Hepatocyte growth factor activation inhibitors (HAI-1 and HAI-2) regulate HGF-induced invasion of human breast cancer cells
Article first published online: 23 MAR 2006
Copyright © 2006 Wiley-Liss, Inc.
International Journal of Cancer
Volume 119, Issue 5, pages 1176–1183, 1 September 2006
How to Cite
Parr, C. and Jiang, W. G. (2006), Hepatocyte growth factor activation inhibitors (HAI-1 and HAI-2) regulate HGF-induced invasion of human breast cancer cells. Int. J. Cancer, 119: 1176–1183. doi: 10.1002/ijc.21881
- Issue published online: 5 JUN 2006
- Article first published online: 23 MAR 2006
- Manuscript Accepted: 2 JAN 2006
- Manuscript Received: 19 JUL 2005
- Breast Cancer Campaign
- Cancer Research Wales
Hepatocyte growth factor (HGF) plays a plethora of roles in cancer metastasis and tumour growth. The interaction between tumour cells and their surrounding stromal environment is a crucial factor regulating tumour invasion and metastasis. Stromal fibroblasts are the main source of HGF in the body, and release HGF as an inactive precursor (pro-HGF). HGF activator (HGFA), matriptase, urokinase-type plasminogen activator and hepsin are the main factors responsible for converting pro-HGF into active HGF. HAI-1 and HAI-2 are 2 novel Kunitz-type serine protease inhibitors that regulate HGF activity through inhibition of HGFA, matriptase and hepsin action. Recent studies demonstrate that HAI-1 and HAI-2 may also potently inhibit a number of other pro-metastatic serine proteases and therefore have direct bearing on the spread of tumours. Our study examined the potential of these HAI's to suppress the influence of HGF and regulate cancer metastasis. We generated a retroviral expression system that induced HAI expression in a human fibroblast cell line. Forced expression of either HAI-1 or HAI-2 in these fibroblasts resulted in a dramatic decrease in the production of bioactive hepatocyte growth factor (HGF). This reduction in HGF activity subsequently suppressed HGF's metastatic influence on breast cancer cells. To further assess the anti-cancer properties of HAI-1 and HAI-2 we generated recombinant HAI proteins. These recombinant HAI proteins possessed the ability to potently quench HGF activity. We also demonstrate that these recombinant HAI's suppressed fibroblast-mediated breast cancer invasion. An additional ribozyme transgenes study revealed that elimination of HAI-1 and HAI-2 expression, in an MDA-MB-231 breast cancer cell line, significantly enhanced the migratory, proliferative and invasive nature of these breast cancer cells. Overall, our data demonstrates the important roles of HAI-1 and HAI-2 in cancer metastasis, and reveals that these serine protease inhibitors display strong therapeutic potential. © 2006 Wiley-Liss, Inc.
Cancer metastasis is the single most important factor influencing patient mortality. Controlling the metastatic spread of tumours remains a crucial goal for the successful treatment of cancer. HGF plays a well-recognised role in promoting tumour invasion and metastasis, as this factor demonstrates the ability to stimulate proliferation, dissociation, migration and invasion in a wide variety of tumour cells and is also a potent angiogenic factor.1, 2, 3 Therefore, factors such as HGF may play a pivotal role in developing new methods to combat the spread of tumours.
In the body, the main sources of HGF are the stromal tissues; however, some tumour cells themselves have been shown to produce HGF.3, 4 Reports demonstrate that patients with breast cancer have elevated serum HGF levels, and that following removal of the malignant breast tumours, the serum levels decrease.5, 6, 7, 8 This suggests that the primary source of HGF within these patients may have been the tumour cells themselves, or stromal cells within the tumour tissue. HGF levels have also been shown to correlate with disease progression, with the levels rising in cases of recurrence.9, 10 Importantly, 1 study revealed that the immunoreactive level of HGF was a stronger independent predictor of recurrence and survival than that of lymph node involvement.11 Jiang et al.3 summarise the significance of HGF in cancer.
Interactions between tumour cells and their surrounding stromal environment play a key role in modulating the aggressive nature of tumour invasion and metastasis.12, 13, 14 HGF is synthesised and released by stromal fibroblasts as an inactive precursor, pro-HGF, and requires site-specific cleavage to function as a biologically active cytokine.15, 16 A number of proteases have been suggested as possessing HGF-converting properties, but the initial factor was a serine protease known as HGF activator (HGFA).17 The liver is the main source of HGFA in the body, although we have demonstrated that a variety of cancer cell lines and tissues also produce HGFA.18, 19 This activation of HGF is the key step that governs the influence of HGF in cancer metastasis. Therefore, suppression of pro-HGF conversion may limit the potent effects of HGF. Studies have described 2 serine protease inhibitors with the ability to bind to HGFA, and block its HGF-activating properties. These inhibitors were termed HGFA inhibitor type-1 and type-2 (HAI-1 and HAI-2).20, 21 HAI-1 and HAI-2 are regulators of HGF action, and may therefore limit the pro-metastatic effects of HGF on tumour cells. Crucially, Kunitz domain-1 of HAI-1 also demonstrates the ability to potently inhibit the action of variety of other serine proteases involved in cancer metastasis including matriptase, hepsin, plasma kallikrein and trypsin.22, 23, 24 Additional factors that possess pro-HGF converting ability include pro-metastatic factors such as matriptase, hepsin and urokinase-type plasminogen activator.25, 26, 27
The significance of HGF in human cancer metastasis is well established.3 However, the roles of HAI-1 and HAI-2 in the body are still unclear; these inhibitors may play multiple roles, and have been linked to a variety of physiological processes. Very little is known about the regulation of HGF activity and the interaction between the HGF activators (HGFA and matriptase) and the HGF activator inhibitors (HAI-1 and HAI-2). Our previous data reveal that the degree of HAI-2 expression is inversely correlated with tumour spread in human breast cancer.19 We suggest that these inhibitors play an important role in cancer progression. Presently, the true function of these Kunitz-type serine protease inhibitors requires clarification. Our study used a variety of methods to examine the potential of HAI-1 and HAI-2 to act as anti-cancer agents through their ability to block HGF activation, and also assessed their function and role in the metastatic cascade.
Previously, we have demonstrated that the human MRC5 fibroblast cell line produces high levels of the HGF and HGFA proteins. However, these fibroblasts display little or no HAI presence with which to limit HGF activity.18 The current study employed a retroviral system to force the expression of HAI-1 and HAI-2 within these human fibroblasts. The ability of these modified fibroblasts to influence breast cancer cell invasion and migration was assessed through a range of in vitro analyses. We also synthesised biologically-active recombinant HAI-1 and HAI-2 proteins, and assessed the potential therapeutic value of these novel serine protease inhibitors. In addition, we employed ribozyme transgenes to eliminate expression of HAI-1 and HAI-2 in a human breast cancer cell line to further demonstrate the significance of these factors in cancer progression.
Material and methods
Cell lines and culture
LNCAP human prostate cancer cells, HT-115 human epithelial colon carcinoma cells, MDA-MB-231 human breast cancer cells, MDCK canine kidney cells and MRC-5 human fibroblast cells were obtained from the European Collection for Animal Cell Culture (ECACC, Porton Down, Salisbury, UK) apart from the human prostate cancer cell lines that were purchased from the American Type Culture Collection (ATCC, Rockville, MD). pRev-Tet-on was obtained from Clontech (Mountain View, CA).
Construction of retroviral expression systems
Full-length human HAI-1 and HAI-2 cDNAs were obtained from LNCAP and HT-115 human cancer cell lines, respectively, using Herculase enhanced polymerase blend (Stratagene, Amsterdam, Netherlands) which has a 3′–5′ proof reading enzyme (see Table-I for all primer details). Purified PCR products with a Hind III and Cla 1 restriction site on either side of the product were digested prior to ligation into the pRevTRE vector, which was similarly digested using Hind III and Cla 1 (New England BioLabs, Hertfordshire, UK) using T4 DNA ligase protocol (Promega Corporation, Madison, WI). The resulting ligation products were precipitated, purified and used to transform chemically competent bacteria (JM109, Promega). After selecting positive strains with expression sequence correctly inserted, E.coli was amplified, followed by extraction and purification of plasmid using a plasmid extraction kit (Qiagen, Crawley, West Sussex, UK).
|HAI-1 (inc. SnaB1 site)||Forward||GATTACGTATGCCTCGCATCCAAC|
|HAI-1(inc. EcoRV site)||Reverse||GATGATATCTACGAGGGGCCGGTGGTGT|
|HAI-2 (inc. SnaB1 site)||Forward||GATTACGTAATGGCCGCAGCTGTGCGGGC|
|HAI-2 (inc. EcoRV site)||Reverse||GATGATATCTCACAGGACATATGTGTTC|
The PT67 packaging cell line was transfected with pRevTet-On (selection marker neomycin) or pRevTRE + HAI-1 (or HAI-2) (hygromycin resistance), using electroporation (Flowgen, Lichfield, UK). Following selection and identification of positive strains, stable transfectant was used to generate active viral stock which contained infectious pRevTet-On, pRevTRE-HAI-1 or pRev-HAI-2 particles.
Transduction of MRC-5 cells with active retroviral stocks
MRC-5 (40–60% confluency for successful transduction) was co-transduced with pRevTet-On and pRevTRE + HAI-1, pRevTet-On and pRevTRE+HAI-2 or pRevTet-On and pRevTRE for a successive 48 hr. Polybrene (4 μg/ml) was also added to the culture to aid infection during this period. Following an additional 24 hr in plain medium, the cells were treated with selection medium that contained Neomycin (100 μg/ml) and Hygromycin (100 μg/ml), in order to establish a pRev-Tet-On and expression positive stain. Following successful identification of the dually positive cells, the cells were routinely cultured in maintenance medium (with Neomycin 25 μg/ml and Hygromycin 25 μg/ml).
MRC5 cells were then treated with or without doxycycline (10 μg/ ml final) overnight. The presence of HAI transcript was verified using reverse transcription-polymerase chain reaction (RT-PCR) and presence of cellular HAI proteins using Western blotting.
Recombinant HAI-1 and HAI-2 production with a yeast expression system
We wanted to examine the anti-cancer properties of HAI-1 and HAI-2 in vitro. To do this we employed a multi-copy Pichia yeast expression system (Invitrogen, Paisley, UK), to generate large quantities of recombinant HAI-1 and HAI-2 protein. (Note: see the earlier section for detailed descriptions of the techniques employed by this eukaryotic expression system).
The HAI-1 and HAI-2 sequences were amplified from the retroviral vectors (containing the appropriate HAI insert), using a Hi-fidelity amplification kit (ABgene, Surrey, UK), and primers incorporating appropriate SnaB1 and EcoRV restriction sites. The products were subsequently cloned into the pPIC9K plasmid vector. Upon ligation of HAI-1 and HAI-2 into the response vector, these constructs were then transformed into chemically competent bacteria. The pPIC9K + HAI containing or control plasmids were linearised with a Sal 1, before being electroporated (450 volts) into the KM71 yeast strain (3 μg). KM71 was plated onto histidine-deficient agar plates. The colonies that contained high yield of HAI-1 and HAI-2 were grown in BMMY medium supplied with the kit. HAI protein expression was induced following the addition of methanol (5%), and protein concentration calculated using a Bio-Rad DC protein assay kit (Bio-Rad laboratories, Hercules, CA). HAI-1 and HAI-2 presence in the supernatant was confirmed through Western Blotting procedures, using HAI-1 and HAI-2 antibodies developed by our lab.19
Reverse transcription-polymerase chain reaction
cDNA was prepared from 0.25 μg of total RNA with a reverse transcription kit (Sigma, Poole, Dorset, UK). The quality of DNA was verified using β-actin. For all primers used in our study see Table I. PCR was performed in a GeneAmp PCR system 2400 thermocycler (Perkin Elmer, Norwalk, CT). PCR products were then loaded onto a 0.8% agarose gel and electrophoretically separated. The gel was then visualised under ultraviolet light following ethidium bromide staining.
Bioassay analysis of bioactive HGF
This method was essentially the same as that established by Stoker et al.28 and Jiang et al.29 Briefly, MDCK cells were seeded onto 2 96-well plates, at a density of 5,000 cells/well, and incubated for 6 hr. Serially diluted samples of transduced and wild type MRC-5 fibroblast supernatant were then added to the cells, along with recombinant human HGF standards. Following 24 hr incubation in normal cell culture conditions, the cells were fixed (4% formalin), and stained with crystal violet to facilitate assessment of MDCK cell colony scattering. The quantity of active HGF produced by the fibroblast cells was determined by examining which was the highest dilution of supernatant that was capable of inducing cell scattering.
Co-culture tumour cell invasion assay, to examine HGF bioactivity
To a 24-well plate was seeded 10,000 of each different set of transduced MRC-5 cells, along with the wild type MRC-5 cells to act as a control.18 Polycarbonate filter insert (pore size 8 μm)(Becton Dickinson, Labware, Oxford, UK), pre-coated with 50 μg/insert of Matrigel (Collaborative Research Products, Bedford, MA), were then inserted into the wells and added 10,000 MDA-MB-231 cancer cells. This co-culture system was then incubated for 72 hr before the cells were fixed and stained.
Knockout of HAI-1 and HAI-2 expression using ribozyme transgenes
Our previously reported ribozyme system,30, 31 was employed to knockout the expression of HAI-1 and HAI-2 in the MDA-MB-231 breast cancer cell line. Briefly, the secondary structure of human HAI-1 and HAI-2 was generated using Zuker's RNA mFold software. The ribozyme that specifically targets HAI-1 or HAI-2 was generated using touchdown PCR with the primers described in Table I. The resulting correct inserts were purified and cloned into the pEF6/V5-His-TOPO vector, and then electroporated into the MDA-MB-231 breast cancer cell line. MDA-MB-231 cells with HAI-1 and HAI-2 eliminated were termed HAI-1 KO and HAI-2 KO, respectively. Following selection and knockout confirmatory processes described previously,31 the effect of HAI loss from these breast cancer cells was examined through a series of in vitro studies described in this methods section.
The modified MDA-MB-231 breast cancer cells were seeded in a 96-well plate at a density of 7,000 cells/well, and incubated at 37°C for 72 hr. MTT was added in solution to the cells (200 μg/well) and incubated for 4 hr at 37°C, then the cells were lysed with Triton (10%). The intensity of the colour released was determined by a plate reader (Titertek Multiskan, Eflab, Finland), to give a value in absorbance units, which was representative to the number of cells.
The migratory properties of MDA-MB-231 breast cancer cells were assessed to determine the impact of the elimination of HAI-1 and HAI-2 from these cells. This technique to measure cell motility has been described in a previous study.32 The different cell types were seeded at a density of 50,000/well into a 24-well plate and allowed to reach confluence. The layer of cells was then scraped with a fine gauge needle to create a wound of ∼200 μm. The movement of cells to close the wound was recorded and analysed as described previously using a time-lapsed video system.32 After the addition of a treatment, the cells' motile qualities were monitored and recorded on video for 100 min. Wound closure/cell migration was evaluated with motion analysis software (Optimus 6) and results were exported to a spreadsheet (Excel) for further evaluation and interpretation.
We assessed the results obtained using (2-sided) students t-test. Values obtained in the study are given as mean value ± SD. A p < 0.05 was defined as statistically significant.
Retroviral system expression of HAI-1 and HAI-2
HAI-1 and HAI-2 were successfully amplified from the prostate (LNCAP) and colorectal (HT115) cancer cell lines. These products were then successfully cloned into the retroviral vectors, and sequencing (through gene bank database alignment) confirmed the identity of the inserted DNA sequences as the HGFA inhibitors. Following transduction, the modified MRC5 cell lines where termed MRC5CONTROL for MRC5 cells containing the empty vector, MRC5HAI-1+ for the fibroblasts expressing HAI-1 and MRC5HAI-2+ for the HAI-2 expressing fibroblasts. The presence of the HGFA inhibitors in the human fibroblast cell lines was confirmed through RT-PCR (Fig. 1). The specific primers for HAI-1 and HAI-2 were used to confirm the presence of the inhibitor mRNA. Strong bands were produced, that represented either HAI-1 or HAI-2 in the appropriate transduced cells. Our results also confirmed our previous report that little or no HAI mRNA was expressed by wild type MRC-5 fibroblasts.
These results confirmed that the HAI containing retroviral plasmids had been successfully transfected into the PT67 cell line, effectively packaged to produce infectious non-replicatable viruses, and used to infect MRC-5 fibroblasts.
Yeast system production of recombinant HAI-1 and HAI-2 proteins
In this Pichia expression system, HAI-1 and HAI-2 where initially amplified from the retroviral plasmid DNA (Fig. 2 – upper panel), as previous sequencing analysis of this plasmid DNA revealed mutation-free HAI nucleotide sequences. The Hi-fidelity RT-PCR method ensured that the amplified products would be error free and suitable for cloning into a yeast vector following minor modifications (as discussed in the methods). Our amplified version of HAI-1 contained all the main components of the HAI-1 protein, such as the 2 Kunitz domains, and the LDL-like receptor domain, although the majority of the N-terminal domain was absent, resulting in a 795 bp sequence. However, the amplified version of HAI-2 was an exact copy of the complete HAI-2 sequence and its size was 759 bp.
The production and secretion of HAI-1 and HAI-2 proteins by the yeast cells (strain KM71) was confirmed through filtration and purification of the yeast culture medium, and subsequent Western Blotting (Fig. 2 – lower panel).
Bioassay assessment of HGF activity from HAI-expressing fibroblasts
MRC-5 cell supernatant from cultured wild type and transduced fibroblasts was removed following a 4-day incubation period. These samples were examined quantitatively for HGF activity. The amount of bioactive HGF in these samples was defined as units/ml, in which, the units were defined as the highest dilution of supernatant that induced scattering. This value was converted to ng/ml by multiplying it against 0.625 ng/ml, as this was the quantity of HGF standards that induced the same degree of cell scattering.
It was revealed that the wild type and MRC-5CONTROL fibroblasts had in the range of 80 ng/ml of active HGF in the supernatant. The bioactivity of the transduced fibroblast supernatants was far lower than this level. The MRC5HAI-1+ fibroblasts produced 20 ng/ml of active HGF, whereas MRC5HAI-2+ fibroblasts produced ∼10 ng/ml (Table II). The inhibition of HGF processing was also confirmed by Western blot analysis (Fig. 3). Our study has led to the generation of a fibroblast cell line that expresses the HGFA inhibitors. These data have shown that the HAI transduced MRC-5 cells do not produce as much bioactive HGF as the wild type and MRC-5CONTROL fibroblasts. Thus, the presence of the HAI-1 and HAI-2 in these fibroblasts suppresses natural ability of HGFA to activate pro-HGF.
|MRC-5 fibroblasts||Highest dilution of scattering||Bioactive HGF released (ng/ml)|
Breast cancer cell invasion assay (co-cultured with HAI modified fibroblasts)
We used this assay to examine the interactions between the breast cancer cells (MDA-MB-231) and the transduced and wild type stromal fibroblasts (MRC5). The control group revealed a mean invasion of 23.1 ± 3.9 breast cancer cells (MDA-MB-231). The cultured wild-type and MRC-5CONTROL fibroblasts produced a large amount of active HGF in the system, which resulted in a significant increase in the level of breast cancer cell invasion (52.9 ± 7.4 and 60.5 ± 3.3, respectively) compared to the control group (p < 0.0001). However, the presence MRC5HAI-1+ and MRC5HAI-2+ fibroblasts in the system resulted in a dramatic reduction in the number of breast cancer cells invaded (p < 0.0001), compared to the wild type and MRC-5CONTROL fibroblasts. The MRC5HAI-1+ and MRC5HAI-2+ fibroblast treatments showed a mean invasion of 17.8 ± 4.0 and 18.2 ± 5.2 breast cancer cells, respectively. This assay has demonstrated that the pro-HGF normally produced by these MRC5 fibroblasts is not effectively converted to active HGF by these MRC5HAI-1+ and MRC5HAI-2+ fibroblasts (Fig. 4). This suggests that the level of active HGF in the system had been markedly reduced, because of the presence of HAI-1 and HAI-2. Therefore, presence of these HGFA inhibitors significantly suppressed the influence of the wild type MRC5 fibroblasts on breast cancer cell invasion.
Recombinant HAI-1 and HAI-2 suppressed activation of pro-HGF
We employed a bioassay cell scattering technique to assess the amount of active HGF generated by MRC5 fibroblasts in combination with our recombinant HGFA inhibitors (Table III). The supernatant from MRC5 fibroblasts was found to contain 40 ng/ml of active HGF. However, when the cells were cultured in the presence of either HAI-1 or HAI-2 the level of active HGF in the supernatant was found to be ∼20 ng/ml. Interestingly, when the fibroblasts were cultured in a combination of both HAI-1 and HAI-2, the level of active HGF was found to be at a further reduced concentration of 10 ng/ml. These results show that, individually, our HGFA inhibitors were able to reduce the level of active HGF produced by the fibroblasts by 50%. However, HAI-1 and HAI-2 working in tandem suppressed generation of HGF by 75%.
|MRC-5 fibroblast treatment||Highest dilution of scattering||Bioactive HGF in the supernatant (ng/ml)|
Recombinant HAI-1 and HAI-2 inhibited breast cancer cell invasion
This in vitro assay assessed the ability of our recombinant HAI proteins to inhibit MRC5 fibroblast-induced MDA-MB-231breast cancer cell invasion (Fig. 5). The control group demonstrated spontaneous tumour cell invasion (30.7 ± 3.8). However, the breast cancer cells that were co-cultured with MRC5 fibroblasts showed a significant increase in the number of cells invaded (67.1 ± 7.3) (p < 0.0001). Importantly, when either HAI-1 or HAI-2 (final concentration 10 μg/ml) was added to this system in combination with the MRC5 cells, the level of invasion was significantly reduced (41.9 ± 3.6 and 36.6 ± 4.4 respectively) (p < 0.0001). The most dramatic reduction in breast cancer cell invasion was observed following the addition of both HAI-1 and HAI-2 (10 μg/ml) to the fibroblasts (32.7 ± 4.0 cells), this groups degree of invasion was comparable to that of the control group with no stimuli (p < 0.0001).
Once again, these results demonstrate that the presence of HAI-1 and HAI-2 is able to quench the invasive influence of the stromal fibroblasts.
The elimination of HAI-1 and HAI-2 expression significantly increased breast cancer cell proliferation, migration and invasion
MDA-MB-231 breast cancer cells express both HAI-1 and HAI-2.18 We used a ribozyme system to effectively inhibit HAI-1 and HAI-2 expression in these cells. RT-PCR was performed to assess the effectiveness of the ribozyme transgenes ability to suppress HAI expression. HAI-1 and HAI-2 expression was successfully suppressed in the breast cancer cells (Fig. 6a). MDA-MB-231 cells with HAI-1 and HAI-2 eliminated were termed HAI-1 KO and HAI-2 KO, respectively. Results from the proliferation assay demonstrated that the loss of HAI-1 or HAI-2 from these MDA-MB-231 breast cancer cell lines resulted in significantly (p < 0.001) accelerated growth in these cells (Fig. 6b). Therefore, HAI-1 and HAI-2 may play a role in limiting tumour cell proliferation.
Upon suppression of HAI-1 and HAI-2 expression there was also a marked change in the aggressive nature of these breast cancer cells (Fig. 6c and 6d). HAI-1 KO and HAI-2 KO breast cancer cells were significantly more invasive than the control breast cancer cells with HAI's present (p < 0.0001). These results demonstrate that HAI-1 and HAI-2 may naturally limit breast cancer cell invasion.
Elimination of either HAI-1 or HAI-2 also enhanced the motile nature of these breast cancer cells (Fig. 6e). The breast cancer cells demonstrated their motile properties in the wound closure motility assay; however, the absence of either HAI-1 or HAI-2 significantly increased cell migration to close the wound compared to the control group (p < 0.05 and p < 0.001, respectively). This suggests that HAI-1 and HAI-2 may limit the motile properties of tumour cells in human breast cancer.
The last decade has witnessed the rapid increase of knowledge available on the role of HGF and c-Met in human cancer. HGF stimulates, through c-Met coupling, the metastatic spread and angiogenesis of tumours. Therefore, the blockade of HGF signalling has become a strategy to inhibit tumour invasion and metastasis. An increasing number of reports support this theory, as shown by recent NK4 studies. NK4 is an antagonist of HGF-Met coupling, and has demonstrated significant potential as a new anti-cancer agent.33, 34, 35, 36, 37, 38, 39, 40 NK4 is a variant of HGF that competitively blocks HGF binding to the c-Met receptor, thereby reducing HGF-related pro-metastatic effects. Our study approaches the suppression of HGF from a different angle, in that our HAI inhibitors prevent pro-HGF being converted into the active form in the first instance. HAI-1 and HAI-2 are 2 Kunitz-type serine protease inhibitors.20, 21 These inhibitors inhibit the influence of a range of proteases, most notably HGFA and matriptase. Pro-HGF is ineffective as a pro-metastatic factor prior to its interaction and subsequent activation via the HGFA or matriptase proteases.41, 42, 43, 44 The conversion of pro-HGF to the biologically active HGF is the critical limiting step in the HGF regulatory system and will play a key role in the control of metastatic events. Our study examined the properties of HAI-1 and HAI-2 to assess whether these inhibitors represent possible suppressors of cancer metastasis and angiogenesis.
Previous studies have demonstrated that growth, migration and invasion of cancer cells are markedly accelerated through interaction with a broad spectrum of fibroblasts both in vitro and in vivo.12, 13, 14 We have demonstrated that human MRC-5 fibroblasts produce pro-HGF, which is readily converted to active HGF.18 However, these fibroblasts produce little or no HAI's with which to modulate HGF activity, resulting in the uninhibited production of biologically active HGF.18 Therefore, local and mutual interactions between cancer cells and these HGF-secreting stromal fibroblasts are likely to be particularly important in regulating the malignant behaviour of tumour cells.
Our study has led to the generation of a retroviral expression system capable of inducing the expression of HAI-1 and HAI-2, individually, within an MRC-5 fibroblast cell line. Wild type MRC-5 fibroblasts synthesise pro-HGF, and also express HGFA to readily convert this immature factor into the bioactive form of HGF. We report that both wild-type and empty vector control MRC5 fibroblasts produced approximately the same level of bioactive HGF (80 ng/ml) into the culture system. Importantly, the HAI-expressing fibroblasts generated significantly lower levels of active HGF (10–20 ng/ml). In these MRC5HAI-1+ and MRC5HAI-2+ fibroblasts, the actual synthesis of pro-HGF will not be suppressed, instead, the pro-HGF processing required to create active HGF has been significantly reduced. This reduction in pro-HGF processing may be due to the ability of HAI-1 and HAI-2 to inhibit the HGFA, matriptase and hepsin pro-HGF converting action.27 The normal tumour–stromal cell interactions that facilitate breast cancer cell invasion, as demonstrated by the wild-type/empty vector control MRC5 fibroblasts and MDA-MB-231 breast cancer cells in a co-culture invasion assay, were significantly reduced by the presence of HAI-producing fibroblasts in the system.
The importance of suppressing pro-HGF processing was recently highlighted in an interesting study that described, through a single amino acid substitution in the proteolytic site, how an uncleavable form of pro-HGF suppressed tumour growth and dissemination in a mouse model.45 Therefore, the HAI-induced suppression of HGF activity should lead to a reduction in the pro-metastatic influence of HGF. To further assess HAI function, we employed a yeast expression system to generate recombinant HAI proteins. Crucially, the value of the HAI's as potential anti-cancer agents was confirmed through these recombinant HAI-1 and HAI-2 studies. Addition of either rHAI-1 or rHAI-2 to cultured fibroblasts significantly reduced the production of biologically active HGF. This reduction in HGF levels will lessen the impulse for tumours to metastasise. We also demonstrated the potential of the recombinant HAI proteins through a co-culture invasion assay. Addition of rHAI-1 and rHAI-2 into this system markedly reduced breast cancer cell invasion, this may be due to the ability of these HAI's to interact and inhibit the pro-invasive function of matriptase, hepsin and HGFA. Therefore, the natural ability of these fibroblasts to facilitate cancer cell invasion has been suppressed by the addition of our recombinant HAI proteins.
Suppression of HGF activity is due to a shift in the balance between the HGF-converting proteases and the HAI inhibitors. This shift between HGF activation and HGF activation suppression is the crucial step controlling the metastatic influence of HGF, and may represent a method of limiting tumour progression. Our study reveals that the presence of HAI-1 or HAI-2 altered the balance to favour suppression of HGF activity. Therefore, the ability of HGFA, matriptase and hepsin to enhance the potent effects of HGF on tumour cells has been quenched through the presence of HAI-1 and HAI-2. Importantly, the most effective suppression of breast cancer cell invasion was observed when rHAI-1 and rHAI-2 were used in combination. HAI-2 has recently demonstrated the ability to down-regulate matriptase gene expression, which also led to a decrease in the invasive properties of a human ovarian cancer cell line.46 Therefore, these inhibitors appear to possess additional and individual inhibitory properties,23, 46, 47, 48, 49 and may benefit from being deployed in tandem.
Further confirmation of the importance of HAI-1 and HAI-2 in tumour cell invasion was revealed by our ribozyme transgenes system, which was employed to knock down the expression of HAI-1 and HAI-2 in a human breast cancer cell line. We report that the loss of either HAI-1 or HAI-2 resulted in a significantly more aggressive cancer cell phenotype. Breast cancer cells (MDA-MB-231) with HAI-1 or HAI-2 expression knocked down became markedly more invasive, and also revealed enhanced motile and proliferative properties. Interestingly, we report that these cells do not produce pro-HGF,18 so we cannot be enhancing HGF activity through the loss of HAI's. However, this cell line does express matriptase and may express other HAI-regulated proteolytic enzymes, such as hepsin. Dysregulated matriptase and hepsin activity has been shown to promote cancer progression and metastasis.50, 51 These findings are consistent with our previous findings that demonstrated reduced levels of HAI-1 and HAI-2 were associated with poor prognosis from a cohort of 120 breast cancer patient samples.19 The 2 Kunitz domains of both HAI-1 and HAI-2 are the key to their inhibitory aspect against a variety of serine proteases. As yet, the full list of protease targets for these inhibitors is unclear, however, understanding the interaction of these HAI's with a variety of other serine proteases such as HGFA, matriptase, hepsin, plasma kallikrein and trypsin will help elucidate the roles and value of HAI-1 and HAI-2 in cancer metastasis.
In conclusion, our findings demonstrate that HAI-1 and HAI-2 play important roles in controlling the aggressive nature and spread of cancer. We also report that HAI-1 and HAI-2 are serine protease inhibitors that display unique therapeutic potential. Further progress will undoubtedly lead to the application of these advances in the generation of future therapies to prevent the spread of cancer. Presently, we are investigating the significance of these HAI proteins as anti-cancer agents through in vivo analysis, to confirm the anti-cancer effects of these inhibitors in a living model.