1α,25-Dihydroxyvitamin D3-induced Monocyte Antimycobacterial Activity Is Regulated by Phosphatidylinositol 3-Kinase and Mediated by the NADPH-dependent Phagocyte Oxidase*

We investigated the basis for the induction of monocyte antimycobacterial activity by 1α,25-dihydroxyvitamin D3 (D3). As expected, incubation of Mycobacterium tuberculosis-infected THP-1 cells or human peripheral blood, monocyte-derived macrophages with hormone resulted in the induction of antimycobacterial activity. This effect was significantly abrogated by pretreatment of cells with either of the phosphatidylinositol 3-kinase (PI 3-K) inhibitors, wortmannin or LY294002, or with antisense oligonucleotides to the p110 subunit of PI 3-Kα. Cells infected with M. tuberculosisalone or incubated with D3 alone produced little or undetectable amounts of superoxide anion (O⨪2). In contrast, exposure of M. tuberculosis-infected cells to D3 led to significant production of O⨪2, and this response was eliminated by either wortmannin, LY294002, or p110 antisense oligonucleotides. As was observed for PI 3-K inactivation, the reactive oxygen intermediate scavenger, 4-hydroxy-TEMPO, and degradative enzymes, polyethylene glycol coupled to either superoxide dismutase or catalase, also abrogated D3-induced antimycobacterial activity. Superoxide production by THP-1 cells in response to D3 required prior infection with liveM. tuberculosis, since exposure of cells to either killed M. tuberculosis or latex beads did not prime for an oxidative burst in response to subsequent hormone treatment. Consistent with these findings, redistribution of the cytosolic oxidase components p47 phox and p67 phox to the membrane fraction was observed in cells incubated with liveM. tuberculosis and D3 but not in response to combined treatment with heat-killed M. tuberculosis followed by D3. Redistribution of p47 phox and p67 phox to the membrane fraction in response to live M. tuberculosis and D3 was also abrogated under conditions where PI 3-K was inactivated. Taken together, these results indicate that D3-induced, human monocyte antimycobacterial activity is regulated by PI 3-K and mediated by the NADPH-dependent phagocyte oxidase.

either wortmannin, LY294002, or p110 antisense oligonucleotides. As was observed for PI 3-K inactivation, the reactive oxygen intermediate scavenger, 4-hydroxy-TEMPO, and degradative enzymes, polyethylene glycol coupled to either superoxide dismutase or catalase, also abrogated D 3 -induced antimycobacterial activity. Superoxide production by THP-1 cells in response to D 3 required prior infection with live M. tuberculosis, since exposure of cells to either killed M. tuberculosis or latex beads did not prime for an oxidative burst in response to subsequent hormone treatment. Consistent with these findings, redistribution of the cytosolic oxidase components p47 phox and p67 phox to the membrane fraction was observed in cells incubated with live M. tuberculosis and D 3 but not in response to combined treatment with heatkilled M. tuberculosis followed by D 3 . Redistribution of p47 phox and p67 phox to the membrane fraction in response to live M. tuberculosis and D 3 was also abrogated under conditions where PI 3-K was inactivated. Taken together, these results indicate that D 3 -induced, human monocyte antimycobacterial activity is regulated by PI 3-K and mediated by the NADPH-dependent phagocyte oxidase.
As the leading cause of death from any single bacterial infection in the world, tuberculosis represents a global health problem of paramount importance. Current estimates are that one-third of the world's population is infected with the etiological agent Mycobacterium tuberculosis (1) and that the incidence of new cases with active disease is anticipated to rise from the current 7 million per year to 10 million per year by 2015 (2). These statistics highlight the importance of developing new, more effective anti-tuberculous drugs, an effort dependent on acquiring more insights into host-pathogen interactions that determine the outcome of infection.
M. tuberculosis primarily infects mononuclear phagocytes, where it resides and multiplies within a host-derived phagosome (3). Its success as a pathogen is largely attributable to its ability to evade or resist the multiplicity of antimicrobial mechanisms available to this host cell. The macrophage is not only the primary site for M. tuberculosis growth but also ordinarily provides the primary line of host defense against invading pathogens in its role as an effector of innate immunity. Macrophages use varied strategies to kill and destroy invading organisms, including production of reactive nitrogen and oxygen intermediates, phagosome maturation and acidification, fusion with lysosomes, exposure to defensins and host cell apoptosis (4). Augmentation of any of these processes during macrophage activation may contribute to control of disease.
Vitamin D 3 is a steroid hormone that regulates several cellular and physiological responses. The classical mechanism of action of D 3 1 involves genomic signaling where hormone binds to the vitamin D receptor (VDR), a ligand-dependent transcription factor. The D 3 ⅐VDR complex then translocates to the nucleus, where it directly regulates transcription by binding to the vitamin D response element consensus sequence located upstream of D 3 -activated genes (5). Recently, nongenomic signaling in response to D 3 (i.e. cellular responses brought about independent of de novo transcription from a classical vitamin D response element) has been recognized to regulate important cellular processes (6 -10).
Vitamin D 3 is known to possess a variety of immunomodulatory properties including effects on both myeloid and lymphoid cells (11). Among these are its ability to induce the differentiation of immature myeloid cells into more mature monocytes and macrophages (12)(13)(14)(15)(16). Several lines of evidence indicate that D 3 regulates host resistance to M. tuberculosis. D 3 deficiency and vitamin D receptor polymorphisms have been linked to increased susceptibility to M. tuberculosis and Mycobacterium leprae (17)(18)(19). In this regard, D 3 production in vivo is promoted by exposure to ultraviolet light, and this may provide a link between exposure to sunlight and antimycobacterial resistance mechanisms (17,20). In addition, D 3 has been shown to activate mononuclear phagocyte antimycobacterial activity (20,21). Until now, the molecular basis through which D 3 regulates host resistance to M. tuberculosis has not been identified.
Class I phosphatidylinositol 3-kinase (PI 3-K) is a lipid kinase that phosphorylates the 3Ј-position of the inositol ring of phosphatidylinositol (PI) and its derivatives. It is composed of an 85-kDa Src homology 2 domain containing regulatory subunit (p85) and a 110-kDa catalytic subunit (p110). It is a multifunctional signaling molecule that has been implicated in a wide range of cellular processes including nuclear signaling, vesicle transport, organization of the cytoskeleton, cell growth, transformation, and survival (22)(23)(24). It was recently found that PI 3-K is required for D 3 -induced cell differentiation in the human macrophage cell lines THP-1 and U937 and in peripheral blood monocytes (6). In the latter study, vitamin D 3 was observed to activate PI 3-K, and D 3 -induced expression of the monocyte differentiation markers, CD14 and CD11b, required PI 3-K. Furthermore, D 3 treatment induced the formation of a vitamin D receptor⅐PI 3-K complex, representing a novel nongenomic signaling pathway activated by D 3 . In light of these results, this study examined whether PI 3-K is involved in regulating D 3 -induced antimycobacterial activity and investigated the mechanistic basis for this effector function. The results obtained show that D 3 -induced monocyte resistance to M. tuberculosis is regulated by PI 3-K and that this effect is due to activation of the phagocyte NADPH oxidase.
Cell Lines-The human promonocytic cell line THP-1 and murine macrophage-like cell line RAW264.7 were from the American Type Culture Collection. Cell lines were cultured in RPMI 1640 supplemented with 10% fetal calf serum (Hyclone), 2 mM L-glutamine and maintained between 2 and 10 ϫ 10 5 cells/ml. The human myeloid cell line THP-1 was used as a model for M. tuberculosis infection studies because of its similarity to human MDM (26) in M. tuberculosis infection models and its availability for use.
Isolation and Culture of Human Monocyte-derived Macrophages (MDM)-Peripheral blood mononuclear cells were isolated as described previously (27). Mononuclear cells were allowed to adhere for 45 min at 37°C in a humidified atmosphere with 5% CO 2 . Nonadherent cells were removed with three washes with HBSS. Adherent cells were maintained for 3 days at 37°C in a humidified atmosphere with 5% CO 2 prior to use for infections or treatments.
Infections-A virulent strain of M. tuberculosis (H37Rv) was grown to late log phase in Middlebrook 7H9 with OADC (Difco). Aliquots were frozen at Ϫ70°C, and representative samples were thawed and evaluated for colony-forming units (CFUs) on Middlebrook 7H10 agar with OADC (Difco Laboratories). Heat-killed bacteria were prepared by heating at 80°C for 2 h. Formaldehyde-fixed bacteria were prepared by fixation in 30% formaldehyde in methanol for 30 min. UV-irradiated bacteria were prepared by UV irradiation for 16 h. Each treatment reduced the viability of M. tuberculosis by greater than 5 logs. Cells were infected as described previously (28). Briefly, prior to infection, THP-1 cells were seeded at a density of 10 5 /cm 2 in either six-well flat bottom or 10-cm diameter tissue culture plates (Becton Dickinson, Franklin Lakes, NJ) and allowed to adhere and differentiate in the presence of 20 ng/ml PMA at 37°C in a 5% CO 2 humidified atmosphere for 24 h. Cells were washed, and medium without PMA was replenished 4 h prior to the addition of bacteria. Prior to infection, bacteria were opsonized with fresh serum as described previously (28). Monolayers were infected with opsonized M. tuberculosis at a 50:1 ratio. Latex beads were opsonized as for M. tuberculosis and incubated with cells at a ratio of 20:1. After 4 h, noningested bacilli or beads were removed by washing three times with HBSS, and medium was replenished. This resulted in infection of 80 -90% of cells with 1-5 bacteria/cell. Infection was evaluated by Kinyoun staining (BBL Microbiology Systems, Cockeysville, MD). For latex beads, ϳ90% of cells contained beads with 1-7 beads per cell.
Immediately after infection, macrophages were treated with inhibitors at final concentrations as follows: LY, 14 M; Wm, 50 nM, and L-NMMA, 8 M. Cells were incubated with these inhibitors for 15 min at 37°C, 5% CO 2 prior to the addition of D 3 . 4-Hydroxy-TEMPO (0.1 mM), PEG-SOD (100 units/ml), or PEG-cat (100 units/ml) were added during the 4-h infection, and reagents remaining in the culture medium were washed away from treated cells along with noningested bacteria. Heatinactivated PEG-SOD and PEG-cat were used as negative controls for the active enzymes. Vitamin D 3 (1 M) was added to monolayers after treatment with inhibitors and was left in the medium for the remainder of the experiment.
In Vitro PI 3-K Assays-In vitro PI 3-K assays were performed as described previously (6).
Determination of CFUs-Colony-forming units were determined as described previously (20). Bacilli were plated immediately after infection (time 0) and at 2 and 4 or 4 and 7 days after infection. Organisms were released in cold phosphate-buffered saline, 0.1% Triton X-100, serially diluted in Middlebrook 7H9, and 20 l of three dilutions were plated in triplicate on Middlebrook 7H10 agar. CFUs were counted after 14 days of incubation and maintained for 21 days to ensure that no additional CFUs became visible.
Sense and Antisense Phosphorothioate-modified Oligonucleotides-Phosphorothioate-modified oligonucleotides were prepared and incorporated into cells as described previously (6,29). Briefly, phosphorothioate-modified oligonucleotides to the ␣-isoform of the p110 subunit were synthesized and high pressure liquid chromatography-purified by Life Technologies, Inc. Oligonucleotides were phosphorothioate-modified to prevent intracellular degradation and were purified to remove incomplete synthesis products. 21-mers were produced to the human ␣ isoform of the p110 subunit of PI 3-K, including the presumed translation initiation site in both sense and antisense directions with the following sequences: sense (5Ј-ATG CCT CCA AGA CCA TCA TCA-3Ј) and antisense (5Ј-TGA TGA TGG TCT TGG AGG CAT-3Ј). THP-1 cells (5 ϫ 10 6 ) were resuspended in 500 l of RPMI containing 2.5% Lipo-fectAMINE (Life Technologies) and 5 M phosphorothioate-modified oligonucleotides and incubated on a rotary shaker for 4 h at 37°C prior to adherence and differentiation. NO 3 Ϫ Assays-Nitrite secreted by cells was measured using the Griess reagent according to the manufacturer's protocol (Sigma). 10% glycerol and incubated with rotation at 4°C for 1 h. The resulting suspension was filter-sterilized through a 0.22-m filter and contained membrane and membrane-associated proteins. 5Ј-nucleotidase and lactate dehydrogenase (LDH) assays were done to monitor for membrane contamination of cytosol and cytosolic contamination of membrane fractions, respectively.
5Ј-Nucleotidase Assay-5Ј-Nucleotidase activity was measured by the cleavage of phosphate from 5Ј-AMP (1.25 mM) in reaction buffer containing 10 mM HEPES, pH 7.4, 0.15 M NaCl, 2 mM KCl, 2 mM MgCl 2 for 15 min at room temperature. Liberated phosphate was measured by adding malachite green in 0.01% Tween 20 and allowing color to develop for 15 min at room temperature. The amount of phosphate released was determined from a standard curve.
LDH Assays-LDH assays were performed using an LDH assay kit from Sigma.

SDS-Polyacrylamide Gel Electrophoresis and Western
Immunoblotting-SDS-polyacrylamide gel electrophoresis was performed by the method of Laemmli et al. (31). Membranes were developed by ECL as described previously (32).
Apoptosis Assays-Apoptosis was evaluated using the TUNEL assay (Calbiochem) including controls provided with the kit and incubation with actinomycin D as a positive control.
Data Presentation and Statistical Analysis-Data in graphs are expressed as means Ϯ S.E. Statistical analyses for superoxide and nitrite assays were done by an unpaired Student's t test. Comparisons for CFUs were done by analysis of variance for each time point. Differences were considered significant at a level of p Ͻ 0.05.

D 3 -induced Antimycobacterial Activity in THP-1 Cells and Human Peripheral Blood MDM Is PI 3-Kinase-dependent-
Incubation of M. tuberculosis-infected THP-1 cells with the bioactive metabolite of vitamin D 3 resulted in the induction of antimycobacterial activity. In contrast, no effect was observed using either the D 3 precursor, 1-hydroxyvitamin D 3 (1-OH D3) or an inactive analog, 25-hydroxyvitamin D 3 (25-OH D3) (Fig.  1A). Whereas D 3 had no effect on the growth of M. tuberculosis in broth culture (data not shown), the number of CFUs recovered from THP-1 cells 4 and 7 days after treatment with D 3 was 36 Ϯ 3 and 20 Ϯ 4%, respectively, compared with untreated, control cells (Fig. 1A).
Vitamin D 3 has previously been shown to initiate a signaling pathway in human mononuclear phagocytes involving activation of inhibited PI 3-K activity in M. tuberculosis-infected, D 3 -treated THP-1 cells ( Fig. 2A, lanes 4 and 6 versus lane 2). Treatment with either LY or Wm also significantly reduced the antimycobacterial effect of D 3 at both day 4 and day 7 in THP-1 cells and day 2 and day 4 in human peripheral blood MDM (Fig. 2, B and C). The effect of LY or Wm was independent of effects on viability of the THP-1 cells. LDH activity was measured in culture supernatants of M. tuberculosis-infected, D 3 -treated cells 24 h after D 3 treatment. Cells were treated with LY, Wm, or diluent. LDH activity at 25°C was 124.4 units in diluenttreated or LY-treated culture supernatants and 125.9 units in Wm-treated culture supernatants. Analysis of apoptosis in similarly treated cells by TUNEL assay revealed 19.1% TUNELpositive cells in control samples, 18.5% in LY-treated cells, and 17.2% in Wm-treated cells.
Treatment with antisense, but not sense, oligonucleotides to the p110␣ subunit of PI 3-K eliminated the expression of this protein subunit in M. tuberculosis-infected, D 3 -treated THP-1 cells (Fig. 3A, lane A versus lane S). Further, treatment of M.
tuberculosis-infected THP-1 cells with p110 antisense oligonucleotides also significantly attenuated the antimycobacterial action of D 3 . At day 4, M. tuberculosis growth was restored from 37 Ϯ 3 to 78 Ϯ 18% of control cells, and at day 7 growth was restored from 21 Ϯ 2 to 77 Ϯ 9% of control cells (Fig. 3B). Treatment with antisense oligonucleotides to the p110 subunit of PI 3-K␣ did not alter either the infection rate or M. tuberculosis growth in non-D 3 -treated THP-1 cells (Fig. 3B). Furthermore, the reconstitution of CFUs by antisense to the p110␣ isoform was as effective as was the use of either nonspecific inhibitor, LY or Wm. Taken together, these findings suggest that the antimycobacterial action of D 3 is regulated by class I PI 3-K, p85/p110␣. infected THP-1 cells with L-NMMA did not reduce the antimycobacterial action of D 3 (Fig. 4A), and neither THP-1 cells nor human MDM produced nitrite in response to D 3 (data not shown). In contrast, D 3 -induced antimycobacterial activity in M. tuberculosis-infected, murine RAW 274.1 cells was significantly reduced by pretreatment with L-NMMA. Consistent with the implication that NO contributed to D 3 -induced antimycobacterial activity in murine cells, D 3 treatment of infected RAW cells resulted in significant secretion of nitrite (Fig. 4B), and this was markedly reduced in the presence of L-NMMA (Fig. 4B). Also, in contrast to human macrophages, PI 3-K inhibitors did not affect the induction of antimycobacterial activity in infected RAW cells (data not shown). These findings demonstrate that the antimycobacterial action of D 3 is NO-independent in human macrophages, whereas the converse is true in murine cells. Moreover, in the latter, this process is independent of any effects of PI 3-K. tuberculosis-infected, D 3 -treated THP-1 cells from 94 Ϯ 3 to 17 Ϯ 2 nmol/10 6 macrophages/h, while sense oligonucleotides had no effect (Fig. 5B). Similar results were obtained using MDM (Fig. 5C) terial Action of D 3 -To assess the role of the phagocyte oxidative burst in the antimycobacterial action of D 3 , 4-hydroxy-TEMPO, PEG-SOD, and PEG-cat were used to examine whether killing of M. tuberculosis was related to the production of reactive oxygen intermediates. Whereas none of these compounds alone affected the growth of M. tuberculosis in untreated cells, each of them significantly attenuated the ability of D 3 to induce resistance to M. tuberculosis in THP-1 cells and MDM (Fig. 7, A and B). Heat-inactivated PEG-SOD or PEG-cat, however, had no effect on D 3 -induced anti-mycobacterial activity (data not shown). These results indicate that removal of reactive oxygen intermediates, either by scavenging (4-hydroxy-TEMPO) or enzymatically (PEG-cat or PEG-SOD) significantly nullifies the antimycobacterial action of D 3 .

D 3 Induces the Association of Phagosome Oxidase Components with the Membrane Fraction in M. tuberculosis-infected
THP-1 Cells-Phagosome oxidase activity requires recruitment of the cytosolic components, p47 phox and p67 phox , to the membrane for assembly with flavocytochrome b 558 (33,34). The relative distribution of p47 phox and p67 phox in cytosolic and membrane fractions of stimulated cells was evaluated. Incubation of M. tuberculosis-infected cells with D 3 resulted in a marked translocation of both p47 phox and p67 phox to the membrane fraction (Fig. 8A). Untreated and PMA-differentiated THP-1 cells had similar amounts of membrane-associated p47 phox and p67 phox . Incubation of cells with live M. tuberculosis alone for either 2 or 4 h resulted in minimal changes in the membrane association of p47 phox and p67 phox ; however, subsequent treatment with D 3 increased these amounts to 7.6-and 8.1-fold, respectively, over resting cells (Fig. 8A, lane 5 versus lane 1). Exposure of cells to either heat-killed M. tuberculosis alone or followed by treatment with D 3 resulted in minimal or no detectable increases in membrane levels of oxidase components (Fig. 8A, lanes 6 and 7). Treatment of cells with either live or heat-killed M. tuberculosis alone or in combination with D 3 did not bring about detectable changes in cytosolic levels of either p47 phox or p67 phox (Fig. 8B), consistent with the relatively small fraction of p47 phox and p67 phox translocated during NADPH oxidase activation (33,34).

D 3 -induced Translocation of Cytosolic Oxidase Components to the Membrane Fraction Is PI 3-K-dependent-
The redistribution of p47 phox and p67 phox to the membrane fraction in response to infection with M. tuberculosis and treatment with D 3 was markedly reduced in cells pretreated with either of the PI 3-K inhibitors, LY or Wm (Fig. 8, C and D). In M. tuberculosis-infected cells treated with D 3 , the amount of p47 phox associated with the membrane increased 8.9-fold over control cells, and this was reduced to 2.4-and 3.2-fold, respectively, by treatment with LY either prior to infection or prior to D 3 treatment (Fig. 8C, upper panel, lanes 8 and 9 versus lane 4). Similarly, levels of p67 phox associated with the membrane increased 7.1-fold over PMA-differentiated cells, and this was reduced to 2.3-and 1.9-fold, respectively, by treatment with LY either prior to infection or prior to D 3 treatment (Fig. 8C, lower  panel, lanes 8 and 9 versus lane 4). The amount of p47 phox associated with the membrane increased 6.3-fold over PMAdifferentiated cells, and this was reduced to 1.2-fold by treatment with Wm either prior to infection or prior to D 3 treatment (Fig. 8D, upper panel, lanes 8 and 9 versus lane 4). The level of p67 phox associated with the membrane increased 9.3-fold over PMA-differentiated cells, and this was reduced to 2.3-and 2.1-fold, respectively, by treatment with Wm either prior to infection or prior to D 3 treatment (Fig. 8D, lower panel, lanes 8  and 9 versus lane 4). Priming

DISCUSSION
The steroid hormone vitamin D 3 has been linked to host resistance to tuberculosis since the mid-19th century (17), and the active metabolite, 1␣,25-dihydroxyvitamin D 3 has been shown to induce antimycobacterial activity in macrophages (20,21). In addition, it has been found that the antimycobacterial efficacy of cholecalciferol metabolites correlated with their binding affinities to the VDR (21). Other evidence for the role of D 3 in host resistance to tuberculosis has emerged recently. For example, D 3 deficiency has been linked to increased susceptibility to tuberculosis, and evidence suggests that vitamin D receptor polymorphisms confer differential resistance to mycobacterial diseases including leprosy and tuberculosis (18,19,35). Whereas it is clear from these observations that D 3 enhances host resistance to tuberculosis and other myco- bacterial diseases, the underlying mechanism accounting for D 3 -mediated antimycobacterial activity in human cells and its regulation have been obscure.
The findings of the present study indicate that monocyte antimycobacterial activity induced by 1␣,25-dihydroxyvita-min D 3 is mediated by the NADPH-dependent phagocyte oxidase and that this process is regulated by phosphatidylinositol 3-kinase. The conclusion that D 3 -induced antimycobacterial activity is due to activation of the phagocyte oxidative burst is based on three lines of evidence: (i) D 3 treatment of M. tuber- culosis-infected cells leads to a burst of O 2 . production (Fig. 5, A and C); (ii) the reactive oxygen intermediate scavenger, 4-hydroxy-TEMPO, and enzymes that degrade reactive oxygen intermediates (PEG-cat and PEG-SOD) dramatically reduce the D 3 -mediated antimycobacterial effect in both THP-1 cells and MDM (Fig. 7, A and B); and (iii) the phagosome oxidase components, p47 phox and p67 phox , translocate to the membrane fraction in M. tuberculosis-infected cells upon treatment with D 3 , as would be expected during assembly of a functional oxidase (Fig. 8A). Increased production of superoxide by infected cells in response to D 3 was rapid and transient, whereas we were only able to detect a decrease in M. tuberculosis CFUs at either 2, 4, or 7 days following D 3 treatment. This apparently delayed effect could reflect the production of low, undetectable levels of superoxide for longer periods of time. However, we believe a more likely possibility is that D 3 -induced antimycobacterial activity mediated by the NADPH oxidase acts synergistically by rendering M. tuberculosis more susceptible to other host factors that contribute to decreased CFUs observed over a longer time course. This model is consistent with the recent finding that NADPH oxidase-dependent killing of Salmonella in vivo was confined to the first few hours after phagocytosis. Reactive oxygen intermediates were essential to control infection but contributed at the very early stages of a multiphasic host response (36).
The conclusion that the oxidative burst is the principal mechanism by which D 3 mediates antimycobacterial activity in human mononuclear phagocytes is consistent with genetic evidence linking the phagocyte oxidase to host resistance to mycobacteria. Thus, patients with chronic granulomatous disease, a heterogeneous group of genetic disorders involving functional inactivation of the phagocyte oxidase (37), show increased susceptibility to recurrent and life-threatening bacterial and fungal infections including infections due to mycobacteria (38 -41).
Vitamin D 3 has been reported to stimulate directly human peripheral blood monocytes to generate an oxidative burst based upon the finding that it induced peripheral blood monocytes to secrete hydrogen peroxide (42). In the present study, either D 3 or M. tuberculosis acting alone triggered only modest oxidative burst activity. In respect to M. tuberculosis, this finding is consistent with previous reports showing only low levels of phagocyte oxidase activation in response to infection with mycobacteria (43,44). In contrast to the results obtained with either agent acting alone, in the presence of M. tuberculosis infection the oxidative burst in response to D 3 was markedly augmented. These findings indicate that M. tuberculosis infection acted to prime cells to respond to D 3 , and this priming was specific for viable M. tuberculosis, since it was not observed in cells that had ingested either dead bacilli or latex beads.
The precise mechanism for this priming effect by viable bacilli is unclear but probably involves the recruitment of one or more signaling pathways that act in conjunction with PI 3-K to prompt assembly of a functional oxidase complex. The finding that M. tuberculosis LAM was able to condition cells for an enhanced oxidative response to D 3 (Fig. 9) provides a potential mechanism to account for the priming observed in response to infection with live M. tuberculosis. LAM is known to have pleiotropic effects including effects on multiple signaling pathways in macrophages (45)(46)(47). The requirement for infection with live bacilli may reflect a need for de novo synthesis of LAM within cells, for the intracellular release and trafficking of LAM, or both. At the same time, although it was possible to show that LAM had the property of oxidase priming, these results do not establish that it is the actual in vivo priming factor or that it is the only priming factor elaborated by viable M. tuberculosis within cells.
In activated murine macrophages, antimycobacterial activity has been shown to be due, at least in part, to the inducible  synthesis of nitric oxide, and studies in inducible nitric oxide synthase knock out mice have suggested that this is an effector arm of the host response to M. tuberculosis (48,49). In the experiments reported above, detectable levels of nitrite were not produced by D 3 -treated, M. tuberculosis-infected THP-1 cells or MDM, and the NO inhibitor L-NMMA did not affect the induction of antimycobacterial activity by D 3 in THP-1 cells (Fig. 4A). This was in direct contrast to parallel observations made using the murine macrophage-like cell line RAW 274.1, which when infected with M. tuberculosis produced significant amounts of nitrite in response to D 3 (Fig. 4B) and in which the antimycobacterial effect of D 3 was markedly attenuated by prior treatment with L-NMMA (Fig. 4A). These results suggest that at least two independent mechanisms operate to induce antimycobacterial activity in D 3 -treated macrophages. In murine cells, inducible nitric oxide synthase operates independently of PI 3-K, whereas in human macrophages the NADPHdependent phagocyte oxidative burst appears to be predominant and regulated by PI 3-K. Moreover, these findings are consistent with a recent report that examined the mechanisms involved in the antimycobacterial activity of toll-like receptor 2-activated macrophages (50). In these studies, TLR-2-activated murine macrophages demonstrated NO-dependent antimycobacterial activity, whereas TLR-2-activated antimycobacterial activity in human macrophages was NO-independent. No effector mechanism for TLR-2-activated antimycobacterial activity was defined in human cells (50). Our results presented here suggest that one possible effector mechanism is oxygendependent killing via activation of the phagocyte oxidase. This model is supported further by the findings that, when ligandactivated, the D 3 receptor (6) and TLR-2 (51) have in common the property of binding to PI 3-K. As we have shown, in M. tuberculosis-infected, D 3 -treated cells, this results in activation of the phagocyte oxidase. It seems highly likely that this would occur as well in TLR-2-activated cells infected with mycobacteria.
Recently, there has been considerable interest in whether macrophage apoptosis is involved in the host response to mycobacteria and, if so, whether apoptosis could affect mycobacterial viability. Although results from various studies have been inconsistent (52-57), we nevertheless considered the possibility that D 3 -induced antimycobacterial activity could be due to the induction of apoptosis. Rates of apoptosis in THP-1 cells incubated with D 3 alone for either 1 or 4 days were at background levels (1-2%). In cells infected with M. tuberculosis alone, apoptosis was detected at frequencies of 13 and 15%, respectively, at 1 and 4 days postinfection, and these rates were not affected by the addition of D 3 . Similar results were found using human MDM. Treatment with D 3 alone for 1 or 4 days resulted in levels of apoptosis similar to untreated cells (4 or 6%, respectively). Cells infected with M. tuberculosis alone showed apoptosis in 17 and 20% of cells at 1 and 4 days postinfection, and these rates were not affected by the addition of vitamin D 3 . These findings indicate that apoptosis is unlikely to be involved in the induction of antimycobacterial activity by D 3 .
The second principal conclusion drawn from this study is that PI 3-K regulates the antimycobacterial action of D 3 . Several lines of evidence support this argument including the following: (i) PI 3-K is activated by D 3 in THP-1 cells (Fig. 1B; see Ref. 6); (ii) D 3 -induced antimycobacterial activity is abrogated by PI 3-K inhibitors LY and Wm in both THP-1 cells and MDM independent of any cytotoxic effect on the cells (Fig. 2, B and C); (iii) D 3 -induced antimycobacterial activity is abrogated by antisense mRNA to class I PI 3-K (Fig. 3B); (iv) O 2 . production by M. tuberculosis-infected THP-1 cells and MDM in re-sponse to D 3 is inhibited by PI 3-K inhibitors LY and Wm (Fig.  5, A and C); (v) D 3 -induced O 2 . production by M. tuberculosisinfected THP-1 cells is inhibited by antisense mRNA to class I PI 3-K (Fig. 5B); and (vi) pretreatment of THP-1 cells with the PI 3-K inhibitors LY or Wm prevents p47 phox and p67 phox translocation to the membrane fraction upon D 3 treatment (Fig. 8, C and D).
The finding that the antimycobacterial action of D 3 is regulated by PI 3-K represents another important example of nongenomic signaling by this and other steroid hormones. The long standing paradigm for D 3 action has been recognized to involve genomic signaling in which hormone binds to the VDR. The D 3 ⅐VDR complex then translocates to the nucleus, where the ligand-activated transcription factor VDR directly regulates transcription by binding to vitamin D response elements in D 3 -activated genes (5). However, numerous reports have shown that D 3 also acts via nongenomic signaling, where cellular responses are brought about independent of de novo transcription from a classical vitamin D response element (58 -60). Moreover, it was recently shown that D 3 treatment of human monocytes results in the rapid activation of PI 3-K leading to monocyte differentiation (6). In the latter report, a novel nongenomic mechanism of action of D 3 was shown to be regulated by a VDR⅐PI 3-K signaling complex. Similar findings were also recently reported for signaling by estrogen, where hormone treatment induced the formation of a signaling complex including the estrogen receptor-␣ and the p85␣ regulatory subunit of PI 3-K (61). PI 3-K has been shown to regulate a wide range of physiological processes including movement of organelle membranes, cytoskeletal rearrangement, cell proliferation, apoptosis, etc. (23,63). Importantly, PI 3-K has also been shown to regulate the oxidative burst. For example, PI 3-K has been shown to be required for triggering the NADPH oxidase in response to particulate stimuli through activating F c ␣ and F c ␥ receptors in neutrophils (64). Similarly, based upon effects of the PI 3-K inhibitor wortmannin, a role for this enzyme in regulating the production of superoxide by human neutrophils in response to soluble stimuli such as fMLP (65) and PMA (62) has also been suggested. In the present study, PI 3-K was found to be required to trigger the oxidative burst in response to D 3 in M. tuberculosis-infected cells. However, it appears that activation of PI 3-kinase alone is insufficient to bring about full oxidase activation, since treatment of THP-1 cells with D 3 alone activates PI 3-K (Fig. 1B, lanes 6, 8, and 10 versus lanes 5, 7, and  9; see Ref. 6), but at best this elicits only modest superoxide production (Fig. 5, A-C). In contrast, prior infection of cells with live M. tuberculosis primed cells for an enhanced D 3induced oxidative burst, and this was not recapitulated by either latex beads or killed M. tuberculosis (Fig. 6). Consistent with these findings, only in cells infected with live bacilli did D 3 induce translocation of the oxidase components p47 phox and p67 phox to the membrane fraction for oxidase assembly (Fig. 8). These findings suggest that live M. tuberculosis or some factor unique to the phagosome containing viable M. tuberculosis potentiates phagocytes to undergo a vigorous oxidative burst in response to D 3 .
In summary, this report identifies a key mechanism of D 3induced host defense against tuberculosis and a key regulatory pathway required to bring about this effector mechanism. The findings demonstrate that the antimycobacterial activity of D 3 in human macrophages is due to activation of the NADPH oxidase and the production of reactive oxygen intermediates. Furthermore, they show that PI 3-K is necessary but not sufficient for assembly of a functional oxidase in response to D 3 in M. tuberculosis-infected cells. This represents another novel example of nongenomic signaling by vitamin D 3 . Ongoing studies concerned with the basis for M. tuberculosis priming and the additional regulatory pathways involved should provide additional insight into host resistance to mycobacterial disease and the associated inflammatory consequences.