Spring 2000
New York, NY

NIH-Funded AIDS Vaccine Research: A Critical Review
by Gregg Gonsalves

Credits & Acknowledgments
Gregg Gonsalves is the Policy Director for the Treatment Action Group. He is the author of TAG's Basic Research on HIV Infection: A Report from the Front and the co-author of TAG's AIDS Research at the NIH: A Critical Review and Problems with Protease Inhibitor Development Plans. He is a member of the Office of AIDS Research Advisory Council.

The Treatment Action Group (TAG) fights to find a cure for AIDS and to ensure that all people living with HIV receive the necessary treatment, care, and information they need to save their lives. TAG focuses on the AIDS research effort, both public and private, the drug development process, and our nation's health care delivery systems. We meet with researchers, pharmaceutical companies, and government officials, and resort when necessary to acts of civil disobedience, or to acts of Congress. We strive to develop the scientific and political expertise needed to transform policy. TAG is committed to working for and with all communities affected by HIV.

Acknowledgements
First thanks go to Sam Avrett without whom this report never would have happened. Thanks also go to Mark Harrington, Rose McCullough, Bill Snow, David Gold, Mike Powell and Luís Santiago for reading and commenting on various versions of this report. This report was made possible with a generous grant from the Lloyd Foundation. TAG and AIDS Vaccine Advocacy Coalition (AVAC) conducted joint work to gather data, interview researchers, and analyze data about NIH funded research in 1998. In January 2000 AVAC and TAG agreed both organizations could separately use the jointly developed material in any way that supports their core missions. Uses made by either organization may not reflect the position or policies of the other organization. The joint portion of the work was done by Gregg Gonsalves for TAG and by Sam Avrett, for AVAC.

Getting Involved
If you would like more information on how to get involved in advocacy efforts on AIDS vaccine research contact:

Treatment Action Group
350 Seventh Ave., Ste. #1603
New York, New York 10036
Phone: 212.972.9022
Fax: 212.971.9019
E-mail: ggonsalves@msn.com

* * *

Introduction
The first two decades of AIDS vaccine research have been a series of disappointments and setbacks. Indeed, while at the epidemic's outset many predicted that it would be easier to develop a vaccine for HIV than effective treatments, the reverse has been the case. Initial approaches to HIV vaccine development foundered due to the unexpected complexity of HIV-1 as an immunogen and, in part, due to somewhat simplistic research approaches. Now, however, with new insights into HIV pathogenesis, new research tools, new resources, and new commitment from the U.S. government, from research administrators at the National Institutes of Health (NIH), and from the non-governmental and multilateral sectors, AIDS vaccine research is finally getting the attention, resources, and emphasis it deserves.

Our goal in this report is to outline the scientific and practical obstacles in the path of developing a safe, effective, globally deployable AIDS vaccine, to examine the AIDS vaccine research programs funded by NIH -- taking 1998 (the last year for which complete data are available) as a single, comprehensive snapshot -- and to recommend methods of overcoming those obstacles to accelerate the discovery, development, and deployment of an effective vaccine.

We believe that a vaccine is most likely to emerge from a creative and rigorous synthesis of basic research in human and non-human primate immunology and in HIV virology, with animal and clinical studies of vaccine candidates, delivery routes, adjuvants, and the like. We hope that by examining the scientific issues faced in basic, animal, and clinical HIV vaccine research, we can contribute to overcoming the obstacles and thus contribute to a revitalized, accelerated, intensified effort.

Materials & Methods
We surveyed the published scientific literature to get a sense of the progress in the field and the unanswered questions that are facing the scientific community. We also interviewed more than two dozen leading researchers,¹ asking them what they thought were the most pressing scientific issues in vaccine research and what practical problems might be impeding progress towards developing a vaccine for HIV infection. Finally, we pored over the listings and abstracts of grants funded with AIDS vaccine research dollars during federal fiscal year (FY) 1998. We looked at FY 1998 because that was the most recent year for which complete records were available from the NIH's AIDS Research Information System (ARIS) and Computer Retrieval of Information on Scientific Projects (CRISP) databases.² Whenever possible, we have tried to comment on newer projects and programs when they are relevant.

We have divided the report into three major sections: basic research, animal-based research, and clinical trials. Several major recommendations and overarching themes emerged from our analysis. These major themes include advocating for 1) continuing increases in research funding for all three major areas of HIV vaccine research, 2) intensified investment in human immunology, 3) an expanded, more cohesive animal-based research program, and 4) maintaining a strong, scientifically rigorous domestic and international clinical research program.

It is of great concern to us that our findings and recommendations echo many of those made by researchers and advocates about the NIH AIDS vaccine program in the Levine report almost four years ago. While the NIH AIDS vaccine program has grown substantially in recent years and been considerably restructured, there are lingering problems in the program that have yet to be adequately addressed.

Major Themes & Recommendations

Major Theme I: Increase the Overall Investment in Vaccine Research
Finding a way to control the global AIDS epidemic should be one of the world's highest public health priorities. Although behavioral interventions and public health measures such as safer sex campaigns, antiretroviral therapy initiated during pregnancy, and the surveillance of the blood supply remain our best current options, HIV vaccines may be the only cost-effective way to ultimately end transmission of HIV worldwide.

Despite increased attention in the media, speeches from President Clinton, and increased budgets for HIV vaccine research, government and private investor funding for HIV vaccine research is still relatively small. Funding in 1998 to research and develop HIV vaccines by NIH and all other U.S. agencies amounted to only about 10% of overall AIDS research funding. However, much HIV pathogenesis, natural history, and epidemiology research is also likely to contribute to the development of an HIV vaccine.

The $148 million spent by NIH in 1998 for HIV vaccine research was widely distributed, supporting research through 647 grants and contracts to more than 140 institutions and easily more than 1,000 researchers. NIH funding thus encompassed an enormously broad range of researcher ideas. Although widely distributed, NIH funding wasn't randomly scattered. The 1998 NIH effort included several targeted, concerted funding initiatives to address critical research questions, refine research techniques, make new assays and reagents widely available, and support networks and collaborations.

We believe the NIH AIDS vaccine research effort -- including its basic, animal, and clinical aspects -- can utilize continuing funding increases in line with those obtained in the past few years. Where will this money come from? In recent years, NIH has received unprecedented increases in its annual budget. AIDS research too has continued to grow at a robust rate. With the NIH budget set to double by 2002, additional spending on AIDS vaccine research should be possible without reducing resources devoted to other areas of research on HIV infection, or on other high priority basic and clinical research areas.

HIV Vaccine Basic Research

Major Theme II: Increase the Investment in Human Immunology
While great strides have been made in basic research on HIV infection, there are many unanswered questions about what kind of immune response to HIV a successful vaccine will have to induce and about how to engender that response. Right now, we cannot induce long-lasting, broadly cross-reactive, vigorous antibody and cellular immune responses in humans with any of the current vaccine candidates in development.³ While there is a healthy investment in immunology at NIH, most of that work is being done in mice. A robust, new investment in clinical immunology is needed to understand how we can generate effector and memory neutralizing antibody and cytotoxic T-cell responses at mucosal sites.⁴ [This too was a key, still incompletely implemented, recommendation of the 1996 Levine Committee report.]

1. Increase Support for Research on Mucosal Immunity
   NIH should increase resources devoted to mucosal immunology and vaccinology, and should help bring new researchers into the field by initiating new training programs in this area. In addition, clinical research on mucosal delivery of new candidate vaccines should be made a priority by the new Vaccine Trials Network.

   Increase Support for Studies of Antigen Delivery in Primates and Humans

   NIH needs to support additional work on antigen delivery in primates and humans. NIH should expand access to monkeys through supporting new breeding programs and investing in the development and production of reagents for primate experiments. Study sections need to be more generous in the level of funding they will support for primate studies. Resources devoted to clinical and human immunology need to be increased, with a special emphasis on bringing young physicians into the field to study some of these unanswered questions in a clinical setting.

   Increase Support for Studies of Immunological Memory

   With the difficulty in sustaining both antibody and cellular responses with current immunogens, NIH should invest additional resources in the study of immunological memory in humans. NIH should coordinate its non-AIDS work in this area with basic research on HIV vaccines.

   Accelerate Research on Novel Envelope-based Immunogens

   NIH needs to break free of the tether of outdated approaches to virus neutralization using monomeric subunit vaccines in both basic and clinical research. While several groups are working on better envelope-based immunogens, research in this area can be expedited if NIH supports fast-track testing of these new products in macaques and in humans.

HIV Vaccine Animal-based Research

Major Theme III: Create an Expanded, More Cohesive Animal-based Research Program
The SIV and SHIV macaque models offers important tools to address many basic scientific questions about the primate immune system, the immune responses to retroviral infection, and the relative efficacy of different vaccine concepts. Unfortunately, using macaques in research is expensive; these animals are a precious and scarce resource. To date, most of the studies performed in macaques have been small, pilot studies using different assays, different vaccine candidates, different routes of challenge, different challenge viruses, and so on. With small numbers and varying research conditions, it is hard to compare animal-based research studies and get a sense of the big picture. NIH is just now beginning to initiate a large, comparative study of four vaccine approaches in macaques. This is a good first step, but NIH needs to make greater resources available for animal-based research and to encourage a coordinated, rational approach to animal studies. The National Center for Research Resources (NCRR) has been negligent and unresponsive to criticisms from the Levine Committee and others regarding the administration of the Regional Primate Research Centers (RPRCs) and other primate colonies. NCRR should be catalyzing progress in this area rather than defending its bureaucratic turf.

5. Increase Overall Resources for Animal Research
   NIH needs to enhance its support for virological and immunological studies of retroviral infections in primates.

   Increase Resources per Research Study

   Program staff at the National Institute for Allergy & Infectious Diseases (NIAID) and at NCRR need to direct additional funding to support studies with sufficient statistical power to offer reliable answers to research questions. Researchers should design studies to include larger numbers of primates (i.e., more than 3-4 per arm) and ask for this funding in their applications. Study sections need to look favorably on larger primate studies.

   Rationalize the Simian/SIV Model

   Large, comparative studies of vaccines in macaques are critical to assess and optimize candidate immunogens for further testing in humans. NIH should streamline the mechanisms for initiating these trials and provide sufficient resources for their expeditious conduct so that the information gathered can be quickly integrated into decision

   Build Understanding of Protection by Live, Attenuated SIV

   Live, attenuated SIV offers an important model for understanding the correlates of protection conferred by vaccination. More research is needed to look at a wide range of systemic and mucosal immune response in this model.

HIV Vaccine Clinical Trials

Major Theme IV: Maintain a Strong, Scientifically Rigorous Domestic and International Clinical Research Program
NIH has supported clinical research on HIV vaccines for more than a decade. The programs supported have evaluated candidate vaccines supplied by industry for both safety and immunogenicity. So far, NIH has not initiated any phase III efficacy studies (a single, industry-sponsored international study of the VaxGen gp120 subunit envelope protein is ongoing). The current phase I-II NIH-supported clinical trials programs need to be maintained in the United States and abroad; however, these programs need to refocus their efforts on phase I and II studies of novel vaccine concepts rather than pursuing further study of vaccine candidates that have shown little promise in preclinical, animal, or early human testing. In addition, the clinical trials programs should continue developing closer ties to the basic scientific community to ensure more bi-directional collaboration between bench and clinic in HIV vaccine research.

9. Fund Development of New Vaccine Products and Immunization Strategies
   NIH needs to directly fund development of products to feed into clinical evaluation. NIH should expand direct contracting with companies to develop HIV vaccine candidates, such as newer versions of MVA, VEE, AAV, and attenuated salmonella vectors. The NIH should encourage optimization of current vaccine candidates and be willing to discourage work in outdated vaccine concepts that have performed poorly in preclinical, animal or early human studies.

   Maintain a Solid Clinical Research Infrastructure

   NIH should ensure that the new HIV Vaccine Trials Network (VTN) and the vaccine clinical trial activities of the Vaccine Research Center (VRC), the NIH intramural trial site, and the perinatal sites will have the capacity for an adequate number of phase I, II, and III trials during the coming five years. That being said, the clinical research programs for vaccines should be held to the highest scientific standards with adequate oversight through regular review by extramural researchers. The advancement of candidate immunogens along the vaccine clinical development pipeline has to be driven by science and not by political expediency or industry demands.

   Rigorously Evaluate Immunology gt;amp; Basic Science within Clinical Trials

   A coherent plan with interim goals need to be articulated for the clinical trials program so that clinical trials evaluate immunization strategies in a concerted, systematic manner. Expansion of human immunology research, linked to HIV vaccine clinical trials, is needed. Larger, more intensive phase I and II trials should compare vaccine products and immunization strategies and the resulting immune responses.

   Support a Strong International Vaccine Development Effort

   NIH should work more closely with the Centers for Disease Control (CDC), the U.S. Agency for International Development (USAID), and the Department of Defense (DOD) to support the development and maintenance of biomedical research infrastructure in developing countries.

I. HIV Vaccine Basic Research
Basic research provides the groundwork for developing a vaccine against the virus by advancing our understanding of the immune response to HIV. In 1998, NIH funded a large amount of basic research focused on understanding HIV antigens and human responses to those antigens. Major work explored the mechanisms of humoral and cellular immunity, both systemically and at mucosal surfaces. Other important work helped to define the structure and immunogenicity of HIV envelope and other HIV proteins, and to understand the presentation and processing of HIV antigens by the human immune system. About $37 million, or 25%, of the $148 million spent by NIH on HIV vaccine research was dedicated to basic research. This funding mostly came from NIAID (73%), with the remaining funding coming from NCRR (15%), the National Cancer Institute (NCI) (12%) and other institutes. Funding for basic research related to HIV vaccines was largely extramural (95%). More than 70% of the funding went to extramural investigator-initiated grants. Although the intramural AIDS vaccine program at NIH is small, there is a talented cadre of investigators in this area on the NIH campus. The intramural effort is due for significant expansion in the years ahead with the establishment of the new Vaccine Research Center (VRC) in Bethesda. 1998 was an unusual year for grant funding, since two cycles of the AIDS Vaccine Research Committee (AVRC)-initiated Vaccine Innovation Grants (R21s) were funded in this year, accounting for as much funding as HIV vaccine-related R01s (for explanations of the award types, see pages 29-31).

Basic HIV vaccine research needs a vigorous infusion of new funding support. The figures above indicate the relative paucity of research being done on mucosal immunity and antigen presentation ($7.6 million and $3.8 million, respectively). It is also obvious that the R21 innovation grants have helped attract high-quality investigators to selected topics; in the 1998 funding cycle these R21s represented 38% of the NIH basic HIV vaccine research portfolio. These are only two-year awards, however, and NIH needs to consider how best to support longer-term research in these target areas and in others indicated throughout this report.

Scientific Issues

The Mucosal Immune Response
The infection and the immune response to HIV begin together at the frontier between the outside world and the body, which is, in the case of sexual transmission, the mucosal epithelia. HIV can cross the epithelial barrier through a process known as transcytosis or through active transport by M cells which are both notably susceptible to antibody blockade^5,6 although other studies suggest that intraepithelial Langerhans cells are first infected by the virus and initiate its spread. Once past the epithelia, HIV is thought to be picked up by antigen presenting cells (APCs), primarily dendritic cells (DCs). Newer work, however, suggests that HIV directly infects CD4+ T lymphocytes in mucosal associated lymphoid tissue (MALT), establishes a fulminant local infection within a few days, and then spreads quickly throughout the body.^8,9 Emerging data suggest that viral reservoirs are established early during mucosal infection, and thus predict that preventive vaccines must engender a quick and vigorous mucosal immune response.

What kind of mucosal immune response will be needed to block HIV infection, and how might it be achieved? One key issue, still controversial, is whether most mucosal transmission takes place in the form of cell-free or cell-associated virus. Good research exists supporting the physiological relevance of both approaches.^7,10,11 Most likely it is a mix of both mechanisms; however, the relative frequency of each is unclear. Cell-free infection may be easier to block with an antibody response, while a cell-associated infection is likely to require cell-mediated immune responses in addition to antibodies.

Although several non-human primate studies have demonstrated protection from mucosal infection with SIV, SHIV, and HIV-2, none has been able to determine the correlates of this protective immunity.^{12,13,14,15,16} Antibody-mediated immune responses could potentially block HIV, at the epithelial surface or in intraepithelial space before it infects T cells or is picked up for transport to local lymphoid tissues, by neutralizing virus intracellularly or by interfering with assembly of virus particles, if cells are infected at these sites.¹⁷ While intravenous challenge experiments have established that only high neutralizing antibody titers will protect against infection by this route, the characteristics of a successful mucosal antibody response to HIV or SIV remain to be established.¹⁸ Mucosal IgA is likely to be important based on evidence from other viral infections,^19,20 studies of individuals exposed to, but uninfected with HIV^21,23 and in vitro studies.²³ IgG may be useful, as it is plentiful at mucosal sites, and has been shown to neutralize HIV and to mediate antibody-dependent cellular cytotoxicity (ADCC), at least in serum.²⁴

Even if a vaccine-induced antibody response is not sufficient to contain HIV infection by itself, this response may slow the progress of HIV transmission enough to allow vaccine-induced cellular responses to clear the nascent HIV infection. There is some evidence that mucosal cytotoxic T lymphocytes (CTLs) are involved in protection against rectal and vaginal challenge with SIV. Recent work by Cromwell and colleagues²⁵ has shown that immunization with live, attenuated SIV, which protects macaques from infection with pathogenic strains of the virus, is associated with the presence of CTLs expressing the alpha 4/beta 7 intestinal homing integrin in blood, lymph node, duodenum, and colon. By contrast, immunizations with a combination of DNA and modified vaccinia virus Ankara (MVA) vaccines did not elicit these CTLs.

There may be additional mechanisms of protection from HIV infection at mucosal sites. Recent studies demonstrating protection from rectal challenge with SIV after targeted iliac lymph node immunization with recombinant simian immunodeficiency virus gp120 and p27 indicated an association with elevated levels of beta-chemokines, including RANTES, MIP-1 alpha and MIP-1 beta, and a suppressor factor produced by CD8+ T cells.²⁶

Since most HIV infections worldwide occur through sexual contact with infected individuals, the immune response in the genital mucosa will be critical to prevent infection. Yet this area of research remains grossly understudied. Basic principles about the mucosal immunology of the genital tract are not well understood: Where do the humoral and cellular immune responses in this part of the body originate? How long do they last? How do we induce local immune responses? Can we immunize systemically and generate a mucosal response? What cell types are key players, and what are the patterns of cytokine and chemokine expression in a healthy host and during inflammation and infection with other micro-organisms?

Antigen Presentation
Both humoral and cellular immune responses depend on the efficient and proper presentation of antigen to T and B lymphocytes. CD4+ T cells recognize fragments of antigen processed through the endocytic pathway and in complex with major histocompatibility complex encoded class II proteins. CD8+ T cells recognize fragments of antigen processed through the cytoplasmic pathway and in complex with MHC class I encoded proteins. For their part, B cells can recognize unprocessed, intact antigen through their immunoglobulin receptors. Antigens are presented to lymphocytes by specialized antigen presenting cells (APCs) including dendritic cells (DCs), by phagocytic cells such as macrophages, and by B cells themselves, which need feedback from CD4+ T cells to differentiate into plasma cells and secrete antibodies.

HIV subverts the antigen presentation process in a multitude of ways. The HIV-1 tat protein inhibits inhibit the production of MHC class I proteins.²⁷ The HIV-1 nef protein triggers the endocytosis of these proteins from the surface of the cell.²⁸ The HIV-1 vpu protein induces the degradation of these proteins in the endoplasmic reticulum.²⁹ Mutations in HIV can interfere with the binding of viral peptides to MHC proteins or to T-cell receptors (TCR); thus, these mutations can permit viral escape from CTL recognition.³⁰ In addition, these new viral variants also "can differentially activate CTL, act as CTL decoys, induce anergy, distort the CTL repertoire or antagonize the response to normal antigen."³¹

The primary loci for antigen presentation are the lymphoid organs, which are also the primary loci for HIV replication. Follicular dendritic cells (FDCs) in the germinal centers of the lymph nodes and mature germinal center dendritic cells (GCDCs) can transmit virus to CD4+ T cells respectively through virions complexed with antibody and virions captured on Fc receptors.^32,33 In these ways, HIV subverts the machinery of antigen presentation to perpetuate its own spread and survival yet, despite this interference by the virus, antigen presentation remains a key component of the immune response to HIV in infected individuals. The manner in which immunogens are presented to the immune system will likely be an important factor in the success of any AIDS vaccine. DCs are likely to be responsible for initiating the strong CTL responses seen in HIV infection by stimulating CD8+ T cells with HIV peptides complexed with MHC class I proteins. DCs may be infected in MALT, present peptides from noninfectious virions or viral debris, or present peptides from phagocytosed, apoptotic cells.^34,35 Thus, an effective vaccine is likely to require early mobilization of a DC response.

Basic research on antigen presentation offers clues to what kind of vaccine might be best in inducing a protective immune response; however, much of this work to date has been done in mice. Live, attenuated viruses are ideal candidate immunogens because they can present intact, native virions to B cells, inducing neutralizing antibody responses, and shuttling viral proteins into the endocytic pathway for presentation in complex with MHC class II proteins to CD4+ T cells, which drive differentiation and maturation of antibody and CTL responses. Live, attenuated viruses can also infect cells, allowing viral proteins and RNA to enter the cytoplasmic pathway and be presented in complex with MHC class I proteins to CD8+ CTLs. Other forms of vaccines are theoretically less ideal from the standpoint of antigen presentation. Live viral vectors that replicate poorly or for only a single round of replication may not provide enough peptides for uptake by antigen processing mechanisms and may blunt the maturation of DCs by limiting the inflammatory response that accompanies a robust viral infection. DCs play an integral role in DNA vaccination. They present antigen from neighboring, transfected cells and can be directly transfected themselves by the plasmid immunogen. One current challenge is to target more DCs for transfection or to get them to take up more antigen from neighboring cells.^36,37

The Cellular Immune Response
Most viral infections, perhaps with the notable exceptions of hepatitis B and rabies,³⁸ are brought under control by the cellular arm of the immune system with or without the assistance of an antibody response. While the correlates of immunity to HIV infection are still incompletely understood, eliciting cytotoxic T-cell (CTL) responses to HIV is likely to be an important goal for a preventive AIDS vaccine. What evidence do we have that suggests this? Studies in macaques vaccinated with a live, attenuated SIV and protected from challenge with a pathogenic strain of the virus appeared to correlate protection with the presence of SIV-specific CTLs, but not of neutralizing antibody.^39,40,41 More recent studies, however, have shown that SIV-specific CTLs rise after viral challenge in protected monkeys,⁴² which complicates any assumptions about the role of these cells in conferring protection. Thus, the anti-HIV CTL response may be necessary but not sufficient (CD4 T-cell help may be required, for example). Additionally, in both acute and chronic SIV infection, CTLs seem to be critical in controlling viremia. In primary infection, macaques treated with an antibody to CD8+ lymphocytes do not bring initial viremia under control and have a rapid disease course.⁴³ Monkeys with chronic SIV infection have a burst of viremia if they are treated with anti-CD8+ lymphocyte antibodies.⁴⁴ Using the relatively new MHC class I/peptide tetramer-binding assay, the emergence of a single epitope-specific CTL population constituting anywhere from 1.3% to 8.3% of the total peripheral pool of CD8+ T cells was shown to correlate with the clearance of the initial viremia in acute SIV infection of macaques.⁴⁵ Using older CTL assays in studies of acute HIV infection, the presence of env-specific CTLs has been shown to be associated with lower viral loads and delayed disease progression.⁴⁶ In a cohort of HIV-infected hemophiliacs, gag-specific CTLs were similarly associated with delayed progression.⁴⁷ These observations have been confirmed using the new tetramer technology.⁴⁸ The importance of the CTL response in two special populations of individuals-long-term non-progressors (LTNPs) and exposed uninfecteds-has also been extensively documented. In several studies, LTNPs have been found to have strong CTL responses to various HIV epitopes in association with low viral loads.^49,50 EU individuals also seem to have robust CTL responses.^51,52,53 Recently, three HIV-infected individuals who stopped antiviral therapy but maintained viral suppression were shown to have vigorous and broad CTL responses with or without strong neutralizing antibody titers.⁵⁴

Despite the evidence suggesting the potential of a vigorous HIV-specific CTL response to prevent infection with the virus, CD4+ T-cell helper responses generally sustain CTL memory responses and promote active CTL effector responses^55,56,57,58 and may be required for an effective vaccine. The importance of HIV-specific CD4+ T-cell helper responses is underscored by recent work suggesting that these responses help suppress viral replication in LTNPs, and that HIV-specific CD4+ helper cell responses can be rescued from viral destruction by treating acutely infected patients with highly active antiretroviral therapy (HAART).⁵⁹ Another more recent study, however, has shown that these HIV-specific CD4+ cell responses persist in untreated patients with chronic, progressive HIV infection and are not correlated with viral load.⁶⁰ [The two studies used different assays, the first using lymphocyte proliferation assays (LPA) and the second cytokine staining.] These studies complicate our understanding of the role of CD4+ T cells in the immune defense against HIV.

The Antibody Response
Why might it be important to have a potent antibody response to HIV? First of all, important experiments demonstrated that chimpanzees and monkeys can be protected by neutralizing HIV-1 antibodies from infection with lab strains of HIV itself or chimeric SHIV, which carries the core components of SIV and the envelope of primary HIV isolates. In 1992 Emini and colleagues⁶¹ protected chimps from infection with the classic lab strain of HIV-1, HIVIIIB, after passive transfer of neutralizing antibodies directed at the V3 loop of the viral envelope. More recent work by Shibata and colleagues⁶² has shown that purified immunoglobulin (IgG) derived from chimpanzees infected with a dual-tropic strain of HIV could block infection of pigtailed macaques by an SHIV with the envelope of that same primary isolate.

Antibody neutralization and prevention of infection by primary HIV isolates, those strains most closely resembling viruses encountered in naturally occurring infections, has also been shown in hu-PBL-SCID mice. Gauduin and colleagues⁶³ were able to completely protect mice from infection with two primary HIV isolates via passive transfer of a monoclonal antibody, IgG1b12, several hours after viral challenge.

Although protection of animals by antibodies has been accomplished in several settings, repeating the same success in humans will be a difficult task. Why is this the case? Primary isolates of HIV are notoriously difficult to neutralize. Even antibodies recovered from people with HIV have only a weak ability to neutralize their own autologous viruses or a panel of other primary isolates.⁶⁴ There are several reasons for this phenomenon. First, the viral envelope is swathed in relatively non-immunogenic glycoproteins that may provide a shield for the virus against recognition by the immune system.⁶⁵ Second, the envelope proteins of primary isolates of HIV are distinctly different in conformation or structure from strains of the virus that have been passaged in culture in the laboratory. The envelope of lab strains, or T cell line-adapted (TCLA) strains of HIV, seems to have more exposed epitopes or sites for immune attack than the strains that circulate in people with HIV. In fact, only three epitopes have been found to be accessible for neutralization on a broad range of primary viruses-b12, 2G12, and 2F5. Figure 1 below represents the envelopes of TCLA and primary isolates of HIV and the various epitopes that are available for neutralization or are structurally obscured.⁶⁶ The envelope of the primary isolates seems to have parts of the V1/V2 and V3 loops, as well as the CD4 binding site, masked from recognition. Others have also suggested that the V1/V2 and V3 loops might obscure the CD4 binding and the chemokine receptor binding sites, respectively.⁶⁷

Figure 1: Conformational Differences in HIV-1 Membrane Glycoproteins

Third, the humoral immune response in vivo may be directed to viral debris rather than the envelope of intact virions.⁶⁸ Many antibodies from people with HIV bind to unprocessed gp160 envelope precursor, "shed" monomeric gp120 dissociated from the virion/infected cell surface, and gp41 exposed on the virion/infected cell surface after shedding of gp120. Most of these antibodies have extremely low affinity for binding to mature oligomeric envelope.

Most current vaccine candidates elicit at best disappointing levels neutralization to primary HIV isolates. Most subunit vaccines based on monomeric gp120 or gp160 fail to induce neutralizing antibody responses to mature envelope proteins on intact virions from primary isolates because epitope exposure on the recombinant proteins differs significantly from that on circulating virions. Moreover, other vaccine strategies such as live recombinant vectors (e.g., poxviruses) and naked DNA have failed to demonstrate success in generating relevant antibody responses.⁶⁹ Several strategies are being pursued to design immunogens that will induce better humoral immune responses. One is to try to design envelope proteins mimicking the oligomeric native envelope on primary isolates while removing features of the molecule that help it evade the immune response.^70,71 Some labs are trying to remove the hypervariable loops (i.e., V1/2 and V3) to expose hidden epitopes and to redirect the immune response away from these highly mutable regions-with varying degrees of success. , Others are trying to strip away the sugar molecules from the envelope to expose the more immunogenic epitopes below.^72,73 Recently, Nunberg and colleagues at the University of Montana devised an immunogen based on a transitional form of the envelope protein that induced neutralization of 23 out of 24 strains of HIV representing primary isolates from North America, Europe, Africa, India, and Thailand.⁷⁶ Nunberg crafted his immunogen by taking cells expressing an envelope protein from a T-cell-tropic strain of virus from an HIV-infected person in an Amsterdam cohort study and another set of cells expressing both the CD4 and CCR5 receptors, letting them begin to fuse together and then freezing them in this transitional state with formaldehyde. All these efforts are indebted to the recent groundbreaking work elucidating the structure of the viral envelope by several teams of researchers around the country.^77,78,79

In addition to defining targets on primary isolates susceptible to attack by antibodies and designing immunogens that can induce neutralization of these viruses, there may be a need to make sure that these humoral immune responses are relatively long-lasting. In a recent study, the antibodies induced by repeated immunizations with a recombinant gp120 lingered only for 40-60 days.⁸⁰ Live viral or bacterial vectors expressing HIV envelope may be needed in addition to simple recombinant proteins in order to sustain antibody responses over the long haul.

Basic Research Recommendations

1. Increase Support for Studies of Mucosal Immunity
HIV infection is largely (~ 80-90%) a mucosally transmitted disease,¹⁷ and the induction of potent immune responses in the genital tract may be essential in the development of an AIDS vaccine. Still, basic principles about the mucosal immunology of the genital tract are not well understood: Where do the humoral and cellular immune responses in this part of the body originate? How long do they last? How do we induce local immune responses? Can we immunize systemically and generate a mucosal response?

Despite the importance of achieving protection at mucosal sites, most human vaccine studies to date have involved intramuscular or intraveneous inoculation. The few completed studies using mucosal delivery have not shown immunogenicity or antibody production in local secretions.⁸¹ The huge, indigenous microflora in mucosa sites may make it difficult for vaccine candidates with low expression of HIV antigens to engender strong immune responses there; again, a vigorous, local infection may need to be set up to initiate a vigorous immune response.

NIH has several programs in mucosal immunity, including NIAID's Cooperative Mucosal Immunology Group for Investigations on AIDS Vaccines, and other more specific initiatives, focusing on the oral and gastrointestinal mucosa, at the National Institute for Dental Research (NIDR) and the National Institute for Diabetes, Digestive and Kidney Diseases (NIDDK), but the scope of NIH's efforts in this area is dwarfed by the need for increased understanding of immunology at mucosal sites and the need for better tools to assess mucosal responses.

NIH should increase resources devoted to mucosal immunology and vaccinology and should endeavor to bring new researchers into the field by initiating new training programs in this area. Clinical research on mucosal delivery of new candidate vaccines should be made a priority by the new Vaccine Trials Network.

2. Increase Support for Studies of Antigen Delivery in Primates and Humans
What is needed to initiate and sustain an immune response to HIV? How are HIV antigens taken up and expressed? Can this process be upregulated or optimized? How much HIV antigen needs to be expressed and for how long? These questions about antigen presentation and the questions outlined above regarding immunological memory are inextricably linked. While much basic research on antigen presentation is being supported by NIH, the vast majority of this work is being done in mice. In the era of molecular biology, a lot of resources are devoted to antigen identification, cloning genes and expressing protein while far fewer resources are devoted to the process of targeting antigen for delivery to macrophages and dendritic cells.

NIH needs to support additional work in antigen delivery in primates and humans. NIH should expand access to monkeys through supporting new breeding programs and investing in the development and production of reagents for primate experiments. Study sections need to be more generous in the level of funding they will support for primate studies. Resources devoted to clinical and human immunology need to be increased with a special emphasis on bringing young physicians into the field to study these unanswered questions in a clinical setting.

3. Increase Support for Studies of Immunological Memory
While eliciting broadly cross-reactive and long-lived CTL responses is one of the major goals of basic AIDS vaccine research, there is considerable uncertainty about how to accomplish this task. Do persistent HIV antigens need to be provided to maintain active, effector T cells against the virus, or will a robust initial expansion of T cells engender a robust memory response? One thing is certain, we will need new vaccine candidates that can induce a vigorous, initial infection (with HIV, rather than vector, antigens dominating the immune response) because the CTL responses generated with the immunogens furthest along in clinical testing have been disappointing. However, the newer, live vectors such as MVA and AAV and some DNA-based immunogens have shown promising results in monkey models.⁸²

There are many persistent questions about the role of CTLs in mediating protection from HIV infection or disease progression, although the measurement of CTL responses is clearly undergoing a revolution. Researchers can now actually quantitate CTL responses in vivo with the new tetramer and elispot assays, but does the binding of cells to MHC-peptide tetramers correlate with actual CTL killing? New studies suggest there may be striking functional defects even in CTLs that can bind tetramer and stain positive in intracytoplasmic cytokine assays.^83,84 More work is clearly required to refine the new assays to better understand what is going on in the cellular compartment of the immune system, while simultaneously trying to bring standards of measurement to the field.

Recent, persuasive work in mice has been done which establishes that CD8+ and CD4+ memory T cells do not need constant ongoing antigenic stimulation to persist and recognize their cognate antigen (in these cases, murine lymphocytic choriomeningitis virus, LCMV, and pigeon cytochrome).^85,86 The relevance of this work to humans, the nature of the positive signals that allow memory T cells to survive, and whether we can manipulate the cells or innate immune processes to increase their numbers need to be established.

In 1996, the Levine report noted a dearth in funding for human immunology and made a major recommendation to NIH to boost work in this area of research.⁸⁷ In addition, the 1998 Jordan report outlined several areas of scientific opportunity in basic immunology that could advance our search for vaccines for many diseases, including HIV.⁸⁸ NIH has made several important new efforts to increase its efforts in human immunology, but they are still too small. NIH now supports a handful of Centers for Excellence in Human Immunology, but the awards for these sites are limited (approximately $3 million in total funding for seven centers has been allocated for this initiative). A single round of innovation grants in human immunology has also been supported by the OAR with $6 million from its discretionary fund.

With the difficulty in sustaining both antibody and cellular responses with current immunogens, NIH should invest additional resources in the study of immunological memory in humans. NIH should coordinate its non-AIDS work in this area with basic research on HIV vaccines.

4. Accelerate the Development of New Envelope-based Immunogens
Eliciting antibodies that neutralize a broad range of primary isolates is one of the major goals of basic AIDS vaccine research, and will be especially needed if an intransigent pool of virus is established early after transmission. This task may not be accomplished simply by learning how to present envelope proteins in their mature, native form. While the recombinant subunit vaccines produced to date have only offered monomeric proteins in their non-native conformation, just mimicking what is expressed on the viral envelope in vivo is unlikely to elicit a protective response. The envelope may have to be modified to make it more immunogenic and molecularly tinkered with to induce strong, neutralizing responses. Additional structural studies of the envelope in conjunction with cellular receptors could be very important in reaching this goal. In the studies described earlier by Shibata et al., the serum raised against the primary isolate in chimps had to be able to neutralize 100% of the challenge SHIV virus in vitro before protection was achieved in macaques passively infused with the antibody preparation. This daunting figure clearly demonstrates how difficult it will be to develop an effective HIV vaccine whose only mode of activity is the induction of neutralizing antibody responses. Furthermore, if a potent antibody response against vulnerable epitopes develops too slowly, it may not be capable of preventing the virus from gaining a foothold in the body.

NIH has used the Innovations Grants program (expedited two-year awards with a maximum of $150,000 awarded per year) to spur new work on the viral envelope, but it ended up initially rejecting support for Jack Nunberg's pioneering work (which was initially funded by amfAR, the American Foundation for AIDS Research).

NIH needs to break free of the tether of outdated approaches to virus neutralization using monomeric subunit vaccines both in basic and clinical research. While several groups are working on better envelope-based immunogens, research in this area can be expedited by supporting fast-track testing of these new products in macaques and in humans.

* * *

II. HIV Vaccine Animal-based Research

Animal-based HIV vaccine research evaluates the safety, immunogenicity, and efficacy of immunization strategies in animal/virus models that resemble human HIV exposure, infection, and disease as closely as possible. During past years, NIH has annually funded more than $30 million of HIV vaccine research in animal models. In 1998, the funding level rose to $37 million, constituting by far the largest source of worldwide funding for HIV vaccine research in animal models. Use of animal models in HIV vaccine development has greatly increased our understanding of possible immunization approaches, but the effort still needs increased resources, rationalization of the many animal models used, and further investigation of the results seen in current studies.

Scientific Issues

Modeling Human HIV Infection and AIDS
Animal models, particularly primate models, provide an opportunity in HIV vaccine research for defining potential approaches for HIV immunization, for gaining insight into the mechanisms of inducing immune responses, and for acquiring some sense of potential safety and immunogenicity in humans. No viral infection in an animal exactly duplicates human HIV infection and disease, but several models demonstrate similar patterns of viral infection, replication, spread, mechanisms of disease, expression of disease, and immune response. Certain animal models may ultimately predict the dynamics of HIV in humans in immune system control and protection against mucosal or intravenous infection. Until protection is demonstrated in humans, however, any correlation or predictive value of animal protection to human protection will remain unknown.

Of Monkeys, Mice, and Men

So far, the closest model for in vivo human HIV infection and disease has been the infection of macaques with SIV and with chimeric SHIV (made up of the replication machinery and core proteins of SIV with envelope and some regulatory genes from HIV-1).^89,90,91 African SIV isolated from mangabeys causes disease in Asian macaque species, including rhesus, cynomolgus, and pigtailed macaques, in a way similar to HIV in humans. This model, therefore, has been important for work in understanding early viral infection and disease, as well as antiviral immune responses. Furthermore, the envelope structure, core proteins, and coreceptor-facilitated entry of SIV are similar to HIV.⁹² Thus, candidate vaccines can be formulated with SIV antigens for evaluation in the macaque/SIV model, and with both SIV and HIV antigens in the macaque/SHIV model.⁹³

In the macaque/SIV model, there is a range of infectivity, virulence, and disease outcomes. In terms of virulence and disease outcomes, SIV variants range from those causing rapidly lethal disease, such as an SIVsm variant designated PBj14,^94,95 to those causing no short-term disease at all, such as an SIVmac attenuated with genetic deletions.⁹⁶ Pathogenesis might be different in neonate macaques than in adults, as demonstrated in a study of attenuated SIVmac showing lethality in neonates.⁹⁷ Like HIV, SIV is inefficient at infecting across mucosal surfaces.⁹⁸ For evaluation of SIV vaccines in affecting viral replication and disease progression in large groups of macaques, researchers have generally used SIVmac or SIVsm, such as SIVmac251, SIVmac239, SIVsmB670, or SIVsmE660, that cause progressive AIDS-like disease within two to three years. Researchers can also directly evaluate the immunogenicity and efficacy of certain HIV-1 vaccines in macaques by using pathogenic SHIV viruses as disease-inducing challenges.^{99,100,101,102}

Aside from the macaque/SHIV system, no animal model exists to directly evaluate immune responses generated by HIV-1 vaccines against a virus causing an AIDS-like disease. Furthermore, except for reagents for macaques and mice, development of complex and specific reagents to detect immune responses is limited. Nevertheless, work continues in several other animal species. Some HIV-1 strains can cause an AIDS-like disease in chimpanzees^103,104 and more pathogenic strains could be developed, although this model is limited by the high cost, the restricted availability of animals, and ethical concerns. Baboons can be infected in a limited way by HIV-1 and HIV-2.^105,106 Cats are infected with a lentivirus called FIV, goats with a CAEV, and horses with an EIAV. FIV-infected cats sometimes develop an AIDS-like illness. However, these viruses are much more distantly related to HIV than SIV, and background immunology and immune reagents are limited for these species.

Attempts are being made to develop transgenic mice and rabbits.^107,108,109 If successful, this would create a small (and inexpensive) animal model of HIV infection and replication. Because immunologic reagents for mice and inbred strains of mice are readily available, researchers might be able to use transgenic mice in better understanding the mechanisms of antibody, CTL, and other immune responses in controlling HIV infection and replication. It is unclear, however, how predictive small animal studies will be of human immune responses.

Protection Achieved, but How?
Research in animals has demonstrated that vaccines can induce protection against HIV-like infections. In macaques with SIV, a range of vaccine approaches have been able to clear challenge SIV, although only live, attenuated SIV is able to provide this level of protection against the more virulent SIVmac251¹¹⁰ and SIVsmB670.¹¹¹ Viral load reduction or delayed disease progression against these SIV strains has been demonstrated by vaccination with combination DNA plus vaccinia vector,¹¹² vaccinia vector plus subunit,¹¹³ and attenuated vaccinia vector plus subunit vaccines.¹¹⁴

It remains unclear from these studies which immune responses are responsible for protection, but researchers are developing tools to answer questions about the mechanisms of immune protection against pathogenic virus. In the macaque/SIV model, researchers are defining macaque MHC,¹¹⁵ and developing reagents to detect cytokines,¹¹⁶ chemokines,¹¹⁷ costimulatory molecules,¹¹⁸ and other immune responses.

Building Momentum for Discovery & Development
The animal research field is poised to evaluate and compare a range of new HIV immunization strategies in animals, to measure immune responses and outcomes, and to evaluate mechanisms of protection. To move the field forward, momentum must be maintained in three areas of the NIH effort with respect to animal studies:

• Building upon and improving animal models of the human-HIV system, including further development of the macaque/SHIV model, work to MHC type and clone macaques, establishment of specific pathogen-free macaque colonies, and further work on primate and non-primate immunology.
• Use of existing animal models to gain understanding of safety and immunogenicity of newly developed and reformulated candidate immunization strategies.
• Use of existing animal models, particularly the macaque/SIV model, to increase understanding of protection against lentivirus infection.

Figure 2 (above): Utilization of HIV Vaccine Animal Research Resources
Figure 3 (below): Utilization of HIV Immunogens in Animal Models

In 1998, NIH spent about $37 million on HIV vaccine research involving animals. Most of this funding came from NIAID ($22.7 million, or 61%), NCRR ($5.9 million, or 16%), and NCI ($3.66 million, or 10%). NIAID and NCRR accounted for 74% and 21% of primate-related spending. Overall, 1998 funding was balanced between researcher grants (46%) and extramural and intramural contracts (54%). Programmatically, in 1998, nearly 45% ($16.8 million) of the NIH's $37 million effort in animal-based vaccine research was spent in developing and improving animal models of the human/HIV system. $11 million was spent on macaque breeding contracts and macaque immunology through contracts to the Regional Primate Research Centers (RPRCs) and Simian Vaccine Evaluation Units (SVEUs) and awards to other institutions. About $4 million went to development of mouse-based models of HIV infection (Figure 2). About $5.1 million of NIH funding was dedicated to work on live, attenuated SIV and elucidation of mechanisms of protection and potential pathogenesis in adult and neonatal macaques. $15.1 million was allocated for direct evaluation of immunization strategies in animals, including work on mucosal routes of immunization,^{119, 120} work on envelope subunits,^121,122,123 and work on viral vectors (Figure 3).¹²⁴

Animal Research Recommendations

5. Increase Overall Resources for Animal Research
As the above discussion and the basic research section indicate, NIH needs to enhance support for virological and immunological studies of retroviral infections in primates.

6. Increase Resources per Research Study
Despite some efforts within and outside of NIH to address it, the number of animals in any particular HIV vaccine study often remains too limited to demonstrate statistically valid results from any immunization approach. Animal models are an expensive resource, and researchers, funding review committees, and program staff continue to be overly conservative in allocating funding to grants and contracts, particularly for primates.

NCRR and NIAID program staff need to designate additional funding to grants and contracts to support studies with sufficient statistical power to offer reliable answers to research questions. Researchers need to study larger numbers of primates (i.e., more than 3-4 per arm) and request funding for this in their applications. Study sections should look favorably on larger primate studies.

7. Rationalize the Simian / SIV Model
Although ultimate demonstration of comparative immunogenicity and efficacy will have to be done in human volunteers, comparison of vaccine approaches in a standardized primate model would be useful to help further optimize vaccine concepts. But standardization is a challenge. Reliable, standardized assays to measure antibody, CTL, and other immune parameters are needed. Standardized challenge stocks need to be chosen that are "hot" enough to reliably infect and cause disease in a short period of time, but that still mimic the infectivity of HIV in humans. Finally, comparable vaccine antigens and immunization routes need to be selected.

In 1998, NIH funded extensive work to breed specific pathogen-free macaques, to MHC-type macaques, and to develop new SIV and SHIV challenge stocks. NIH also developed a large primate trial to compare different vaccine approaches against each other. Now the primates are out of quarantine and ready, the trial design is complete, and the standardized challenge stock and immune assays have been chosen. The trial will compare gag-protease and env antigens in a wildtype vaccinia vector, an MVA vector, a fowlpox vector, and a DNA construct.

Large, comparative studies of vaccines in macaques are an important tool in assessing and optimizing candidate immunogens for further testing in humans. NIH should streamline the mechanisms for initiating these trials and provide sufficient resources for their expeditious conduct so that the information gathered can be quickly integrated into decision-making regarding human studies.

8. Develop a Clearer Understanding of Protection by Live, Attenuated SIV
In the case of protection of macaques against a pathogenic SIV by immunization with a live, attenuated virus, more research is needed to look at a wide range of systemic and mucosal immune responses. Is there truly sterilizing immunity or transient infection, or does the challenge virus re-emerge? How does control of viral infection happen, and is it correlated with antibody titers, antibody binding and neutralizing capability, timing of early CTLs, CTL epitope specificity and avidity, presence of memory CTLs, other viral suppressive activity, chemokine and cytokine production patterns, some combination of these, or other as yet unknown factors?^125,126 Live, attenuated SIV offers an important model for understanding the correlates of protection conferred by vaccination.

* * *

III. HIV Vaccine Clinical Trials

HIV vaccine clinical trials aim to evaluate the safety, immunogenicity, and efficacy of a range of candidate immunization strategies in human volunteers with the ultimate goal of finding a safe and effective vaccine that can be used around the world. Since 1987, NIH has funded clinical evaluation of HIV vaccines, at about $30 million per year since 1995. Much has been accomplished during the past twelve years of clinical testing of vaccine concepts, but concerns continue to be voiced among researchers and NIH program staff about the limits of the clinical trials effort so far. These concerns include the limited number of vaccine strategies tested to date, the limits to insights gained from past trials, and the amount of work yet to be done to be truly ready for large scale international phase III trials.

Scientific Issues

Safety, Efficacy, Applicability
Ideally, a vaccine to prevent HIV infection and disease should be safe enough to use in millions of people, and effective enough to protect against several routes of infection over a long period of time with as few doses as possible. For global use, an ideal vaccine would be easily administered, inexpensive, and easy to transport and store. To date, most NIH HIV vaccine clinical trials have focused on strategies to prevent infection and disease in adults; some work, however, has evaluated immunization strategies for inducing maternal and infant immune responses to prevent perinatal transmission and HIV disease in children.

Defining and Measuring "Protection"
Efficacy and activity of candidate HIV vaccines can be measured by several clinical and virologic endpoints. If an immunization strategy could completely prevent HIV infection or rapidly clear or control local infection, well-designed, controlled clinical trials carried out in large at-risk populations might be able to demonstrate this in a few years. If an immunization strategy failed to prevent dissemination and establishment of viral infection, but controlled long-term disease progression and transmission, clinical studies would need to follow viral load levels, immune responses, and clinical outcomes in vaccinated-but-infected trial volunteers for many years. In either case, extensive long-term follow-up would be required to define the duration of protection and the long-term safety of the immunization strategy.

Evaluation of potential efficacy of HIV vaccines in clinical trials is complicated by several scientific factors. One is that, in any given clinical trial, an immunization strategy might have a varied protective effect depending on the route of exposure and dose of infecting virus, thus muddying the statistical evidence of short-term protection. The second is that correlates of protection, such as the characteristics of antibody or cellular responses, won't be known definitively until some protection in humans is demonstrated. The third is that long-term evaluation of viral load levels, immune responses, and clinical outcomes in vaccinated-but-infected trial volunteers will be more difficult to interpret in situations where volunteers have broad access to effective antiviral treatments.

Data from vaccine clinical trials provide important information for basic research and product development. Information gained from clinical trials about vaccine administration, antigen processing and expression, and resulting immune responses will help further refine and optimize immunization strategies.

The First Thirteen Years of Trials
Clinical trials of candidate HIV vaccines were begun by NIH through the AIDS Vaccine Evaluation Group (AVEG) in late 1987. The first trials focused on the safety and immunogenicity of recombinant HIV envelope proteins and recombinant vaccinia vector expressing gp160.^127,128 During the ensuing thirteen years, the AVEG initiated more than 30 Phase 1 trials evaluating more than three dozen vaccines and vaccine combinations.¹²⁹ The most extensive clinical studies evaluated recombinant gp120 and gp160 envelope subunit proteins, recombinant peptides with B and T cell epitopes, and vaccinia and canarypox vectors. Trials were also done of recombinant p24 and p55 proteins, lipopeptide conjugate, and salmonella and DNA vectors expressing envelope and core antigens. These were tested with a range of adjuvants, delivery routes, and dosing schedules. Immune responses were measured with standard assays for antibody binding, neutralization, fusion inhibition, CD4 binding inhibition, antibody-dependent cellular cytotoxicity (ADCC), and CTL and lymphocyte proliferation.

During 1998, more than ten NIH-sponsored clinical protocols were launched or continued in the United States, concurrent immunology work was done using clinical trial samples and data, and preparation was done for phase 1 and 2 clinical trials in Uganda and elsewhere.

The largest of these NIH sponsored trials, a Phase 2 trial begun in 1997, enrolled 420 participants in order to evaluate the safety and immunogenicity of a combination of recombinant canarypox vaccine (containing genetic copies of HIV-1 env, gag, and protease) with boosts of recombinant gp120. Through the AVEG, NIH also funded follow-up analyses of a 1992-94 phase 2 trial of gp120 (AVEG protocol 201),¹³⁰ and follow-up of all vaccine trial volunteers with break-through infections.¹³¹

Five phase 1 trials were initiated or continued to evaluate canarypox vectors: a low-dose vs. high-dose vCP205 with rgp120 SF2 (AVEG 022 and 022A), vCP205 administered mucosally (AVEG 027), vCP205 with GM-CSF adjuvant (AVEG 033), vCP 300 with rgp120 (AVEG 026), and a comparative trial of three canarypox products vCP205 vs. vCP1433 vs. vCP1452 (AVEG 034). The intensity of the NIH effort on canarypox virus vectors is due partly to the availability of the products from a company, Pasteur-Merieux Connaught (PMC), but also due to the fact that canarypox vectors so far induce the best CTL responses seen so far in humans. These CTL responses are far from optimal, with only 15 to 30% of volunteers showing CTL responses at any one time during studies. A lot more needs to be known about the duration and kinetics of these CTL responses and their relation to protective efficacy. More research is being done to evaluate immune responses generated by canarypox vectors with multiple genetic inserts, administered mucosally, and in combination with other vaccine products.

Beyond canarypox vectors, the AVEG network conducted trials to identify new immunization strategies that are safe and generate strong, durable CTL responses. These included a phase 1 safety and immunogenicity trial of a Therion vaccinia vector with env/gag/pol inserts (AVEG 014C), a phase 1 trial of an Apollon gag/pol DNA vaccine (AVEG 031), a phase 1 trial of an orally administered attenuated salmonella vaccine (expressing env) in combination with gp120 (AVEG 028), and a phase 1 trial of the adjuvant QS21 with two different gp120 vaccines (AVEG 036). Perinatal work through the Pediatric AIDS Clinical Trials Group (PACTG), the Women and Infants Transmission Study (WITS), and General Clinical Research Center (GCRC) sites coded as HIV vaccine research evaluated the safety and immunogenicity of gp120 vaccines in mothers and infants.¹³² This work demonstrated safety of gp120 in infants less than six months old, and stronger antibody responses than adults to low doses of gp120.

Researchers also continued to develop and refine new assays for use in clinical trials to define humoral, cellular, and other immune responses to HIV, and to define in vivo processing and expression of viral vector and DNA vaccine antigens. Clinical researchers are now proposing phase 1 and 2 studies to define the comparative safety and immunogenicity of newer envelope designs, such as proteins derived from primary isolates, newer viral vector or DNA constructs with multiple epitopes (env, gag, pol, nef), and various prime-boost schedules and doses of vector-subunit combinations.

Envelope is Safe
In the first decade of clinical trials, evaluation of recombinant envelope subunit proteins had found that:

• Immunization with HIV envelope proteins is safe in humans.¹³³
• Recombinant envelope subunits can induce antibodies that recognize native envelope structure.¹³⁴
• Recombinant subunit vaccines can induce functional antibody, with neutralizing activity and fusion inhibition.¹³⁵
• Antibodies induced by the tested envelope subunit vaccines do not neutralize primary isolates in vitro, and have differences in specificity to antibodies found in natural infection.¹³⁶
• Immune responses in gp120 vaccinees with subsequent breakthrough HIV infection do not appear to affect early viral load.¹³⁷

Some Vectors are Safe and Can Induce Cellular Responses
Clinical trials of viral vector vaccines demonstrated:

• Recombinant vaccinia and canarypox vector vaccines are safe, can induce CTL responses, and can augment antibody responses to subunit vaccines.^{138,139,140,141}
• Antibodies induced by candidate vector vaccines do not neutralize primary isolates in vitro, and levels of neutralizing antibody responses induced by current vaccines are 5 to 10 times lower against T cell line adapted viruses than levels found in HIV-infected persons.¹⁴²

Trial Sites are Being Prepared
Finally, establishment of HIV vaccine trial sites and development of volunteer cohorts by the NIH HIV Vaccine Efficacy Trials Network (HIVNET) had adequately shown that suitable high-risk and low-risk cohorts can be recruited for HIV vaccine trials in the United States and in Africa, Asia, the Caribbean, and South America. Clinical infrastructure, laboratory infrastructure and staff, and trained investigators were funded through the international HIVNET in many countries for Phase 3 efficacy trials. HIVNET sites identified, recruited, and retained volunteers at high risk for HIV infection, and provided high quality risk-reduction interventions. Finally, long-term political relationships, critical to the success of HIV vaccine research, have been established by NIH with governments, researchers and communities in several countries.

Clinical Research Recommendations

9. Fund Development of Novel Vaccine Products & Immunization Strategies
The major limitation of the clinical trials effort is the lack of products and immunization concepts to be tested. Major advances were made in 1997 and 1998 in understanding HIV envelope structure and function, and yet the clinical trials networks are still reliant on recombinant gp120 envelope subunit products that pre-date these findings. NIH has funded pre-clinical work for years on novel viral and bacterial vectors, such as Venezuelan equine encephalitis virus (VEE), modified vaccinia virus Ankara (MVA), bacillus Calmette-Guérin (BCG), herpesvirus vector, attenuated salmonella, adeno-associated virus (AAV), and yellow fever virus, but the clinical trial networks have focused trials on canarypox vectors as the only vaccine vector with a fully developed and company-supported development plan. Vaccine companies have recently increased investment in DNA and vector vaccines.

NIH needs to directly fund development of products to feed into clinical evaluation. NIH should expand direct contracting with companies to develop HIV vaccine candidates, such as newer versions of MVA, VEE, AAV, and attenuated salmonella vectors. The NIH should encourage optimization of current vaccine candidates and be willing to discourage work in outdated vaccine concepts that have performed poorly in preclinical, animal or early human studies.

10. Maintain a Solid Clinical Research Infrastructure
During the coming decade, NIH clinical trial programs need to continue rapid phase 1 and 2 assessment of a broader range of candidate immunization strategies for adult and perinatal use, push the basic science to increase comparative understanding of elicited immune responses, and prepare for phase 3 evaluation of candidate vaccines in the US and internationally. This will require an ongoing commitment to a productive, efficient clinical trials infrastructure.

The total 1998 funding from NIH designated for HIV vaccine clinical research was approximately $42 million, about 28% of the total vaccine program. Of this, about $11 million went to the AVEG network, about $11 million to the US HIVNET, $7.6 million to the international HIVNET and international clinical training programs, and about $2.2 million to perinatal and pediatric clinical immunization studies. An additional $5.2 million was steered to establishing the new NIH Vaccine Research Center. (Table 3) More than 90% of 1998 clinical trial funding came from NIAID, with most of the remainder from NCRR to support for adult and pediatric trial activities through General Clinical Research Centers (GCRCs).

NIH should ensure that the new HIV Vaccine Trials Network and the vaccine clinical trial activities of the Vaccine Research Center, the NIH intramural trial site, and the perinatal sites will have the capacity for an adequate number of phase 1, 2 and 3 trials during the coming five years. That being said, the clinical HIV vaccine research programs should be held to the highest scientific standards with adequate oversight through regular review by extramural researchers. The advancement of candidate immunogens along the vaccine clinical development pipeline must be driven by science and not political expediency or industry requests.

11. Strengthen Links between Clinical Trials, Immunology & Basic Science
Clinical research goals continue to be heavily influenced by advances in basic science. Elucidation of the HIV envelope 3-D structure in 1998 helped to reformulate the thinking on what vaccine antigens might induce primary isolate neutralization.^143,144 As canarypox vectors were able to elicit cross-clade cellular responses, optimism about the feasibility of developing vaccines to induce CTL responses has increased, although -- as we noted above -- those particular CTL responses were far from optimal.

The first era of HIV vaccine clinical trials was characterized by a limited understanding of HIV pathogenesis and human immunology, less sophisticated assays, and limited numbers of volunteers in each phase I vaccine study. Thus, the two dozen phase I clinical trials described above yielded some conclusive (negative) results, but provided few insights for new vaccine development efforts beyond learning what doesn't work. Researchers are challenged to design phase I trials that can generate new understanding about how to elicit potentially protective immune responses.

NIH funding supports important immunology research on the CTL, antibody, and cytokine immune responses of phase I trial volunteers, including analysis of cross-clade activity of antibodies and CTLs from combination vaccines,¹⁴⁵ further development of assays to measure CTL responses,¹⁴⁶ and further analysis of antibody neutralization capabilities.¹⁴⁷ More studies in the clinical setting are needed to learn about mucosal and other routes of vaccine administration, in vivo antigen processing and expression, and mechanisms and quantitative measures of humoral, cellular, and mucosal immune responses.

A coherent plan with interim goals need to be articulated for the clinical trials program, so that clinical trials evaluate immunization strategies in a concerted, systematic manner. Expansion of human immunology research, linked to HIV vaccine clinical trials, is needed. Larger and more intensive phase I and II trials should compare vaccine products and immunization strategies and the resulting immune responses.

12. Support a Strong International Vaccine Development Effort
An important NIH effort during past years has been the support of the first-ever phase I HIV vaccine trial in Africa, and support of clinical trial site preparation in Asia, Africa, Latin America, and the Caribbean.¹⁴⁸ During 1998, NIH funding built on-site laboratory capacity in Uganda to evaluate CTL responses in vaccine volunteers. Through its international HIVNET network, NIH has supported preparation for a phase II trial in Brazil, Haiti, and Trinidad and Tobago, and supported characterization of cohorts, CTL responses, and viral strains in many countries. This international clinical trials effort is significant for several reasons. The safety, efficacy, and practicality of HIV immunization strategies will need to be studied in populations with different disease burden, nutritional status, access to health care, level of exposure to HIV, and other environmental factors. More needs to be learned about the effectiveness of vaccine-induced immune responses against various routes of HIV exposure and to different strains of HIV. Finally, vaccine research efforts by the countries most affected by the global AIDS epidemic must overcome barriers of infrastructure for intensive immunologic research and clinical care, and must also secure long-term political and economic support.

International clinical trial site readiness takes years of investment, training, and relationship-building. The continuity of NIH's investment and support of other countries' clinical trial infrastructure has been an important part of the successful launch of a phase I trial in Uganda, and of the preparedness of trial sites in Thailand, Trinidad, Haiti, Brazil, India, South Africa, and elsewhere. NIH should be careful that the recompetition of clinical trial sites does not undermine the years of investment already committed to those countries.

NIH should work more closely with the Centers for Disease Control (CDC), the United States Agency for International Development (USAID) and the Department of Defense (DOD) to support the development and maintenance of high-quality biomedical research infrastructure in developing countries.

* * *

IV. The NIH AIDS Vaccine Research Program

Anatomy of Fiscal Year 1998 NIH Funding for HIV Vaccine Research and Development
In 1998, NIH invested about $148 million in HIV vaccine research and development. This HIV vaccine funding was dispersed to intramural research activities at NIH (19%) and externally (81%) through 374 awards to 135 research institutions.