Keratinocyte Growth Factor Administration Attenuates Murine Pulmonary Mycobacterium tuberculosis Infection through Granulocyte-Macrophage Colony-stimulating Factor (GM-CSF)-dependent Macrophage Activation and Phagolysosome Fusion*

Background: Mycobacterium tuberculosis is a major cause of morbidity and mortality worldwide, and chemotherapeutic treatments are not always effective. Results: Keratinocyte growth factor (KGF) administration results in an epithelium-dependent, GM-CSF-mediated induction of microbicidal programs in alveolar macrophages. Conclusion: KGF induces GM-CSF and enhances clearance of M. tuberculosis from the lungs of mice. Significance: KGF may be a useful adjunctive strategy for treatment of infection with M. tuberculosis. Augmentation of innate immune defenses is an appealing adjunctive strategy for treatment of pulmonary Mycobacterium tuberculosis infections, especially those caused by drug-resistant strains. The effect of intranasal administration of keratinocyte growth factor (KGF), an epithelial mitogen and differentiation factor, on M. tuberculosis infection in mice was tested in prophylaxis, treatment, and rescue scenarios. Infection of C57BL6 mice with M. tuberculosis resulted in inoculum size-dependent weight loss and mortality. A single dose of KGF given 1 day prior to infection with 105 M. tuberculosis bacilli prevented weight loss and enhanced pulmonary mycobacterial clearance (compared with saline-pretreated mice) for up to 28 days. Similar effects were seen when KGF was delivered intranasally every third day for 15 days, but weight loss and bacillary growth resumed when KGF was withdrawn. For mice with a well established M. tuberculosis infection, KGF given every 3 days beginning on day 15 postinoculation was associated with reversal of weight loss and an increase in M. tuberculosis clearance. In in vitro co-culture experiments, M. tuberculosis-infected macrophages exposed to conditioned medium from KGF-treated alveolar type II cell (MLE-15) monolayers exhibited enhanced GM-CSF-dependent killing through mechanisms that included promotion of phagolysosome fusion and induction of nitric oxide. Alveolar macrophages from KGF-treated mice also exhibited enhanced GM-CSF-dependent phagolysosomal fusion. These results provide evidence that administration of KGF promotes M. tuberculosis clearance through GM-CSF-dependent mechanisms and enhances host defense against M. tuberculosis infection.

Mycobacterium tuberculosis, the causative agent of tuberculosis (TB), 2 infects one-third of the world's population and results in up to 1.2 million deaths per year (1). Strains resistant to all major anti-TB drugs have emerged, and novel approaches to therapy are needed. Inhaled M. tuberculosis bacteria impact the airway epithelial lining layer and are internalized into the phagosomes of alveolar macrophages (AMs) (2)(3)(4). Infected AMs secrete TNF␣ and chemokines that coordinate the recruitment of T lymphocytes into lung granulomas, which function to contain the infected cells and to facilitate the execution of microbicidal programs (5). T lymphocytes within the lesions secrete IFN-␥, GM-CSF and other cytokines that promote effective intracellular killing by AMs through both oxidative and nonoxidative mechanisms (6). Fusion of the macrophage phagosome with lysosomes is a tightly regulated event that exposes the bacterium to the acidic pH and digestive enzyme milieu within the lysosome and thereby promotes bacterial killing. M. tuberculosis can subvert host innate immune responses and survive within macrophages through a variety of adaptive mechanisms, including inhibition of phagolysosome fusion or acidification of the phagosome and scavenging of iron, leading to latency or progressive pulmonary infection (4). For the billions of individuals with latent TB, recrudescence years after the primary exposure is the most common source of new infections in the more than 9 million patients with active TB worldwide (1,7). Treatment strategies focused on activating or reactivating the microbicidal activities of the AMs, including phagosome-lysosome fusion, are therefore intuitively appealing.
Keratinocyte growth factor (KGF) is a potent epithelial mitogen that is known to contribute to epithelial repair in several organs. KGF is expressed by mesenchymal cells and binds to FGFR2b receptors that are almost exclusively restricted to the epithelium (8). Previous studies indicate that KGF protects the lung from various insults, such as hypoxia, acid instillation, and intratracheal bleomycin (9 -16) and enhances host survival following Pseudomonas aeruginosa-induced lung injury in vivo (17,18). We have recently reported that exogenous KGF results in activation of AMs and enhanced clearance of Gram-negative bacteria from murine lungs at least in part by inducing the secretion of GM-CSF from the pulmonary epithelium and GM-CSF receptor-dependent activation of the STAT5 signaling pathway in AMs (19).
The objective of the current study was to determine whether KGF would promote pulmonary host defense against pulmonary M. tuberculosis infection in mice. We provide evidence that KGF enhances the clearance of M. tuberculosis through GM-CSF-induced activation of AM, nitric oxide-dependent killing, and an increase in phagosome-lysosome fusion.

EXPERIMENTAL PROCEDURES
Animals-Female C57BL/6 mice aged 6 -8 weeks were purchased from Jackson Laboratories (Bar Harbor, ME) and housed under pathogen-free conditions in the animal facility of the University of Cincinnati College of Medicine. Mice were given ad libitum access to sterilized food and water. M. tuberculosis-infected mice were housed in a separate BSL3 animal facility. Animals were infected with M. tuberculosis by the intranasal route and treated with human recombinant KGF (Palifermin, a gift of Biovitrium UMEA) or saline. All animal procedures were approved by the University of Cincinnati Institutional Animal Care and Use Committee.
Culture and Isolation of M. tuberculosis-The virulent Erdman strain of M. tuberculosis was obtained from the ATCC (ATCC 35801). The GFP-expressing M. tuberculosis Erdman strain used for phagolysosome fusion experiments was a gift from Dr. Vojo Deretic (University of New Mexico). M. tuberculosis was cultivated on 7H11 agar plates and harvested in RPMI 1640 containing 10 mM HEPES. To minimize problems with clumping, suspensions were gently sonicated in a GenProbe bath sonicator. Inoculum size and mycobacterial burden were confirmed by serial dilution and determination of colony-forming units (cfu) on agar plates. To assess M. tuberculosis growth from the infected mice, the lungs were removed aseptically at specified time points, cut into small pieces, and homogenized. Viable M. tuberculosis in the lung tissue homogenates was quantified by serial dilution, plating in duplicate onto 7H11 Middlebrook agar in 6-well plates, incubation at 37°C in a 5% CO 2 incubator for 3 weeks, and counting of cfu (20). The plates were incubated for an additional 2 weeks to detect any slower growing M. tuberculosis colonies. The results are expressed as means Ϯ S.E.
Cultured Cells-MLE-15 cells (gift of J. Whitsett, Cincinnati Children's Hospital Medical Center) are an immortalized cell line derived from the lung tumors of transgenic mice expressing the simian virus 40 (SV40) large T antigen under the transcriptional control of the human surfactant protein C (SP-C) promoter (21). MLE-15 cells retain many characteristics of alveolar type II cells, including epithelial cell morphology, microvilli, cytoplasmic multivesicular bodies, multilamellar inclusion bodies, and expression of SP-A, SP-B, and SP-C (22). MLE-15 cells were grown in RPMI 1640 (Gibco) containing 2% fetal bovine serum, 0.5% insulin-transferrin-sodium selenite (Sigma), 5 mg/liter transferrin, 10 mM HEPES, 10 Ϫ8 M ␤-estradiol, and 10 Ϫ8 M hydrocortisone. RAW 264.7 (RAW) cells (American Type Culture Collection, Manassas, VA) are a mouse macrophage cell line that was originally established from a tumor induced by the Abelson murine leukemia virus. RAW cells were maintained in DMEM (Invitrogen) supplemented with 10% heat-inactivated FBS (Invitrogen) and cultured at 37°C in a 5% CO 2 atmosphere. MLE-15 cells were plated at 25% confluence in DMEM (Gibco) in the lower well of transwell plates and were pretreated with KGF (100 ng/ml) or medium alone overnight. RAW cells were seeded at 25% confluence in the upper well of transwell dishes. On the third day, when the RAW cells were ϳ50 -60% confluent, they were incubated with M. tuberculosis at a multiplicity of infection of 10:1 for 2 h, washed, and co-incubated in transwell co-cultures with MLE-15 cells. After 5 days, the RAW cells were harvested, lysed, and plated onto 7H11 Middlebrook agar plates. M. tuberculosis cfu were quantified as outlined above.
ELISA-GM-CSF concentrations were measured in cell-free supernatants of culture samples or in lysates of freshly isolated AMs using a commercially available ELISA kit (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions and were expressed as pg/ml of the culture supernatant or lysate. In some experiments, NG-nitro-L-arginine methyl ester (L-NAME) was added to the medium to inhibit nitric-oxide synthase.
Isolation of Murine Alveolar Macrophages-Mice were sacrificed, and bronchoalveolar lavage (BAL) was performed by up to 10 cycles of instillation and aspiration of 1 ml of Hanks' balanced salt solution with 0.6 mM EDTA (23). The BAL fluid was centrifuged at 1500 ϫ g for 5 min, and the cells were resuspended in lysis buffer containing 0.15 M NH 4 Cl with 0.01 M KHCO 3 , washed, and resuspended in 0.9% NaCl. Cells were quantified using a hemocytometer, and cell differentials were determined by examination of cytopreparation smears (Cytospin II, Shandon Southern Instruments, Inc., Sewickley, PA) after processing with the Hema 3 staining system (Fisher). Viability of AM was determined by trypan blue exclusion.
Infection of Mice with M. tuberculosis-Mice were infected i.n. with Erdman M. tuberculosis (2.5 ϫ 10 1 to 1 ϫ 10 6 ) suspended in 50 l of Hanks' balanced salt solution in a class III biohazard safety cabinet, and maintained in the BSL3 animal facility at the University of Cincinnati. Mice were treated with intranasal saline (PBS) or KGF (5 mg/kg) as indicated and monitored for weight change and vital status or sacrificed for the determination of cfu in lung homogenates. The initial inoculum that was used to infect the mice was verified by plating onto 7H11 Middlebrook plates.
Assessment of Phagosome-Lysosome Fusion-To examine the co-localization of M. tuberculosis with phagosomes, AMs from KGF-or saline-treated mice were plated onto slides at 1 ϫ 10 5 AMs/well in DMEM for 4 h. After incubation for 1 h at 37°C in 5% CO 2 , GFP-M. tuberculosis Erdman was added at a multiplicity of infection of 10:1 for 2 h. For the neutralization of the GM-CSF-dependent phagolysosome fusion, anti-GM-CSF antibody (eBioscience, San Diego, CA) was added (100 g/well) to AMs. After washing the monolayer, the AMs were fixed serially with 1 and 2% paraformaldehyde for 10 min each. The monolayers were permeabilized with 100% methanol for 5 min, washed three times with Dulbecco's PBS, and blocked overnight at 4°C with PBS containing normal goat serum, 10% heatinactivated FCS (Hyclone), and 5 mg/ml BSA. After washing, the monolayers were incubated with mouse anti-human LAMP-1 antibody (University of Iowa Hybridoma Facility, Iowa City, IA) for 1 h, followed by incubation with mouse IgG conjugated with Alexa Fluor 555 (1:5000 dilution; Molecular Probes). After washing, the slides were separated from the wells, overlaid with a coverslip and mounting medium, and sealed with nail polish. The slides were viewed using a confocal microscope (Olympus FV1000). On average, 200 cell-associated bacilli contained within ϳ70 macrophages were counted from randomly selected confocal images in each test group. Results are expressed in percentage values of phagosome-lysosome fusion and shown as the mean Ϯ S.E. of three independent experiments.
Statistical Analysis-All experiments were conducted in triplicate with a minimum of three experiments. The results were expressed as means Ϯ S.E. or S.D. The statistical difference between the control and experimental data was determined by Student's t test or analysis of variance with paired comparisons. A p value of Ͻ0.05 was considered to be statistically significant.

Infection by M. tuberculosis Results in Inoculum
Size-dependent Mortality in C57BL/6 Mice-Experiments were performed to determine the optimal M. tuberculosis inoculum size for subsequent experiments. A group of nine mice received Erdman M. tuberculosis i.n. at doses ranging from 10 2 to 10 9 bacteria and were observed daily for up to 35 weeks (Fig. 1). Mice that received 1 ϫ 10 9 or 1 ϫ 10 8 cfu of M. tuberculosis survived for up to 8 or 12 weeks, respectively, whereas those that received 1 ϫ 10 4 to 1 ϫ 10 2 M. tuberculosis survived beyond 32 weeks. We chose an inoculum of 10 5 organisms/ mouse for most experiments, except where otherwise indicated. Collectively, these data indicate that infection with the M. tuberculosis (Erdman) strain results in a dose-dependent mortality in C57BL/6 mice.
KGF Protects Mice from Weight Loss and Enhances Clearance of M. tuberculosis from Mouse Lungs-The effect of KGF pretreatment on body weight and pulmonary burden of M. tuberculosis bacilli was assessed. The dose and route of KGF administration were based on parameters shown to enhance pulmonary clearance of Gram-negative bacteria in a recent study from our group (19), which demonstrated a 2.5-fold increase in AMs recovered by BAL and type II cell proliferation without appreciable histological evidence of inflammation. Mice were treated with a single intranasal dose of KGF (5 mg/kg) or saline and then inoculated i.n. with 10 5 M. tuberculosis bacilli 24 h later. The mice that received saline lost more than 20% of their baseline body weight by day 20; in contrast, the mice that received KGF gained weight to about 20% above baseline by day 20 (n ϭ 4 mice/ group, p Ͻ 0.05) ( Fig. 2A). At day 30, the lungs were harvested, and cfu in the lung homogenates were quantified (Fig. 2B). KGF pretreatment resulted in a 1.9 Ϯ 0.25-fold reduction in the number of cfu in the lungs compared with the saline-treated controls (p Ͻ 0.05). There were no differences in mean weights of control mice treated with KGF or saline in the absence of infection (data not shown). Collectively, the data indicate that a single prophylactic dose of KGF prevents weight loss and enhances the clearance of M. tuberculosis from the lungs of mice.
To determine whether serial treatment with KGF protects mice from weight loss and progressive mycobacterial infection, mice were inoculated i.n. with 10 5 M. tuberculosis bacilli and then treated with PBS or KGF i.n. every 3 days for 15 days (Fig.  3). At that point, saline-treated mice were divided into two groups: (a) continued saline administration every third day through day 45 (n ϭ 4) (saline only) or (b) crossed over to receive KGF every third day through day 45 (n ϭ 4) (saline then KGF). Similarly, mice that received KGF for the first 15 days were either given (a) KGF treatment every third day for 30 additional days through day 45 (KGF only) or (b) saline every third day from day 15 through 45 (KGF then saline).
The saline-only mice lost ϳ33% of their body weight (Fig.  3A). In contrast, the KGF-only mice gained weight, plateauing at about 10 -11% above their baseline weight by day 15. The difference in body weight in the two groups was significant at day 20 (p Ͻ 0.001) and day 45 (p Ͻ 0.001). In the saline-only group, M. tuberculosis cfu in lung homogenates increased by ϳ2.6-fold over the interval from 15 to 45 days, from 3.5 Ϯ 0.6 ϫ 10 6 bacteria at 15 days to 7.1 Ϯ 0.6 ϫ 10 6 bacteria at 30 days and 9.3 Ϯ 0.4 ϫ 10 6 bacteria at 45 days (Fig. 3B). The KGF-only mice had a lower burden of M. tuberculosis in the lung on day 15 compared with saline-treated mice and only a 1.6-fold log increase in the number of organisms over the following 30 days (from 2.0 Ϯ 0.1 ϫ 10 6 bacteria on day 15 to 3.2 Ϯ 0.1 ϫ 10 6 bacteria on day 30 and 4.1 Ϯ 0.2 ϫ 10 6 bacteria on day 45). We concluded that continuous treatment with KGF for 45 days protects mice from weight loss and reduces the growth of M. tuberculosis in the lungs.
Crossover experiments were performed to determine the effect of KGF initiation on established M. tuberculosis infection and the effect of withdrawing KGF on the subsequent course of infection. In "saline then KGF" mice, the weight fell upon saline treatment by 20% through day 15, but after serial KGF administration every 3 days commenced on day 16, weight plateaued and began to increase again (Fig. 3A). In the first 2 weeks after KGF administration commenced, the burden of organisms continued to increase, and it more than doubled from 3.5 Ϯ 0.6 to 7.4 Ϯ 0.2 ϫ 10 6 cfu from day 15 to 30. By day 45 however, M. tuberculosis cfu in the lung had fallen to 6.4 Ϯ 0.2 ϫ 10 6 cfu (Fig. 3B). In "KGF then saline" mice, the weight was stable for approximately the first 30 days, despite KGF exposure for only the first 15 days, but at that point, it began to decrease. At 2.0 Ϯ 0.01 ϫ 10 6 bacteria on day 15, the bacterial burden in the KGF treatment animals was lower at the same time point than in the "saline alone" mice. After KGF withdrawal, M. tuberculosis growth was similar to that in "KGF only" mice through day 30 but then accelerated relative to that group over the interval from day 30 to 45, from 3.4 Ϯ 0.2 ϫ 10 6 to 6.2 Ϯ 0.1 ϫ 10 6 bacteria. We concluded that KGF can reduce M. tuberculosis growth in an established infection model and that withdrawal of KGF is associated with progression of M. tuberculosis infection.
KGF Enhances the Clearance of M. tuberculosis Inoculated at Doses That More Closely Mimic Inhalational Exposure-To more closely model natural infection with M. tuberculosis in humans, we measured the effect of KGF on the bacterial burden in the lungs after low dose M. tuberculosis inoculation. Mice were inoculated i.n. with 1 ϫ 10 2 Erdman M. tuberculosis. After 30 weeks, mice were given intranasal saline or KGF (5 mg/kg) every third day for 2 weeks. Four weeks after commencement of the intervention, the lungs were harvested and homogenized, and cfu were quantified. We found that KGF enhanced the clearance of M. tuberculosis by Ͼ5-fold (n ϭ 4, p Ͻ 0.001) (Fig.  4). These data indicate that KGF enhances the clearance of M. tuberculosis in an established infection model following a  low dose inoculation that more closely mimics inhalational tuberculosis.
Finally, to determine whether KGF treatment enhances the survival of mice that harbor a heavy burden of M. tuberculosis (as occurs in advanced disease), mice were inoculated with 1 ϫ 10 8 Erdman M. tuberculosis and then treated with intranasal KGF or saline every third day for 2 weeks. Vital status was monitored on a daily basis over a 50-day period starting from the day of initial infection. As shown in Fig. 5, 50-day survival was significantly greater in the KGF-treated group (75%) than in the saline-treated group (34%) (n ϭ 9, p ϭ 0.033). These data indicate that KGF protects mice from a lethal intranasal inoculum of M. tuberculosis.
A Co-culture Model of Alveolar Epithelial Cells and Macrophages Demonstrates That GM-CSF-dependent Killing of M. tuberculosis Is Induced by KGF Administration-We have previously reported that KGF enhances pulmonary epithelial GM-CSF production, which in turn activates AMs and enhances bacterial clearance (19). To determine whether GM-CSF-dependent AM activation contributes to the enhanced M. tuberculosis clearance seen in KGF-treated mice, we con-ducted a series of in vitro co-culture experiments using macrophage (RAW) and alveolar pneumocyte (MLE-15) cell lines (Fig. 6). MLE-15 cells were plated in the lower well of transwell plates and pretreated with KGF (100 ng/ml) or medium alone overnight. RAW cells were infected with M. tuberculosis at a multiplicity of infection of 10:1, washed, and added to the upper well. After 5 days of co-culture, the RAW cells were harvested, lysed, and plated onto 7H11 Middlebrook agar plates, and M. tuberculosis cfu were quantified. We found that KGF treatment of MLE-15/RAW co-cultures reduced the growth of AMassociated M. tuberculosis ϳ2.5-fold (p Ͻ 0.05) (Fig. 6A). The GM-CSF content of the medium supernatants under various culture conditions is shown in Fig. 6B. These data indicate that MLE-15 cells produce a small amount of GM-CSF under basal conditions, which is incrementally augmented by M. tuberculosis infection of AM, by the addition of KGF to the medium bathing MLE-15 cells, and by co-culture with uninfected AMs.   Co-culture of M. tuberculosis-infected RAW cells with KGFtreated MLE-15 cell monolayers resulted in an ϳ7-fold increase in the level of GM-CSF in culture medium, relative to the levels in the culture medium of unstimulated, uninfected cells (p Ͻ 0.05). Conditioned medium (CM) from KGF-treated MLE-15 cells was then used to determine whether GM-CSF was the factor responsible for the enhanced killing of M. tuberculosis (Fig. 7). Incubation of M. tuberculosis-infected RAW cells with CM reduced the growth of intracellular M. tuberculosis by about 2-fold. This effect of CM was reversed by a neutralizing antibody to GM-CSF but not by an isotype control antibody. In addition, the mycobactericidal effect of CM was blocked by the nitric oxide synthase inhibitor, L-NAME. Incubation of either anti-GM-CSF antibody or L-NAME with the RAW cells without CM did not have any effect on the M. tuberculosis growth (data not shown). These results indicate that KGF treatment of MLE-15 cells induces GM-CSF production that results in a decrease in M. tuberculosis burden in co-cultured RAW, at least in part through nitric oxide-dependent mechanisms. These data support an important role for GM-CSF and nitric oxide in the KGF-mediated anti-mycobacterial action of AMs.
Effect of KGF on Phagosome-Lysosome Fusion in M. tuberculosis-infected AMs-We also examined the effect of KGF on phagosome-lysosome fusion in M. tuberculosis-infected AMs (Fig. 8). AMs were isolated 24 h after treatment with KGF or saline and incubated with GFP-expressing M. tuberculosis. Following phagocytosis, green fluorescent M. tuberculosis bacilli were visible within phagosomes using confocal microscopy. The late endosomal/lysosomal compartment was visualized by staining with a red fluorescent LAMP-1 antibody. Co-localization of the red and green signals indicative of phagosome-lysosome fusion was quantified by counting yellow fluorescent signals. AMs isolated from KGF-treated mice showed increased co-localization of GFP M. tuberculosis with LAMP-1 (Fig. 8, A  and B). The percentage of co-localized GFP M. tuberculosis was 28.9 Ϯ 3.8% in KGF-treated mouse AMs compared with 10.9 Ϯ 1.3% in saline-treated mouse AMs (p Ͻ 0.005) (Fig. 8C). We have previously found that lysates of AMs isolated from mice 1 h after in vivo KGF administration contained elevated GM-CSF levels, albeit lower than type II cell lysates from the same animals (19). In the prior report, we also found that type II cells but not AMs produced GM-CSF following in vitro exposure to KGF, a result that was not surprising given that the KGF receptor is known to be restricted to the epithelium. In order to assess the levels of GM-CSF that may be introduced into the phagolysosomal fusion assay, we harvested AMs by BAL 24 h after treating mice with saline or KGF and prepared lysates for GM-CSF quantitation by ELISA. GM-CSF levels associated with AMs from KGF-treated mice were ϳ3-fold higher than the levels associated with AMs from saline-treated mice (p Ͻ 0.01) (Fig. 9A). To determine the role of GM-CSF in the enhanced phagolysosomal fusion induced by KGF, we isolated AMs 24 h following saline or KGF treatment and performed the phagolysosomal fusion assay as described previously with the exception that isolated AMs were incubated with no antibody, with isotype antibody, or with anti-GM-CSF antibody prior to the addition of GFP-expressing M. tuberculosis. Phagolysosomal fusion was enhanced ϳ3-fold in AMs isolated from KGF-treated mice compared with those from saline-treated animals in the absence of the anti-GM-CSF antibody, whereas neutralization of GM-CSF abolished the KGF-induced enhancement of  phagolysosomal fusion (Fig. 9, B and C). At least one prior report has demonstrated an increase in GM-CSF expression in human AMs 24 h following in vitro M. tuberculosis infection (24). To investigate the possibility that M. tuberculosis was inducing GM-CSF expression in our system, we performed experiments in which we infected cultured AMs isolated from KGF-or saline-treated animals with GFP-expressing M. tuberculosis and assessed GM-CSF expression 24 h later in medium by ELISA and in cells by RT-PCR. We did not detect GM-CSF protein in the assay medium from either group, and although we did detect a very low level of gene expression in AMs from KGFand PBS-treated animals, there was no significant difference between the groups (not shown). Collectively, these results indicate that KGF promotes GM-CSF-dependent phagosome-lysosome fusion that, together with enhanced nitric-oxide production, promotes the killing of intracellular M. tuberculosis in AMs.

DISCUSSION
We found that KGF administered prior to M. tuberculosis infection or serially postinfection protects mice from weight loss and reduces the burden of M. tuberculosis bacilli in lung tissue (Figs. 2 and 3). KGF was effective in enhancing the pulmonary clearance of M. tuberculosis after a low dose inoculum that more closely mimics inhalational exposure (Fig. 4) and in attenuating mortality from a lethal inoculum of M. tuberculosis (Fig. 5). We also found that serial KGF administration every 3 days protected against weight loss and reduced bacillary growth, either when begun prior to infection or in mice with well established infection, and that withdrawal of KGF from M. tuberculosis-infected mice resulted in a decline in weight and increased M. tuberculosis burden (Fig. 3). The mechanisms of enhanced clearance of M. tuberculosis were GM-CSF-dependent and included augmentation of both nitric oxide production and phagosome-lysosome fusion (Figs. 6 -9). These findings suggest that administration of KGF (as adjunctive therapy) may be a plausible strategy for enhancing alveolar host defense during pulmonary M. tuberculosis infection.
Inhaled M. tuberculosis is readily phagocytosed by AMs and internalized into phagosomes (5). The bacteria are highly adapted to this intracellular niche and are able to subvert host  's t test). In other experiments, isolated AMs were allowed to adhere for 4 h to overnight in a tissue culture chamber slide and then incubated with no antibody, isotype antibody, or anti-GM-CSF antibody for 1 h. The AM monolayer was infected with GFP-expressing M. tuberculosis for 3 h and then washed to remove unbound M. tuberculosis bacilli. AMs were fixed, permeabilized, and stained with the lysosomal marker LAMP-1. Phagolysosomal fusion was assessed using a confocal microscope by quantifying the percentage of organelles exhibiting yellow staining consistent with co-localization of the green and red fluorophores (B). Representative images (ϫ1200 magnification) of AMs isolated from KGF-treated mice and assessed for phagolysosomal fusion in the presence of no antibody, of isotype antibody, or of anti-GM-CSF antibody are shown (C). Phagolysosomal fusion results are reported as mean Ϯ S.E. of triplicate samples and are representative of two independent experiments. **, p Ͻ 0.01 (one-way analysis of variance). DIC, differential interference contrast. defenses and proliferate intracellularly. Effective killing normally requires phagosome-lysosome fusion and a robust oxidative burst. These antimicrobial functions are regulated by a complex milieu of cytokines and chemokines that are produced by macrophages and lymphocytes in response to signals from the infected AMs (5). We have previously shown that KGF administration results in enhanced pulmonary clearance of Gram-negative bacteria through induction of GM-CSF and GM-CSF-mediated activation of phagocytic and oxidative mechanisms in AMs. (19) Given the central importance of GM-CSF in M. tuberculosis infection (25)(26)(27)(28)(29)(30), we postulated that KGF would also augment host defense against this intracellular bacterium.
KGF administration enhanced M. tuberculosis clearance and blocked weight loss for a short period even after a single prophylactic dose (Fig. 2). These data suggest that KGF can prime alveolar defenses to resist infection. Serial KGF treatment attenuated weight loss but did not completely block the growth of M. tuberculosis. Interruption of KGF treatment resulted in resumption of weight loss and mycobacterial growth after an interval of about 15 days. KGF treatment begun in mice with established infection resulted in weight gain and a fall in the bacterial burden after a delay of about 2 weeks (Fig. 3B).
Administration of GM-CSF has been shown to enhance killing of M. tuberculosis in mice (31,32). In addition, the absence of GM-CSF has been reported to increase the severity of M. tuberculosis infection in GM-CSF null animals (29). The use of GM-CSF as an adjuvant for the BCG vaccine has been shown to enhance protective immunity against TB (27,33).
Our data suggest that GM-CSF at least partially mediates KGF-induced M. tuberculosis clearance. Using in vitro co-culture methods, we demonstrated that KGF treatment of MLE-15 cells induced GM-CSF and enhanced the killing of M. tuberculosis in co-cultured RAW cells (Fig. 6). Conditioned medium from KGF-treated MLE-15 cells also reduced the growth of M. tuberculosis in RAW cells in a manner that was blocked by anti-GM-CSF antibody (Fig. 7). We provide evidence that augmented oxidative killing via nitric oxide-dependent pathways (Fig. 7) and phagosome-lysosome fusion (Fig. 8) are important mechanisms by which KGF enhances clearance of M. tuberculosis, with the caveat that substantial differences in nitric oxide production exist between human and murine macrophages (34) and between primary cells and cell lines (35). Phagolysosome fusion was enhanced in AMs isolated from KGF-treated but not PBS-treated mice and was inhibited in AMs infected with M. tuberculosis ex vivo by the addition of anti-GM-CSF antibody (Fig. 9, B and C). The source of GM-CSF driving phagolysosome fusion in this assay was also explored. We found that excess levels of GM-CSF were associated with AMs isolated from mice following KGF treatment (Fig. 9A). We did not find evidence that GM-CSF produced by M. tuberculosisinfected AMs during the course of the assay mediated the enhanced phagolysosomal fusion. These data support our prior contention that type II cells are the principal source of alveolar GM-CSF following KGF administration (19).
Effects of KGF that are potentially distinct from those of GM-CSF, such as those on alveolar epithelial antimicrobial protein expression (36), collectin expression (37), barrier integrity, per-meability, wound healing, and fluid transport (38), may be playing a role in KGF-enhanced survival and bacterial clearance following challenge with M. tuberculosis.
The Food and Drug Administration approved the use of KGF for chemotherapy-induced mucositis in December of 2004 (39). Clinical trials with KGF demonstrated reduced duration of oral lesions and salivary gland toxicity in patients with head and neck cancers treated with chemoradiation, diminished duration and severity of mucositis during bone marrow transplantation conditioning, and reduced incidence and severity of mucositis after chemotherapy for metastatic colorectal carcinoma (reviewed by Barasch et al. (40). Adverse effects were transient and mild to moderate in severity; the major side effects were skin flushing, temporary elevation of serum lipase and amylase, and hypersalivation. Phase I and II clinical trials have demonstrated that administration of KGF in doses up to 80 g/kg/day is well tolerated with minimal side effects, and a recent randomized controlled trial indicates that even single dose KGF before each chemotherapeutic cycle reduces the incidence and severity of mucositis. A recent study supports the concept that parenteral KGF induces GM-CSF production and AM phagocytosis, enhancing bronchoalveolar lavage fluid antimicrobial activity in healthy volunteers (41). Food and Drug Administration approval, commercial availability, and favorable side effect profiles enhance our enthusiasm for exploring KGF as a potential therapeutic agent for TB, perhaps as an adjunctive agent to treat latent TB or active infections caused by drugresistant strains, including multidrug-resistant or extremely drugresistant strains.
We conclude that KGF enhances the clearance of M. tuberculosis in prophylactic treatment and rescue murine models by a mechanism that includes paracrine activation of AMs by epithelium-derived GM-CSF and enhancement of nitric oxide production and phagosome-lysosome fusion. Future experiments are needed to determine whether KGF acts synergistically with conventional chemotherapeutic agents to eradicate pulmonary infections due to M. tuberculosis.