Nuclear Cathepsin F Regulates Activation Markers in Rat Hepatic Stellate Cells

Hepatic Stellate Cell Isolation and HSC Cell Lines

Primary HSCs were isolated from Wistar rats (∼400 g) using a recently published pronase/collagenase perfusion protocol (Weiskirchen and Gressner, 2005). The protocol was approved by the Institutional Animal Care and Use Committee (IACUC). The cells were resuspended and seeded into 75-cm² culture flasks using high glucose DMEM containing 10% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin. Purity was assessed using vitamin A autofluorescence. The rat HSC cell line HSC-T6 (Vogel et al., 2000) was a gift from Dr. Scott Friedman (Mount Sinai School of Medicine, New York). The cell line CFSC-8B (Greenwel et al., 1993) was kindly provided by Dr. Marcus Rojkind from the George Washington University Medical Center.

Primary cells and all cell lines used were cultured at 37°C in a 5% CO₂ humidified incubator. All cell culture reagents were purchased from Invitrogen (Carlsbad, CA).

Establishment and Characterization of the HSC-2 Cell Line

HSC-2 is a spontaneous immortalized clonal cell line derived from rat primary HSCs in our laboratory. The primary cells were isolated from a male Wistar rat (in-house animal colony) as mentioned above and passaged several times before clonal selection by limiting dilution. HSC-2 cells were passaged twice weekly for almost a year with no apparent changes in their phenotypic characteristics determined by immunofluorescence and Western blot. The karyotyping service was performed by National University Hospital (Singapore).

First-Strand cDNA Synthesis, RACE, Plasmid Construction, and Transfection

Total RNA from primary HSCs was used for first-strand cDNA synthesis, followed by Rapid Amplification of cDNA Ends (BD SMART RACE cDNA Amplification kit, Takara Bio, Tokyo, Japan) as described elsewhere (Lim et al., 2008). The following rat cathepsin F–specific RACE primers were used: 5′-RACE: 5′-AGATCGGTCCTTTCTGGGCCAGCCAGG-3′ and 3′-RACE: 5′-ACCAGGGCATGTGTGGCTCCTGCTGG-3′. The locations of these primers are nt 1170–1196 and nt 858–883, respectively, according to accession number EU253481. The overlapping 5′ and 3′ sequences determined by sequencing were assembled and aligned using the software Vector NTI Suite 9 (Invitrogen). The rat cDNA was cloned by PCR using primers sense (HindIII) 5′-AAGCTTTCGACCTCGCTATGGCGCTCC-3′ and antisense (XhoI) 5′-CTCGAGTCAATGGTGATGGTGATGATGGTTCACCACTGCCGAACTGG-3′containing a His(6)-tag. The resulting PCR product was cloned into the HindIII/XhoI site of the pcDNA4/TO vector. The plasmid was stable transfected into HSC-2 cells using Lipofectamine 2000 (Invitrogen).

Antibodies for Immunofluorescence Studies and Western Blot

Fibronectin (F-7387), laminin (L-9393), vinculin (V-9131) and SMAA-Cy3 (C-6198) were from Sigma (St. Louis, MO). Procollagen type I (sc-25974), cathepsin F (sc-13987), retinoic acid receptor (RAR) α (sc-551), RARβ (sc-552), RARγ (sc-7387), and protein disulfide isomerase (PDI; sc-20132) were from Santa Cruz Biotechnology (Santa Cruz, CA). SC35 (ab11826) and cystatin B (ab53725) were from AbCam (Cambridge, United Kingdom). SMAA (M0851) and glial fibrillary acidic protein (GFAP; Z-0334) were from DakoCytomation (Glostrup, Denmark). Cathepsin F (NB100–1784) for immunofluorescence (IF) on cells, in vivo staining, and Western blot was from Novus Biologicals (Littleton, CO). The secondary antibodies were anti-goat FITC (Pierce, Rockford, IL), anti-mouse FITC (Sigma), and anti-rabbit Alexa555 (Invitrogen). All peroxidase conjugated antibodies were from Santa Cruz Biotechnology.

Immunofluorescence Imaging

The IF staining of cell lines and liver sections was performed as described previously (Maubach et al., 2007) using two different anti-cathepsin F antibodies. Rat liver perfused with 4% paraformaldehyde (PFA) was embedded in cryomedia and cut into 15–20 μm cryosections using the cryostat CM3050 S (Leica, Heidelberg, Germany). Images were taken with the LEICA RMB-DM epifluorescence microscope (see Figure 2A and Supplementary Figures S1 and S2) or LEICA TCS SP2 equipped with the DM6000 (LEICA; see Figures 2B and 3). We used the Zoom In tool of the LEICA TCS SP2 to make more detailed scans for certain areas of interest (nuclei).

SDS-PAGE and Western blot

Cells were harvested by scraping in PBS. The cytosolic and nuclear extracts were prepared using either the NE-PER Nuclear and Cytoplasmic Extraction kit (Pierce) or the Nuclear Extraction kit (Active Motif, Carlsbad, CA). Total protein was prepared by resuspending the cells in 63 mM Tris/HCl containing 1% SDS (pH 6.8) supplemented with protease inhibitors (Pierce) and 10-min heating at 95°C. Protein concentration was estimated using the BCA protein assay kit (Pierce). Forty micrograms of sample was separated in a 4–12% Bis-Tris gel (Invitrogen) and transferred onto nitrocellulose membrane (Hybond-ECL, GE Healthcare, Chalfont St. Giles, Buckinghamshire, United Kingdom). The membrane was developed with ECL Plus (RPN2132, GE Healthcare). All antibodies were incubated in blocking solution (5% nonfat milk in TBS) for 1 h at room temperature. Fifteen and ten micrograms of protein was analyzed in Western blot for SMAA and cystatin B, respectively, following the same procedure. Semiquantitative analysis was performed using the ImageJ software (W. Rasband, NIH; http://rsb.info.nih.gov/ij/).

Enzyme Activity Assay

The cathepsin enzyme assay was performed similarly to earlier described protocols (Wang et al., 1998; Shi et al., 2000). Briefly, the assay was performed in 100 mM potassium phosphate buffer, pH 6.5, containing 2.5 mM DTT and 2.5 mM EDTA. The substrates used were Z-Leu-Arg-MCA (Z-LR-MCA) and Z-Arg-Arg-MCA (Z-RR-MCA; Peptides International, Louisville, KY) at a final concentration of 50 μM, where Z corresponds to benzyloxycarbonyl and MCA is 4-methylcoumaryl-7-amide. Five micrograms of cytosolic or nuclear protein extract diluted in assay buffer was incubated for 1 h at 37°C. The liberated fluorescence was measured using the Tecan Safire II (Tecan, Zurich, Switzerland) fluorescence plate reader at λ_ex 380 nm and λ_em 460 nm. A standard curve for MCA was used to calculate the amount of released fluorophore in nanomole per microgram protein per hour at 37°C.

The inhibition of enzymatic activity was studied using AEBSF, 1 mM; acetyl-pepstatin, 200 nM; cathepsin L inhibitor IV and V, 100 nM (Calbiochem, La Jolla, CA); CA-074 at 1, 10, 100 nM, and 1 μM; epoxomicin, 10 μM; Ac-YVAD-CMK, 10 μM; E-64, 10 μM (Peptides International); and LHVS, 5 and 50 nM (kindly provided by Prof. H. A. Chapman, Cardiovascular Research Institute, University of California, San Francisco). The inhibition is expressed as percent activity remaining compared with the uninhibited reaction.

Treatment of Culture-activated HSCs and HSC-2

Primary HSCs cultivated for 8 d (in vitro activated) and HSC-2s were seeded into 75-cm² culture flasks and grown overnight at conditions described above. At a confluence of 60–70%, cells were treated with a combination of 100 μM palmitic acid and 5 μM retinol for short-term (3 h) and long-term (48 h) analysis. As control, the cells were treated with DMSO only. During the 48-h treatment, fresh medium containing retinol/palmitic acid was replaced after 24 h. The same procedure applied to the treatment with 10 μM E-64d (Calbiochem) for 3 h. For small interfering RNA (siRNA) experiments, HSC-2 were transfected with two cathepsin F– or cystatin B–specific siRNAs at a final concentration of 10 nM and 120 μl HiPerfect reagent according to manufacturer's protocol (Qiagen, Hilden, Germany). Mock control refers to transfection with HiPerfect reagent only. The two siRNAs used were Rn _LOC361704_2 and Rn_LOC361704_3 for cathepsin F and Rn_Cstb_1 and Rn_Cstb_4 for cystatin B. Cystatin B siRNA was transfected 24 h before treatment with retinol/palmitic acid.

Quantitative RT-PCR

Total RNA isolation and quantitative RT-PCR using the ABI 7500 Fast Real Time PCR System (Applied Biosystems, Foster City, CA) was performed as described previously (Maubach et al., 2007). One microgram of total RNA was reverse-transcribed in a total RT reaction volume of 100 μl using reagents and conditions described in the Taqman kit (N808-0234). We used the Taqman probes for cathepsin F (Rn01450705_g1), cathepsin L (Rn00565793_m1), cathepsin B (Rn00575030_m1), cathepsin K (Rn00580723_m1), cystatin B (Rn00754988_g1), cystatin C (Rn01415507_g1), SMAA (Mm01204962_gH), and collagen type I (Rn01463849_g1) in a total reaction volume of 10 μl. All samples were normalized using 18S rRNA or β-actin.

Statistics

All quantitative results were presented as mean ± SEM. Experimental data were analyzed using two-tailed Student's t-test assuming equal variances. p < 0.05 was considered significant.

Characterization of HSC-2 Cell Line

The karyotype analysis after 40 passages revealed a mean number of 43 chromosomes (Supplementary Figure S1A). On average, one copy of chromosomes 3, 5, 6, 11, 12, 15-17, and 22 was lost. Chromosome 7 showed an additional copy. The cells exhibited a myofibroblast-like phenotype (Supplementary Figure S1B) and expressed fibronectin, procollagen type I, laminin, and vinculin (Supplementary Figure S1, C–F). Western blot analysis revealed the expression of RARα, β, and γ (Supplementary Figure S1G). A comparison of SMAA expression by immunofluorescence staining among different HSC cell lines showed prominent fibrils in HSC-2 (Supplementary Figure S2, A–C). The Western blot confirmed a high expression of SMAA in HSC-2 compared with CFSC-8B and HSC-T6 (Supplementary Figure S2D).

Cloning and Immunodetection of Rat Cathepsin F

The cDNA for the rat cathepsin F was cloned from primary HSCs using the rapid amplification of cDNA Ends (RACE) method. The cDNA sequence data have been submitted to the GenBank database under accession number EU253481 and the deduced amino acid sequence is shown in Figure 1. The rat full-length sequence shares 94 and 73% similarity at the amino acid level with the mouse and human sequence, respectively. It has five potential N-glycosylation sites and the ERFNAQ motif is present.

Surprisingly, we detected cathepsin F in the nuclei of primary HSCs cultivated for 3 d by immunofluorescence microscopy (Figure 2A). In addition, the expression of cathepsin F in vivo was biased toward HSCs positively stained for GFAP around the sinusoidal area of the rat liver (Figure 2B, arrowheads in merged image), confirming the expression of cathepsin F by HSCs in the liver. The nuclear localization of cathepsin F was confirmed in the HSC-2 cell line by co-staining with the nuclear speckle marker SC35 and Z-sections using two different anti-cathepsin F antibodies (Figure 3, A and B). In contrast, although cathepsin B was also detected in the nucleus, it did not co-localize with SC35 (Figure 3C). In the Western blot for HSC-2, two bands were detected in the nuclear and the cytosolic fraction corresponding to the mature and proform of cathepsin F (Figure 4A, inset). The intermediate band observed could belong to a truncated proform of the protein. The purity of the nuclear fraction was assessed by the absence of PDI, an endoplasmic reticulum marker protein.

Enzyme Activity Assay

To substantiate our immunofluorescence and Western blot findings, we assayed the cathepsin activity (including cathepsin F) using a fluorescence-labeled peptide substrate (Z-LR-MCA), which was used in previous studies for cathepsin F (Wang et al., 1998; Shi et al., 2000). A clearly discernible enzymatic activity was observed in the nuclear fraction of different HSC cell lines including HSC-2 (Figure 4A). The EDTA in the assay buffer prevents activity from metal-dependent proteases like matrix metalloproteinases (MMPs) and calpains. Strong inhibition by E-64 confirmed a cysteine protease activity, even though a robust inhibition by the caspase inhibitor Ac-YVAD-CMK was also observed (Figure 4B).

To determine the extent of Z-LR-MCA cleavage by cathepsin B and L, we used two approaches. The contribution of cathepsin L activity was tested using two commercially available inhibitors. Except for the cathepsin L inhibitor IV in the nuclear fraction, we found no significant decrease in activity (Figure 4C). To assay for the contribution of cathepsin B activity, we utilized the cathepsin B–specific inhibitor CA-074 and a cathepsin B–specific substrate Z-RR-MCA, in addition to Z-LR-MCA. The differences in the CA-074 concentration-dependent inhibition for both substrates used showed that cathepsin F contributes to the cytosolic and more importantly to the nuclear activity (Figure 4D). In addition, the inhibitor N-morpholinurea-leucine-homophenyl alanine-vinylsulfone-phenyl (LHVS) decreased the nuclear activity by 30 and 70% at 5 and 50 nM, respectively (Supplementary Figure S3A).

Treatment of HSC-2 with Retinol/Palmitic Acid

Earlier observations in our lab have indicated that the cytosolic and nuclear activity changes during the process of HSC culture activation (in vitro). The cytosolic activity increased during transdifferentiation until day 10 after isolation and decreased afterward. In contrast, the nuclear activity decreased steadily from day 0 onward (data not shown).

Therefore, we decided to use conditions previously described that result in a deactivation of HSCs (Abergel et al., 2006) to study the influence of the nuclear activity on HSC activation markers SMAA and/or collagen type I. HSC-2 was treated with a combination of retinol/palmitic acid for 3 and 48 h. Short-term (3 h) treatment with retinol/palmitic acid resulted in an insignificant increase of SMAA and collagen type I mRNA (Figure 5A). In contrast, the results after 48 h revealed a significant down-regulation of SMAA and collagen type I mRNA (Figure 5B), which was confirmed on the protein level using Western blot (Figure 5, C and D), although the change for collagen is less pronounced. The transcripts of all cathepsins investigated were up-regulated at both time points, except for cathepsin B which was up-regulated after 48 h only (Figures 5, A and B). Consistent with this finding, the activity increased in the nuclear and cytosolic fraction after 3 h retinol/palmitic acid treatment (Figure 6A, 3 h). Surprisingly, there was a significant decrease in the cytosolic and an almost 50% reduction in the nuclear activity after 48 h (Figure 6A, 48 h). This behavior was confirmed using culture-activated primary HSCs (Supplementary Figure S3B). In search for a possible explanation to account for the down-regulation of activity after 48 h, we investigated the mRNA levels of two endogenous inhibitors, cystatin B and C. A strong up-regulation of cystatin B and C transcripts was observed after 48 h (Figure 6B). Furthermore, immunofluorescence staining of HSC-2 showed a nuclear localization of cystatin B, which also partially co-localized with the nuclear marker SC35 (Supplementary Figure S4). This observation agrees with earlier findings (Riccio et al., 2001). We performed siRNA experiments in order to test the hypothesis that cystatin B is responsible for the inhibition of the nuclear activity. Our data show that indeed the knockdown of cystatin B (Figure 6D) before retinol/palmitic acid treatment led to an increased nuclear activity, compared with the retinol/palmitic acid treatment, almost to the level of untreated cells (Figure 6C). A significant increase in SMAA transcript was observed, though not for cathepsin F mRNA (Figure 6D). No consistent results on collagen type I regulation were obtained (data not shown).

Effect of E-64d Inhibition, Cathepsin F siRNA Transfection and Overexpression of Cathepsin F

To further evaluate the importance of the nuclear activity in affecting the regulation of SMAA and/or collagen type I, we treated the cells with E-64d, a cell-permeable cysteine protease inhibitor. E-64d treatment led to a down-regulation of cathepsin F and K mRNAs, whereas cathepsin L and B mRNA levels remained unchanged. In addition, the mRNA for SMAA and collagen type I significantly decreased and increased respectively (Figure 7A). The enzyme activity was drastically inhibited in the cytosolic and nuclear fraction (Figure 7B).

E-64d is not specific for cathepsin F as it also abrogates the nuclear activity of cathepsin B. In consequence, we treated HSC-2 for 48 h with different cathepsin F siRNAs to decrease the mRNA and the nuclear activity of cathepsin F. Figure 7, C and D, illustrate that a knockdown of cathepsin F mRNA and protein indeed led to a significant down-regulation of SMAA and collagen type I mRNA, as well as a decrease in nuclear activity. On the protein level, we did not observe any significant changes for both SMAA and collagen type I (data not shown). The overexpression of rat cathepsin F in HSC-2 resulted in a significant increase in the mRNA levels of cathepsin F, collagen type I, and cystatin B; however, a decrease for SMAA was observed (Figure 8A). Interestingly, the cytosolic activity was higher, but the nuclear activity was lower in the rat cathepsin F overexpressing-HSC-2 compared with the HSC-2 (Figure 8B). The effect on the protein level was more prominent for SMAA, but a robust up-regulation of collagen type I was also noted (Figure 8, C and D).

Nuclear Localization of Cathepsin F

The cathepsins belong to a class of lysosomal proteases that are involved in many different processes in the cell (for review see Dickinson, 2002; Mohamed and Sloane, 2006; Brix et al., 2007). In our present investigation, we attempt to identify a possible role of cathepsin F during the transdifferentiation of hepatic stellate cells. In the course of this work, we established a new rat HSC cell line named HSC-2. This cell line exhibits features characteristic of myofibroblast-like HSCs (Supplementary Figures S1 and S2). Furthermore, HSC-2 also expresses the retinoic acid receptors RARα, β, and γ, which are relevant for the responsiveness to retinoids.

To establish the expression of the full-length cathepsin F in HSCs, we cloned the cDNA using RACE. Rat cathepsin F features the catalytic triad (Cys, His, and Asn), as well as the ERFNAQ motif in a very long propeptide (Figure 1). The rat sequence is highly similar to the mouse sequence, but has one potential N-glycosylation site less. The most striking difference between the rodent and human sequence is a 24-amino acid gap, starting 11 amino acids after the end of the predicted cystatin-like folding of the propeptide.

Strong evidence from our initial results showed that cathepsin F could also be found in the nucleus of HSCs (Figure 2A). The expression of cathepsin F by HSCs in the liver was confirmed using co-immunostaining with GFAP of rat liver sections (Figure 2B, arrowheads). In agreement with a nuclear localization of cathepsin F, we found a co-localization with SC35, a marker for nuclear speckles (Figure 3, A and B, arrowheads). In contrast, cathepsin B does not stain for the same compartment, suggesting different functions of the two cathepsins. Whether this localization implies a function of the cathepsin F in the process of mRNA splicing is hypothetical, although the proteolytic processing of splicing factors is known (Casciola-Rosen et al., 1994; Shav-Tal et al., 2000).

Cathepsin F gives rise to a nuclear protease activity, which we observed in different HSC cell lines (Figure 4A). Consequently we demonstrated that this activity belongs to a cysteine protease of the papain family by showing its strong inhibition using E-64 (Figure 4B), which is specific for cysteine proteases like cathepsins and calpains, but not caspases. We can exclude calpains because of the EDTA contained in the reaction buffer. The inhibition by the caspase inhibitor Ac-YVAD-CMK (Figure 4B) can be explained by the observation of Rozman-Pungercar et al. (2003), which showed that a number of caspase inhibitors are also good inhibitors of cathepsins. The fact that LHVS (Supplementary Figure S3A), specific for cathepsin S at low nanomolar concentrations and inhibits cathepsin F completely at 50 nM (Shi et al., 2000), did not inhibit the nuclear activity to the same extent as E-64 hints at the presence of more than one cathepsin, probably cathepsin B. Nevertheless, using a cathepsin B-specific inhibitor and substrate, we were able to show convincingly that cathepsin F contributes significantly to the nuclear activity (Figure 4D). The observed activities are not inhibited by cathepsin L inhibitors, except for the inhibitor IV, which resulted in a significant decrease of the nuclear activity (Figure 4C). This is probably due to a cross-inhibition of this inhibitor with other related cathepsins.

The mechanism responsible for the translocation of cathepsins into the nucleus is still unclear. The proposed translational initiation at a downstream start site and thereby the deletion of the signal peptide, which leads to a cytosolic accumulation of cathepsin B and L, cannot alone explain the transport into the nucleus (Goulet et al., 2004; Bestvater et al., 2005). Instead, Bestvater et al., (2005) proposed a targeting signal, within a hierarchy of signals (signal peptide, mitochondrial targeting signal, and nuclear targeting signal), for cathepsin B, which facilitates the nuclear import. With this in mind, we noted that the first publication on cathepsin F described a shortened molecule, without a signal peptide, exhibiting a normal activity (Wang et al., 1998). This truncation could lead to a nuclear import of cathepsin F. A SUMOplot prediction (Grammatikoff et al., 2004) reveals a possible sumoylation site (with low probability) for rat cathepsin F. A nearby RKK motif could serve as a nonclassical nuclear localization signal. Thus, one could speculate a possible SUMO-dependent translocation of a truncated cathepsin F into the nucleus. This hypothesis has to be studied further. Another interesting possibility is a glycosylation-dependent import of cathepsins into the nucleus (Monsigny et al., 2004). The prerequisite for this process is a cytosolic localization, which implicate again a truncated cathepsin.

Regulation of Cathepsins, SMAA, and Collagen Type I by Retinol/Palmitic Acid

Observed changes in the cytosolic and nuclear activity during HSC activation process (unpublished data) led us to the assumption that a cathepsin F activity could be involved in this process. We hypothesized that the processing of a nuclear target could somehow influence the regulation of activation markers, such as SMAA and collagen type I. Using primary cells to study the activation process has the disadvantage of a great variability in the cell source; therefore, we used the cell line HSC-2. A previous study reported that the treatment with a combination of retinol/palmitic acid mimics the deactivation (reversal of activation) of HSCs (Abergel et al., 2006). We treated HSC-2 for different times with retinol/palmitic acid, assuming a deactivation of HSCs would cause changes in the activation marker expression and in the detected cathepsin activity. Three-hour treatment did not significantly influence the SMAA and collagen type I mRNAs (Figure 5A). On the other hand, after 48-h treatment with retinol/palmitic acid, we observed a down-regulation of these two mRNAs (Figure 5B), in accordance with data from Abergel et al. (2006). Concurrently, we also studied the regulation of cathepsins F, K, L, and B. With the exception of cathepsin B at 3 h, all cathepsins mRNAs are strongly up-regulated upon treatment with retinol/palmitic acid (Figure 5, A and B), which led to a corresponding increase in cytosolic and nuclear activity. Unexpectedly, after 48 h we observed a dramatic decrease especially in the nuclear activity (Figure 6A). The same effect was observed using in vitro–activated primary HSCs (Supplementary Figure S3B), showing that HSC-2 resembles activated HSCs. After this observation, we hypothesized that an endogenous inhibitor of cysteine proteases could be involved and examined the levels of two cystatins during the same treatment. The mRNAs of cystatin B and C remained unchanged after 3 h, but significantly increased after 48 h (Figure 6B), providing strong grounds that indeed the protease activity in both the cytosolic and nuclear fractions might be affected by endogenous inhibition. Cystatin B is a type 1 cystatin and is located mostly intracellularly, whereas cystatin C is secreted (Turk and Bode, 1991). The immunofluorescence images in Supplementary Figure S4 revealed a partial colocalization with SC35 and hence indicate that cystatin B could be responsible for the inhibition observed. Furthermore, an inhibition of cathepsin F by cystatin B has also been described (Shi et al., 2000). Therefore, we performed knockdown experiments for cystatin B during retinol/palmitic acid treatment to test our hypothesis. Our data clearly show that a decrease in cystatin B transcript and protein (Figure 6D) abolished the effect of retinol/palmitic acid treatment on the nuclear activity (Figure 6C). The change in nuclear activity, even under retinol/palmitic acid treatment, facilitates an up-regulation of SMAA mRNA (Figure 6D), confirming the interplay between nuclear activity and transcriptional regulation of SMAA. In addition, the insignificant changes in cathepsin F transcript, compared with the retinol/palmitic acid treatment, prove that the reduction of cystatin B accounts for the increased nuclear activity.

On the basis of these results, we established that cathepsins F, K, L, and B and cystatin B and C can be regulated by a combination of retinol/palmitic acid. To corroborate this observation, we searched a transcription factor database using TESS (Schug, 2003) for RAR-binding sites, responsible for a regulation by retinoids. There is at least one binding site for either RARβ or RARγ in the first 2000 base pairs upstream of the transcription initiation site of the four cathepsins and cystatin B and C. Although the idea of retinol/palmitic acid regulating the cathepsins and inhibitors simultaneously may seem somewhat contradictory at first, this phenomenon reflects the necessity of a tight regulation of the cathepsins. Considering the spatial separation of enzyme and inhibitor, this regulation should not compromise the normal lysosomal function of cathepsins.

Transcriptional Regulation of SMAA and Collagen Type I by the Nuclear Activity

To further examine the direct effect of the nuclear activity on the regulation of activation markers (SMAA and collagen type I), we treated HSC-2 with a cell-permeable cysteine protease inhibitor, E-64d, for 3 h. Assuming that the regulation observed for the activation markers by retinol/palmitic acid is elicited by the decrease in enzymatic activity, we should be able to see a similar effect upon E-64d treatment. Indeed we observed a significant down- and up-regulation of SMAA and collagen type I transcript, respectively, whereas the cytosolic and nuclear activity is almost completely inhibited (Figure 7, A and B). This observation suggests that the inhibition of the protease activity has differential effects on SMAA and collagen type I. Possible explanations could be that different targets of the nuclear activity are involved; the same target has diverse regulatory effects on SMAA and collagen type I or different cysteine protease activities are affected. Interestingly we found a down-regulation of cathepsin F and K mRNA, which could be another indication for the role of the nuclear activity as a transcriptional regulator.

To ascertain the cathepsin F nuclear activity as the nuclear activity responsible for the regulation of SMAA and collagen type I, we used cathepsin F–specific siRNA to decrease the nuclear activity related to this enzyme. After 48-h treatment with cathepsin F siRNA, we were able to observe a down-regulation of cathepsin F mRNA and protein (Figure 7C) as well as a decrease in cathepsin F activity in the nuclear fraction (Figure 7D). As in all other experiments, we also noted here that the cytosolic activity was less affected than the nuclear activity. This could be due to a greater compensation by other cathepsins or an increased stability of the cathepsin F in the lysosomes. Nevertheless, the decrease in nuclear activity gives rise to a significant down-regulation of SMAA and collagen type I transcripts (Figure 7C). This finding is somewhat surprising because it seems to contradict in part the E-64d results. Possible reasons could be the involvement of other cysteine protease activities or the regulation of another protease activity by the cathepsin F nuclear activity itself. It is also plausible that the cathepsin F nuclear activity is regulated by a different cysteine protease. A candidate could be the cathepsin B, which is also present in the nucleus as shown (Figure 3C). When we overexpressed rat cathepsin F in HSC-2, we found intriguing that the up-regulation of cathepsin F (Figure 8A) led to a down-regulation in nuclear activity (Figure 8B). Taking the up-regulation of cystatin B into account, the findings were similar to the retinol/palmitic acid treatment, except for the up-regulation of the collagen type I mRNA. The different observations among the retinol/palmitic acid, E-64d, cathepsin F siRNA and overexpression of cathepsin F studies could be attributed to the complexity of the regulation of these two activation markers, especially in the case of retinol/palmitic acid.

Our report established the existence of a nuclear cathepsin F in hepatic stellate cells, which is active and localizes in the nuclear speckle compartment within the nucleus. We have presented convincing data to support the hypothesis that the nuclear cathepsin F activity and its regulation by cystatin B is involved in the transcriptional regulation of two HSC activation markers. The results that treatment with retinol/palmitic acid regulates cathepsins and cystatins, with regard to their transcripts and the cathepsin F nuclear activity in particular, implicates a regulatory function of retinyl palmitate (the storage form of retinol in HSCs) in HSCs. Another important finding that the nuclear activity can be regulated by cystatin B highlights the significance of the nuclear activity. It demonstrates that intracellular cystatin B not only prevents unwanted cytosolic protease activity due to lysosomal leakage but also participates in transcriptional regulation. Although in this report the nuclear localization of cathepsin F has been shown for the rat HSCs, we have evidence that the same observation applies to human HSCs (LI90) and rat and human glioblastoma cell lines (unpublished data). Hence, it seems that the nuclear cathepsin F is not only confined to one cell type specifically, but could be a more widespread phenomenon. Further studies on the nuclear activity should concentrate on identifying the nuclear target(s) and other functional implications on HSCs, particularly in vivo.