Intravital Imaging of Attenuated Auxotrophic Mycobacterium tuberculosis in Murine Lung during Early Intravascular Infection

Summary

Here, we present a protocol for intravital imaging of an attenuated BSL-2-approved Mycobacterium tuberculosis. strain in the murine lung using a permanent lung window. This approach enables longitudinal, single-cell-resolution analysis of early host-pathogen interactions, including bacillary localization, vascular dynamics, and macrophage uptake within the same lung microenvironment.

Abstract

Intravital microscopy enables direct visualization of dynamic cellular processes within intact tissues. However, its application to Mycobacterium tuberculosis (Mtb) has been limited by Biosafety Level 3 (BSL-3) containment requirements and the technical challenges of stabilizing the lung for high-resolution imaging. Here, we present a protocol that combines the thoracic Window for High-Resolution Imaging of the murine Lung (WHRIL) with mc²7902, a genetically defined triple-auxotrophic Mtb strain approved for use under BSL-2 conditions. We also describe the generation of mc²8471, a tdTomato-expressing derivative of mc²7902, its preparation for intravenous infection, and application for intravital imaging in reporter mice. This system enables real-time visualization of early infection dynamics, including rapid bacillary entry into the pulmonary vasculature, aggregation, dissemination into the lung parenchyma, and macrophage uptake in the same lung microenvironment over the first 3 days post-infection. By enabling longitudinal imaging of bacilli and host–pathogen interactions at single-cell resolution, this approach overcomes key limitations associated with conventional imaging and high-containment models. This protocol provides a practical BSL-2-compatible platform for real-time intravital imaging of attenuated mycobacteria in the lung and establishes a foundation for mechanistic studies of bacterial physiology, host recognition, and immune-mediated clearance.

Introduction

High-resolution intravital microscopy has transformed the study of dynamic cellular behavior within intact tissues by enabling direct visualization of cell migration, vascular interactions, and immune surveillance at single-cell resolution^1,2,3,4. While these approaches are widely used in cancer research^3,5, their application to pulmonary infections has been limited by respiratory motion, tissue fragility, and biosafety constraints. Recent advances in thoracic imaging platforms, including vacuum-stabilized and permanently implanted optical windows, have addressed many of these technical challenges by providing stable, motion-free imaging of the pulmonary microenvironment over extended periods^6,7,8,9,10. Permanently implantable windows, in particular, are powerful tools as they enable repeated visualization of vascular flow, immune cell dynamics, and single-cell behavior over hours to weeks, and have been instrumental in defining mechanisms of cancer metastasis, vascular permeability, and immune surveillance^11,12.

Despite these advances, intravital imaging has not previously been applied to Mycobacterium tuberculosis (Mtb) infection in the in vivo mammalian lung at single-cell resolution. The closest precedents come from intravital imaging of mycobacterial infections in other settings. Real-time imaging of Mycobacterium marinum infection in zebrafish embryos¹³ revealed early macrophage interactions and granuloma initiation, while two-photon imaging of BCG-induced granulomas in the mouse liver^14,15 revealed macrophage and T cell dynamics during granuloma development and demonstrated limited antigen presentation and T cell effector function within granulomatous lesions. These studies established foundational insights into mycobacterial host-pathogen interactions but did not access the intact mammalian lung^8,16.

Intravital imaging approaches have also been applied to the lung in other infectious and biological contexts, including general pulmonary imaging⁸, studies revealing that patrolling alveolar macrophages can conceal bacteria from the immune system¹⁷ and characterizing the immediate myeloid response to SARS-CoV-2 infection in the human lung¹⁶. Intravital imaging has similarly been applied to influenza-infected lungs, enabling direct visualization of dynamic immune-cell behavior within infected lung tissue^18,19.

Whole-body optical imaging approaches using fiber-optic microendoscopic excitation²⁰ and reporter enzyme fluorescence technology^21,22 have enabled sensitive detection of Mtb in mouse lungs, but cannot detect individual or small clusters of bacilli, track bacterial aggregation dynamics, or visualize direct interactions between mycobacteria and host cells within intact lung tissue. High-containment intravital imaging approaches have been implemented for infected lungs in BSL-3 settings, including SARS-CoV-2 infection²³ and Mtb-focused²⁴ multiphoton imaging platforms. These studies demonstrate the feasibility of imaging infected lungs under containment, but such setups remain technically demanding and are not widely accessible. Work with virulent Mtb requires BSL-3 containment, posing significant engineering and biosafety challenges for intravital imaging^23,24. As a result, most studies of early Mtb infection rely on static endpoint assays that cannot capture the spatial and temporal dynamics of bacterial behavior within the lung.

The development of genetically defined, BSL-2-approved triple-auxotrophic Mtb strains provides a unique opportunity to overcome these limitations. The mc²7902 strain, a pantothenate-leucine-arginine Mtb auxotroph derivative of H37Rv, is fully attenuated, fails to replicate in immunocompromised mice, retains classical acid-fast staining, exhibits normal phage susceptibility, and can be safely handled outside of BSL-3 facilities²⁵. These properties make mc²7902 an ideal surrogate for establishing intravital imaging approaches and probing early host-pathogen interactions under reduced biosafety containment.

Here, we apply the Window for High-Resolution Imaging of the Lung (WHRIL) with mc²8471, a fluorescent tdTomato-expressing derivative of the triple-auxotrophic Mtb strain mc²7902, to visualize early Mtb infection in the intact murine lung. Experiments were performed in immunocompromised Rag2^-/- mice²⁶ crossed with MacBlue mice²⁷, in which the Csf1r promoter drives expression of enhanced cyan fluorescent protein (ECFP) in monocytes, microglia, and subsets of dendritic cells and macrophages. The MacBlue transgene enables clear visualization of mononuclear phagocyte populations during early Mtb infection. We chose to utilize Rag2-deficient mice to focus the system on the earliest innate host-pathogen interactions and to avoid the added complexity of adaptive immune responses during the short imaging window examined here. Importantly, this technique is not restricted to Rag2-deficient mice and can also be performed in immunocompetent animals.

By combining a permanent lung window with a BSL-2-compatible fluorescent attenuated Mtb strain, this platform enables serial intravital imaging of bacilli-host cell interactions in the same lung microenvironment at single-cell resolution over multiple days. Although demonstrated here for early infection, the platform itself is not inherently restricted to early time points and could, in principle, be applied to later stages depending on the experimental model and study design. Since the successful application of this imaging platform depends on the use of the specific fluorescent bacillary preparation described here, we included the essential steps required to generate and prepare that reagent as part of the complete protocol. This platform provides direct, real-time imaging of Mtb behavior in vivo and establishes a foundation for future mechanistic studies of bacterial physiology, host recognition, and immune-mediated clearance.

Protocol

All procedures were conducted in accordance with institutional biosafety and animal care guidelines and were approved by the Albert Einstein College of Medicine Animal Care and Use Committee (IACUC). The strain mc²8471 is a pYUB1169 transformant of mc²7902, a triple-auxotrophic Mtb strain approved for BSL-2 use. All work involving live bacteria was performed using standard precautions for handling mycobacteria.

1. Construction of the BSL-2-approved tdTomato-expressing Mtb strain mc²8471

1. Generation of tdTomato-expressing BSL-2-safe mc²8471 by plasmid transformation
   1. Transform the BSL-2-approved triple-auxotrophic strain mc²7902 with the episomal mycobacterial plasmid pYUB1169, which contains the tdTomato gene expressed from the G13 promoter, to generate the fluorescent derivative mc²8471.
   2. Inoculate 100 µL of mc²7902 (optical density at 600 nm (OD600 nm) of 0.7-1.0) in 10 mL of sterile Middlebrook 7H9 supplemented with OADC enrichment (10% (v/v)), glycerol (0.2% (v/v)), D-pantothenate (24 mg/L), L-leucine (50 mg/L), L-arginine (200 mg/L), sodium propionate (0.1 mM), and tyloxapol (0.05% (v/v)) in a 30 mL inkwell bottle.
   3. Incubate the culture, slowly shaking (100 rpm) at 37 °C until it has reached an OD600 nm of 0.7-1.0.
   4. Transfer the cells into a 15 mL conical centrifuge tube and centrifuge for 10 min at 3,000 x g at room temperature. Discard the supernatant and resuspend the cells in 10 mL of sterile 10% glycerol containing 0.05% (v/v) tyloxapol (wash #1). Repeat centrifugation and washes 2x. Resuspend the cell pellet in 0.175 mL of 10% glycerol containing 0.05% (v/v) tyloxapol and add 2 µL of pYUB1169 DNA (100-200 ng of DNA).
   5. Transfer mc²7902 and pYUB1169 into a 0.2 cm electroporation cuvette. Wait 10 min. Tap the cuvette 2x to 3x to remove air bubbles and electroporate at 2,500 mV, 1,000 Ω, 25 µF, time constant 20 ms, at room temperature.
   6. Add 1 mL of sterile Middlebrook 7H9 supplemented with OADC enrichment (10% (v/v)), glycerol (0.2% (v/v)), D-pantothenate (24 mg/L), L-leucine (50 mg/L), L-arginine (200 mg/L), sodium propionate (0.1 mM), and tyloxapol (0.05% (v/v)) to the cuvette and transfer the contents of the cuvette to a 15 mL conical centrifuge tube. Incubate the tube while shaking at 37 °C overnight.
   7. Plate 0.1 mL of the transformation mixture onto a Middlebrook 7H10 agar plate supplemented with OADC enrichment (10% (v/v)), glycerol (0.2% (v/v)), D‑pantothenate (24 mg/L), L‑leucine (50 mg/L), L‑arginine (200 mg/L), sodium propionate (0.1 mM), and kanamycin (20 mg/L). Centrifuge the remaining transformation mixture, discard 0.9 mL of the supernatant, and resuspend the cell pellet in the residual supernatant. Plate the resuspended cells onto a second Middlebrook 7H10 agar plate containing the same supplements and antibiotics. Incubate the plates at 37 °C for 3–4 weeks. Pink colonies indicate successful transformants.
   8. Pick one isolated pink colony and grow it at 37 °C, slowly shaking, in 5 mL of sterile Middlebrook 7H9 supplemented with OADC enrichment (10% (v/v)), glycerol (0.2% (v/v)), D-pantothenate (24 mg/L), L-leucine (50 mg/L), L-arginine (200 mg/L), sodium propionate (0.1 mM), tyloxapol (0.05% (v/v)) and 20 mg/L kanamycin in a 30 mL inkwell bottle.
2. Preparation of mc²8471 for intravenous infection
   1. Grow 10 mL of mc²8471 at 37 °C, slowly shaking, in sterile Middlebrook 7H9 supplemented with OADC enrichment (10% (v/v)), glycerol (0.2% (v/v)), D-pantothenate (24 mg/L), L-leucine (50 mg/L), L-arginine (200 mg/L), sodium propionate (0.1 mM), tyloxapol (0.05% (v/v)), and 20 mg/L kanamycin in a 30 mL inkwell bottle to mid-log phase (OD600 nm should be between 0.7 and 1.0).
   2. Transfer the cells into a 15 mL conical centrifuge tube and centrifuge 10 min at 3,000 x g at room temperature. Discard the supernatant and resuspend the cells in 10 mL of sterile phosphate-buffered saline containing 0.05% (v/v) tyloxapol (PBS-T, wash #1). Repeat centrifugation and washes 2x. Resuspend the cell pellet in 4 mL of PBS-T and sonicate the cells at power 100 W, 2x 1,000 J.
   3. Measure the optical density of the culture and dilute it in PBS-T for a final concentration of 5 x 10⁶ CFUs per injection, using the approximate conversion that an OD600 nm of 1 corresponds to 1 x 10⁸ CFU/mL. Because this relationship can vary with strain and growth conditions, OD600 nm should be calibrated empirically against plated CFU values to ensure accuracy and reproducibility.

2. Permanent lung window implantation in mice for intravital imaging

1. Prepare a Rag2KO MacBlue mouse for intravital imaging. In this study, a 10-week-old male mouse weighing 30 g was used. However, mice of either sex, 8-12 weeks of age, and within the expected weight range for that age are suitable for this procedure. This strain expresses enhanced cyan fluorescent protein (ECFP) in macrophages on a Rag2-knockout background and was generated by crossing B6.129S6-Rag2^tm1FwaN12 mice (Rag2, Model RAGN12) with Tg (Csf1r-GAL4/VP16,UAS-ECFP)^1Hume/J mice (Stock No. 026051).
2. Implant the Window for High-Resolution Imaging of the Lung (WHRIL) 24 h before bacterial infection to provide stable optical access to the lung and allow recovery before imaging.
3. Induce anesthesia with 5% isoflurane in an induction chamber. Confirm adequate depth of anesthesia before beginning any procedure and monitor the animal throughout surgery and imaging to prevent pain, distress, or respiratory compromise.
4. Administer pre-operative analgesia: inject 10 µL (0.1 mg/kg) of buprenorphine diluted in 90 µL of sterile PBS subcutaneously.
5. Apply ophthalmic ointment to both eyes to prevent corneal drying during anesthesia.
6. Tie a 2-0 silk suture around the base of a 22G catheter, leaving ~2 inch tails for securing the catheter.
7. Intubate the mouse with the suture-tied catheter. Visualize the tracheal opening by transillumination and confirm placement using an inflation bulb (bilateral chest rise).
   NOTE: Improper intubation can cause airway trauma or inadequate ventilation. Proceed only after correct placement is confirmed by bilateral chest expansion.
8. Secure the intubation catheter by tying the 2-0 silk suture around the snout after successful intubation. Connect the catheter to the ventilator.
9. Place the mouse in right lateral decubitus on a heated surgical platform to expose the left thorax, then secure the forelimbs, hindlimbs, and back with tape to maintain stable positioning throughout the procedure.
10. Reduce isoflurane to 3% for the remainder of the surgery. Remove hair from the left upper thorax using depilatory cream.
11. Prepare the surgical field with an antiseptic solution. Lift the skin and make a 5-10 mm circular incision with a biopsy punch (33-35-SH) approximately 7 mm left of the sternum and 7 mm above the subcostal margin, centered over the 6^th to 7^th intercostal space.
12. Inspect the exposed field for superficial vessels. If division of a visible vessel is necessary, cauterize both ends to maintain hemostasis. Remove the overlying soft tissue to expose the ribs clearly before entering the thoracic cavity.
13. Grasp and elevate the 6th or 7th rib with forceps and gently pierce the intercostal muscle between them using blunt micro-dissecting scissors (rounded side facing the lung) to enter the thoracic cavity.
    NOTE: Enter the thoracic cavity gently and under direct visualization to avoid lacerating the lung, intercostal vessels, or surrounding tissues.
14. Apply short, gentle bursts of compressed air at the intercostal opening to transiently deflate the adjacent lung surface and separate it from the chest wall before creating the window opening.
15. Position the biopsy punch on the cutting tool and guide the cutting tool through the intercostal opening, keeping it parallel to the chest wall. Punch a 5 mm circular defect in the rib cage.
16. Place a 5-0 silk purse-string suture ~1 mm from the edge of the circular defect, passing between adjacent ribs to encircle the opening.
17. Seat the window frame so the edges of the chest wall defect fit into the groove of the frame, then tighten and secure the purse-string suture to lock the frame in place.
18. Load cyanoacrylate gel adhesive into a 1 mL syringe. Gently lift the frame to separate it from the lung, then apply a thin layer of cyanoacrylate adhesive to the undersurface of the frame.
19. Dry the surface of the lung using a continuous, gentle stream of compressed air until the surface appears matte.
    NOTE: Use only a gentle stream of air and avoid prolonged drying, as excessive force or over-drying can injure the lung surface and impair window adhesion.
20. Using positive end expiratory pressure (PEEP), inflate the lung and then adhere the window to the frame. Release the PEEP and keep pressure on the window frame for 20-30 s.
    NOTE: Excess pressure during inflation or frame placement can damage the lung or compromise the seal. Apply only the minimum pressure needed to achieve gentle apposition.
21. Place a second 5-0 silk purse-string suture <1 mm from the skin incision edge. Tuck the skin beneath the frame, ensure that any excess skin is captured under the window frame or trimmed away, then cinch the purse-string suture tightly and tie securely.
22. Using a vacuum pickup, apply a very thin layer of adhesive to the underside of the 5 mm coverslip, scraping off excess on the edge of a rectangular cover glass. Again, dry the lung surface with a continuous, gentle stream of compressed air.
23. Position the circular coverslip within the recess of the window frame at a shallow angle (~15°) above the lung surface. Briefly increase airway pressure to inflate the lung, then rotate the coverslip into a parallel position so that the lung surface opposes the underside of the glass. Maintain gentle pressure for approximately 25 s while the adhesive sets, then detach the vacuum pickup from the coverslip.
    NOTE: Avoid trapping bubbles or applying excess adhesive during coverslip placement, as either can reduce image quality, disrupt the seal, or injure the lung surface.
24. Apply a small amount of liquid cyanoacrylate at the interface between the coverslip and the metal frame to reinforce the airtight seal.
25. Attach a sterile needle to a 1 mL insulin syringe. Insert the needle just below the xiphoid process and advance toward the left shoulder through the diaphragm into the thoracic cavity. Gently aspirate to remove residual intrathoracic air.
    NOTE: Advance the needle carefully to avoid puncturing the lung, heart, or major vessels. Aspirate gently and only until residual air is removed.
26. Turn off isoflurane and continue ventilation with 100% oxygen until a near-normal waking respiratory rate is nearly resumed. Do not leave the animal unattended during recovery. If normal spontaneous respiration does not resume promptly, provide appropriate supportive care or euthanize according to institutional guidelines.
27. Remove the securing suture, extubate the mouse, and place it in a clean, warmed cage for recovery. Monitor the animal until fully ambulatory; mice that fail to regain normal respiration should be euthanized according to institutional guidelines.
28. Administer post-operative analgesia: inject 10 µL (0.1 mg/kg) of buprenorphine diluted in 90 µL of sterile PBS subcutaneously.

3. Intravital imaging

1. Allow mice to recover after window implantation before imaging. Perform imaging 1 to 5 days after surgery.
2. Induce anesthesia with 3% isoflurane, then maintain at 1.5% for the duration of imaging.
3. Insert the custom stage plate-window frame adapter between the skin and the outer rim of the lung window frame and secure a standard microscope stage plate using adhesive tape.
4. Place the mouse with the custom stage plate-window frame adapter attached on the microscope translation stage. Maintain body temperature at ~37 °C using a forced air heated chamber with temperature sensor feedback loop.
5. With the mouse on the stage plate, gently pull back the skin around the eye on the injection side to expose the medial canthus and inject FITC-dextran (150 kDa; 20 mg/mL, 100 µL) retro-orbitally to label the vasculature.
6. Acquire snapshot, z-stack, or time-lapse images of the lung vasculature prior to bacterial injection using the following parameters.
   1. For confocal microscopy settings: Bit depth: 12-bit, Frame rate: 2 fps, Image size: 512 × 512 pixels, Pixel size: 0.497 µm/pixel, 0.994 µm/pixel, Z-step size: 2 µm, Total z-depth: 50–70 µm, Frame averaging: 3, Scan speed: 2 µs/pixel (977 lines/s), Laser excitation wavelengths: 405 nm, 488 nm, 559 nm, Laser power at specimen: 230 µW (405 nm), 75 µW (488 nm), 40 µW (559 nm), Detection filters: 430-470 nm (Blue), 505-540 nm (Green), 575-675 nm (Red), Detector gains: 800 V for Blue, 650 V for Green, 700 V for Red, Pinhole diameter: 60 µm or 1 Airy unit for 559 nm
   2. For multiphoton microscopy, use the same imaging parameters as those used for confocal microscopy, except that a 920 nm excitation laser was used with 25 mW laser power at the sample.
7. Retro-orbitally inject 5 x 10⁶ CFUs tdTomato-expressing mc²8471.
   NOTE: Perform retro-orbital injections only with appropriate training and restraint, as improper technique can cause ocular injury, bleeding, or failed delivery of the inoculum.
8. Acquire snapshot, z-stack, and time-lapse images of tuberculosis and fluorescently labeled cells and vasculature using the imaging parameters in 3.6.1 and 3.6.2. Adjust the gain of all detectors to maximize signal while avoiding detector saturation.
9. After imaging, turn off isoflurane and remove the mouse from the microscope. Keep the mouse warm and monitor until full recovery.
10. Repeat imaging at the following time points after mc²8471 injection: 0-60 min, 6 h, and 3 days. For each imaging session, re-inject FITC-dextran (100 µL, retro-orbitally) to visualize the vasculature, then acquire images as described in steps 3.8-3.9.

Results

Following electroporation and kanamycin selection, isolated colonies of mc²8471 exhibited bright red fluorescence, confirming stable tdTomato expression. mc²8471 bacilli were expanded, prepared as a single-cell suspension, and administered intravenously to mice bearing surgically implanted WHRIL. Intravital imaging was performed to visualize early Mtb infection dynamics in the lung microenvironment. For this study, baseline images of the lung vasculature were acquired before infection (Figure 1A), followed by intravital images within 60 min, at 6 h, and at 3 days after mc²8471 injection into a Rag2^-/- × MacBlue mouse (Figure 1B-D, Figure 2A-B, Figure 3A-C, Supplementary Video 1, Supplementary Video 2, Supplementary Video 3). These time points were selected to capture the earliest stages of bacillary localization, redistribution, and macrophage association in the lung. Within a few minutes following intravenous infection with mc²8471, bacilli could be observed within blood vessels, illustrating how rapidly the inoculum reaches and localizes within the lung vasculature (Figure 1B-D, Supplementary Figure 1A-C). At 1 h post-infection, bacillary fluorescent signal remained stationary and associated with the pulmonary vasculature, appearing either as discrete puncta or as larger clustered structures (Figure 2A, Supplementary Figure 1A-C).

By 6 h post-Mtb infection, bacillary localization had changed markedly, with bacilli found widely disseminated throughout the lung parenchyma in regions where no macrophages or blood vessels could be visualized (Figure 2B, Supplementary Figure 2A-C), a pattern indicative of loss of local perfusion.

By 3 days post-infection (Figure 3A-C, Figure 4A-C), bacilli were markedly reduced compared to the 6 h time point, indicating substantial bacterial clearance from the lung during this period. When detected with high-resolution images (Figure 3A-C, Figure 4A-C, Supplementary Video 3, and Supplementary Video 4), bacteria were found predominantly within macrophages, with multiple bacilli per cell (multiplicity of infection greater than one). This shift from extracellular aggregates at 6 h to predominantly intracellular bacteria by day 3 demonstrates successful phagocytic uptake and suggests that bacillary replication cannot outpace immune clearance at this early stage.

Unless otherwise noted, all imaging data presented in this manuscript were acquired using confocal microscopy. To demonstrate compatibility with multiphoton imaging, complementary datasets were collected using two-photon microscopy under conditions comparable to the confocal parameters described in section 3 (Figure 4A-C; Supplementary Video 4). In addition, two-photon 2D time-lapse imaging was used to visualize dynamic macrophage movements (Supplementary Video 5).

Together, these observations demonstrate that the WHRIL platform enables direct visualization of sequential early events in Mtb infection, including vascular delivery, aggregation, parenchymal dissemination, and macrophage uptake, at single-cell resolution in vivo.

Figure 1: Early intravascular appearance of mc²8471 in the lung immediately after intravenous injection. All panels show representative snapshot images. (A) Representative snapshot of the lung vasculature before bacterial injection. (B) Representative snapshot image acquired immediately after intravenous injection, showing the appearance of tdTomato-expressing mc²8471 within the pulmonary vasculature; white arrows indicate representative bacilli (magenta). (C) Representative image showing mc²8471 within lung blood vessels. (D) Magnified view of the boxed region from (C). Orange = FITC-labeled 150 kDa Dextran marking the vasculature. Magenta = tdTomato-expressing mc²8471. Sky Blue = CFP-labeled macrophages. Scale bars = 50 µm. Please click here to view a larger version of this figure.

Figure 2: Time-dependent changes in mc²8471 localization in the lung after intravenous infection. All panels show representative snapshot images. (A) Representative image obtained within 60 min of injection showing early stationary vascular-associated localization of mc²8471 near pulmonary blood vessels. (B) Representative snapshot image at 6 h post-injection showing clustered mc²8471 signal in a region with loss of visible dextran, indicative of loss of perfusion. White arrows denote representative regions of pulmonary occlusion. (C) Representative snapshot at 3 days post-injection showing a marked reduction in bacillary burden. Orange = FITC-labeled 150 kDa Dextran. Magenta = tdTomato-expressing mc²8471 Mtb. Sky Blue = CFP-labeled macrophages. Scale bars = 100 µm. Please click here to view a larger version of this figure.

Figure 3: Association of mc²8471 with macrophages at 3 days post-injection. (A–C) Representative higher-magnification snapshot images showing remaining bacilli in close association with CFP-labeled macrophages at 3 days post-injection. Orange = FITC-labeled 150 kDa Dextran marking the vasculature. Magenta = tdTomato-expressing mc²8471. Sky Blue = CFP-labeled macrophages. Scale bars = 20 µm. Please click here to view a larger version of this figure.

Figure 4: Association of mc²8471 with macrophages at 3 days post-injection (two-photon imaging). (A–C) Representative high-magnification snapshot images showing remaining bacilli in close association with CFP-labeled macrophages at 3 days post-injection. Orange = FITC-labeled 150 kDa Dextran marking the vasculature. Magenta = tdTomato-expressing mc²8471. Sky Blue = CFP-labeled macrophages. Scale bars = 20 µm. Please click here to view a larger version of this figure.

Supplementary Figure 1: Early intravascular appearance of mc²8471 in the lung immediately after intravenous injection. (A–C) Representative snapshots of the lung vasculature acquired immediately after intravenous injection and within 60 min post-injection. Orange = FITC-labeled 150 kDa dextran (vasculature). Magenta = tdTomato-expressing mc²8471. Sky Blue = CFP-labeled macrophages. Scale bars = 100 µm.Please click here to download this file.

Supplementary Figure 2: Early intravascular appearance of mc²8471 in the lung at 6 h post-injection. (A–C) Snapshots of the lung vasculature acquired 6 h after intravenous injection. White arrows denote representative regions of pulmonary occlusion. Orange = FITC-labeled 150 kDa dextran (vasculature). Magenta = tdTomato-expressing mc²8471. Sky Blue = CFP-labeled macrophages. Scale bars = 100 µm.Please click here to download this file.

Supplementary Video 1: Original z-stacks immediately after intravenous injection. Z-stack acquired immediately after intravenous injection of mc²8471 (<60 min post-injection); a representative 2D snapshot from this z-stack is shown in Figure 1C. Orange = FITC-labeled 150 kDa dextran (vasculature). Magenta = tdTomato-expressing mc²8471. Sky Blue = CFP-labeled macrophages. Scale bars are displayed in the z-stacks.Please click here to download this file.

Supplementary Video 2: Original z-stacks at 6 h post-injection. Z-stack acquired 6 h post-injection; a representative 2D snapshot from this z-stack is shown in Figure 2B. Orange = FITC-labeled 150 kDa dextran (vasculature). Magenta = tdTomato-expressing mc²8471. Sky Blue = CFP-labeled macrophages. Scale bars are displayed in the z-stacks.Please click here to download this file.

Supplementary Video 3: Original z-stacks at 3 days post-injection. Z-stack acquired 3 days post-injection; a representative 2D snapshot from this z-stack is shown in Figure 3C. Orange = FITC-labeled 150 kDa dextran (vasculature). Magenta = tdTomato-expressing mc²8471. Sky Blue = CFP-labeled macrophages. Scale bars are displayed in the z-stacks.Please click here to download this file.

Supplementary Video 4: High-resolution two-photon z-stack imaging. A two-photon z-stack was acquired from the lung surface to a depth of 42 µm (2 µm step size) in a mouse at 3 days post-injection of mc²8471. Multiple bacilli are visible throughout the stack, and most appear to be located within macrophages. This high-resolution two-photon z-stack enables precise assessment of bacillus position relative to CFP-labeled macrophages. Orange = FITC-labeled 150 kDa dextran (vasculature). Magenta = tdTomato-expressing mc²8471. Sky Blue = CFP-labeled macrophages.Please click here to download this file.

Supplementary Video 5: Two-photon time-lapse imaging. Two-photon 2D time-lapse images were acquired from a mouse at 3 days post-injection of mc28471 for 8 min 52 s at 11.1 s intervals. A single macrophage containing mc²8471 is visible. Dynamic movements of macrophages and negatively contrasted cells within the blood vessels are also observed. Orange = FITC-labeled 150 kDa dextran (vasculature). Magenta = tdTomato-expressing mc²8471. Sky Blue = CFP-labeled macrophages. Time stamps (seconds; mm:ss format) are shown in the upper-left corner of the movie.Please click here to download this file.

Discussion

This platform addresses a genuine gap in tuberculosis research by enabling direct visualization of bacilli and host-cell interactions within the intact mammalian lung under BSL-2 conditions. Prior in vivo imaging strategies, including whole-body optical detection using fiber-optic microendoscopic excitation²⁰, reporter enzyme fluorescence imaging²¹, and near-infrared fluorescent protein-expressing Mtb²², enabled sensitive longitudinal monitoring of bacterial burden and drug responses, but they measure only aggregate bacterial loads and cannot resolve individual and small clusters of bacilli, track aggregation dynamics, or reveal spatial relationships between mycobacteria and host cells within intact tissue architecture. Although BSL-3-compatible multiphoton systems have demonstrated proof-of-principle for imaging virulent Mtb in the lung²⁴, such implementations require specialized containment infrastructure that remains inaccessible to most laboratories. The present work builds on a lineage of intravital mycobacterial imaging that began with real-time visualization of M. marinum infection in zebrafish¹³ and continued with two-photon imaging of BCG granulomas in the mouse liver^14,15, extending these approaches for the first time to the visualization of bacilli and host-cell interactions within the intact mammalian lung. By combining the WHRIL permanent lung window with mc²8471, a fluorescent tdTomato-expressing derivative of the triple-auxotrophic BSL-2-approved Mtb strain mc²7902, this platform overcomes the containment barriers that have limited prior approaches. The result is serial detection and tracking of bacilli within the intact lung microenvironment over multiple days under standard BSL-2 conditions.

Using this system, we directly visualized early infection events previously accessible only through inference, including rapid vascular delivery of bacilli, early aggregation, dissemination into the parenchyma, and uptake by pulmonary macrophages. Although this study employs intravenous inoculation of an attenuated strain rather than aerosol delivery of virulent Mtb, the biological behaviors observed mirror key early steps thought to occur during pulmonary infection. In aerosol infection models, early productive infection has been shown to occur predominantly in airway-resident alveolar macrophages²⁸, providing an important point of comparison for the macrophage associations observed here. Intravenous delivery was therefore chosen as a controlled and reproducible route for studying these interactions under BSL-2 conditions, enabling mechanistic insights that would otherwise require highly specialized BSL-3 containment infrastructure. Extending this platform to aerosol infection with virulent Mtb under BSL-3 conditions remains an important future direction, and proof-of-principle BSL-3 multiphoton systems²⁴ show that such adaptations are technically feasible.

Several aspects of the protocol can be adapted to different experimental contexts. The inoculum of 5 x 10⁶ CFU was selected to ensure reliable detection of bacilli during the early post-infection period; lower doses more closely approximate physiological aerosol exposure but may make localization of bacteria technically challenging as the organisms are cleared from the lung, and this trade-off should guide dose selection for future applications. Window placement is not anatomically targeted, which introduces sampling constraints and animal-to-animal variability when quantifying infectious foci, and investigators should account for this when designing quantitative experiments. The use of Rag2-deficient mice can be substituted with immunocompetent animals when the research question requires evaluation of T and B cell responses. Similarly, while the protocol employs a laser-scanning confocal microscope, it can be implemented with multiphoton microscopy. Although multiphoton imaging typically offers advantages in tissue penetration depth, the lung's highly scattering architecture, with its many oil-water-air interfaces, results in minimal practical difference in penetration depth between the two modalities in this setting, and the choice should be guided by the specific experimental question and available instrumentation.

The current protocol carries limitations that must inform the interpretation of results. Most fundamentally, it employs an attenuated, auxotrophic Mtb strain delivered intravenously, which does not fully recapitulate aerosol infection with virulent Mtb, and imaging is restricted to the superficial subpleural lung region accessible through the WHRIL window, so whole-lung infection dynamics cannot be assessed. Quantitative features such as dissemination kinetics and immune cell recruitment should therefore be interpreted within the constraints of this attenuated intravenous model. Because Rag2-deficient mice lack mature T and B cells²⁶, adaptive immune responses cannot be evaluated in this setting, and observations should be understood as reflecting the earliest innate-phase interactions rather than the full spectrum of host responses to infection. This limitation can be addressed directly, as the technique is not inherently restricted to immunodeficient animals and can be performed in immunocompetent mice. A further interpretive limitation concerns the assignment of intracellular versus extracellular bacillary localization: unless imaging studies are specifically designed to achieve sufficient lateral and axial resolution, to use a well-characterized point spread function, and to employ an adequate z-stack step size, the position of bacteria relative to host-cell boundaries cannot be definitively resolved, and interpretation should remain conservative. Finally, because the WHRIL provides access to a fixed subpleural region, observations reflect local infection dynamics within that field and should not be assumed to represent the whole lung.

Troubleshooting and Limitations
Successful application of this protocol depends on careful attention to two independent sources of technical failure. The first is the failure of lung window implantation. Motion artifacts during imaging are a direct consequence of incomplete stabilization of the lung surface resulting from suboptimal window placement. Inadequate sealing of the coverslip will be immediately apparent, as mice will be unable to recover from surgery. Complete removal of intrathoracic air, proper use of positive end-expiratory pressure during coverslip attachment, and thorough but gentle drying of the lung surface before adhesion are each critical for achieving a stable, well-sealed window. It is essential that the lung surface be completely dry before adhesive application, as the most common cause of unsuccessful coverslip attachment is residual moisture on the lung surface prior to apposition with the frame or coverslip. Furthermore, only a very thin layer of adhesive should be applied; excess adhesive should be scraped off prior to coverslip placement, as thick adhesive layers degrade image quality. With repeated imaging sessions, exudate from the skin surrounding the window frame may congeal and prevent proper seating of the fixturing plate used to immobilize the mouse on the microscope stage; placing a moistened tissue over the window for 10 to 15 min will soften the exudate and allow proper placement of the plate. Injection volumes should be limited to a maximum of 50 µL at a time, as excess injected volume can cause the lung surface to excrete water and detach from the coverslip. Finally, some small amount of imaging drift may be observed immediately after placing the mouse on the microscope stage, arising from relaxation of the animal's body or thermal expansion of stage components; allowing approximately 30 min for equilibration before beginning image acquisition will minimize this effect, and any residual drift can be corrected computationally.

The second source of failure is the preparation of the bacterial inoculum. Incomplete dispersion of mc²8471 leads to artifactual aggregation prior to injection, directly confounding interpretation of early in vivo clustering events. Adequate sonication and inclusion of tyloxapol throughout the washing steps minimize clumping and improve reproducibility. Fluorescence intensity must be verified before infection, since reduced tdTomato signal can indicate plasmid loss or suboptimal growth conditions that would undermine detection of bacilli in tissue.

A technical constraint of the imaging setup is that confocal microscopy of the lung is limited in depth penetration; deeper bronchioles, vessels, and parenchymal structures are not accessible through the WHRIL window. For studies that require definitive assignment of intracellular versus extracellular bacillary localization, imaging should be planned in advance to ensure sufficient lateral and axial resolution, a well-characterized point spread function, and z-stack acquisition with adequate step size. These imaging conditions are achievable but are not guaranteed by default and must be deliberately built into the experimental design.

Finally, as this study was intended as a proof-of-principle demonstration of the application, all imaging data were obtained from one live mouse (n = 1) in a longitudinal experiment designed to assess time-dependent changes in mc²8471 localization in the lung following intravenous injection. Therefore, additional studies with larger cohort sizes are needed to confirm the reproducibility and generalizability of these observations. At each time point (immediately after injection, 6 h post-injection, and 3 days post-injection), multiple regions accessible through the permanent lung window were imaged. Representative images from these regions are provided in Supplementary Figure 1 and Supplementary Figure 2.

Despite these constraints, the platform opens a range of experimental investigations that have been previously inaccessible. Mtb undergoes profound metabolic remodeling during infection, including shifts toward host lipid and cholesterol utilization²⁹, yet how these physiological states influence bacterial morphology, aggregation, trafficking, or immune recognition in vivo remains unclear; because this system permits real-time visualization of bacilli within intact lung tissue, it can be directly applied to address these questions. The platform equally enables systematic dissection of genetically defined mutants with altered cell-envelope or virulence properties, including acid-fast-deficient mutants³⁰, non-cording variants that may influence aggregation or vascular occlusion, and ESX-1-deficient strains resembling BCG³¹, with mc²7902-derived strains remaining BSL-2-compatible²⁵ and therefore accessible to serial high-resolution imaging. A particularly compelling application is the direct visualization of immune-mediated bacterial killing. The conditionally persistent mutant mc²7901, which persists in immunodeficient hosts but is sterilized by adaptive immunity in immunocompetent mice³², provides a tractable substrate for such studies, and combining aerosol infection of fluorescent mc²7901 with intravital lung imaging could enable real-time visualization of when and where immune-mediated sterilization occurs at the single-cell level. This system also enables direct comparison of early bacillary fate in unimmunized versus mc²7902-immunized mice, a vaccination regimen known to induce exceptionally strong protection (unpublished data), potentially revealing whether sterilization correlates with altered macrophage behavior, accelerated intracellular killing, or changes in spatial organization within the lung, and thereby distinguishing immune mechanisms that restrain bacterial growth from those that achieve true clearance. In summary, this work provides a practical BSL-2-compatible platform for real-time intravital imaging of attenuated mycobacteria in the lung, establishes a foundation for mechanistic studies of bacterial physiology, host recognition, and immune-mediated clearance, and lays the groundwork for future extension to virulent Mtb under BSL-3 conditions.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
15 ml conical tube	Corning	430790
2-0 silk tie	Ethicon, Inc	LA55G
5 mm coverslip	Electron Microscopy Sciences	72296-05
5 mm disposable biopsy punch	Integra	33-35-SH
5-0 silk purse-string suture	Ethicon, Inc	640G
Blunt micro-dissecting scissors	Roboz	RS-5980
Buprenorphine	Hospira	0409-2012-32	Injection of 10 μL buprenorphine (0.3 mg/mL stock solution) diluted in 90 μL of sterile PBS (final dose = 0.1 mg/kg)
Compressed air canister	Falcon	DPSJB-12
Cyanoacrylate adhesive	Loctite	234790
D-calcium pantothenate	Acros	243305000
ECM 630 Electroporation System Complete, with Safety stand	BTX Harvard Apparatus	45-0051
Electroporation cuvette	Fisher Scientific	FB102
FITC-Dextran	Sigma-Aldrich	FD150S-1G	150 kDa
Glycerol	Fisher Chemicals	G33-1
Graefe forceps	Roboz	RS-5135	To grasp and elevate the rib during the lung window implantation procedure
Isoflurane	Pivetal	78949580
Isoflurane vaporizer	SurgiVet	VCT302
Kanamycin	Goldbio.com	K-120-25
L-arginine monohydrochloride	Sigma-Aldrich	A5131-100G
L-leucine	Sigma-Aldrich	L8000
MacBlue mice	Jackson Labs	Strain #: 026051
Middlebrook 7H10 Agar	Difco	262710
Middlebrook 7H9 broth	Difco	271310
Middlebrook OADC Enrichment	Becton Dickinson	212351
Murine ventilator	Kent Scientific	PS-02	PhysioSuite
Nair	Church & Dwight Co., Inc.	40002957
PBS	Corning	21-031-CV
PETG media inkwell bottle	Thermo Scientific	342020
PhysioSuite	Kent Scientific	PS-5623	To monitor the mouse’s tidal volume via intubation during surgery
Plasmid pYUB1169	Custom made plasmid
Rag2 KO mice	Jackson Labs	Strain #: 008449
Small animal lung inflation bulb	Harvard Apparatus	72-9083
Sodium propionate	Thermo Scientific	A17440.36
Sonicator	Fisher Scientific	FB705
Sorval Centrifuge X pro series	Thermo Scientific	75009020
Spectrophotometer Genesys 10 UV	Thermo Spectronic	335903
SurgiSuite	Kent Scientific	SURGI-M02	Multi-Functional Surgical Platform for Mice, with Warming
Tracheal catheter	Exelint International	26746	22G
Tyloxapol	Sigma-Aldrich	T0307-50G
Vacuum pickup system metal probe	Ted Pella, Inc.	528-112