Safety and Immunogenicity of an HIV Adenoviral Vector Boost after DNA Plasmid Vaccine Prime by Route of Administration: A Randomized Clinical Trial

Background In the development of HIV vaccines, improving immunogenicity while maintaining safety is critical. Route of administration can be an important factor. Methodology/Principal Findings This multicenter, open-label, randomized trial, HVTN 069, compared routes of administration on safety and immunogenicity of a DNA vaccine prime given intramuscularly at 0, 1 and 2 months and a recombinant replication-defective adenovirus type 5 (rAd5) vaccine boost given at 6 months by intramuscular (IM), intradermal (ID), or subcutaneous (SC) route. Randomization was computer-generated by a central data management center; participants and staff were not blinded to group assignment. The outcomes were vaccine reactogenicity and humoral and cellular immunogenicity. Ninety healthy, HIV-1 uninfected adults in the US and Peru, aged 18–50 were enrolled and randomized. Due to the results of the Step Study, injections with rAd5 vaccine were halted; thus 61 received the booster dose of rAd5 vaccine (IM: 20; ID:21; SC:20). After the rAd5 boost, significant differences by study arm were found in severity of headache, pain and erythema/induration. Immune responses (binding and neutralizing antibodies, IFN-γ ELISpot HIV-specific responses and CD4+ and CD8+ T-cell responses by ICS) at four weeks after the rAd5 booster were not significantly different by administration route of the rAd5 vaccine boost (Binding antibody responses: IM: 66.7%; ID: 70.0%; SC: 77.8%; neutralizing antibody responses: IM: 11.1%; ID: 0.0%; SC 16.7%; ELISpot responses: IM: 46.7%; ID: 35.3%; SC: 44.4%; CD4+ T-cell responses: IM: 29.4%; ID: 20.0%; SC: 35.3%; CD8+ T-cell responses: IM: 29.4%; ID: 16.7%; SC: 50.0%.) Conclusions/Significance This study was limited by the reduced sample size. The higher frequency of local reactions after ID and SC administration and the lack of sufficient evidence to show that there were any differences in immunogenicity by route of administration do not support changing route of administration for the rAd5 boost. Trial Registration ClinicalTrials.gov NCT00384787


Introduction
While significant challenges exist in the search for a safe and effective HIV vaccine [1], an important part of the discovery process is testing in humans for safety and immunogenicity. In the development of HIV vaccines, improving immunogenicity while maintaining safety is critical. One factor that can influence safety and immunogenicity is the route of administration. A significant increase in immunogenicity through use of a particular route may allow for a greater chance of demonstrated efficacy, as well as fewer or lower doses used, which can affect the cost of vaccine development.
Administration of vaccines into the skin or subcutaneous tissue may be more immunogenic or provide a different pattern of immune responses than administration by the intramuscular route. The skin is one of the largest organs of the body and the most common site for manifestations of immune reactions [2]. The skin plays critical roles in both innate immunity, as a physical barrier to pathogens, and in adaptive immunity [3]. Dermal immunization attempts to induce an immunologically efficacious response by providing antigen to a variety of cells, including keratinocytes and dendritic cells (DC). After maturation, Langerhans cells (dendritic cells found mainly in the epidermis) and dermal DC (found mainly in the dermis) can migrate to draining lymph nodes where presentation of antigens to T cells can initiate a variety of immunological responses [4,5]. In contrast, intramuscular vaccination delivers antigen to a place with fewer professional antigenpresenting cells [6,7]. Thus, it is possible that different routes of administration may produce differences in T-cell memory or effector populations and drive differences in trafficking patterns of lymphocytes responding to HIV vaccines. Furthermore, dermal immunization may provide an advantage over intramuscular immunization if lower doses of the vaccine can be utilized with similar or improved immune responses. Finally, dermal immunization could more effectively overcome any dampening effects of pre-existing immunity to vaccine vectors.
Studies of a variety of vaccines have found that intradermal vaccination can be just as effective as, or more effective than, intramuscular vaccination, using doses several fold lower [7][8][9][10][11][12] but this advantage may be influenced by other factors, such as age of the host. Subcutaneous dosing has been found to be comparable to intramuscular dosing in terms of immunogenicity [10,11]. In many of these studies, the frequency of local reactions to vaccines given by the intradermal or subcutaneous route were higher than when given intramuscularly, but usually mild and transient. There have been no overall differences in systemic reactions or serious adverse events [7][8][9][10][11][13][14][15].
Using vaccines with demonstrated immunogenicity in multiple clinical trials [16][17][18], the objective of this studywas to compare the effect of routes of administration on safety and immunogenicity of a prime-boost regimen of two HIV vaccines: a DNA vaccine prime given intramuscularly via the needle-free BiojectorH and a recombinant replication-defective adenovirus type 5 vaccine boost given one of three routes: intramuscularly, intradermally, or subcutaneously (HIV Vaccine Trials Network (HVTN) Protocol # 069).

Study design and procedures
The protocol for this trial and supporting CONSORT checklist are available as supporting information; see Checklist S1 and Protocol S2. The study was designed as a multicenter, open label, randomized trial. Starting in November 2006, 90 participants were randomized to one of three groups designated by the route of administration of the rAd5 vaccine (30 participants in each group). The randomization was stratified by country (United States and Peru) using a fixed block size of 6 to ensure balance across arms. Assignment to group was via a web-based randomization system managed by the central data management center. Participants and site staff were not blinded as to the group assignment. Randomization was not stratified by Ad5 neutralizing antibody titer. DNA vaccine injections were planned to be given at months 0, 1, and 2 months and rAd5 vaccine at 6 months. Follow-up visits were conducted two weeks after each injection and at 7, 9 and 12 months post-enrollment.
Local and systemic reactogenicity assessments were performed following each vaccination for up to 3 days and were graded as mild (transient or minimal symptoms), moderate (symptoms requiring modification of activity), severe (incapacitating symptoms resulting in bed rest and/or loss of work or social activities) or lifethreatening. Any adverse events were reported for the entire study duration for individual participants and coded for their relationship to study product (not, possibly, probably or definitely related).
In September 2007, the Step Study involving the MRKAd5 HIV-1 gag/pol/nef vaccine was stopped due to futility since the results to date indicated that the vaccine did not prevent HIV-1 infection nor reduce early viral level [19]. Furthermore, there seemed to be an increase risk of HIV infection among male vaccinees who had prior neutralizing antibodies against adenovirus type 5. Based on this observation, in October 2007, leadership of the HVTN 069 protocol, HVTN and Division of AIDS (DAIDS) decided to halt all injections of the rAd5 vaccine in this protocol.

Study participants
After written informed consent, participants were screened for eligibility and willingness to participate at five HVTN sites in the United States and Peru: Birmingham, AL; Boston, MA; New York, NY; Seattle, WA; Iquitos, Peru. Potential participants were drawn from a large pool of participants screening at the study sites for multiple HVTN vaccine clinical trials. Based on laboratory tests, medical history, physical examinations and interview questions, healthy, HIV-1-uninfected adults, aged 18-50 years were enrolled. In some studies, preexisting immunity to adenovirus serotype 5 (Ad5) has been found to decrease immune responses to recombinant adenoviral serotype 5 vaccines [16,20]. On the other hand, in regions of the world most affected by HIV, 75-99% of populations are seropositive for Ad5 neutralizing antibodies [21]. Thus, to more closely reflect the situation in many areas of the world, only volunteers possessing detectable levels of neutralizing antibodies against Ad5 (titer $1:12) were enrolled. All participants were counseled about HIV risk reduction and pregnancy prevention and assessed about potential social impacts at each study visit.

Ethics
The study protocol was approved by the institutional review boards of all participating study sites: University of Alabama at Birmingham

Vaccines
The DNA vaccine (VRC-HIVDNA009-00-VP) was composed of 4 closed, circular DNA plasmids. One plasmid was designed to express clade B HIV-1 Gag/Pol/Nef polyprotein. The other 3 plasmids were designed to express HIV-1 Env glycoprotein from clade A, clade B, and clade C. The recombinant adenoviral vector product (VRC-HIVADV014-00-VP) (rAd5) was a replication defective, combination vaccine containing a mixture of 4 recombinant serotype 5 adenoviral vectors. Each vector expresses 1 of the 4 HIV antigens-clade B GagPol polyprotein, clade A Env, clade B Env, and clade C Env-in a 3:1:1:1 ratio.
The vaccines were administered in a prime-boost combination with 4 mg of the DNA vaccine prime administered intramuscularly (IM) as one 1 ml injection via the needle-free BiojectorH. The boost rAd5 vaccine was administered in one of three ways: one 1 ml injection of 1610 10 PU intramuscular (IM), one 0.1 ml injection of 1610 11 PU intradermal (ID), or one 1 ml injection of 1610 10 PU subcutaneous (SC).

Humoral assays
Binding antibodies by ELISA. Anti-Gag and Anti-Env binding antibody responses were determined by validated ELISAs. Sera from cryopreserved samples were tested in duplicate in microtiter plates (NUNC) coated with gp140 Env protein (Con S gp140, a group M consensus, provided by Drs. Liao and Haynes, Duke University). Sera were diluted and incubated with the antigens bound to the plate. The plates were washed with an automated and calibrated plate washer (Bio-Tek). Response was considered positive if the difference in duplicate antigen-containing and non-antigen-containing wells (OD antigen -OD non-antigen) had an optical density (OD) greater than or equal to an OD of 0.2 with background subtracted and the OD was $3 times the baseline OD (M2 plate reader, Molecular Devices). Standard curves were generated from the plot of absorbance (450) against the log of serum dilution and sigmoidal curves were fit using a four-parameter logistic equation (Softmax Pro) [22,23].
Neutralization assay. Neutralization was measured as a function of reductions in luciferase reporter gene expression after a single round of infection in TZM-bl cells as described [24,25]. TZM-bl cells were obtained from the NIH AIDS Research and Reference Reagent Program, as contributed by John Kappes and Xiaoyun Wu. Briefly, 200 TCID50 of virus was incubated with serial 3-fold dilutions of test sample in duplicate in a total volume of 150 ml for 1 hr at 37uC in 96-well flat-bottom culture plates. Freshly trypsinized cells (10,000 cells in 100 ml of growth medium containing 75 mg/ml DEAE dextran) were added to each well. One set of control wells received cells + virus (virus control) and another set received cells only (background control). After a 48 hour incubation, 100 ml of cells was transferred to a 96-well black solid plates (Costar) for measurements of luminescence using the Britelite Luminescence Reporter Gene Assay System (PerkinElmer Life Sciences). Neutralization titers were the dilution at which relative luminescence units (RLU) were reduced by 50% compared to virus control wells after subtraction of background RLUs. Response to an isolate was considered positive if the titer was $25. An assay stock of uncloned HIV-1 MN was produced in H9 cells and titrated in TZM-bl cells. Assay stocks of molecularly cloned Bal.26 (clade B, tier 1A), 92RW020.2 (clade A, tier 2) and 97ZA012.29 (clade C, tier 2), all matched to the vaccine strains and used as Env-pseudotyped viruses were prepared by transfection in 293T cells and were titrated in TZM-bl cells as described [25].
End of study diagnostic EIA testing. At the last study visit, participants were screened by commercial EIA (Abbot Laboratories HIV-1/HIV-2 rDNA, Biorad Genetic Systems HIV-1/HIV-2 Plus O and Biorad Genetic Systems rLAV EIA).

Cellular assays
IFN-c ELISpot assays. Validated IFN-c ELISpot assays [26] were conducted using previously cryopreserved PBMC stimulated ex vivo with pools of peptides 15 amino acids in length at a final concentration of 1 mg/ml. The peptide sequences were designed to incorporate the most frequent HIV 10mers from the Los Alamos National Laboratory online database (http://www.hiv.lanl.gov/ content/sequence/HIV/mainpage.html). Nine global PTE (potential T-cell epitope) peptide pools encompassing 1407 total peptides (three pools for Pol PTE and Env PTE, two pools for Gag PTE and one pool for Nef PTE peptides) were used [27]. PBMC incubated with media alone were used in the negative control wells, and PBMC treated with PHA were used in the positive control wells. Responses were measured as the number of spot-forming cells per million (SFC/ 10 6 ) PBMC and expressed as geometric means; the criteria for positive and negative responses were defined as previously described [28].
Intracellular cytokine staining (ICS) assays. Intracellular cytokine staining (ICS) assays were performed by flow cytometry using previously cryopreserved PBMC to determine both HIVspecific CD4+ and CD8+ T-cell responses [29]. In summary, thawed PBMC were incubated overnight and then stimulated for six hours with HIV-1 peptide pools in the presence of Brefeldin A (Sigma, St. Louis, MO) and 1 mg/ml each of 2CD28 and CD49d antibodies (BD Biosciences, San Jose, CA). The pools of HIV peptides were 15 amino acids in length and used at a final concentration of 1 mg/ml. The peptide sequences were designed as described above. Duplicate wells of PBMC incubated with DMSO were used as the negative control, and SEB (Sigma, St. Louis, MO) stimulation was used as the positive control. A previously validated 8-color ICS protocol was used to detect live IFN-c-and IL-2-secreting CD3+/CD8+ and CD3+/CD4+ HIVspecific T cells. For the flow cytometric analysis, the specimens were collected from 96-well plates using the High Throughput Sampler (HTS, BD) device on a BD LSRII and then analyzed using FlowJo software (Treestar, Inc; OR) and LabKey Flow [30]. Positive responses and criteria for evaluable responses were determined as previously described [29] and were based on background measurements and the number of T cells examined. Since separate criteria are applied for CD4+ and CD8+ cells, the total number of specimens included in each ICS analysis can differ between the CD4+ and CD8+ T-cell evaluations.

Statistical analysis
The primary goal was to determine if ID or SC route of administration of the rAd5 boost was superior to IM administration in eliciting vaccine-induced HIV-specific T cell responses at four weeks after the rAd5 boost. Because IM administration is the most common method to administer these products, IM administration would continue to be used with this product unless there is a large improvement in immune response.
Sample size. The original design of 30 participants per group had 82% power to detect a 2-fold difference in mean response magnitudes between two groups for a two-sided two-sample t-test with a Type I error rate of 0.05. Due to the suspension of rAd5 vaccinations, the sample size was reduced (61 of 90 enrolled participants) and data were instead analyzed per protocol using overall tests to determine if there were any differences among the three routes of administration. If any significant differences were found, then pair-wise comparisons were conducted and p-values adjusted accordingly.
Safety assessments. Safety data from enrolled participants were analyzed according to initial randomization assignment. For reactogenicity, the number and percentage of participants experiencing each type of reactogenicity sign or symptom was tabulated by severity. For a given sign or symptom, each participant's reactogenicity was counted once under the maximum severity after each injection. Kruskall-Wallis test was used for testing overall difference in terms of severity of symptoms by treatment group for each sign or symptom separately. When there was a significant overall difference, pairwise comparisons (IM vs ID, IM vs SC and ID vs SC) were performed and Bonferroni multiplicity adjustment was applied. Adverse experiences were tabulated using MedDRA preferred terms. The number and percentage of participants experiencing each specific adverse experience were tabulated by severity and by relationship to treatment. Each participant's adverse experience was counted once under the maximum severity or the strongest recorded causal relationship to study product. Boxplots of local laboratory values were generated for baseline values and for values measured during the course of the study to present the distribution of these data. Kruskal-Wallis test was used to test for differences in terms of severity of adverse events by treatment group.
Immunogenicity. Fisher exact test was conducted to compare if the response rates among the three routes of administration differed within each assay. For the comparison of magnitudes of responses, the non-parametric Kruskal-Wallis test was performed to determine if there were any overall differences among the three routes.

Enrollment and follow-up
Between November 2006 and October 2007, 90 participants were enrolled and randomized to one of the three study groups at sites in the US and Peru. The median age of participants was 27.0 years; 41% were women, 40% were non-Hispanic Whites, 32% were Hispanic and 18% non-Hispanic Black (Table 1). There were no significant differences in gender, race/ethnicity or age by study group. All participants received the first dose of DNA vaccine, 88 (99%) received the second dose of DNA vaccine, 84 (93%) received the third dose of DNA vaccine and 61 (68%) received the booster dose of rAd5 vaccine. Of the 29 missing the rAd5 booster dose (IM: 10, ID: 9, SC: 10), 22 (75.9%) were due to the suspension of vaccination after the Step Study results were released. Of the remaining 7 participants, 3 missed due to inability to schedule vaccination within the target window, 1 due to receipt of yellow fever vaccine (ID group) and 1 due to military enlistment (SC group). One discontinued vaccination after 2 DNA doses due to hypoesthesia which was deemed to be mild and probably not related to study vaccine (ID group). The numbness resolved within 20 days and the study participant declined further injections. The final participant discontinued vaccination due to a decrease in hemoglobin levels after only one dose of DNA vaccine (SC group). The decrease was deemed as mild and not related to study product but the study safety team decided to discontinue vaccination. Hemoglobulin levels returned to baseline values within 22 days. Retention at the last study visit at 12 months was 94% (Figure 1).

Vaccine Safety
Overall, the vaccines were well-tolerated. Systemic reactogenicity, including malaise, myalgia, headache, nausea, vomiting, chills and arthralgia was generally mild, with 39% of participants reporting no systemic symptoms and 34% reporting mild symptoms. After the three DNA doses, there were no significant differences in local or systemic reactogenicity by study arm. In contrast, after the rAd5 boost, significant differences by study arm were found in severity of headache (p = 0.01), pain (p = 0.04) and erythema and/or induration (p,0.0001) ( Table 2). Pairwise comparisons found significantly more severe headaches and pain in the SC group compared to the ID group (adjusted p = 0.0167 for both). Erythema and/or induration was significantly more severe in the ID group compared to IM group (adjusted Adverse events were reported by 83 (92.2%) of participants; 80 of those participants experienced mild or moderate events. Two participants had severe adverse events (one with abnormal weight loss and one with alanine aminotransferase increase) and one with a life-threatening event (malaria and fetal loss). None of these  adverse events were determined to be related to study product. There were no significant differences in severity of adverse events by study arm (p = 0.60).

Immunogenicity
Humoral responses. Binding antibodies by ELISA at four weeks after the rAd5 boost are reported for 56 (91.8%) of the 61 participants who received the rAd5 boost (18 in the IM group, 20 in the ID group and 18 in the SC group). Five participants were not included in the analysis (IM = 2, ID = 1, and SC = 2) since their study visits occurred outside of the target windows. The response rates in the three study arms were 66.7% ( Cellular responses. Vaccine-induced HIV-1 specific T-cell responses were measured at 2 weeks after the third DNA vaccination and 4 weeks after the rAd5 vaccine boost. IFN-c ELISpot assays were completed on 57 participants at two weeks after the third DNA dose and on 51 at four weeks after the rAd5 boost. The frequency of IFN-c ELISpot HIV-specific responses after the DNA immunization was 42.1% and after the rAd5 vaccine boost was 41.2% (Table 3). There were no significant differences (p = 0.82) in the percent responding at 4 week post rAd5 boost by study group: 46.7% in the IM group (7/15; 95% CI: 24.8, 69.9), 35.3% in the ID group (6/17; 95% CI: 17.3, 58.7) and 44.4% in the SC group (8/18; 95% CI: 24.6, 66.3). The specificity of the ELISpot responses was heavily biased towards Env. Figure 3 provides the boxplots of IFN-c HIV-specific responses by study visit and study arm for each of the antigens (Env, Gag, Nef and Pol). The median magnitude of Env-specific responses in the 24 positive responders post DNA priming was 116 SFC/10 6 PBMCs (range: 66-418) and 177 SFC/10 6 PBMCs (range: 62-2204) in the 21 positive responders after rAd5 boost. There was no significant difference in ELISpot magnitude based on delivery route (Env: p = 0.40; Gag: p = 0.11; Nef: p = 0.17; Pol: p = 0.10).
For the ICS assay, CD4+ T-cell responses are reported on 54 (88.5%) of the 61 participants at two weeks after the third DNA dose and 49 (80.3%) at four weeks after the rAd5 boost. HIV-1specific response rates for CD4+ T-cells expressing IL-2 and/or IFN-c using the ICS assay were found to be lower after rAd5 boosting compared to response rates after initial DNA priming (Table 4). Response rates at 4 weeks post rAd5 boost were not significantly different (p = 0.67) by study arm: 29.4% in the IM group (5/17; 95% CI: 13.3, 53.1), 20.0% in the ID group (3/15; 95% CI: 7.0, 45.2) and 35.3% in the SC group (6/17; 95% CI: 17.3, 58.7). Env was most frequently recognized after DNA priming and after rAd5 boosting, responses broadened somewhat to include Gag (Figure 4).

Discussion
This study was designed to determine the effect of route of administration of a boost on safety and immunogenicity of a primeboost regimen of two HIV vaccines in Ad5-seropositive volunteers. The main limitation of this study is the reduction of sample size, and thus limited power, due to the discontinuance of vaccinations after the results of the Step Study were available. rAd5 immunizations were resumed only in studies that excluded participants who were found to have pre-existing immunity to Ad5, and a DNA prime-rAd5 boost regimen is now being tested in an efficacy trial among Ad5 seronegative participants [31]. Resuming vaccination with rAd5 was not implemented in this protocol.
Overall, the vaccines were found to be well tolerated, adding to the overall safety profile of these vaccines [18,[32][33][34]. A greater frequency and level of severity of erythema and induration were observed among those receiving rAd5 vaccine by the ID and SC routes compared to the IM route. This finding is consistent with studies of other vaccines delivered through the ID and SC routes, including influenza, hepatitis A and B, modified vaccinia Ankara, and anthrax [7][8][9][10][11]13,15].
Varying the route of administration of the rAd5 boost to increase immunogenicity was of particular interest since pre-existing immunity to adenovirus type 5 has been shown to diminish the immune response to rAd5 vaccines [35,36]. Given higher rates of some reactogenicity and not sufficient evidence to show that there is a difference by route of administration, adopting a SC or ID route would likely not aid in addressing prior Ad5 immunity.  Other studies of the effect of administration route have been reported for other candidate HIV vaccines. Bansal et al [37] compared prime vaccination with a DNA vaccine delivered by ID and IM route, boosted by a protein vaccine delivered by IM. They found that the ID group had the lowest magnitude of Env-specific CD4 T-cell responses after the DNA prime but there were no  differences after the protein boost. Another study compared ID to IM route of an HIV-lipopeptide candidate vaccine and found that while more local reactions were observed with the ID route of administration, HIV-specific CD8+ T-cell responses were similar between groups [38]. The effect of administration route was tested in this study using vaccines with demonstrated immunogenicity in multiple clinical trials [17,18,39]. In one study among participants in the United States, of whom only a minority had pre-existing immunity to Ad5 in contrast to this study, the DNA prime (using either 4-or 6plasmid vaccines) combined with rAd5 boost was more highly immunogenic than observed in this study [18]. Another study using the 4-plasmid DNA and rAd5 vaccines used in the current study found a somewhat higher frequency of IFN-c ELISpot responses (58% vs. 41%) and higher CD8+ T-cell response rates (44% vs. 32%) but lower CD4+ T-cell response rates (17% vs 29%) [40]. However, many of the participants had no pre-existing Ad5 immunity. In studies using the 6-plasmid DNA priming, HIVspecific T cell responses were found at slightly lower frequencies among participants with pre-existing Ad5 immunity in studies conducted in East Africa [32,33]. In this study utilizing the 4plasmid DNA for priming in participants who were exclusively Ad5 seropositive, the relatively low T-cell response was not unexpected. A high frequency of HIV-specific antibody responses was seen, validating the immunogenicity and stability of the product.
In summary, limited by the reduced sample size, this study does not provide sufficient evidence to show that there were any differences in immunogenicity by route of administration. With a higher frequency of headache, pain and erythema and/or induration after ID and SC administration, this study does not support changing route of administration for the rAd5 boost.