**1. Introduction**

According to the latest reports, there are currently 37 million people infected with HIV, the majority of whom live in sub-Saharan Africa [1]. Despite the existence of anti-retroviral therapy, 1.8 million people were newly infected in 2017, and almost one million people died due to HIV-related disease [1]. Developing a safe, efficacious, and accessible HIV vaccine would be the optimal solution for the prevention of HIV-1 infection as well as reduction of HIV-related diseases. Evidence demonstrating the role of T-cell responses directed against HIV-1 in the control of viral replication is growing [2,3].

HIV-1-specific CD8<sup>+</sup> T-cells have been detected in exposed seronegative individuals, and CD8<sup>+</sup> cytotoxic T-lymphocytes (CTL) responses targeting HIV-1 Gag have been associated with reduced viral loads in infected individuals [4–6].

The "HIVACAT T-cell immunogen" (HTI) was designed to cover T-cell targets, against which T-cell responses are predominantly observed in HIV-1-infected individuals with low HIV-1 viral loads [7]. This immunogen has been manufactured as naked DNA, as well as introduced into viral vaccine vectors and an immunization regimen consisting of three injections of DNA encoding HTI followed by a boost with modified vaccinia Ankara (MVA)-vectored HTI has been shown in C57BL/6 mice to result in the induction of HIV-1 specific T-cell responses to most antigen regions included in its design. High magnitudes of HIV-1 specific T-cells were also induced in macaques following three injections of DNA.HTI and two injections of MVA.HTI [7]. It is well known that delivering a boosting injection of an immunogen using a viral vector such as MVA can boost the response of DNA immunization and increase the magnitude of responses. Such heterologous prime-boost regimens have been used for a number of immunogens in the context of HIV [8].

Upon intramuscular delivery, DNA vaccines are thought to induce cellular immunity through antigen synthesis and presentation to T-cells. The immunity is induced by the direct transfection of antigen-presenting cells (APCs) or through cross-presentation by APCs [9,10]. Similar to DNA, the *Mycobacterium bovis* strain, Bacillus Calmette-Guérin (BCG), when used as a vaccine vector also induces immunity by targeting APCs, although this occurs not through transfection, but rather through infection. Recombinant BCG is a promising live attenuated bacterial vaccine vector for inducing T-cell immunity, as it is a slow-growing organism that provides a persistent low-level antigenic exposure upon the infection of macrophages and APCs, and could drive effector and memory T-cell responses [11]. In the context of anti-tuberculosis (anti-TB) immunity, the pathogen for which BCG currently is licensed as a vaccine, immune responses in humans are predominantly T helper 1 (Th1)-cell mediated. Infected dendritic cells migrate to lymph nodes and activate antigen-specific CD4<sup>+</sup> T-cells in the presence of cytokines such as interleukin (IL)-18 and IL-12 [12]. The T-cell repertoire induced by vaccination is broad, targeting multiple mycobacterial antigens [13]. BCG is safe and is currently administered to 80% of infants in countries where it is part of the national childhood immunization program [14]. Aside from the induction of anti-TB immune responses, vaccination has been linked to decreased mortality due to other infections in infants [15,16]. This phenomenon has been related to the enhanced monocyte function observed in humans following vaccination, which is characterized by CD11b, TLR-4 expression, and increased cytokine production [17].

Recombinant BCG (rBCG) has been utilized as a vector to express HIV-1/Simian Immunodeficiency Virus (SIV) antigens and assessed regarding the induction of specific T-cell responses in several animal models [18]. The need for adjuvants is overcome as its cell-wall peptidoglycans and lipoproteins act as an adjuvant on their own [19–22]. BCG also has several advantages as a live vaccine vehicle, it is easy to mass-produce with low cost, and is heat stable [23]. It is also suitable for neonates, as vaccination is not affected by maternal antibodies [24,25]. Finally, it has a proven safety record after having been delivered as a TB vaccine to over three billion individuals [26]. As with DNA vaccination, rBCG-based HIV vaccines on their own may induce low-level specific immune responses, and thus they are often combined with virally vectored immunogens in heterologous prime-boost vaccination regimens where they have been shown to increase HIV-1 specific immune responses [27–30]. However, a rBCG expressing HIV-1 group M consensus Env vaccine on its own was shown to induce comparable immune responses both in the female reproductive tract and lungs when compared with adenovirus prime/recombinant vaccinia virus boost immunization [31]. The applicability of rBCG as a priming agent was demonstrated by Ami *et al.*, where priming with rBCG and boosting with a replication-deficient vaccinia virus strain expressing SIVgag was able to confer protection in Cynomolgus macaques against mucosal challenge with pathogenic SHIV. Interestingly, effective protection was not achieved in animals receiving the opposite combination or the vaccine modalities delivered on their own [32]. These data support the use of rBCG as a priming vector for an HIV-1

vaccine, whereby the use of rBCG could potentially strengthen or qualitatively modify the immune response when combined with DNA and/or a viral vector expressing an HIV-1 immunogen.

Here, we present the construction and characterization of recombinant BCG expressing the HTI immunogen, the BCG.HTI2auxo.int vaccine, harboring the integrated 2auxo expression cassette in their chromosome. The HTI immunogen has previously been assessed in mice and macaques when delivered as a prime-boost regimen vectored by DNA and MVA [7]. We demonstrate that BCG.HTI2auxo.int delivered in combination with ChAdOx1.HTI increases the HIV-1 specific T-cell responses in adult BALB/c mice. Priming with wild type BCG (BCGwt) was also shown to similarly increase the magnitude of the response, but significantly decreased the breadth of the T-cell responses. Furthermore, we demonstrate that priming with rBCG, in some mice, can alter the immunodominance profile of the vaccine-induced T-cell response.
