Why did HAART improve the prognosis of AIDS?

Jean Umber
Professor of Organic Chemistry
Académie de Nancy-Metz
Lorraine, France

Translated from the French by Jean Umber and edited by David Crowe.
If you are fluent in French, English and Organic Chemistry and can improve this translation, please contact us.
We apologize for any translation-related errors.

When, on May 13th 2000, I received an invitation to sign a petition of support for South African President Thabo Mbeki, I did not expect to discover that behind this desire for re-evaluation of the relationship between the HIV and AIDS there was a debate suppressed for 16 years by the “political-medical-scientific” dictatorship

I was stunned but, after about three days I tried to verify all this information.

What amazed me, and convinced that there was something worth investigating, was the strength of the arguments of ‘rethinkers’, and in particular the numerous references they cited, that I naturally tried to verify, compared to the poverty of the information provided by establishment officials, who generally condemn ex cathedra these rethinkers, using the argument accepted by the majority of people by fear: “By criticizing the official dogma, you prevent the patients from looking after themselves and thus you lead them unto death.”

This argument reminds me of that of medieval inquisitors.


1. Observations

One of the aspects of the controversy which most struck me was the almost unanimous rejection of AZT by these rethinkers. Having a solid training in organic chemistry, I wanted to know more about it, and I was flabbergasted when I discovered the structure of this molecule: It presented a functional group, the azide, the in vitro action of which we know: It releases dinitrogen N2 , in the presence of certain catalysts, and leaves behind[added by drc] a very reactive structure, the nitrene, capable of reacting with many biological structures.

One of the most complete documents concerning the toxicity of the AZT was compiled by Anthony Brink, a South African lawyer, and mentor of president Mbeki.

Moreover numerous in vitro and in vivo studies confirm AZT’s activity.

First of all consider the studies of Handlon and Oppenheimer in 1988, published in Pharmaceutical Research (5: 297-9):

The ability of thiols to reduce 3'-azidothymidine (AZT) to 3'-aminothymidine has been investigated. Incubation with glutathione, dithiothreitol (DTT), or mercaptoethanol at pH 7.2 and 37 degrees C leads to quantitative reduction of the azido moiety to an amine. The reaction is first order in AZT and first order in reducing agent (mono- or dithiol). The second-order rate constants are 2.77 x 10-3, 6.55 x 10-5, and 6.35 x 10-6 M-1 s-1 for the dithiothreitol, glutathione, and mercaptoethanol reductions, respectively. The thiol reduction of alkyl azide to amine under mild conditions is a synthetic method particularly suitable for water-soluble azido compounds that are sensitive to catalytic hydrogenation. The potential for the mono- or dithiol-mediated reduction of alkyl azides under biological conditions must be considered when conducting studies of azido drugs.

It is interesting to note that as far back as 1988 we knew that “therapies” containing an azido grouping reacted without the help of enzymes with the main reducing reserve of the liver and the cells: glutathione and that the reaction is quantitative in vitro (in the conditions of the living cell).

To go further and see if other compounds form in this reaction (which is well and truly a reaction of oxido-reduction, which displeases certain biologists), Reardon et al. repeated the experiment in 1994 and published their results in The Journal of Biological Chemistry (269 (23): 15999-16008):

3'-Azido-3'-deoxythymidine (AZT), AZT 5'-monophosphate, and AZT 5'-triphosphate (AZTTP) were reduced by dithiothreitol with second-order rate constants of 2.30 x 10-3, 1.50 x 10-3, and 7.46 x 10-4 M-1 s-1, respectively. Handlon and Oppenheimer reported that AZT is quantitatively reduced by thiols to 3'-amino-3'-deoxythymidine (Handlon, A. L., and Oppenheimer, N. J. (1988) Pharm. Res. (N.Y.) 5, 297-299). In the present report, multiple products of this reaction were identified by the techniques of UV spectroscopy, phosphate analysis, coelution with authentic standards from reversed-phase high pressure liquid chromatography, two-dimensional NMR spectroscopy, and mass spectrometry. The product mixture from reduction of AZT 5'-monophosphate at pH 7.1 and 25 degrees C was composed of 2,3'-anhydro-beta-D-threo-thymidine 5'-monophosphate (6.4%), 3'-amino-3'-deoxythymidine 5'-monophosphate (19.6%), beta-D-threo-thymidine 5'-monophosphate (6.8%), thymine and 3-amino-2,3-dideoxyribal 5-monophosphate (8.9%), beta-D-threo-thymidine 3',5'-cyclic monophosphate (9.1%), 3'-deoxy-2',3'-didehydrothymidine 5'-monophosphate (31.5%), and 3',5'-anhydro-beta-D-threo-thymidine (17.8%). Thymine and 3',5'-anhydro-beta-D-threo-thymidine were also products of reduction of AZT and AZTTP. Furthermore, the nucleosides of the above monophosphates were products of reduction of AZT, and the corresponding triphosphates were products of reduction of AZTTP. The product ratios were dependent on the level of phosphorylation of AZT and on the pH of the reaction. Mechanisms for formation of these products are proposed.

Many studies also show that AZT affects all the elements of the redox system of the cell: for example Cretton et al., 1991 ; Cretton & Sommadossi , 1993 ; Eagling et al. , 1994 ; etc.

What is very interesting in the Reardon document quoted above, is that the most plentiful reduced compound is the 3'-deoxy-2'-deoxy-2',3'-didehydrothymidine-didehydrothymidine 5'-monophosphate-monophosphate, which is the monophosphate of a well known “anti-AIDS medicine”: d4T. Reardon looked for the mechanisms which lead to the formation of all these compounds. They are essentially mechanisms of reduction, substitution and elimination known well by the organic chemists. This means that these reactions will take place in vivo.

Moreover the conclusion of Reardon :

Finally, it has not been demonstrated that reduction products of AZTMP, AZTDP, or AZTTP are formed intracellularly. These issues are currently under investigation.

found a solution recently when Becher and Pruvost showed conclusively that this reduction of AZT to d4T occurs in vivo. Although using the mechanistic considerations of Reardon (that I consider partially erroneous from a strictly chemical point of view) they clearly show that one of the metabolites of AZT ( zidovudine ) is d4T (stavudine ) and from that draw conclusions about the association between AZT and d4T, that all clinicians reject.

However, they do not clarify the real mechanism of this reduction. What is the reducer, in particular?

In their publication they give the results obtained from diverse patients, treated with a mixture of nucleoside analogues. 

It is clear that the unique cases where d4T is present correspond to the combination of AZT and of 3TC (3-thiacytidine, also called Lamivudine). We can thus ask whether the reduction that occurs with AZT, which is a priori caused by the cellular thiols (glutathione) does not also occur with lamivudine.

Studying this topic cleared up for me a fact that no rethinker should refute: Here is the evolution of the cases of AIDS in the United States since 1980:

Note that, from 1993, the US modified the definition of the disease AIDS. It was previously necessary to suffer from an AIDS-defining disease to be diagnosed but, from 1993, it was only necessary to be HIV-positive and to have a count of CD4 cells lower than 200 cells per microliter.

It is clear that the diagnosis of the real AIDS rapidly fell in 1992-3, declines which became more marked in 1995-6. An even more drastic decline in mortality began in 1996.

How can one interpret this evolution if HAART was ineffective?

Treatments against AIDS, since the beginning until 1995, essentially contained antibiotics (e.g. bactrim or metronidazole) or antifungals, and, from 1986-87, AZT. The increasing popularity of these treatments coincided with an increase in the cases of AIDS and a little later, an increase in deaths. The decline in cases in 1992-3 coincides exactly with the decrease of the dosage of AZT in prescriptions, thanks to the impact of the Concorde study which showed that delaying the administration of the AZT increased slightly the probability of survival in 3 years.

On the other hand, in 1995, the combination AZT-3TC (combivir) appeared, which is still use today. In fact, most of the drug regimes from that time on contained 3TC.

This trend is strongly connected to the strong decrease in mortality of people diagnosed with AIDS as noted in various articles, including 3TC le bon élève de la classe (“3TC is the best student in the class”) and 3TC une synergie pleine de promesse (“A synergy full of promise”).

We notice moreover that the scientists are only “noticing” this synergy, without being able to really explain it. It is almost mystic.

An analogue of 3TC, emtricitabine (aka emtriva or FTC), presents the same advantages. Today we find practically no triple therapy without 3TC or FTC.

Studies were made to determine, among the different nucleoside analogues, which have most impact on the health of the patients, and invariably we find 3TC or FTC included.

For example, this paper of Saag et al ( 2004 ) concludes:

Once-daily emtricitabine [FTC] appeared to demonstrate greater virological efficacy, durability of response, and tolerability compared with twice-daily stavudine when used with once-daily didanosine and efavirenz.

It is same strange to think that a substance (emtricitabine) supposed to work in the same way as another (stavudine) has an impact so different.

2. Why this efficiency?

First of all, compare the chemical structures of all the antiretrovirals which we have discussed :

What do we notice? Who distinguishes the nucleosides with little activity (stavudine or didanosine) from lamivudine?

It is obviously the presence of sulfur in 3', that replaces a CH2 or a CH. It is not a coincidence that we also find this sulfur in emtricitabine. In a more general way, as we are going to be able to notice later, the compounds which are profoundly going to modify the evolution of the disease are acetals or thioacetals [?].

Let us return to the oxidizing effects of AZT and other substances, such as bactrim or metronidazole. It has been demonstrated that they react with the reducing reserve of cells, namely glutathione, the antioxidant effect of which is connected to a thiol group, of structure R-S-H (Sulfur has the same degree of oxidation).

Now it is very important to note that the disease AIDS (deficiency of CD4 and related opportunist diseases) is definitely connected to a deficiency of glutathione. De Rosa & al. demonstrated this link, and even showed that the supplementation with biological thiols (derived from cysteine) increased the life expectancy of people using AZT

Numerous studies show the importance of these sulphur-containing molecules. For example, that of Kalebic & Kinter:

At 15mM, GSH, GSE and NAC supressed both HIV production and total viral protein synthesis.

The authors also indicate that the combination of these thiols with AZT has a “cooperative” antiviral effect – like the synergistic effect of 3TC!

The presence of HIV was determined by the activity of the reverse transcriptase enzyme, a method which according to Papadopoulos et al is not enough to demonstrate the presence of a virus, but does show the existence of oxidative stress.

3. How can we chemically compare 3TC with thiols?

As regards the redox properties, lamivudine is a cellular reducer (as is glutathione).

The proof is that the main metabolite excreted is S-oxy-lamivudine


RXLIST supplies us with some explanations:


Metabolism of lamivudine is a minor route of elimination. In man, the only known metabolite of lamivudine is the trans-sulfoxide metabolite. Within 12 hours after a single oral dose of lamivudine in 6 HIV-infected adults, 5.2% ± 1.4% (mean ± SD) of the dose was excreted as the trans-sulfoxide metabolite in the urine. Serum concentrations of this metabolite have not been determined.

A chemist will notice that this structure represents one 3-oxathiolane, that is a mono-sulphurated acetal. These compounds can be synthetized in the absence of water (anhydrous conditions) and are stable in the water in pH > 6 approximately. On the other hand, they are hydrolyzed more or less quickly in acid pH.

The site Organic-chemistry.org allows us to go a little farther. It shows that the structure of lamivudine is intermediate between that of a dithiolane and that of a dioxolane. (see protecting groups)

In particular, we notice that the first ones are the most difficult to hydrolyze in pH 1 (that is the stomach pH). Logically, lamivudine will be at least as quickly hydrolyzed as dithiolanes. Now what do we notice?

(NB: the legend is found here : )


At pH=1 and at room temperature, dithiolanes react slowly (some reactions of hydrolysis have a duration of from 1 to 4 hours according to the proportion of water. This reaction is also called deprotection). This duration is long enough so that in the stomach a significant amount of lamivudine can be hydrolyzed :

It forms an aldehyde, the glycolic aldehyde and a mercaptoalcohol. This last one (1-cytosyl-2-mercaptoéthanol) is moreover, in theory, unstable and would be transformed to cytosine and 2-mercaptoethanal HS-CH2-CHO. This hydrolysis is strongly catalysed by electrophiles, as indicated in Carey et Sundberg (Chimie Organique Avancˇe, Tome 2, page 691), in particular the nitrosyl ion NO+, certainly commonly present in patients suffering from oxidative stress

The 1-cytosyl-2-mercaptoéthanol and the 2-mercaptoethanal are both thiols, and the second one an aldehyde, and may thus partially replace the deficient thiols, or rather, according to me, avoid the oxidation of the glutathione in sulfates by AZT or an other oxidative stuff. To note as well, while dibrome (Br2) oxidize quickly the dithiolane (Electrode potential of Br2/Br - = 1.09 V), it can be expected a stronger oxidation by azides (the potential of the electrode HN3/(NH4+ + N2) [structure of HN3 close to AZT] is sharply superior, equal at 1.96V) (Values given in the “Usuel de chimie générale et minérale” from Maurice Bernard & Florent Busnot [Dunod])[??? DO NOT UNDERSTAND AT ALL ??? Is it better?],  Besides aldehydes are very reducing (more reducing than thiols), because the potential of the electrode acid/aldehyde is 0.09 V, that of the electrode RSSR/2 RSH being 0.3 V approximately.

Some published studies lend credence to this hypothesis, in particular, this study of Fletcher et al.

First, zidovudine has been shown in vitro to reduce the amount of lamivudine triphosphate formed in phytohemagglutinin-stimulated PBMC and U937 cells. We hypothesize that the probable mechanism of this inhibition is the ability of zidovudine to increase dCTP pools in stimulated and resting cells, which results in feedback inhibition of deoxycytidine kinase activity.

Manifestly, AZT decreases the amount of lamivudine in vitro. It only makes sense to look for a possible reaction between these compounds. And effectively, we have just shown this evident possibility, the one being reducing (lamivudine), the other oxidizing (AZT).

We can also foresee the forming of stavudine and sulfates, the ultimate stage of the oxidation of the sulphurated compounds. These sulfates are, moreover, known to be generative of strong diarrheas..

4. Other studies involving acetals or thiocetals

A recent study (1999) brings more support to this hypothesis:

“Anti-Human Immunodeficiency Virus Type 1 Activity, Intracellular Metabolism, and Pharmacokinetic Evaluation of 2'-Deoxy-3'-Oxa-4'-Thiocytidine” – De Muys, Gurdeau, Nguyen-Ba. Antimicrobial Agents and Chemotherapy, August 1999, 43(8): 1835-44. 0066-4804/99

It is clearly indicated by this study that the virologic impact of the dOTC is more important than that of the 3TC, and nevertheless, the amount of dOTC  not metabolized is much less than that of the 3TC: with less “substance” in the cell, there is an more important pharmacodynamic effect. The article does not clarify the nature of the formed metabolites, the molar mass clearly  smaller  than that of the dOTC because their retention time in Gas Chromatography is shorter, but it is possible to think that it is their presence which is going to give more efficiency to the dOTC. One of these metabolites could be for example the 2-cytosyl-2-mercaptoethanol, a thiol that is certainly very reducing, which would be obtained much more easily from the dOTC than the 1-cytosyl-2-mercaptoéthanol from the 3TC can be, doubtless because the hydrolysis begins with the protonation of the oxygen, easier in the dOTC where this oxygen is less hindered.

Naturally, the presence of this metabolite infers the presence of glycolic aldehyde.

Another article  of 1999, also resulting from the laboratory of Mark Wainberg, studies the impact of the corresponding dioxolanes. He notices that these are capable of operating there where the other synthetic nucleosides failed because of diverse "mutations".

However, the ability to react towards the water of dioxolanes, such as indicated here, is extremely important, to the point that even at pH=4, they are slowly hydrolyzed. At pH=1, the hydrolysis occurs in 5 minutes.
How can we explain its positive impact on AIDS when the hydrolysis does not release thiols here?
The only plausible hypothesis takes into account the fact that in every case about which we have just spoken (dioxolanes, oxathiolanes, even dithiolanes), there is liberation of glycolic aldehyde HOCH2-CHO. The speed of this liberation depends on the speed of hydrolysis of the heterocycle. This one varies in the following way:

dioxolane > oxathiolane substituted near the sulfur > oxathiolane substituted near the oxygen > dithiolane.

We notice that it is roughly the evolution of the positive impact of these molecules on the "virologic" factors.

Besides, the aldehyde oxidase is capable of reducing numerous nitrogenous oxidizers, as well as shown by this study :

Aldehyde oxidase is a molybdenum cofactor-containing soluble enzyme present in the liver and other tissues of several mammalian species. Despite its name, aldehyde oxidase is involved not only in the oxidation of aldehydes to carboxylic acids but also in the oxidation of nitrogen-containing heterocyclic compounds and the reduction of nitro-aromatic compounds, isoxazole, and isothiazole ring systems.... Compared to the cytochrome P450 system, this offers a much simpler situation in which to address potential drug interactions using in vitro approaches.

We conceive here that a compound which releases an aldehyde is capable, thanks to the aldehyde oxidase, of eliminating the danger resulting from nitroaromatic compounds (metronidazole for example) or isoxazoles (bactrim)

The reducing character of these compounds containing cycles with two 1,3-chalcogenes (oxygen, sulfur, even selenium) is demonstrated thus amply, and would result essentially from the aldehyde moiety released and additionally of the forming of thiols.

5. Conclusion

Other studies are naturally necessary, but it seems completely reasonable to envisage, especially considering the issue of oxidative stress dear to the Perth Group and to Heinrich Kremer (to see " Answers to the President Mbeki "), that lamivudine and the other acetals capable of releasing an aldehyde largely eliminates the oxidizing effects of AZT, and prevents it from eliminating the cellular reducing reserve, already depleted by the disease AIDS itself.

Moreover, some studies lead us to conclude that AZT does not act by terminating the growth of proviral DNA, but rather by direct oxidation of proteins of reverse transcription and the RNA. Indeed, if it was really about stopping this replication only the thymidine containing an azide in 3' would be capable of it. Now, these studies show undoubtedly a not unimportant "virologic" activity (twice less important than that of the AZT) for the ADRT (4'-azidothymidine) (opposite), which contains on more OH in 3'. These both compounds thus act as azides, with all that it implies regarding toxicity and oxidizing properties. [? Had a hard time understanding this paragraph and adapting the translation ? Is it better?]

Finally, we can wonder whether the high mortality attributed to AIDS (or HIV) at the end of the 1980s was not really due to the excessive use of AZT (e.g. these studies from Ruengpung Sutthend).