This study was approved and performed in accordance with the ethical guidelines of the Institutional Animal Care and Use Committee at National Jewish Health (approval #AS2574-01-20). All mice were obtained from Jackson Laboratories (Bar Harbor, ME, USA) and bred at National Jewish Health. LysM-Cre (B6.129P2-Lyz2tm1(cre)Ifo/J; stock number 004781), Csf1r-cre (FVB-Tg(Csf1r-icre)1Jwp/J; stock number 021024), and CX₃CR1-ERT2 (B6.129P2(C)-CX₃CR1tm2.1(cre/ERT2)Jung/J; stock number 020940) mice were crossed with R26-stop^fl/fl-TdTomato (B6.Cg-Gt(ROSA)26Sortm14(CAG-tdTomato)Hze/J; stock number 007914) mice. Experimental animals from these three crosses were hemizygous for both genes. hCD68-rtTA (B6.Cg-Tg(CD68-rtTA2S*M2)3Mpil/Mmjax; stock number 32044-JAX) and R26-M2rtTA (B6.Cg-Gt(ROSA)26Sortm1(rtTA*M2)Jae/J; stock number 006965) mice were crossed with tetOn-GFP (Tg(tetO-HIST1H2BJ/GFP)47Efu/J; stock number 005104) mice. Experimental animals from hCD68-rtTA x tetOn-GFP were homozygous for both genes, while R26-rtTA x tetOn-GFP were hemizygous for both genes. Both male and female mice were used in these experiments.

Normal diet chow was replaced with doxycycline chow (Teklad laboratory diets, NJ, USA) for 1 week, then mice were analyzed on day 7 (pulse). Doxycycline chow was dosed at 625 mg doxycycline/kg chow, designed to provide 2–3 mg of doxycycline/day based on consumption of 4–5 g chow/day. Additional analyses were performed after mice were returned to normal diet for 1, 2, or 4 weeks (wait). Mice were allowed to eat chow ad libitum. Tamoxifen (Sigma) suspended in neutral oil was given at 0.25 mg/g of body weight by intraperitoneal (IP) injection. Mice were injected on day 1, again on day 4, then analyzed on day 7 (pulse), Additional analyses were performed 1, 2, or 4 weeks after the last tamoxifen injection (wait).

Mice were euthanized by CO₂ inhalation or IP injection of Fatal Plus (Vortech Pharmaceuticals, Dearborn, MI, USA). Peritoneal lavage (PL), bronchoalveolar lavage (BAL), lung, blood, bone marrow, and spleen were collected on ice in the following order. PL was collected in 10 mL of PBS containing 0.5 mM EDTA, pelleted, and resuspended in HBSS containing 0.3% BSA and 0.3 mM EDTA. BAL was collected in 5 mL of PBS containing 0.5 mM EDTA, pelleted, and resuspended in HBSS containing 0.3% BSA and 0.3 mM EDTA. Blood was collected by cardiac puncture using a syringe coated with 50 µL of 100 mM EDTA, then placed directly into 10 mL of PharmLyse buffer (BD Biosciences, San Jose, CA, USA) for 20 min on ice. After 20 min, blood cells were washed and resuspended in HBSS containing 0.3% BSA and 0.3 mM EDTA. Lungs were perfused by injecting 10 mL PBS through the right ventricle, after which lung tissue was visibly blanched. Lung tissue was finely chopped with a razor blade then incubated at 37°C in 1 mL of 0.4 mg/mL Liberase TM (Roche, Indianapolis, IN, USA) in RPMI for 25 min. Following incubation, tissue was pipetted rapidly up and down to further disaggregate, then filtered through 100 µm filters. Lung cells were pelleted and residual RBC were lysed for 30 s using 1 mL of BD PharmLyse (BD Biosciences, San Jose, CA, USA). BAL, lung, blood, and PL cells were washed in HBSS and resuspended in HBSS containing 0.3% BSA and 0.3 mM EDTA. Spleen tissue was chopped and mechanically dissociated, then filtered through 100 µm filters. Bone marrow was flushed from the cavities of the femur and tibia with cold PBS and mechanically dissociated, then filtered through 100 µm filters. Residual RBC in spleen and bone marrow were lysed for 10 min using 1 mL of Gey’s Solution (0.155 M NH4Cl 10 mM KHCO3 aqueous). The hemolysis wash was repeated. Spleen and BM cells were washed and resuspended in PBS containing 2% FBS, 2 mM EDTA, and 0.09% NaN₃.

Flow cytometry was performed on single cell suspensions. Cells were protected from light and incubations were performed on ice. Single cell suspensions were treated with unlabeled CD16/CD32 for 30 min to block non-specific FcyR-mediated binding. Cells were stained with surface antibody panels to identify myeloid populations for 1 h, and then washed. When necessary, cells were incubated with streptavidin-conjugated secondary antibody for another 45 min, and then washed. HBSS containing 0.3% BSA and 0.3 mM EDTA was used as a buffer for all incubations. Single cell suspensions of BAL, lung, blood, and PL were analyzed using an LSRII flow cytometer (BD) and FlowJo software (Tree Star, Ashland, OR, USA). Spleen and bone marrow cells were analyzed using a YETI cell analyzer (Propel Labs). IM1, IM2, IM3, and γδ T cells were the smallest populations examined; data collection was performed such that >400 events for each of these populations were analyzed, for most populations >5,000 events were analyzed. Primary antibodies were used in various combinations as follows (source/clone): unlabeled CD16/32 (eBioscience/93), Ly6G (BD/IA8), MHCII (BD/114.15.2), F4/80 (ebioscience/BM8), CD45 (BD/30-F11), Ly6C (eBioscience/HK1.4), Siglec-F (BD/E50-2440), CD11c (eBioscience/N418), CD11b (eBioscience/M1/70), CD103 (eBioscience/2E7), CD115 (eBioscience/AFS98), CD64 (Biolegend/X54-5/7.1), MerTK (R&D/BAF591), CD3 (Biolegend/17A2), CD19 (eBioscience/eBio1D3), NK1.1 (Biolegend/PK136), γδTCR (eBioscience/eBioGL3), EpCAM (Biolegend/G8.8), CD31 (Biolegend/390), Thrombomodulin (R&D/AF3894), CD206 (Biolegend/C068C2), Tim4 (Biolegend/F31-5G3), and Gr-1 (BD/RB6-8C5). All primary antibodies were used at a 1:400 dilution except EpCAM (1:1,000), and MerTK (1:100). Streptavidin-conjugated FITC, AF647, and PE secondary antibodies (Jackson Immunolaboratories) were used at 1:200.

LPS (Escherichia coli 055:B5; List Biological Laboratories, Campbell, CA, USA) was administered in a dose of 20 µg in 50 µL of PBS. LPS was instilled intratracheally (IT) using a modified gavage needle. Mice were sedated with isoflurane (Baxter, Deerfield, IL, USA) before IT instillations. Reporter expression was assessed in BAL, lung, and blood 6 days after LPS instillation.

To improve our understanding of conditional and conditional inducible systems within the lung and aid in experimental design and interpretation, we assessed the specificity and strength of four different Cre driver lines for targeting macrophages. The two conditional lines selected, LysM-Cre and Csf1r-Cre, are the most highly published “macrophage targeting” lines in the literature (3, 9, 28). The two inducible lines, CX₃CR1-ERCre and CD68-rtTA, were developed more recently and have been used less extensively. However, based on previous work, we suspected that both lines might be appropriate for targeted studies of IM. Moreover, the two inducible lines have the advantage of strict temporal control. All conditional and conditional inducible lines were crossed with various reporter lines to mark Cre activity, discussed in detail below. We began by assessing reporter expression in the homeostatic lung.

Flow cytometry was used to assess reporter expression in cells from digested lung tissue (Figure 1A). Using surface expression of CD45, cells were split into hematopoietic (CD45⁺) and non-hematopoietic (CD45⁻) populations. Within non-hematopoietic cells, we identified epithelial cells as EpCAM⁺, and endothelial cells as EpCAM⁻, CD31⁺ and thrombomodulin⁺. To identify our populations of interest in hematopoietic cells, we first excluded neutrophils using Ly6G. Within Ly6G⁻ cells, macrophages were identified as double-positive for CD64 and MerTK, and non-macrophages as low or negative for CD64 and MerTK expression. From the non-macrophage population, lymphocyte populations were segregated. B cells were identified by their high expression of CD19⁺. CD3⁺ was used to identify T cells that were further identified as γδ T cells or classical αβ T cells by expression of the γδ TCR. Within remaining CD19⁻CD3⁻ cells, NK cells were identified as NK1.1⁺. Also within the non-macrophage population, CD64⁻ and MerTK⁻ cells, DCs were identified by their high expression of MHCII and CD11c, and then further subdivided into CD11b⁺ DC and CD103⁺ DC by expression of their eponymous markers. The macrophage population identified by MerTK and CD64 was divided into CD11c^hi ResAM and CD11b^hiCD11c^hi/lo IMs. IMs were then divided into subsets, first using CD11c and CD206 to identify IM3 (CD206^low), and then remaining cells (CD206^hi) were defined as IM1 and IM2 using expression of MHCII.

To assess the efficiency and specificity of the Csf1r-Cre and LysM-Cre conditional systems, each strain was crossed with R26-stop^fl/fl-TdTomato reporter animals. In the resultant offspring, cells that express Cre excise the floxed stop codon upstream of the TdTomato construct and the fluorescent protein is expressed. Both Csf1r-Cre and LysM-Cre were strongly activated in ResAM and IM, with nearly 100% efficiency (Figures 1B,C). However, further examination of Csf1r-Cre demonstrated that this conditional line targeted nearly all leukocytes in the lung, including lymphocytes, with greater than 80% efficiency (Figure 1B). In contrast, LysM-Cre minimally target lymphocytes, although it did target approximately half of Ly6C^hi monocytes, CD11b⁺ DC and CD103⁺ DC (Figure 1B) (9). Of significant importance, LysM-Cre targeted nearly a quarter of lung epithelial cells (Figure 1B). Previous studies have suggested that these targeted epithelial cells are alveolar type II cells (1). In comparison, Csf1r-Cre did not target lung epithelium or endothelium. Data presented in Figure 1B are also summarized in Table 1 and representative histograms for all cell populations are shown in Figure S1 in Supplementary Material.

As a next step, the inducible targeting systems were explored. First, CX₃CR1-ERCre mice were crossed with R26-stop^fl/fl-TdTomato reporter mice. In these animals, cells express a mutated Cre that is directed by the CX₃CR1 promoter and requires tamoxifen as a co-factor for activity. Accordingly, when tamoxifen is administered, Cre recombinase becomes active in all CX₃CR1 expressing cells, thereby excising the floxed R26-stop floxed codon and enabling production of TdTomato. In a similar fashion, CD68-rtTA mice were crossed with tetOn-GFP mice. In these animals, rtTA acts as a transcription factor for the tetOn promoter, but requires the presence of doxycycline for activity. Thus, GFP is expressed in CD68 positive cells following administration of doxycycline. The duration of tamoxifen or doxycycline administration will affect the efficiency of targeting as the percentage of target cells affected increases over time. We have previously reported that CD68-rtTA reporter expression increases to a maximum level over 6 days (29). We therefore assessed CX₃CR1-ERT2 and CD68-rtTA reporter mice after 1 week of tamoxifen or doxycycline.

CX₃CR1-ERCre and CD68-rtTA efficiently targeted IM, with 90 and 75% efficiency, respectively (Figures 1B,C). However, ResAM remained reporter negative in both lines, as we have shown previously for CD68-rtTA (29). In addition to IMs, both CX₃CR1-ERCre and CD68-rtTA targeted lung DCs (Figure 1B). CX₃CR1-ERCre also targeted a significant portion of NK cells, although targeting of other lymphocytes remained low (Figure 1B). Reporter expression was not observed in endothelial or epithelial cells for either line. The data presented in Figure 1B is also summarized in Table 1 and representative histograms for all cell populations are shown in Figure S1 in Supplementary Material.

During homeostasis, Ly6C^hi monocytes constitutively traffic through the lungs (30). During inflammation, emigration of these cells from the circulation is increased and it is believed that they give rise to RecMΦ. Neutrophils and other leukocytes also emigrate from the bloodstream during inflammation. Therefore, we assessed activation of Csf1r-Cre, LysM-Cre, CX₃CR1-ERCre, and CD68-rtTA in blood leukocytes. Blood was obtained by cardiac puncture, red cells were lysed, and reporter expression was measured by flow cytometry (Figure 2A). First, CD45⁺ leukocytes were identified. Neutrophils were identified using Ly6G, and eosinophils distinguished using Siglec-F. Monocytes were identified from remaining cells as CD115⁺CD11b⁺. B and T cells were identified within the remaining pool as CD19⁺ or CD3⁺, respectively.

Ly6C^hi and Ly6C^lo monocyte subsets were targeted by all four lines, with the highest percentage of cells targeted by Csf1r-Cre (Figures 2B,C). Csf1r-Cre, LysM-Cre, and CD68-rtTA were activated in neutrophils, and both Csf1r-Cre and LysM-Cre were activated in eosinophils (Figure 2B). Only Csf1r-Cre was completely activated in lymphocytes (Figure 2B). The data presented in Figure 2B is also summarized in Table 1 and representative histograms for all cell populations are shown in Figure S2 in Supplementary Material.

Although resident tissue macrophages of other organs do not directly migrate to the lung, macrophages of other organs, particularly the bone marrow and spleen, regulate hematopoiesis, and impact circulating leukocyte levels (31–33). We examined reporter expression of Csf1r-Cre, LysM-Cre, CX₃CR1-ERCre, and CD68-rtTA-driven lines in bone marrow, spleen, and PL using flow cytometry (Figures S3A,B,E in Supplementary Material). Macrophages in all three tissues were targeted by all four drivers (Figures S3C,D,F,G in Supplementary Material).

A distinct advantage of inducible Cre systems such as CX₃CR1-ERCre and CD68-rtTA is that the timing and duration of tamoxifen or doxycycline administration can be used to enhance cell selectivity and specificity. As an example, we explored the kinetics of dosing in both inducible strains to refine targeting of IM. Mice were treated with either tamoxifen or doxycycline for 1 week, then the drug was withdrawn and reporter expression was examined 1, 2, and 4 weeks later.

In the lung tissue of both CX₃CR1-ERCre and CD68-rtTA mice, IM1 and IM2 reporter expression remained stable after tamoxifen or doxycycline was withdrawn, with only a 5% decrease in reporter positive cells over 4 weeks (Figures 3A,B). IM3 showed a steeper rate of declining reporter expression, with a 20% decrease in reporter positive cells (Figures 3A,B). As anticipated, ResAM were reporter negative at all times. Importantly, the percentage of reporter positive DC dropped to less than 10% in CX₃CR1-ERCre mice by 4 weeks and to less than 30% in CD68-rtTA mice. Reporter expression was similarly low in lymphocytes isolated from the lungs, with the exception of NK cells in CX₃CR1-ERCre mice, in which nearly 25% remained positive (Figures 3C,D).

Reporter expression in neutrophils and monocytes from peripheral blood was also lost after 4 weeks of tamoxifen or doxycycline withdrawal (Figures 3E,F). Virtually, all of the neutrophils and Ly6C^hi monocytes were TdTomato reporter negative, whereas approximately 20% of Ly6C^lo monocytes remained TdTomato positive in both strains. These findings support the concept that neutrophils and Ly6C^hi monocytes have relatively short lifespans, and that turnover of neutrophils and Ly6C^hi monocytes are relatively high (30). In addition, since excision of floxed alleles is permanent in affected cells (i.e., cells stay reporter positive for life), the observation that the IM3 population gradually loses reporter expression suggests that at least a portion of these cells are replaced by monocyte precursors, supporting previous reports using parabiotic mice (14). Taken together, these findings show that dosing of tamoxifen and doxycycline can be manipulated to enhance targeting of IMs in CD68-rtTA and CX₃CR1-ERCre mice.

Inflammation has been reported to alter Cre activity, and during inflammation a new population of macrophages is present in the lung airways: RecMΦ. This macrophage population is believed to arise from infiltrating monocytes maturing in situ, and can be distinguished from ResAM by high expression of CD11b (Figure 4A) [and low expression of Siglec-F (17, 34)]. We assessed Csf1r-Cre and LysM-Cre reporter mice following LPS injury for the potential changes to Cre activity and measured reporter expression in recruited macrophages. RecMΦ were targeted by both Csf1r-Cre and LysM-Cre with near 100% efficiency (Figures 4B–D). Efficiency for all three IM populations was also high (>95%; Table 2). However, as seen during homeostasis, other leukocyte populations were also highly targeted in Csf1r-Cre mice and to a lesser extent in LysM-Cre animals (Table 2).

We have previously reported that ResAM do not activate CD68-rtTA in vivo after 21 days of doxycycline, or in vitro when doxycycline was provided in culture media for 4 days (29). However, it was unclear whether this restriction reflected a failure to activate the CD68 promoter or resulted from decreased rtTA activity. To assess this, we obtained R26-rtTA mice (35), where rtTA is expressed under control of the Rosa26 promoter, which is ubiquitously expressed by all leukocytes including ResAM (36). Both CD68-rtTA and R26-rtTA mice were crossed with tetOn-GFP mice, and resulting offspring were administered doxycycline for 10 months. Notably, ResAM from CD68-rtTA mice remained GFP negative, although the majority of myeloid cells (including IM) were GFP positive (Figures 5A,B). Strikingly, ResAM from R26-rtTA mice also failed to express GFP, whereas >95% of other leukocytes were GFP positive (Figures 5A,C). We conclude that ResAM are resistant to the activation of rtTA-driven systems. Accordingly, the CD68-rtTA system can be exploited to target-specific lung tissue macrophage populations and to distinguish the roles they play during both homeostasis and in lung disease.