Recombinant cathepsin S propeptide attenuates cell invasion by inhibition of cathepsin L–like proteases in tumor microenvironment

Cathepsin L–like proteases are closely related members of the C1 family of papain-like cysteine proteases and include cathepsins L, K, S, and V (1). Normally localized in the lysosomes, specific roles for these proteases have been established, including their contribution in the processes of antigen presentation (2, 3), bone resorption (4), and extracellular matrix remodeling (5, 6).

One of the main mechanisms to regulate the activity of the cysteine cathepsins is their synthesis as inactive zymogens. The NH₂-terminal propeptide domain of the cathepsin L–like proteases is an inhibitor of the active enzyme in its latent form, which is then removed on its activation (7, 8). Using purified proteases in peptidolytic assays, the ability of the propeptide to act as reversible inhibitors of their cognate protease has been shown (9, 10). Moreover, due to the conserved homology in their sequence, each propeptide can inhibit other cathepsins within this distinct subfamily of cathepsin L–like proteases with low nanomolar potencies (8–12).

In addition to their normal physiologic roles, the aberrant activity of cathepsin L–like proteases has been shown to promote tumorigenesis (13–15). Evidence showing the overexpression, secretion, and inappropriate activity of these proteases suggests that they contribute to the hydrolysis of extracellular matrix directly through their ability to degrade components such as laminin, fibronectin, and collagen (16, 17), thus facilitating the invasive nature of the tumor cells through their microenvironment.

To collectively examine the role of the extracellular cathepsin L–like proteases in tumor cell invasion, we have generated a recombinant human cathepsin S propeptide (rCatSPP). The inhibitory activity of this rCatSPP was determined by steady-state fluorimetric assays, which revealed its ability to selectively inhibit the cathepsin L–like subfamily of lysosomal cysteine proteases. Further demonstration of the ability of the rCatSPP to act as an inhibitor of this subfamily was furnished in a series of in vitro tumor cell invasion assays. Manipulation of the rCatSPP to contain a COOH-terminal IgG Fc-domain enhanced the stability of the recombinant protein and its application to invasion assays showed a significant increase in effectiveness in comparison to the original rCatSPP species.

The U251mg, MDA-MB-231, PC3, and HCT116 cell lines were obtained from the American Type Culture Collection. U251mg and MDA-MB-231 cells were cultured in DMEM, whereas PC3 and HCT116 cell lines were cultured using RPMI 1640 and McCoy's 5A medium, respectively, supplemented with 10% FCS and 2% l-glutamine. All cells were maintained in a humidified environment containing 5% CO₂ at 37°C.

The open reading frame for CatSPP (residues 17-113 of human cathepsin S) was amplified from a human spleen cDNA library (Origene) using primers CatSPP forward (5′-TTTTTTGGATCCCAGTTGCATAAAGATCCTAC) and CatSPP reverse (5′-TTTTTTGTCGACCCGATTAGGGTTTGA) containing BamHI and SalI restriction sites, respectively (as italicized). The expected band of 330 bp was visualized by agarose electrophoresis and subsequently cloned into the pQE-30 bacterial expression vector (Qiagen), incorporating an NH₂-terminal hexahistidine tag [(His)₆ tag] to enable downstream purification. Positive clones were identified by colony PCR and sequence aligned to accession no. M90696. A single verified clone was used in all subsequent experiments. Similarly, the same PCR fragment pertaining to CatSPP was inserted into a modified pRSET vector (Invitrogen) containing the open reading frame for murine IgG_2b Fc CH2 and CH3 domains. Expression from this ligated vector resulted in the generation of the CatSPP with an NH₂-terminal (His)₆ tag as before and including the COOH-terminal murine Fc-domain. Positive clones were verified, as before, by DNA sequencing.

A positive clone was transformed into TOP10F′ Escherichia coli cells and cultured until reaching mid-log phase (A₅₅₀ = 0.5; 37°C). Expression of rCatSPP was induced by the addition of IPTG (1 mmol/L) and propagated for a further 4 h before harvesting. Cell pellets were resuspended and lysed in 50 mmol/L NaH₂PO₄ (pH 8.0) containing 8 mol/L urea, 300 mmol/L NaCl, and 10 mmol/L imidazole. The crude denatured lysate was clarified by centrifugation (10,000 × g, 60 min at 4°C) before application to an immobilized metal affinity chromatography (IMAC) column charged with Ni²⁺ ions (HiTrap 1 mL column; GE Healthcare). Nonspecifically bound material was washed from the column using 50 mmol/L NaH₂PO₄ (pH 8.0) containing 8 mol/L urea, 300 mmol/L NaCl, and 20 mmol/L imidazole followed by on-column refolding by reduction of the urea from 8 to 0 mol/L over 200 column volumes. Refolded column-bound material was washed with a further 20 column volumes of 50 mmol/L NaH₂PO₄ (pH 8.0), 300 mmol/L NaCl, and 20 mmol/L imidazole then eluted with 50 mmol/L NaH₂PO₄ (pH 8.0), 300 mmol/L NaCl, and 250 mmol/L imidazole. Protein fractions were collected, desalted into PBS, and analyzed by SDS-PAGE and Western blotting to determine purity and integrity. Concentrations of the rCatSPP were determined using the BCA method (Pierce) and stocks of purified rCatSPP were stored at -20°C before use.

Similarly, expression of the CatSPP-Fc chimera was done using BL21(DE3)pLysS cells, whereby the cells were cultured and recombinant protein was induced again using IPTG at 37°C for 4 h. Cells were harvested and the rCatSPP-Fc protein isolated and purified via its NH₂-terminal (His)₆ tag using the same procedure for the rCatSPP described above.

Fluorimetric assays were done in triplicate in 96-well microtitre plates in the presence of 100 mmol/L sodium acetate, 1 mmol/L EDTA, 0.1% Brij, and 1 mmol/L DTT (pH 5.5) using cathepsins S, L, K, V, and B (Calbiochem). Cathepsin S activity was monitored using the fluorogenic substrate carbobenzloxy-l-valinyl-l-valinyl-l-arginylamido-4-methyl coumarin (Cbz-Val-Val-Arg-AMC; 25 μmol/L). Similar assays for cathepsins L, K, and V were done using carbobenzloxy-l-phenylalanyl-l-arginylamido-4-methyl coumarin (Cbz-Phe-Arg-AMC; 5 μmol/L for cathepsin L and 25 μmol/L for cathepsin K and V), whereas assays for cathepsin B were done using carbobenzloxy-l-arginylamido-l-arginylamido-4-methyl coumarin (Cbz-Arg-Arg-AMC; 25 μmol/L) as substrates. Purified rCatSPP or rCatSPP-Fc was added to assays at various concentrations and the resultant fluorescence was measured using a Cytofluor 4000 spectrofluorimeter with excitation at 395 nm and emission at 460 nm wavelengths. Fluorimetric assays using the fluorescently labeled DQ-elastin and DQ-gelatin (Invitrogen) were done as above (100 μg/mL), with excitation and emission wavelengths of 480 and 530 nm, respectively.

The relative level of cathepsin L–like subfamily activity in the human malignant cell lines was also determined by fluorimetric assays. Conditioned medium was retained and cell lysates were prepared using a 100 mmol/L sodium acetate lysis buffer containing 100 mmol/L NaCl and 0.1% Triton X-100. The cell suspension was sonicated to ensure complete cell disruption, and after centrifugation, cell lysates and conditioned medium samples were subjected to BCA analysis to determine protein concentration. The relative levels of cathepsin L–like activity were measured by incubating samples (100 μg/mL) with 0.5 mol/L MES buffer (pH 6.0) containing 2 mmol/L DTT. The substrate Cbz-Val-Val-Arg-AMC (5 μmol/L) was added, and the rate of substrate hydrolysis at 37°C was monitored every 60 s over a period of 60 min.

RNA was extracted from the mammalian cell lines using the Absolutely RNA RT-PCR Miniprep kit (Stratagene). Reverse transcription-PCR was done using the One-Step RT-PCR kit (Qiagen) under the following conditions: 50°C for 30 min, 95°C for 15 min, and 35 cycles of 94°C for 1 min, 55°C for 90 s, and 72°C for 1 min followed by a final incubation at 72°C for 10 min. Primer sequences for target genes were as follows: cathepsin S forward GGGTACCTCATGTGACAAG and reverse TCACTTCTTCACTGGTCATG, cathepsin L forward ATGAATCCTACACTCATCCTTGC and reverse TCACACAGTGGGGTAGCTGGCTGCTG, cathepsin K forward ATGTGGGGGCTCAAGGTTCTGC and reverse TCACATCTTGGGGAAGCTGGCC, cathepsin V forward ATGAATCTTTCGCTCGTCCTGGC and reverse TCACACATTGGGGTAGCTGGC, and actin forward ATCTGGCACCACACCTTCTACAATGAGCTGCG and reverse CGTCATACTCCTGCTTGCTGATCCACATCTGC.

In vitro invasion assays were done as described previously (18). Briefly, cells were incubated in the presence of predetermined concentrations of rCatSPP, rCatSPP-Fc, or appropriate controls, and invasion plates were incubated at 37°C and 5% CO₂ for 24 h. Noninvaded cells were removed and invaded cells were fixed in Carnoy's fixative for 15 min. After drying, the nuclei of the invaded cells were stained with Hoechst 33258 (50 ng/mL) in PBS for 30 min at room temperature. The chamber insert was washed twice in PBS and mounted in Citifluor, and invaded cells were viewed with a Nikon Eclipse TE300 fluorescent microscope. All assays were done in triplicate and 10 digital images of representative fields from each of the triplicate membranes were taken using a Nikon DXM1200 digital camera at magnification of ×20. Results were analyzed using Lucia GF 4.60 software by Laboratory Imaging. The results were expressed using untreated control cells (0 nmol/L) as 100% cell invasion, and all other readings were expressed as percentage ± SD cell invasion. The Student's t test was used to determine the statistical significance of the reduction in invasion observed in the presence of the rCatSPP. To determine EC₅₀ values for rCatSPP and rCatSPP-Fc, relative rates of invasion were subjected to nonlinear regression analysis using Prism software.

Cytotoxicity of the rCatSPP was determined using the MTT cell viability assay. Briefly, cells were seeded at 1 × 10⁴ per well of a 96-well microtiter plate in 200 μL medium. Appropriate concentrations of rCatSPP (200 nmol/L), control protein, or vehicle-only control were added to the cells and incubated at 37°C and 5% CO₂ for 24, 48, and 72 h, respectively. After incubation, conditioned medium was removed and 0.5 mg/mL MTT (200 μL) was added to the wells and incubated for a further 2 h at 37°C and 5% CO₂. The MTT reagent was carefully removed and the insoluble formazan crystals were dissolved in 100 μL/well DMSO. Absorbance was measured at 570 nm and the results were expressed as the percentage cell viability relative to vehicle-only control. All tests were done in quintuplicate.

To compare the relative stabilities of the rCatSPP and rCatSPP-Fc, 500 nmol/L of each protein were incubated with HCT116 cells and stability of the proteins was evaluated by Western blotting of the cell supernatants over 24 h. Supernatants were resolved by SDS-PAGE using 4% to 12% bis-Tris gels and MOPS running buffer and proteins were subsequently transferred to polyvinylidene difluoride membranes and immunoblotted using an anti-polyhistidine tag horseradish peroxidase–conjugated antibody (1:3,000; Sigma).

To enable the generation of rCatSPP, the DNA sequence encoding the CatSPP (residues 17-113) was amplified and cloned into the bacterial expression vector, enabling the expression of the domain with an NH₂-terminal (His)₆ tag at the expected molecular weight of ∼16 kDa. Expression of the rCatSPP from TOP10F′ E. coli cells was shown to be inducible and regulated by the addition of IPTG through the analysis of cell lysates by SDS-PAGE and Coomassie blue staining (Fig. 1A). The presence of the (His)₆ tag was confirmed by immunoblotting with an anti-polyhistidine horseradish peroxidase–conjugated antibody (Fig. 1B). The (His)₆-tagged protein was expressed insolubly (data not shown); therefore, after harvesting, cells were lysed and proteins were denatured in 8 mol/L urea phosphate buffer before application to a Ni²⁺-charged IMAC column. On-column refolding was done by the complete removal of the chaotropic agent over a linear gradient from 8 to 0 mol/L. The column bound material was subsequently eluted using an imidazole gradient (0-250 mmol/L), resulting in the generation of two elution peaks (Fig. 1C). Analysis of eluted fractions from the second purification peak by SDS-PAGE identified the presence of the purified rCatSPP (Fig. 1D).

The inhibitory activity of the refolded rCatSPP was determined initially against recombinant, mature human cathepsin S, employing a steady-state fluorimetric assay. Typical progress curves for the cathepsin S–catalyzed hydrolysis of Cbz-Val-Val-Arg-AMC in the presence of varying concentrations of the rCatSPP are shown in Fig. 2A (main graph). These progress curves are indicative of the action of a slow tight-binding reversible inhibitor. The progress curves were subjected to nonlinear regression analysis (19) where the production of fluorescence [P] over time can be represented by the following equation:

In this equation, v_s is the final steady state rate and v₀ is the initial steady state rate for the formation of product. The term d is a displacement term that represents the concentration of product at t = 0. The apparent first-order rate constant k_obs is a reflection of the time-dependent nature of the transition between the initial and the steady state rates. Using GraFit software, data points from the progress curves were fitted to Eq. (1) and values of v_s were then determined for each inhibitor concentration employed. These were then used to generate Morrison plots (v_s against [I]; Fig. 2A, inset) from which K_{i (observed)} was determined. This was then corrected to account for competing substrate, as shown in Eq. (2), giving a true K_i value of 2.5 ± 0.12 nmol/L against cathepsin S.

The ability of the rCatSPP to inhibit the peptidolytic activities of recombinant human cathepsins K, L, and V were also determined by fluorimetric assays (Fig. 2B-D, main graphs). The final steady-state rates from the resultant progress curves for the inhibition of these proteases with rCatSPP were also calculated and used to generate additional “Morrison plots” (Fig. 2B–D, insets). The rCatSPP was found to function as a potent inhibitor of each of the cathepsin L–like proteases, exhibiting K_i values of 13.9 ± 2.4 nmol/L, 17.6 ± 1.3 nmol/L, and 4.3 ± 0.6 nmol/L for cathepsins L, K and V, respectively. No discernable inhibitory activity against the more distantly related cathepsin B was observed (data not shown).

The low nanomolar K_i values obtained in this investigation are comparable with those reported previously for the inhibition of the cathepsin L–like proteases by their respective propeptides, but there are some minor differences between the reported values and those determined herein. However, these can be attributed to the differences in the composition of assay buffers used in the present and previous studies. Previously, Guay et al. showed that the replacement of a 50 mmol/L citrate assay buffer by a 50 mmol/L MES buffer led to the total abolition of the inhibition of cathepsins L, S, and K by the rCatKPP (9). In these experiments, the fluorimetric assays were done in a 100 mmol/L sodium acetate buffer (pH 5.5), as has been published previously (20, 21), and is suitable for use with the each of the cathepsin L–like cysteine cathepsins.

The ability of the rCatSPP to attenuate the elastinolytic and collagenolytic activity of cathepsin S was determined by fluorimetric assays using the fluorescently labeled (BODIPY) elastin and gelatin proteolytic substrates, DQ-elastin and DQ-gelatin. The assays were done using activated cathepsin S in the presence of increasing concentrations rCatSPP. The rCatSPP appears to have the ability to attenuate the cleavage of both DQ-elastin and DQ-gelatin (Fig. 3). A reduction in the ability of cathepsin S to cleave elastin was observed, with 60% reduction is elastinolytic activity observed in the presence of the rCatSPP (200 nmol/L; Fig. 3A). Likewise, a 80% reduction in cathepsin S–mediated collagenolytic activity was also observed in the presence of the rCatSPP (200 nmol/L; Fig. 3B).

Next, we investigated if the in vitro inhibition of the cathepsin L–like proteases by the rCatSPP would result in blockage of tumor cell invasion. Cell lines representative of different types of invasive malignancy were selected: PC3 (prostate), HCT116 (colorectal), U251mg (astrocytoma), and MDA-MB-231 (breast). The presence of the different cathepsin L–like subfamily proteases within the panel of human malignant cell lines was verified by reverse transcription-PCR (Fig. 4A). The presence of mRNA for each protease was clearly evident in all four cell lines, with the exception of cathepsin V. Cathepsin V expression was almost negligible in the MDA-MB-231 breast carcinoma cell line. In addition, the activity of the cathepsin L–like proteases was determined by fluorimetric assay in the conditioned medium and cell lysates (Fig. 4B), further showing that these proteases are expressed in and secreted from these tumor cell lines.

The human malignant cell lines were examined in a series of in vitro invasion assays to ascertain the extent to which the secreted cathepsin L–like proteases promote tumor cell invasion through their inhibition by the rCatSPP. Tumor cells were incubated with the rCatSPP (0-200 nmol/L) for 24 h to determine the relative rates of invasion through the Matrigel-coated Transwells. As controls, an unrelated recombinant protein produced from the same bacterial expression vector (200 nmol/L) and a vehicle-only buffer control were also incubated with the cells under identical conditions. The rCatSPP was found to attenuate the rate of tumor cell invasion in a dose-dependent fashion in each of the cell lines examined (Fig. 4C). The greatest attenuation of invasion was observed with the HCT116 colorectal carcinoma cells, with a 58% reduction in invasion observed in the presence of 200 nmol/L of the rCatSPP (P < 0.0001; Fig. 4C, first panel). A significant inhibition of tumor cell invasion was also observed for the other cell lines, with a 48%, 47%, and 35% reduction in invasion observed for PC3, MDA-MB-231, and U251mg cells in the presence of 200 nmol/L of the rCatSPP (P < 0.01, <0.0001, and <0.001, respectively; Fig. 4C, remaining panels).

To confirm that the observed reduction in tumor invasion was not simply due to antiproliferative or cytotoxic effects of the rCatSPP, MTT assays were done. For example, HCT116 cells were incubated with the rCatSPP (200 nmol/L), control protein (200 nmol/L), or vehicle-only control for 24, 48, and 72 h and the number of viable cells was counted/compared for each treatment. No significant effects on cell viability by the rCatSPP or control protein were detected; therefore, the anti-invasive effects observed were solely due to the inhibition of the cathepsin L–like proteases, which facilitate tumor invasion (Fig. 4D).

Based on the significant anti-invasive effects observed with the rCatSPP, we asked if the effects could be improved via fusion with a COOH-terminal murine IgG_2b Fc-domain. Fusion of small therapeutically useful peptides and proteins is useful to stabilize the protein by increasing its half-life, generally as a result of increased size in the resultant fusion protein (22). The recombinant CatSPP-Fc chimera was expressed and purified from bacterial cultures by the NH₂-terminal (His)₆ tag using the identical procedure as the original rCatSPP species using a Ni²⁺-charged IMAC column. The purification profile of the rCatSPP-Fc revealed two main peaks (Fig. 5A) and analysis of the elution fractions from the second peak by SDS-PAGE identified the presence of a highly purified protein species with a molecular weight of ∼46 kDa, as expected for the rCatSPP-Fc (Fig. 5B).

Fluorimetric assays were used to confirm that the rCatSPP-Fc had retained its ability to act as a potent inhibitor of the cathepsin L–like cathepsins and that the COOH-terminal Fc-domain had no negative effects on the kinetics of inhibition. Consequently, rCatSPP-Fc (0-200 nmol/L) was incubated with human cathepsin S in the presence of Cbz-Val-Val-Arg-AMC (25 μmol/L) and the resultant progress curves revealed a dose-dependent inhibition of cathepsin S activity (Fig. 5C). Subsequent Morrison plots revealed that the inhibition constant for the rCatSPP-Fc was reduced ∼3.5-fold with an observed K_i of 8.9 ± 2.5 nmol/L in comparison with 2.5 ± 0.12 nmol/L for the original rCatSPP. These effects are probably due to some slight changes in the conformation of the propeptide domain as a result of fusing the Fc-domain onto its COOH terminus. As a control, NH₂-terminally labeled (His)₆-tagged Fc-domain protein was expressed, purified, and tested for inhibitory activity using the same fluorimetric assay to confirm that the Fc-domain had no effects on inhibition. It was found that incubation of 200 nmol/L of this control Fc protein with human cathepsin S showed no discernable effects on cathepsin S activity (Fig. 5D).

Having determined that the rCatSPP-Fc could potently inhibit the peptidolytic activity of cathepsin S, albeit at slightly reduced potency, its ability to attenuate tumor cell invasion was investigated in parallel with the original rCatSPP species. Incubation of MDA-MB-231 cells with the rCatSPP (0-250 nmol/L) confirmed the dose-dependent effects observed earlier, with 44% reduction in tumor cell invasion in the presence of 250 nmol/L rCatSPP (Fig. 6A, top). However, a significant reduction in tumor cell invasion was observed in the presence of much lower inhibitor concentrations of the rCatSPP-Fc; a 46% reduction in invasion was observed in the presence of only 5 nmol/L of the rCatSPP-Fc, with up to 75% reduction observed in the presence of 50 nmol/L (Fig. 6A, bottom).

The relative inhibition/reduction of tumor cell invasion in the presence of the two propeptide species were subjected to nonlinear regression analysis using Prism software, with sigmoidal dose-response curves plotted to elucidate the relative EC₅₀ values for each inhibitor. The rCatSPP was found to have an EC₅₀ value of 78.0 nmol/L, whereas the increased potency of the rCatSPP-Fc was clearly evident from the EC₅₀ value calculated as 8.3 nmol/L (Fig. 6B).

Finally, to ascertain if the addition of the Fc-domain had enhanced the stability of the protein, resulting in the improvement in EC₅₀, the rCatSPP and the rCatSPP-Fc proteins (500 nmol/L) were incubated with HCT116 cells for 24 h. Conditioned medium was collected and assessed by SDS-PAGE with Western blotting using an anti-polyhistidine tag antibody. The results of these studies are shown in Fig. 6C. The presence of the rCatSPP is clearly evident in the supernatant at the zero time point but cannot be detected beyond this point (Fig. 6C, top). In comparison, the stability of the rCatSPP-Fc is clearly enhanced by the addition of the COOH-terminal Fc-domain. The rCatSPP-Fc is clearly present in the zero time point and is still evident after 24 h, although levels appear reduced in a time-dependent manner (Fig. 6C, bottom). Finally, whole-cell lysate samples were analyzed by Western blot to see if either propeptide species had been taken up by the cells, but detection of the proteins via their (His)₆-tag proved negative (data not shown).

Several cysteine cathepsins have been implicated in the invasion and metastatic processes of a range of different malignancies (15, 23, 24). The overexpression and secretion from tumor cells of these proteases frequently correlates with increasing grade of tumor invasiveness (18, 25, 26). Normally, only active at the low pH of activated lysosomes, the cathepsin L–like proteases were originally perceived to have only limited roles in the higher pH of the tumor microenvironment. However, demonstration that at neutral pH cathepsin S can maintain activity (27) and that cathepsin L activity can be stabilized by p41 extracellularly (28) suggest that secreted cathepsin L–like species might play a role in cell invasion. Indeed, a substantial amount of evidence points toward their role in the degradation of elastin, collagen, laminin, and other components of the extracellular matrix (29–33). Furthermore, their role in the coactivation and activation of other proteases, such as the metalloproteases, has also been suggested (34).

Aberrant cathepsin S expression was shown to be significantly greater in astrocytoma and prostate carcinoma sections in comparison with normal tissues (18, 35, 36), whereas cathepsin L overexpression has been identified in a wide range of human malignancies including melanomas, meningiomas, and colorectal carcinomas (37–39). Cathepsin K overexpression has been associated with invasive lung adenocarcinomas (40) and breast carcinomas (41) and also identified as the principal protease in giant cell tumor of the bone (42). Cathepsin V was originally identified in colorectal and breast carcinomas as well as certain ovarian and renal cell carcinomas (24) and is currently one of only two proteases employed in a multigene signature assay to predict recurrence of tamoxifen-treated, node-negative breast carcinomas (43).

Recently, it has been shown that there is involvement of several cysteine cathepsins in tumorigenesis where increases in their expression were observed in a transgenic mouse model of pancreatic carcinoma. Expression of cathepsins L and S and also the more distantly related cathepsins C, H, Z, and B were all increased in the murine angiogenic islets and tumors. Similar effects were also observed in a second mouse model of human cervical carcinogenesis, suggesting that it may be a general phenomenon rather than restricted to particular tumor types (14).

A more comprehensive investigation determining the distinct roles of the cysteine cathepsins in multistage tumorigenesis was done using a RIP1-Tag2 mouse model of pancreatic islet cell carcinogenesis and specific cathepsin-null mice. This study implicated cathepsins B and S in tumor formation and angiogenesis, whereas cathepsins B and L had roles in tumor cell proliferation and tumor growth. Interestingly, the null mutation of cathepsins B, L, or S leads to a perturbation in the progression of the invasive carcinoma, suggesting a possible nonredundant role of each of these proteases (15).

In this study, we have developed a broad-spectrum inhibition strategy to examine the role of cathepsin L–like proteases secreted from tumor cells. The generation of recombinant cathepsin propeptides has been shown previously, where the production of these recombinant proteins resulted in the formation of insoluble inclusion bodies, which necessitated laborious purification and refolding procedures using a GdnHCl concentration gradient (8) or through affinity chromatography and subsequent cleavage using a GST-fusion construct (9). Here, we applied an alternative simplified strategy for the production of the CatSPP recombinant protein using IMAC and on-column refolding, which we have employed previously for the characterization of CatVPP (12). These previous propeptides were kinetically characterized in vitro, showing nanomolar inhibitory activities against the cathepsin L–like proteases (9–12), which we have similarly found with the rCatSPP species in this current investigation. Using the specific example of cathepsin S, we showed that the prodomain can block its cleavage of nonfibrillar collagen (gelatin) and elastin in in vitro assays, consistent with previous reports that these proteases are involved in the degradation of the extracellular matrix and in agreement that their inhibition will attenuate invasion.

Based on these results, we proceeded to investigate if the rCatSPP could actually be employed in cell-based assays to inhibit the cathepsin L–like proteases in the tumor microenvironment and block cell invasion processes. We observed the significant diminution of cell invasion across a range of cancer cell lines, representative of some of the common forms of tumor, that is, astrocytoma and breast, colorectal, and prostate carcinomas. The propeptide is a small basic polypeptide with limited secondary structure (10) and therefore is likely to have a relatively short half-life in these cell-based invasion assays due to breakdown by other protease species. As anticipated, the fusion of the polypeptide with a COOH-terminal murine IgG Fc-domain significantly improved its half-life and thus its anti-invasive potency. Furthermore, we were unable to detect either prodomain species in whole-cell lysates, showing noninternalization of intact active protein, indicating the effect we were observing was most likely specific to inhibition of extracellular cathepsin L–like activity only.

Our data would suggest that this approach for inhibition of the cathepsin L–like proteases could form the basis of a novel therapeutic strategy to inhibit tumor progression. The synthesis of small-molecule inhibitors for the cathepsin L–like proteases, both reversible and irreversible, has been well documented (44–47). However, the therapeutic application of such compounds is limited by toxicity concerns caused by their possible collateral inhibition of intracellular cathepsin L–like proteases in normal tissues (48). Therefore, alternative strategies that can target only secreted proteolytic activities as observed in pathologies such as malignancy are attractive. Furthermore, inhibitors that have broad specificity against this entire subfamily of proteases may prove more useful due to redundancy in function that has been shown from gene knockout studies (49, 50). Consequently, an appropriate next step will be the evaluation of this fusion protein to attenuate tumor burden and associated tumorigenic processes in in vivo studies. Given that this compound is peptide based, its application in animal models such as the RIP1-Tag2 mouse is essential to ensure that efficacious bioavailability and acceptable pharmacokinetics can be achieved.