| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Molecular Physiology and Biological Physics,
Center for Cellular Signaling,
Department of Radiation Oncology,
|| Department of Health Evaluation Sciences, University of Virginia, Charlottesville, Virginia, USA; and
American Institute for Diseases of the Prostate, Charlottesville, Virginia, USA
1Correspondence: Department of Molecular Physiology and Biological Physics, University of Virginia, 1300 Jefferson Park Ave., P.O. Box 800736, Jordan Hall, Charlottesville, VA 22908-0736, USA. E-mail: avs5u{at}virginia.edu or aps2n{at}virginia.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: myosin phosphatase • G-protein • endothelium • tumor necrosis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. It is well known that cancer progression and metastasis require cell motility, entry into the blood (or lymphatic) stream followed by exit, breakdown of extracellular matrix, and growth in a "fertile" environment, supported by neo-angiogenesis. Given the multiplicity of pathways involved in cancer progression, increasing attention is being paid to targeting these pathways with noncytotoxic drugs, alone or in combination with cytotoxic agents (2
, 3)
. Two of the major target pathways supporting tumor growth and metastasis are neo-angiogenesis (4
5
6
7
8)
and matrix breakdown by matrix metalloproteinases (MMPs) (3
, 6
, 9
10
11
12)
.
Several properties of the Ras subfamily-Rho family GTPase proteins indicate they and their effectors may be on pathways that are suitable candidates for anti-cancer targeting (13
, 14)
. Rho GTPases are required for Ras-mediated oncogenic transformation (15
, 16)
convey increased motility (17
, 18)
and invasiveness (17
, 19
20
21)
to experimental tumors. RhoA and possibly RhoC are overexpressed in some human malignancies (reviewed in ref 14
), and nuclear localization of RhoA increases with progression of malignant phenotypes of human prostate cancers (22)
.
The RhoA effector Rho-kinase (ROK) enhances cell motility through the mechanism shown in Fig. 1
.
|
When activated by GTP·RhoA, ROK phosphorylates the myosin phosphatase regulatory subunit MYPT1 and inhibits, probably through variable mechanisms, the catalytic subunit responsible for dephosphorylating and inactivating myosin II, a major motor protein required for cell motility in eukaryotic cells (reviewed in refs 23
, 24
). By targeting this mechanism, the strongly selective pyridine derivative ROK inhibitor Y-27632 (25)
inhibited invasiveness of several animal cancers (26)
and of xenotransplants of the highly aggressive PC3 human prostate cancer (17)
. Y-27632 showed unexpected, but significant, anti-angiogenic activity (17)
. ROK inhibitors increase the resistance of the endothelial barrier (27
, 28)
. In the case of tumors metastasizing to bone, such as prostate (29
, 30)
and breast cancers, the positive regulation of osteoclastic activity by Rho (31
, 32)
suggests that inhibition of Rho signaling may inhibit metastatic bone destruction. In the present study we explored, in vitro and in vivo, potential anti-angiogenesis and anti-tumor mechanisms and activities of Wf-536, another pyridine derivative ROK inhibitor (25)
. We anticipated that treatment of an established, large, and disseminated tumor would unlikely be effective and that rational therapy with agents inhibiting matrix breakdown and cell migration would be most effective when applied early. Therefore, we initiated therapy at an early stage of growth of human prostate cancer, PC3 xenotransplants, to determine 1) whether the putative mechanism of myosin II regulation by ROK, phosphorylation of MYPT1, is detectable in human cancer and endothelial cells; 2) the effect of inhibition of ROK on the earliest stage of angiogenesis; 3) the combined effects of two agents expected to act on multiple pathways, the ROK inhibitor Wf-536 and the matrix metalloproteinase (MMP) inhibitor Marimastat, on growth of PC3 xenotransplants; and 4) the effect of combining these drugs with a cytotoxic agent, Paclitaxel, because preclinical and clinical studies suggest the potential of taxanes for prostate cancer therapy (33
, 34)
. A significant and unexpected finding was the beneficial synergism between MMP and ROK inhibition.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
95% confluence in F12K medium (Invitrogen) supplemented with 10% FBS, endothelial cell growth supplement (ECGS, Sigma, St. Louis, MO), and heparin (Sigma).
Myosin phosphatase and myosin light chain phosphorylation
PC3 and HUVECs cells were cultured at 37°C RPMI medium supplemented with 10% FBS for 16 h with or without Wf-536, 2 or 10 µM. Cells lysed in lysis buffer with 1% SDS and 50 µg of total protein were immunoblotted with 1:2000 diluted polyclonal IgG rabbit anti-phospho-MYPT1 antibody (Thr696, Upstate Biotechnology, Lake Placid, NY) or 1:1000 polyclonal IgG sheep anti-MYPT1 antibody (Upstate Biotechnology). The blots were reprobed with a 1:2000 rabbit anti-phospho LC20 antibody (S19, Upstate Biotechnology) or a 1:5000 mouse anti LC20 antibody (Sigma). Anti-rabbit or anti-sheep IgG-HRP-labeled secondary antibody (Pierce, Rockford, IL) was used at 1:10,000 dilution, anti-mouse IgM-HRP (CalBiochem, San Diego, CA) at 1:5000 dilution, and reactive bands detected by ECL were quantitated by densitometry (Amersham Biosciences, Molecular Dynamics, Sunnyvale, CA). Phosphorylation (%) indicates relative band intensity of phospho-protein/total protein with phosphorylation in the absence of Wf-536 defined as 100%.
In vitro angiogenesis assays
Three-dimensional vacuole tube formation assays were performed as described (35
, 36)
with minor modifications. HUVECs were cultured to 95% confluence in F12K medium (Invitrogen) supplemented with 10% fetal bovine serum (FBS, Invitrogen), endothelial cell growth supplement (ECGS, Sigma), and heparin (growth medium, Sigma). In some experiments, cells were pretreated for 3 days with 10 ng/mL Clostridium botulinum exotoxin (C3, CalBiochem). HUVECs were rinsed with PBS, trypsinized, and resuspended at a density of 2.1 x 106 cells/mL in growth medium without FBS. Fibrinogen (Sigma) and cells were added to thrombin (Sigma) yielding 10 mg/mL fibrin, 2.5 x 104 cells, and 1 unit/mL thrombin. The plate was gently shaken using the mix function on a plate reader and the clot was solidified at 37°C, 5% CO2 for 15 min. F12K Medium was supplemented with 20% charcoal stripped FBS (Hyclone, Logan, UT), heparin, and ECGS. Various concentrations of Wf-536, C3, and 600 nM sphingosine-1-phosphate (S1P, Biomol, Plymouth Meeting, PA) were added to the medium; 100 µL was added to each appropriate well and the plate incubated for 24 h, at which time the gels were fixed with 3% glutaraldehyde. Fixed gels were mounted on glass slides random fields viewed with phase interference microscopy and the percentage of cells with vacuoles was determined in triplicate by two blinded investigators.
Endothelial cell cord formation was assayed by adding 2 x 105 HUVECs in serum-free media (SFM) to 100 µL of polymerized Matrigel/well in 24-well plates and incubated for 24 h in the presence or absence of 10 µM Wf-536, 10 µM Marimastat, or a combination of 10 µM Wf-536 plus 10 µM Marimastat, each in triplicate. In some experiments, 400 nM S1P was present to stimulate cord formation (17
, 36
, 37)
. Specimens were fixed in 3% paraformaldehyde. To evaluate the effects of drugs, each well was first viewed under dark field illumination on a dissecting microscope that allowed imaging of the entire well, because complexity of cord formation varied from center to periphery. Actin was stained with TRITC-labeled phalloidin (Sigma) at a 1:1000 dilution and cord formation was photographed in random areas midway between the center and periphery of each well on a Zeiss fluorescent microscope with a 5x objective and a Nikon COOLPIX 995 camera.
Endothelial sprout formation was assayed in 3-dimensional collagen gels as described (38)
. Collagen gels were prepared by mixing seven parts 3 mg/mL rat tail collagen (BD Biosciences, Bedford, MA) with two parts 0.1 M NaHCO3 and one part 10x concentrated M199. Collagen gel (100 µL) was placed into 48-well plates and allowed to polymerize overnight at 37°C. Gels were overlaid with 300 µL of serum-free 3 F12K media supplemented with 40 ng/mL vascular endothelial growth factor (VEGF; R&D Systems, Minneapolis, MN), 40 ng/mL hepatocyte growth factor (HGF, Cell Sciences, Norwood, MD), ECGS (15 mg/500 mL, Sigma), heparin (50 mg/500 mL, Sigma), 100 units/mL penicillin, and 100 µg/mL streptomycin (Invitrogen). 1 mm2 pieces of rabbit thoracic aorta were placed endothelium face down on the collagen gel and gels were overlaid with three drops of unpolymerized collagen. After polymerization, the gels were overlaid with 300 µL of the above media without or with 5 or 10 µM Wf-536 or Marimastat or a combination of 5 µM each and 10 µM each. The extent of sprout formation was measured with a calibrated grid in the eyepiece of a phase interference light microscope.
Cell morphology
HUVECs (passage number <12, 2x105 cells/well in the growth media indicated above plus 10% FBS) were plated onto cover glasses in a 12-well plate and incubated at 37°C overnight with or without 5, 10, or 25 µM Wf-536, fixed with 3% paraformaldehyde in PBS, and permeabilized with 0.5% Triton X-100. Actin was stained with TRITC-labeled phalloidin at 1:1000 dilution (Molecular Probes, Eugene, OR). Cells were rinsed in PBS and mounted in 1 mg/mL DABCO anti-bleach (1,4 diazabicylclo [2.2] octane, Sigma).
Cell migration assays
HUVEC migration was assayed in 24-well format cell culture inserts 8.0 µm pore size, polyethylene terephthalate membrane (BD Biosciences, Bedford, MA), coated for 2 h with 5 µg/mL fibronectin (BD Biosciences) dissolved in PBS at room temperature for 2 h or with Matrigel (BD Biosciences) at 200 µg/mL for 2.5 h, according to a method modified from (36)
. HUVECs were released from the culture flask with trypsin and washed with SFM. 5 x 104 cells in 200 µL SFM were allowed to migrate for 4 h at 37°C. Before addition of the HUVECs to the upper chamber, they (except untreated controls) were treated for 15 min with 10 µM Wf-536, 10 µM Marimastat, or 10 µM Wf-536 plus 10 µM Marimastat. The lower chambers contained 750 µL of SFM plus 10 nM EGF and 60 nM S1P as chemoattractants, and drug concentrations matched with the upper wells. After incubation, cells on the upper side of the transwell insert were removed with a cotton swab and the migrating cells were fixed with methanol and stained with crystal violet. Cells on three randomly selected areas on each insert were counted at 400x magnification. Experiments were run twice with four inserts for each condition. The number of migrating cells was normalized to the average number of cells in the control as 100%. The distribution of cells on the filter was homogeneous. PC3 cell migration assays were carried out as above, but membranes were uncoated and cells were treated with 10 and 25 µM Wf-536 or Y-27632.
In vivo experiments: PC3 xenotransplants
The effect of MMP inhibitor (Marimastat) and ROK inhibitor (Wf-536) alone or combined
SCID beige immunocompromised adult male mice, in two series of four groups each, were implanted with subcutaneous (s.c.) 4 wk (0.25 µL/h) osmotic pumps (Alzet pumps, Durect Corp., Cupertine, CA) dispensing sterile 75% DMSO solution (controls) or 10 mg · kg-1 · day-1 Wf-536 plus 6.5 mg·kg-1 · day-1 Marimastat in 75% DMSO or, as single agents, Wf-536 or Marimastat in the above doses. The two series combined included 18 controls and 18 animals receiving the combination (Wf-536 plus Marimastat) and 10 animals each receiving a single treatment (Wf-536 or Marimastat). Two days after implantation of pumps, 106 PC3 human prostate cancer cells in 300 µL chilled Matrigel basement membrane matrix were injected s.c. at the ventral midline, 1 cm below the sternum. After 4 wk, the original pumps were replaced with new pumps filled as above. Mouse body weights and tumor volumes measured with calipers were recorded weekly and animals were killed for moribund appearance or weight loss exceeding 15%.
The effect of Wf-536 and Marimastat on tumor necrosis
The s.c. PC3 tumor model was used to determine tumor histology at 4 wk in 10 animals treated with single agents or a combination of Wf-536 (10 mg · kg-1 · day-1) plus Marimastat (6.5 mg · kg-1 · day-1) administered as described above. At 4 wk, the animals were killed, tumors excised, and processed for histology. The tumors were cut in half, with one-half cut again at right angles through the center; thin sections were cut from the exposed faces. Tumor micrographs were digitized and the necrotic areas as percentage of total tumor area were determined.
The effects of Paclitaxel alone, with Wf-536, Marimastat, or Marimastat-Wf-536 combination
106 PC3 cells in Matrigel were implanted s.c. 2 days before initiating treatment with s.c. osmotic pumps. Six treatment arms were divided into 1) controls (n=10) (75% DMSO, pump); 2) controls with intraperitoneal (i.p.) injection of Intralipid (the hydrolyzed soybean oil vehicle for Paclitaxel, Sigma) (n=20); 3) control (75% DMS0, pump) plus i.p. Paclitaxel (10 mg·kg-1 · day-1) (n=20); 4) Wf-536 (pump) plus i.p. Paclitaxel (n=20); 5) Marimastat (pump) plus i.p. Paclitaxel (n=20); and 6) the combination: Wf-536 plus Marimastat (pump) plus i.p. Paclitaxel (n=20). Paclitaxel 10 mg · kg-1 · day-1 in Intralipid was delivered by i.p. injections over 5 consecutive days during wk 2 and 6. Intralipid, a well-tolerated diluent of a clinical anesthetic, was used to deliver Paclitaxel; we found in preliminary experiments that Paclitaxel 12.5 mg · kg-1 · day-1 or 6.25 mg · kg-1 · day-1 in Cremophor reduced tumor size, but a large number of mice died from unknown cause usually within 1 wk of i.p. injection. Cremophor, a castor oil derivative, is reported to cause anaphylactoid reactions, weight loss and rapid death (33
, 39
, 40)
. The outcomes of the two control groups (pumps only or pumps plus i.p. Intralipid) were not significantly different and were combined for final statistical analysis.
Tumor volume assessment
The tumors’ major and minor diameters were measured with calipers weekly. Tumors were typically uniform in shape and approximated the form of a prolate spheroid whose volume was calculated as 4/3
x major radius x (minor radius) (2)
.
Statistics
A random coefficient regression model with fixed effects for time, group, group by time interactions, and a random effect for time was used to model log tumor volume for each experiment (41)
. The model assumes each animal has a unique profile of tumor growth over time, such that observations of tumor size on the same animal are correlated over time and each animal contributes information about the overall group profile for the group to which it belongs. Changes in tumor volume were assumed to be linear, and include a potential change in slope at wk 4 when the infusion pump change occurred in all animals. No correction for "missing" (due to animals that died or were killed for tumor-related reasons) was incorporated. The Tukey-Kramer (42)
adjustment for multiple comparisons was applied to P values and confidence intervals for between group comparisons at 4, 6, and 8 wk, as well as for analysis of tumor necrotic areas and comparison of cell migration. Analysis of variance (ANOVA) was used for analysis of tumor necrotic areas and comparison of cell migration; between group differences were adjusted with Tukey-Kramer adjustment for multiple comparisons.
Proportional hazards regression (43)
was used to compare the length of time until natural death or killing of the animals. Death during surgical implantation or exchange of an infusion pump was not considered an outcome of interest and those animals were treated as censored. Hazard ratios indicate the rate comparison in the number of deaths per unit time between two groups. Experiments were analyzed using the week number of death (Table 1
) or the actual day of death for all animals (Table 2
). The Pocock (44)
adjustment for repeated significance testing was applied to the confidence intervals.
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), Wf-536 significantly decreased in PC3 cells and HUVECs, phosphorylation of the inhibitory site Thr696 (45)
of MYPT1 as well as phosphorylation of myosin LC20 (Fig. 2)
. In PC3 cells, Wf-536 (2 and 10 µM) decreased P-MYPT1 by 35% (P<0.003) and 70% (P<0.0003), respectively, whereas P-LC20 decreased by 60% (P<0.003) and 85% (P<0.0003). In HUVECs Wf-536 (10 µM) decreased P-MYPT1 by 42% (P<0.001) accompanied by a 72% decrease in P-LC20 (P<0.03).
|
In vitro angiogenesis
Endothelial cells suspended in 3-dimensional fibrin lattices form intracellular vacuoles (Fig. 3
A). These cells migrate and assemble into branching tube-like structures with lumens thought to arise by fusion of the individual cell vacuoles, recapitulating in vivo angiogenesis. Vacuole formation in individual cells and the formation of tube-like structures (Fig. 3B, C
) were significantly inhibited by 5 and 10 µM Wf-536 (Fig. 3D, E
). Endothelial cells treated with Wf-536 frequently displayed long processes (Fig. 3E
). C3 exotoxin that ADP-ribosylates and inhibits RhoA markedly inhibited vacuole and tube formation (Fig. 3F
). S1P enhanced tube formation (Fig. 3B, C
), but this was not quantitated. Wf-536 potently inhibited the formation of vacuoles and tubes in the S1P treated preparations. Data for vacuole formation are summarized in Fig. 3G
.
|
The effects of Wf-536 and Marimastat, either alone or in combination, on cord formation by endothelial cells on Matrigel are shown in Fig. 4
. When viewing the surface of each well with a dissecting microscope with dark field, patterns under different conditions were clearly distinguishable but similar for each replicate. Stained with TRITC phalloidin under fluorescence microscopy, the cord arms with Marimastat treatment were somewhat shorter than controls (Fig. 4)
. Wf-536 (10 µM) much more dramatically shortened cord arms, which were disrupted with irregular outlines, and had long thin strands, similar to those induced by Y-27632 (17)
. The effects of combining 10 µM Wf-536 with 10 µM Marimastat were even more extreme (Fig. 4)
.
|
Endothelial proliferation and sprout formation from aortic explants embedded in collagen gels were inhibited by Wf-536 and Marimastat (Fig. 5
). Marimastat (5 or 10 µM) inhibited sprout formation only modestly, whereas Wf-536 inhibited sprout formation by 30 and 40% at 5 and 10 µM, respectively. Sprout formation was inhibited 80–95% by combining the drugs at 5 and 10 µM, indicating a marked degree of drug synergism.
|
Cell morphology
Overnight treatment of HUVECs with Wf-536 induced significant changes in cell morphology even at the lowest dose used, 5 µM (Fig. 6
). The control cells had a typical spread out morphology with extensive stress fibers, which were absent in treated cells. Wf-536 induced irregular cell boundaries with multiple processes, some terminating in long strands (Fig. 6B
). Actin was pronounced at the cell periphery, particularly at the ruffling edges of treated cells.
|
Cell migration
Cell migration of HUVECs in transwell assays, in which insert membranes were coated with fibronectin or Matrigel, was retarded by Wf-536, Marimastat, and a combination of the drugs (Fig. 7
). S1P and EGF were present in all lower wells to stimulate sufficient migration during the 4 h assay. The surface distribution of cells was homogeneous in controls and treated wells, and the difference between treated migrating cells and controls was obvious by inspection. Inhibition of migration by each treatment was highly significant (P<0.0001) and slightly less for Marimastat on fibronectin (P<0.001). The largest effect, 43% inhibition, was found with the combination treatment on Matrigel coated membranes. PC3 cell migration across uncoated inserts was inhibited by 80 and 87% by 10 and 25 µM Wf-536, respectively (P<0.0001, data not shown).
|
In vivo effects on PC3 xenotransplants
The effect of early treatment with ROK and MMP inhibitors, singly or in combination, on tumor volume and survival
This experiment was carried out over 8 wk and followed tumor growth of 106 PC3 cells in 300 µL Matrigel inserted s.c. 2 days after implantation of osmotic pumps delivering diluent (controls), Wf-536, or Marimastat or the two-drug combination (see Materials and Methods). The overall test for the effect of group on tumor volume indicates that group profiles over time were significantly different (P<0.001). Tumor volumes (Table 1
and Fig. 8
) were highly significantly smaller in the group treated with the combination of MMP and the ROK inhibitor than in animals treated with either inhibitor alone or untreated controls at wk 4, 6, and 8, P < 0.004, 0.0003 and 0.005, respectively. Given the number of survivors, we consider 6 wk data as statistically most robust (Table 1)
. Single agents Wf-536 or Marimastat did not have a statistically significant effect on tumor volumes. Survival data showed an increased survival benefit with combination therapy compared with controls (P<0.03 and Hazard Ratio of 2.7).
|
Tumor necrosis induced by Wf-536 and Marimastat
At 4 wk after tumor implantation, tumor necrosis increased with each treatment. The percent necrotic area/tumor cross section was 3.6-fold greater in the animals treated with the combination of Wf-536 plus Marimastat than in untreated controls (P=0.009) (Fig. 9
). Marimastat treatment alone increased tumor necrosis P < 0.03 whereas a positive trend with Wf-536 did not reach statistical significance at this 4 wk point. For comparison, total tumor volume at 4 wk (Table 1)
was smaller (P=0.004) with the combination treatment compared with controls. Allowing for the increased necrosis with treatment, viable tumor volumes corrected for necrotic volume would be much smaller than measured with external calipers. A typical necrotic area in a tumor obtained from an animal receiving combination treatment is shown in Fig. 9
.
|
The effect of Paclitaxel alone or in combination with the MMP and ROK inhibitors
In animals whose treatment began 2 days after PC3 tumor implantation, Marimastat plus Wf-536 combined with Paclitaxel caused a highly significant (P<0.0001 for each time point) reduction in tumor volume compared with controls (Table 2)
. This combination was significantly more effective than Paclitaxel alone (P<0.0001) by 6 and 8 wk. There was only minimal or no significant tumor growth detectable during the initial 8 wk of active treatment with the combination of Paclitaxel plus MMP and ROK inhibitors, whereas other groups showed significant tumor growth beginning 5 wk after tumor implantation (Fig. 10
A). Even well after termination of therapy at 8 wk, tumors treated with the combination remained smaller than tumors of the only other surviving group, the one treated with Paclitaxel plus Marimastat (Figs. 10A, C
). Discontinuation of Paclitaxel appeared to stimulate tumor growth (Figs. 10A, B
).
|
Survival advantage of treatment with combination plus Paclitaxel compared with controls or with treatment with Paclitaxel alone was highly significant (P=0.0006 and P=0.004, respectively). The combination treatment, Wf-536 plus Marimastat plus Paclitaxel, showed a trend of improved survival compared with treatment with Marimastat plus Paclitaxel (hazard ratio 1.52), but this difference did not reach statistical significance (Fig. 10C
).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
ROK-mediated phosphorylation of MYPT1
The presence of phosphorylated MYPT1 in PC3 and HUVEC cells, its reduction and of myosin LC20 phosphorylation by Wf-536 are evidence of the operation of a major cellular mechanism of Rho-ROK (Fig. 1)
. This result supports the conclusion that ROK and myosin II phosphatase modulate the motility of cancer and endothelial cells and, through it, metastasis and neo-angiogenesis. ROK phosphorylation of threonine 696 on MYPT1 inhibits myosin phosphatase catalytic activity (45)
, although another site, serine 854, is phosphorylated in MDCK cells stimulated with phorbol esters (46)
. In avian tissues, regulation may be MYPT isoform dependent (47)
.
In vitro effects of Wf-536 and Marimastat: Inhibition of angiogenesis and cell migration
The earliest phase of angiogenesis is thought to be the formation in endothelial cells, of vacuoles coalescing into the lumen of newly formed blood vessels (35
, 48
49
50)
. This very early phase of angiogenesis was inhibited by the bacterial exoenzyme C3 that ADP ribosylates and inactivates the Rho·GTPase; inhibition of vacuolation by Wf-536 implicates ROK RhoA in the early phase of angiogenesis (Fig. 4)
. Cdc42 and Rac without Rho have been reported to mediate for vacuole and lumen formation (35)
, perhaps reflecting the use of different agents to stimulate angiogenesis. Chemotaxis of endothelial cells is differentially regulated: S1P-induced migration is inhibited by C3, dominant negative Rho, or Y-27632 (37)
but VEGF-induced migration is not. Thrombin present in the fibrin clot stimulates cell migration, a component of lumen formation, through the Rho/ROK pathway (51)
and increases MLC20 phosphorylation in HUVECs (52)
. Thus, there are multiple signaling pathways targetable for inhibiting tumor angiogenesis and migration (53)
, and they may vary during tumor progression. We have previously shown that organization of endothelial cells into cord-like structures on Matrigel is inhibited by a related ROK inhibitor, Y-27632 (17)
. Wf-536 similarly markedly inhibited cord formation, whereas Marimastat had a slight but reproducible effect, possibly by preventing matrix breakdown. Cord formation was most markedly inhibited by the combination of Wf-536 plus Marimastat, suggesting that cell migration and MMP activity participate in the process. MMP activity of endothelial cells increases in 3-dimensional collagen lattices (54)
. Wf-536 induced a marked change in the morphology of HUVECs resulting in cells with long tails, previously seen in PC3 cells treated with Y-27632 and consistent with the requirement for myosin II-based contraction to release focal contacts at the rear of migrating cells (17
, 55)
. Wf-536 inhibited the S1P plus EGF-stimulated migration of HUVECs, further implicating ROK-modulated myosin II activity in angiogenesis. The MMP inhibitor decreased cell migration through both substrates. A combination of Marimastat and Wf-536 was more effective than the single agents although not additive, possibly reflecting the nonlinearity of this qualitative assay. The ability of Marimastat to inhibit matrix breakdown and up-regulation of VEGF by MMPs (57)
and of Wf-536 to prevent the earliest phase of angiogenesis (present study) and the migration of tumor (17)
and endothelial cells may have been responsible for the beneficial, enhanced inhibition of angiogenesis in the in vitro assays and on sprout formation. If correct, this conclusion supports the use of combined therapies directed against multiple targets.
Inhibition of tumor growth in vivo
Inhibition of in vivo tumor growth by MMP and ROK inhibitors was highly significant both without a cytotoxic agent and in combination with Paclitaxel. A revealing observation was that even without Paclitaxel, the combination of Wf-536 plus Marimastat inhibited tumor growth much more effectively (see Results) than either inhibitor alone, reducing tumor size by >twofold vs. untreated controls. In fact, this noncytotoxic therapy was equally or more effective in reducing tumor size as cytotoxic treatment with Paclitaxel (compare Fig. 8
with Fig. 10A
) and as effective as Paclitaxel (12.5 mg · kg-1 · day-1 in Cremaphor; data not shown). This may be related to the only moderate sensitivity of PC3 tumors to near maximally tolerated doses of Paclitaxel (33)
. Given the extensive tumor necrosis present in combination-treated animals (Fig. 9)
, it is probable that the treatment significantly reduced the number of viable tumor cells capable of multiplication, invasion, and, eventually, metastasis. Treatment with Wf-536 and Marimastat increased survival compared with untreated animals (hazard ratio=2.7), but 7 of 18 treated animals died by wk 8 (Fig. 8)
(compared with the 1 survivor of 17 in the control group). This loss of animals may have been the consequence of significant tumor growth at 8 wk despite therapy and/or complications of the surgery for implantation of osmotic pumps. Nevertheless, it will be important to determine the combined toxicity levels of matrix metalloproteinase and ROK inhibitors at their minimal effective (rather than maximum tolerated) dose before introduction into clinical trial.
Tumor necrosis and inhibition of in vivo tumor growth were likely the result of the anti-angiogenic effects of ROK and MMP inhibition. In addition to the anti-angiogenesis affects of Wf-536 (see above), Marimastat may have contributed to anti-angiogenesis (56)
, possibly by inhibiting up-regulation of VEGF by MMPs (57)
. The tumor necrosis of xenotransplants observed in animals treated with another MMP inhibitor (58)
supports this possibility.
Combining Paclitaxel with Wf-536 and Marimastat was remarkably effective in inhibiting tumor growth and improving survival. Of potential clinical importance was that the rapid tumor progression following termination of treatment with Paclitaxel alone after 8 wk was absent in animals that received adjunct treatment with the Wf-536 and Marimastat combination that conferred a highly significant (P=0.0004) survival benefit compared with untreated controls and animals treated with Paclitaxel alone (P=0.003). The Marimastat plus Paclitaxel combination significantly limited tumor growth and reduced survival hazard. Combining Paclitaxel with another MMP inhibitor (AG3340) had similar benefits in the treatment of Bf16–10 tumor (58)
. These results support the potential value of MMP inhibitors as adjuncts with taxane and/or other cytotoxic therapy for prostate cancer and perhaps other cancers. In our study, the combination of the ROK (Wf-536) and MMP (Marimastat) inhibitors was most successful in reducing tumor growth. Preclinical studies suggest that inclusion of other agents (8
, 59
60
61)
may be effective as adjuncts to cytotoxic drugs in therapeutic combinations.
Clinical implications
Used alone, the ROK inhibitor Y-27632 improved survival in an earlier study in which intercardiac injection of PC3 cells resulted in extensive macroscopic metastases inhibited by the drug (17)
. The two studies are not strictly comparable, as the present approach did not produce macroscopic metastases. However, the survival benefit of combining Wf-536 and Marimastat (present study) was significant. It will be important to determine in a metastatic model whether the combination of a taxane with ROK and MMP inhibitors inhibits tumor dissemination and improves survival. Also important for future clinical considerations is that treatment of animals was limited by their intolerance for osmotic pumps over 4 wk (17)
or 8 wk (present study) requiring two surgeries and anesthesia. Wf-536 and Marimastat are suitable for oral delivery for more prolonged therapy in humans. Marimastat has already been used in clinical trials, and the low animal toxicity of ROK inhibitors (25)
suggests their suitability for human use. The clinical implications of our results support the potential value of combining (possibly multiple) adjunct therapies with cytotoxic agents for the early treatment of human prostate cancer and probably other malignancies.
|
|
ACKNOWLEDGMENTS |
|---|
Received for publication July 24, 2002. Accepted for publication October 15, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y. Mizuno, E. Isotani, J. Huang, H. Ding, J. T. Stull, and K. E. Kamm Myosin light chain kinase activation and calcium sensitization in smooth muscle in vivo Am J Physiol Cell Physiol, August 1, 2008; 295(2): C358 - C364. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-L. Lin, A. Chiang, D. Chang, and S.-Y. Ying Loss of mir-146a function in hormone-refractory prostate cancer RNA, Marc |
