| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Cell and Tumor Biology |
1 Rigel, Inc., South San Francisco, California and 2 University of Michigan School of Dentistry, Ann Arbor, Michigan
Requests for reprints: Sacha J. Holland, Rigel, Inc., 1180 Veteran's Boulevard, South San Francisco, CA 94080. Phone: 650-624-1283; E-mail: sholland{at}rigel.com.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Angiogenesis, a physiologic and strictly regulated process in healthy adults, is aberrantly induced by hypoxic tumors thus securing an adequate blood supply that feeds tumor growth and facilitates metastasis. The assembly of functional vessels requires the precise coordination of multiple cellular events and interactions (3). Compelling evidence shows that vascular endothelial growth factor (VEGF) can initiate angiogenic sprouting in vivo, increasing permeability and supporting endothelial cell proliferation and survival (4). Sprouting migration of activated endothelial cells is accompanied by up-regulation of integrin matrix factor receptors (e.g., 
vß3 and 5ß1). Nascent endothelial tubes thus formed are stabilized by pericytes and basement membrane formation. Targeted inhibition of angiogenesis using an anti-VEGF antibody (bevacizumab) has shown efficacy in treating human cancer (5). However, as tumors frequently advanced while patients were receiving therapy, inhibition of VEGF signaling alone may not be sufficient to block neovascularization in human malignancies (6). Blocking processes complementary to those driven by VEGF may therefore be important in fully inhibiting angiogenesis in human disease.
Retroviral-based functional screening has been established as an effective, versatile, and unbiased method for identifying novel drug targets and regulators of specific cellular and/or disease processes (7). The use of retroviral expression vectors allows the efficient delivery and stable expression of complex and diverse libraries of genetic effectors (e.g., cDNAs and peptides) in a multitude of cell types including primary human cells. This allows design of screening assays in cell types that reflect physiologically relevant disease responses. We reasoned that this approach would be well suited to probing a complex and technically challenging cellular function such as cell migration.
As part of an effort to develop novel therapeutic strategies for cancer treatment, we developed a unique retroviral-based functional genetic screening protocol to discover genes that regulate cell migration in primary human endothelial cells. We identified Axl as a novel regulator of endothelial cell haptotactic migration towards the matrix factor vitronectin. Axl (UFO/ARK/Tyro7) is a receptor tyrosine kinase (RTK) that is stimulated by the 
76-kDa secreted protein Gas6 (growth arrest specific 6; refs. 8, 9). We show that Axl signaling affects multiple cellular behaviors required for neovascularization in vitro and regulates angiogenesis in vivo. Furthermore, we show that loss of Axl expression in tumor cells blocks the growth of solid human neoplasms in an in vivo MDA-MB-231 breast carcinoma xenograft model. This suggests a model in which Gas6-Axl signaling supports tumor-stroma interactions necessary for tumor growth and angiogenesis.
Taken together, these data indicate that Axl signaling can independently regulate neovascularization and tumor growth and thus represents a novel target class for tumor therapeutic development.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cell culture. All cells were cultured at 37°C, 5% CO2. Phoenix A cells (Dr. Gary Nolan, Stanford) were maintained in DMEM supplemented with 10% fetal bovine serum (FBS), antibiotics, and glutamine. MDA-MB-231 human breast epithelial carcinoma cells (American Type Culture Collection, Rockville, MD) were maintained in F12K supplemented with 10% FBS. Primary human umbilical vein endothelial cells (HUVEC), human dermal microvascular endothelial cells (HMVEC), and pulmonary artery smooth muscle cells (PASMC) were cultured as vendor's recommendations (Cambrex, Walkersville, MD).
Retroviral haptotaxis screen. The HUVEC green fluorescent protein (GFP)-cDNA library was prepared in the retroviral CRU5-GFP fusion vector (10). mRNA was isolated from VEGF, basic fibroblast growth factor, and epidermal growth factor–stimulated HUVECs with poly-dT magnetic beads (Dynal, Lake Success, NY) and normalized with biotinylated first strand HUVEC cDNA and streptavidin beads (Invitrogen, Carlsbad, CA). Double-stranded cDNA was synthesized using standard methods (Smart cDNA System, Clontech/BD Biosciences, Mansfield, MA) and cloned into the library vector bidirectionally using BstXI adapters (Invitrogen). The complexity of the final library was 1.5 x 106. Retroviral packaging cells (Phoenix A) were transfected with the cDNA library to produce infectious virus sufficient to infect 7 x 106 subconfluent HUVEC at 10% transduction efficiency in T75 flasks. GFP-expressing HUVECS were sorted by fluorescence-activated cell sorting (FACS) and replated. The following day, the cells were harvested and migrated thrice in 75-mm Boyden chambers (Transwell, Corning-Costar, Acton, MA) coated on the underside with 15 µg/mL vitronectin (Chemicon, Temecula, CA). Each migration was conducted for 3 hours. Nonmigrating cells were harvested with 0.05% Trypsin for 90 seconds at 37°C and remigrated. The cells were infected with wild type Moloney murine leukemia virus (MMTV, American Type Culture Collection) before the third migration step to produce library clone viruses from the enriched nonmigratory cell population. The final selected nonmigratory HUVEC population was plated and library viruses were harvested the following day and used to infect PG13 packaging cells at a low multiplicity of infection (0.5%). Single GFP+ PG13 cells were arrayed into microtiter plates by FACS. cDNA inserts were isolated by reverse transcription-PCR (RT-PCR, Clontech/BD Biosciences) from purified retrovirus (12). To confirm the library insert phenotype, naive HUVECs were transduced using supernatants from each PG13 clone. GFP-expressing cells were purified by FACS and tested in individual haptotaxis assays (see below).
Small interfering RNA design and transfection. Luciferase control (GL2), VEGFR2 (D-003148-05), and custom small interfering RNA (siRNA) oligos were purchased from Dharmacon (Lafayette, CO). siRNA sequences (Supplementary Table 1) were designed according to established criteria and examined to ensure minimal homology with other sequences. For siRNA transfection, 60,000 HUVECs were plated per well in a six-well plate. The following day, cells were transfected with 3.2 µg siRNAs using Oligofectamine reagent in Opti-MEM (both Invitrogen) for 4 hours after which 3x volume of culture medium containing 6% FBS was added.
Constructs. The shRNA vector (EFS-U3/U6) was derived from the TRA retroviral vector (10). The EFS-U3/U6 vector comprises retroviral elements required for stable integration into the genome of infected cells, a modified U6 RNA promoter (–230 to +1) and terminator (TTTTT) embedded into the NheI site of the TRA vector 3' long terminal repeat for conditional expression of hp-siRNA and an internal EF1- expression cassette (11) that drives expression of a destabilized version (COOH-terminal PEST sequence) of the Renilla GFP (dsRMG) for independent monitoring of transfection/infection efficiencies. Specific shRNA vectors were created by PCR. COOH-terminally myc-tagged Axl cDNA (480-3140) was expressed from an IRES-GFP vector (10) derivative where GFP is replaced with Renilla GFP.
Transfection/retroviral infection. Phoenix A cells were transfected using the calcium phosphate method (12). Approximately 30 hours after transfection, the medium was changed to growth medium for the cells to be infected, supplemented with 10% FBS. Infectious supernatant was collected 48 hours after transfection. Target cells were exposed to supernatant containing 5 µg/mL protamine sulfate for 6 hours to overnight before being returned to regular growth medium.
Taqman. Total RNA was isolated using RNeasy Mini Kit (Qiagen, Valencia, CA) and quantified using Ribo Green Quantification kit (Molecular Probes, Eugene, OR). Human Gas6 primers and probe were designed using Primer Express software: sense primer, ATGTGGCAGACAATCTCTG; reverse primer, ACAGCATCCCTGTTGACCT; probe, AGCTGGCGCGGAATCTGGTCA. Reactions were run using a one-step RT-PCR kit (Qiagen) in an ABI Prism 7900HT sequence analyzer. An 18S rRNA reaction kit (ABI, Foster City, CA) was used as the reference/loading control reaction. Data was normalized to 18S rRNA and plotted as percentage of luciferase control.
Western blot. Cells were lyzed in PLC lysis buffer containing 1% Triton X-100 (13). Protein concentrations were determined using the bicinchoninic acid protein assay kit (Pierce, Rockford, IL) and equalized. SDS-PAGE and Western blotting were carried out according to standard procedures.
Boyden chamber haptotaxis assay. Migration assays were done 48 hours after siRNA transfection, or 1 to 2 days after sorting of infected populations. The undersides of 8 µm 24-well Transwell filters (Costar) were coated with 15 µg/mL vitronectin (Chemicon). Six replicate wells were run per sample. The following day, unbound vitronectin was eluted in migration medium [0.5% bovine serum albumin (BSA, Sigma, St. Louis, MO) in EGM-2 medium (Cambrex)]. Migration medium (500 µL) was added into the lower chamber. HUVECs resuspended in migration medium at 2,000 cells/300 µL per well were added to the upper chamber. Cells were allowed to migrate for 3 hours at 37°C after which they were removed from either the upper or lower surface of the membrane (three wells each) using a cotton swab. Following fixation in 3.7% formaldehyde, swabbing was repeated and nuclei were stained using 2 µg/mL 4',6-diamidino-2-phenylindole. Cells remaining on the upper (nonmigrated cells) or lower (migrated cells) surface of the membrane were imaged using a Cellomics Arrayscan instrument and processed using ImagePro software. Migration is presented as mean haptotactic index (% nonmigrated, control / % nonmigrated, test) ± SD: 1, no effect; >1, enhanced migration; <1, retarded migration. Representative examples of at least two replicate experiments are shown. Statistical analysis was done using the 
2 test.
Proliferation assay. Cells were seeded sparsely in six-well dishes, infected with retroviral supernatants, and maintained in complete endothelial cell medium. The percentage of infected, GFP-positive cells in the population was monitored at regular intervals by FACS. Cells were passaged during the course of the experiment as necessary to keep them subconfluent. Data was analyzed using FlowJo software. The relative number of GFP-positive cells for all samples at the beginning of the experiment (day 3) was set to 100%. Subsequent changes in relative % GFP over time were normalized to the in-day vector control value.
Coculture assay. Early-passage HUVECs (
p6) were mixed 1:10 with PASMC in HUVEC growth medium (EGM-2) and seeded onto collagen-coated (Sigma) LabTekII chamber slides (Nalge Nunc, Rochester, NY). Medium was replaced every 2 days. Five days after seeding, the cocultures were washed in PBS, fixed using –20°C/70% ethanol at room temperature for 15 minutes, blocked in PBS + 5% BSA, and stained using FITC-conjugated anti-CD31 antibodies or Rhodamine-conjugated UEA-1 lectin (1:200) to reveal endothelial cell tubes. Cultures were photographed and mean tube formation over 16 images covering an entire representative well was quantified using ImagePro software. Representative examples of at least two replicate experiments are shown. Statistical analysis was done using the t test.
In vivo angiogenesis assay. HMVECs were infected with shRNA vectors and GFP-positive cells were enriched by FACS. Poly L-lactic acid (PLLA) sponge matrices (5 x 5 x 1 mm) were seeded with 700,000 sorted HMVEC cells in a 1:1 mixture of EGM-2 MV and Matrigel (BD Biosciences). Up to four sponges were implanted s.c. in female CB-17 severe combined immunodeficient (SCID) mice according to Nör et al. (14). Sponges were harvested after 14 days for biochemical or immunohistochemical analysis.
To specifically label perfused human endothelial cells in the sponges, 200 µL of rhodamine-conjugated UEA-1 lectin (1 µg/µL in sterile 0.9% NaCl) was injected into the tail vein 30 minutes before sacrifice. Excised sponges were homogenized using a handheld mincer (Biospec Products, Inc., Bartlesville, OK) in 200 µL of modified radioimmunoprecipitation assay buffer. Fluorescence in filtered (0.45-µm GHP syringe filter, Pall-Gelman Corp., Inc., East Hills, NY) supernatant was quantified on a Fluorskan machine with a 530/590-nm excitation/emission filter pair and normalized to plasma lectin content. Tie-2 levels present in sponge lysates (100 µL) were measured using a human Tie-2 ELISA (R&D Systems) and normalized to protein content.
Statistical analysis between groups was done using the Mann-Whitney rank sum test.
Xenograft assay. MDA-MB-231 cells were infected with shRNA vectors and GFP positive populations were enriched to 98% to 100% by FACS. RNA interference (RNAi)–mediated knockdown of cell surface Axl expression was confirmed on the day of implantation. Cells (1 x 107) were injected in 0.2 mL of F12K medium + 10% FBS/high concentration Matrigel (1:1 ratio, BD Biosciences) into the right flank of female CB-17 SCID mice. Animals in which tumors measured at least 0.3 to 0.5 cm were selected for the study. Tumor volume [V = (tumor length x tumor width2) / 2] was measured every 2 to 3 days for 30 days, or until it reached a volume of 2,000 mm3. Tumor growth was expressed as V/Vo, where V = tumor volume and Vo = initial volume of tumor. Mean V/Vo and SDs were calculated using Labcat Software. Statistical analysis between groups was done using the t test.
Immunohistochemistry. Staining reagents were from BioCare. Heat-induced epitope retrieval was done on 4 µm paraformaldehyde/formalin-fixed, paraffin-embedded tissue sections using Medical Decloaking Chamber in Borg high-pH buffer for 2 minutes followed by rinses in Hot Rinse Solution to remove residual paraffin. Nonspecific staining was eliminated using SNIPER blocking reagent. Primary antibody (mouse monoclonal anti-human CD31; 1:50 or mouse monoclonal anti-human Axl, 1:100) incubations were followed by biotinylated goat anti-mouse/rabbit secondary (BioCare) and then alkaline phosphatase or horseradish peroxidase–conjugated streptavidin. Development was done according to the manufacturer's instructions. Slides were counterstained with hematoxylin, mounted, and coverslipped. Brightfield pictures were taken using a Nikon Eclipse 400 microscope with a Spot Insight QE imaging system.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Two hundred fifty four independent library inserts were identified. Library clone #420 encoded a fragment from the COOH-terminal region of the RTK Axl (nucleotides 2884-2964), which was oriented in the reverse direction indicating that it may function as an antisense effector.
Genetic effectors targeting Axl impair vitronectin haptotaxis. To reconfirm that the isolated clone #420 virus could reproduce the phenotype selected for in the screen, retroviral supernatant from clone 420 was used to infect naive HUVECs. GFP-expressing cell populations were enriched by FACS and assayed by migration towards 15 µg/mL of vitronectin. HUVECs infected with the Axl genetic effector indeed showed reduced haptotactic migration towards vitronectin (Fig. 1B).
We used siRNA-mediated silencing to independently verify that Axl knockdown could inhibit vitronectin haptotaxis. Compared with HUVECs transfected with a control siRNA-targeting luciferase, those transfected with either of two independent siRNAs targeting Axl (Fig. 1C) phenocopied the effect of the Axl antisense screening insert and exhibited reduced migration towards vitronectin (Fig. 1D). Furthermore, GFP-expressing retroviral vectors designed to produce shRNAs corresponding to these Axl-silencing sequences generated a similar reduction in haptotaxis (data not shown). Axl expression in HUVECs is thus required for haptotaxis towards vitronectin.
Axl regulates human umbilical vein endothelial cell proliferation. An important characteristic of angiogenic endothelial cells is their high proliferative capacity. As Gas6 is a weak mitogen in some cell types, we asked whether Axl signaling regulates HUVEC proliferation. HUVECs infected with Axl shRNA vectors or a positive control construct expressing the cell cycle inhibitor p21 (22) exhibited a marked growth disadvantage compared with controls (Supplementary Fig. 1). Furthermore, although we did not detect significant apoptosis in Axl shRNA-expressing cell populations under the assay conditions, Axl silencing also reduced DNA replication as assayed by bromodeoxyuridine incorporation (data not shown). Taken together, these data show that Axl expression is important in driving HUVEC proliferation.
Axl knockdown impairs endothelial tube formation in vitro. We next addressed whether Axl regulates endothelial tube formation. To mimic the complex matricellular environment encountered by angiogenic endothelial cells in vivo, we designed a HUVEC/primary PASMC coculture branching morphogenesis/tube formation assay (Fig. 2A). Endothelial morphogenesis in this assay is VEGF dependent and proliferation independent (see Supplementary Fig. 2).
|
50 % in HUVECs transfected with Axl siRNAs (Fig. 2B, ii, iii, v, and vi and C; data not shown) compared with those transfected with the luciferase control siRNA (Fig. 2B, i and iv and C). These results show that Axl is involved in the regulation of endothelial tube morphogenesis. Endothelial cell Gas6 expression and Axl catalytic activity are required for regulation of proangiogenic processes. We next asked whether ligand-stimulated Axl catalytic activity is required for its proangiogenic functions. HUVECs synthesize the Axl ligand, Gas6 (23). Two independent siRNAs designed to target Gas6 depressed HUVEC Gas6 message levels and reduced HUVEC haptotaxis towards vitronectin compared with controls (Supplementary Fig. 3). Gas6 silencing thus phenocopies the effect of Axl knockdown on vitronectin haptotaxis.
We next focused on the catalytic activity of Axl. Overexpression of WT Axl in HUVECs (Fig. 3A) induced a growth advantage relative to controls (Fig. 3B). Conversely, overexpression of a catalytically inactive Axl mutant (AxlKD; K567E substitution; ref. 24; Fig. 3A) conferred a 30% growth disadvantage over the course of the assay (Fig. 3B). Lastly, sorted, GFP-positive WT Axl or GFP-infected HUVECs were seeded into a coculture assay. Overexpression of WT Axl led to 40% increase in the number of endothelial cell tubes formed compared with control vector (Fig. 3C, ii and v and D). In contrast, overexpression of AxlKD did not enhance tube morphogenesis (Fig. 3C, iii and vi and D). Similar trends were noted for fiber length and branching (data not shown). The autocrine/paracrine ligand-induced kinase activity of Axl therefore controls several distinct proangiogenic processes and thus represents an attractive target for small-molecule inhibition.
|
|
Fluorescent UEA-1 lectin, which binds specifically to human endothelial cells, was injected into the mouse bloodstream before sacrifice. The relative fluorescence in implant lysates compared with mouse plasma provides a measure of functional circulation in the human neovasculature. Figure 4C (ii) shows a marked reduction (77% for Axl-2 and 58% for Axl-4) in lectin fluorescence (i.e., implant perfusion) in Axl shRNA samples compared with controls, indicating a functional impairment in human neovessels lacking Axl expression.
Vessel morphology was assessed by anti-human CD31 staining of implant sections. Control vector–infected HMVECs formed small capillaries with obvious lumens (Fig. 4D, i, arrow). Moreover, developing blood vessels seem to have fused normally to form larger vessels (Fig. 4D, ii, arrow). In contrast, vessels generated by HMVECs infected with Axl shRNA vectors seem consistently smaller than controls and lack patent lumens (Fig. 4D, iii and iv).
These data show that Axl expression in endothelial cells is important for the formation of functional blood vessels in vivo.
Inhibition of Axl expression reduces growth of MDA-MB-231 breast carcinoma cells in a xenograft assay. Axl is overexpressed in a large variety of human tumors and has transforming activity in murine NIH3T3 and 32D cells (25–27), suggesting the possibility that Axl may act as an oncogene. However, functional evidence that Axl can drive the growth or progression of human tumors in vivo has thus far not been forthcoming.
The human breast tumor line, MDA-MB-231, which expresses high levels of Axl (Supplementary Fig. 4; ref. 28), was infected with Axl shRNA vectors. We did not detect significant changes in the proliferation or apoptosis of Axl-shRNA-infected MDA-MB-231 cells in culture compared with controls (data not shown). GFP-positive cells were enriched by FACS and implanted s.c. into SCID mice. Compared with vector controls, MDA-MB-231 cells infected with Axl shRNA vectors showed a markedly impaired ability to grow as xenografts in immunocompromised mice (Fig. 5B). The extent of inhibition correlated with the degree of Axl knockdown, assayed by immunostaining for Axl protein in tumor sections harvested at the end of the study (Fig. 5C). Whereas the mechanism underlying these observations in xenografts is not understood at present, these results show that, in addition to regulating angiogenesis, Axl plays a key role in driving the growth of human tumor cells in vivo.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The receptor tyrosine kinase Axl was first identified from the DNA of chronic myelogenous leukemia patients due to its transforming activity in vitro (25, 26). Expression of Axl is detected in many cell types but is prominent in the vasculature (in both endothelial cells and vascular SMCs, VSMC) and in cells of the myeloid lineage (29–31). Axl and its two close relatives Mer/Nyk and Sky (Tyro3/Rse/Dtk) all bind and are stimulated to varying degrees by Gas6 (8, 9). Axl signaling in culture protects cells from serum starvation– or tumor necrosis factor--induced apoptosis and mediates ligand-induced chemotaxis and cell differentiation (24, 31–33). However, Axl–/– mice exhibit no overt phenotype and the physiologic function of Axl in vivo is not clearly established (34). The overexpression of Axl and/or its ligand, Gas6, has been reported in a wide variety of solid human tumor types (ref. 35 and references therein) and myeloid leukemias (30). Axl signaling is also functionally implicated in the response to vascular injury and kidney disease (29, 36). Axl may thus potentially represent a therapeutic target for diverse pathologic conditions.
Our data suggest that Axl acts as a central facilitator of endothelial cell activation. Axl may support pathologic neovascularization as impaired developmental angiogenesis has not been reported in Axl–/– mice. Axl and Gas6 are coexpressed in endothelial cells and silencing of either gene impairs vitronectin haptotaxis. Thus, a constitutive autocrine/paracrine signaling loop may exist in these cells (ref. 37; data not shown). Axl may modulate activated endothelial cell functions through one of several possible mechanisms: (a) Axl signaling in endothelial cells may support the activated state, potentially by altering gene expression, thus maintaining a proliferative, motile phenotype. Axl inhibition may therefore alter the angiogenic balance, driving endothelial cells towards a more quiescent, nonmotile, differentiated state. (b) Axl signaling modulates integrin function. Angelillo-Scherrer et al. recently showed that Gas6-dependent signaling results in phosphorylation of ß3 integrin in platelets (38) resulting in defective IIbß3-dependent "outside-in" signaling in Axl–/– platelets. Similar crosstalk has been described between growth factor receptors (e.g., VEGFR2 and FGFR) and integrins (e.g., vß3 and vß5; ref. 39). Integrin-mediated substrate adhesion is required for cell migration, proliferation, and survival of growth factor–stimulated cells. Thus, blocking Axl signaling may alter the function of certain integrins necessary for the growth, migration, and survival of endothelial cell and tumor cells. (c) Constitutive Gas6 stimulation may maintain the steady state activity of Axl effectors such as Rac and Akt (40, 41) that control matrix- and chemoattractant-induced cell motility. The defective lumen morphology observed in vessels lacking Axl may reflect regulation of the Rho/Rac-dependent fusion of intracellular vacuoles thought to control lumen generation (42, 43). Akt is a positive regulator of angiogenesis and is central to both Gas6- and VEGF-induced endothelial cell survival (41, 44). Moreover, the pivotal role of Akt in tumor development is well documented (45).
Axl knockdown diminished vitronectin haptotaxis, proliferation, and blunted endothelial cell tube morphogenesis in vitro. Conversely, overexpression of WT Axl but not AxlKD enhanced endothelial cell growth and tube formation. Whereas we cannot at present rule out a contribution from homophilic adhesive interactions via the ligand-binding domain (46), these observations show the kinase dependency of proangiogenic Axl signaling and imply that this process is a classic target for small-molecule drug inhibition.
Gallicchio et al. recently reported that Gas6 antagonizes VEGFR2-dependent angiogenic processes (47). Low Gas6 concentrations (0.01 ng/mL) blocked VEGFR2 phosphorylation and VEGF chemotaxis, whereas high concentrations (>20 ng/mL) had little effect and >40 ng/mL Gas6 impaired tube formation in the Matrigel assay (47). We have not studied VEGF-induced chemotaxis; however, tube formation in endothelial cell/SMC cocultures is VEGF-dependent (see Supplementary Information), although both endothelial cells and SMCs express Gas6 (data not shown; refs. 29, 37). Axl knockdown impairs endothelial tube formation and overexpression of WT but not catalytically inactive Axl enhances endothelial cell tube morphogenesis. Furthermore, Axl silencing in an in vivo angiogenesis model produces an inhibitory phenotype similar to VEGFR2 knockdown. To reconcile these observations, one could address whether Axl silencing blocks Gas6-mediated VEGFR2 inhibition.
Pronounced Axl expression has been documented in a variety of human cancers (ref. 35 and references therein). Furthermore, Axl possesses transforming activity in NIH3T3 and 32D cells (25–27). In endothelial cells, Axl controls diverse processes including growth, survival, migration, and morphologic differentiation linked to the proliferative and invasive changes required for angiogenic activation (31, 37, 41). These mechanisms are also central to tumor formation and progression in vivo. Although no significant impairment of growth was noted in culture, sustained, RNAi-mediated inhibition of Axl expression in tumor cells blocked the growth of solid human neoplasms in an in vivo MDA-MB-231 human breast cancer xenograft model. This experiment provides functional evidence that Axl signaling may contribute to driving human tumor growth. The recently reported induction of the protumorigenic cytokine interleukin-8 (IL-8) by Axl family RTKs (48) is compatible with a model in which Gas6-Axl signaling enhances tumor growth and supportive angiogenesis in part via the local induction of cytokine secretion. Confluent, serum-deprived MDA-MB-231 cells up-regulate Gas6 message (data not shown) consistent with the finding that these cells produce IL-8 under serum starvation conditions (49). Correspondingly, we have observed expression of human Gas6 in MDA-MB-231 xenograft samples (data not shown).
In addition to angiogenic endothelial cells and neoplastic cells, solid tumors offer perivascular cells as potential cellular therapeutic targets (50). Disruption of pericyte function using platelet-derived growth factor receptor inhibitors synergizes with VEGFR2 inhibition in endothelial cells resulting in enhanced inhibition of angiogenesis and tumor (51). Interestingly, Axl expression is up-regulated in VSMCs of developing blood vessels (31) and Axl-Gas6 signaling plays key roles in activation of SMCs after vascular injury (29). Whereas the function of Axl in pericytes or VSMCs in developing vessels is not established, inhibition of Axl may also potentially disrupt the support function of such perivascular cells providing a third level at which Axl targeting could intervene to suppress tumor growth (Fig. 6). We have recently discovered small-molecule inhibitors of Axl signaling (data not shown). We hypothesize that disruption of Axl signaling using a small-molecule inhibitor may simultaneously affect the endothelial cell, pericyte, and tumor cell compartments, independently targeting angiogenesis and tumor growth and providing a highly effective method to treat solid human tumors.
|
|
Acknowledgments |
|---|
We thank Hua Wang for statistical analysis.
|
Footnotes |
|---|
Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
Received 3/24/05. Revised 7/19/05. Accepted 8/ 4/05.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
gene. J Biol Chem 1994;269:29831–7.This article has been cited by other articles:
|
|
 |
|
  A. Gustafsson, D. Martuszewska, M. Johansson, C. Ekman, S. Hafizi, B. Ljungberg, and B. Dahlback Differential Expression of Axl and Gas6 in Renal Cell Carcinoma Reflecting Tumor Advancement and Survival Clin. Cancer Res., July 15, 2009; 15(14): 4742 - 4749. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Grosso, A. Puissant, M. Dufies, P. Colosetti, A. Jacquel, K. Lebrigand, P. Barbry, M. Deckert, J. P. Cassuto, B. Mari, et al. Gene expression profiling of imatinib and PD166326-resistant CML cell lines identifies Fyn as a gene associated with resistance to BCR-ABL inhibitors Mol. Cancer Ther., July 1, 2009; 8(7): 1924 - 1933. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-X. Zhang, P. G. Knyazev, Y. V. Cheburkin, K. Sharma, Y. P. Knyazev, L. Orfi, I. Szabadkai, H. Daub, G. Keri, and A. Ullrich AXL Is a Potential Target for Therapeutic Intervention in Breast Cancer Progression Cancer Res., March 15, 2008; 68(6): 1905 - 1915. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Hutterer, P. Knyazev, A. Abate, M. Reschke, H. Maier, N. Stefanova, T. Knyazeva, V. Barbieri, M. Reindl, A. Muigg, et al. Axl and Growth Arrest Specific Gene 6 Are Frequently Overexpressed in Human Gliomas and Predict Poor Prognosis in Patients with Glioblastoma Multiforme Clin. Cancer Res., January 1, 2008; 14(1): 130 - 138. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Revil, J. Toutant, L. Shkreta, D. Garneau, P. Cloutier, and B. Chabot Protein Kinase C-Dependent Control of Bcl-x Alternative Splicing Mol. Cell. Biol., December 15, 2007; 27(24): 8431 - 8441. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. C. Abba, V. T. Fabris, Y. Hu, F. S. Kittrell, W.-W. Cai, L. A. Donehower, A. Sahin, D. Medina, and C. M. Aldaz Identification of Novel Amplification Gene Targets in Mouse and Human Breast Cancer at a Syntenic Cluster Mapping to Mouse ch8A1 and Human ch13q34 Cancer Res., May 1, 2007; 67(9): 4104 - 4112. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-D. Lay, C.-C. Hong, J.-S. Huang, Y.-Y. Yang, C.-Y. Pao, C.-H. Liu, Y.-P. Lai, G.-M. Lai, A.-L. Cheng, I.-J. Su, et al. Sulfasalazine Suppresses Drug Resistance and Invasiveness of Lung Adenocarcinoma Cells Expressing AXL Cancer Res., April 15, 2007; 67(8): 3878 - 3887. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. D.M. Collett, A. P. Sage, J. P. Kirton, M. Y. Alexander, A. P. Gilmore, and A. E. Canfield Axl/Phosphatidylinositol 3-Kinase Signaling Inhibits Mineral Deposition by Vascular Smooth Muscle Cells Circ. Res., March 2, 2007; 100(4): 502 - 509. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Nakamura, I. J. Fidler, and K. R. Coombes Gene Expression Profile of Metastatic Human Pancreatic Cancer Cells Depends on the Organ Microenvironment Cancer Res., January 1, 2007; 67(1): 139 - 148. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Vajkoczy, P. Knyazev, A. Kunkel, H.-H. Capelle, S. Behrndt, H. von Tengg-Kobligk, F. Kiessling, U. Eichelsbacher, M. Essig, T.-A. Read, et al. Dominant-negative inhibition of the Axl receptor tyrosine kinase suppresses brain tumor cell growth and invasion and prolongs survival PNAS, April 11, 2006; 103(15): 5799 - 5804. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Jenkins, V. Markovtsov, W. Lang, P. Sharma, D. Pearsall, J. Warner, C. Franci, B. Huang, J. Huang, G. C. Yam, et al. Critical Role of the Ubiquitin Ligase Activity of UHRF1, a Nuclear RING Finger Protein, in Tumor Cell Growth Mol. Biol. Cell, December 1, 2005; 16(12): 5621 - 5629. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
