The therapeutic HIV Env C5/gp41 vaccine candidate Vacc-C5 induces specific T cell regulation in a phase I/II clinical study

More than 30 years after its discovery, Human Immunodeficiency Virus type 1 (HIV) remains a global challenge with more than 36 million people infected worldwide [1]. Despite the achievements of effective antiretroviral therapy (ART), patients with access to ART control viremia but are not cured. Chronic immune activation is reduced, but persists [2] and is associated with non-AIDS morbidity and mortality [3, 4], an increasing concern in the long term care of HIV patients. Search for alternative treatment strategies are therefore warranted.

Vacc-C5 is a peptide-based therapeutic vaccine candidate that aims to induce non-neutralising antibodies (AB) to the 5th constant domain (C5) of the HIV-1 envelope glycoprotein gp120 in conjunction with part of the transmembrane glycoprotein of gp41. The presence of AB towards the C5 region has previously been shown to correlate with a slower disease progression [5‐7]. Analysis of serum from treatment naïve patients has demonstrated higher levels of non-neutralising anti-Vacc-C5 AB levels in HIV natural viral suppressors with moderate viral loads (<10,000 copies/ml) than in patients with higher viral loads (>10,000 copies/ml). Furthermore, longitudinal analysis of HIV patients has revealed an inverse correlation between anti-Vacc-C5 AB levels and disease progression [8]. The C5 domain has sequence similarity to peptide binding sites on HLA molecules [9] and can bind both defined class II and class I restricted peptides, bindings that are blocked with anti-C5 monoclonal AB [10]. It has therefore been proposed that exposed C5 could amplify chronic immune activation, either directly by cross-signalling via HLA-like motifs or by presentation of peptides that facilitate alloactivation-like responses [11]. Our hypothesis was that any detrimental effects of C5 explain the inverse relationship between anti-C5 AB levels and clinical progression, and that enhancement of non-neutralising anti-C5 AB by vaccination with Vacc-C5 could reduce exposure of C5.

Against this background, a phase I/II trial was conducted with the primary objective to evaluate safety of Vacc-C5 at three dose levels administered either intradermally or intramuscularly with two different adjuvants while on effective ART. Local low dose granulocyte-macrophage colony-stimulating factor (GM-CSF) facilitates dermal dendritic cell maturation and migration to the lymph nodes for antigen presentation [12, 13] and has been used in former studies with intradermal administration of the peptide-based HIV Gag p24 vaccine Vacc-4× [14, 15] and with peptide-based cancer vaccines [16]. Alhydrogel, the only adjuvant approved and marketed for worldwide use (e.g. pneumococcal vaccine Prevnar and HPV vaccine Gardasil), was applied in the intramuscularly administered vaccine.

The secondary objectives were to explore humoral responses to Vacc-C5, HIV-associated immune activation and T cell immunogenicity. T cell responses to HIV antigens, such as C5, may be strongly regulated in chronic HIV infection [17‐19], and immune regulatory mechanisms to HIV vaccine antigens may therefore be important to consider before and during therapeutic HIV vaccination [20, 21]. We have previously assessed an exploratory parameter for vaccine antigen-induced T cell regulation, based on simultaneous blockade of the two potent, soluble cytokines downregulating HIV-specific T cell proliferation, a key feature of effector T cells [22, 23]. Our approach does not intend to differentiate the mechanistic pathways of immune regulation, but rather measure the net impact on proliferative capacity of effector T cells to simplify implementation in a clinical trial setting.

This parameter of immune regulation was addressed also in the present study, and whereas only marginal cellular and humoral immune responses were induced, vaccine-specific regulation markedly increased after completed vaccination. These results may illustrate the complexity of therapeutic HIV vaccination in general and may also hopefully encourage further attention on assessments of immune regulation.

In this phase I/II trial the therapeutic vaccine candidate Vacc-C5 was administered over six months with initial immunisation and two booster periods in the presence of ART. The primary objective was to evaluate safety of Vacc-C5 at three dose levels with two different adjuvants, while the secondary objectives were to explore humoral responses to Vacc-C5, HIV-associated immune activation and T cell immunogenicity. Immune regulation of T cell proliferative responses was an optional exploratory objective. Vacc-C5 was well tolerated, and all 36 patients completed the study. Despite rather low pre-existing anti-Vacc-C5 AB levels at baseline, only marginal changes in humoral responses to Vacc-C5 were found. Overall, both AB levels and T cell IFN-γ ELISPOT responses decreased throughout the study. The two modes of administration were comparable, with no significant differences in Vacc-C5-induced responses between them. The higher doses appeared disadvantageous as AB levels declined in the MD group, and a reduction in IFN-γ responses was seen in the HD group.

Although intra- and inter-assay variations might cause some of the minor changes in anti-Vacc-C5 AB levels, the main conclusion is that marginal changes were induced with the doses, schedule and adjuvants tested. It should be noted that Vacc-C5 has been shown to induce robust humoral responses in two classical animal models; in sheep using Alhydrogel and in rabbits using rh-GM-CSF as adjuvants (Bionor Pharma, data on file). Although not tested in HIV seronegative volunteers for regulatory and ethical reasons, Vacc-C5 was expected to induce humoral responses in humans as well. Moreover, if consistent enhancement of Vacc-C5 AB levels had been observed, additional characterisation of Vacc-C5-specific Ig isotype and subclass responses could have been assessed, although no reference to such information is available from reports on anti-C5 AB levels and slow disease progression [5‐7].

Despite the minimal changes in anti-Vacc-C5 AB levels, some interesting data regarding T cell responses were observed. It should be noted that the T cell responses to Vacc-C5 were generated by pooled peptides representing the Env vaccine heterodimer, and did not differentiate responses between the two short Env regions. Overall, Vacc-C5 induced only small changes in cellular responses from baseline to end of study. After the first booster period, however, CD8⁺ T cell responses significantly increased, but were subsequently reduced after the second booster period. Strikingly, Vacc-C5-induced cytokine-mediated immune regulation was the only parameter that markedly increased during the trial. We assessed a parameter for regulation by blocking the two inhibitory cytokines IL-10 and TGF-β in Vacc-C5-stimulated cell cultures. In previous studies, this parameter has been associated with rapid disease progression [25] and inconsistent boosting effects of the HIV p24-based Vacc-4× [22]. In a recent study, where Vacc-4× was administered intranasally, immune regulation could explain most cases of T cell anergy towards HIV p24 as well as the variable final responses [23]. In the present study, increased immune regulation might partially explain the coinciding declining T cell responses after the second booster period (Figs. 4a and 6a) and possibly the lack of anti-Vacc-C5 AB responses as well. Our data find support in a recent primate prime and boost vaccination model which has shown that a SIV vaccine can induce suppressor cells and thereby dampen protective CD8⁺ T cell responses [26].

Immune regulation may be quantified in different ways. In previous papers [22, 23, 25], we have defined antigen-induced cytokine-mediated regulation as a parameter which also included control cultures with and without mAbs to IL-10 and TGF-β. Net regulation is a more simplified assessment, as reported by others [27] and illustrated in Fig. 5. Net regulation might be easier to comprehend, and is probably a more precise term to use in the discussions of this phenomenon. A variant of net regulation is fold increase, i.e. the ratio between proliferative responses in stimulated cultures with and without blockade. Here we explored immune regulation in either way with generally identical results.

Our approach is supported in part by Brockman et al. [27] who reported that blockade of the IL-10 pathway increased proliferative T cell responses to HIV antigens in viremic patients, whereas elite controllers and patients on effective ART failed to respond to the blockade. After immunisation with Vacc-C5, we found that blocking of both IL-10 and TGF-β significantly increased the proliferative capacity to Vacc-C5 in these non-viremic patients. We have previously shown that this parameter of T cell regulation is HIV antigen-specific and not an inhibitory response to T cell activation in general. In an HIV positive individual, regulation may be high to one specific HIV antigen but low or absent to other HIV antigens and non-HIV antigens. Moreover, regulation induced by different HIV antigens are not correlated in a particular sample [25, 28]. Therefore, it is likely that the regulation induced by Vacc-C5 is vaccine-specific and not a global T cell regulation induced by the vaccine.

We chose to assess regulation by blocking two inhibitory cytokines and quantify the total inhibitory effect on Vacc-C5-specific T cell proliferation. Since proliferation is a fundamental feature of effector T cells, the impact of regulation on T cell proliferative capacity may provide additional information, relevant for choosing patients expected to benefit from therapeutic vaccination.

In the era of an HIV cure, there has been a renewed interest in immunotherapy [29, 30]. Efficient HIV-specific CD8⁺ cells able to eliminate reactivated latently infected cells are essential to put HIV into remission [31], but these cells are hampered in chronic HIV infection [32, 33]. Immunotherapy may therefore be required for an effective functional HIV cure, but a possible challenge may be induction of immune regulation.

Both proliferative responses and immune regulation differed considerably from one patient to another throughout the study. Consistent with previous results [22, 23], our data highlight the importance of individualised immunotherapy. Moreover, they support the view that both proliferative responses and regulation should be monitored in therapeutic immunisation trials, because T cell proliferation and immune regulation induced by a defined HIV antigen are generally not correlated [25, 28].

This dose escalation study, with safety as primary objective, has limitations, particularly small size of each adjuvant group. Regardless, it was possible to analyse secondary endpoints in subgroups, but with the risk of type II statistical errors. Moreover, the study was planned and approved with no unimmunised or adjuvant only control groups included. Another limitation may be the selection of patients for proliferative assays. Although proliferation and regulation were not tested in all participants, these patients were comparable to the rest of the cohort regarding clinical characteristics, dose and administration as well as AB levels and IFN-γ ELISPOT responses at baseline and end of study. The regulatory assays could have been expanded to explore the individual contribution by the two inhibitory cytokines and to include other well characterised regulatory pathways by blockade of T cell receptors like CTLA-4 and PD-1. Still, all the above mentioned regulatory pathways are recently addressed by our group, and the combined blockade of IL-10 and TGF-β was found to best identify net regulation induced by Env antigens [28]. Finally, more direct methods to assess regulation might have provided even more definitive data, e.g. separation of T regulatory cells by cell sorting and add-back experiments with the sorted cells. However, these in-depth analyses would require substantial PBMC and resources unrealistic to implement in a clinical trial setting.

The development of Vacc-C5 was based on the hypothesis that the C5 region is involved in chronic immune activation and that blocking this domain with non-neutralising AB may lead to reduced immune activation and disease progression [10, 11]. However, in this study including patients on ART, increased levels of anti-Vacc-C5 AB were not induced, and the activation markers HLA-DR/CD38 and the two chosen pro-inflammatory mediators IP-10 and IL-6 remained unchanged. Since the C5 domain shows sequence similarity to HLA-DR, the HLA phenotype of study participants may potentially have influenced the results. Finally, it might be possible that Vacc-C5 administered with another adjuvant or with a carrier molecule could have generated AB with effects on HIV-associated immune activation.

In this study investigating safety and immunogenicity of the therapeutic vaccine candidate Vacc-C5, only marginal immune responses were safely induced, whereas vaccine-specific regulation clearly increased after completed vaccination. It is said that we learn more from failure than success (Samuel Smiles 1812–1904, Scottish author and reformer). Despite the absence of enhanced immune responses, this study might still provide insights to improve future therapeutic vaccine designs and immunotherapy and encourage further focus on immune regulation in therapeutic vaccination.