June 27, 1995

The editor of Nature, John Maddox, has issued a published invitation to "Peter Duesberg and his associates... to comment" on two new studies by Wei et. al.[1] and Ho et. al.[2] that he feels lend strong support to the hypothesis that HIV causes AIDS [3]. Maddox credits us for having identified two paradoxes of this hypothesis, (i) "Duesberg was quick to point to a paradox... [that] there was no dramatic increase of the frequency of infected T-cells as infection gave way to overt disease", and that (ii) "Duesberg is right to have argued all along that the ususally slow decline of CD4+ cells [T-cells] is not consistent with... a specific cytotoxic viral mechanism." [3]

According to Maddox, "the new divelopments are (or should be) an embarrassment for Duesberg", because they "resolve the paradox". But we do not see any reason why a scientist should be embarrassed for having pointed out paradoxes in the past, which ever way these paradoxes are subsequently solved. We also object to rhetoric personalizing a scientific debate. However, it is embarrassing that in the name of science clinical, public health, journalistic, and political decisions have been made in the past, based on a hypothesis that-we all agree now-was unproven at that time.

Since the HIV-AIDS hypothesis makes many assumptions that are paradoxical, if not bewildering, for pre-HIV virologists, and since the new studies do not clearly define the HIV hypothesis, we shall first state the hypothesis and then explain why, in light of these "new" studies, it remains paradoxical.

In 1984 it was proposed that the retrovirus HIV can cause such diametrically different diseases as Koposi's sarcoma, pneumonia, dementia, diarrhea, and weight loss [4,5]. All of these diseases and over two dozen more are now collectively called acquired immunodeficiency syndrome (AIDS) [6], if antibody to HIV is present. But many of these diseases, including Kaposi's sarcoma, lymphoma, dementia and weight loss, are neither consequences of, nor consistently associated with, immunodeficiency [7,8]. For example, Kaposi's sarcoma and dementia have been diagnosed in male homosexuals whose immune systems were normal [9-13]. As a cause of these diseases HIV was proposed to follow an entirely unprecedented course of action:

1) HIV was proposed to cause immunodeficiency by killing T-cells. But retroviruses do not kill cells [14,15].

2) Within weeks after infection, HIV would reach moderate to high titers of 10-10^4 infectious units per ml blood [16], sufficient to induce antiviral immunity and antibodies (a positive "AIDS-test"). According to Shaw, Ho and their collaborators, HIV activity is "rapidly and effectively limited" by this antiviral activity [17,18]. Prior to antiviral immunity, HIV would neither kill T-cells nor cause AIDS [16,19]. But all other viruses are primarily pathogenic prior to immunity; the reason vaccination protects against disease. Not one virus exists that causes damage only after it is neutralized by antiviral immunity [20,21].

3) On average 10 years after HIV is neutralized, the virus is postulated to cause AIDS diseases [5,22]. But all other viruses typically cause disease within days or weeks after infection, because they replicate exponentially with generation times of 8 to 48 hours [20,23,24].

4) As a consequence of antiviral immunity, the virus titer is undetectably low prior to and even during AIDS [25-29]. Only in rare cases HIV titers are as high as in asymptomatic, primary infection [16,30]. But in all other viral diseases the virus titer is maximally high when viruses cause disease [20,21].

5) Antiviral immunity would typically restrict HIV-infected lymphocytes to less than 1 in 500-- prior to and even during AIDS [14,26,27,30-32]. But all other viruses infect more cells than the host can spare or regenerate when they cause disease [20,21].

The hypothesis fails to shed any light on the causation of non-immunodeficiency AIDS diseases, like Kaposi's sarcoma, dementia, lymphoma and weight loss, which make up 39% of all American AIDS cases [8,33]. Today this HIV-AIDS hypothesis stands unproven and has failed to produce any public health benefits [34-36].

The new studies are claimed by two News and Views articles from Maddox [3] and Wain-Hobson [43] to resolve the paradox, (1) how HIV kills T-cells, (2) how HIV causes AIDS, and (3) why HIV needs 10 years to cause AIDS. But we argue that the new studies have failed to resolve any of these paradoxa; in fact they have added new ones:

(1) Until HIV appeared on the scene, retroviruses did not kill their host cells. This is the reason they were considered possible tumor viruses. Since retroviruses integrate their genes into the chromosome of the host, they can only replicate as long as the host survives integration and remaines able to express integrated viral genes. Therefore a cytocidal retrovirus would be suicidal. Indeed, HIV proved to be non-cytocidal. It is mass-produced for the "AIDS-test" in immortal T-cells in culture at titers of 10^6 infectious units per ml [37,38]. Luc Montagnier and others have confirmed that HIV does not kill T-cells [39-42]. Hence the claim that HIV causes AIDS by killing T-cells is paradoxical.

The new papers have indeed resolved this paradox by shifting the paradigm: According to Maddox, T-cells "that harbour virus will be killed off very soon"-not by HIV-but by the immune system. Also consistent with a non-cytocidal virus, Wei et al. report that "the average half-life of infected PBMCs [peripheral blood mononuclear cells] is very long and of the same order of magnitude as the half-life of uninfected PBMCs". But, paradoxically, the same investigators also report that "the life span of virus-producing cells is remarkably short (t1/2=2+/- 0.9 days)", although these cells are in the same system as their long-lived HIV-infected peers [1]. Ho et al. state that "there is virus-and immune-mediated killing of CD4 lymphocytes" [2]. According to the News and Views article by Simon Wain-Hobson, "an intrinsic cytopathic effect of the virus is no longer credible" [43].

It is consistent with this "new view of HIV" that there is no correlation between virus titers and T-cell counts in the patients that Wei et.al. and Ho et. al. have studied. In some of Ho et. al.'s patients, i.e. #303 and #403, a 100-fold variation in virus titers corresponds to no changes in T-cell counts. In Wei et al.'s patients 100-fold variations in virus titers correspond to only 0.25 and 3-fold variations in T-cell counts-hardly a correlation to prove that HIV kills T-cells.

Since HIV is no longer viewed as a T-cell killer, the above paradox is solved. However, if T-cell killing via antiviral immunity were the cause of AIDS, we would have a bigger HIV-AIDS paradox than before. Since only 1 in 500 T-cells are ever infected, and most of these cells contain latent HIV not making viral proteins [25,26,30,44], only less than 1 in 500 T-cells could ever be killed by antiviral immunity.

(2) Until HIV appeared on the scene, the patogenicity of a virus was a direct function of the number of virus-infected cells: the more infectious virus there was, the more cells were infected, and the more pathogenic an infection was.

But in typical AIDS patients HIV is so rare, that even leading AIDS retrovirologists from the US, like Robert Gallo, and in the UK, like Robin Weiss, failed for years to isolate HIV from AIDS patients [45,46]. Likewise, virus-infected cells are so rare that they could not be found by George Shaw, the senior investigator of the new study by Wei et.al., Gallo and their collaborators in most AIDS patients [27]-- until the rare proviral DNA could be amplified with the polymerase chain reaction (PCR) [31,44,47].

Although the new studies never mention the percentage of infected T-cells, Maddox confirms the status quo: "the scarcity of T-cells from which virus can be recovered in test-tube experiments is consistent with the assertion that the immune system is in overdrive from the onset of infection by HIV". But the new studies claim on average 10^5 of "free virus" [1] of "plasma virion" per ml blood [2] in AIDS patients. That should be enough virus to eliminate all remaining T-cells of these patients, 10^5 per ml, within the two days HIV needs to replicate [48]-- unless, as Maddox suggest, the "new techniques for assaying the low levels of virus involved were also necessary" [3] (amplifying viral RNA with the polymerase chain reaction) possibly because no infectious HIV could be detected by conventional infectivity tests.

Indeed, Wei et. al. acknowledge "substantial proportions of defective or otherwise non-infectious virus". "To determine whether the viral genomes represented in total viral nucleic acid correspond to infectious virus..." they had to resort to the same techniques that the "old HIV hands", as Wain-Hobson calls them [43], had used to isolate HIV fom rare infected lymphocytes of AIDS patients: "We cocultivated PBMCs... with normal donor lymphoblasts in order to establish primary virus isolates". Shaw together with some of the investigators of Wei et. al. had shown in 1993 how to convert "plasma viral RNA" to infectious virus. They concluded that the "quantitative competitive PCR" is "as much as 60,000 times more sensitive" [49] than infectious virus [16,19]. Divide 10^5 "plasma viral RNA" units by 60,000 and you have 1.6 infectious units per ml, a number that is consistent with numerous previous reports (see above). Ho and a different group of collaborators just published a paper in which they show that over 10,000 "plasma virions", detected by the "branched DNA signal-amplification assay" used in their Nature paper, correspond to less than one (!) infectious virus [50].

In other words, there is no evidence for infectious virus in Wei and Ho et. al.'s patients. Wei and Ho et. al. had apparently detected non-infectious virus that had been neutralized by "the immune system [that] reacts hyperactively to HIV infection"-just as Maddox suggests. Infectious virus was only obtained by activating latent HIV from a few infected cells out of millions of mostly uninfected cells from a given AIDS patient. Such virus activation is only achieved by growing cells in culture away from the hyperactive immune system of the host, just as the "old HIV hands" used to do it, when they tried to isolate HIV from AIDS patients [45,46]. Thus the paradox of too few viruses to cause immunodeficiency remains unresolved.

In view of the evidence that there are no more than 1.6 infectious HIVs per ml blood in Wei's and Ho's patients, one wonders whether the 10^5 viral RNAs per ml are real or are an artifact reflecting inherent difficulties in quantifying the input number of "plasma viral RNA" molecules after many rounds of amplification by the PCR. The problem with the quantification of the input RNAs-after 30 to 50 rounds of amplification by the PCR [51]-- is like calculating the number of original settlers in America, from the current number of Americans and their current growth rates. But even if the 10^5 "plasma viral RNAs" per ml were real, it is hard to guess where they came from in view of "the scarcity of T-cells from which the virus can be recovered..." acknowledged by Maddox [3].

However, the apparent lack of infectivity of the "free virus" or "virions" [2] resolves the paradox of the coexistence of 10^5 T-cells with 10^5 plasma viral RNAs per ml blood in Ho et al.'s and Wei et. al.'s AIDS patients [1]. Even HIV cannot kill T-cells that it cannot infect. The fact that over 99% of T-cells in persons with AIDS are not infected by HIV [14,26,27,31,32,44], is definitive evidence that there is no infectious HIV in typical AIDS patients. Clearly, in AIDS patients with 1.6 infectious HIV units per ml, something other than HIV must cause AIDS.

In earlier efforts to resolve the paradox, that there is too little HIV in AIDS patients to cause AIDS, both groups have observed huge discrepancies between virus titers and AIDS symptoms. In 1993, Shaw and colleagues have described otherwise identical AIDS patients of which 5 contained 0 infectious HIV per ml, and 22 contained between 5 and 10^5 [16,19]. In 1989, David Ho et.al. have also described 40 AIDS patients with virus titers ranging from less than 1 to 10^5 infectious units per ml [30]. In 1993, Ho et. al. even reported 12 AIDS patients, including 8 who had AIDS "risk factors", who were totally HIV-free: "Specific antibody assays, viral cultures, and polymerase chain reaction (PCR) techniques" for HIV were all negative. Their T-cell counts ranged from 3 to 308 per ml [52].

There is only one consistent hypothesis to reconcile the bewildering ranges of HIV titers in Ho's and Shaw's patients, that the role of the virus in AIDS-HIV is a passenger virus, rather than the cause of AIDS. Indeed, non-correlation between the tiers of a virus and disease, and between the very presence of a virus and disease, is one of the hallmarks of a passenger virus. Both Ho et. al. and Shaw et. al. have failed to understand that rare correlations between a virus-at-high-titer and a disease are the hallmark of a passenger virus, and that consistent correlations between a virus-at-high-titer and a disease are the hallmark of a causative virus [8,53,54]. They have, contrary to their claims, established HIV as a passenger virus of AIDS patients.

(3) Until HIV appeared on the scene, the latent period from infection to disease was a function of the generation time of a virus. A virus that replicates in 2 days and produces 100 viruses per generation would cause disease in about two weeks-provided that there is no antiviral immunity. This is because 100 viruses infect 100 cells producing 100x100 or 10,000 viruses 2 days later. Within 14 days of such exponential growth 10^14 cells-the equivalent of a human body-would be infected. Therefore the latent periods of pathogenic retroviruses, like Rous sarcoma virus, and non-retroviruses like flu, measles, mumps, herpes, hepatitis, mononucleosis, chicken pox are all 7 to 14 days [23]. Since HIV replicates in 2 days, like all other retroviruses [48], and since according to Ho an infected cell produces over 1000 viruses per 2 days [32], HIV should cause AIDS-if it could cause AIDS-just as fast as other viruses.

Yet, as Maddox points out, the failure of HIV to cause AIDS within weeks after infection presents another paradox for the HIV-AIDS hypothesis, "...the usually slow decline of CD4+ cells is not consistent with what one would expect from a specific cytotoxic viral mechanism". Indeed, both studies confirm the paradox. Since the AIDS patients contain 10^5 "free viruses/virions" and 10^5 T-cells per ml plasma, the plasma of these patients should be T-cell free within 2 days, the generation time of HIV. But Ho et. al.report that the T-cells of AIDS patients are either steady or even increasing over 1 month, and Wei et. al. report that the T-cells of their patients remain either steady or decline slowly over 5 to 8 months [1,2].

Even if there are 50-timies more T-cells in hidden reservoirs-as Ho et. al. report--, they, too should be infected within two weeks, because according to Wei et. al. the "plasma viral RNA" titer can rise two orders of magnitude within two weeks. In fact, the ability of HIV to increase from 10^3 "plasma viral RNA" units to 10^5 units per ml described by Wei et. al. should only be a fraction of the real "dynamics of the infection in people by HIV" [3], since it occurred despite the presence of two DNA chain terminators, AZT and ddI, used as anti-HIV drugs in addition to a new coded antiviral drug.

Therefore it remains paradoxical that-dated from the time of HIV infection-AIDS occurs at entirely unpredictable times, currently estimated to average 10 years [5]. To determine whether the currently unpredictable time from HIV infection to AIDS can be reconciled with a viral mechanism at all, one needs to know whether HIV kills T-cells, how much infectious virus there is, and the percentage of infected cells at a given time. Since the new studies by Wei et.al. and Ho et.al. provide none of these data, all new calculations "on the dynamics of the infection of people by HIV... in the process of understanding the disease called AIDS" are worthless.

However, the hypothesis that HIV is a passenger virus provides a consistent explanation for the unpredictable time intervals between HIV infection and AIDS. It is one hallmark of a passenger virus, that the time of infection is unrelated to, and independent of the time when a disease occurs-just as with HIV and AIDS. Another hallmark of a passenger viurs is that its titer and even its presence are not correlated with disease- just as was shown above form HIV and AIDS.

The simplest interpretation of the slow decline of T-cells in Ho's and Wei's AIDS patients is a non-viral cause, e.g. long-term intoxication [7]. Take for example the slow decline of liver cells in long-term alcoholics or of lung cells in long-term smokers.

Maddox seems concerned that "reporting of the new event has been curiously selective". Perhaps even science reporters begin to wonder how much further the virus-AIDS hypothesis can be stretched to explain its most obvous failures and inconsistencies: Why is there no vaccine? Why does American/European AIDS stay in the classical risk groups, male homosexual, intravenous drug users and transfusion recipients? Why do AZT-treated HIV-positives get AIDS? [55,56] Why do 918 HIV-positive male homosexuals who had "avoided experimental medications on offer" and chose to abstain or significantly reduce their use of recreational drugs..." remain AIDS-free, long-term survivors? [57] Why did the T-cells of 29% of 1020 HIV-positive male homosexuals and former intravenous drug users from the placebo arm of a clinical AZT trial increase up to 22% over two years-despite the presence of HIV? [58] Why did the T-cells of 14 out of 31 HIV-positive hemophiliacs treated with highly purified factor VIII increase up to 25% over three years-despite the presence of HIV? [59] Why is there not a single study showing that HIV-positive 20 to 50-year-old men or women who are not drug users or recipients of transfusions ever get AIDS? [60]

Why did neither Ho et. al. nor Wei et. al. identify the risk groups their patients came from or indicate whether they had Kaposi's sarcoma, dementia, or diarrhea or lymphoma? Can they exclude that recreational drugs used by AIDS risk groups, like nitrate inhalants, amphetamines, and cocaine are immunotoxic or carcinogenic? [61] Why is it that among 10 long-term (10 to15 years) survivors of HIV recently described by Ho et. al. [50] "none had received antiretroviral therapy..."? Can Wei et. al. and Ho et. al. exclude that the DNA chain terminators, AZT and ddI, that their patients received in addition to the new experimental drugs, do not play any role in the "slow decline of CD4+ cells"? Are they aware that the manufacturer of AZT says in the Physician's Desk Reference that "it was often difficult to distinguish adverse events possibly assoiated with zidovudine [AZT] administration from underlying signs of HIV diseases..."? [62] Are they aware that the DNA chain terminators were developed 30 years ago to kill growing human cells for chemotherapy, not as anti-HIV drugs?

It seems to us that the "new developments" of Wei et. al. and Ho et. al. are a Mayday of AIDS virologists-rather than a "virological mayhem." [43]

(NOTE-Exponents are typed in the following manner: 10 to the power of 4 appears as 10^4)

Acknowledgments: We thank Serge Lang (Yale University), Siggi Sachs (UC Berkeley) and Russel Schoch (UC Berkeley) for critical comments. Supported by the Council for Tobacco Research, USA, and private donations.

References:

[1] Wei, X., et. al., _Nature_ 373, p. 117-122 (1995)

[2] Ho, D.D., et al., _Nature_ 373, p. 123-126 (1995)

[3] Maddox, J. _Nature_ 373, 189, p. 189 (1995)

[4] Gallo, R.C., et. al., _Science_ 224, p. 500-503 (1984)

[5] Institute of Medicine, _Confronting AIDS- Update 1988_ (National Academy Press, Washington DC, 1988)

[6] Centers for Disease Control and Prevention, _Morb Mort Weekly Rep_ 41(No. RR17), p. 1-19 (1992)

[7] Duesberg, P.H., _Pharmacology & Therapeutics_ 55, p. 201-277 (1992) [8] Duesberg, P.H., _Int. Arch. Allergy Immunol_ 103, p. 131-142 [9] Murray, H.W., Scavuzzo, D.A., Kelly, C.D., Rubin, B.Y. & Roberts, R.B._Arch. Intern. Med._ 148, p. 1613-1616 (1988)

[10] Spornraft, P., et.al., _Br. J. Dermatol_ 335, p. 1-9 (1988)

[11] Friedman-Kien, A.E., et.al., _Lancet_ 335, p. 168-169 (1990)

[12] Kaldor, J.M., Tindall, B., Williamson, P. Elford, J & Cooper, D.A., _Journal of Acquired Immune Deficiency Syndromes_ 6, p. 1145-1149 (1993)

[13] Bacellar, H., et.al, _Neurology_ 44, p. 1892-1900 (1994)

[14] Blattner, W.A., Gallo, R.C. & Tein, H.M. _Science_ 241, p. 514-515 (1988)

[15] Duesberg, P.H. _Science_ 241, p. 514-516 (1988)

[16] Piatak, M. et al., _Science_ 259, p. 1749-1754 (1993)

[17] Daar, E.S., Moudgil, T., Meyer, R.D. & Ho, D.D. _N. Eng. J. Med._ 324, p. 961-964 (1991)

[18] Clark, S.J., et.al. _N. Engl. J. Med._ 324, p. 954-960 (1991)

[19] Duesberg, P.H. _Science_ 260, p. 1705 (1993)

[20] Mims, C. & White, D.O. _Viral Pathogenesis and Immunology_ (Blackwell Scientific Publications, Oxford, 1984)

[21] Freeman, B.A. _Burrows Textbook of Microbiology_ (W.B. Saunders Co., Philadeophia, 1979)

[22] Weiss, R.A. _Science_ 260, p. 1273-1279

[23] Fenner, F., McAuslan, B.R., Mims, C.A., Sambrook, J. & White, D.O. _The Biology of Animal Viruses_ (Academic Press, Inc., New York, 1974)

[24] Gross, L _Oncogenic Viruses_ (Pergamon Press, Oxford, 1970)

[25] Harper, M.E., Marselle, L.M., Gallo, R.C. & Wong-Staal, F. _PNAS_ 83, p. 772-776 (1986)

[26] Simmonds, P., et.al. _J. Virol._ 64, P. 864-872 (1990)

[27] Shaw, G.M., et. al. _Science_ 226, p. 1165-1171 (1984)

[28] Weiss, S.H., et. al., Science_ 239, p. 68-71 (1988)

[29] Falk, L.A., Jr. _NEJM_ 316, p. 1547-1548 (1987)

[30] Ho, D.D., Moudgil, T. & Alam, M _New Engl. J. Med_ 321, p. 1621-1625 (1989)

[31] Schnittman, S.M., et.al. _Science_ 245, p. 305-308 (1989)

[32] Ho, D. _New Engl. J. Med_ 322, p. 1467 (1990)

[33] Centers for Disease Control _HIV/AIDS Surveillance_, Year-end edition, p. 1-23 (1993)

[34] Benditt, J. & Jasny, B. _Science_ 260, p. 1219, 1253-1293 (1993)

[35] Fields, B.N. _Nature_ 369, p. 95-96 (1994)

[36] Swinbanks, D. _Nature_ 370, p. 494 (1994)

[37] Rubinstein, E. _Science_ 248, p. 1499-1507 (1990)

[38] Karpas, A., Lowdell, M., Jacobson, S.K., & Hill, F _PNAS_ 89, p. 8351-8355 (1992)

[39] Langhoff, E., et.al., _J. Clin. Invest._ 84, p. 1637-1643 (1989)

[40] Lemaitre, M., Guetard, D., Henin, Y., Montagnier, L. & Zerial, A., _Res. Virol_ 141, p. 5-16 (1990)

[41] Hoxie, J.A., Haggarty, B.S., Rakowski, J.L., Pillsbury, N. & Levy, J.A., _Science_ 229, p. 1400-1402 (1985)

[42] Anand, R., et.al., _Lancet_ ii, p. 234-238

[43] Wain-Hobson, S. _Nature_ 373, (1995)

[44] Pantaleo, G., et.al. _Nature_ 362, p. 355-358 (1993)

[45] Cohen, J. _Science_ 259, p. 168-170 (1993)

[46] Weiss, R. _Nature_ 349, p. 374 (1991)

[47] Embretson, J., et.al. _Nature_ 362, p. 359-352 (1993)

[48] Weiss, R., Teich, N., Varmus, H. & Coffin, J. _Molecular Biology of RNA Tumor Viruses (Cold Spring Harbor Press, Cold Spring Harbor, NY, 1985)

[49] Cohen, J. _Scince_ 260, p. 292-293 (1993)

[50] Cao, Y., Quin, L., Zhang, L., Safrit, J. & Ho, D.D. _N Engl. J. Med_ 332, p. 201-208 (1995)

[51] Raeymaekers, L. _Analytical Biochemistry_ 214, p. 582-585 (1993)

[52] Ho, D.D., et.al. _NEJM_ 328, p. 380-385 (1993)

[53] Duesberg, P.H. _Science_ 260, p. 1705 (1993)

[54] Duesberg, P.H. _N. Engl. J. Med_ 322, p. 1466 (1990)

[55] Seligmann, M., et.al., _Lancet_ 343, p. 871-881 (1994)

[56] Goedert, J.J., et.al. _Lancet_ 344, p. 791-792 (1994)

[57] Wells, J. _Capital Gay_ August 20, p. 14-15 (1993)

[58] Hughes, M.D., et.al. _J. Infectious Diseases_ 169, p. 28-36 (1994)

[59] Seremetis, S.V., et.al. _Lancet_ 342, p. 700-703 (1993)

[60] Duesberg, P.H. _Science_ 267, p. 313 (1995)

[61] Haverkos, H.W. & Dougherty, J.A. _Health Hazards of Nitrite Inhalants_(U.S. Dept. of Health & Human Services, Washington DC, 1988)

[62] Physicians' Desk Reference, _Retrovir_ p. 742-746 (Medical Economics Co., Orandell, NJ, 1994)