A Quantitative Method for the Study of HIV-1 and Mycobacterium tuberculosis Coinfection

Abstract Mycobacterium tuberculosis and human immunodeficiency virus-1 (HIV-1) syndemic interactions are a major global health concern. Despite the clinical significance of coinfection, our understanding of the cellular pathophysiology and the therapeutic pharmacodynamic impact of coinfection is limited. Here, we use single-round infectious HIV-1 pseudotyped viral particles expressing green fluorescent protein alongside M. tuberculosis expressing mCherry to study pathogenesis and treatment. We report that HIV-1 infection inhibited intracellular replication of M. tuberculosis and demonstrate the therapeutic activity of antiviral treatment (efavirenz) and antimicrobial treatment (rifampicin). The described method could be applied for detailed mechanistic studies to inform the development of novel treatment strategies.

Mycobacterium tuberculosis and human immunodeficiency virus-1 (HIV-1) syndemic interactions are a major global health concern. Despite the clinical significance of coinfection, our understanding of the cellular pathophysiology and the therapeutic pharmacodynamic impact of coinfection is limited. Here, we use single-round infectious HIV-1 pseudotyped viral particles expressing green fluorescent protein alongside M. tuberculosis expressing mCherry to study pathogenesis and treatment. We report that HIV-1 infection inhibited intracellular replication of M. tuberculosis and demonstrate the therapeutic activity of antiviral treatment (efavirenz) and antimicrobial treatment (rifampicin). The described method could be applied for detailed mechanistic studies to inform the development of novel treatment strategies.
Mycobacterium tuberculosis is an obligate, acid-fast, intracellular bacillus, causing the human disease tuberculosis. Human immunodeficiency virus (HIV-1; HIV) is a lentivirus that causes acquired immune deficiency syndrome (AIDS) in humans. Both diseases are among the leading causes of death worldwide [1].
People with HIV are up to 27-times more likely to develop active tuberculosis than people without HIV and, in 2020, of the 1.3 million deaths attributed to tuberculosis, 214 000 were among people with HIV [2]. Infection with either pathogen can reactivate latent infection of the other, and antimicrobials targeting M. tuberculosis are known to interact with antivirals targeting HIV [3].
Despite the clinical significance of HIV and M. tuberculosis coinfection, our understanding of the cellular pathophysiology and therapeutic pharmacodynamics is limited. This is in part due to the absence of quantitative methods required for study of coinfection. The development of M. tuberculosis and HIV that carry fluorescent molecules, potentially allows for the simultaneous quantification of both pathogens using fluorescence-based modalities suitable for single-cell study and drug discovery [4,5].
We have previously used single-round infectious pseudotyped viral particles (PVPs) to monitor antibody neutralization of Ebola virus and for the study the glycolipid composition of pathogenic and nonpathogenic M. tuberculosis [6,7]. Here, we have applied a similar approach to study M. tuberculosis and HIV coinfection in THP-1 monocyte-derived macrophages. We utilized fluorometry, high-content imaging, and flow cytometry to assess the impact of antiviral and antimicrobial treatment on HIV and M. tuberculosis coinfection.

HIV and M. tuberculosis Coinfection and Treatment
PVPs were added to macrophages at a concentration 30 ng/well in FluoroBrite DMEM, supplemented with 10% HI-FBS and L-glutamine (200 µL/well), as appropriate. PVPs lacking the envelope protein (ΔEnv), making them unable to enter target cells, were included at 30 ng/well as a negative control. Concentrations were selected based on our previously published work [7]. After 24 hours, cell media were removed, and the macrophages infected with M. tuberculosis at a multiplicity of infection (MOI) of 1:5. The MOI was selected based on our previously published work [8]. After 24 hours, the cells were washed to remove extracellular baccili. Efavirenz in FluoroBrite DMEM at 10 µg/mL was added to THP-1 cells 2 hours prior to infection with HIV and remained in the culture media for the duration of the experiment, as appropriate. Rifampicin in FluoroBrite DMEM at 10 µg/mL was added 24 hours after the addition of M. tuberculosis, as appropriate.

Confocal Laser Scanning Microscopy Live Imaging
Plates containing infected cells were sealed with Breath-EASIER sealing membranes (Sigma) and wrapped in parafilm. Plates were then wiped with 5% surfanios, then 70% ethanol and transferred to the confocal microscope (LSM 880; Zeiss). Z-stacks were obtained at 60× magnification.

Fluorometry
Infected plates were sealed with Breath-EASIER sealing membranes (Sigma) and wrapped in parafilm. Plates were then wiped with 5% surfanios, then 70% ethanol and transferred to the plate reader (Varioskan LUX). Fluorescence was measured at 395/409 nm (excitation/emission) for green fluorescent protein and 587/610 nm for mCherry.

Flow Cytometry
THP-1 cells were washed twice with prewarmed phosphatebuffered saline (PBS) and then detached from the plate using icecold PBS and transferred to screw caped tubes. Cells were stained for viability (Pacific Blue; LifeTechnologies), washed, and then incubated in 5% paraformaldehyde at room temperature for 2 hours. Cells were washed once, resuspended in ice-cold PBS, and stored in the absence of light until acquisition using a FACS LSR II flow cytometer (BD Biosciences). Standard procedures were used to maintain the instrument and quality control performed daily. A compensation matrix was created, and sequential cell isolation used to identify populations using FlowJo version 10 (Treestar Inc).

Coinfection of THP-1 Macrophages by HIV and M. tuberculosis
We first confirmed by confocal microscopy that intracellular coinfection had been established ( Figure 1A). For cultures infected with HIV-VSV-G-PVP and M. tuberculosis, as well as for cultures infected with HIV-BAL-PVP and M. tuberculosis, coinfection was observed amongst THP-1 macrophages ( Figure 1A). Visual inspection revealed that HIV-VSV-G-PVP had a higher infectivity rate than the HIV-BAL-PVP ( Figure 1A).
Fluorometry was utilized to study the impact of HIV infection on M. tuberculosis growth kinetics. Over the course of 96 hours, in the absence of HIV-PVP, intracellular replication of M. tuberculosis was observed ( Figure 1B). In cultures where HIV-VSV-G or HIV-BAL-PVP infection had previously been established, there was no detectable growth of M. tuberculosis ( Figure 1B).

Single-Cell Assessment of M. tuberculosis Growth Kinetics in Macrophages Infected With HIV
To improve assay resolution, we performed single-cell analysis by flow cytometry. A sequential gating strategy was used to identify THP-1 cells infected with virus and/or bacilli (Figure 2A). In cultures singularly infected with M. tuberculosis, approximately 31% of viable cells were infected with M. tuberculosis and, in cultures where HIV infection had been established prior to the addition of M. tuberculosis, approximately 15% of viable cells were infected with M. tuberculosis (P = .0378; Figure 2B). The addition of M. tuberculosis had no impact on the frequency of HIV infection ( Figure 2B).

Single-Cell Assessment of Drug Activity in Coinfection
We next set out to demonstrate the utility of this method for the assessment of antiviral and/or antimicrobial treatment. Incubation with efavirenz was shown to block HIV-PVP infection ( Figure 2C); approximately 5% of viable cells were singularly infected with HIV in untreated control cultures and no cells were infected in cultures treated with efavirenz (P = .0015); approximately 0.4% of viable cells were simultaneously infected with HIV and M. tuberculosis in untreated control cultures and no cells were identified in cultures treated with efavirenz (P = .0083; Figure 2C). Efavirenz had no impact on the frequency of cells singularly infected with M. tuberculosis ( Figure 2C).
Incubation with rifampicin was shown to have a significant impact on the intracellular burden of M. tuberculosis ( Figure 2C); approximately 14% of viable cells were singularly infected with M. tuberculosis in untreated control cultures and approximately 8% were infected in cultures treated with rifampicin (P = .0236); approximately 0.4% of viable cells were simultaneously infected with HIV and M. tuberculosis in untreated control cultures and approximately 0.07% were identified in cultures treated with rifampicin (P = .0123; Figure 2C). Rifampicin was also shown to reduce the frequency of cells singularly infected with HIV; approximately 5% of viable cells were infected in untreated control cultures and approximately 0.7% of viable cells were infected in cultures treated with rifampicin (P = .0031; Figure 2C).

DISCUSSION
Here, we describe a quantitative platform suitable for the study HIV and M. tuberculosis coinfection. The described method may be applied to study disease pathogenesis and could serve  to inform the development of novel treatment strategies targeting either pathogen alone, or both pathogens simultaneously.
Interestingly, and consistent with published literature [9], prior exposure to one pathogen was observed to influence the response to another subsequently encountered pathogen. Specifically, infection of macrophages with either HIV-VSV-G-PVP or HIV-BAL-PVP, prior to M. tuberculosis infection, resulted in sustained inhibition of intracellular growth of M. tuberculosis.
Pathogens can modulate cellular metabolism through direct interaction with Toll-like receptors via pathogen-associated molecular patterns or through indirect activation of innate immune signaling pathways. Because it is well-established that HIV infection drives M1 polarization in vitro [10], it is reasonable to hypothesize that proinflammatory polarization may have impacted the capacity for M. tuberculosis to replicate intracellularly. Interestingly, because this effect was observed for both HIV-VSV-G-PVP and HIV-BAL-PVP, it can be assumed that the mechanism is independent of the viral envelope. Given that M. tuberculosis is also known to differently influence macrophage polarization [11], the order in which cells encounter either HIV or M. tuberculosis could very well influence the dynamics of disease progression and further study is warranted.
It would be an oversimplification to attempt to infer any link between the observations made here and clinical disease progression. The model presented here is unlikely to provide insight into long-term disease progression-it does, however, represent a unique platform to study the earliest stages of coinfection where HIV is already established, and where M. tuberculosis is subsequently encountered. An appreciation of recent evidence concerning macrophage heterogeneity, in respect to both development and metabolism, indicates that the true pathophysiological representation of polarization is far more complex [12]. Whilst simplistic definitions of M1 and M2 macrophage polarization are useful, data presented elsewhere have demonstrated that, in reality, multiple distinct macrophage activation states exist across as spectrum [12].
Recent evidence suggests that the developmental origin of macrophages can determine their responses to infection stresses. In support, metabolic and transcriptional studies have demonstrated that embryonically derived alveolar macrophages and hematopoietically derived interstitial macrophages respond differently to M. tuberculosis [13]. Despite this, the significance of the observations made here should not be overlooked-reduced intracellular replication of M. tuberculosis is likely to prolong the time taken to sterilize the intracellular space through antimicrobial treatment and may increase risk of treatment failure.
We have demonstrated the potential utility of this platform for assessing the activity of antivirals and antimicrobials targeting HIV and M. tuberculosis. Efavirenz is known to block the integration of viral genomic RNA and was shown to completely block HIV infection. Similarly, rifampicin resulted in a reduction in the frequency of macrophages infected with M. tuberculosis. Interestingly, the addition of rifampicin was also shown to impact the frequency of cells identified as being singularly infected with HIV. Rifampicin is known to inhibit the DNA-dependent RNA polymerase of bacteria and viruses and has previously been shown to inhibit viral assembly of DNA viruses and the reverse transcriptase, RNA-directed DNA polymerase, of the Rous sarcoma virus [14,15].
It was striking that off-target effects were observed following treatment with single compounds. Given that tuberculosis and HIV treatment requires administration of multiple compounds simultaneously, we would encourage further work to understand the cellular pathophysiological impact of typical combination therapeutic strategies in the context of coinfection. Our platform affords an opportunity to assess novel combination strategies that seek to investigate potential synergistic interactions between antiviral and antimicrobial as well as host-directed immunomodulatory therapeutics. Further value could be added through inclusion of concentration-responses analysis and/or through use of primary host cells, including monocyte-derived macrophages and primary macrophages isolated from healthy volunteers and people with HIV. It is possible that macrophages with distinct developmental origins could show distinct phenotypes with respect to M. tuberculosis infectivity and intracellular growth.