Mycobacterium tuberculosis-Infected Hematopoietic Stem and Progenitor Cells Unable to Express Inducible Nitric Oxide Synthase Propagate Tuberculosis in Mice

We show that bone marrow stem and progenitor cells containing Mycobacterium tuberculosis induce hallmarks of tuberculosis if killing of intracellular bacteria is defective. Understanding the relative contribution of these cells in tuberculosis latency and reactivation could inform novel host-directed therapies.

Tuberculosis represents a devastating health problem with 10.4 million new cases and 1.7 million deaths in 2016 globally [1]. One compounding factor in efforts to control the global spread of tuberculosis is the ability of the causative bacterium, Mycobacterium tuberculosis, to cause latent tuberculosis infection in human hosts where bacteria persist in the absence of clinical signs of tuberculosis, but in the presence of an M. tuberculosis-specific immune response [2]. Individuals with latent tuberculosis infection remain at risk of developing active tuberculosis during their lifetime [3].
The circumstances of bacterial persistence during latent tuberculosis infection remain largely enigmatic. Recently, M. tuberculosis has been detected within mesenchymal [4] and hematopoietic [5] stem cell compartments in mice and humans. Both hematopoietic stem cells (HSC) and mesenchymal stem cells (MSC) are multipotent; HSCs giving rise to lymphoid and myeloid cell lineages in the blood [6] and MSCs giving rise to adipocytes, chrondrocytes, and osteoblasts, amongst other cell types [7]. HSCs or MSCs recovered from humans, with either latent tuberculosis infection or having been successfully treated for pulmonary tuberculosis, contain M. tuberculosis in a predominantly uncultivatable form [4,5]. Evidence for the pathological context of carriage of M. tuberculosis within HSC or MSC compartments is currently lacking [8,9]. In the present work, we used a murine model of tuberculosis where mice lack inducible nitric oxide synthase 2 (NOS2), an enzyme critical for defense against intracellular M. tuberculosis, with the aim of addressing the pathological context of this niche [8,9].

Mice
All animal experiments were approved by the State Office of Health and Social Affairs Berlin (Landesamtes für Gesundheit und Soziales Berlin; approval number G0055/88). C57BL6 wildtype (wt) and C57BL6 Nos2 −/− mice (Charles River Laboratories) were bred in our facilities at the Max Planck Institute for Infection Biology, Berlin. For infection, 8-week old mice were anesthetized via intraperitoneal injection of ketamine (65 mg/ kg), acepromazine (2 mg/kg), and xylazine (11 mg/kg) and 10 4 M. tuberculosis organisms were injected into the ear dermis.

Harvest of Bone Marrow and FACS Sorting
Mice were sacrificed 4 weeks postinfection by cervical dislocation and femora and tibiae were removed from hind legs. Bone marrow cells were harvested by flushing femora and tibiae with PBS. Bone marrow cells were separated into lin + and lin − fractions using a lineage cell depletion kit for mice according to manufacturer's instructions (Miltenyi Biotec). Lin − Sca1 + cells were purified from lin − cells using an anti-Sca1 Microbead kit (fluorescein isothiocyanate) (Miltenyi Biotec). For subsequent identification of cell fractions, lin − cells were stained with the following antibodies: c-Kit (2B8), Sca1 (D7), CD150 (TC15-12F12.2) (eBioscience). Stained cells were sorted to >98% purity by fluorescence activated cell sorting (FACS) using an Aria II flow cytometer (BD Biosciences).

IS6110 PCR
For polymerase chain reaction (PCR) analysis, lin + and lin − , LSK CD150 + and LSK CD150 − cellular fractions were pelleted, resuspended in ultrapure water, and heat-treated at 80°C for 20 minutes. PCR was performed directly on heat-treated samples using the primers 5′-CGTGAGGGCATCGAGGTGGC-3′ and 5′-GCGTAGGCGTCGGTGACAAA-3′ to amplify a 245-base pair fragment located within the IS6110 insertion element present in the H37Rv genome.

Bone Marrow Transfers
Harvested whole bone marrow cells, purified lin + or Lin − Sca1 + cell preparations were pelleted and resuspended in PBS at 10 7 /mL, and 100 µL containing 10 6 cells of whole bone marrow cells, purified lin + cell preparations, or either 5 × 10 5 or 5 × 10 4 Lin − Sca1 + cell preparation, was transferred into untreated C57BL6 wt or C57BL6 Nos2 −/− via the tail vein. Mice receiving cell preparations were monitored for signs of weight loss, sacrificed at 8 weeks post-transfer, and spleen, lung, and bone marrow were harvested. Organs were homogenized in PBS-Tween 80 0.05% (vol/vol), plated at suitable dilutions on 7H11 agar plates supplemented with Middlebrook OADC Enrichment (Difco), and incubated at 37°C. Colonies growing on agar plates were enumerated after ≥3 weeks.

Histology
Mice were sacrificed 8 weeks after cell transfer and lung and spleen tissue were fixed in PBS containing 4% w/v paraformaldehyde overnight at room temperature. Sections of formalin-fixed, paraffin-embedded tissue, 2-3 μm thick, were deparaffinized and subjected to hematoxylin and eosin (H&E) staining.

RESULTS
We have previously described an experimental murine model of M. tuberculosis infection where Nos2 −/− mice are able to recapitulate hallmarks of human tuberculosis, including necrotizing granuloma pathology in the lung [10,11]. As previously reported using this model, where an infectious dose of 10 4 viable M. tuberculosis is injected into the dermis, we observed here systemic infection in Nos2 −/− mice with cultivatable M. tuberculosis in the spleen and lung at day 28 postinfection. Dermally infected wt mice showed cultivatable M. tuberculosis in the spleen at day 28 postinfection. We did not detect cultivatable M. tuberculosis from 5 × 10 7 total bone marrow cells at day 28 postinfection from either wt or Nos2 −/− mice ( Figure 1A).
To assay bone marrow cell populations from infected mice for presence of M. tuberculosis that could not be cultured, we sorted bone marrow cells into distinct compartments and probed DNA purified from these cells for the presence of IS6110, an insertion element present exclusively in M. tuberculosis complex strains, using PCR [12]. We separated cells based on expression of lineage markers to obtain a lin + cell population and a negatively selected population lacking these markers (lin − ), enriched for hematopoietic stem and progenitor cells (HSPCs). Using PCR, we identified IS6110 by PCR in lin − but not lin + cell populations of the bone marrow ( Figure 1B). To ascertain whether infections of lin − cell populations were particular to dermal infection with M. tuberculosis, we examined the bone marrow of wt and Nos2 −/− mice 28 days after aerosol challenge with M. tuberculosis. In this setting, both wt and Nos2 −/− mice harbored approximately 10 6 M. tuberculosis in the lung and 1-100 culturable M. tuberculosis colony forming units (CFUs) in the bone marrow. We identified M. tuberculosis in both lin − and lin + populations using IS6110 PCR, suggesting that presence of M. tuberculosis in lin − cells was not particular to dermal infection with M. tuberculosis.
We next investigated whether M. tuberculosis present in lin − cellular fractions could propagate infection when delivered to naive recipient mice. We transferred 10 6 bone marrow cells, containing approximately 1 × 10 4 HSPCs harboring 5-10 M. tuberculosis stainable with rhodamin-auramin, from dermal-infected wt and Nos2 −/− mice to naive recipient mice via the tail vein of recipient mice and monitored recipients for signs of infection. Finally, we evaluated the extent of pathology in Nos2 −/− recipients resulting from adoptive transfer of Nos2 −/− bone marrow cells from M. tuberculosis-infected donors. Organs were harvested 8 weeks after cell transfer and analyzed using hematoxylin and eosin staining. We observed demarcated granulomas ( Figure 2C and D), similar to those observed after infection of Nos2 −/− mice via the dermal route with viable M. tuberculosis [11]. Spleens contained significant regions of cellular necrosis ( Figure 2E) compared to uninfected controls ( Figure 2F), indicating active tuberculosis.
Taken together, the bacterial load and pathology reveal that adoptive transfer of Nos2 −/− bone marrow cells, with M. tuberculosis present predominantly in HSPCs, can infect Nos2 −/− naive mice leading to the hallmarks of tuberculosis.

Discussion
We show that M. tuberculosis-infected HSPCs in bone marrow are involved in propagating systemic hallmarks of the primary infection after adoptive cell transfer to naive mice, contingent on the inability of these cells to express NOS2. Intriguingly, although tuberculosis was propagated after adoptive transfer of 5 × 10 5 HSPCs as microbead-purified Lin − Sac1 + cells (Supplementary Figure 3A-C), which contain a mixed population of LT-HSCs, ST-HSCs, and MPPs, transfer of 5 × 10 4 infected LT-HSCs to Nos2 −/− mice did not result in tuberculosis (data not shown). We attribute this to the potential involvement of multiple bone marrow stem cell populations in triggering of tuberculosis from the bone marrow niche. Although it has previously been shown that the vast majority of M. tuberculosis in HSC is in a noncultivatable state [5] and although the experiments presented here have not detected cultivatable M. tuberculosis in bone marrow cells from the M. tuberculosis-infected Nos2 −/− mice used as donors, we would not rule out that low numbers of cultivatable M. tuberculosis in HSPC were the origin of cultivatable M. tuberculosis in the Nos2 −/− recipients. Further characterization of the phenotypic status of M. tuberculosis in bone marrow stem cells, beyond that of cultivability on supplemented agar, could support our model as a system by which resuscitation of M. tuberculosis and transition to tuberculosis from latent tuberculosis infection can be studied further.
NOS2 is not expressed in resting cells [13] and thus NOS2 is not active in HSCs. NOS2 is induced by inflammatory cytokines in MSCs and controls growth of intracellular M. tuberculosis in vitro [14] and, similarly, could also be induced in more mature hematopoietic progenitors to kill intracellular M. tuberculosis, as they develop from HSC. Thus, wild-type bone marrow would retain M. tuberculosis only in the resting LT-HSC, as we have seen previously in M. tuberculosis-infected wild-type bone marrow cells [5]. However, bone marrow of Nos2 −/− mice might harbor M. tuberculosis in more-differentiated progenitors, such as MPP, as we have shown ( Figure 1D). Moreover, tumor necrosis factor-alpha (TNF-α) augments NOS2 expression in MSCs in synergy with interferon-gamma (IFN-γ) in mice [15]. Because anti-TNF-α treatment is a known trigger of tuberculosis in humans, it is tempting to speculate that blocking of TNF-α may also abrogate the ability of later HSC progenitors and their lineage-positive cellular progeny, as well as MSCs, to control intracellular M. tuberculosis.
Our experiments support the essential role of NOS2 in control of M. tuberculosis in mice. This control might be exerted by the intracellular expression of NOS2 in hematopoietic cells, as well as through the interaction of the hematopoietic cells with NOS2-expressing endothelial and mesenchymal cells in hematopoietic niches. In contrast, the essential role of NOS2 for protection in human tuberculosis is more controversial and it is therefore tempting to speculate that additional antimycobacterial mechanisms could control M. tuberculosis reactivation in the early stages of progression from latent tuberculosis infection to active tuberculosis disease. Dissection and contextualization of the relative contribution of the individual cell types in this process could promote rational design of host-directed tuberculosis therapies in the future.