Following the establishment of HIV infection, a relatively long, asymptomatic period characteristically ensues during which there is a variable but progressive decline in both overall immune competence and the ability to contain the virus immunologically. In 1987 it was hypothesized that the development of AIDS in chronically infected patients might be delayed or prevented by active immunization with an appropriate immunogen (1). Following evidence of safety and immunogenicity in rabbits, guinea pigs, and chimpanzees (2), a number of clinical studies were undertaken beginning in 1987 for the purpose of testing this hypothesis. The immunogen employed for these studies was a gp 120-depleted, inactivated preparation of HIV-1 emulsified in mineral oil (incomplete Freund's adjuvant, IFA), referred to as HIV-1 Immunogen in IFA. Since then, others have employed alternative HIV antigen preparations with the same objective (3).
We report here on USC Study 1A/1B (4,5), initially a 2-year investigation of safety, immunogenicity, and clinical progression which has now accrued clinical and laboratory observations in 25 patients over a period of up to 6 years.
Twenty-five patients were enrolled in studies 1A/1B from November 1987 to September 1988. Two of these patients had asymptomatic HIV infection (CDC II), while 23 had biopsyproven persistant generalized lymphadenopathy (PGL). All subjects were HIV-seropositive homosexual men, with a median age of 38 years (range 25-59). Aside from PGL in 23, additional HIV-related symptoms were present in eight (32%) at study entry, consisting of oral candidiasis in two, oral hairy leukoplakia in four, and both conditions in two. The median time between diagnosis of PGL or other AIDS-related conditions (ARC) and study entry was 3.2 years. In addition, review of all lymph node biopsies was performed as part of routine data review for this study. The retrospective review of lymph node biopsy on one patient (SC015), obtained 4 years prior to study entry, revealed foci of Kaposi's sarcoma (KS) in tissue which had previously been diagnosed as PGL alone. This patient did not develop an opportunistic infection (OI) during the follow-up period. He was included in calculations of clinical progression when new KS lesions appeared and the disease became clinically apparent.
At study entry, no patient had an acute intercurrent infection other than genital herpes, and none had taken corticosteroids or investigational agents during the preceding 30 days. No patient had taken zidovudine (AZT) or other known, effective antiretroviral therapies at any time prior to study entry. A consent form approved by the Human Research Committee of the University of Southern California School of Medicine was signed by all patients. The study was conducted under the auspices of the Food and Drug Branch of the State of California.
At its inception in November of 1987, the objectives of this open-label, single center study were to assess the safety of HIV-1 Immunogen, to determine if an immunologic response to HIV could be induced, and to determine any effect upon HIV viral burden. The original design provided for doses at 0, 3, and 6 months, with follow-up to 2 years. Patients were evaluated at baseline, weekly for 4 weeks, biweekly for the next month, then monthly until month 12. Thereafter, patients were evaluated every 3-6 months. Subsequent amendments provided for repeated immunizations at 3-6-month intervals to the end of year 3 of the study, with continued clinical observation and laboratory assessment of immune status and viral burden as described below.
This report details laboratory observations recorded from the time of study entry until study termination as defined by: (a) voluntary withdrawal or discontinuation; (b) development of an AIDS-defining condition; or (c) enrollment into USC study 1G (occurring between January 27, 1992, and May 6, 1992), a continuation study in which inoculations of HIV-1 Immunogen have been resumed.
Clinical observations are reported for all patients through February 1, 1994, regardless of their study termination date. All clinical observations are included in calculations of clinical progression rates.
Preparation and Use of the Immunogen
The immunogen studied is obtained by concentration and purification from the supernatant fluid of HZ321-infected Hut-78 cells (6). This HIV viral strain was initially isolated at the Centers for Disease Control (CDC) from the serum of a 26-year-old woman from northern Zaire, who died of an illness strongly suggestive of AIDS in 1978. Confirmation of HZ321 as an HIV-1 isolate was demonstrated by immunologic assay, virus dot-blot, Southern hybridization, and molecular characterization (6,7). In preparing the immunogen for use, envelope gp120 is depleted during freezing and thawing and the sucrose gradient purification process. Viral inactivation is accomplished through sequential application of beta-propiolactone (7) and 60Co irradiation (8). Immunogen was prepared for administration in two ways: for intramuscular inoculations, as 100 μg total protein emulsified with IFA in a final volume of 1.0 ml and, for intradermal inoculations, as 100 μg total protein in 0.5 ml of saline.
The first group of nine patients who enrolled in November 1987 received their initial 100 μg inoculation intramuscularly in IFA, followed by a 100 μg intradermal dose in saline and 100 μg intramuscular doses in IFA thereafter. A second group of 10 patients entered the study in January 1988 and received their first 100 μg dose intradermally in saline followed by 100 μg intramuscular doses in IFA thereafter. Six additional patients enrolled in September 1988 and received all 100 μg inoculations by the intramuscular route in IFA. Since no differences in response or toxicity were observed among the various routes of administration, all subsequent doses of HIV-1 Immunogen were administered by intramuscular injection for the remainder of the study in all patients. Among the total group of 25 patients, a median of eight doses (range two to 10) were administered over the first 4 years of this study. The majority of patients received three or four inoculations per year during years 1 and 2, and zero or one inoculation during year 3. Only five patients received a dose after year 3 as part of this protocol. Intradermal delayed-type hypersensitivity (HIV-DTH) skin testing was initiated in January 1988. The small doses of immunogen administered during the skin testing were given in addition to the 100 μg immunizations as outlined above.
HIV-DTH Skin Testing
Three immunogen concentrations were prepared for intradermal HIV-DTH skin testing: 0.1 μg, 1 μg, and 10 μg total protein per 0.1 ml of saline. These small doses of intradermal immunogen were not considered as immunization but rather as skin tests. All three concentrations were administered simultaneously, with saline alone used as a control. The mean induration of HIV-DTH response (average of the largest diameter and its perpendicular diameter) was measured at approximately 48 h.
Additional Skin Testing
The Merieux CMI Multitest (Institut Merieux, Lyon, France) was utilized to assess cellular immune competence at baseline and periodically throughout the study (9). This simultaneously applied panel contains the following seven antigens: tetanus, diphtheria, streptococcus, tuberculin, candida, trichophyton, and proteus. A vehicle (glycerin) control is included. As noted within the published guidelines for the CMI Multitest, induration measuring 2 mm or greater in average diameter 48 h after application was considered indicative of response to a particular antigen. All DTH skin testing procedures, including application, measurement, and recording were performed consistently throughout the study by two designated study personnel (K.M., J.A.).
Assessment of Safety
At each physician evaluation, patients were asked to report all signs and symptoms noted since the last visit. They were questioned regarding pain, redness, or swelling at the injection site following the most recent inoculation and were specifically questioned regarding common signs and symptoms of drug hypersensitivity reactions such as fever and rash. A complete blood count, routine chemistries, urinalysis, and physical examination were performed at each visit.
HIV-1 Western blots were performed at irregular intervals as a qualitative measure of overall humoral responses to HIV. Levels of anti-p24 antibody were determined by endpoint titration in an enzyme immunoassay (EIA). Initially, antibody titers were determined based upon serial 10-fold dilutions of fresh patient sera, which revealed no significant changes. In 1990, an EIA was developed employing serial twofold dilutions of sera, with an assay variance of ±1 dilution (2). This EIA was utilized to retest cryopreserved sera from prior to 1990 and fresh serum thereafter. Briefly, 96-well microtiter plates were coated overnight with 500 ng per well of recombinant HIV p24 (Pharmacia Genetic Engineering, La Jolla, CA, U.S.A.), then blocked with 5% bovine serum albumin. An HIV-1-positive control serum (No.X-1015) was obtained from the New York Blood Center and utilized throughout as a positive control. Patient sera were coarsely screened at baseline using serial 10-fold dilutions beginning at 1 × 10-2. Serial twofold dilutions were then tested beginning at the last positive 10-fold dilution. Plates containing serum dilutions were incubated at 68°-74°F for 1 h. Three washings were performed using phosphate-buffered saline/0.5% Tween-80. Horseradish peroxidase-conjugated goat-anti-human IgG (Kirkegaard and Perry Laboratories, Gaithersburg, MD, U.S.A.) was utilized for detection of bound antibody. Following a 30-min incubation with substrate (2,2'-azino-di[3-ethylbenzthiazolinesulfonate]), the colorimetric reaction was stopped by the addition of 1% sodium dodecyl sulfate and absorbance determined at 405 nm using a Nippon 2000 Immunoreader (Intermed). The endpoint is taken as the last dilution to register an optical density (OD) of 0.050-0.200. A consistent change of fourfold or greater from baseline was shown to be greater than assay variability.
p24 Antigen Determinations
HIV p24 antigen was measured using both a standard assay (Coulter, Hialeah, FL, U.S.A.) and, retrospectively on cryopreserved specimens, the standard assay modified by acid (glycine-HCl) pretreatment to dissociate immune complexes (Coulter ICA kit) (10,11). All assays were run in triplicate; 10 pg was the threshold for positivity, as determined by a standard curve. Additional internal standards of 20, 100, and 200 μg of P24 antigen were also included in each assay.
Lymphocyte Surface Markers
Flow cytometric phenotyping of peripheral blood mononuclear cells (PBMC) was performed by the USC Flow Cytometry Laboratory, using whole-blood lysis cell preparations and direct two-color staining (RD1/FITC) with fluorochrome-conjugated monoclonal antibodies (Coulter): CD14(MO2)/CD45 (KC56), IgG1/IgG2a, CD4(T4)/CD3(T3), CD8(T8)/CD3(T3) (12).
Initially and during the first 18 months of study, the ability to isolate HIV from PBMCs was evaluated by cocultivating Ficoll-Hypaque-purified PBMC from 20 ml of patient blood with phytohemagglutinin antigen (PHA)-stimulated PBMC from a seronegative donor. In April 1989, an improved coculture assay was instituted (13). Briefly, 5 × 106 PHA-stimulated PBMCs from an uninfected donor were cocultivated with 1 × 107 patient PBMCs in RPMI-1640 medium (Cellgro, Herndon, VA, U.S.A.), supplemented with 20% fetal bovine serum and 5% v/v interleukin-2 (Cellular Products, Buffalo, NY, U.S.A.). Cocultures were incubated in 75-ml flasks at 37°C in a 5% CO2, 95% air atmosphere with sampling of supernatant fluid and refeeding every 7 days, with the addition of 3 × 106 cryopreserved PBMCs from the same seronegative donor as the original cocultivation, to a maximum of 35 days. Positivity was assessed by measurement of HIV p24 antigen in supernatant fluid using a standard assay (Coulter).
In 1991 a polymerase chain reaction (PCR) assay was developed for detection and quantitation of HIV-DNA. Quantitation was based on coamplification with HIV-1 (30 cycles) and globin (15 cycles) primers (14). Both amplicons were 32P-end-labeled, and quantitation was obtained by comparison to HIV-1 (10-1,000 copies) and globin (2 × 105 to 3.2 × 106 cells) standard curves. After gel electrophoresis, cpm from both amplicons were quantitated using the AMBIS Radioanalytic Imaging System (San Diego, CA, U.S.A.). Globin controls were used to normalize the HIV-1 copy numbers per 4 × 105 PBMCs. The precision of the assay (reproducibility of replicates) has been previously reported (14). The 95% tolerance intervals were calculated and the limits of these intervals (percent change) established such that a 60% increase or a 38% decrease in the amount of HIV-1 DNA copies represent significant changes, greater than assay variability. These limits were employed in the analysis of copy number over time, using cryopreserved PBMCs from study subjects, with all specimens on each subject run simultaneously. DNA copy numbers were estimated per 4 × 105 CD4 cells. The DNA-PCR copy number was considered to have increased if PCR values were 60% or more above baseline on two consecutive measurements. The DNA-PCR copy number was considered to have decreased if PCR values were 38% or more below baseline on two consecutive determinations. Data from 19 patients are available for analysis. The median number of PCR determinations was four (range two to six).
Methods of Data Analysis
To analyze measures of HIV infection and various indicators of disease progression listed in Tables 1 and 2, baseline values and changes from baseline levels were summarized. For purposes of analysis, the logarithm of the CD4+ count and the logarithm of the number of DNA copies (PCR) were used. The average percent change per year in CD4+ count (last column in Table 2) was based on the slope of the simple linear regression for logarithm of CD4+ count taken as a function of time from first inoculation. For purposes of summarizing, p24 antibody titers were classified as (a) four times the baseline value on at least two occasions (increase), (b) half the baseline value on at least two occasions (decrease), or (c) in between (stable). Trends for changes in the number of PCR copies (adjusted to the CD4+ count) were defined as follows: if the number of PCR copies was 60% or greater than the baseline value on at least two consecutive occasions, the trend was labeled as “increase”; if the number decreased by 38% or more from baseline, the trend was labeled as “decrease”; otherwise, the trend was labeled “stable.” For the HIV skin test, a DTH response was considered positive if the induration was 9 mm or greater in the average of the largest diameter and its perpendicular. This value was chosen since a very clear separation between the groups was apparent at this level of response (Fig. 1). All laboratory values obtained during the course of treatment and follow-up on the 1A/1B study are included in this summary of results. For patients who began AZT on-study (Table 3), all values are included in the summaries and no adjustments have been made.
The Kaplan-Meier product-limit method (15) was used to estimate the probability of disease progression as a function of time from first inoculation, and the progression rates were estimated as the number of events (i.e., progressions) divided by the total time at risk (i.e., on treatment and follow-up) for progression.
To compare patients with an HIV-DTH response of ≥9 mm with those with a lesser or no response and to compare patients who experienced progression to AIDS to those who did not, Fisher's exact test, the two-sample t test, or the Wilcoxon test were used (16,17). All reported p values are two-sided; all intervals constructed (exact or based on large sample results) are 95% confidence intervals.
Twenty-five patients were enrolled in studies 1A/1B from November 1987 to September 1988. The median follow-up interval for general laboratory analysis from the time of study entry was 3.8 years (range 0.4-4.5 years) and that for clinical follow-up was 5.8 years (range 2.8-6.2 years). Reasons for study withdrawal were as follows: voluntary withdrawal (SC006 and SC007), disease progression (SC009 and SC015), noncompliance (SC083 and SC082), and development of either KS (SC031, SC001, SC014) or OI (SC024, SC005). Patients who were withdrawn from the study for any reason were followed for clinical progression and survival, as detailed below. The remaining patients requested further treatment and were enrolled in the USC-1G follow-up study between January 27, 1992, and May 6, 1992.
Symptoms attributed to inoculations of HIV-1 Immunogen in IFA were generally mild and transient and were reported by 20 of 25 patients (80%). Symptoms included local pain (44%), generalized muscle aches (44%), local soreness (24%), “feeling warm” (20%), “flu-like symptoms” (20%), fatigue (8%), and headache (8%). Signs of adverse reactions included erythema (32%) and/or induration (16%) at the injection site and faint generalized maculopapular rash (20%). Adverse reactions typically resolved within 24-48 h without treatment, though one subject experienced pain and bruise at the injection site, which lasted for 14 days, and a second developed a mass at the injection site, which lasted 18 days. There were no adverse laboratory or serious adverse clinical experiences attributed to the test material.
Cellular Immune Responses
The immunogen skin test (HIV-DTH) first became available in January 1988 and was applied at baseline to 15 patients. Twelve demonstrated no response, while three showed some degree of reactivity, with induration measuring 3, 3, and 6 mm, respectively, at 48 h. Eleven of 15 (including patients 015 and 018) were responsive to various antigens in the Merieux panel applied the same day, indicating that the lack of HIV-DTH responsiveness was not due to global anergy. Patients 014, 017, 023, and 070 were anergic to the Merieux panel at baseline. Two of these (017, 023) were subsequent responders to the HIV/DTH skin test, while the other two were nonresponders over time.
Over the course of the observation period, 12 patients displayed a maximum of 10-30 mm of induration (mean 14 mm) at one or more time points, as demonstrated in Fig. 1. These responses were more likely to occur at the 10 μg concentration of HIV skin test antigen (eight of 12 responders), while four of 12 responded to the 1.0 μg dose of DTH-HIV antigen. The first positive test was noted after an average of four inoculations (range one to seven). In a random sampling of six patients, induration was confirmed to represent DTH by histochemical and immunofluorescent staining of punch biopsy specimens, which revealed characteristic vasocentric lymphocytic infiltrates of predominantly CD4-positive T-cell phenotype (18; Fig. 2).
The HIV-DTH-responsive group generally had less advanced HIV disease at baseline compared with the HIV-DTH-nonresponsive group, as judged by mean CD4 counts, p24 antigenemia, HIV-DNA copy number by PCR, and anti-p24 antibody levels (Table 1). The HIV-DTH-responsive group was also skin tested more often than the HIV-DTH-nonresponsive group, in part due to earlier termination in the latter group. More inoculations with HIV-1 Immunogen were administered to the HIV-DTH responders (median nine, range five to 10) than to the HIV-DTH nonresponders (median six, range two to 10), though this difference was not statistically significant (p = 0.12, Wilcoxon).
T-lymphocyte function was further assessed in vivo using a panel of standard antigens, the Merieux CMI Multitest. This panel was applied at baseline to all 25 patients, with 24 being evaluable, since one patient did not return for 48-h follow-up. Among the 16 patients showing ≥2 mm of induration to one or more antigens at baseline, five lost all responses by the last measurement, six showed increased numbers of responses and size of induration, and responses in five were intermittent and generally unchanged. Eight patients were found to have ≤1 mm of induration to each of the seven test antigens at baseline. Over the period of follow-up, two of these eight showed consistently increased numbers of responses and size of induration, and six patients became intermittently responsive to various antigens. Of interest, only two of the eight patients who were anergic to the Merieux panel at baseline were HIV-DTH nonresponders over time (SC007 and 014), while the other six (SC001, 003, 008, 017, 023, 070) all responded to DTH-HIV testing after immunization.
Among the 12 patients with an HIV-DTH response ≥9 mm, five had increased DTH responses to the Merieux panel over time, five had stable (unchanged) responses, one experienced a loss of response, and data were not available on one. Among the remaining 13 patients without an HIV-DTH response, three experienced improved DTH responses to the Merieux panel, six remained stable, and four lost responsiveness.
Humoral Immune Responses
Humoral antibody data are available for 23 patients. Assay of cryopreserved sera was performed using serial twofold dilutions. Over the course of the study, fourfold increases in antibody titers were detected in nine patients (39%), while 13 patients maintained titers at least equal to or greater than baseline and one patient had a decrease in titers. At the time of last p24 antibody analysis, fourfold increases were maintained in six of these nine patients (Table 1).
Changes in the serologic profile of patients were assessed by qualitative Western blot at a 1:100 serum dilution. While the majority of patients exhibited similar profiles pre- and postimmunization, qualitative increases in reactivities were seen in eight patients, restricted to gag gene products. Five patients, p24 antigenemic at baseline, showed a decrease in reactivity to p24 and p17.
CD4 Lymphocyte Counts
As demonstrated in Table 2, CD4 cell counts were stable during the first 2 years of study. Thus, the geometric mean CD4 cell count prior to treatment was 375/dl (95% CI: 305-461), at 1 year 381/dl (95% CI: 306-474), and at 2 years 311/dl (95% CI: 240-402). By the end of the study period, the geometric mean CD4 count had dropped to 212/dl (95% CI: 146-306), representing a significant decrease from baseline (p = 0.001). When the slopes of the trend of CD4 cell counts over time were calculated for each patient, the median slope translated to an overall decrease of 10.7% (95% CI: -29.1% to -9.5%) per year in CD4 count. The differences in the patterns of distribution and change of CD4% levels in high and low HIV-DTH responders are shown in Fig. 3.
Viral culture data are available for 19 patients, using methods which were available at the outset in 1987. Except for sporadic negative cultures which occurred in five patients (all HIV-DTH responders), all patients remained culture-positive during the period tested. Employing the revised coculture technique (13) beginning in April 1989, 22 subjects were serially studied to determine the time to first p24 antigen positivity in culture. As shown in Fig. 4, HIV-DTH responders took longer than nonresponders to become HIV culture-positive. Thus, of 106 HIV viral cultures performed in the 12 HIV/DTH responders, 47 (44%) required greater than 7 days to demonstrate culture positivity, while only three of 68 cultures (4%) in the HIV/DTH-nonresponsive group required longer than 7 days to become positive.
Measurement of p24 antigen by glycine-HCl dissociation was also performed on samples at baseline and during the course of treatment in 19 patients. During the course of study, seven of the 19 had increasing antigenemia, eight remained antigen-negative, one remained stable, and three had decreasing antigenemia (Table 1).
Assessment of the number of virally infected PBMCs by PCR-DNA revealed geometric mean copy numbers of 212 (95% CI: 126-356) at baseline and 420 (95% CI: 202-874) at last follow-up. Interestingly, the 10 patients who developed an initial AIDS-defining illness, in addition to patient SC015, who progressed from initial AIDS, demonstrated a median increase of 323% in HIV-1 DNA copy number over baseline, compared to a 24% median increase in those who did not develop AIDS. The distribution of HIV-1 DNA copy number in relation to HIV-DTH responsivity is shown in Fig. 5.
A total of 18 AIDS-defining events have occurred in 10 of the 25 (40%) patients over the course of 5.8 years of follow-up. An eleventh patient with KS at baseline experienced clinical progression of KS. Details of the clinical courses are presented below and in Table 3.
Clinical progression to AIDS was limited to the development of KS in three patients (SC014, SC031, SC078). One patient (SC001) developed KS, followed by Mycobacterium avium intracellulare (MAI) infection at 4.7 years, cytomegalovirus (CMV) at 5.1 years, and death at 6.1 years. OIs alone occurred in three patients, all of whom died (SC005, SC024, SC007); one patient (SC021) developed both OI and HIV dementia; one (SC006) developed both OI and central nervous system (CNS) lymphoma; and one (SC009) developed KS, dementia, and OI; all of these subsequently died. The patient with KS at study entry (SC015) progressed with KS at 1.3 years but did not develop OI. Minor evidence of clinical progression was noted in eight other patients, consisting of oral hairy leukoplakia in two, oral candidiasis in five, and both conditions in one. Four patients without a clinical AIDS-defining diagnosis experienced a persistent drop in CD4 count to under 200, or <14%.
For the entire group of 25 patients, the probability of remaining alive and free of AIDS-defining clinical conditions or progression was 96% (±4%) at 1 year, 88% (±7%) at 2 years, 84% (±7%) at 3 years, 68% (±9%) at 4 years, and 60% (±10%) at 5 years.
During the course of this study (March 1990), AZT was approved for use in HIV-infected patients with less than 500 CD4 cells/mm3. At that time, patients were given the opportunity to take this antiretroviral compound. Sixteen patients (64%) eventually chose to start AZT at a median of 2 years from study onset (range 1-4 years) and remained on AZT therapy for a median of 1.3 years. All maximum HIV-DTH responses were observed prior to initiation of AZT therapy, except in one patient (SC001) who was DTH-nonresponsive prior to AZT therapy and became responsive thereafter. Of interest, seven of 10 patients who developed AIDS had been on AZT prior to onset of the AIDS-defining illness. Serum p24 antigen levels were noted to diminish following initiation of AZT in five subjects.
Association of Laboratory Measurements and Clinical Outcome
Although the 11 patients who progressed or died presented with slightly more unfavorable laboratory measurements at the start of treatment, compared to the 14 patients who are alive and still free of AIDS, the majority of these differences were not statistically significant: the (geometric) mean baseline CD4 count was 320 for those who progressed or died vs. 425 for those who did not (t test, p = 0.17); the (geometric) mean baseline PCR-DNA copy number was 264 for those who had an unfavorable outcome vs. 162 for those who had a favorable outcome (t test, p = 0.34); median p24 antibody titer was 400 for those who failed and 4,000 for those who did not (Wilcoxon signed rank test, p = 0.03); and in terms of the serum p24 antigen levels, seven of 10 of the progressors were antigenemic while two of nine patients currently doing well were antigenemic at baseline (Fisher's exact test, p = 0.07).
By the end of the study period, those patients who had progressed had lower CD4 counts than nonprogressors (129 vs. 314, t test, p = 0.001), higher PCR-DNA copy numbers (907 vs. 179, t test, p = 0.15), and lower p24 antibody titers (800 vs. 5,200, Wilcoxon signed rank test, p = 0.08) and were more likely to be antigenemic (8/11 vs. 4/10, Fisher's exact test, p = 0.20).
Of interest, nine of the 11 clinical progressors were among the low HIV-DTH responders, while 10 of the 14 patients who are alive and free of AIDS progression were among the higher HIV-DTH responders (Fisher's exact test, p = 0.015; Table 4). In general, patients with higher HIV-DTH responses presented with more favorable laboratory measurements, though these differences were not statistically significant, except for p24 antibody titer. At the end of the study period, the 12 patients with higher HIV-DTH responses demonstrated more favorable laboratory values compared to those who remained HIV-DTH-unresponsive (Figs. 2, 3). Rates of HIV progression per year per 100 patients were 3.1 (95% confidence interval: 0.8-12.4) for the higher DTH responder patients vs. 17.6 (95% confidence interval: 9.2-33.9) for the low/non-HIV-DTH responders. No relationship was seen between DTH reactivity to the Merieux panel and subsequent clinical course in either the HIV-DTH-responsive or -nonresponsive group. Thus, additional DTH-Merieux reactivity was seen in five of 11 progressors, while four lost specific antigen reactivity and two remained unchanged. Of the 14 nonprogressors, two remained unchanged in terms of DTH-Merieux reactivity, while seven gained reactivity to various antigens and five lost reactivity.
The purpose of this communication is to report on the safety and immunogenicity of HIV-1 Immunogen and on the virologic, CD4 cytologic, and clinical patterns observed in 25 patients with PGL, asymptomatic HIV infection, and other ARCs. These subjects received multiple doses of the immunogen and were observed over a period of up to 6 years (median 5.8 years).
No significant adverse side effects of immunization with HIV-1 Immunogen in IFA were observed. The majority of patients developed mild acute symptoms following injection of the immunogen, including pain, soreness, swelling, and/or redness at the injection site. However, there was no evidence suggesting long-term HIV viral activation following immunization, as reflected in either p24 antigen levels or HIV PCR-DNA copy numbers. Further, as indicated by serial laboratory analyses or clinical rates of progression, there was no suggestion that immunization was associated with acceleration of progression toward AIDS.
HIV-1 Immunogen, which was shown to be immunogenic in preclinical studies, also appears to have been associated with the development of specific cell-mediated (HIV-DTH) responses. Thus, HIV-DTH responses were observed in 12 of the 25 patients (48%) following immunization. Additionally, fourfold or greater increases in p24 antibody levels were observed in nine of 23 patients (39%), though without a control group this can only be assumed to be secondary to immunization. Three subjects (SC008, SC016, SC031) developed both HIV-DTH and p24 antibody responses.
An association was observed between clinical progression of HIV infection and the degree of HIV-DTH responsiveness observed postimmunization; HIV-DTH responders generally had less advanced HIV disease at baseline than nonresponders, as judged by CD4 cell count, p24 antibody levels, and prevalence of HIV-related symptoms and signs. Six cases of OI and seven deaths occurred in the group of low HIV-DTH responders, while only one individual in the higher HIV-DTH responder group developed an OI, at approximately 5 years after study entry, and subsequently died. The high vs. low HIV-DTH responders did not differ in the incidence of KS (two of 12 vs. four of 13, respectively). Of note, we observed no correlation between DTH response to the Merieux panels and to HIV-1 Immunogen when used as a skin test antigen. Nonspecific DTH reactions have previously been noted to correlate with HIV disease progression (19); the lack of such correlation in the current study could result from the fact that Merieux Multitest evaluations in HIV-infected patients have not always provided consistent results (20).
In addition, virologic and CD4 cellular observations were more favorable in the HIV-DTH-responsive patients. Negative cultures for HIV were observed on one or more occasions in only five patients, all of whom were in the HIV-DTH-responsive group, while this group also demonstrated a higher proportion of cultures requiring more than 7 days to become positive. A sevenfold mean increase in HIV-1 DNA copy number was observed in the group of patients who progressed to an AIDS-defining illness or died as compared to a 1.7-fold mean increase in those who did not progress. These data suggest that a longitudinal increase in viral burden as assayed by quantitative PCR HIV-1 DNA is correlated with clinical progression, as demonstrated by others (21-25). In regard to CD4 cells over time, greater stability of counts was apparent among the high HIV-DTH responders, with final values approximately 68% of baseline in this group, vs. 47% of baseline in the HIV-DTH low responders.
After a median of eight inoculations given to this small group of patients with asymptomatic HIV infection or PGL, in 32% of whom other evidence of symptomatic HIV disease was present, the probability of progression to AIDS (either KS or OI) was 4% at 1 year, 12% at 2 years, 16% at 3 years, and 40% at 5 years. Since this study was designed and executed as a phase I safety and immunogenicity study, without a control population, no conclusions can be made regarding clinical efficacy, which will require long-term, placebo-controlled clinical studies.
The relationship observed in this study between HIV-DTH responsiveness and virologic, CD4 cellular, and clinical courses suggests that HIV immunotherapy may be expected to be most effective in patients who are capable of developing HIV-DTH responsiveness following immunization. Since the development of HIV-DTH reactivity following immunization cannot be used for stratification purposes in future studies which include unimmunized controls, it would be useful to determine whether there are other immunologic factors which correlate with the capacity to develop such responsiveness. Studies by Clerici and colleagues (26,27) suggest that differences in T-helper (TH) cell subset predominance might provide such a correlate. Thus, TH1 helper cells, which promote primarily cell-mediated immune responses, predominate early in the course of HIV infection, whereas TH2 cells, which promote humoral responses, become predominant as individuals progress towards AIDS, in parallel with a loss of in vitro responsiveness to recall antigens (26,27). If a correlation exists between TH1 predominance and the ability to develop HIV-DTH reactivity following immunization, then TH cell subset status might serve as a basis for prognosis and stratification in future trials. Furthermore, such a correlation would suggest the desirability of exploring measures designed to enhance or maintain a TH1-predominant state (27,28).
In summary, HIV-1 immunogen has been administered to 25 patients, who were followed over 6 years without evidence of significant harmful effects. Twelve of the 25 became HIV-DTH-responsive, and a fourfold or greater increase in p24 antibody level occurred in nine. Six cases of OI and seven deaths occurred in the 13 low HIV-DTH responders, while only one of 12 in the higher responder group developed an OI and subsequently died. Further studies are indicated to determine clinical efficacy of the HIV Immunogen as well as the significance of the apparent correlation between HIV-DTH responsivity and a more favorable clinical course.
Acknowledgment: The authors sadly acknowledge the death of Dr. Jonas Salk on June 23, 1995; Dr. Salk was thoroughly involved in the preparation of this manuscript, and was entirely responsible for the concept of this therapeutic approach. The authors express their sincere gratitude to Michael Ascher, M.D., H. W. Sheppard, Ph.D., Carl Hansen Ph.D., and Donna Traclindo, M.S. of the California Department of Health; Steven Brostoff, Ph.D., Karen G. Burnett, Ph.D., Chris Duffy, Wieslawa Giermakowska; Sharon L. Liu, M.P.H., Ana Marchese, Robert E. Peters, Ph.D., and Jay R. Savary of the Immune Response Corporation; Herbert B. Slade, M.D. of Rhone-Poulenc Rorer; Merril Gersten, M.D. of the Jonas Salk Foundation; and Joan Abrahamson of the Jefferson Institute for their support and assistance in initiating and conducting this clinical trial, the conduct of assays on trial samples, and the collection and analysis of data.