Simvastatin induces degradation of the extracellular matrix in human leiomyomata: novel in vitro, in vivo, and patient level evidence of matrix metalloproteinase involvement

Abstract

Objectives:

To assess the effect of simvastatin on uterine leiomyoma growth and extracellular matrix (ECM) deposition.

Design:

Laboratory analysis of human leiomyoma cell culture, xenograft in a mouse model, and patient tissue from a clinical trial.

Setting:

Academic research center.

Patient(s):

Tissue culture from human leiomyoma tissue and surgical leiomyoma tissue sections from a placebo-controlled randomized clinical trial.

Intervention(s):

Simvastatin treatment.

Main Outcome Measure(s):

Serum concentrations, xenograft volumes, protein expression, and apoptosis markers.

Results:

Mice xenografted with 3-dimensional human leiomyoma cultures were divided as follows: 7 untreated controls, 12 treated with activated simvastatin at 10 mg/kg body weight, and 15 at 20 mg/kg body weight. Simvastatin was detected in serum from mice injected at the highest dose. Xenograft volumes were significantly smaller (mean 53% smaller at the highest concentration, p < 0.05). There was dissolution of compact ECM, decreased ECM formation, lower collagen protein expression, and increased apoptosis in xenografts. Membrane type 1 matrix metalloproteinase (MT1-MMP) was increased in vitro and in vivo. Matrix metalloproteinase 2 (MMP2) and low density lipoprotein receptor related protein 1 (LRP1) were increased in vitro.

Conclusions:

Simvastatin exhibited antitumoral and proapoptotic activity, with extracellular matrix degradation and decreased leiomyoma tumor volume in vivo. Activation of the MMP2/MT1-MMP/LRP1 pathway may explain these findings.

Introduction

Uterine fibroids (UFs), also known as uterine leiomyomas, are the most common tumor affecting women with an overall prevalence of up to 77% (1). Approximately 25% to 50% of women develop clinically significant symptoms (2) which are associated with decreased quality of life (3). These benign tumors are seen in over 50% of ultrasound studies (4) and have been found in more than 70% of hysterectomy specimens (5,6). In the United States, there are approximately 200,000 hysterectomies for UFs performed annually which accounts for 37% of all hysterectomies (7). Nationally, this disease is estimated to cost up to $34.4 billion annually (8). Definitive therapy consists of hysterectomy (9), but factors such as desired uterine preservation, desired future fertility, or poor functional status may preclude hysterectomy. Non-surgical treatment options exist with varying indications, risks, and benefits; however, novel therapies are needed.

The hallmark of uterine leiomyomatous disease is abundant and disorganized extracellular matrix (ECM) deposition (10). Any agent that can regulate formation and/or lead to ECM deterioration may have therapeutical potential. One such therapeutic candidate is simvastatin, as it is known to be antiproliferative and proapoptotic to leiomyoma cells in vitro with an associated decrease in ECM protein expression (11–13). Antiproliferative properties and decreased xenograft tumor volumes have also been shown in vivo (14), however, the effect on ECM proteins in vivo has not been studied.

Simvastatin influences signal transduction pathways commonly associated with leiomyoma fibrosis (13,15,16). In vitro, standard leiomyoma tissue culture models are used to characterize the genomic and proteomic phenotypes characteristic of leiomyoma cells (17). However, they are insufficient for analyzing ECM formation and maintenance. In vivo models provide better insight into the impact of long-term simvastatin exposure on ECM maintenance (18). Furthermore, in vivo models can also help assess the effects of therapies on tumor size, angiogenesis, and apoptosis (19). For this reason, we utilized an established human leiomyoma xenograft murine model (18) and hypothesized that simvastatin would decrease ECM protein deposition, decrease tumor size, and increase apoptosis. Furthermore, we characterized cellular mechanisms associated with the effects of simvastatin to further our understanding of critical molecular pathways associated with uterine leiomyomata.

Materials and Methods

In Vivo Methods

Animals

NOD/SCID female mice (Strain Code 394) were purchased from Charles River Laboratories (CRL, Frederick, MD). The mice were 8–9 weeks of age and weighed between 22–26g. Mice were ovariectomized at CRL. Studies performed were approved by the Uniformed Services University of the Health Sciences (USUHS) Institutional Animal Care and Use Committee (protocol# OBG 20–023).

Hormone Pellet Implantation

Combined hormone pellets [17β-estradiol, E2 (0.1mg); progesterone, P4 (25mg)/90 day release] were obtained from Innovative Research of America (IRA; Sarasota, FL). The pellets were surgically implanted subcutaneously and laterally in the nape between the ear and shoulder of the mice.

3D Culture Growth

Well-established myometrial and leiomyoma cell lines from patient-matched tissues were used (20). Cell lines had previously been immortalized using PA317 LXSN 16E6E7 (American Type Culture Collection, CRL-2203), encoding HPV-16 (human papillomavirus) E6 and E7 gene open reading frames (20). Cells were cultured in fresh complete medium: Dulbecco’s Modified Eagle’s Medium and Ham’s F-12 Nutrient Mixture (DMEM/F12 with phenol red and with 15mM HEPES buffer; ThermoFisher Scientific, Waltham, MA) with 10% Fetal Bovine Serum (FBS, HyClone) and 1X Normocin (Invivogen, San Diego, CA).

Cells were grown in a 3D rat collagen I matrix (Cultrex, 5 mg/mL; R&D Systems, Minneapolis, MN) following an established protocol with minor modifications (17,19,21,22). All working solutions were kept on ice because collagen polymerizes at room temperature and all procedures were performed in a biosafety hood. Briefly, the collagen gels were prepared at 3 mg/mL concentration. For 3.5 mL total volume: 2.1 mL of collagen solution, 0.35 mL of 10X sterile phosphate buffered saline (PBS, pH 7.4), and 0.0485 mL of 1N sodium hydroxide (NaOH) were mixed in a cold 50 mL tube. Myometrium or leiomyoma cells were added at a concentration of 5×10⁵ cells/mL collagen gel, making sure that volume of cell suspension was 10% or less (≈0.35 mL) of the final solution. Cold complete medium was used to make the total volume. The mixture was plated in 96-well cell culture plates (0.1 mL/well) and allowed to polymerize for at least 30 minutes at 37°C. Complete medium was added to the culture and the 3D cultures were maintained in a humidified 95% air/5% CO₂ incubator at 37°C. Culture medium was replaced every day. The cultures were grown for 7–10 days and demonstrated a network of activity growing spindle shaped cells, before implantation in the flanks of mice.

Subcutaneous Implantation of 3D Cultures

Growth of 3D cultures in 96-well cell culture plates optimized the size of the initial implant in the mice. Immediately before implantation the 3D cultures of leiomyoma and myometrium were washed with sterile PBS (pH 7.4). Excess liquid was absorbed with Kimwipes. The 3D cultures were implanted subcutaneously (SC) in the murine flanks under anesthesia using sterile technique. Each mouse received 1–2 leiomyoma grafts in the right flank and 1 patient-matched myometrial graft in the left flank per an Institutional Animal Care and Use Committee (IACUC) approved protocol.

Simvastatin Activation

Simvastatin was activated from its lactone prodrug form (SIM) to its active β-hydroxyacid (aSIM) form as previously described with minor modifications (23). 5 mg simvastatin was dissolved in 100 μL of ethanol then 150 μL of 0.1 M NaOH was added. The solution was incubated at 50°C for 2 h and then neutralized to pH 7.0 with 1 M hydrochloric acid. The resulting solution was diluted to 1 mL with distilled water, and aliquots were stored at −80°C.

Intraperitoneal Simvastatin Injections

After 3 days of acclimation, the mice were randomized into 4 groups: 7 untreated controls, 12 treated with activated simvastatin at 10 mg/kg body weight (aSIM10), and 15 at 20 mg/kg body weight (aSIM20). These doses are typical for murine studies (14) and are higher than those required for humans due to increased 3-hydroxy-3-methylglutaryl-CoA reductase expression after statin administration (24). Untreated controls did not receive injections. Mice in the 10 and 20 mg/kg treatment groups received daily intraperitoneal injections at the same time of day for 4-weeks. Injections were alternated between the lower right and left abdominal quadrants. Injection volumes were < 0.05 mL/dose.

Collection of Xenografts

At the end of the experimental period, the mice were anesthetized and euthanized. Xenografts were measured in 2 dimensions using calipers then each graft was excised sharply along with surrounding mouse skin. The tissue removed from the mice was placed in 10% buffered formalin for immunohistochemical studies.

Serum Simvastatin Measurements by Liquid Chromatography Mass Spectrometry (LCMS)

Blood was collected prior to xenograft placement and prior to euthanization (1 hour after the last simvastatin injection) using the Microvette CB system (Kent Scientific). The initial draws and draws from untreated mice were used as controls. The serum was separated, aliquoted, and stored at −80°C.

Stock solutions of simvastatin (SIM), d6-simvastatin (dSIM), simvastatin acid (SIMA), and d6-simvastatin acid (dSIMA) were prepared in acetonitrile with concentrations in the mg/mL range. Stocks were diluted in a 1:1 water/acetonitrile solution with 0.2% formic acid to concentrations ranging from 0.5 to 2 μg/mL and used to make an internal standard (IS) mixture consisting of 400 ng/mL each of dSIM & dSIMA in 1:1 water/acetonitrile with 0.2% formic acid. Mouse serum samples were processed by taking 40 μL of serum and adding 4 μL of the IS mixture and then 116 μL of acetonitrile to precipitate proteins. This diluted the original SIM and SIMA concentrations 4-fold. Samples were then vortexed, incubated on ice for 15 minutes (with additional vortexing at 7.5 and 15 minutes), centrifuged at 17,000 g for 5 minutes at 4°C to separate precipitated proteins, and 128 μL of supernatant was removed.

High-Performance Liquid Chromatography (HPLC) was carried out on an Agilent 1200 series HPLC system using a YMC-Triart C18 column (2 × 100 mm, 5 μm particle size, 12 nm pore size) with a YMC-Triart C18 guard column (2.1 × 10 mm; same particle and pore sizes) with a flow rate of 300 μL/min at 30°C. The HPLC solvents were A = water, B = acetonitrile (both with 0.2% formic acid). The HPLC gradient consisted of a ramp with MS detection from 50% to 100% solvent B in 10 minutes. Solvent composition was isocratically held at 100% B for 4 minutes to flush the column of any residual non-polar material. The solvent was then ramped back down to the original 50% B composition over 1.5 minutes and held there for 8 minutes to re-equilibrate the column before the next injection. A two-way valve diverted the HPLC flow from the MS ion source to waste during the non-polar column flush, the ramp-down, and the re-equilibration.

Mass spectrometry (MS) was performed using a SCIEX 4000 Q-Trap triple quadrupole/linear ion trap spectrometer. To optimize detection, IS stock solutions were infusion into the spectrometer’s Turbo-V ESI source at rates between 10 and 25 μL/min using a Harvard Apparatus syringe pump (model 55–2222). SIM and SIMA both gave sufficient signal in positive ion mode.

Initially, 10 μL of the supernatant was injected into the LCMS system. SIM and SIMA signals were either very low or undetectable, so 100 μL of supernatant for each sample was dried using a Savant SPD1010 SpeedVac System (Thermo Scientific) and resuspended in 31.25 μL of 1:1 water/acetonitrile/0.2% formic acid thus re-concentrating the sample 3.2-fold. Then 10 μL of the resuspended serum supernatant was injected into the LCMS using the foregoing conditions.

SCIEX Analyst 1.7 software was used for experimental control and analysis. Source/gas parameters and multiple reaction monitoring transitions are presented in supplementary tables 1 and 2.

Clinical Patient Tissue

Surgical leiomyoma tissue sections were obtained after scheduled hysterectomy or myomectomy from participants in a phase II, double-blinded, placebo-controlled randomized clinical trial designed to study the effects of a 12-week course of oral daily simvastatin 40 mg on uterine fibroids in menstruating patients ages 18 to 55 years-old experiencing heavy menstrual bleeding, pelvic pain, or compressive symptoms secondary to uterine leiomyomata (Johns Hopkins University School of Medicine IRB#: IRB00149869; ClinicalTrials.gov identifier: NCT03400826; PI: Mostafa Borahay). At least one fibroid greater than 3cm in diameter diagnosed by ultrasound or magnetic resonance imaging was required to participate with an unlimited maximum number of leiomyomata. Patients that experienced pregnancy, lactation, hormonal treatment, rapid tumoral growth, suspected leiomyosarcoma, or urgent surgery were excluded from the study. We received tissue from the placebo and treatment groups.

In Vitro Experimental Design

Experimental design for zymography and western blot studies were similar except for the differences noted in individual sections below. Cells were plated in 6-well plates at a concentration of 2×10⁴ cells/well and maintained in complete medium at 37°C and 5% CO₂. Once the monolayer cultures reached approximately 60% confluence, complete media was replaced with charcoal-stripped media for 24 hours [DMEM/F12 (no phenol red) with 10% charcoal treated FBS and 1X Normocin]. The cells were treated with activated simvastatin as detailed below. Control cells were exposed to the highest concentration of the solution used for activation of simvastatin. Protein lysates were collected as previously established(25) with minor modifications. Cell monolayers were washed with ice-cold PBS before lysis using the Pierce RIPA radioimmunoprecipitation assay with Halt protease and phosphatase inhibitors, and ethylenediaminetetraacetic acid (EDTA). The samples were scraped, collected, and prepared for either zymography or western blotting as detailed below. Protein concentrations were measured using the Pierce bicinchoninic acid assay (BCA) protein assay kit (ThermoFisher Scientific). Aliquots of the lysate and concentrated media were stored at −80°C.

Gelatin Zymography and Analysis

Using the experimental design above, leiomyoma cells were plated, exposed to activated simvastatin at 0, 0.1, 1, 2, and 4 μM for 24 hours, lysed, scraped, collected, vortexed for 5 seconds, and centrifuged at 13,000 g for 20 min at 4°C. For zymography of conditioned media, activated simvastatin treatment was performed in serum free media. Conditioned media (2–4 mL) was collected before the addition of the compound and at the end of 24-hour treatment. To remove cellular debris, the media was centrifuged at 1,000 g for 10 min at room temperature. The supernatant was collected and 1X protease and phosphatase inhibitor was added. The media was concentrated to 0.5–0.7mL using Amicon Ultra-4 centrifugal filter units (MilliporeSigma). Protein measurement was done as described above.

The previously described zymography protocol (19) was adapted. Novex 10% Zymogram Plus (gelatin) protein gels (ThermoFisher Scientific) were loaded with 10μg protein per lane and run for 90 min at 125 mV in Tris-glycine SDS running buffer. Gels were placed in renaturing buffer for 45 min and then in developing buffer overnight at 37°C (Novex Zymogram, ThermoFisher Scientific). The gels were stained with 0.5% Coomassie Blue G-250 for 2 hours and imaged. The gels were de-stained in distilled water for 1 hour or with a 30% methanol 10% acetic acid solution to reveal clear bands of gelatinolytic activity. Gels were digitally photographed and subjected to band densitometry analysis using ChemiDoc XRS+ imaging system with ImageLab software (QualityOne software 2.1; Bio-Rad) to give semi-quantitative evaluation of the enzymatic activity of pro-MMP2 (64 kDa) and active MMP2 (60 kDa). Data was presented as relative density units of adjusted volume data relative to control pro-MMP2. To assess the amount of protein that was loaded per lane we transferred the section of gel containing COX IV (17 kDa) onto a nitrocellulose membrane and followed the western blotting protocol below.

Western Blotting (WB)

Using the experimental design above, leiomyoma cells were plated and exposed to activated simvastatin at 0, 0.001, 0.01, 0.1 and 1 μM for 4, 6, 24, and 48 hours. Cells were lysed, sonicated, and centrifuged at 14,000 g for 20 minutes at 4°C to pellet large cellular debris. Then protein was quantified, denatured, and stored at −80°C. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and western immunoblotting were performed as previously described (25). Blocking was performed with 10% milk for 30 minutes to 1 hour. Incubation of primary antibodies was performed overnight at 4°C and of secondary antibodies for 1 hour at room temperature unless otherwise noted: MT1-MMP (MMP14, Abcam ab271840, 1:750), LRP1 (Abcam ab92544, 1:5,000), COX IV (Cell Signaling Technology 5247S, horseradish peroxidase [HRP] conjugated, 1:2,000 for 1 hour at room temperature), anti-rabbit (Abcam ab205718, HRP conjugated, 1:10,000). Experiments were repeated a minimum of 3 times with multiple replicates.

Immunohistochemistry (IHC)

Xenografts were either cut using scissors or punched out using a tissue/biopsy punch (6mm) and directly transferred to 10% buffered formalin. The samples were submitted for sectioning, hematoxylin & eosin (H&E) staining of representative sections, and paraffin embedding to the Biomedical Instrumentation Center at USUHS. Slides were deparaffinized twice in xylene for 10 minutes each then rehydrated in graded ethanol washes (95%, 80%, and 70%) for 3 minutes each followed by placement in distilled water for 5 minutes. Immunohistochemical staining was performed using the Vectastain Elite ABC-HRP kit (Vector Laboratories, Newark, CA) according to the product manual with the minor modifications as previously detailed (26). Endogenous peroxidases were blocked using BLOXALL (Vector Laboratories) for 1 hour. Antigens were retrieved with a pH 6 citrate buffer heated to 95°C for 45 minutes. Slides were washed in 1X Tris-buffered saline (TBS) between steps. A 1X TBS, 3% bovine serum albumin, and 0.1% Tween-20 solution was used for antibody and blocking dilution. Sections were blocked for 1 hour with 150 μL normal serum in 10mL of diluent. They were incubated in primary antibodies overnight at 4°C: MT1-MMP (MMP14, Abcam ab271840, 1:400). They were incubated for 30 minutes using 1 μL secondary biotinylated antibody and 3 μL normal serum per 200 μL of diluent. Incubation in DAB substrate (3–3’-diaminobenzidine; Vector Laboratories) for 5 minutes was used for peroxidase staining. Sections were counterstained with Gill’s Hematoxylin Solution No. 3 (Sigma-Aldrich) and bluing reagent then dehydrated, cleared in xylene, and coverslip mounted with Permount (Fisher Scientific).

Connective tissue was stained using Abcam trichrome stain kit (ab150686) following the manufacturer’s recommendations. After deparaffination and rehydration, slides were incubated in Bouin’s fluid, rinsed, stained with Weigert’s iron hematoxylin, rinsed, de-stained with Biebrich scarlet/acid fuchsin, differentiated with phosphomolybdic/phosphotungstic acid, stained with aniline blue, rinsed, de-stained in acetic acid, dehydrated, cleared in xylene, and mounted.

Apoptosis

Methyl green pyronin-Y (MGP-Y) was used to identify apoptotic cells. MGP-Y is a histochemical stain which is used in oral pathology for this indication (27). Apoptotic cells shrink, cytoplasm stains deep pyronin red, and nuclei undergo chromatin condensation (pyknosis) with methyl green staining (28). Prior to staining, slides were deparaffinized and rehydrated as above. Sections were then incubated with MGP-Y solution (Sigma-Aldrich) for 5 minutes and rinsed in distilled water. Tissue slides were blotted with Whatman 3 mm blotting paper (MilliporeSigma) until completely dry before placing in acetone for 10 seconds. Slides were dipped in equal parts acetone/xylene then in several changes of 100% xylene prior to mounting.

Statistical Analysis

Peak area averages and standard deviations were calculated for LCMS samples that gave observable and quantifiable peaks. For xenograft volume calculation we used formula V = (W² × L)/2, where V is tumor volume in mm³, and W and L are tumor width and length, respectively, in mm. We reported absolute xenograft volumes and percent difference compared to controls [(V-V_ctr)/V_ctr x 100 where V_ctr is the average control volume] as the initial amount of tissue varied between the experiments. Student t-test, one way ANOVA with multiple comparisons, and Kruskall-Wallis tests were used for statistical evaluation, where appropriate. Logarithmic transformation of tumor volumes was performed prior to ANOVA to satisfy statistical test requirements. P-values < 0.05 were considered significant. Zymography analysis was performed with QualityOne 2.1 (Bio-Rad) and data is presented as relative density units. WB analysis was performed using ImageLab v6.1 (Bio-Rad) and data is presented as fold difference of relative density units between treated and untreated samples normalized to COX IV.

Histology Score (H-score) Determinations

We modified an H-score used primarily for nuclear staining in breast cancer samples (26,29). Photographs of stained tissue sections were taken using a Zeiss Axiocam Imager M2 light microscope. Images were captured, saved, and randomly placed in a PowerPoint presentation for review with the use of 64 bit Stereo Investigator Bright Field software (MBF Bioscience). Each slide was photographed to produce 3 random images at 40X magnification.

Six investigators were asked to evaluate slides. Each observer anonymously reviewed and scored each image on a scale of 0 (no staining) to 4 (intense staining for >90% of the slide) based on a set of three parameters: absence or presence of staining, stain intensity, and percentage of the slide being stained. These scoring criteria are based on consensus scoring agreed upon by our lab. The H-score for each slide was assigned using the following formula:

$$\left[0\times \left(0/\sum \right)+1\times \left(1/\sum \right)+2\times \left(2/\sum \right)+3\times \left(3/\sum \right)+4\times \left(4/\sum \right)\right]\times 100$$

Where “Ʃ” equals the total of the numbers recorded by each of the observers for each of the slides evaluated. A final score gives a range from 0 to 400, for each respective slide.

Results

Serum simvastatin acid levels are similar to human therapeutic ranges

In order to confirm that simvastatin treatment resulted in levels seen in humans (2.7–5.6 ng/mL seen after oral simvastatin 40 mg administration (30,31)), we assessed simvastatin acid concentration in treated mice serum. We found that mice treated at 20 mg/kg had on average 3.3 ± 2.7 ng/mL of simvastatin acid in serum. Measurements for the 10 mg/kg and untreated control groups were below the threshold of quantitation.

Simvastatin decreased xenograft volume

Xenografts were smaller in size in the 20 mg/kg treatment group (Figure 1). Mean tumor volumes ± standard errors of the mean were 16.29 ± 2.79 mm³ (controls), 9.87 ± 2.28 mm³ (10 mg/kg), and 7.67 ± 1.59 mm³ (20 mg/kg). Tumors from mice treated at 20 mg/kg were 53% smaller than controls and achieved statistical significance (p < 0.05). Those from mice treated at 10 mg/kg were smaller by 39%, but this finding did not achieve statistical significance. Absolute changes in xenograft volumes compared to controls were 6.42 ± 2.28 mm³ (10 mg/kg) and 8.63 ± 1.59 mm³ (20 mg/kg). Myometrial xenografts were not statistically different in size when compared to controls (data not shown).

Figure 1. Decreased xenograft volume after simvastatin treatment.

Xenograft tumor volumes were decreased by activated simvastatin treatment which achieved statistical significance at the 20 mg/kg body weight dose (p < 0.05). Number of analyzed samples: 5 controls, 5 aSIM10, 12 aSIM20. The box and whisker plots show the medians and means (line and dot within each box, respectively).

*, p < 0.05; ns, not significant (p ≥ 0.05)

Simvastatin treatment resulted in loose and decreased xenograft ECM formation

Leiomyomas primarily consist of excessive and aberrant ECM (10), so we hypothesized that decreased size was likely due to decreased ECM. Human leiomyoma xenografts in murine models have typically developed dense compact matrices (18). Also, simvastatin treatment of leiomyoma cells in vitro decreased ECM collagen protein expression (11). Using our xenograft model, we found that simvastatin treatment led to formation of loose ECM and decreased ECM protein expression (Figure 2, 3). In the absence of simvastatin, spindle shaped leiomyoma cells appear separated by ample compact ECM (Figure 2A–B) with evidence of angiogenesis within the xenograft (Figure 2C–D). In the simvastatin treatment groups, xenografts had leiomyoma cells with little or no intervening ECM of a loose translucent quality (Figure 3A–B). The xenografts from the mice treated at 20 mg/kg were devoid of the compact matrix and had decreased inter-nuclear distances when compared to controls (Figure 3C–D). These findings suggest that ECM was degraded in vivo in the treatment groups. To determine the mechanism of decreased ECM concentration, we next evaluated ECM components known to be disrupted in human leiomyomas.

Figure 2. Abundant compact ECM in control xenografts.

Histopathologic assessment of subcutaneous human leiomyoma xenografts in mice not exposed to simvastatin. A: 10x magnification H&E of subcutaneous leiomyoma xenograft. B: 20x magnification of human leiomyoma xenograft with dense compact matrix (CM) and leiomyoma cells. The distance between cells is attributed to matrix formation. C: 40x magnification of human leiomyoma xenograft showing spindle cells and compact matrix (CM). D: Detailed spatial localization of small blood vessels (BV), leiomyoma cells and dense, continuous “opaque” matrix in human leiomyoma xenograft at 100x magnification in oil.

BV, blood vessel; CM, compact matrix; H&E, hematoxylin and eosin; SkM, skeletal muscle

Figure 3. Loose and decreased xenograft ECM after simvastatin treatment.

Histopathologic characterization of human leiomyoma xenograft in mice treated with simvastatin at 20 mg/kg body weight. A: 10x magnification H&E of subcutaneous leiomyoma xenograft. B: At 20x magnification, the xenograft demonstrates degradation of the ECM as loose matrix (LM) and no evidence of compact matrix. Cells are in closer proximity due to dissolution of the ECM. C: 40x magnification of the xenograft loose matrix with apparent matrix degradation and decreased inter-nuclear distances. D: 100x oil magnification highlights close juxtaposition of leiomyoma cell nuclei and translucent quality of the loose matrix.

H&E, hematoxylin and eosin; LM, loose matrix; SkM, skeletal muscle

Xenograft loose matrix demonstrated decreased collagen protein expression

Collagen in leiomyoma xenografts had previously been shown to be present after 4 weeks (19). We therefore hypothesized that simvastatin would affect collagen expression. Using trichrome staining, we found that simvastatin led to decreased expression in the loose ECM. Control xenografts had compact collagen-rich ECM (Figure 4A–B) while xenografts from mice treated at 20 mg/kg had both compact and loose matrix components (Figure 4C–D). Compared to the dense matrix, leiomyoma cells were arranged in closer proximity in the loose matrix. Furthermore, the loose matrix was more translucent, discontinuous, and filamentous than the compact matrix (Figure 4E–F).

Figure 4. Decreased collagen protein in xenografts after simvastatin treatment.

Total collagen trichrome stain in human leiomyoma xenografts: control and simvastatin treated mice. Total collagen stained blue, and de novo synthesized collagen stained more intense blue (arrows). Cells, blood vessels, and dermal skin appendages (hair follicles and sweat glands) stained red. The mouse papillary dermis served as an internal positive control for collagen staining. A-B: Untreated control group xenografts showed compact matrix. C-F: Simvastatin 20 mg/kg group xenografts. Spindle shape cells were seen in both compact matrix (CM) and loose matrix (LM). The loose matrix stained intensely blue suggestive of de novo collagen formation (arrows). At 20x and 40x magnifications, leiomyoma nuclei are seen with decreased internuclear distances. At 100x magnification in oil, the discontinuous filamentous loose matrix is best appreciated. This area is well vascularized containing small blood vessels (BV).

BV, blood vessel; CM, compact matrix; LM, loose matrix; SkM, skeletal muscle

MMP2 protein activation increased with Simvastatin in vitro

To characterize the loose ECM and associated decreased collagen protein expression, we hypothesized that simvastatin induced ECM degradation by matrix metalloproteinase 2 (MMP2). MMP2 is the most abundant matrix metalloproteinase associated with uterine leiomyomata (32). It is involved in ECM remodeling, collagen degradation, signal transduction (33), and it may be altered by pathologic and therapeutic factors (34,35). As MMP2 can be found in both intra- and extra-cellular compartments (36), we considered in vitro methods only, so that we may separate these compartments.

Using zymography, we found that simvastatin treatment was associated with increased MMP2 activity in cells and in surrounding media. Both pro- and active-MMP2 were increased in cells treated with activated simvastatin (0.1–4 μM for 24 hours; Figure 5A). Furthermore, pro-MMP2 was secreted to the growth medium with increasing concentration of activated simvastatin up to 2 μM. Active MMP2 was only observed in the media at the highest concentrations of activated simvastatin (2 and 4 μM). No bands for active MMP2 were seen in the control media or in media treated at the lowest concentrations of activated simvastatin (0–1 μM; Figure 5B).

Figure 5. MMP-2 activity increased in cells and media after simvastatin treatment in vitro.

Matrix metalloproteinase 2 (MMP2) activity by zymographic analysis. A: Pro-, and active-MMP2 activity from cell lysates increases with activated simvastatin treatment in a concentration-dependent manner. B: Pro-, and active-MMP2 activity in conditioned growth media. Pro-MMP2 activity increased up to 2 μM of activated simvastatin. Active-MMP2 activity was only present at the highest activated simvastatin concentrations (2.0 and 4.0 μM). No band was seen for active MMP2 at the lower concentrations or in the control media. MMP2 activity levels are presented relative to the control pro-MMP2 in each graph. Representative conditioned growth media zymography gel images shown.

MMP2, matrix metalloproteinase 2

MT1-MMP protein expression increases with simvastatin in vitro and in vivo

To further characterize the mechanism of ECM degradation, we evaluated Membrane type 1 matrix metalloproteinase (MT1-MMP aka MMP14) concentration, as it is also known to degrade extracellular collagen and to activate MMP2 (37,38). We found that simvastatin treatment increased MT1-MMP protein expression in vivo and in vitro. In humans, simvastatin treatment was associated with a significantly increased MT1-MMP protein expression compared to placebo (H-scores 350 ± 12.7 and 138 ± 11.4, respectively; p < 0.001; Figure 6A). Representative immunohistochemical slides illustrate this difference: compared to placebo (Figure 6B) simvastatin-treated participants had more intense staining and in more cells (Figure 6C). Similar findings were seen in xenografts from untreated controls (Figure 6D) versus mice treated with activated simvastatin at 20 mg/kg (Figure 6E). We also found that simvastatin was associated with increased MT1-MMP protein expression in our cell lines in a dose dependent manner that achieved statistical significance at the highest (1 μM) concentration from 4 to 48 hours (p < 0.05 compared to vehicle controls; Figure 6F). The greatest increase was observed with activated simvastatin at 1 μM at 24 hours (1.88 ± 0.05 fold; Supplemental Table 3).

Figure 6. MT1-MMP protein expression increased after simvastatin treatment in vivo and in vitro.

Expression of membrane type-1 matrix metalloproteinase (MT1-MMP) in treated leiomyoma tissue and human leiomyoma xenografts by immunohistochemistry. A: Box and whisker plot of MT1-MMP expression by H-score determination: placebo, 138 ± 11.4, n=4; simvastatin, 3.50 ± 12.7, n=4 (p < 0.001). B: Representative section of placebo-treated leiomyoma tissue stained with anti-MT1-MMP antibody demonstrated 20–30% of cells expressing this protein. C: Representative section of simvastatin-treated leiomyoma tissue stained with anti-MT1-MMP demonstrated intense cytoplasmic staining in >50% of cells present. D: Leiomyoma xenograft from untreated controls demonstrated staining for MT1-MMP in 10–20% of cells present. E: Leiomyoma xenografts from mice treated with activated simvastatin showed >50% of leiomyoma cells staining for MT1-MMP expression. F: Western blot analysis showed increased MT1-MMP protein expression for cells treated with activated simvastatin at 1 μM (p < 0.05). Vehicle controls served as the reference and results were normalized to COX IV.

BV, blood vessel; MT1-MMP, membrane type-1 matrix metalloproteinase; *, p < 0.05; ***, p < 0.001

LRP1 protein expression increases with simvastatin in vitro

To further characterize why activation MMP2 was induced by simvastatin, we hypothesized that low density lipoprotein receptor related protein 1 (LRP1 aka CD91) protein expression would be increased as LRP1 is an endocytic receptor that had been shown to induce MMP2 expression (39). We found that activated simvastatin treatment increased LRP1 protein expression in a time- and dose-dependent manner (Figure 7). This increase was statistically significant in cells treated with activated simvastatin at 0.1 and 1 μM for 24 hours (1.48 ± 0.06 and 2.11 ± 0.07 fold up-regulation compared to vehicle controls, respectively; p < 0.05). They were also increased compared to the 6-hour simvastatin treatment (Supplemental Table 4).

Figure 7. LRP1 protein expression increases with simvastatin treatment in vitro.

Activated simvastatin treatment increased low density lipoprotein receptor related protein 1 (LRP1) protein expression in a time- and dose-dependent manner. A: Western blot analysis. Increased LRP1 protein expression was statistically significant in cells treated with activated simvastatin at 0.1 and 1 μM for 24 hours (p < 0.05). Vehicle controls served as the reference and results were normalized to COX IV. B: Representative western blots.

aSIM, activated simvastatin; COX IV, cytochrome c oxidase IV; LRP1, low density lipoprotein receptor related protein 1; *, p < 0.05 compared to vehicle controls

Apoptosis increases with simvastatin in human leiomyoma xenografts

Since simvastatin is known to activate the caspase-3 apoptotic cascade and induce calcium-dependent apoptosis of human leiomyoma cells in vitro (11,12), we hypothesized that simvastatin would be associated with increased apoptosis in our xenograft model in vivo. A terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay would have given a quantitative analysis of apoptotic activity, but we used MGP-Y staining due to the small amount of evaluable tissue. Using MGP-Y histochemical staining, we found that xenografts from simvastatin treatment groups had evidence of apoptosis compared to controls (Figure 8).

Figure 8. Apoptosis.

Evidence of apoptosis in clinical trial-derived leiomyoma tissue and in leiomyoma xenografts: methyl green pyronin-Y stain (MGP-Y). A-B: Leiomyoma tissue from placebo-treated trial participants demonstrated bright red cells with nuclear detail characterized by identification of nucleoli. C-D: Tissue from simvastatin-treated trial participants demonstrated apoptotic cells that stained deep red (arrows). Note obliterated nuclear detail and chromatin fragmentation. E-F: Leiomyoma xenografts from mice receiving activated simvastatin at 20mg/kg body weight simvastatin showed numerous apoptotic cells (arrows).

Discussion

Our findings demonstrate that human leiomyoma xenografts exposed to simvastatin experienced growth inhibition and altered ECM protein production in vivo. Specifically, we found that mice injected with activated simvastatin demonstrated reduced tumor volume of leiomyoma xenografts, ECM degradation, increased cell apoptosis, and activation of pathways contributing to ECM breakdown. Furthermore, clinical trial-derived human leiomyoma tissue showed evidence of regulatory protein expression thought to contribute to matrix disintegration in vivo.

As xenografts in our study demonstrated evidence of collagen synthesis de novo, our results suggested that simvastatin decreased xenograft volume by breaking down matrix collagen. Active MMP2 was known to degrade extracellular collagen(40), MT1-MMP was known to proteolytically convert MMP2 from pro- to active-MMP2 (37), and LRP1 was known to induce MMP2 expression (39). Furthermore, MMP2 and MT1-MMP had been implicated in uterine fibroid pathology (34,35,41,42). Therefore, induction of the MMP2/MT1-MMP/LRP1 pathway by simvastatin can explain the observed architectural changes associated with decreased fibrosis: loosely formed matrices, discontinuous collagen fibers, and decreased cellular spatial arrangement. Finding that simvastatin decreased leiomyoma xenograft volume in vivo is consistent with prior published results (14). Also consistent are the anti-fibrotic effects which had been seen after myocardial infarction (43) and the pro-apoptotic effects seen in human leiomyoma (12), colorectal cancer (44), glioblastoma, neuroblastoma, lung adenocarcinoma, and breast cancer cells (45). We observed pro-apoptotic effects in both clinical trial-derived and xenografted tissues. However, we did not see the evidence of loose matrix formation in clinical trial-derived tissues (results not shown). This suggested that simvastatin’s effect may be targeting newly formed or newly forming ECM.

The methodology of combining in vitro cell culture, in vivo xenograft, and clinical trial tissue protein expression work is a major strength of this study as are the consideration of many time points for cell culture data. Still, we presume that further time points would add to our current knowledge. Finding that serum simvastatin acid levels were comparable to those seen in humans after oral simvastatin treatment strengthens the translational relevance of our findings, however, we observed high variability among internal standard measurements which may limit generalizability. We were limited to 4 weeks of treatment in our murine model as mice demonstrated discomfort and tissue induration at the injection sites. Furthermore, at least 2 mice per simvastatin treatment group experienced localized bleeding at hormone pellet insertion sites and hepatic extramedullary hematopoiesis. Other limitations include that high variability in xenograft volume changes for the vehicle treatment group precluded using these controls for comparison which is why we compared against untreated controls. Lastly, MMP2 results were derived only from in vitro studies so further in vivo data for this pathway may help further elucidate the molecular mechanisms involved.

Conclusions

Simvastatin exhibits antitumoral, proapoptotic, and anti-ECM properties in vivo which may be explained by the activation of the MMP2/MT1-MMP/LRP1 pathway. Our study adds to the building evidence supporting the clinical efficacy of simvastatin in the setting of uterine leiomyoma.

F&S Science Clinical Quick Take

Simvastatin exhibits antitumoral and proapoptotic activity leading to extracellular matrix degradation and decreased leiomyoma tumor volume in vivo. The activation of the MMP2/MT1-MMP/LRP1 pathway may explain these findings.