TLR2-Modulating Lipoproteins of the Mycobacterium tuberculosis Complex Enhance the HIV Infectivity of CD4+ T Cells

Co-infection with Mycobacterium tuberculosis accelerates progression from HIV to AIDS. Our previous studies showed that M. tuberculosis complex, unlike M. smegmatis, enhances TLR2-dependent susceptibility of CD4+ T cells to HIV. The M. tuberculosis complex produces multiple TLR2-stimulating lipoproteins, which are absent in M. smegmatis. M. tuberculosis production of mature lipoproteins and TLR2 stimulation is dependent on cleavage by lipoprotein signal peptidase A (LspA). In order to determine the role of potential TLR2-stimulating lipoproteins on mycobacterial-mediated HIV infectivity of CD4+ T cells, we generated M. smegmatis recombinant strains overexpressing genes encoding various M. bovis BCG lipoproteins, as well as a Mycobacterium bovis BCG strain deficient in LspA (ΔlspA). Exposure of human peripheral blood mononuclear cells (PBMC) to M. smegmatis strains overexpressing the BCG lipoproteins, LprF (p<0.01), LprH (p<0.05), LprI (p<0.05), LprP (p<0.001), LprQ (p<0.005), MPT83 (p<0.005), or PhoS1 (p<0.05), resulted in increased HIV infectivity of CD4+ T cells isolated from these PBMC. Conversely, infection of PBMC with ΔlspA reduced HIV infectivity of CD4+ T cells by 40% relative to BCG-infected cells (p<0.05). These results may have important implications for TB vaccination programs in areas with high mother-to-child HIV transmission.


Introduction
Despite significant improvements in expanding access to global HIV prevention and treatment services in the last decade, an estimated 2 million people were newly infected with HIV in 2014 [1]. A majority of these new infections occurred in tuberculosis-endemic countries, where BCG vaccination at birth is virtually universal. The success of effective behavioral interventions to prevent HIV infection, such as condom use, has been limited by suboptimal uptake in at-risk populations. Universal treatment of all HIV-infected individuals with antiretroviral therapy (ARV) could halt new transmission, but identifying, linking and retaining most HIV-infected individuals in care is challenging [2]. Biological determinants of HIV susceptibility are complex and not fully understood, but could improve HIV prevention strategies. For example, the removal of HIV target cells in the male foreskin through circumcision is a simple one-time procedure that reduces the risk of HIV by 60% [3]. Treatment of sexually transmitted infections also reduces HIV risk through a variety of mechanisms, such as restoring integrity of the mucosal barrier, and reducing the number of activated immune target cells at sites of HIV exposure [4,5,6]. Minimizing inflammation is important because activated CD4 + T cells are much more susceptible to infection than resting T cells [7,8].
HIV pathogenesis and susceptibility is associated with immune activation [9]. HIV preferentially infects activated T cells, which compared to quiescent cells, express a large number of host factors required for HIV replication [10,11]. Co-infection with pathogens common in HIV-endemic areas has been postulated to have an impact on HIV susceptibility [7,12]. Compared to Europeans, Africans and Asians have elevated systemic immune activation markers, which could drive regional disparities in HIV rates around the world [12,13]. Using a single cycle HIV infectivity assay, we have shown that ex vivo infection of peripheral blood mononuclear cells (PBMC) with Mycobacterium tuberculosis or M. bovis BCG (MTB complex), but not with M. smegmatis, enhances susceptibility of CD4+ T cells to HIV infection independent of T cell activation state. Cells exposed to M. smegmatis were found to express lower levels of TLR2, and their ability to become infected with R5-tropic pseudovirus was reduced upon siRNAmediated silencing of TLR2 [14]. In TB infection, lipoproteins are important mediators of TLR2 activation. The lipoprotein signal peptidase, LspA, is an important virulence factor necessary for mycobacterial production of mature lipoproteins [15,16]. In the absence of LspA, cleavage of the conserved N-terminal "lipobox" of prolipoproteins does not occur, preventing them from being processed into mature lipoproteins [16,17,18]. Here, we show that the enhanced HIV infectivity of CD4+ T cells derived from M. bovis BCG-infected PBMC requires LspA. Furthermore, the overexpression of M. bovis BCG-derived lipoproteins, including LprF, LprH, LprI, LprP, LprQ, MPT83, and PhoS1, by M. smegmatis results in significantly greater ex vivo HIV infection of human PBMC-derived CD4+ T cells. The ability of these lipoproteins to increase the HIV infectivity of CD4+ cells is reversed by chemical inhibitors of TLR2 signaling.
Addition of the dual TLR2/TLR4 inhibitor, OxPAPC, significantly reduced the HIV infectivity of CD4+ T cells following infection of PBMC with M. bovis BCG, while addition of the TLR4 inhibitor, CLI-095, showed no effect (Fig 1C), confirming our previous findings that M. bovis BCG infection of PBMC increases the HIV infectivity of CD4+ T cells through a TLR2-dependent mechanism [14]. Addition of the TLR2 agonist zymosan to M. smegmatis cultures prior to PBMC infection increased HIV infectivity of isolated CD4+ T cells 6.2-fold relative to untreated M. smegmatis cultures (p<0.01; Fig 1C). Whereas the addition of OxPAPC to the M. smegmatis knock-in strains overexpressing lprP or mpt83 during PBMC infection reversed the enhanced HIV infectivity of isolated CD4+ T cells to levels similar to those following infection with wild-type M. smegmatis, addition of CLI-095 had no such effect on PBMC infected with the mpt83 knock-in strain ( Fig 1C). OxPAPC alone did not alter HIV uptake by CD4+ T cells uninfected by mycobacteria (data not shown). Taken together, these results support our

Deficiency of lipoproteins reverses M. bovis BCG-mediated induction of HIV infectivity of CD4+ T cells
LspA cleaves the signal peptide from the C-terminus of mycobacterial proteins during maturation [16]. Mycobacterial strains lacking LspA fail to express mature lipoproteins and exhibit impaired TLR2 signaling [15,16]. We generated a M. bovis BCG strain deficient in the lspA gene (ΔlspA) using a targeted suicide plasmid [22]. Pre-infection of PBMC with ΔlspA resulted in significantly reduced HIV infectivity of isolated CD4+ T cells relative to those infected with wild-type M. bovis BCG (Fig 2, p = 0.013), although HIV infectivity of CD4+ T cells following PBMC infection with the mutant remained significantly higher than that observed following that with wild-type M. smegmatis (p<0.01).

Discussion
Although vaccination with M. bovis BCG is contraindicated in HIV-infected individuals due to the risk of disseminated disease [23], this vaccine continues to be administered around the world to neonates at risk for acquiring HIV through contaminated breast milk of HIV-infected mothers. Therefore, a deeper understanding of how M. bovis BCG infection enhances the HIV infectivity of CD4+ T cells may lead to safer TB vaccines, especially in HIV co-endemic areas. Previous work from our group identified a TLR2-mediated increase in HIV infectivity of CD4+ T cells following infection of PBMC with M. bovis BCG compared to M. smegmatis [14]. This phenomenon was independent of T cell activation markers CD38 and HLA-DR. Here, we further elucidate the mechanism of increased infectivity, identifying a key role for lipoproteins. Interestingly, although M. smegmatis expresses several TLR2-stimulating lipoproteins and LspA [24], these do not appear to modulate CD4+ T cell-HIV interactions in the same way as their M. tuberculosis complex counterparts. Thus, the overexpression of M. bovis BCG lipoproteins in M. smegmatis leads to increased HIV infectivity of PBMC-derived CD4+ T cells. Conversely, the infection of PBMC with a strain of M. bovis BCG unable to produce mature lipoproteins resulted in significantly reduced HIV infectivity of CD4+ T cells. We chose to evaluate a ΔlspA M. bovis BCG mutant for several practical reasons. First, there is significant redundancy in TLR2 signaling by mycobacterial lipoproteins, such that disrupting individual lipoproteins may not impact ex vivo HIV susceptibility. Because LspA is required for cleavage of prolipoprotein signal peptides prior to trafficking of all lipoproteins to the cell wall, the ΔlspA M. bovis BCG mutant allowed for a more comprehensive assessment of the role of lipoproteins on HIV susceptibility of CD4+ T cells. LspA-deficient M. tuberculosis mutants lack cell wall-associated lipoproteins and exhibit decreased TLR2 agonist activity [15,16]. In addition, LspA is required for full M. tuberculosis virulence in macrophages and in the mouse model [16]. However, further studies are required to evaluate the immunogenicity of the ΔlspA M. bovis BCG mutant relative to the wild-type strain.
There are several important limitations to our study. While disruption of lipoprotein processing reduced HIV infectivity of CD4+ T cells, this reduction did not achieve levels observed following exposure to M. smegmatis. This suggests that there are non-lipoprotein factors influencing the BCG-mediated increase in HIV infectivity of CD4+ T cells. Glycolipids of the mycobacterial cell wall, particularly phosphatidylinositol mannoside 6 (PIM6) [25], have also been implicated in enhancing TLR2-mediated HIV replication in T cells and may be important to explore in our model. It should also be noted that the M. smegmatis lipoprotein knock-in strains used in the current study express these lipoproteins at much higher levels than does wild-type M. bovis BCG, and we cannot implicate a specific role for any individual lipoprotein. It is highly likely that M. tuberculosis complex-specific lipoproteins and LspA modulate interactions between HIV and CD4+ T cells beyond simply stimulating TLR2. For example, an lspA-deficient M. tuberculosis mutant showed impaired resistance to malachite green, suggesting a cell wall permeability defect [REF], which might alter macrophage functions in a TLR2-independent manner and reduce susceptibility of CD4+ T cells to HIV infection. Furthermore, growth kinetics and other, non-lipoprotein-related genetic differences between the nonpathogenic M. smegmatis and M. bovis BCG may influence HIV infectivity of CD4+ T cells through differential induction of host defense mechanisms, such as cytokine production and cell death pathways. Finally, it is difficult to extrapolate our findings from an ex vivo model of HIV susceptibility to HIV risk in human populations. Interestingly, recent work suggests that immunization of infant rhesus macaques with M. bovis BCG renders them more susceptible to SIV infection following repeated challenge (Kristina De Paris, personal communication). Given that BCG is the most commonly used vaccine worldwide [26], even a small effect on maternal-to-child transmission of HIV could be of public health significance.
Our study demonstrates that lipoproteins expressed by virulent mycobacteria may activate immune pathways that enhance T-cell susceptibility to HIV. Future studies will focus on extending our ex vivo observations to clinically relevant animal models, such as humanized mice and nonhuman primates, as well as testing the protective efficacy of M. bovis BCG ΔlspA against aerosol challenge with virulent M. tuberculosis relative to the wild-type and complemented strains in guinea pigs. Finally, TLR2-independent pathways responsible for the M. bovis BCG-mediated increase in HIV infectivity of CD4+ T cells will be further explored. Given the magnitude of the TB epidemic, particularly in the regions of the world with the highest rates of HIV, these results highlight the need to measure HIV risk when evaluating the safety of TB vaccines at a population level.

Ethics Statement
Written informed consent was obtained from all subjects prior to entry into the study. The study was approved by the Johns Hopkins University School of Medicine Institutional Review Board. All studies were conducted in accordance with institutional and federal guidelines and regulations on human subjects research.

Subjects
HIV-seronegative healthy individuals (ages 25-35) without prior history of BCG vaccination or TB infection were recruited at Johns Hopkins University School of Medicine. All subjects were confirmed to have negative tuberculin skin tests prior to entry into the study.

Cell Isolation and Culture
Peripheral blood mononuclear cells (PBMC) were isolated from healthy volunteers and antigenic stimulation of cells was performed by incubating cells with whole bacteria or antigen for 48-72 hours at 37°C with 5% CO 2 . For infectivity assays, unstimulated cells were used as negative control and PHA (50 μg/ml) stimulated cells were used as positive control to ensure that cells were responsive to in vitro stimulation. Inhibition of TLR signaling was achieved through incubation of PBMC with OxPAPC or CLI-095 (40 μg/ml).

Bacterial Strains and Mycobacterial Infection
M. bovis BCG (Copenhagen), M. tuberculosis CDC1551, and M. smegmatis MC 2 155 strains procured from ATCC were used for stimulation of PBMC at an MOI of 1:4. The bacterial strains were grown to mid-logarithmic phase in 7H9 broth (Difco, Sparks, MD) supplemented with 10% oleic acid-albumin-dextrose-catalase (OADC, Difco), glycerol, 0.05% Tween 80 at 37°C on a shaker. At the time of stimulation, bacteria were pelleted and resuspended in complete RPMI and co-cultured with isolated PBMC in complete RPMI supplemented with IL-2 at 37°C for 24-48 hours.

Single-cycle HIV-1 Infection Assay
After stimulation, the cells were washed with fresh RPMI and CD4+ cells were isolated from stimulated PBMC by negative selection using a CD4+ T cell isolation kit (Miltenyi biotech; GmbH) according to the manufacturer's instructions. Isolated CD4+ cells were infected with a single-cycle replication-competent, green fluorescent protein (GFP)-expressing proviral construct pseudotyped with CCR5-tropic or CXCR4-tropic envelope (NL43-deltaEnvGFP), as previously described [27]. Briefly, stimulated CD4+ cells were spinoculated for 2 hrs with the virus. The infected cells were incubated at 37°C for 72 hours. Fluorescence emitted by the cells carrying GFP-expressing virus was quantified using Cellquest Pro software (Beckton Dickinson) to calculate percent infectivity.

Construction of mutant strains
M. smegmatis strains overexpressing individual lipoproteins were generated in the pGS400H plasmid backbone [28] using the Gateway system (Invitrogen, USA). This construct placed the coding sequence of the lipoprotein gene under the control of a constitutive Hsp60 promoter [29]. These constructs contain a hygromycin selectable marker and integration occurred at the attB site as a single copy. The gene-specific primers pairs used for cloning are listed in Table 1.
The M. bovis BCG strain lacking lspA was generated using a conditionally replicating mycobacteriophage, as previously described [30]. Briefly, Upper (886 bp) and lower (1125bp) allelic exchange substrates (AES) were amplified from CDC1551 and M. bovis BCG genomic DNA before ligation into pCR2.1-TOPO. Following sequencing, AES was subcloned into pYUB854. After transduction, BCG was plated onto Middlebrook 7H10 agar with ADC and hygromycin (50 μg/ml) and colonies screened for the absence of lspA.

Statistical Analysis
Statistical significance was determined by students T Test. A p-value of < .05 was considered statistically significant.