PerspectiveSuccess in Translational Medicine

Twenty-Five Years of Translational Medicine in Antiretroviral Therapy: Promises to Keep

See allHide authors and affiliations

Science Translational Medicine  07 Jul 2010:
Vol. 2, Issue 39, pp. 39ps33
DOI: 10.1126/scitranslmed.3000749


The year 2010 marks the 25th anniversary of modern antiretroviral drug discovery and development. In the early 1980s, AIDS was almost always a lethal disease with an appalling clinical course characterized by severe opportunistic infections and unusual forms of cancer. Since that time, starting with zidovudine (AZT) and related 2′,3′-dideoxynucleosides, the causative retrovirus, now called HIV-1, went from being an untreatable infectious agent to being the target of highly active antiretroviral therapy (HAART). This degree of progress refuted early prophesies of therapeutic futility and represents a striking example of translational medicine. Here I review foundational discoveries and events in HIV-1/AIDS research and explore lessons for future translational medicine efforts.

The only way of discovering the limits of the possible is to venture a little way past them into the impossible.

—Sir Arthur C. Clarke

The path to modern treatment against HIV-1 started approximately 25 years ago (17), and currently available therapies have demonstrably reduced the morbidity and mortality rates of AIDS, in many ways without precedent (Fig. 1). This remarkable reduction in the death rate from HIV-1/AIDS was mediated by the suppression of HIV-1 replication, the ensuing improvements in CD4+ T cell counts, and substantial protection of the immune system made possible by highly active antiretroviral therapy (HAART)—which typically includes a combination of three or more antiretroviral medications from at least two different drug classes. Progress has not been confined to resource-rich countries, particularly with the advent of fixed-dose combinations of generic antiretroviral drugs, which are administered on relatively convenient schedules for greater utility in resource-poor countries (1, 8). Physicians are now secure in the knowledge that their patients will benefit from an array of therapeutic options, and many do not recall a time when there was any other reality. However, this knowledge did not come easily. The presumed futility of antiretroviral therapy and the hopelessness this engendered in patients, physicians, and scientists alike are perhaps now unremembered history. One may ask if there are lessons in this history to guide us in the future.

Fig. 1. Visualizing success.

Shown are trends in annual death rates due to the nine leading causes among individuals 25 to 44 years of age in the United States, 1987–2006. In 1994 and 1995, HIV-1 disease ranked as the leading cause of death among people 25 to 44 years of age. In 1995, HIV-1 disease caused about 32,000 deaths, or 20% of all deaths in this age group. The ranking of HIV-1 disease fell to fifth place from 1997 through 2000 and to sixth place from 2001 through 2006, drops that were associated with combination antiretroviral therapies. In 2006, HIV-1 disease caused about 5000 deaths, or 4% of all deaths in this age group. For comparison with data for 1999 and later years, data for 1987–1998 were modified to account for The International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD-10) rules instead of ICD-9 rules. CREDIT: Data are from the U.S. Centers for Disease Control (1, 45).



In 1981, HIV-1, expressed clinically as AIDS, proclaimed itself in the United States in the form of reports that ominously described a rare cancer, Kaposi’s sarcoma (KS), and uncommon opportunistic infections, such as Pneumocystis carinii, among previously healthy young gay men in New York and San Francisco (9, 10). At about that time, my colleagues and I at the National Cancer Institute (NCI) provided medical care for one of the first patients with this lethal, terrifying, and, at the time, nameless syndrome at the Clinical Center of the National Institutes of Health (NIH). For the public, there was an atmosphere of fear, compounded at times by a loss of faith in the competence or sincerity of government agencies; and in the scientific community, acrimony was driven by contradictory theories on causation and therapy, often as erroneous as they were deeply held. The enormous progress over the past 25 years in developing new therapies, with diverse mechanisms of action, against the causative retroviral agent (1133) was unimaginable at the time.

The existence of animal retroviruses—RNA viruses replicated by reverse transcriptase (RT)—was already well known and widely accepted. However, the importance of pathogenic human retroviruses, or even their very existence, remained topics of significant controversy. Indeed, the following assumptions complicated the task of discovering and developing antiretroviral drugs to attack the causative agent of AIDS: (i) Active (replicating) retroviruses did not exist in human beings. (ii) If human retroviruses did exist, they were not involved in the pathogenesis of human diseases. (iii) If human retroviruses were involved in human disease pathogenesis, they played at most a minor role in the general public health that was limited to rare subacute T cell leukemias, tropical spastic paraparesis, or related neurological syndromes, as in the case of human T cell leukemia virus–1 (HTLV-1). And (iv) even if retroviruses were involved in other human diseases, they were by their very nature inherently untreatable, because of their capacity to integrate into the host genome and/or rapidly mutate as a result of error-prone RT. The situation was more complex still: Even when retroviral causation was finally established, most retrovirologists had rather little experience in clinical drug development and generally believed that a vaccine program was a more important priority than antiretroviral drug discovery. Together, these factors built a barrier to progress in the prevention, diagnosis, and treatment of AIDS.

The discovery of HIV-1 affected almost every aspect of medicine and the public health (11). Indeed, the formal proof, obtained in 1984, that a new human retrovirus was the cause of AIDS and the virtually instantaneous development of an effective screening test for blood donors were astonishing scientific achievements (12). There is no doubt that the rapid application of this knowledge saved countless lives. However, the realization that a previously unknown retrovirus was capable of causing a deadly human disease fostered a sense of futility or therapeutic nihilism in clinical researchers and patients alike. The belief that retroviruses were, by definition, not amenable to therapy remained strong and hampered what clinical investigators, using the available science and technology, could accomplish or even attempt. Moreover, while education, community outreach, and other strategies for prevention were greatly advanced by the identification of HIV-1 as the cause of AIDS, reliance on primary prevention alone would still leave millions of infected adults and children to their fate.

Working in collaboration with scientists at private pharmaceutical companies, Duke University, and other academic centers, my colleagues and I at NCI had the privilege of providing the first evidence that meaningful inhibition of HIV-1 replication could be accomplished, initially in the lab and then in the clinic (17, 15, 26). The story is not about laboratory discoveries alone, but rather illustrates the ability to apply such discoveries to clinical care in one continuous motion. Not uncommonly, these translational processes involve discrete groups separated by geography, administration, and divergent priorities or senses of urgency.

Starting in 1985, the first antiretroviral drugs taken into the clinic were nucleoside reverse transcriptase inhibitors (NRTIs) in the form of 2′,3′-dideoxynucleosides, of which AZT (3′-azido-2′,3′-dideoxythymidine or zidovudine) is a prime example (1).


In 1984, shortly after the discovery of HIV-1, our central hypothesis, or simplifying assumption, was that protecting uninfected immune cells by using virally targeted small molecules to block retroviral infection in vivo, and the attendant T cell destruction, would permit meaningful recovery and restoration of immune function—or at least prevent further loss of immunity. In hindsight, this view seems self-evident, but 25 years ago it was not. We further hypothesized that it was not necessary, as many thought-leaders believed, to eradicate the virus in order to achieve a durable clinical benefit. The daunting and often self-imposed “requirement” of total viral eradication fostered skepticism about the nascent experimental antiretroviral therapy programs that did not take this view. This skepticism in turn strengthened the belief that the best course to treat patients with AIDS was to provide supportive care, while attempting to manage opportunistic infections and neoplasms. In contrast, we believed that if viral replication were suppressed, that alone would drastically reduce the capacity of the virus to cause harm, as long as the suppression held. The latter perspective gave us a realistic and practical focus on drug development with attainable goals for one laboratory. We did not use animal models, because none were accepted or considered reliable at the time, and awaiting their development was not an acceptable option. In any event, it seemed likely that antiretroviral activity in our in vitro systems, including the protection of specific helper CD4+ T cells, would correlate with activity in patients, albeit not perfectly.

We followed these ideas in a newly established antiretroviral drug discovery and screening program in my laboratory (one of the very few in the world with the technical proficiency and willingness to do so at that time). Though we did not use the name, this was a translational medicine team in every sense. The program led to the discovery of the broad HIV-1 RT-targeting properties of a series of novel 2′,3′-dideoxynucleoside congeners of both purines and pyrimidines, one of which is AZT, the first such compound to be used in the clinic. After anabolic phosphorylation reactions by host-cell (not viral) enzymes (13), these NRTIs function via competitive inhibition and chain termination against the HIV-1 DNA polymerase (RT) (14). The earliest agents (now generics) still play an important role as ingredients of antiretroviral therapy combinations, but more important, they pierced a critical barrier, allowing other drugs and, more significantly, drug combinations to follow.

AZT and related drugs held substantial promise for five reasons: (i) These drugs were active against widely divergent HIV-1 isolates in vitro. (ii) There was a large differential between the drug concentration needed to inhibit HIV-1 replication and the viral cytopathic effect and the amount of drug that was toxic to uninfected T cells and monocytes. (iii) Normal antigen- and mitogen-driven T cell activation showed a comparable differential between the dose needed to inhibit viral replication and the dose that interfered with normal function, as did immunoglobulin production by indicator B cells. (iv) The ratio of dideoxynucleoside triphosphate to cellular deoxynucleoside triphosphate was a critical factor in determining the antiretroviral activity of AZT and related compounds in target cells, providing additional support for the likely mechanism of action. And (v) an understandable structure-activity relationship for the class was evident quite quickly, providing insights into the development of second-generation nucleoside/nucleotide analogs (17, 24). Most importantly, we observed clinical activity in our first phase 1 (single-center, nonrandomized) trials that involved the administration of AZT and related drugs to patients with AIDS (1, 5, 15, 25, 26); these observations included increases in the numbers of circulating helper-inducer CD4+ T cells, an improvement of cytotoxic T cell response to influenza virus–infected autologous cells, conversion from anergy to positive delayed hypersensitivity skin-test reactions, clearance of fungal infections without specific antifungal treatment, and other signs of improved immune function. In many cases, the increase in circulating CD4+ T cells became evident after the second week of therapy (5).

We also found that peripheral blood mononuclear cells infected with HIV-1 became undetectable at therapeutic concentrations of AZT (5). At this time, there were no modern molecular diagnostic companions to antiretroviral therapy, such as viral RNA load assays and viral genotyping for drug resistance. Therefore, we developed the earliest dideoxynucleosides without the benefit of such companion diagnostics (1, 26).

We quickly learned that AZT had excellent oral bioavailability and penetration into the cerebrospinal fluid, with clearly evident improvements of AIDS-dementia complex and related neurologic conditions (2629), all features that would be important in deciding whether to move an experimental agent into clinical trials as a therapy for AIDS. These results prompted us to commit to more advanced studies with AZT and, importantly, to also explore clinical applications of other dideoxynucleosides that showed good antiretroviral activity in our laboratory systems, both alone and in combination (1, 30). We then discovered that other dideoxynucleosides that suppressed HIV-1 replication in our laboratory tests had activity in the clinic and a favorable therapeutic index for long-term administration; thus AZT was neither unique nor an anomaly. None of these drugs represented a cure by any means, nor were their virological, immunological, and clinical effects consistently durable when they were used as single agents. A standard of care based entirely on sequential antiretroviral monotherapy was not a long-term solution. Yet combining agents did not prove difficult, and by 1990 (less than 5 years after the first discoveries), we and other researchers saw clear signs in the clinics that AIDS likely could be changed from a rapidly fatal disease to a manageable illness (1, 26).


AZT and related 2′,3′-dideoxynucleosides were rapidly advanced into prospective, randomized, multicenter clinical trials endorsed by the National Institute of Allergy and Infectious Diseases (NIAID) and private pharmaceutical companies, initially using clinical end points (primarily survival) because no surrogate markers were then accepted. For AZT, the clinical scientists at Burroughs-Wellcome (the corporate sponsor, now GlaxoSmithKline), along with several highly dedicated academic investigators (31), made groundbreaking contributions in 1986 by rapidly advancing AZT into registration-seeking clinical trials after the initial laboratory and clinical observations (1, 2, 4, 5) and by strongly advancing a drug to treat an untreatable disease by means of a history-making head-to-head comparison of AZT to placebo treatment. They showed enormous resolve. After all, HIV-1 was still new, and one ruefully common presumption was that treatment directed at this virus—which features proviral integration and enormous mutability during active error-prone replication—was destined to fail, or worse, cause serious harm. At this early stage of antiretroviral drug development, without a placebo-controlled trial, a consensus on the safety and efficacy of AZT could not have been achieved. In turn, the lack of such a consensus would likely have been characterized as confirmation of the futility of targeted antiretroviral therapy in general.

The randomized controlled trial promptly showed a significant survival advantage for patients treated with AZT versus placebo, together with improvements in clinical, virological, and immunological responses (31, 32). These improvements were prefigured in the first patient enrolled in our AZT study at the NIH Clinical Center and others who followed (5). The randomized clinical trial led to approval by the U.S. Food and Drug Administration (FDA) and by health ministries in other countries, with unprecedented speed. This changed everything (33). Retrovir (zidovudine, AZT) was approved in the United States on 19 March 1987. Our results describing zidovudine’s in vitro activity against HIV-1 had only been communicated on 28 June 1985, less than 2 years before the FDA approval. And we published the results of the first clinical study barely 1 year before FDA approval (2, 5).

The improvement of immunological function mediated by antiretroviral therapy also had unanticipated benefits against a previously unknown herpesvirus linked specifically to the pathogenesis of KS, one of the pathognomonic features of the AIDS pandemic. We now know that human herpesvirus 8 (HHV8 or KSHV) is a g-2 herpesvirus present in virtually all forms of KS. There was, at least in part, a temporal association between clinical improvements of KS, the reemergence of HHV8/KSHV-specific cellular immunity, and the suppression of HHV8/KSHV viral replication in patients receiving effective antiretroviral therapy (34, 35). These observations may explain the dramatic drop in the incidence of KS following the introduction of HAART (Fig. 2). This was not predicted in 1985 and is another reminder that basic research and clinical investigation inform one another—perhaps an essential feature of what translational medicine means for the public health.

Fig. 2. KS in SF.

Shown are the annual incidences of KS among white men in San Francisco. The introduction of combination antiretroviral therapies (HAART) was associated with a precipitous drop in the incidence of KS in nine population-based cancer registries sponsored by the NCI Surveillance, Epidemiology, and End Results (SEER) Program (1, 46). CREDIT: Data are from the SEER database.


By 1988, 4 years after the identification of HIV-1 as an “untreatable” new retrovirus, it was credible to conclude that “we now face a totally different future” (33). The primary question at that point was no longer whether HIV-1 could ever be successfully treated using the science and technology already at hand, but rather how fast more therapies could be developed by the pharmaceutical industry. Subsequent events have confirmed this optimism. Members of the first generation of NRTIs were eventually joined by nonnucleoside RT inhibitors (NNRTIs) (16) and viral protease inhibitors (1719). Then, still later came a range of agents targeting other phases of the HIV-1 life cycle, including inhibition of (i) the viral/host-cell fusion step for gp41-mediated entry into the host cell (20, 21); (ii) other early events related to entry of the virus, such as with CCR5 coreceptor antagonists (21); and (iii) the viral integrase, which functions to permit viral double-stranded DNA’s integration into the host-cell genome (22). Eventually, it became possible to combine available agents from different drug classes into what is now called HAART, which is simultaneously a therapeutic term of art and a declaration of definitive progress. A recent scientific review named current therapy against HIV-1 infection a “triumph for modern medicine” (23). Suffice it to say, the word “triumph” was rarely used in the early days of this story.


The environment in which a research program exists is immensely important to its success. Perhaps the location of our research group within the intramural program of NCI had special benefits for antiretroviral drug discovery and development. First, over many years, NCI had a consistent and nearly unique commitment to search for the viral causation of cancer in general and to identify human retroviruses specifically. This commitment led to the discovery and characterization of HTLV-1, the first pathogenic human retrovirus, by Gallo and his co-workers in 1980 (11). Second, there was a long-standing interest in unraveling the relationship between immunodeficiency diseases and cancer. Third, the NCI placed a high priority on novel drug development and had considerable expertise in and resources for clinical pharmacology and toxicology, including the intracellular metabolism (activation) of nucleoside analogs, which were used in treating some cancers. Fourth, the leadership endorsed the philosophy that, in order to discover the limits of the possible, discovery efforts sometimes must extend into the territory of the impossible (as per the philosophy of Sir Arthur C. Clarke), and these nebulous boundaries can shift. Last, but certainly not least, NCI strongly supported translational medicine (before that term was widely used). For example, my group was able to directly initiate proof-of-concept clinical studies based on our own laboratory observations without requiring transfer of the project to another group. This process was facilitated by an NIH tradition of locating research labs adjacent to clinical wards and strongly encouraging, and in many cases requiring, physician-investigators to adopt a philosophy of translating knowledge from their own research lab benches to their own clinics and hospital wards, then taking clinical observations back into the research lab for further study. At every critical juncture, the leadership of both NCI and NIAID strongly endorsed this view. I cannot overstate how important this interactive research network was and still is. Without a doubt, this environment made it possible for my research group to play a role in the discovery and development of the first three antiretroviral drugs approved by the FDA (1, 26).


We have made considerable progress in treating HIV-1/AIDS, and ironically one of the most important current challenges is to avoid triumphalism (Table 1). HIV-1 is by no means vanquished. It is now estimated that over 33 million people are living with HIV-1 infection worldwide, the majority of whom reside in sub-Saharan Africa. Moreover, 2.7 million new HIV-1 infections occur globally each year (with more than 50,000 in the United States alone).

Table 1. Challenges associated with the current paradigm of antiretroviral agents.


View this table:

Twenty-five years ago, integration of the provirus into the host genome was one of the retroviral features expected to render antiretroviral therapy futile. That clearly is not the case. However, viral latency and reservoirs impervious to treatment cause viral rebound after the discontinuation of treatment, and therefore, therapy must be given for a lifetime. There is no well-documented example of a cure for AIDS in any patient. In addition, viral resistance to many available drugs has emerged. Last but not least, it is quite troubling that cross-species transmissions (transpeciation) of HIV-1 and related viruses have been documented in modern times (1). One may reasonably ask: How many anonymous pathogenic retroviruses-in-waiting exist? To compound the matter, there is the very recent case of such retroviral transpeciation in a Cameroonian woman living in Paris (36). This transpeciation involved the gorilla retrovirus SIVgor (now proposed as a “founding” member of a new HIV-1 group P), and several, but not all, commonly used viral RNA load assays did not properly quantify the virus and missed it (36). Similar issues involving the underquantification of HIV-1 non-B subtypes have been reported (37). Such situations have considerable consequences for individual patients, public health, and translational research programs seeking new treatments. The evolution of new and more virulent retroviruses could in principle occur, and these new pathogens could initially evade epidemiological surveillance, feature resistance to the known antiretroviral drug classes, and outpace even the remarkable capacity of the scientific enterprise to develop improved versions of drugs in the current paradigm. Furthermore, all of these issues should be weighed and considered in light of the inherent capacity of HIV-1 to diversify on an astonishing scale. Consider this: The diversity of influenza worldwide in any given year is typically comparable to the diversity of HIV-1 sequences found within a single infected individual at any one point in time (38). Moreover, cardiac and metabolic complications may accompany the prolonged use of current antiretroviral therapies, and this might create medical and financial burdens in the long term, especially on resource-poor nations (1). Progress made can become progress undone.


The primary focus of antiretroviral drug development over the past 25 years has been on classic small molecules that target products of the viral pol gene: RT, protease, and more recently integrase. Yet there is no shortage of ideas for a new era of antiretroviral therapy (1, 23, 39, 40). Fortunately, many of the currently available drugs are compatible with therapy requirements in resource-poor nations, and that should remain one goal in any new paradigm of drug development.

Several proposals for a renewed commitment to paradigm-shifting research and translational medicine are worth mentioning. Richman et al. have made an innovative set of proposals to establish a “collaboratory” to study HIV latency that is operated as a joint venture between NIH, the pharmaceutical and biotechnology industries, and academic medical centers (23). In addition, perhaps mechanisms modeled after the Specialized Program of Research Excellence (SPORE) used in cancer research could be adapted and expanded as centers of translational medicine in HIV-1/AIDS and/or the malignancies associated with this condition. The latter would include KS and high-grade B cell lymphomas, such as Burkitt’s lymphoma.

One specific mission for such programs could be to explore radically new methods of treatment, especially projects that may have unacceptable or indeterminate timelines and substantial risks for return on investment. As but one example of such a new method, it might be possible to deploy small RNA molecules directed against HIV-1 viral promoter regions to durably silence genes at the point of transcription [transcriptional gene silencing (TGS)] in the nucleus (41, 42). This process differs from canonical RNA interference, which involves posttranscriptional gene-silencing pathways. TGS specifically enlists the host cell’s chromatin remodeling system to epigenetically silence the integrated provirus and suppress replication while the modified chromatin remains stable. A short (and renewable) host-cell exposure to these small RNAs could result in durable changes that are epigenetically transmitted to daughter cells. Strategies based on transcriptional regulation by promoter-targeted RNAs or related mechanisms would not necessarily be confined to viral genes. Such RNA-mediated transcriptional silencing could also operate in human cells against virtually any normal or mutated gene, including chemokine receptors or classic oncogenes, and accordingly could have deep clinical ramifications for the treatment of cancer and many other diseases. In the case of HIV-1, TGS would not seek to end viral latency but rather to make it a permanent or semipermanent state; this would be accomplished by targeting not the virus but instead the host chromatin architecture and function [converting (open) euchromatin to (closed) heterochromatin], thereby potentially achieving a functional cure. If the virus cannot replicate and key genes cannot be transcribed, HIV-1 will probably do no harm, and at a minimum, the reliance on current HAART could be diminished substantially.

These kinds of long-term translational medicine endeavors, in which the probability of success or the time to completion cannot be precisely quantified, would likely be beyond the reach of current drug development programs sponsored by either private pharmaceutical companies or NIH. Paradigm shifts, almost by definition, contain an element of surprise. For a period of approximately 10 years after the discovery of AZT and related drugs as therapeutic agents, the annual age-adjusted rate of death due to HIV-1/AIDS in the United States climbed inexorably (Fig. 1), until suddenly there was a peak and then a precipitous drop in mortality in the mid-1990s (1). And yet there had been, all along, step-by-step progress that ultimately culminated in the development and widespread availability of HAART. If denominated in years of life, modern antiretroviral therapy has saved millions of life-years and transformed medicine (1). But along the way, unrealistic or impatient policy-makers might have concluded that antiretroviral therapy was a failure—with profoundly negative consequences.

Finally, it may be possible to expand the scope of certain existing NIH intramural scientific programs beyond their current structure and administration, including joint appointments with institutions in regions with very high HIV-1 prevalence. There are excellent precedents or foundation stones for such activities (43, 44). As one example, in the 1960s, the NCI and Makerere University Medical School in Kampala, Uganda, established a collaboration to perform multidisciplinary research in endemic (Epstein-Barr virus–associated) Burkitt’s lymphoma, which became a curable childhood tumor. In another example, NIAID scientists have been active in Mali for many years, and currently NIAID staff scientists are deployed in the capital at the University of Bamako, advancing research in tropical and infectious diseases in endemic areas, including HIV-1/AIDS. It is possible that a greatly expanded network of such collaborations in HIV-1/AIDS could, in effect, create an NIH “intramural program without walls” in the regions of the world in which the biology of newly discovered retroviruses has special global public health implications. Each of these models for translational medicine could allow participants to engage in laboratory research, new drug or diagnostic discovery, epidemiological surveillance, proof-of-concept clinical trials, mentoring, and investigator travel exchanges on a broad scale.

We have made considerable progress in the past 25 years. However, pathogenic retroviruses and our response to them remain unfinished business.


  • Citation: S. Broder, Twenty-five years of translational medicine in antiretroviral therapy: Promises to keep. Sci. Transl. Med. 2, 39ps33 (2010).

References and Notes

  1. Disclosure: S. Broder is a former director of the NCI (1989–95) and currently an employee of Celera Corporation, which has commercial programs in molecular diagnostics for HIV-1 and other medical conditions, including heart disease and cancer. He is also an inventor or co-inventor on U.S. government patents for dideoxynucleosides discovered in his laboratory while employed within the intramural program of NIH. All rights, title, and interest to those patents were assigned to the government, which gives a part of the royalties it receives to its employee inventors, under terms of the Federal Technology Transfer Act of 1986.
View Abstract

Navigate This Article