|Home | About | Journals | Submit | Contact Us | Français|
Kaposi's sarcoma-associated herpesvirus (KSHV) is the pathological agent of Kaposi's sarcoma (KS), a tumor characterized by aberrant proliferation of endothelial-cell-derived spindle cells. Since in many cancers tumorigenesis is associated with an increase in the activity of the cathepsin family, we studied the role of cathepsins in KS using an in vitro model of KSHV-mediated endothelial cell transformation. Small-molecule inhibitors and small interfering RNA (siRNA) targeting CTSB, but not other cathepsins, inhibited KSHV-induced postconfluent proliferation and the formation of spindle cells and foci of dermal microvascular endothelial cells. Interestingly, neither CTSB mRNA nor CTSB protein levels were induced in endothelial cells latently infected with KSHV. Secretion of CTSB was strongly diminished upon KSHV infection. Increased targeting of CTSB to endosomes was caused by the induction by KSHV of the expression of insulin-like growth factor-II receptor (IGF-IIR), a mannose-6-phosphate receptor (M6PR) that binds to cathepsins. Inhibition of IGF-IIR/M6PR expression by siRNA released CTSB for secretion. In contrast to the increased cathepsin secretion observed in most other tumors, viral inhibition of CTSB secretion via induction of an M6PR is crucial for the transformation of endothelial cells.
Kaposi's sarcoma-associated herpesvirus/human herpesvirus 8 (KSHV/HHV8) belongs to the gammaherpesvirus subfamily and is the pathological agent of Kaposi's sarcoma (KS) and of lymphoproliferative disorders in B cells (20). KS is a mesenchymal tumor generally targeting the skin and soft tissue organs. The sarcoma is composed of interweaving bands of vascularizing spindle cells of endothelial origin, frequently associated with infiltrating inflammatory cells (8, 35). AIDS has been most commonly associated with KS, although other forms of KS exist independently of human immunodeficiency virus infection, e.g., classic, endemic, and iatrogenic KS (15, 35). Antiretroviral treatments against human immunodeficiency virus have effectively reduced the prevalence of AIDS-associated KS; however, the incidence of iatrogenic KS has increased 400% relative to the population growth in North America. This is most likely due to an increase in the number of organ transplant recipients (46).
The development of several in vitro systems that rely on infection of primary or immortalized endothelial cells by KSHV has greatly advanced our understanding of virus-host cell interaction during tumorigenesis (31). Dermal microvascular endothelial cells stably transfected with the oncogenes E6 and E7 of human papillomavirus (E-DMVEC) are immortalized but remain contact inhibited, grow as discrete monolayers with a cobblestone phenotype, and enter senescence if not passaged after attaining confluence (37). Upon infection with KSHV, E-DMVEC develop a spindle cell phenotype, lose contact inhibition, form foci when cultured postconfluence, and acquire the ability to form colonies in soft agar (37). As in KS tumors, viral gene expression in E-DMVEC is largely restricted to open reading frame 71 (ORF71 [vFLIP]), ORF72 (vCyclin), ORF73 (latency-associated nuclear antigen [LANA]), and kaposin, and only a few cells spontaneously enter the lytic cycle of viral replication. Despite the restricted viral gene expression program, latent infections are associated with a dramatic reprogramming of the cellular transcriptome and proteome (4, 38, 39, 55). Several virus-induced host cell proteins are required to support postconfluent growth, including c-Kit, hemoxygenase, RDC-1, Neuritin, and the insulin receptor (30, 39, 42, 43).
One of the hallmarks of KS is aberrant neoangiogenesis by proliferating spindle cells, resulting in abnormal vasculature. Neoangiogenesis involves the degradation of extracellular matrix by proteases, a process that frequently involves cysteine proteases of the cathepsin family (6, 21, 22, 53). Despite this important role of cathepsins in many cancers, the role of cathepsin function in KS development has not been studied so far. Cathepsins are part of the papain subfamily within the cysteine protease superfamily. Most cathepsins are ubiquitously expressed in lysosomes and play a role in protein degradation and processing. In addition to these housekeeping functions, cathepsins are implicated in a wide variety of diseases, including cancer (53, 54). In particular, cathepsin B (CTSB) is linked to a number of human cancers, including prostate carcinoma, breast cancer, and brain tumors (19, 27, 49, 51, 57).
In an effort to characterize cathepsin involvement in KSHV-induced tumorigenesis, we now demonstrate that CTSB is an essential factor for KSHV-induced postconfluent growth of E-DMVEC whereas other cathepsins tested did not play an essential role. Unexpectedly, secretion of CTSB was inhibited in latently infected endothelial cells. CTSB is retained by the insulin-like growth factor-II receptor/mannose-6-phosphate receptor (IGF-IIR/M6PR), which we previously reported to be transcriptionally induced during KSHV-mediated transformation of E-DMVEC (43). Therefore, in contrast to the increased secretion of CTSB observed in several other tumor models, CTSB is retained in the endosomal/lysosomal compartment in KS, where it likely regulates the processing of growth factors or growth factor binding proteins. Given the recent progress in developing CTSB inhibitors for cancer treatment (10), our data suggest that such inhibitors might be a novel treatment for KS.
KSHV-infected E-DMVEC were established and maintained as previously described (37). KSHV-infected DMVEC were used in experiments when >90% of the cells expressed LANA-1. CA-074 (catalog no. 205530), CA-074ME (catalog no. 205531), and hydroxy-2-naphthalenyl-methyl phosphonic acid trisacetoxymethyl ester [HNMPA-(AM3)] (catalog no. 397100) were obtained from EMD Biosciences (San Diego, CA). Antibodies against cathepsin S (catalog no. IM1003), cathepsin L (catalog no. 219387), and CTSB (catalog no. IM27L) were purchased from EMD Biosciences (San Diego, CA). A mouse monoclonal antibody against calreticulin (SPA-601) was purchased from Stressgen (Victoria, British Columbia, Canada). Purified human native CTSB (catalog no. JA7741) was obtained from EMD Biosciences (San Diego, CA). A polyclonal antibody against IGF-IIR/M6PR (H-20) was obtained from Santa Cruz Biotechnology (Santa Cruz, CA).
cDNA synthesis, hybridization to HG_U133A and B arrays (Affymetrix), and signal intensity normalization were performed at the Oregon Health and Science University (OHSU) microarray core facility (www.ohsu.edu/gmsr/amc). GeneChip data were analyzed with Arrayassist (Stratagene). Comparisons were made between passage-matched KSHV-infected E-DMVEC and mock-infected E-DMVEC in four separate experiments. A master data table consisting of expression values extracted from all CEL files was created by the probe logarithmic error intensity estimate (PLIER) method. To define the baseline from which changes in gene expression data were determined for KSHV-infected samples, a virtual chip comprising all mock-infected control expression data was used. Significant changes with a P value of 0.01 were determined by analysis of variance with variance stabilization and no P value correction.
For immunoprecipitation, cells were washed twice with ice-cold phosphate-buffered saline (PBS) and lysed in NP-40 buffer (1% NP-40, 150 mM NaCl, 10% glycerol, 20 mM Tris-HCl [pH 8.0], 1 mM EDTA [pH 8.0], 0.2% sodium dodecyl sulfate [SDS]) containing protease inhibitors (Roche Diagnostics; Indianapolis, IN), 1 mM NaVO4, and 1 mg/ml pepstatin. Lysates were cleared, and the protein concentration was determined. For IGF-IIR/M6PR, 500 μg of whole-cell lysate protein was immunoprecipitated with 10 μg of an anti-IGF-IIR/M6PR antibody and incubated overnight at 4°C with rocking. A 50-μl volume of a protein A/G-agarose bead slurry (Upstate) was added for 45 min with rocking at 4°C. Three washes were performed, and the pellet was boiled in 2× SDS sample buffer. The beads were spun down, and the supernatant was separated by 10% SDS-polyacrylamide gel electrophoresis. For Western blotting, cells were directly lysed in 1× SDS sample buffer (100 mM Tris-Cl [pH 6.8], 2% SDS, 10% glycerol) without dyes or with dithiothreitol-2-mercaptoethanol and were subsequently boiled for 5 min. Protein content was normalized and determined using a bicinchoninic acid protein assay kit (catalog no. 23227; Pierce, Rockford, IL). Before resolution of samples by 10% SDS-polyacrylamide gel electrophoresis, 100 mM dithiothreitol and 0.01% bromophenol blue were added to the lysates and samples were boiled for 5 min. Gels were transferred to nylon membranes, blocked for 20 min with 10% milk in 1× PBS supplemented with 0.1% Tween 20, and probed with the respective antibodies. Bound antibodies were detected by enhanced chemiluminescence (Pierce) and exposed to film.
Quantitative PCR was performed on an ABI7700 sequence detection system (Applied Biosystems, Foster City, CA). Target gene expression was compared to expression of the glyceraldehyde-3-phosphate dehydrogenase housekeeping gene. Total RNA was purified from primary DMVEC using RNeasy spin columns (QIAGEN Inc, Valencia, CA). Human tissue was a generous gift from H. Koon and B. J. Dezube and was prepared as previously described (42). Total RNA was treated with DNase I (DNase free; Ambion, Austin, TX) before synthesis of cDNA with random hexamers and Superscript III (Invitrogen, Carlsbad, CA). The following primers were selected by using Primer Express software (Applied Biosystems): IGF-IIR/M6PR 6333F (GCAGAAGCTGGGTGTCATAGG), 6420R (5′-CACGGAGGATGCGGTCTTAT), IGF-IR 2184F (GGAGGAGGCTGAATACCGC), and 2252R (5′-TCAGGTCTGGGCACGAAGAT). Reactions were performed using SYBR green PCR core reagents. Relative expression values for uninfected versus KSHV-infected cells and for normal skin versus KS tissue were calculated by the comparative CT method (28). Dissociation curves were performed on human reference RNA after each amplification run to control for primer dimers. For analysis of CTSB splicing, the following primers were used as previously published for PCR: CTSB full length fwd (5′-CAGCGCTGGGCTGGTGTG), CTSB(−2) fwd (5′-CAGCGCTGGGTGGATCTA), CB(−2/3) fwd (5′-CAGCGCTGGGCCGGGCAC), and CB EXON 11 rev (5′-CCAGGACTGGCACGACAGGC). PCR to determine CTSB isoforms was performed by standard approaches (95°C for 5 min; 30 cycles of 95°C for 1 min, 55°C for 1 min, and 72°C for 2 min; and 72°C for 10 min).
KSHV-infected E-DMVEC were seeded the day prior to small interfering RNA (siRNA) treatment at 30% confluence in 35-mm-diameter polystyrene dishes (Corning). On the day of treatment, for each sample, 3 μl Oligofectamine (Invitrogen) was mixed with 12 μl OptiMEM (GIBCO-Invitrogen) for 5 min at room temperature. A mixture of 200 nM SMARTpool siRNA (Dharmacon) and 180 μl OptiMEM was added to the Oligofectamine mixture and incubated for 20 min at room temperature. For IGF-IIR/M6PR siRNA treatment, single sense strand siRNAs were employed: AGACCAGGCUUGCUCUAUAUU (On-Targetplus set 6), CAACUUUCCUCCAUCACAAUU (On-Targetplus set 7), and CGAUACCUCUCAAGUCAAAUU (On-Targetplus set 8) (catalog no. J-010601-06; Dharmacon). Cells were washed once in OptiMEM, and 800 μl OptiMEM was added to the dish prior to addition of 200 μl siRNA mixture. Cells were incubated for 6 h before receiving a second treatment of siRNA as before. Cells were incubated overnight, and the siRNA mixture was removed and replenished with 2 ml of complete medium. siRNA-treated cells were incubated for 4 to 6 days to determine the peak of downmodulation, at which point protein levels were assayed by Western blotting. For long-term observation of phenotypic changes, siRNA-treated cells were treated again with siRNA 1 week after the first treatment to ensure efficient knockdown. One day after the second treatment series, CTSB activity was measured by a zymogen assay.
KSHV-infected E-DMVEC were seeded at 30 to 50% confluence on 35-mm-diameter dishes (Corning) and treated either with siRNA or with the CTSB small-molecule inhibitor CA-074 or CA-074ME for a 2-week period. Inhibitors (10 μM) were added during the change of medium (every 2 days), whereas a second round of siRNA treatment was performed 7 days after the first treatment to maintain the knockdown of specified genes. For siRNA rescue, 10 ng/ml of native human CTSB was added every 2 days after initial siRNA treatment and continued for 2 weeks total. Cells were examined for spindle cell formation, loss of contact inhibition (a result of adherent foci in the monolayer), or virus-induced cytopathic effect by microscopy. Cells were washed twice in PBS and fixed in 3.7% paraformaldehyde for 20 min before another wash and addition of a coverslip over 80 μl glycerol.
Cells were washed twice with 1× PBS and fixed with 3.7% formaldehyde (in PBS) for 20 min at room temperature. Cells were then washed twice with 1× PBS and permeabilized with 0.1% Triton X-100 (in PBS) for 5 min. Subsequently, cells were washed three times with 2% PBA (bovine serum albumin in PBS). To measure KSHV infection after inhibitor treatment, cells were probed with a primary rabbit polyclonal antibody against LANA-1 (ORF73) (a gift from Bala Chandran, Rosalind Franklin University, Chicago, IL) at 1:500 for 2 h at 37°C. Cells were washed three times with 2% PBA and then incubated with a goat anti-rabbit secondary antibody conjugated with Alexa Fluor 594 (Molecular Probes, Eugene, OR) at 1:500 for 45 min at 37°C. Images were analyzed with a Nikon fluorescence microscope. For live staining, KSHV-infected E-DMVEC were seeded at confluence on Lab-Tek II 1.5 borosilicate chambered coverglass slides (catalog no. 155382; Electron Microscopy Sciences, Hatfield, PA). The next day, cells were either left untreated or treated with either CA-074 (10 μM), CA-074ME (10 μM), or HNMPA-(AM3) (50 μM) overnight before being probed with GB-111 FL (1 μM) for 60 min. Images were acquired by Aurelie Snyder of the OHSU Molecular Microbiology and Immunology Research Core Facility (http://www.ohsu.edu/research/core) with an Applied Precision DeltaVision RT image restoration system. This includes the API chassis with a precision-motorized XYZ stage, an Olympus IX-71 inverted fluorescent microscope with standard filter sets, mercury illumination with an API light homogenizer, a Nikon CoolSnap camera, and DeltaVision software. Deconvolution using the iterative constrained algorithm of Sedat and Agard and additional image processing were performed on a Linux OS platform with the SoftWoRx program. Further technical specifications are available from Applied Precision.
Cell viability by ATP determination was measured according to the manufacturer's protocol (Molecular Probes; catalog no. A-22066). KSHV-infected E-DMVEC were treated with either CA-074 (10 μM), CA-074ME (10 μM), or HNMPA-(AM3) (100 μM) for 4 days in 35-mm-diameter dishes and were lysed in 1% Triton X-100 for 10 min. A d-luciferin-luciferase mixture (5 μl per sample) was added, and the level of ATP was immediately read in a luminometer.
Apoptosis in uninfected and KSHV-infected E-DMVEC was analyzed with the R&D Systems annexin V-fluorescein isothiocyanate (FITC) apoptosis detection kit (catalog no. TA4638). Cells were washed once in PBS, plated in serum-free medium, and either treated overnight with staurosporine (1 μM), CA-074 (10 μM), or CA-074ME (10 μM) or left untreated. Cells were trypsinized and washed in 2% PBA before each sample was resuspended in 100 μl annexin reagent (10 μl 10× binding buffer, 10 μl propidium iodide, 1 μl annexin V-FITC, and 79 μl double-distilled H2O) for 15 min at room temperature. A 400-μl volume of 1× binding buffer was added before samples were analyzed by flow cytometry.
Uninfected or KSHV-infected E-DMVEC were seeded at 3 × 105 cells (for preconfluence) or 5 × 105 cells (for postconfluence) per 35-mm-diameter dish. Cells were grown in the presence or absence of CA-074 (10 μM), CA-074ME (10 μM), or HNMPA-(AM3) (50 μM) for 5 days before being trypsinized, stained with trypan blue, and counted. For the postconfluence assay, cell numbers were normalized to cell numbers in a confluent monolayer of uninfected E-DMVEC (set at 100%) and are presented as percentages of the number of uninfected E-DMVEC.
Forty thousand KSHV-infected E-DMVEC, either treated with CA-074 or CA-074ME or left untreated for 14 days, were plated in 1.5 ml endothelial-SFM growth medium (11111-044; Invitrogen) supplemented with 10% human AB serum, 25 μg/ml endothelial cell growth supplement, and 0.4% melted agarose onto a 3-ml bottom layer of 0.5% agarose medium per well of a six-well dish. The cells were fed several drops of medium every 3 days, and colonies were photographed after 3 weeks.
CTSB activity was measured with the InnoZyme CTSB activity assay kit (catalog no. CBA001; EMD Biosciences, San Diego, CA). To test inhibition by CA-074 and CA-074ME on native CTSB, 10 μM of either inhibitor was combined with 50 ng of native CTSB enzyme in vitro. The fluorogenic substrate Z-Arg-Arg 7-amino-4-methylcoumarin (AMC) was added, and the release of AMC was assayed as recommended by the manufacturer. To measure CTSB enzymatic activity in uninfected and KSHV-infected E-DMVEC lysates and supernatants, cells were seeded in 35-mm-diameter dishes (Corning) in 200 μl medium. Dishes were rocked constantly to cover all cells with medium. The medium was collected, and cells were lysed as suggested in the protocol. Fluorescence was measured using the Flexstation plate reader and analyzed using Softworx Pro (version 4.0.1) software (Molecular Devices, Sunnyvale, CA).
Since cathepsins have been implicated in a variety of human cancers, we wanted to assess the role of cathepsins in KSHV-mediated tumorigenesis by using the E-DMVEC in vitro model. We initially monitored postconfluent growth of E-DMVEC latently infected with KSHV in the presence or absence of the broad-spectrum cysteine protease inhibitor (2S,3S)-trans-epoxysuccinyl-l-leucylamido-3-methylbutane ethyl ester loxistatin (EST). KSHV-infected E-DMVEC were seeded at 30% confluence and treated with EST (58 μM) for 2 weeks; the inhibitor was refreshed every third day. While untreated cells develop a spindle cell phenotype and form foci, EST prevented postconfluent growth and caused KSHV-infected E-DMVEC to remain cobblestone-like as a single monolayer (Fig. (Fig.1A).1A). Similar observations were made with a series of other broad-spectrum cysteine protease inhibitors (data not shown). These studies suggested that cysteine proteases, presumably cathepsins, are important for the continuous growth of KSHV-infected E-DMVEC when the cells reach confluence.
To determine which cathepsins are expressed in KSHV-infected DMVEC, the gene expression profiles for all known cathepsin transcripts were extracted from complete genome expression arrays (Affymetrix U133A and U133B). Results from four independent experiments with KSHV-infected E-DMVEC cultures were compared with results for their passage- and age-matched cultures of uninfected E-DMVEC. Cathepsin transcripts were scored with the Stratagene algorithm and the PLIER method. Whereas a number of cathepsins were expressed in both uninfected and KSHV-infected E-DMVEC, only cathepsin S, represented by two Affymetrix probe sets, was induced upon KSHV infection (Fig. (Fig.2).2). This expression profile was confirmed at the protein level by using specific antibodies in immunoblotting (Fig. (Fig.1B1B).
Using a series of more selective cathepsin inhibitors, we narrowed down the list of potential candidates responsible for the inhibitory effects observed with broad-spectrum inhibitors. In particular, the vinyl sulfone LHVS (N-morpholinurea-leucine-homophenylalanine-vinylsulfone-phenyl) has specificity for cathepsins S, L, and B (9, 41, 50). KSHV-infected E-DMVEC were treated with LHVS (25 μM) as described above. As with broad-spectrum inhibitors, inhibition of postconfluent growth was observed (Fig. (Fig.1A),1A), suggesting that at least one of these proteases is important for KSHV-mediated transformation of E-DMVEC.
To distinguish which of these three cathepsins is necessary for KSHV-induced tumorigenesis, SMARTpool siRNAs were employed against each candidate protease. KSHV-infected E- DMVEC either were treated separately with SMARTpool siRNAs against cathepsin S, L, or B or were mock treated with a Cy3-labeled siRNA against luciferase. siRNA-treated KSHV-infected cultures were monitored over 2 weeks for spindle cell and focus formation. Transfection efficiency was visualized with the Cy3 label conjugated to the luciferase siRNA and was estimated to be consistently around 95%. Western blot analysis of siRNA-treated cultures revealed that all three SMARTpool siRNAs efficiently reduced target protein levels (Fig. (Fig.3A).3A). Furthermore, cathepsin L siRNA inhibited its target cathepsin but did not affect protein expression of CTSB, and vice versa. Despite efficient gene knockdown, cathepsin S and L SMARTpool siRNAs did not impede KSHV-induced tumorigenesis. Compared to that for untreated cells, there was a slight reduction in the level of focus formation in cathepsin L and cathepsin S siRNA-treated cells, probably due to the multiple rounds of transfection, since a reduction was also observed with Cy3-labeled control siRNA (not shown). In contrast, CTSB siRNA dramatically inhibited spindle cell and focus development in KSHV-infected E-DMVEC (Fig. (Fig.3B).3B). These data suggested that CTSB was responsible for the inhibition of KSHV-mediated E-DMVEC transformation by cathepsin inhibitors.
We initially hypothesized that secreted or extracellular matrix-associated CTSB was involved in tumorigenesis by degrading extracellular matrix components (25, 33, 49). To test this hypothesis, we examined whether postconfluent growth of CTSB siRNA-treated cells could be rescued by exogenous addition of native human CTSB to the medium of KSHV-infected E-DMVEC. Even after prolonged treatment with 10 ng/ml native human CTSB, the siRNA treatment still inhibited postconfluent growth of spindle cells (Fig. (Fig.4A),4A), suggesting that endogenous rather than secreted CTSB is essential for KSHV-mediated transformation.
The role of intracellular versus extracellular CTSB was further examined by using two different versions of the activity-based CTSB suicide inhibitor CA-074. CA-074 is unable to cross the cell membrane and consequently can inhibit only extracellular CTSB activity. A methyl ester-modified version of CA-074, proinhibitor CA-074ME, can cross the membrane and inhibit intracellular CTSB, although the modification reduces selectivity for CTSB (34). To ensure that both versions of CA-074 inhibited CTSB activity, they were tested against native human CTSB in vitro. At a concentration of 10 μM, the inhibitors were combined with native human CTSB in the presence of the synthetic fluorogenic CTSB substrate AMC. The release of AMC was measured by fluorescence. Both versions of CA-074 successfully reduced CTSB activity compared to that with no treatment (Fig. (Fig.4B).4B). To examine whether CA-074 or CA-074ME inhibited KSHV-mediated transformation of E-DMVEC, uninfected or KSHV-infected E-DMVEC cultures were treated with 10 μM of the respective CTSB inhibitors over a period of 2 weeks. Untreated KSHV-infected E-DMVEC developed a spindle cell morphology and formed foci (Fig. 4C1). The impermeant version of the inhibitor, CA-074, was unable to inhibit spindle cell and focus formation (Fig. 4C3). In contrast, the development of spindle cell morphology was inhibited by CA-074ME (Fig. 4C2). This result supports the conclusion that intracellular rather than secreted CTSB is required for DMVEC transformation by KSHV.
To locate enzymatically active CTSB in live cells, a fluorescein-conjugated activity-based probe, GB-111 FL, which binds to both active cathepsins L and B, was employed. Prior to probing with GB-111 FL, KSHV-infected E-DMVEC were treated with CA-074 (10 μM) or CA-074ME (10 μM). As a control, a noncompetitive inhibitor of the insulin receptor, HNMPA-(AM3), was used. As shown in Fig. 4D1, untreated cells were probed with GB-111 FL, indicating that active cathepsins L and B are localized to small vesicles. In Fig. 4D2, E-DMVEC treated with the noncompetitive inhibitor HNMPA-(AM3) showed no difference in staining pattern from untreated cells. Probe staining of most cells treated with the impermeant inhibitor CA-074 was similar to that for untreated cells. Some cells did display reduced vesicular staining (Fig. 4D3). This could be due to the ability of unmodified CA-074 to enter the endocytic compartment during long-term treatment, as has been previously documented (49). The amount that is internalized does not seem to be effective at inhibiting postconfluent growth. In contrast, the membrane-permeant inhibitor CA-074ME completely eliminated vesicular staining (Fig. 4D4), a finding consistent with its efficient inhibition of postconfluent growth. We conclude from these findings that CTSB inhibitors can efficiently neutralize CTSB activity and that intracellular active CTSB is essential for in vitro transformation of DMVEC by KSHV.
In our cell culture system, most but not all E-DMVEC are latently infected with KSHV. However, spindle cell-containing foci forming upon postconfluent growth are enriched for latently infected cells (43). Therefore, we wanted to explore whether treatment with a CTSB inhibitor changed the ratio of infected to uninfected E-DMVEC. Cells were stained for LANA-1, which tethers the viral genome to the cellular chromosome, thus giving a characteristic punctate staining pattern. KSHV-infected E-DMVEC were either left untreated or treated for 14 days with either CA-074 (10 μM) or CA-074ME (10 μM) and then stained with a polyclonal antibody against LANA-1 (Fig. (Fig.5A).5A). CA-074ME treatment resulted in a dramatic, 89% reduction in the number of LANA-1 positive cells relative to that of untreated KSHV-infected E-DMVEC (Fig. (Fig.5A,5A, center and top), whereas the imperment inhibitor CA-074 did not change the ratio of LANA-1-positive to LANA-1-negative cells (Fig. (Fig.5,5, bottom). The number of LANA-1-positive cells was quantified by counting individual nuclei as shown in Fig. Fig.5B.5B. These observations indicated that the postconfluent growth advantage of LANA-1 positive KSHV-infected E-DMVEC is reduced when CTSB activity is inhibited with CA-074ME.
To address whether CTSB inhibitors are toxic to cells, an ATP cell viability assay was performed. KSHV-infected E-DMVEC either were cultured in the presence of CA-074 (10 μM) or CA-074ME (10 μM) or were left untreated for 2 weeks. As a positive control, a previously identified inhibitor of KSHV-induced tumorigenesis, HNMPA-(AM3), was used at a concentration (100 μM) that is toxic to the cells (43). Cells reached confluence and formed spindle cell-containing foci, except when treated with CA-074ME. Treated cells were lysed, and ATP was quantitatively determined by a luciferin bioluminescence assay. A decrease in the ATP concentration was seen only with HNMPA-(AM3), as previously reported (43). In contrast, neither CA-074 nor CA-074ME seemed to inhibit cell viability (Fig. (Fig.6A6A).
Additionally, CTSB inhibitor-treated cells were assayed by flow cytometry for phospholipid flipping as an indication of apoptosis by using an annexin V-FITC apoptosis detection kit. Cells and supernatants were collected after 14 days of treatment with CA-074 (10 μM) or CA-074ME (10 μM) and were stained with annexin V-FITC and propidium iodide. The staining pattern distinguishes between early-apoptotic cells (positive for annexin V) and late-apoptotic cells (positive for propidium iodide and annexin V). While neither treated (uninfected [Fig. 6B3 and B4] and KSHV infected [Fig. 6B7 and B8]) nor untreated (Fig. 6B1 and B5) E-DMVEC showed any significant change in the number of viable cells, a large double-positive cell population was visible upon treatment with the apoptosis-inducing compound staurosporine (Fig. 6B2 and B6). The ATP viability assay and the annexin V stain show that the CTSB inhibitor concentrations used are not toxic to cells.
Next, we examined whether CTSB inhibitors affected the proliferation of KSHV-infected E-DMVEC prior to confluence. Cells were seeded at a low density and allowed to grow to confluence while treated with either CA-074 (10 μM), CA-074ME (10 μM), or HNMPA-(AM3) (100 μM). Cells were trypsinized, stained with trypan blue, and counted using a hemocytometer. Treatment with either version of the CTSB inhibitor CA-074 did not reduce cell numbers (Fig. (Fig.7A).7A). As expected, HNMPA-(AM3) reduced cell numbers. These results support the notion that proliferation of KSHV-infected E-DMVEC is not affected by CTSB inhibitors.
To determine whether the postconfluent growth of KSHV-infected E-DMVEC was affected, cells were seeded at confluence and treated with either CA-074 (10 μM), CA-074ME (10 μM), or HNMPA-(AM3) (50 μM) for 2 weeks. Cells were trypsinized, stained with trypan blue, and counted using a hemocytometer. Inhibition of intracellular CTSB activity with the methyl ester version of CA-074 dramatically reduced the number of cells in culture. This result is similar to that for the positive-control HNMPA-(AM3) treatment (at 50 μM). It was shown previously that at this nontoxic concentration, HNMPA-(AM3) efficiently inhibited KSHV-induced tumorigenesis (43). It appears that CA-074ME similarly targeted KSHV-infected cells to inhibit tumorigenesis. In contrast, KSHV-mediated postconfluent proliferation was not affected by the unmodified version of CA-074 but was comparable to that for untreated KSHV-infected E-DMVEC (Fig. (Fig.7B7B).
KSHV-infected E-DMVEC have the ability to form colonies in soft agar, a hallmark of transformation (37, 43). To establish whether KSHV-infected E-DMVEC were able to form colonies upon CTSB inhibition, cells either were treated with CA-074 (10 μM) or CA-074ME (10 μM) or were left untreated for 14 days before being transferred to soft agar. Both cells treated with CA-074 and untreated KSHV-infected E-DMVEC efficiently formed colonies in soft agar. In contrast, the membrane-permeant CTSB inhibitor CA-074ME inhibited colony formation (Fig. (Fig.7C).7C). The total number of colonies formed under each condition is quantified in Fig. Fig.7D.7D. Taken together, the results suggest that CTSB is not required for E-DMVEC proliferation prior to confluence. Instead, CTSB is essential for supporting postconfluent and anchorage-independent growth.
Our data suggest that KSHV infection in E- DMVEC does not alter CTSB gene expression or protein levels, but intracellular CTSB is essential for postconfluent growth. To determine whether intracellularly retained CTSB was enzymatically active, we compared extracellular and intracellular CTSB activity between uninfected and KSHV-infected E-DMVEC using a CTSB zymogen assay. Cells were cultured overnight in a minimal amount of medium, after which both the supernatants and whole-cell lysates were collected. Lysates and supernatants were mixed with the synthetic fluorogenic CTSB substrate AMC, and released AMC was measured by fluorescence. While most of the active CTSB in uninfected E-DMVEC was secreted, almost all of the active CTSB in KSHV-infected E-DMVEC was retained intracellularly (Fig. (Fig.8A8A).
One possible reason for the differential distribution of CTSB was alternative splicing of the CTSB transcript, which contains 13 exons (1, 2, 32, 57). Tumors can contain different species of CTSB mRNA, which can have a dramatic effect on translational efficiency (1) or on cellular trafficking of the translated protein (32). To examine CTSB mRNA splicing, total RNA from uninfected and KSHV-infected E-DMVEC was isolated and analyzed by RT-PCR. Primers were designed against the most dominant CTSB mRNA products as identified in other tumor models: full-length CTSB, CTSB without exon 2 [CB(−2)], or CTSB without exons 2 and 3 [CB(−2/3)] (32). The dominant species in E-DMVEC appeared to be CTSB mRNAs missing exon 2 or both exons 2 and 3, while the full-length transcript was absent. The same splicing pattern was observed in both KSHV-infected and uninfected E-DMVEC (Fig. (Fig.8B),8B), suggesting that differential splicing is most likely not responsible for the subcellular distribution of CTSB.
Fluorescent probes revealed a perinuclear vesicular staining pattern for active CTSB consistent with an endosomal/lysosomal localization (Fig. (Fig.4D).4D). To further determine whether CTSB located to these compartments, we disrupted endocytosis by neutralizing the lysosomal pH with NH4Cl. Treatment with NH4Cl did not inhibit the enzymatic activity of CTSB in vitro (Fig. (Fig.9A).9A). Instead, the absolute intracellular CTSB activity was reduced in both uninfected and KSHV-infected E-DMVEC to the same extent (Fig. (Fig.9B),9B), consistent with a lysosomal/endosomal targeting of CTSB.
Lysosomal targeting of CTSB occurs via M6PRs. The major receptor for sorting CTSB to lysosomes is the IGF-IIR/M6PR (14, 44). Interestingly, we previously observed that IGF-IIR/M6PR expression was highly induced in E-DMVEC latently infected by KSHV (43). Moreover, IGF-IIR/M6PR is highly expressed in KS tissue (Fig. (Fig.9C).9C). To determine whether CTSB associated with IGF-IIR/M6PR, we immunoprecipitated IGF-IIR/M6PR and probed for CTSB using specific antibodies. CTSB was clearly present in IGF-IIR/M6PR immunoprecipitates (Fig. (Fig.9D).9D). The KSHV-mediated increase in IGF-IIR/M6PR expression resulted in a concomitant increase in the level of coimmunoprecipitated CTSB.
To assess whether induced IGF-IIR/M6PR expression was responsible for the increased CTSB retention in the endosomal/lysosomal compartment in KSHV-infected cells, we treated KSHV-infected DMVEC with siRNA against IGF-IIR/M6PR. As shown in Fig. Fig.9E,9E, treatment with siRNA to IGF-IIR/M6PR efficiently reduced mRNA expression levels and had no effect on IGF-IR mRNA. Moreover, cell-associated CTSB activity was significantly reduced in cells treated with IGF-IIR/M6PR siRNA (Fig. (Fig.9F).9F). Therefore, we conclude that IGF-IIR/M6PR is responsible for retaining active CTSB intracellularly by targeting CTSB to the endosomal/lysosomal compartment.
To determine whether the IGF-IIR/M6PR-mediated retention of CTSB was essential for postconfluent growth and focus formation, we treated KSHV-infected DMVEC with IGF-IIR/M6PR siRNA. Cells were seeded at 30% confluence and incubated with three different siRNAs designed against IGF-IIR/M6PR. One week after the first treatment, cells were treated a second time with siRNA. At the end of 2 weeks, E-DMVEC treated with control siRNA (Cy3-labeled siRNA to luciferase) displayed the KSHV-mediated phenotypic change to a spindle cell phenotype. In contrast, all three individual IGF-IIR/M6PR-specific siRNAs efficiently inhibited spindle cell and focus formation (Fig. (Fig.10).10). Taken together with the finding that IGF-IIR/M6PR retained CTSB and that CTSB-specific siRNA inhibits KSHV-mediated postconfluent growth, these data strongly suggest that retention of CTSB by IGF-IIR/M6PR is essential for the development of the KSHV-induced tumorigenic phenotype.
CTSB involvement in tumorigenesis has been recognized in a variety of human tumors (48, 49, 51, 57), but this is the first time that CTSB has been implicated in KS. Latent infection by KSHV transforms immortalized DMVEC in a characteristic way that is reminiscent of observations of KS tumors: formation of LANA-positive spindle cells, loss of contact inhibition, continued proliferation postconfluence, formation of disorganized foci, and the ability to grow in soft agar. An essential function for CTSB in this transformation of E-DMVEC by KSHV was indicated by the finding that each of these steps was inhibited by the membrane-permeant methyl ester-modified CTSB inhibitor CA-074ME but not by the impermeant inhibitor CA-074. Both versions of CA-074 are specific for CTSB in vitro (Fig. (Fig.4B);4B); however, the methyl ester modification reduces the specificity of CA-074ME for CTSB so that it also reacts with cathepsin L in tissue culture and in vivo (34). Treatment with siRNA specific for cathepsin L, B, or S revealed that only the inhibition of CTSB expression prevented KSHV-mediated postconfluent growth. These data strongly suggest that the inhibition of CTSB activity, but not that of cathepsin L activity, was responsible for the effect of CA-074ME. Interestingly, others have shown that treatment with the antisense strand of CTSB efficiently inhibited tumor invasion and angiogenesis in vitro and in mice (23, 25).
It has been documented in other tumor models that intracellular active CTSB can facilitate tumor invasion by acting intracellularly instead of being secreted (16, 49). For E- DMVEC latently infected with KSHV, we observed increased intracellular retention of CTSB concomitant with an increase in intracellular protease activity. CTSB splicing, transcript levels, and protein levels were unchanged by KSHV, with two dominant splice products being present, both missing exon 2 and one also missing exon 3 (Fig. (Fig.8B).8B). The absence of exon 2 has previously been linked to a 15-fold increase in translational efficiency (1). CTSB lacking the sequence encoded by both exons does not enter the secretory pathway, since its signal sequence is also missing (32). Probing for active CTSB revealed that the majority of CTSB activity is located in vesicles, suggesting that CTSB is able to enter the vesicular sorting system and is likely sorted to the endosomal/lysosomal compartment. This is also supported by the finding that cell-associated CTSB activity is depleted upon disruption of lysosomal trafficking with NH4Cl (Fig. (Fig.9).9). Therefore, E-DMVEC latently infected by KSHV display an increased diversion of CTSB from the exocytic route to the endolysosomal compartment.
Our data further suggest that increased endolysosomal targeting of CTSB resulted from increased retention by IGF-IIR/M6PR, which is transcriptionally induced in KSHV-infected cells. IGF-IIR/M6PR is not associated with any known downstream signaling cascade, because the receptor does not possess any signal transduction capacity. One of the functions of IGF-IIR/M6PR is to regulate IGF-II levels (36, 45). Increased release of IGF-II due to a loss of IGF-IIR/M6PR function has been implicated in the promotion of tumor growth in a variety of cancers (13, 14, 29, 40, 56). In addition, IGF-IIR/M6PR is known as the major cation-independent M6PR in endocytosis and intracellular trafficking of mannose 6-phosphate-tagged proteins, including endolysosomal proteases (13). The increased release of active proteases due to decreased expression of IGF-IIR/M6PR can contribute to tumor invasion of surrounding tissues. This has been observed for human breast carcinoma, where increased secretion of cathepsin D correlated with a reduction in IGF-IIR/M6PR expression (3, 14, 29).
In contrast to the IGF-IIR/M6PR downregulation observed in many cancers, we previously noted an upregulation of IGF-IIR/M6PR in the context of a general KSHV-mediated dysregulation of the IGF system (43). We now demonstrate that this transcriptional increase of IGF-IIR leads to increased retention of CTSB. This conclusion is supported by coimmunoprecipitation and by the finding that siRNA treatment against IGF-IIR/M6PR depleted the concentration of active CTSB in KSHV-infected E-DMVEC (Fig. (Fig.9F).9F). Furthermore, inhibition of IGF-IIR/M6PR expression by siRNA inhibited KSHV-mediated spindle cell and focus formation. Therefore, intracellular retention of CTSB by IGF-IIR is essential for KSHV-mediated postconfluent growth.
What is the potential role of intracellular CTSB in the development of KS spindle cells? Lysosomal CTSB is released into the cytosol in the presence of extrinsic stimuli of apoptosis, particularly tumor necrosis factor and tumor necrosis factor-related ligands, and cytosolic CTSB can act as an effector protease of programmed cell death in tumor cells (5, 12, 26). CTSB has also been identified as an inhibitor of p53-dependent apoptosis in a genome-wide screen of p53 modulators (17); however, death receptor-mediated apoptosis is inhibited by the KSHV protein vFLIP, which is expressed during latency (11). Moreover, if CTSB were promoting apoptosis, we would expect that inhibition of CTSB would prevent apoptosis. Inhibition of CTSB activity did not change the level of apoptosis in KSHV-infected cultures (Fig. (Fig.6B).6B). Consequently, regulation of apoptosis is likely not to be involved in the observed function of CTSB in KS.
Another possible effect of CTSB inhibition could be interference with growth-promoting signal transduction cascades. This includes the regulation of IGF-dependent signaling, since cathepsins have been implicated in the cleavage of IGF-binding proteins (45). Indeed, we observed that cathepsins cleave the KSHV-induced IGFBP-2 (our unpublished observations). Therefore, it is conceivable that CTSB is involved in the previously described regulation of the IGF system by KSHV (43). Alternatively, CTSB could be involved in regulating other KSHV-induced protumorigenic host cell or viral factors. For instance, CTSB controls the processing and release of vascular endothelial growth factor (58), which plays an important role in KS (19).
Finally, CTSB could be involved in regulating the activity of matrix metalloproteases (MMP) (21) or other proteolytic processes. MMP and cysteine proteases play an important role in tumor invasion and angiogenesis by degrading various components of the extracellular matrix (24). CTSB has been shown to activate MMP (21). CTSB activity is generally a tightly controlled process, and deregulating CTSB during tumor growth may facilitate angiogenesis, invasion, or even loss of contact inhibition (18, 44, 47). Interestingly, MMP-1 and MMP-9 are induced in KSHV-infected E-DMVEC and have been shown to be essential for degrading various components of the extracellular matrix (A. V. Moses, unpublished observations). Moreover, an MMP inhibitor is in clinical trials for KS (7).
An imbalance in protease activity has been described for a wide range of diseases, including rheumatoid arthritis, osteoarthritis, cancer, neurological disorders, osteoporosis, and lysosomal storage diseases (21, 22, 53). Not surprisingly, 4 of the 11 characterized human cathepsins have been crystallized in an effort to aid structure-based drug design (52). Based on the data presented here, we suggest that CTSB would be a good target for treating KS.
The work reported in this article was supported by grants 5R01-CA099906 and 5R01-CA94011-01 (to A.V.M. and K.F., respectively). M.B. was supported by grants R01-EB005011 and U54-RR020843. P.P.R. was supported by Molecular Hematology Training Grant T32 HL007781-13 and by an award from the American Heart Association.
Microarray assays (and/or data analysis) were performed in the Affymetrix Microarray Core/Spotted Microarray Core/Bioinformatics and Biostatistics Core of the OHSU Gene Microarray Shared Resource. We thank Aurelie Snyder of the OHSU-MMI Research Core Facility (http://www.ohsu.edu/research/core) for help with microscopy and Bala Chandran for his kind gift of the anti-LANA polyclonal antibody.
Published ahead of print on 16 May 2007.