Hepatocyte growth factor activator inhibitor type 1 (HAI-1) is a transmembrane protease inhibitor that regulates the activities of membrane-bound and extracellular serine proteases. HAI-1 has two Kunitz-type inhibitor domains with the N-terminal Kunitz domain (KD1) responsible for inhibiting known target proteases. Previously, we reported that knockdown of HAI-1 in the human pancreatic carcinoma cell line SUIT-2 resulted in epithelial to mesenchymal transition. To evaluate the role of HAI-1 in metastasis, we examined the metastatic capability of SUIT-2 cells that did or did not stably express HAI-1 short-hairpin RNA in an experimental pulmonary metastasis assay using nude mice. The extent of pulmonary metastasis was verified by histological examination and direct measurement of human cytokeratin 19 mRNA levels. One week after injecting SUIT-2 cells into mouse tail veins, apparent metastatic colonization was observed in 36% (4/11) of mice injected with HAI-1-knockdown SUIT-2, whereas none (0/11) of the control mice were positive for metastasis. After 2 weeks the metastasis positive ratios were 80% (4/5) and 40% (2/5), and after 4 weeks the ratios were 82% (9/11) and 45% (5/11) for HAI-1-knockdown and control SUIT-2 cells, respectively. Thus, loss of HAI-1 promoted pulmonary metastasis. Co-injection of recombinant KD1 abolished metastasis produced by HAI-1-knockdown SUIT-2 cells after 1 week. Moreover, recombinant KD1 restored E-cadherin levels in HAI-1 knockdown SUIT-2 cells and reduced their invasiveness in vitro. These data indicate that HAI-1 regulates pulmonary metastasis of SUIT-2, and KD1 may have therapeutic application for inhibiting metastatic cancer cell spreading. (Cancer Sci 2011; 102: 407–413)
Hepatocyte growth factor activator inhibitor type 1 (HAI-1), encoded by the serine protease inhibitor Kunitz type 1 (SPINT1) gene, is a membrane-associated Kunitz-type serine protease inhibitor expressed on the basolateral surface of most epithelial cells and placental cytotrophoblasts.(1–3) HAI-1/SPINT1 is also expressed in various types of epithelial tumors, although the biological role of HAI-1/SPINT1 in tumor progression remains largely unknown. HAI-1/SPINT1 is composed of an extracellular region containing an N-terminal cysteine-rich region, the first Kunitz domain (KD1), a low-density lipoprotein receptor-like domain and the second Kunitz domain (KD2), followed by a transmembrane region and a short cytoplasmic region.(2) HAI-1/SPINT1 functions on the cell surface as a serine protease inhibitor.(4) The extracellular portion can be secreted by protease-mediated ectodomain shedding, and the secreted HAI-1/SPINT1 (sHAI-1) containing KD1 also exhibits potent protease inhibitor activity.(5–7) However, sHAI-1 carrying both KD1 and KD2 shows reduced affinity for target proteases, possibly due to an inter-Kunitz interaction that occurs in the secreted form.(4,6,7) Therefore, KD1 is the major functional domain of HAI-1/SPINT1. Previous studies have shown that HAI-1/SPINT1 potently inhibits the action of a variety of membrane-bound and secreted serine protease. The membrane-bound serine proteases matriptase/ST14, hepsin/TMPRSS1 and prostasin/PRSS8 are well known targets(2,8) and TMPRSS4 may also be a target.(9) The secreted proteases hepatocyte growth factor activator (HGFA), kallikrein 1-like peptidase 4 (KLK-4), KLK-5 and trypsin are strongly inhibited by HAI-1/SPINT1.(10,11) All of these proteases targeted by HAI-1/SPINT1 are reported to be involved in carcinogenesis and tumor progression.(8,10)
Previously, we reported that knockdown of HAI-1/SPINT1 resulted in epithelial to mesenchymal transition (EMT) in the human pancreatic carcinoma cell line SUIT-2 and the human adenocarcinoma cell line HLC-1.(9) This phenotypic alteration is likely mediated by membrane-bound serine proteases such as matriptase/ST14 and TMPRSS4 through smad-interacting protein 1 (SIP1, also known as ZEB2) upregulation.(9) The HAI-1/SPINT1 KD-induced EMT was accompanied by increased invasiveness in vitro, but the effect of HAI-1/SPINT1 knockdown in vivo was more complex. After subcutaneous transplantation in nude mice of either HAI-1/SPINT1-knockdown or control SUIT-2 cells, the tumor growth rate was lower in the knockdown cells and the incidence of pulmonary metastasis was comparable between knockdown and control cells. However, the number of metastatic colonies per mouse was higher in HAI-1/SPINT1- knockdown SUIT-2 treated mice as long as metastasis was allowed to be established.(9) Therefore, the role of the HAI-1/SPINT in metastatic spreading in vivo remains unclear. Another issue to address is whether HAI-1/SPINT1 has clinical applications. As KD1 is primarily responsible for inhibiting most proteases targeted by HAI-1/SPINT1, a recombinant soluble version of KD1 may have therapeutic implications for inhibiting cancer cell invasion and metastasis.
In the present study we examined the role of HAI-1/SPINT1 in establishing metastatic colonization in the lung using the human pancreatic cancer cell line, SUIT-2, with or without stable expression of HAI-1/SPINT1 short-hairpin RNA (shRNA). Quantification of metastatic SUIT-2 cells was performed with one-step nucleic acid amplification (OSNA) of human cytokeratin 19 (CK19) mRNA. We revealed that knockdown of HAI-1/SPINT1 promotes pulmonary metastasis in an experimental pulmonary metastasis assay. We also tested the effect of recombinant KD1 on the invasive and metastatic capabilities of SUIT-2 cells.
Materials and Methods
Cell lines. The human pancreatic adenocarcinoma cell lines SUIT-2(12) and SUIT-4 were kindly provided by Dr Takeshi Iwamura (Junwakai Memorial Hospital, Miyazaki, Japan). A high CK19-expressing human cell line, LC-2/ad, was established in our laboratory.(13) HeLa cell line, pancreatic cancer cell lines (MIA Paca2, PANC-1) and bile duct carcinoma cell lines (TGBC1TKB, TGBC2TKB) were obtained from Riken Cell Bank (Tsukuba, Japan). Cells were cultured in Dulbecco’s modified Eagle’s minimum essential medium (DMEM) containing 10% fetal bovine serum (FBS).
Antibodies and recombinant protein. The following primary antibodies were used in the present study: anti-human HAI-1/SPINT1 goat polyclonal antibody (R&D Systems, Minneapolis, MN, USA), anti-E-cadherin mouse monoclonal antibody (Takara Bio, Shiga, Japan) and anti-β-actin mouse monoclonal antibody (AC-74; Sigma, St Louis, MO, USA). The cDNA for recombinant sHAI-1 proteins consisting of the N-terminal portion and KD1 (sHAI-1/KD1) of mature HAI-1/SPINT1 was obtained as previously reported.(6) The cDNA was subcloned into an expression vector pcDNA3.1-mycHis-A (Invitrogen, Carlsbad, CA, USA), and the plasmid was stably transfected into CHO cells. The proteins were purified from the conditioned medium of the transfected cells by column chromatography using nickel and anti-myc antibody columns, dialyzed against phosphate-buffered saline (PBS) and used in the experiments.
Knockdown of HAI-1/SPINT1. The construction of the HAI-1/SPINT1 knockdown vector using a short hairpin RNA (shRNA) expression retroviral vector pSINsihU6 (Takara Bio) and infection of the vectors were described previously.(14) The target sequence of the SPINT1 gene was 5′-GGGCAGGCATAGACTTGAAGG-3′ and the scrambled sequence was 5′-GGGACGGAGATTTCACGAGGA-3′. Infected cells were subcultured at an appropriate density in fresh DMEM containing 0.5 mg/mL G418 (Nacalai Tesque, Kyoto, Japan) for isolation of G418-resistant cells.
Reverse transcription-PCR. Total RNA were extracted with the Trizol reagent (Invitrogen), followed by DNase I (Roche Diagnostic, Tokyo, Japan) treatment and phenolchloroform-isoamylalcohol extraction. For reverse-transcription PCR (RT-PCR), 3 μg total RNA was reverse transcribed with a mixture of oligo(dT) and random primer and processed for each PCR reaction as described previously.(15) The primer sequences for SIP1/ZEB2, Snail, Slug, Twist, matrix metalloprotease 9 (MMP-9) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were described previously.(9)
In vitro Matrigel invasion assay. To evaluate the invasive capacity of SUIT-2 in vitro, Matrigel invasion was assessed using ThinCert tissue culture inserts (pore size 8 μm; Greiner Bio-one, Tokyo, Japan) coated with 25 μg of Matrigel (Invitrogen) per filter. Cells (1 × 105 in DMEM, 0.1% bovine serum albumin) were placed in the upper compartment, and the lower compartment contained 10 μg/mL of fibronectin (BD Bioscience, Bedford, MA, USA) or 5% FBS as a chemoattractant. After incubation (24 h), cells on the upper surface of the filter were wiped off with a cotton swab. The cells on the lower surface were stained with hematoxylin and counted in 10 randomly selected fields (200-fold magnification).
Quantification of CK19 mRNA by OSNA. Quantification of human CK19 mRNA was performed using the OSNA method.(16) Briefly, cells or tissue slices were homogenized in lysis buffer for 90 s (Lynorhag; Sysmex, Kobe, Japan) and centrifuged for 1 min at 10 000 × g. Human CK19 mRNA was then amplified by reverse-transcription loop-mediated amplification in an RD-100i apparatus (Sysmex). Automated amplification with a ready-to-use reagent kit (Lynoamp; Sysmex) was performed directly from the sample homogenate according to the manufacturer’s instructions. To optimize detection in mouse lung tissues, various dilutions (×1 to ×100) of the homogenate sample were tested. A standard positive control sample containing 5 × 103 copies/μL of CK19 mRNA and a negative control sample without CK19 mRNA were used for calibration in every assay. The results of the assay were expressed as the number of CK19 mRNA copies per microlitre.
Experimental pulmonary metastasis assay. All of the animal work was carried out using protocols approved by the University of Miyazaki Animal Research Committee, in accordance with international guidelines for biomedical research involving animals. Cells (5 × 105/100 μL medium) that either were or were not premixed with sHAI-1/KD1 (200 μg/mL) were injected into nude mice tail veins (BALB/cAJc1-nu). One, 2 or 4 weeks after injection, the mice were killed, the lungs excised and each right or left lung divided into two blocks. The distal half of the right lung and the proximal half of the left lung were used for OSNA analysis and the remaining blocks were fixed in 4% formaldehyde in PBS for 24 h before processing for histological evaluation with hematoxylin–eosin (HE) stain.
Immunoblot, immunocytochemistry and immunohistochemistry. Cultured cells (60–70% confluency) were incubated without or with sHAI-1/KD1 (10–100 μg/mL) in growth medium for 24 h. After washing in PBS, cellular proteins were extracted using CelLytic (Sigma). For immunoblots, extracted proteins (4.5 μg) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis under reducing conditions, transferred onto Immobilon membranes (Millipore, Bedford, MA, USA), and processed as described previously.(9) For immunocytochemistry, cells were seeded into a Lab-Tek chamber slide (Nalge Nunc International, Tokyo, Japan) and incubated with or without sHAI-1/KD1 (100 μg/mL) for 24 h. The cells were then fixed with 4% paraformaldehyde in PBS for 15 min and processed for washing and blocking as described,(9) followed by incubation with anti-E-cadherin antibody for 1 h at room temperature. After washing with PBS, the cells were incubated for 1 h at room temperature with Alexa Fluor 488-conjugated goat anti-mouse IgG (Invitrogen) in PBS, washed with PBS, counterstained with 4′, 6-diamidino-2-phenylindole (DAPI; Sigma), and visualized with Axio Imager.A2 (Carl Zeiss MicroImaging, Tokyo, Japan). Immunohistochemistry for E-cadherin was performed as described previously.(9)
OSNA of human CK19 mRNA to evaluate SUIT-2 metastasis in mouse lungs. Stable expression of HAI-1/SPINT1 shRNA in SUIT-2 cells (SUIT-2-KD) resulted in knockdown of HAI-1/SPINT1 expression, which was accompanied by reduced E-cadherin expression compared with control SUIT-2 cells expressing a scrambled RNA sequence (SUIT-2-sc; Fig. 1A). Thus, knockdown of HAI-1/SPINT1 induced an EMT phenotype in SUIT-2 cells as reported previously.(9) We next analyzed the effect of HAI-1/SPINT1 knockdown on metastatic pulmonary colonization of SUIT-2 cells with an experimental metastasis assay using nude mice. In order to perform a quantitative analysis of metastasis, we used OSNA of human CK19 mRNA. OSNA detection of CK19 mRNA is now widely used in clinical practice to detect sentinel lymph node metastasis in breast cancer patients.(17) SUIT-2 cells express CK19 mRNA abundantly, and a linear correlation was observed between the CK19 mRNA levels and cell numbers (Fig. 1B). Indeed, carcinoma cell lines derived from the pancreas and the biliary tract consistently expressed CK19 mRNA (Fig. 1C). The copy numbers of CK19 mRNA per each parent SUIT-2, SUIT-2-sc or SUIT-2-KD cell line were measured, as it is possible that EMT induced by HAI-1/SPINT1 knockdown could result in reduced cytokeratin expression in SUIT-2-KD. As shown in Figure 1D, the level of CK19 mRNA from SUIT-2-sc was comparable to parent SUIT-2 cells, showing a mean copy number per single cell of 10.7 × 102 and 9.3 × 102, respectively. On the other hand, SUIT-2-KD expressed 6.1 × 102 copies of CK19 mRNA/cell, indicating a 40% reduction in mRNA levels compared with SUIT-2-sc (P <0.05, paired Wilcoxon test). However, the SUIT-2-KD CK19 mRNA level was still sufficient for measurement by OSNA. Next, we checked the effect of tissue extracts from mouse lung on the measurement of SUIT-2 CK19 mRNA by OSNA. Unexpectedly, the mouse lung tissue contained substance(s) that interfered with the OSNA reaction, but this interference could be prevented by diluting the extracted sample 10-fold, which permitted quantitative analysis of SUIT-2 CK19 mRNA in mouse lung lysates (Fig. 1E).
Enhanced pulmonary metastasis following knockdown of HAI-1/SPINT1. To test the effect of HAI-1/SPINT1 knockdown on metastasis, SUIT-2-KD cells and SUIT-2-sc cells were injected into the tail veins of nude mice and their pulmonary metastasis capability was compared. The mice were killed 2 or 4 weeks after the injection and lungs were excised for evaluation with OSNA and histology as described in the Materials and Methods section. A comparison of histological analysis and CK19 mRNA levels determined by OSNA showed agreement between the results obtained from these two assays (Fig. 2A). Two weeks after injection, the incidence of metastasis positive animals (i.e. positive in both OSNA and histological analyses; Fig. 2A, closed circles) was 40% (2/5) and 80% (4/5) in mice injected with SUIT-2-sc and SUIT-2-KD cells, respectively. After 4 weeks, the ratios were 45% (5/11) and 82% (9/11) for SUIT-2-sc and SUIT-2-KD cells, respectively (Fig. 2A). Therefore, knockdown of HAI-1/SPINT1 expression enhanced the metastatic colonization of SUIT-2 cells in pulmonary tissues. Histological analysis indicated a less differentiated morphology and reduced E-cadherin expression in HAI-1/SPINT1 knockdown cells (Fig. 2B). The estimated numbers of metastatic tumor cells, which were calculated using mean CK19 mRNA copy numbers per cell (Fig. 1A), are summarized in Table 1. Although the incidence of metastasis was higher in SUIT-2-KD treated mice compared with SUIT-2-sc, the cellular growth of SUIT-2-KD after metastasis was comparable, and even somewhat slower than SUIT-2-sc cells, based on the ratio of estimated metastatic cell numbers in the fourth week to those in the second week.
Table 1. Estimated metastatic cell numbers in lungs
2 weeks after injection (n =5) (mean of metastasis-positive cases)
4 weeks after injection (n =11) (mean of metastasis-positive cases)
Each value ×103 represents the estimated metastatic cell numbers in the lungs of each mouse. Positive cases with both one-step nucleic acid amplification (OSNA) and histological analyses (i.e. metastasis positive) are indicated in bold. The mean value of metastasis-positive case is in parentheses. 0, no metastasis; <1, faintly positive by OSNA, but negative with histology.
Reversion of E-cadherin expression in HAI-1/SPINT1 knockdown SUIT-2 cells by recombinant KD1. HAI-1/SPINT1 has two extracellular Kunitz domains. KD1 is responsible for the inhibition of its major target proteases, and in sHAI-1, KD2 interferes with the affinity of KD1 for proteases.(4,6,7) Therefore, we prepared recombinant sHAI-1 protein consisting of only the N-terminal portion and KD1 (sHAI-1/KD1) (Fig. 3A) and examined the effect of sHAI-1/KD1 on HAI-1/SPINT1 knockdown-induced EMT in SUIT-2. The activity of sHAI-1/KD1 was confirmed by its inhibitory activity against recombinant human HGFA (data not shown). The addition of sHAI-1/KD1 to cultured SUIT-2-KD cells suppressed the HAI-1/SPINT1- knockdown-induced upregulation of SIP1/ZEB2, a repressor of E-cadherin transcription (Fig. 3B) in a dose-dependent manner, which in turn induced recovery of E-cadherin expression (Fig. 3C). The expression of other E-cadherin repressors such as Snail, Slug and Twist were not significantly altered (Fig. 3B). HAI-1/SPINT1 knockdown also induced MMP-9 expression as reported previously,(9) and the upregulated MMP-9 mRNA level was also suppressed by the addition of sHAI-1/KD1 (Fig. 3B). SUIT-2-KD cells showed enhanced Matrigel invasion in vitro and the invasion was significantly suppressed by sHAI-1/KD1 (Fig. 3D). Cellular viability was not affected by the addition of sHAI-1/KD1 (data not shown).
KD1 suppresses metastatic colonization of HAI-1/SPINT1 knockdown SUIT-2 cells. Lastly, we examined the effect of sHAI-1/KD1 on the metastatic pulmonary colonization of SUIT-2 cells in vivo. As pharmacokinetic analysis of KD1 injected in mice revealed that this protein has a short half-life time in circulation,(18) we premixed the cells with sHAI-1/KD1 (200 μg/mL) before the tail vein injection. One week after the injection, the extent of pulmonary colonization was analyzed by OSNA and histopathological evaluation. As shown in Figure 4A, even 1 week after injection, a distinct CK19 mRNA signal was detectable in 36% (4/11) of mice injected with the SUIT-2-KD cells, whereas none of the control SUIT-2-injected mice showed positive reactions by OSNA (0/11). Histology revealed that several small nests of SUIT-2-KD cells formed around the alveolar walls and in the blood vessel intima (Fig. 4B). In addition, one OSNA-negative case having positive histology (a small metastatic lesion in the vascular intima) was present in the SUIT-2-KD group. In a few OSNA-negative or only faintly positive cases, tiny clusters consisting of a few tumor cells were observed after extensive histological analysis (Fig. 4A, gray circles, and Fig. 4C). Notably, pretreatment of SUIT-2-KD with sHAI-1/KD1 significantly prevented the pulmonary implantation of SUIT-2-KD cells (Fig. 4A,D).
Taken together, these observations indicate that HAI-1/SPINT1 has suppressive effects on the initial pulmonary colonization of SUIT-2 cells that is mediated by the N-terminal region that includes the KD1 domain.
To obtain insights into the role of HAI-1/SPINT1 in cancer progression, we investigated the effect of HAI-1/SPINT1 knockdown on experimental pulmonary metastasis of the human pancreatic carcinoma cell line SUIT-2. Our results showed that loss of HAI-1/SPINT1 enhanced metastatic pulmonary colonization of SUIT-2 cells, which could be prevented by pre-treating the cells with recombinant sHAI-1 containing KD1 domain. Therefore, cellular HAI-1/SPINT1 has a suppressive role in establishing pulmonary metastasis. Protease-protease inhibitor interactions in the pericellular microenvironment are an important determinant in cancer cell biology because they regulate extracellular matrix degradation and the activity of various bioactive molecules on and around the cell membrane. HAI-1/SPINT1 is a unique serine protease inhibitor that has a transmembrane domain near its C-terminus and is expressed on the basolateral surface of polarized epithelial cells and also of carcinoma cells.(1,2) As such, HAI-1/SPINT1 is an important regulator of serine protease activities in the pericellular microenvironment, and consequently may significantly modulate cancer cell biology.
Several lines of evidence suggest that HAI-1/SPINT1 has a significant role in tumor progression, particularly in the process of cellular invasion. Clinicopathological studies indicated that reduced expression of HAI-1/SPINT1 is likely involved in the progression of prostate, breast, gastric and gynecological cancers.(19–23) Knockdown of HAI-1/SPINT1 induces EMT and enhanced migration in pancreatic and lung adenocarcinoma cell lines in vitro.(9) Enhanced invasion after loss of HAI-1/SPINT1 expression was also observed in breast and prostatic carcinoma cells in vitro.(24,25) Engineered overexpression of HAI-1/SPINT1 reduced the invasiveness of glioblastoma and endometrial cancer cells in vitro,(23,26) and induced mesenchymal to epithelial transition of endometrial cancer cells.(26) This anti-invasive effect of HAI-1/SPINT1 is likely to be mediated by its interaction with target protease(s).(9,22–24) Using nude mice in an experimental pulmonary assay, the present study indicates for the first time that HAI-1/SPINT1 has a suppressive role in the hematogenous metastatic colonization of carcinoma cells in vivo.
Of interest was the observation that recombinant sHAI-1 consisting of the N-terminal region and KD1 domain (sHAI-1/KD1) significantly suppressed pulmonary metastasis of HAI-1/SPINT1 knockdown cells. Consistently, an in vitro study indicated that E-cadherin expression could be restored and invasiveness suppressed by addition of sHAI-1/KD1 to HAI-1/SPINT1 knockdown SUIT-2 cells. Thus, it is reasonable to postulate that the anti-metastatic effect of HAI-1/SPINT1 is mediated by interaction(s) between the KD1 domain and its target protease(s). Indeed, overexpression of HAI-1/SPINT1 with mutation at the P1 position of the reactive site of KD1 failed to suppress the invasion of glioblastoma cells in vitro, while wild-type HAI-1/SPINT1 suppressed the invasion significantly.(26) Moreover, the significant antitumor effect of PEGylated KD1 on prostatic cancer cells overexpressing the membrane serine protease hepsin/TMPRSS1 was recently reported.(18) While SUIT-2 cells do not express detectable levels of hepsin,(9) our previous study suggested that other transmembrane serine proteases, such as matriptase and TMPRSS4, may be targets of the HAI-1/SPINT 1 and could be involved in the EMT phenotype of SUIT-2.(9) Alternatively, deregulated pericellular activity of HGFA may be involved in the enhanced metastasis of HAI-1/SPINT1-knockdown SUIT-2 cells, as HGFA is an efficient activator of the proform of hepatocyte growth factor (HGF),(10) and HGF and its receptor c-Met are reported to be involved in metastatic spreading of SUIT-2 cells.(27)
In summary, the result of the present study suggests that HAI-1/SPINT1 suppresses initial metastatic colonization of carcinoma cells, probably via its KD1 domain. Further studies regarding the therapeutic applications of HAI-1/SPINT1, particularly of KD1, for metastasis prevention will be required.
This work was supported in part by Grant-in-Aid for Scientific Research no. 20390114 (H. Kataoka) and for Young Scientists no. 22790384 (T. Fukushima) from the Ministry of Education, Science, Sports and Culture, Japan. The authors thank Ms Y. Tobayashi for her help with the histological studies.