Introduction
The obesity epidemic has been closely linked with an increased incidence of non alcoholic fatty liver disease (NAFLD) [
1]. NAFLD results from accumulation of fat in the liver which alters the local liver microenvironment leading to changes in lipid profiles, recruitment of inflammatory cells and changes in cytokine expression [
2,
3]. The changes in the tissue microenvironment can affect cell-cell interactions and influence the development of both primary and metastatic tumors. Epidemiological studies show that NAFLD has been linked to an increase in the risk for development of primary liver cancer [
4,
5] but very little is known about the effect of steatosis on tumor metastasis to the liver. The liver is a frequent site of metastasis for several cancers such as colorectal cancer, breast and pancreatic cancer. In addition, obesity is an independent risk factor for the development of these tumor types among others [
6]. Previous studies from our lab have shown that mice with high fat diet induced steatosis have an increase in metastatic burden compared to mice with normal livers [
7]. Metastasis represents the end-stage of cancer progression and is responsible for most cancer related deaths [
8]. Improved understanding of the metastatic tumor microenvironment is important in devising therapies to impact this stage of the disease.
Matrix metalloproteinases (MMPs) are a family of zinc dependent proteases that are capable of cleaving various components of the extracellular matrix. Apart from matrix molecules, they can also cleave adhesion proteins, activate growth factors such as TGF-β and VEGF as well as release and activate cytokines [
9,
10]. MMPs are produced by tumor cells, stromal cells and infiltrating inflammatory cells and facilitate host-tumor interactions. Members of the MMP family have been associated with progression through multiple stages of cancer, from initiation to acquisition of metastatic properties [
11,
12]. However, MMPs have also been shown to have anti-tumorigenic functions and are important for normal developmental and wound healing responses [
13]. Early clinical trials broadly targeting the MMP family as a whole, led to unintended clinical side effects. Improved understanding of specific roles for individual MMPs and development of pharmaceutical agents that selectively target individual MMPs may be part of effective therapeutic strategies in the future [
14].
MMP13 is an interstitial collagenase that is capable of cleaving multiple collagens, preferentially collagen II, as well as other matrix substrates such as gelatin, fibronectin, and aggrecan relevant to tumor metastasis [
15]. This protease has been previously linked to the progression of fibrotic liver disease [
16]. In addition, increased expression of MMP13 has been associated with poor prognosis in patients with colorectal cancer metastasis to the liver [
17]. MMP13 was identified as a part of the breast cancer metastasis signature and was associated with decreased overall survival and metastasis in breast cancer and renal cell carcinoma [
18-
20]. Furthermore, stroma-derived MMP13 was found to be involved in the growth of liver, lung, brain and heart metastases of melanoma cells [
21]. We thus hypothesize that both stromal and tumor derived MMP13 play an important role in modulating the liver microenvironment and facilitate the establishment of liver metastasis.
Discussion
The liver is a common site of metastasis for several types of cancer [
26] and identification of molecular effectors that can prevent metastasis to the liver is of important clinical relevance. NAFLD is increasingly becoming recognized as the most common cause of liver disease and it is associated with increased risk of development of primary liver cancers, even prior to establishment of cirrhosis [
1]. Progression of NAFLD results in changes in the liver microenvironment and alterations in the extracellular matrix, which effects cancer progression and outcome [
27,
28]. The MMPs are an important class of proteases that can alter the extracellular matrix and influence the microenvironmental integrity. There is considerable evidence supporting the role they play at different steps of malignant tumor metastasis including tumor cell intravasation and extravasation [
29]. We initially evaluated the steatotic microenvironment for alterations in the gene expression levels for a panel of MMPs that have been previously associated with cancer progression. Our results demonstrate that MMP13, an interstitial collagenase, is elevated in the steatotic liver in our murine model as well as in human patient samples with NAFLD. Other groups have additionally shown that MMP13 is elevated in fibrotic liver disease [
16] and could be relevant in other settings such as hepatitis as well. Several studies also link elevated levels of MMP13 in either the stroma or tumor cells with cancer progression [
17,
18,
20,
21], however there are reports that alternately suggest a protective role for MMP13 [
30,
31]. There still remains a limited understanding of the role of MMP13 in the liver microenvironment and its influence on the establishment of hepatic metastases. Here, we evaluated the role of both stromal and tumor cell derived MMP13 on the establishment of metastases in the liver.
Using mice genetically deficient in MMP13 we have shown that loss of stromal MMP13 leads to a significant decrease in the tumor burden in the liver. This effect was seen in both normal and steatotic livers, suggesting that elevation of MMP13 in the liver plays a role in tumor metastasis. Although we saw a significant decrease in tumor burden in the
Mmp13 deficient mice, differences in proliferation or apoptosis were not observed within the tumors, nor were any differences in the vascularity of the tumors. This suggested that the difference in metastatic burden was due to early events in the metastatic cascade. Similar levels of metastatic cell dissemination and early cell survival were observed in wildtype compared to
Mmp13 deficient livers, yet the loss of host MMP13 diminished the ability of tumor cells to extravasate. Several factors can influence the ability of tumor cells to adhere to the vascular walls and extravasate into the surrounding tissues. Studies have shown that changes in expression level and structure of collagen surrounding the vasculature are important in the ability of tumor cells to extravasate [
31-
33]. Since MMP13 is a collagenase, we evaluated the level of collagen I, II and IV in the liver but found that mice lacking MMP13 had no significant differences in collagen mRNA expression and immunofluorescence staining pattern compared to the wildtype mice. Further, we did not observe any differences in fibrotic disease progression as shown by trichrome staining of murine liver sections (Additional file
1: Figure S6).
Although we did not observe changes in collagen in the liver, MMP13 may mediate release or activation of critical factors involved in tumor cell extravasation. MMP13 can cleave, release and activate cytokines thereby altering recruitment and or activation of inflammatory cells such as neutrophils and macrophages that have been shown to facilitate tumor cell extravasation [
34-
37]. We evaluated changes in the lymphoid and myeloid lineage inflammatory cell sub populations within the liver of wildtype and
Mmp13−/− mice but did not see significant differences within the inflammatory cell subpopulations. However, the activation state and function of the immune cells could alternately explain the differences in tumor burden observed without being reflected in absolute cell numbers.
MMP13 expression has been observed in invasive malignant tumors such as breast carcinomas, squamous cell carcinomas (SCCs) of the head and neck and vulva, primary and metastatic melanomas, hepatocellular carcinoma cell lines [
38] and transitional cell carcinoma of the urinary bladder [
39]. Colorectal cancers can be classified as either the most commonly seen classical nonmucinous adenocarcinomas (AC) or as the less frequent subtypes of mucinous adenocarcinomas (MAC) and signet-ring cell carcinomas (SC) [
40]. Recent studies showed that the nonmucinous ACs had a higher tumor derived MMP13 protein levels than MACs and SCs, which are more invasive and have a higher frequency of lymph node metastasis than nonmucinous ACs [
41]. MACs and SCs have different characteristics from classical ACs and might have additional mutational changes that circumvent the requirement of MMP13 and affect their ability to invade and metastasize [
40]. Further, levels of related MMPs including MMP1, MMP8, MMP14, MMP2 and MMP9 which have previously been shown to affect tumor cell invasion were not evaluated in these studies and could compensate for MMP13 levels [
42]. Our results indicate that both the MC38 murine colon cancer line and the HCT116 colon cancer cells express MMP13. Knockdown of
Mmp13 utilizing lentiviral shRNA resulted in a decreased ability of tumor cells to migrate and invade
in vitro and correspondingly led to the development of fewer metastasis
in vivo. Coupled together, both tumor cell and host derived MMP13 promote the establishment of metastases in the liver.
The use of selective MMP13 inhibitors may be an important step to control tumor growth and metastasis to the liver, however in the past, the clinical use of MMP inhibitors has been hampered by their lack of specificity [
43]. Fortunately, progress is being made toward developing specific inhibitors and a few MMP13 selective inhibitors are currently being studied [
19,
44]. MMP13 is an ideal candidate with relatively low expression in normal adult tissues that will limit unwanted side effects when targeted [
15]. It will therefore be important to determine whether MMP13 selective inhibitors can pharmacologically block metastasis to the liver.
Methods
Patient samples
Liver biopsies from consented patients undergoing bariatric surgery were collected and flash frozen for protein or RNA collection and further de-identified in accordance with a protocol approved by Vanderbilt’s Internal Review Board (VUMC IRB#120829). Samples were reviewed by a Vanderbilt University pathologist and classified as either Normal (<5% Steatosis), Steatosis, Steatohepatitis or NAFLD related cirrhosis based on histology. De-identified formalin-fixed, paraffin-embedded liver tissue sections from patients with colorectal cancer metastasis to the liver were obtained from the Vanderbilt Translational Pathology Shared Resource.
Mice
8 week old C57bl/6 J male mice were obtained from Jackson Research Laboratories (Bar Harbor, ME) and a breeding pair of MMP13 null mice was obtained (Laboratory of Dr. Zena Werb, UCSF). Mice were housed in a level 6 animal facility at Vanderbilt University. Mice were fed either regular chow, a 13.5% fat diet (5001, LabDiet) or fed a 42% calories from fat diet (TD.88137, Harlan Teklad)
ad libitum for 3 months at which time point we have shown that wildtype mice develop prominent steatosis [
7].
Cell lines
MC38 murine colon cancer cells lines, syngeneic to C57Bl/6 background were provided by Dr. Steven Libutti, NCI. HCT116 human colorectal carcinoma cell line was obtained from ATCC (CCL-247). Cell lines were grown in culture conditions of 10% fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA) in Dulbecco’s Modified Eagle Media (Gibco BRL, Carlsbad, CA) at 37°C and 5% CO2, and harvested at 75% confluence for experimental studies.
Western blot
Protein lysates were prepared with RIPA lysis buffer. Protein concentration was determined by the Pierce BCA protein assay (Thermo Scientific). 30 μg of protein was loaded into each well and separated on a 10% SDS-PAGE gel. Proteins were transferred onto a nitrocellulose membrane, subsequently blocked with 3% milk, and then incubated with anti MMP13 antibody (Abcam, ab39012) overnight at 4°C. The blots were washed and then incubated with secondary antibody (IR700 conjugated donkey anti rabbit IgG) for 1 h at room temperature. Blots were washed and then imaged with Licor Odyssey scanner.
qRT-PCR
RNA was extracted using a combined Qiazol extraction and subsequent Qiagen RNeasy mini kit. One microgram of total RNA was reverse transcribed using the high capacity cDNA reverse transcription kit (Applied Biosystems) and real-time PCR was performed using specific primers for mouse MMPs with GAPDH as control (Qiagen, QT00116116, QT00107751, QT00110012, QT00113540, QT00108815, QT00115521, QT00099729, QT01658692, QT00098945, QT01064308). Real-time PCR was performed with the iQ SYBR green supermix kit (Bio-Rad) according to the manufacturer’s instructions and measured via a CFX96 real time PCR detection system (Bio-Rad). Experiments were done in triplicate with three replicates per sample. Fold-change was determined relative to normal wildtype samples and calculated using GAPDH levels as a reference.
Histology and immunohistochemistry
Liver tissue was formalin fixed, embedded in paraffin blocks, and cut into 6 μm sections. For histology, sections were rehydrated with xylenes and a decreasing ethanol series and then stained with Mayer’s Hematoxylin (Sigma) and Eosin. Hydrated sections were boiled in a citric acid solution (10 mM trisodium salt dihydrate pH 6.0, 0.5% Tween-20) for 8 minutes to unmask antigen. Slides were stained with primary antibody (MMP13: sc-12363, Ki67: ab15580, Cleaved caspase 3: cell signaling D175, vWF: Dako A0082) and incubated overnight at 4°C, washed and incubated with biotinylated secondary antibody (Vector Labs), processed with the ABC Vectastain kit (Vector Labs) and developed in chromogen solution (0.1 M Tris–HCl pH 7.4, 1.125 mM diaminobenzidine, 0.01% H2O2). Slides were counterstained with Mayer’s Hematoxylin Solution (Sigma), dehydrated with ethanols and mounted with permount. Slides were imaged with a Q Imaging Micropublisher color digital camera mounted to a Zeiss Axioplan 2 microscope using MetaMorph software for acquisition.
Experimental liver metastasis was carried out as previously described [
7]. Briefly, 5-month-old wildtype or
Mmp13 knockout mice that had been on diet for 3 months were injected with 5 × 10
5 MC38 parental or
Mmp13 knockdown cells into the spleen and allowed to perfuse to the liver for 3 minutes before splenectomy. Mice were maintained on respective diets and sacrificed 14 days post-injection. At the time of sacrifice the mice were weighed, the livers were removed and weighed, and the livers were processed for histology as described above. Graphical representation of metastatic burden was calculated using GraphPad software to compare total liver weights and percent liver weight relative to total animal weight. Further, in-depth quantitative analysis of tumor burden was assessed on the left lateral lobe of the liver, that was fixed and cut sagittally into four parts, paraffin-embedded, sectioned and stained for H&E. Slides were scanned at 20X using an Ariol® SL-50 scanner. Ariol software was used to quantify the size and number of tumors per section.
To determine the number of tumor cells extravasating in the liver, we utilized the methodology previously described by Martin et al. [
25], with some modifications. Briefly, mice were injected intrasplenically with 1×10
6 cell tracker red (Life Technologies) labeled MC38 tumor cells. At 24 and 48 hours post injection, mice were anesthetized using isoflurane and placed on mechanical ventilation at 60 breaths/minute through surgical tracheostomy. 100 μl of 0.5 mg/ml 488-tomato lectin vascular label (Vector Laboratories, DL-1174) was injected into the spleen and allowed to circulate throughout the body for 6 minutes. Next, the abdominal aorta was cut to provide outflow and livers were perfused with PBS using gravity pressure through the heart and spleen. After the effluent had cleared, the liver was resected en-bloc, placed in a glass bottom dish with #1.5 glass (In Vitro Scientific) and immediately imaged.
Images were acquired with a LSM780 confocal microscope (Carl Zeiss Inc.) using a Fluar 40X oil objective with NA = 1.30 at room temperature. 488 nm and 561 nm laser lines were used to simultaneously excite fluorescence from the vasculature cell tracker labeled tumor cells. The band-pass emission filters for the two channels were set as follows: 499 – 552 nm for the green fluorescence, and 602 – 747 nm for the red fluorescence. The pixel size was set to 0.692 μm and the dwell time was 12.6 μs. Images were acquired every 0.700 μm for every z-stack. The laser power of the laser lines was independently adjusted for each stack, between 0.1% and 3%, to avoid saturated pixels and to maintain a good signal to noise ratio through all the planes. The bit depth of the images was set to 12 bit. Image analysis was performed using Fiji [
45]. The tridimensional structure of the vasculature was reconstructed from each z-stack using the Tubeness function [
46] and overlaid onto the red channel to identify a cell as extravasating or not. Percentage of cells extravasating at each time point was determined from over 30 individual tumor cells per mouse (n = 3 per group).
Establishment of stable MMP13 knockdown cell lines
Pooled lentiviral particles targeting three different regions of murine Mmp13 (Open Biosystems: V2LMM_28573, V2LMM_ 34601, V2LMM_37490) and control particles were used to transfect MC38 cells. Control non-target and specific human MMP13 shRNA lentiviral particles targeting three different regions of human MMP13 (sc-41559) were used to transfect HCT116 cells. Post-transfection, shRNA-expressing cells were selected with puromycin. Knockdown was confirmed in resulting clones with qRT-PCR and western blot analysis.
MTT
Proliferation of cells was determined using the MTT assay. Briefly, 10×104 cells were plated in each well of a 96 well plate and allowed to attach overnight. Cells were serum starved for 24 hours followed by changing to DMEM containing 10% FBS for 24 hours. 20 μl of MTT reagent (5 mg/ml, Sigma) was added to each well and incubated at 37°C for 2 hours after which the media was aspirated and the remaining MTT formazan crystals dissolved in 100 μl of isopropanol. Absorbance at 570 nm was read using a Victor3 V 1420 Multilabel Plate Counter. Experiments were carried out in triplicate with 5 replicates per plate.
Transwell migration and invasion assays
Cell migration and invasion were assessed by a modified Boyden assay using 24 well multi-well inserts (BD) with 8 μm pore PET membrane. Mmp13 knockdown and control MC38 or HCT116 cells (1×105 cells/chamber) were seeded in DMEM plus 1% FBS directly on the insert membrane (Migration) or resuspended in 100 μl of 1 mg/ml of basement membrane matrix (BD) (Invasion) and seeded on the membrane. DMEM with 10% FBS was added to the lower chamber and cells were allowed to migrate through the filter for 16 h at 37°C in 5% CO2 for MC38 cells and 48 h for HCT116 cell lines. MC38 cells were allowed to invade for 24 h and HCT116 cells for 72 hours. Cells on the lower surface of the membrane were fixed in 100% methanol, stained with Dapi, and imaged. Experiments were carried out in triplicate with three replicates per experiment and the number of cells migrated/invaded per 10X field was determined from five random fields per well.
All animal experimental procedures and protocols were carried out in accordance to the ARRIVE guidelines. Experimental procedures and protocols were approved by the Vanderbilt university medical center IACUC protocol #M/09/216 and performed according to the institutional ethical guidelines for animal care and use. Human tissue samples were collected under the National Cancer Institute (NCI) best practices and CHTN standard operating procedures and approved by the Vanderbilt Internal Review Board (VUMC IRB #120829).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
All authors contributed to the design of the experiments. AM performed experiments, analyzed data and drafted the manuscript. AU participated in confocal microscopy and analysis for experiments in Figure
3. MVS and LG conceived the study, participated in its design and coordination and helped draft the manuscript. All authors read and approved the final manuscript.