initial CD4 cell counts, more prominent rises were obtained.
Nevertheless, the authors claimed: “This demonstrates convincingly that the
CD4+ lymphocyte depletion seen in AIDS is primarily a consequence of the destruction
of these cells induced by HIV-1, not a lack of their production.” They explained
that such an inverse correlation would be expected if T-cell proliferation
were governed by some kind of homeostatic mechanism. From the inverse correlation,
it was estimated that the minimum number of CD4+ cells in blood produced or
destroyed each day ranged from 4.3 x 106 to 109 x 106,
with a mean of 35.1 x 106. As the blood lymphocyte pool is about
2 percent of the total population, the overall CD4+ lymphocytes turnover in
the patients was calculated to vary from 0.2 x 109 to 5.4 x 109
cells per day, with a mean of 1.8 x109 cells per day. This number
of cells was about the same as the number of putative viruses produced (and
cleared) each day, far too many cells killed for the number of viruses produced.
Things didn’t add up.
The increase in CD4+ lymphocyte
counts following ABT-538 administration was also modelled linearly, and using
the same arguments as for the decline in viral load, the minimum estimates
of total CD4+ lymphocytes production or destruction rates at baseline were
determined to vary between 0.1 x 109 to 7.8 x 109 cells
per day with a mean of 2.6 x 109 cells per day, sufficiently close
to the estimate above.
The authors commented that the CD4+ lymphocyte depletion seen
in advanced HIV-1 infection “may be likened to a sink containing a low water
level, with the tap and drain both equally wide open.” As the regenerative
capacity of the immune system is not infinite, it is not difficult to see
why the sink eventually empties (when CD4+ cells are all depleted).
Now comes the crucial conclusion
that has justified the “hit hard, hit early” [14] strategy of HAART that has
gone so disastrously wrong for otherwise healthy HIV+ individuals: “It is also evident from this analogy that our primary
strategy to reverse the immunodeficiency ought to be to target virally mediated
destruction (plug the drain) rather than to emphasize lymphocyte reconstitution
(put in a second tap).”
And: “We believe our new
kinetic data have important implications for HIV-1 therapy and pathogenesis.
It is self evident that, with rapid turnover of HIV-1, generation of viral
diversity and the attendant increased opportunities for viral escape from
therapeutic agents are unavoidable sequelae. Treatment strategies, if they
are to have a dramatic clinical impact, must therefore be initiated as early
in the infection course as possible, perhaps seen during seroconversion…”
The Shaw
study [12] used the protease inhibitors ABT-538, L-735.524, or the non-nucleoside
reverse transcriptase inhibitor Nevirepine on a total of 22 patients, as part
of a phase I/IIA clinical trial, and came to the same conclusions. In addition,
it found drug resistant mutant viruses in all subjects soon after treatment
started. The lowest point of viral load was at two weeks in all subjects after
treatment started, when the CD4+ cell count rose to a peak. Thereafter, viral
load increased rapidly, despite increased drug dosage, and by week four, 100
percent of the virus in blood was drug resistant. The CD4+ cell counts dropped
more slowly, and were back to baseline within 6-20 weeks.
Critics faulted the Shaw study for the same unwarranted assumptions
that underlie the Ho study. Neither study included a control group. The clinical
outcomes of the drugs on the patients were not reported, so it was impossible
to tell whether the patients benefited from the transient reduction in viral
load or the transient increase in CD4+ cells. The mathematical model had no
contact with the observations other than dubious fitting of a straight line
through two or three data points [15].
The Ho/Shaw model began to
unravel almost as soon as it was proposed, but the “hit hard hit early” HAART
approach continued at least until 2001 when the US government’s expert panel
on anti-HIV therapy finally recommended restricting the prescription of anti-HIV
drugs for as long as possible for people without symptoms, on account of the
serious side effects [3] (Unraveling AIDS, Chapter
7).
“The final nails in the coffin” of Ho/Shaw models
Mario Rodoerer at Stanford University Beckman Center, writing in News
and Views of the February 1998 issue of Nature
Medicine commented [15] that two papers published in the same issue
[16, 17] “provide the final nails in the coffin for models of T cell dynamics
in which a major reason for changes in T cell numbers is the death of HIV-infected
cells [i.e., the Ho/Shaw models].”
The papers presented extensive data on the remodelling of
the T cell compartment in HIV-infected individuals after treatment with HAART.
Throughout the early stages of HIV infection, CD4+ cells decline, whereas
the total CD8+ cells expand. However, the application of flow cytometry techniques
that accurately identified subsets of T cells showed that this increase in
CD8+ cells is made up entirely of memory and activated T cells, while naïve
T cells (precursor of memory and activated T cells) declined at the same rate
as naïve CD4+ cells (precursor of memory and activated CD4+ cells). Activated
T cells are found only in peripheral tissues - the spleen and lymph nodes
– and their expansion in the blood in HIV-infected individuals indicated an
active immune response even during the later stages of disease.
Within weeks after starting HAART, there were significant
increases in the number of B cells, and of CD4+ and CD8+ cell in the blood,
but these were only memory cells that can maintain long-term residence in
lymph nodes, and not naïve T cells, which do not dwell in lymph nodes and
do not immediately respond to HAART.
Essentially, the studies
provided evidence for the ‘redistribution hypothesis’: the increase in CD4+
cell counts observed shortly after the start of HAART are T lymphocytes redistributed
from the lymph nodes, and not produced by cell proliferation. During active
viral replication and the concomitant cellular immune response, a large number
of B and T cells may be trapped in peripheral sites (for example, by antigen,
cytokine or chemokine signals). After initiation of HAART, when HIV is effectively
removed from the system, the immune response begins to resolve and cells pour
out of the inflamed lymph nodes back into the blood.
The first study [17] suggested
that the degree of T lymphocyte trapping increases as disease progresses.
That would explain why the response to HAART tends to be greater in individuals
with lower CD4+ cell counts.
Functional recovery of the T cell compartment is only complete
when the repertoire of T cell receptors is restored, so that potentially all
antigens can be recognized. The decrease in naïve and memory T cell populations
during disease progression means that the repertoire becomes increasingly
restricted, finally resulting in immunodeficiency.
The second study [18] confirmed earlier findings that the
T cell receptor repertoire in HIV-infected individuals is significantly different
from the normal distribution found in healthy adults. This is due to a loss
of unique T cell clones and an expansion of antigen-specific clones caused
by an over-representation of certain receptor types.
In individuals responding to HAART, the number of naïve T
cells slowly increases over a six-month period after initiation of HAART and
a reconstitution of the T cell repertoire can take place (but see later).
Notably, this reconstitution occurs only in individuals who show reductions
in viral load in response to HAART. It is also likely that failure of HAART,
which occurs in many patients over time, will also be accompanied by a re-initiation
of cell losses and repertoire restriction.
AIDS and an over-stimulated and unbalanced immune system
The use of radioisotope labelling has enabled researchers to identify different
populations of T lymphocytes in the human body [19]. There are long-lived and
short-lived cells, and the size of the total T lymphocyte pool appears to be
regulated mainly at the level of the long-lived cells. During the course of
an antigen-driven cell proliferation response, some T cells differentiate into
effector cells that clear the antigens from the body, and typically have a short
life span. Others become memory T cells, which, by contrast, are long-lived
and serve as reservoirs for subsequent activation by antigen to proliferate
and produce effector cells. Naïve T cells also have a long life span. In advanced
HIV-1 infection, a much higher proportion of T cells are short-lived, compared
to healthy controls, and effective HAART tends to restore the values towards
the normal. Advanced HIV-1 infection greatly reduces the percentage and total
number of CD4+ cells that are long-lived. Because these cells represent the
regenerative source of newly formed CD4+ effector T cells, their loss may underlie
the immunodeficiency of HIV-1 disease. These abnormalities may not be present
in early HIV-1 infection and may represent a marker of disease stage.
However, many questions remain unanswered [20]. Why is HIV
so uniquely powerful, among chronic viruses, in inducing a chronic state of
immune activation? And why is the HIV- induced immune activation is so disruptive
of the proper overall functioning of the immune system?
Of course, there remains
the lingering doubt that HIV is not actually causing the disease
A radical new model is needed
None of the models so far has taken into account the role of nutrition
in AIDS or AIDS-like diseases, and the ability of good nutrition to reverse
or delay disease progression [3] (Unraveling AIDS, Chapters
15-17). In particular, AIDS is a disease in which the immune system is out
of balance, not only in being chronically activated, but also in the predominance
of the humoural (type 2) at the expense of cellular (type 1) immunity [3]
(Unraveling AIDS, Chapter
12).
All HIV models so far have considered the CD4+ cells
as a single entity. But it has been known for some that the pool of CD4+ cells
(commonly known as T-helper or Th cells) contained two different subsets:
Th 1, responsible for cell-mediated immunity and Th 2, responsible for extracellular
or humoural immunity. The majority of the CD4+ Th 1 cells reside in the peripheral
blood and it is their depletion that occurs in the progression to AIDS [1
and references therein]. Th2 cells reside mainly in the bone marrow and to
a lesser extent in the lymph nodes, and do not appear to become depleted in
the progression to AIDS. If anything they have been observed to increase [21].
As AIDS progresses there
appears to be a gradual shift from Th1- to Th2-dominance, which is why patients
experience mainly fungal and mycobacterial infections, but very few “classical”
bacterial diseases. Furthermore, elevated levels of antibodies, including
autoantibodies, are characteristic of all AIDS patients, as consistent with
an increase in Th2 subset. Contrary to what one might expect, HIV is expressed
primarily in Th0 (precursor of Th1 and Th2) and Th2 cells and is scarcely
Expressed in the Th1 subset [22]. Yet it is the Th1 cells that are depleted,
whereas the cells in which HIV prefers to reside do not decrease. So what
mediates the Th1 to Th2 shift, and how can it be prevented or reversed so
as to restore balance to the immune system?
Culshaw [1] suggests using bifurcation theory, a branch of
mathematics that deals with changes in critical parameters that determines
major or abrupt changes, such as the commitment of Th0 to become either Th1,
or Th2.
One crucial component in the Th1 to Th2 shift is the release
of nitrous oxide (NO) from the cell-mediated arm of the immune system [23].
NO can diffuse though cell membranes without the help of receptors in cell-cell
communication, and nitrogen oxides are regulated by the oxidative state of
the immune cells. Excessive oxidation negatively affects immune function through
the production of cytokines from the immune cells. Oxidative processes are
counterbalanced by reduction, which is accomplished by sulphur-containing
molecules that serve as electron donors, the main one is glutathione, a tripeptide
consisting of cysteine, glutamine and glycine. Glutathione is found in both
the reduced (GSH) and the oxidized (GSSG) form. The ratio of GSH:GSSG has
been shown to be important in regulating Th1/Th2 balance [24, 25]. If the
GSH:GSSH ratio declines, Th2 cells are preferentially made from Th0, thus
resulting in Th2 dominating at the expense of Th1 cells.
HAART causes a transient
increase in T-cell counts in the peripheral blood because it damages B-cells
as they mature and disrupts antibody production. Unable to make contact with
antibody-producing B-cells in the bone marrow, the CD4+ Th2 cells return to
the peripheral blood. So, although CD4+ T cell count increases, the cells
are ineffective against opportunistic infections. This gives rise to the phenomenon
of ‘immune reconstitution syndrome’, in which patients experience the “irony”
of an increase in opportunistic infections after initiating HAART therapy
[26].
Culshaw sketches out an alternative mathematical model based
on the GSH:GSSG ratio and Th1/Th2 balance, which are crucial in the development
of AIDS, and proposes that HIV itself need not even be included as a variable.
The proposed model tracks the Th0, Th1 and Th2 subsets of the T-cell pool
over time, with the ratio of GSH:GSSG as a possible bifurcation parameter.
As GSSG increases, and the ratio declines, a greater proportion of the Th0
cells mature into Th2 cells and are diverted from the Th1 pool. Such a model
could enable researchers to determine the critical ratio of GSH:GSSG below
which a shift to Th2-dominance occurs.
There are several advantages to such a model. First, it replaces
viral load measurements, which have not been shown to have good clinical predictive
value as a therapeutic endpoint. Second, determining a critical ratio of GSH:GSSH
rather than a critical value gets around the problem of variability among
individual patients. Finally, the model explicitly considers the Th1/Th2 ratio,
an important measure in the progression to AIDS that has been largely neglected
in theoretical modelling.
Circumstance evidence in favour of such a model is that selenium
and other antioxidants appear to be effective in preventing and treating AIDS
[3] (Unraveling AIDS,
Chapter 17). Another advantage of Culshaw’s new model is that it can make
direct contact with nutritional status, an important determinant in disease
progression.
Other ISIS articles about HIV and AIDS
| 
