GLI2 Transcription Factor Mediates Cytokine Cross-talk in the Tumor Microenvironment*

Tumor cells interact with their surrounding microenvironment to survive and persist within the host. Cytokines play a key role in regulating this crosstalk between malignant cells and surrounding cells in the microenvironment. Although this phenomenon is clearly established, the molecular mechanisms mediating this cellular event remain elusive. Here, using as a model bone marrow stromal cells, we describe a novel signaling mechanism initiated by CCL5 in these cells leading to up-regulation of immunoglobulin secretion by malignant B cells. CCL5 increases IL-6 expression and secretion in bone marrow stromal cells. IL-6 in turn induces Ig secretion by malignant B cells. Analysis of the mechanism reveals that CCL5 signaling induces GLI2 through a PI3K-AKT-IκBα-p65 pathway and requires GLI2 transcriptional activity to modulate IL-6 expression and Ig secretion in vitro and in vivo. Together, these results identify a novel signaling pathway mediating the stromal-cancer cell interactions, leading to increased Ig production by malignant cells.

The tumor microenvironment plays an important role in the development and maintenance of cancer cells. Because the microenvironment is an integral part of cancer cell function, it is difficult to dissociate it from cancer cells. Malignant cells within each organ/compartment depend on signaling from the surrounding microenvironment to survive and proliferate within the host. This interaction has been shown to be important in numerous malignancies and can promote, for instance, resistance to chemotherapeutic drugs (1,2). One such interaction is that of the malignant B cells and bone marrow stromal cells. Malignant B cells at various stages of development depend on signaling from the surrounding microenvironment to survive and proliferate within the host and promote chemoresistance (3)(4)(5)(6). Bone marrow stromal cells provide malignant B cells with external signals such as soluble growth factors, anti-gens, cytokines, and cell-cell interactions (7)(8)(9)(10)(11). Therefore, a better understanding of the crosstalk between malignant cells and stromal cells will be essential for the development of new therapies targeting B cell malignancies.
Here, using Waldenström macroglobulinemia (WM) 3 bone marrow stromal cells as a model to study the mechanisms regulating Ig secretion by malignant cells, we describe a novel pathway initiated by the cytokine CCL5 in stromal cells leading to up-regulation of Ig production by malignant B cells. We demonstrate that CCL5 promotes IL-6 expression and secretion in stromal cells by regulating the GLI2 transcription factor via a CCR3/PI3K-AKT-IB␣-p65 pathway. GLI2 is an effector of the cellular function of the Hedgehog pathway, a well established oncogenic pathway in numerous neoplasms (12)(13)(14). Interestingly, characterization of this regulatory mechanism shows that GLI2 does not require an active Hedgehog pathway to modulate the expression of IL-6. Using a combination of reporter and chromatin immunoprecipitation assays, we demonstrate that GLI2 binds to and activates the IL-6 promoter in stromal cells. Increased IL-6 present within the tumor microenvironment binds to the IL-6 receptor on malignant B cells and activates signaling cascades, leading to an increased Ig secretion by malignant B cells in vitro and in vivo. Moreover, we demonstrate that this regulatory mechanism is not exclusive to WM but is rather a more general event present in other B cell malignancies such as multiple myeloma (MM) and monoclonal gammopathy of undetermined significance (MGUS). Taken together, our results identify a novel CCL5-GLI2-IL-6 axis in the stromal microenvironment regulating Ig secretion by malignant cells. Therefore targeting this novel axis may provide a therapeutic benefit to patients with malignancies associated with increased Ig production.

EXPERIMENTAL PROCEDURES
Cell Lines and Patient Samples-The IgM-secreting cell line BCWM.1 (15)(16)(17)(18) was a kind gift from Dr. Steve Treon (Dana Farber Cancer Institute, Boston, MA). HS-5 stromal cells were obtained from the ATCC. The Saka T (19) stromal cell line (referred to here as Saka) was a kind gift from Dr. David Roodman (University of Pittsburgh, PA). Primary cells were isolated from patients who provided written informed consent. This study was approved by the Mayo Clinic Institutional Review Board. Bone marrow aspirates from WM, MGUS, and MM were used to generate mesenchymal stromal cells (referred to here as stromal cells (sc)). For B cell isolation, total bone marrow cells were used to isolate CD19ϩ and CD138ϩ cells concurrently by positive selection. Cells were then counted and used for experiments as indicated in the manuscript. 20 ϫ 10 6 negative cells/100 mm tissue culture plates were cultured in RPMI containing 10% FBS and antibiotics. Non-adherent cells were washed off after 3 days, and adherent mesenchymal stromal cells were cultured until monolayers were confluent. Primary stromal cell cultures were characterized by FACS and microscopy analysis and found to express the established stromal markers, as reported previously (20 -22) (CD54, CD59, CD105, and CD146), and negative expression of lymphoid and myeloid markers (CD45, CD16, CD18, CD33, CD38, CD40L, and CD14). Optical analysis demonstrated that stromal cells have the characteristic fibroblast-like morphology. Multiple plates were generated for each patient used, depending on number of cells available after sorting for B cells. It is important to point out that these cells were not used beyond passage 2 and maintained the phenotype throughout the length of the study.
Each cell line was optimized for transfection efficiency using a plasmid expressing GFP to determine transfection conditions. HS-5 cells were transfected by electroporation at 280 V for 25 ms (BTX, Harvard Apparatus, Holliston, MA). BCWM.1 cells were electroporated at 240 V for 25 ms. Primary WM stromal cells were transfected using Lipofectamine (Invitrogen) following the manufacturer's recommendations. All transfections were done for 36 h unless otherwise noted. For electroporation experiments, 4 ϫ 10 6 cells were placed in each cuvette. For each treatment, 2 g of IL-6 promoter reporter and 10 g of 8XGLI reporter were used. For overexpression experiments, 5 g of CCR1 or CCR3 and 18 g of GLI2 were used. For knockdown of expression, 18 g of shGLI2 was used. For transfection of multiple constructs, the same concentrations were used, except for GLI2 rescue (Fig. 2C), where cells were transfected with 2 g of IL-6 reporter, 18 g of shGLI2, and 2 g of GLI2. For CCR3 rescue experiments, 5 g of CCR3 or empty vector were transfected into stable cell lines. For GLI2 bypass experiments, 18 g of GLI2 were transfected into stable cell lines. In all experiments, equal amounts of plasmid DNA was added by adding empty vectors to appropriate controls. For primary patient stromal cells, 50 -100 ϫ 10 5 cells were transfected with 0.5 g of IL-6 promoter reporter construct in triplicate wells.
Stromal cells were transduced with lentiviral particles following the manufacturer's protocol. After 48 h, cells were counted and used to set up experiments as outlined below.
Lentivirus Infection and Coculture Experiments-To determine the effect of GLI2 knockdown on IL-6 and IgM secretion in a coculture system, 0.1 ϫ 10 6 lentivirus-infected cells were plated in triplicate wells and cocultured with 0.5 ϫ 10 6 serumstarved BCWM.1 cells in 24-well plates in RPMI containing 0.5% BSA. After 2 days in coculture, supernatants were harvested and used to determine the levels of IL-6 and IgM by ELISA. For HS-5 and Saka cells, 0.05 ϫ 10 6 cells were plated in triplicate wells in 24-well plates, either alone or in the presence of 0.5 ϫ 10 6 BCWM.1 cells, for 48 h. Supernatants were then harvested and used to determine the levels of IL-6 and IgM in the culture supernatants by ELISA.
Luciferase Assay-Cells were grown and transfected as indicated above. For luciferase reporter assays, 2 ϫ 10 6 cells were plated in triplicate in 6-well plates in medium containing 10% FBS for 36 h. Samples were harvested and prepared for luciferase assays following the manufacturer protocol (Promega, Madison, WI). To control for intersample variations in transfection efficiency, the total protein for the samples in each well was quantitated using the Bio-Rad protein assay, and luciferase readouts were normalized to protein content. Relative luciferase represents luciferase readouts/protein concentration normalized to control cells within each experiment.
ELISA-ELISA plates (Nunc Maxisorp, Nalge Nunc International, Rochester, NY) were used to quantitate IL-6 and IgM levels. IL-6 levels were quantitated using a human IL-6 ELISA (R&D Systems), following the manufacturer's recommendations. IgM levels were quantitated using a human IgM ELISA (Bethyl Laboratories, Inc., Montgomery, TX), following the manufacturer's recommendations. For both ELISA kits, plates were developed with Turbo TMB-ELISA (Thermo Scientific, Rockford, IL). The reaction was stopped by addition of 1 N H 2 SO 4 , and results were measured with a plate reader (Molecular Devices, Palo Alto, CA) and analyzed using SoftMax Pro 5.2 software.
Semiquantitative RT-PCR-Total RNA was extracted using TRIzol reagent (Invitrogen). A total of 1-2 g of RNA was reverse-transcribed using SuperScript III reverse transcriptase (Invitrogen). A portion of the total cDNA was amplified by PCR using 94°C denaturation, 57°C annealing, and 72°C extension temperatures. Positive and negative strand primers and the number of cycles used for amplification of each mRNA species were as follows: IL-6: 30 cycles, TGACAAACAAATTCGGTA-CATCC and AATCTGAGGTGCCCATGCTAC; GAPDH: 27 cycles, GACCTGACCTGCCGTCTAGAAAAA and ACCAC-CCTGTTGCTGTAGCCAAAT; GLI2: 35 cycles, CAAGGAT-TCCTGCTCATGGG and AGTGGCTGCCGCGTACTT; SMO: GAGAGTTCTGGATGTCTGGCTCA and ACTCTG-GGAACTGTCACCTCTGC; and CCR3: GGAGGCATTTCC-ACACTCTG and ATCTGCCCAGGTGCATGAG, respectively. Amplified products were visualized under UV illumination following electrophoresis on ethidium bromidestained-agarose gels. Amplification of the appropriate gene fragments was assured by comparison with molecular weight markers run on the same gel.
ChIP Assay-HS-5 (5 ϫ 10 6 ) cells were cross-linked with 1% formaldehyde for 15 min at 25°C, harvested in radioimmune precipitation assay lysis buffer (150 mM NaCl, 1.0% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 50 mM Tris-Cl) and sheared to fragment DNA to ϳ 500 bp. Samples were then immunoprecipitated using a GLI2 antibody (R&D Systems) or agarose beads alone at 4°C overnight. Following immunoprecipitation, samples were washed and eluted using the chromatin immunoprecipitation kit (R&D Systems) following the manufacturer's instructions. Cross-links were removed at 65°C for 4 h, and immunoprecipitated DNA was purified (Qiagen, Valencia, CA) and subsequently amplified by PCR. PCR was done using three primer sets for the three areas containing GLI binding sites in the IL-6 promoter sequence ( Fig. 2A): Sense (1), TAGTCTAG-AGCCCATTTGCATGAG; and antisense TTCAGGGCAGA-AAGGGGGAGAATA. Sense (2), ATTCCCAAGGGGTCAC-TTGGGAGA; and antisense ACTGAGTTTCCTCTGACTC-CATCG. Sense, (3) CACTTTTCCCCCTAGTTGTGTCTT; and antisense GGATTTCCTGCACTTACTTGTGGA. PCR products were visualized by 2% agarose gel. To determine p65 binding to the GLI2 promoter, 5 ϫ 10 6 HS-5 cells were lysed and treated as indicated above. Samples were immunoprecipitated as indicated previously using a p65 antibody (Santa Cruz Biotechnology) or agarose beads. PCR was done using three primer sets for the three areas containing p65 binding sites on the GLI2 promoter sequence ( Immunoblotting-To determine protein levels of GLI2, 0.5 ϫ 10 6 HS-5 cells were serum-starved overnight and then treated with or without CCL5. After 24 h, cells were lysed and analyzed by SDS-PAGE. For signaling studies in the presence or absence of CCL5, cells were plated as indicated previously, serumstarved overnight, and then treated with CCL5. At the indicated times after treatment, cells were lysed, and lysates were used to determine the phosphorylation of Akt. In Vivo Experiments-Six 8-week old athymic nude mice were obtained from Harlan Laboratories (Indianapolis, IN). Mice were irradiated with 2 gray and allowed to rest overnight. Mice were then subcutaneously injected with 5 ϫ 10 6 BCWM.1 and 0.5 ϫ 10 6 lentivirus-infected HS-5 cells (containing either shGLI2 or scramble). BCWM.1/HS-5 cells were premixed prior to injection, spun, and then reconstituted in 100 l PBS/mouse. This experiment was done three times. All experiments were terminated after 14 days. Tumor sizes were measured, and tumors were harvested and immediately fixed to stain for IgM by immunohistochemistry. Blood was also collected from individual mice into serum-separation tubes and spun to collect serum for determining the IgM levels in mice sera. In addition, we utilized a second model using severe combined immunodeficient (SCID) mice (Charles River Laboratories, Inc., Wilmington, MA), which were subcutaneously implanted with human fetal bone (Applied Bioscience Resources, Alameda, CA) as described previously (15,25,26). Briefly, human fibula was cut into 3-mm fragments and implanted into SCID mice subcutaneously. Two days later, mice were subcutaneously injected with 5 ϫ 10 6 BCWM.1 and 0.5 ϫ 10 6 lentivirus-infected HS-5 cells (containing either shGLI2 or scramble) immediately next to the bone. After 2 weeks, blood was also collected from individual mice into serum-separation tubes and spun to collect serum to use to determine the IgM levels in mice sera.
Immunohistochemical Staining of Mouse Tumors-Mouse tumors were fixed in formalin (10%) and then paraffin-embedded. Paraffin-embedded specimens were cut into 4-m sections. The sections were deparaffinized in xylene and rehydrated with water through a graded alcohol series. The sections were pretreated with 10 mM citrate buffer (pH 6.0) for 30 min, cooled for 5 min, and then rinsed well with cold running tap water. To block endogenous peroxidase activity, sections were incubated for 10 min in methanol/H 2 O 2 and then rinsed with tap water. To block nonspecific staining, sections were incubated with PBS with 5% human serum for 10 min. Anti-human IgM (Dako, Carpinteria, CA) (10 g/ml) were added to the slides for 30 min. The slides were rinsed with APK wash solution (Ventana Medical Systems, Inc., Tuscon, AZ), incubated with biotinylated goat anti-rabbit IgG (1:300) for 15 min, and rinsed with APK. SA-HRP (1:300) was added to all slides for 15 min, followed by rinsing with APK. Staining was visualized using 3,3Ј-diaminobenzidine (Dako Cytomation) and counterstaining with hematoxylin. The slides were coverslipped and mounted with Cytoseal 280 (Stephens Scientific, Kalamazoo, MI). All slides were observed with light microscopy (Olympus AX70, 200ϫ aperture 0.46, 400ϫ aperture 0.75, 600ϫ aperture 0.80; Olympus America, Melville, NY) with images being captured with a SPOT RT camera and software (Diagnostic Instruments, Burlingame, CA). A, IL-6 secretion by serum-starved stromal cells from HS-5 (0.01 ϫ 10 6 /well), Saka, or WM patients (0.1 ϫ 10 6 /well, P1-P3) were cultured in the presence or absence of CCL5 (500 ng/ml) for 24 h. B, relative changes in luciferase activity in HS-5 stromal cells transfected with IL-6 promoter-luciferase reporter. Cells were allowed to rest overnight in the absence of serum and then incubated in presence or absence of CCL5 (500 ng/ml) for 12 h. A similar set of cells were treated with CCL5 (500 ng/ml) for 24 h, and RNA was subsequently extracted. IL-6 and GAPDH gene expression was analyzed by RT-PCR as described under "Experimental Procedures." C, relative changes in luciferase activity in HS-5 or Saka cells transfected with IL-6 promoter-luciferase reporter and either a control vector or CCR3 expression vector for 36 h. A similar set of cells was used to determine expression levels by RT-PCR. D, HS-5 and Saka cells with stable knockdown of CCR3 were generated as described under "Experimental Procedures." Cells were then transfected with IL-6 promoter-luciferase reporter for 36 h. RNA was isolated from cells stably expressing shCCR3 or scramble (Scr) and used to determine expression by RT-PCR. E, relative changes in luciferase activity in HS-5 and Saka stromal cells stably transfected with either Scr or shCCR3 and then transfected with IL-6 promoter-Luciferase reporter and Ϯ the CCR3 expression construct. Bars represent the mean Ϯ S.E. of triplicate experimental wells.  Statistical Analysis-Comparisons between groups were based on 2 tests for nominal variables. The Wilcoxon ranksum test or the Kruskal-Wallis test was used for continuous variables. For all statistical tests, p Ͻ 0.05 was considered significant. Analysis was performed on Statview software (SAS Institute, Inc., Cary, NC).

CCL5 Modulates IL-6 Expression and Secretion in WM Stromal Cells-CCL5
is a known cytokine serving as an autocrine or paracrine growth factor for tumor and stromal cells, and is a mediator of the complex networks existing between tumor cells and the stromal microenvironment (27)(28)(29)(30). However, the molecular mechanism underlying this cellular event remains  (2)) or scramble (Scr) shRNA control in the presence or absence of the GLI2 expression construct. A similar set of cells was used to determine the expression levels by RT-PCR. D, effect of GLI2 knockdown (shGLI2) on the viability (left) and proliferation (right) of HS-5 stromal cells infected with lentivirus. Cell viability and proliferation were assessed by annexin V/propidium iodide staining and thymidine incorporation, respectively, as described under "Experimental Procedures." Cell viability and proliferation were normalized to Scr. E, HS-5 stromal cells were lysed, and a chromatin immunoprecipitation (ChIP) assay was performed. PCR was done using three primer sets for the three areas containing GLI binding sites in the IL-6 promoter sequence (region 1, region 2, and region 3). F, relative changes in luciferase activity in HS-5 stromal cells transfected with the IL-6 promoter-luciferase reporter and either Scr or shGLI2 in the presence or absence of the CCR3 expression vector. G, HS-5 and Saka cells stably expressing Scr or shCCR3 were transfected with the IL-6 promoter-luciferase reporter and either GLI2 or the empty vector for 36 h. All experiments were done in triplicate wells and repeated at least three times with similar results. Bars represent mean Ϯ S.E. elusive. Here, using WM primary stromal cells (WMsc) and stromal cell lines HS-5 and Saka, we demonstrated that CCL5 stimulation resulted in an increase in IL-6 secretion by stromal cells (Fig. 1A). IL-6 is a proinflammatory cytokine shown to be important for the biology of normal and malignant B cells (10,(31)(32)(33)(34)(35)(36)(37)(38)(39). Characterization of this phenomenon revealed that CCL5-induced IL-6 secretion was accompanied by an increase in IL-6 mRNA levels and promoter activity (Fig. 1B). CCL5 mediates its cellular functions via three known receptors (CCR1, CCR3, and CCR5). All the studied stromal cells express CCR1 and CCR3 receptors but not CCR5 (data no shown). Interestingly, activation of the CCL5 signaling cascade via overexpression of CCR3 (Fig. 1C) but not CCR1 (data not shown) increased IL-6 promoter activity in stromal cells. Conversely, stable knockdown of CCR3 expression led to a reduction in IL-6 promoter activity and expression in HS-5 and Saka cells (Fig.  1D). These results were validated by rescuing the knockdown using an RNAi-resistant CCR3 cDNA (Fig. 1E). Together, these results support a novel role for CCL5-CCR3 pathway in the modulation of IL-6 in stromal cells and suggest the transcriptional activation of its promoter as a potential regulatory mechanism for IL-6 expression.
CCL5 Requires an Intact GLI2 Transcriptional Activity to Modulate IL-6 Expression-Transcription factors are essential mediators of signaling-induced gene expression. Bioinformatics analysis identified GLI proteins (GLI1, 2, and 3) as candidate mediators of CCL5 activation of the IL-6 promoter in stromal cells ( Fig. 2A). These transcription factors are known effectors of the Hedgehog pathway and essential for regulation of the cellular functions mediated by this signaling cascade (14,40,41). The IL-6 promoter contains eight candidate binding sites for GLI transcription factors within the Ϫ1267 bp upstream of the transcriptional starting site ( Fig. 2A). To determine a possible involvement of GLI proteins in CCL5-IL-6 interplay, we used a combination of luciferase assays and shRNA approaches and found that GLI2 is the mediator of CCL5-IL-6 interplay in stromal cells. GLI2 is expressed in WMsc, HS-5, and Saka cells (Fig. 2B). Using two different shRNA vectors to target GLI2, we found that GLI2 knockdown resulted in a reduction in IL-6 promoter activity (Fig. 2C). These results were validated using a rescue of the knockdown using a GLI2 RNAi-resistant cDNA (Fig. 2C). The effect of GLI2 knockdown on the regulation of IL-6 was not due to an effect in growth or survival of stromal cells. GLI2 knockdown did not affect stromal cell growth or survival (Fig. 2D). Using ChIP, we confirm binding of endogenous GLI2 to the IL-6 promoter. Fig. 2E shows GLI2 occupancy of this regulatory sequence in HS-5 cells, thus supporting the fact that IL-6 is a novel direct target of the GLI2 transcription factor.
To confirm the requirement of GLI2 in CCL5-CCR3 axismediated IL-6 activation, stromal cells were cotransfected with  (2)) or the Scr control. Cells were then transfected with the IL-6 promoter-luciferase reporter, and 36 h post-transfection, luciferase activity was determined. A similar set of cells were used to determine SMO expression by RT-PCR. All experiments were done in triplicate wells, and bars represent average Ϯ S.E. C, HS-5 cells (1 ϫ 10 6 cells/well) were plated in 6-well plates and allowed to adhere overnight. Cells were serum-starved overnight and treated with CCL5 (500 ng/ml) for 6 or 24 h, and RNA was extracted and used for GLI2 expression by RT-PCR. D, HS-5 cells were plated (1 ϫ 10 6 cells/well) in a 6-well plate, allowed to adhere overnight, serum-starved, and then treated with CCL5 for 24 h. Lysates were used to determine GLI2 and ␤-actin expression by immunoblotting. E, WMsc were serum-starved overnight and stimulated with CCL5. At the indicated time points, RNA was extracted, and RT-PCR was done using GLI2-and GAPDH-specific primers. F, HS-5 and Saka cells were transfected with a CCR3 expression construct or the empty vector for 36 h in the presence of serum. RNA was extracted and used to determine GLI2 expression by RT-PCR. G, RNA was generated from HS-5 and Saka cells stably expressing Scr or shCCR3 and used to determine GLI2 expression by RT-PCR. . After 36 h, GLI2 expression was determined by RT-PCR. E, bioinformatics analysis of the GLI2 promoter identifies candidate p65 (p) binding sites. Gray lines represent amplicons for the three primer sets used that span the different p65 binding sites. Forward and reverse primer start sequences are indicated. TSS, transcription start site. F, lysates from HS-5 stromal cells were chromatin-immunoprecipitated (ChIP) with an Ig control or p65 antibody, DNA was purified, and PCR was done using primers for areas containing p65 binding sites in the GLI2 promoter sequence as described under "Experimental Procedures." G, HS-5 cells were transfected with either the p65 expression construct or the empty vector (Ctrl). After 36 h, GLI2 expression was determined by RT-PCR. the CCR3 expression vector as well as the shRNA targeting GLI2 and IL-6 reporter luciferase construct. We found that stromal cells with decreased levels of GLI2 showed impaired IL-6 promoter activation by CCR3 (Fig. 2F). Moreover, GLI2 overexpression was able to rescue the effect of CCR3 knockdown on the IL-6 promoter in HS-5 and Saka cells, thus further demonstrating that GLI2 is downstream of the CCL5-CCR3 axis (Fig. 2G).
CCL5 Regulates GLI2 Activity in a Hedgehog-independent Manner in Stromal Cells-GLI transcription factors are regulated by Hedgehog and mediate many of its cellular functions during development as well as disease (42)(43)(44). Interestingly, overexpression of a constitutively active Smoothened (caSMO), the signaling component of the Hedgehog receptor complex, did not activate the IL-6 promoter in HS-5 cells (Fig. 3A), and knockdown of SMO using two different shRNA constructs targeting SMO did not affect IL-6 promoter activity (Fig. 3B). These results prompted us to examine the mechanism of regulation of GLI2 by CCL5. HS-5 cells incubated with CCL5 show a rapid and persistent induction of GLI2 mRNA, peaking at 6 h and remaining at high levels up to 24 h (Fig. 3C). This also resulted in an increase in GLI2 protein levels in HS-5 cells (Fig.  3D). A similar pattern of GLI2 modulation was observed in WMsc treated with CCL5 (Fig. 3E) or HS-5 and Saka cells over-expressing the CCR3 receptor (Fig. 3F). In addition, HS-5 and Saka cells with stable knockdown of CCR3 expression had a reduction in GLI2 mRNA expression (Fig. 3G), further supporting a role for CCL5/CCR3 axis in modulating GLI2 levels in stromal cells.
CCL5 is known to regulate numerous signaling pathways. In HS-5 cells, treatment with this chemokine distinguishably activates the PI3K/AKT as measured by phosphorylation of AKT 30 min after addition of CCL5 (Fig. 4A). To better define the role of these molecules in GLI2 induction, we used a combination of well established activators or inhibitors of the PI3K-AKT pathway and determined GLI2 expression. As shown in Fig. 4B, constitutively active PI3K and AKT mutants increase the expression of GLI2 in HS-5 cells. These mutants can also rescue GLI2 expression in HS-5 cells with stable CCR3 knockdown (Fig. 4C). Conversely, dominant negative molecules antagonizing these pathways effectively blocked CCR3 induction of GLI2 in stromal cells (Fig. 4D). Further analysis of the mechanism identified the IB-p65 axis as a candidate regulator of GLI2 levels in stromal cells. Overexpression of IB␣ dominant negative (super-repressor) (IB␣ (S 3 A) ) impaired the CCR3-mediated increase in GLI2 expression (Fig. 4D). The ChIP assay shows that endogenous p65 binds to the candidate site proximal to the transcriptional initiation site in the GLI2 promoter (Fig. 4, E and F) and that p65 overexpression leads to increased GLI2 mRNA expression in HS-5 stromal cells (Fig.  4G). Taken together, these results identify a novel pathway that uses GLI2 as the effector for its cellular function in the tumor microenvironment and expands the repertoire of signaling cascades that, in addition to Hedgehog, regulate GLI transcriptional activity.
IL-6 Promotes IgM Secretion by Malignant B Cells-Our data indicates that activation of the CCL5-CCR3-GLI2 axis increases IL-6 secretion within the tumor microenvironment. IL-6 is a known modulator of the biology of normal and malignant B cells, including immunoglobulin secretion (36,38,45). We initially examined the ability of IL-6 to modulate IgM secretion in WM. First, we confirmed the expression of the IL-6 receptor by flow cytometry on CD19ϩ CD138ϩ WM patient malignant cells as well as the BCWM.1 cell line. The IL-6 receptor was uniformly expressed on all the analyzed cells (Fig. 5A). Interestingly, we found a dose-dependent increase in IgM secretion by BCWM.1 cells in response to IL-6 ( Fig. 5B). Similar to recombinant IL-6 treatment, activation of the pathway using the gp130 expression construct, the signaling component of the IL-6 receptor complex, leads to increased secretion of IgM by BCWM.1 cells (Fig. 5C). Conversely, BCWM.1 cells pretreated with a blocking antibody for the IL-6 receptor (␣IL-6R) show impaired IgM secretion by recombinant IL-6 ( Fig. 5D). Furthermore, malignant cells isolated from WM bone marrow biopsy specimens secreted IgM in response to IL-6 in a similar fashion to the BCWM.1 cells (Fig. 5E).
To confirm the biological relevance of this newly identified signaling axis in modulating stromally secreted IL-6, we used a coculture system of stromal cells and the BCWM.1 cell line. Because an intact GLI2 is required for CCL5-induced IL-6 expression and functions downstream of the CCR3 signaling cascade, we infected WMsc, HS-5 cells, and Saka cells with a lentivirus containing shRNA targeting GLI2 or a scramble control, cocultured them with BCWM.1 cells, and measured IL-6 secretion after two 2 days. As expected, knockdown of GLI2 in WMsc or the HS-5 and Saka cell lines decreased IL-6 secretion by stromal cells (p Ͻ 0.0001) (Fig.  6A). Similar results were obtained when IgM secretion by BCWM.1 cells was examined after coculture with WMsc and HS-5 and Saka cells infected with lentivirus containing shRNA targeting GLI2 or a scramble control (p ϭ 0.0016) (Fig. 6B). To confirm the in vitro coculture findings, we used two in vivo mouse models. In the first model (16,46), we infected HS-5 cells with lentivirus-targeting GLI2 or scrambled shRNA and subcutaneously coinjected them with BCWM.1 cells at a ratio of 1:10 into athymic nu/nu mice. After 2 weeks, mice were sacrificed. Tumors and serum were harvested and used to determine IgM levels by immunohistochemistry or ELISA. Similar to the in vitro data, mice injected with HS-5 cells lacking GLI2 had lower IgM levels in the tumors (Fig. 6C) and serum (p ϭ 0.0355) (D) compared with mice injected with HS-5 cells with an intact GLI2. This effect on IgM secretion was not due to a change in tumor size (Fig. 6E) or number of BCWM.1 cells (data not shown), as there was no significant difference in these parameters after 2 weeks between the two experimental groups. In the second mouse model, we subcutaneously implanted human fetal bone (15,25,26) into the flanks of SCID mice and then injected BCWM.1 cells and HS-5 cells infected with lentivirus containing either scramble shRNA or GLI2 shRNA. Similar to the subcutaneous model, there was a significant decrease in IgM in the sera from mice injected with HS-5 cells lacking GLI2 (p ϭ 0.0488) (Fig. 6F). In summary, these data further support a central role of GLI2 in this newly identified axis and confirm the biological significance for this pathway in the regulation of IgM secretion in WM malignant cells. To test whether our newly identified cascade is unique to WM or more generally applicable to other malignancies of B cell origin, we generated primary stromal cultures from multiple myeloma (MMsc) and monoclonal gammopathy of undetermined significance (MGUSsc) patients and used them to determine the role of the CCL5-GLI2 axis in modulating IL-6 and Ig secretion. Similar to WMsc, MMsc and MGUSsc treated with CCL5 showed an increased IL-6 secretion upon stimulation with this cytokine (Fig. 7A). MMsc and MGUSsc were infected with lentivirus containing shRNA targeting GLI2 or scramble control shRNA and then cocultured with BCWM.1 cells for 2 days in the presence of CCL5. We found that decreased GLI2 expression in these stromal cells resulted in reduced IL-6 secretion (p Ͻ 0.0001) (Fig. 7B). Furthermore, knockdown of GLI2 resulted in impaired IgM secretion in coculture (p Ͻ 0.0001) (Fig. 7C). These results further support the role of the CCL5-GLI2-IL-6 axis in the modulation of Ig secretion and suggest that this pathway is a common feature in plasma cell malignancies with dysregulated Ig production.

DISCUSSION
Accumulating evidence has shown that the bone marrow tumor microenvironment is critically important for the initia-tion, maintenance, and progression of B cell malignancies. Cellular interactions within this tumor microenvironment are essential for the regulation of the neoplastic phenotype, including Ig production and secretion by malignant B cells (7,47). Although this is a recognized phenomenon, the complex interplay of these cellular compartments and the mechanisms mediating cytokine regulation of Ig secretion in B cell malignancies are poorly understood. The data presented here show that a novel pathway initiated by CCL5 in stromal cells leads to an increase in IL-6-dependent Ig secretion by malignant cells (Fig.  7D). We found that CCL5 increased IL-6 secretion by stromal cells within the tumor microenvironment through activation of the IL-6 promoter. Analysis of the mechanism regulating IL-6 expression identified GLI2 as the mediator of CCL5-induced activation of IL-6. GLI2 belongs to the GLI family of transcription factors, which are known effectors of the Hedgehog signaling pathway, a cascade that plays a major role during development (43) as well as carcinogenesis (14,42,44,48). Activation of these transcription factors occurs when the Hedgehog family ligand binds to PATCHED, a 12-transmembrane receptor. The binding blocks an inhibitory effect that PATCHED has over SMO, a signaling component of the Hedgehog receptor complex, so that SMO can signal to activate GLI transcription factors (12)(13)(14). These proteins in turn bind to promoter sequences leading to regulation of Hedgehog target genes (49 -51). In this study, we demonstrate that GLI2 activity is regulated in a SMO independent manner by CCL5. We have identified All experiments were done in triplicate wells, and bars represent average Ϯ S.E. D, model for CCL5-GLI2-IL-6 regulation of Ig production in B cell malignancies. Elevated CCL5 levels activate the CCR3 receptor, triggering a signal that increases the expression of the GLI2 transcription factor via a PI3K-AKT-NF-B pathway. Then, GLI2 binds to and activates the IL-6 promoter, leading to an increase in IL-6 expression and secretion by stromal cells. The IL-6 present in the tumor microenvironment binds to and activates the IL-6R on the malignant B cells and leads to increased IgM secretion.
GLI2 as a novel transcriptional target of this signaling cascade. Analysis of the mechanism revealed that CCL5 promotes GLI2 activity by increasing its expression via the PI3K-AKT-NF-B pathway. Thus, these data identify a novel pathway that uses GLI2 as the effector for its cellular function and expands the repertoire of signaling cascades that, in addition to Hedgehog, regulate GLI transcriptional activity in tumors.
Studies to date have addressed the role of CCL5 and GLI2 proteins in malignant cells using various cancer models (52)(53)(54)(55)(56)(57)(58). We show here that this novel signaling axis involving CCL5 and GLI2 is also important to support cells within the stromal microenvironment (Fig. 7D). The importance of stromal cells is well established in many neoplasms, including malignancies of B cell origin (7,47,53,59). Coculture and coinjection of stromal cells with decreased GLI2 levels leads to a reduction in both IL-6, a novel target of this transcription factor, and, consequently, IgM secretion, therefore supporting the notion that the CCL5-GLI2-IL-6 axis in stromal cells is important for Ig secretion by B cells. Indeed, loss of GLI2 in the stromal cells leads to a decrease in IgM secretion by plasma cells in vitro and in vivo. These findings may fuel future experiments focused on more carefully examining the direct role of this axis in the microenvironment and potentially shed more light on the molecular mechanisms underlying the development and maintenance of plasma cell malignancies.
In summary, these studies outline a novel pathway initiated by the chemokine CCL5, promoting the production of the proinflammatory cytokine IL-6 in WM. With an established role in the promotion of Ig secretion in normal B cells (60) and IgG secretion in MM (61), here we showed that IL-6 also promoted IgM secretion by malignant cells in WM and other Ig-related diseases. Our observations in this Ig-related disorder apply to MGUS and MM, where a decrease in GLI2 expression in the stroma of these malignancies also leads to a reduction in IL-6 secretion as well as IgM secretion by malignant B cells. This study expands our understanding of the mechanisms regulating Ig secretion by a signal initiated by the tumor microenvironment. Therefore, therapeutic targeting of the CCL5-GLI2-IL-6 axis in the stromal cells may provide a useful tool to inhibit Ig secretion in Ig-related disorders, including WM, MGUS, and MM.