Abstract
Inflammation is characterized by an excess of cell proliferation often leading to fibrosis and sclerosis with subsequent loss of organ function. We hypothesized that these features may be ameliorated by induction of cell cycle arrest and apoptosis as result of therapy with matrix metalloproteinase (MMP) inhibitors. In our study, mesangial cells and experimental mesangial proliferative glomerulonephritis provided the model of inflammation. First, we investigated the effect of the MMP inhibitor BB-1101 in anti-Thy1.1 nephritis. The numbers of apoptotic glomerular cells in nephritic rats increased about 4 and 6 times as a result of BB-1101 therapy, observed 11 and 14 days after induction of disease, respectively. Subsequently, rat mesangial cells were exposed to an MMP inhibitor in vitro. Fluorescence-activated cell sorter analyses of cells exposed to RO111-3456 demonstrated a dose-dependent cell cycle arrest in the G0/G1phase associated with increased expression of statin. The cell cycle arrest was followed by apoptosis as investigated by terminal deoxynucleotidyl transferase (TdT)-mediated deoxyuridine triphosphate (dUTP) biotin nick-end labeling (TUNEL) and acridine orange/ethidium bromide stainings, as well as by annexin V binding. The induction of p53, p21, and bax, but not the Fas/FasL pathway appeared to play an important pathogenetic role. In summary, MMP inhibitors induce cell cycle arrest followed by apoptosis in mesangial cells. These features help to explain the anti-inflammatory effects of these compounds, such as reduction of mesangial cell proliferation and attenuation of extracellular matrix accumulation. In conclusion, induction of cell cycle arrest with subsequent apoptosis may offer new perspectives in the therapy of inflammation even beyond kidney diseases.
Inflammation is often accompanied and characterized by an unwanted excess of cell proliferation that may ultimately lead to fibrosis and sclerosis. Sclerosis of renal tissues, such as the glomerulus, is an important cause of end-stage kidney disease in humans. Therapy of end-stage kidney disease requires demanding and expensive methods, such as dialysis and renal transplantation.
In the kidney, there exists a whole spectrum of proliferative forms of glomerular inflammatory disorders. Mesangial proliferative glomerulonephritis is a particularly prominent example of a difficult to treat inflammatory renal disease (Couser, 1999). Increased mesangial cell proliferation also plays an important role in other glomerular diseases, such as diabetic nephropathy. Therefore, we have chosen the mesangial cell and experimental mesangial proliferative glomerulonephritis as a model for the investigation of a new anti-inflammatory therapy.
Induction of cell cycle arrest and apoptosis represents an established method to treat malignant disorders (Shapiro et al., 1999). A similar approach may also be useful for the therapy of inflammation (Gao et al., 1998). We used matrix metalloproteinases (MMP) and their synthetic inhibitors to evaluate such a strategy. In mesangial cells, inhibition of MMP expression and activity has profound effects beyond the degradation of extracellular matrix (ECM) proteins. The constitutive synthesis of mesangial cell MMP-2 was greatly reduced with antisense RNA produced by an episomally replicating vector or with specific anti-MMP-2 ribozymes expressed by a retroviral transducing vector (Turck et al., 1996). The transfected or retrovirally infected cells reverted from a proliferative, inflammatory phenotype to the quiescent state that occurs in the normal renal glomerulus. The distinct differences included changes in synthesis of ECM proteins, loss of activation markers, and an almost total stop of proliferation (Turck et al., 1996).
These studies were extended by the use of a synthetic MMP inhibitor. Exposure of mesangial cells to Ro 31-9790 in vitro resulted in a dose-dependent and reversible inhibition of cell proliferation, associated with a reduction in the expression of α-smooth muscle actin (Steinmann-Niggli et al., 1997).
Elevated MMP expression and activity occur in inflammatory disorders of the kidney. After induction of anti-Thy 1.1 nephritis in rats, mesangiolysis is followed by an increase in activation and proliferation of mesangial cells, associated with augmented expression of MMP-2 (Lovett et al., 1992; Marti et al., 1994). Therefore, we treated these nephritic rats with a synthetic MMP inhibitor. Total glomerular cell count reflecting the degree of mesangial cell proliferation and ECM accumulation were significantly reduced (Steinmann-Niggli et al., 1998). Consequently, glomerular histology was markedly improved and proteinuria showed a clear tendency toward a decrease.
It is known from studies with tumor cells that synthetic MMP inhibitors may induce cell cycle arrest or even promote apoptosis (Burke et al., 1997; Erba et al., 1999). Therefore, we hypothesized that these compounds may exert similar effects in nontumor cells also. In the present study, we used BB-1101 to demonstrate the induction of MMP inhibitor-related mesangial cell apoptosis in experimental mesangial proliferative glomerulonephritis. Subsequently, we studied the mechanisms of this phenomenon in cultured mesangial cells by the use of RO111-3456, a closely related compound.
Materials and Methods
Animals and Antibodies.
Male Wistar rats were obtained from the local animal facilities at our hospital. Approval for rat studies was attained from the commission for animal studies, a local government agency. Anti-Thy1.1 IgG (OX-7) was prepared as described previously (Steinmann-Niggli et al., 1998). A mouse monoclonal anti-statin IgM was provided by Dr. Eugenia Wang, McGill University, Quebec, Canada (Wang and Pandey, 1995). Anti-c-myc antibody has been described previously (Brunner et al., 2000).
MMP-Inhibiting Agents.
The MMP inhibitor BB-1101 (Mr = 389.5 g/mol) was obtained from British Biotech Pharmaceuticals, Oxford, UK. According to the instructions by the manufacturer, BB-1101 is designed for in vivo studies and is only for limited use in in vitro experiments because of low aqueous solubility in cell culture medium. BB-1101 was suspended in PBS/0.1% Tween 80 (v/v) and sonicated for 5 min before injection. BB-1101 was used in a dose of 30 mg/kg of body weight per day, given by single daily intraperitoneal injections, as described (Steinmann-Niggli et al., 1998).
For in vitro studies, we used the MMP inhibitor RO111-3456 and captopril. RO111-3456 (Mr = 425.89 g/mol) represents a peptide-based hydroxamic acid derivative, very similar to BB-1101. RO111-3456 was received as a gift from Dr. J. Caulfield, Roche Bioscience, Palo Alto, CA. For mesangial cell culture experiments, concentrations of 5 to 100 μM RO111-3456 in a final concentration of 0.03% DMSO were used.
For both inhibitors, BB-1101 and RO111-3456, the IC50 concentrations for MMP inhibition are in the nanomolar range (personal communication by the manufacturer;Steinmann-Niggli et al., 1998).
Captopril (Mr = 217.29 g/mol; Bristol-Myers Squibb AG, Baar, Switzerland), a well known angiotensin-converting enzyme inhibitor, inhibits MMP-2 and MMP-9 activity also by interacting with the zinc ion at their active sites (Sorbi et al., 1993; Trocme et al., 1998). Importantly, it represents an agent with a radically different structure than the hydroxamic acid-based MMP inhibitors. Captopril was prepared as stock solution of 250 mM in culture medium, and final concentrations of 1 to 30 mM were used for cell culture experiments and for zymography, exactly as described (Sorbi et al., 1993; Trocme et al., 1998).
Anti-Thy1.1 Nephritis: Experimental Design.
Four groups of male Wistar rats (150 g of body weight at day 0) were studied (n = 54): group A, healthy rats (n = 9); group B, pretreated healthy rats (n = 9); group C, nephritic rats (n = 18); group D, pretreated nephritic rats (n = 18). MMP inhibitor BB-1101 (groups B and D) or the respective amount of solvent (groups A and C) was given intraperitoneally once daily from day −2 to +14, as described (Steinmann-Niggli et al., 1998). Anti-Thy1.1 nephritis was induced at day 0 (groups C and D) as described previously (Steinmann-Niggli et al., 1998); healthy control rats (groups A and B) received the respective amount of PBS only. Nephrectomy for histological analyses was performed at days +4, +8, +11, and +14 after induction of disease. Blood levels of BB-1101 at day +11 were analyzed by British Biotech Pharmaceuticals as reported (Steinmann-Niggli et al., 1998).
Histological Analyses.
Renal tissues were fixed for 24 h in 5% buffered formalin, dehydrated, and embedded in paraffin. Subsequently, kidneys were cut longitudinally into 2-μm sections for periodic acid Schiff reaction, TUNEL, and α-smooth muscle actin staining, as described previously (Baker et al., 1994; Ziswiler et al., 1998).
Glomerular cross sections containing only a minor portion of the glomerular tuft were not included in the analysis of apoptotic cell counting. These investigations were performed using a video camera mounted on a Leitz microscope (Dialux 20; Leitz, Wetzlar, Germany) and a color monitor. Cell nuclei were counted manually to determine total cell number and TUNEL-positive cells per glomerular cross section (n = 50/animal), as described previously (Steinmann-Niggli et al., 1998). Electron microscopy for the detection of mesangial cell apoptosis was performed, as described previously (Baker et al., 1994).
Mesangial Cell Cultures.
Rat mesangial cells were isolated and propagated as described previously (Lovett et al., 1992;Steinmann-Niggli et al., 1997, 1998). Cell synchronization occurred in culture medium containing 1% FCS for 24 h, if required. Subsequently, culture medium containing 10% FCS was supplemented with either 5 to 100 μM RO111-3456 in 0.03% DMSO, 0.03% DMSO only, or 1 to 7 mM captopril (Trocme et al., 1998), and was added to cell cultures for periods of up to 24 h. Mesangial cell proliferation was assessed by manual cell counting.
Mesangial Cell Necrosis.
The effect of RO111-3456 and captopril on mesangial cell viability and necrosis was assessed by light microscopy, trypan blue exclusion, and release of lactate dehydrogenase (LDH) in cell culture supernatant (Ziswiler et al., 1998).
Gelatin Zymography.
SDS-polyacrylamide gel electrophoresis (PAGE) was performed on a 10-well, 10% polyacrylamide minigel containing 0.1% gelatin (w/v), as described by us in detail (Lovett et al., 1992; Steinmann-Niggli et al., 1997, 1998). For inhibition studies, the developing buffer was supplemented with 1, 50, 100, or 500 nM RO111-3456 in 0.003% DMSO, or with 0.003% DMSO only. In a separate series of experiments, the buffer was complemented by the addition of captopril in final concentrations of 0, 15, and 30 mM (Sorbi et al., 1993), respectively.
TUNEL- and Acridine Orange/Ethidium Bromide (AO/EB) Staining.
Mesangial cells were grown on glass coverslides to reach subconfluency. TUNEL assays of cells exposed to RO111-3456 were performed using an in situ cell detection kit (catalog number 1684817; Boehringer-Mannheim, Mannheim, Germany). For AO/EB staining of RO111-3456- and captopril-treated cells, 5 μl of freshly prepared AO/EB solution (100 μg/ml AO and 100 μg/ml EB in PBS) was added, and apoptosis was assessed immediately using an inverted fluorescence microscope (Baker et al., 1994; Amarante-Mendes et al., 1998).
Assessment of DNA Content by Flow Cytometry.
Cell cycle analyses and quantification of apoptosis in mesangial cells exposed to RO111-3456 and captopril were investigated by FACS analyses using propidium iodide (PI) staining of nuclear DNA as published (Nicoletti et al., 1991; Healy et al., 1998).
Suspensions of 2 × 106 synchronized mesangial cells were analyzed by FACS (Becton Dickinson, Franklin Lakes, NJ) with laser excitation at 488 nm using an emission 639-nm band pass filter to collect the red PI fluorescence. Forward and side light scatter were also recorded as indices of cell size and granularity, respectively. The percentages of cells in the various phases of the cell cycle, namely, sub-G0/G1, G0/G1, S, and G2/M, were assessed using the Modfit program (Becton Dickinson).
FACS Analyses of Annexin V Binding.
After exposure to RO111-3456, apoptotic mesangial cell death reflected by phosphatidylserine externalization was assessed by analyses of annexin V binding (annexin V-FITC kit, catalog number BMS306FI; Bender MedSystems, Boehringer Ingelheim Bioproducts, Heidelberg, Germany) (Amarante-Mendes et al., 1998). Cells were counterstained with PI to distinguish between apoptotic and late apoptotic/necrotic cells.
Western Blot Analyses of Statin, c-Myc, Fas/FasL, Caspase-3, p53, p21, and bax.
Subconfluent mesangial cells were exposed to 100 μM RO111-3456 for periods of 12 and 24 h, as indicated.
For analyses of Fas, FasL, and caspase-3, aliquots of 50 μg of whole cell extracts were resolved by 12% SDS-PAGE. In the cases of statin, c-myc, p53, p21, and bax samples of 100 μg of whole cell extracts were used for 10% SDS-PAGE (Wang and Pandey, 1995). The gels were blotted to a polyvinylidene fluoride microporous membrane (Immobilon-P; Millipore, Bedford, MA) for 1 h with 120 mA. Thereafter, the blots were blocked for 30 min in 10% (w/v) nonfat dry milk in TBS 0.1% (v/v) Tween 20 (TBS-T).
Following two washes in TBS-T, blots were incubated for 1 h at room temperature with 1:1000 polyclonal rabbit anti-Fas or anti-FasL IgG (M-20 and N-20; Santa Cruz Biotechnology Inc., Santa Cruz, CA), 1:500 rabbit polyclonal anti-caspase-3 IgG (H-277; Santa Cruz Biotechnology Inc.), 1:1000 monoclonal mouse anti-p53 IgG (Ab-2; Oncogene Science Diagnostics, Inc., Cambridge, MA), 1:1000 polyclonal rabbit anti-p21 IgG (H-164; Santa Cruz Biotechnology Inc.), or 1:1000 monoclonal mouse anti-bax IgG (6A7; PharMingen, Becton Dickinson, Bâle, Switzerland). Incubations for 14 h at 4°C were carried out for 1:500 monoclonal mouse anti-statin IgM (S-44; E. Wang, McGill University), 1:2000 monoclonal mouse anti-actin (Amersham Pharmacia Biotech UK Ltd., Little Chalfont, UK), and 1:2000 polyclonal rabbit anti-mouse anti-c-myc (Brunner et al., 2000) diluted in 2% nonfat dry milk/TBS-T.
Subsequently, blots were washed extensively in TBS-T and incubated with a goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology Inc.) for analyses of Fas, FasL, caspase-3, and p21 or with a goat anti-mouse horseradish peroxidase-conjugated secondary antibody (Bio-Rad Laboratories AG, Glattbrugg, Switzerland) for the remaining examinations, including actin as the internal standard. Finally, membranes were developed using enhanced chemiluminescence (Amersham Pharmacia Biotech UK Ltd.).
For in vivo analyses of p53, glomeruli were harvested from healthy, nephritic, and BB-1101-treated nephritic rats at days +8 and +11. Aliquots containing 40 μg of protein from nuclear extracts (Kobet et al., 2000) of the respective glomeruli were subjected to Western blot analyses for p53 as described above. Scanning densitometry of the respective Western blots was performed as described previously (Steinmann-Niggli et al., 1997).
Assessment of Caspase-8 Activity.
Caspase-8 activity of RO111-3456-exposed mesangial cells was determined using a fluorometric protease assay kit based on the detection of cleavage of the substrate IETD-AFC (Caspase-8/Flice Fluorometric Assay kit; BioVision, Palo Alto, CA). Thereafter, fluorescence was analyzed in a spectrofluorometer (SPECTRAmax GEMINI; Molecular Devices, Sunnyvale, CA) using a 400-nm excitation filter and a 505-nm emission filter.
Inhibition of Caspase Activity by Acetyl-Asp-Glu-Val-Asp-Aldehyde (DEVD-CHO).
The caspase inhibitory peptide DEVD-CHO (Bachem, Bubendorf, Switzerland) was dissolved in water and used at concentrations of 0 to 200 μM, as reported previously (Ortiz et al., 1998). DEVD-CHO was added to cell culture medium together with 100 μM RO111-3456 or 0.03% DMSO only. Apoptosis was assessed 24 h later by FACS analyses using annexin V-FITC/PI staining, as described above. Positive control consisted of 50 ng/ml anti-Fas antibody (CH-11; Upstate Biotechnology, Lake Placid, NY)-treated Jurkat cells, as described (Zhou et al., 1999).
Immunocytology for Statin.
Cells were harvested after exposure to 100 μM RO111-3456 or to 0.03% DMSO for 12 h. Thereafter, staining of cytocentrifuge cell samples was performed as described (Hirt et al., 1997).
Statistical Analyses.
In vitro results were expressed as mean ± S.D., and comparison of means was done by the Student'st test. In vivo data are given as median (percentile P25 to percentile P75); analyses were performed by ANOVA (with the Bonferroni adjustment). For all experiments, probability of error (P values) <0.05 was regarded to be significant.
Results
Apoptosis of Mesangial Cells in Vivo
Increased Rates of Apoptosis in Experimental Glomerulonephritis.
Kidney sections of healthy, healthy pretreated, nephritic, and pretreated nephritic rats were used for histological analyses. Therapy consisted of daily applications of BB-1101 resulting in blood levels of approximately 100 nM, equivalent to multiples of MMP IC50 concentrations, as reported previously (Steinmann-Niggli et al., 1998). We selected this compound, since we have used it successfully in anti-Thy1.1 nephritis in the past (Steinmann-Niggli et al., 1998). Furthermore, there were no guidelines or published data available on the in vivo use of RO111-3456.
Apoptotic glomerular cells were identified by TUNEL staining, as depicted in Fig. 1A. Analyses were performed at days +4, +8, +11, and +14 after induction of nephritis by counting TUNEL-positive apoptotic cells in randomly selected glomerular cross sections. Results are shown in Fig. 1B. At all analyzed time points, there was an increased number of apoptotic cells in pretreated compared with untreated nephritic animals with a peak at day +11. The numbers of apoptotic cells per 50 glomerular cross sections expressed as median and percentile (P25/P75) were as follows: 0.2 (0/0.5) and 0.5 (0/1) in healthy rats, 2.5 (1.7/3.3) and 1 (1/1.4) in nephritic rats, and 8.4 (4.2/15.0) and 6.4 (5.3/6.8) in pretreated nephritic rats at days +11 and +14, respectively. The values of increased apoptosis in pretreated nephritic versus untreated nephritic animals were highly significant at both time points (P< 0.05). Healthy rats were not affected by BB-1101 therapy (data not shown).
As published previously, the MMP inhibitor treatment led to a significant decrease in glomerular cellularity and ECM deposition (Steinmann-Niggli et al., 1998). Therefore, the increase in apoptotic cells expressed in relation to total glomerular cell content was even more pronounced. In this respect, treated nephritic animals showed 5.6 and 3.8% apoptotic glomerular cells on days +11 and +14, respectively. Untreated nephritic animals displayed over 4.5 and 6 times less apoptotic cells, corresponding to values of only 1.2 and 0.6% on these time points. Importantly, TUNEL-positive apoptotic cells were predominantly located in the mesangium and apoptosis of tubulo-interstitial cells remained negligible (data not shown).
Apoptotic cells in both groups of nephritic animals were further characterized. The predominant mesangial cell apoptosis was confirmed by TUNEL and α-smooth muscle actin double-staining and by electron microscopy. Figure 2 demonstrates apoptotic mesangial cells in BB-1101-treated nephritic rats (Fig. 2A) and in nephritic untreated animals (Fig. 2B).
Therefore, we decided to study the mechanism of the proapoptotic effect of MMP inhibitor on the cellular level using cultured mesangial cells. However, we were unable to use BB-1101 for these in vitro analyses, as mentioned above. We chose to use the MMP inhibitor RO111-3456, the later version and successor of Ro 31-9730. The latter compound, in identical concentrations, was already used by us for mesangial cells in the past (Steinmann-Niggli et al., 1997).
Cell Cycle Arrest and Apoptosis of Mesangial Cells in Vitro
Inhibition of MMP Activity.
Conditioned medium of rat mesangial cells contained gelatinolytic activity exclusively derived from MMP-2. The zones of lysis in the zymogram were inhibited in a dose-dependent manner by exposure to RO111-3456, as demonstrated in Fig. 3A. Gelatinolytic activity in the gel strips clearly decreased following exposure to as little as 1 nM inhibitor and it almost completely disappeared after incubation with 500 nM RO111-3456. Similar inhibitor concentrations were used by us previously (Steinmann- Niggli et al., 1997). In addition, captopril lead to a dose-dependent and complete inhibition of MMP activity (Fig.3B). All analyses were performed in duplicates with two identically treated samples on each gel strip.
Inhibition of Mesangial Cell Proliferation.
Subconfluent mesangial cells were exposed to RO111-3456 in concentrations of 0 to 100 μM for 24 h. The addition of this MMP inhibitor to the culture led to a dose-dependent decrease in cell proliferation, as shown in Fig. 3C. An inhibition of 50% of cell proliferation occurred at a concentration of approximately 40 μM inhibitor and 61% inhibition at the maximal concentration of 100 μM.
Exclusion of Mesangial Cell Necrosis.
Compared with untreated control cells, viability of mesangial cells exposed for 24 h to either 100 μM RO111-3456, 0.03% DMSO, or 7 mM captopril was not impaired, as analyzed by light microscopy and trypan blue exclusion (viability in all experiments >95%). Furthermore, LDH levels in mesangial cell culture supernatant expressed as percentage of total LDH release remained at constant levels: 6.1 ± 0.4% in control cells and 6.4 ± 0.4% in cells exposed to 100 μM RO111-3456 in 0.03% DMSO. LDH release was also not influenced by 0.03% DMSO alone, and by 7 mM captopril (data not shown). Therefore, both MMP-inhibiting agents caused no obvious signs of cell necrosis.
Cell Cycle Progression.
For FACS analyses of nuclear DNA content by PI staining, subconfluent and synchronized mesangial cells were exposed to medium supplemented with 100 μM RO111-3456 or 0.03% DMSO only for time periods of 6, 12, and 24 h. Synchronization was realized by a previous 24-h incubation with culture medium containing 1% FCS. Subsequently, mesangial cells were stimulated with medium containing 10% FCS and the respective amount of inhibitors for further 24 h. All experiments were performed in triplicates.
Mesangial cells treated with 100 μM RO111-3456 remained to a high extent, about 80%, in G0/G1 phase and showed only limited cell cycle progression, equivalent to a marked decrease in proliferation, as depicted in Fig. 4, A and B. In contrast, uninhibited control cells progressed to S phase, reflecting increased DNA synthesis and cell proliferation.
Analyses of Statin and c-Myc.
The effects of 100 μM RO111-3456 on mesangial cell cycle arrest in the G0/G1 phase were confirmed by the analyses of statin, a 57-kDa protein expressed by truly quiescent cells in the G0 phase (Turck et al., 1996). MMP inhibitor-treated cells displayed a distinctively higher level of statin after an incubation period of 12 h, as shown by Western blot analyses (Fig. 5, A and B) and immunocytology (Fig. 5C). Correspondingly, the synthesis of c-myc, a G1 phase-specific gene, somewhat decreased as a result of the inhibitor exposure (Fig. 5, A and B) (Wang and Pandey, 1995; Brunner et al., 2000). Therefore, the expected inverse regulation of statin and c-myc expression was observed due to cell cycle arrest before the initiation of apoptosis (Wang and Pandey, 1995).
TUNEL and AO/EB Staining.
TUNEL-positive, apoptotic mesangial cells were not present in the untreated control group (Fig.6A). Exposure of cells to even 50 μM RO111-3456 for 24 h resulted in a distinct increase of TUNEL-positive cells, as shown in Fig. 6B.
Chromatin condensation, which may occur independent of DNA fragmentation is another characteristic feature of ongoing apoptosis (Amarante-Mendes et al., 1998). This event of apoptosis was investigated by AO/EB staining of mesangial cells treated with RO111-3456 for 24 h. A dose-dependent increase in apoptotic cells was obtained. Untreated and 0.03% DMSO only-treated control cells showed no signs of apoptosis (Fig. 7A). RO111-3456 in a concentration of 100 μM showed mostly late apoptotic cells with condensed nuclei; hardly any necrotic cells were detectable (Fig. 7B). Concentrations of 50 μM RO111-3456 showed overall fewer apoptotic cells; there were less late apoptotic but more early apoptotic cells visible. RO111-3456 in a concentration of 15 μM showed almost only viable cells with very few early apoptotic cells (data not shown).
Captopril treatment of mesangial cells also revealed some dose-dependent signs of apoptosis, but distinctively less than RO111-3456. The maximal concentration of 7 mM showed mostly early signs of apoptosis, as shown by bright green dots in the nuclei as a consequence of chromatin condensation and nuclear fragmentation (Fig.7C). Lower levels of 1 and 3.5 mM demonstrated only few signs of apoptosis (data not shown). Necrotic cells remained negligible in all cases.
Annexin V Binding.
An early event during apoptosis is the externalization of phosphatidylserine, a phospholipid normally restricted to the inner leaflet of the plasma membrane (Healy et al., 1998). This apoptotic event can be monitored using annexin V, a phosphatidylserine-specific binding protein.
To quantify early and late events in the course of apoptosis, mesangial cells were stained with annexin V-FITC and PI, as depicted in Fig.8. Living cells are double negative; apoptotic cells first display annexin V binding only but at a later stage acquire double positivity for annexin V and PI as a result of late apoptosis or even necrosis. Treatment of mesangial cells with 100 μM Ro 111-3456 for 24 h induced a significant increase in annexin V-positive/PI-negative cells from 3 to 22%, consistent with induction of early apoptotic cell death. To a lesser extent, elevated numbers of annexin V/PI double-positive cells were also observed, possibly reflecting late apoptotic cells, although induction of some necrosis in this population could not be totally excluded (Fig. 8, A and B).
Dose-Dependent Induction of Apoptosis.
As an estimate of apoptosis, we calculated the percentage of cells in the “sub-G0/G1 phase” (Erba et al., 1999). This sub-G0/G1 peak represents a population of cells with reduced DNA content due to DNA fragmentation (Healy et al., 1998). To avoid potential cytotoxicity, we did not use inhibitor concentrations higher than 100 μM and incubation times longer than 24 h.
Figure 9A depicts representative flow cytometric DNA histograms where the amount of fluorescence emitted is directly proportional to the amount of DNA present in the cells. In contrast to control cells, RO111-3456 treatment at the highest dose of 100 μM led to the appearance of a distinct sub-G0/G1 peak in the DNA histogram. Figure 9B demonstrates the distinct dose dependence of RO111-3456 exposure and subsequent apoptosis with a maximum of 23.4 ± 4.5% apoptotic cells.
Cell Cycle Arrest and Apoptosis over Time.
RO111-3456 treatment in the maximal concentration of 100 μM for 6, 12, and 24 h in medium containing 10% FCS caused time-dependent increases in the percentage of the cells in the G0/G1 phase that was paralleled by decreases of cells in the G2/M (mitotic/dividing) phases, as depicted in Fig.10 and Table1. These findings explain the dose-dependent antiproliferative effect of RO111-3456 in mesangial cells. Already after a 6-h exposure of MMP inhibitor, nonsynchronized mesangial cells started to accumulate in G0/G1 phase, representing cell cycle arrest. DNA fragmentation after 6 and 12 h remained minimal and only reached significant values after 24 h. Therefore, it can be concluded that cell cycle arrest occurs first and the induction of apoptosis is a subsequent event.
Captopril given for 24 h in medium containing 10% FCS caused a dose-dependent cell cycle arrest, as reflected by an accumulation in the G0/G1 phase that reached values almost identical to RO111-3456 (Table2). However, as shown above, apoptosis remained low compared with the MMP inhibitor.
Fas/FasL and Caspases-3/-8.
Alterations in Fas and FasL expression as a result of 100 μM RO111-3456 treatment were investigated following incubation periods of 12 and 24 h. At both time points, Western blot analysis failed to show an increase in the expression of 45-kDa Fas protein, 37-kDa FasL protein, and caspase-3 in the MMP inhibitor-exposed cells compared with untreated controls (data not shown). Furthermore, there was only a slight tendency toward an approximately 1.5-fold increase in caspase-8 activity in MMP inhibitor-treated cells at 24 h, as analyzed by fluorometric measurements (data not shown).
Inhibition of Caspases by DEVD-CHO.
The peptide DEVD-CHO inhibits caspase-3 and related caspases (Ortiz et al., 1998). Apoptosis induced by RO111-3456 was attempted to be inhibited by this peptide. Whereas anti-Fas-induced apoptosis in Jurkat cells was dose dependently inhibited by DEVD-CHO as expected, DEVD-CHO in concentrations of 0 to 200 μM failed to attenuate 100 μM RO111-3456-induced apoptosis in mesangial cells. Results are summarized in Table3. Therefore, we had no evidence that the Fas/FasL and probably also the TNF-α/TNF-αR pathways played a major role in the antiproliferative effect of the MMP inhibitor.
Analyses of p53, p21, and bax.
Mesangial cells were exposed to 100 μM RO111-3456 for 12 h. At this time point distinct cell cycle arrest of mesangial cells was observed.
Subsequently, synthesis of p53, p21, and bax was investigated by Western blot analyses of mesangial cell extracts. In accordance with cell cycle arrest and subsequent apoptosis, RO111-3456 caused a distinctive up-regulation of p53, bax, and p21. Results are depicted in Fig. 11A.
The analysis of p53 was also performed in vivo. Compared with nephritic rats, the BB-1101 therapy caused an approximately 3-fold increase in glomerular expression of p53 at day +8 (Fig. 11, B and C). At day +11, results were less impressive, but still a clear tendency toward an inhibitor-induced increase in p53 expression was obtained (data not shown).
Discussion
The present study describes the profound effects of synthetic MMP inhibitors on cell cycle arrest and apoptosis in glomerular mesangial cells. MMPs are zinc-dependent metalloendopeptidases belonging to the collagenase supergene family. Primarily based on substrate specificity, they are classified into several groups, such as interstitial collagenases (MMP-1, -8, -13), gelatinases (MMP-2, -9), stromelysins (MMP-3, -7, -10, -11), and membrane type-MMPs (membrane type-MMP-1, -2, -3, -4) (Cuvelier et al., 1997). MMPs are secreted in a latent form as pro-MMP and extracellular activation occurs in a complex manner by conformational change and by proteolytic action of proteinases, such as plasmin and membrane-type matrix metalloproteinase (Davies et al., 1992). MMP activity is regulated by natural inhibitors, mainly, the tissue inhibitors of metalloproteinases (tissue inhibitor of metalloproteinase-1 to -4) (Nagase, 1997). The regulation of ECM metabolism is probably the main function of these proteolytic enzymes and their inhibitors. Recently, a broad spectrum of very effective, low-molecular-weight MMP inhibitors was developed by the addition of a zinc-chelator, mostly a hydroxamate, to a peptidyl moiety based on MMP substrates (Vincenti et al., 1994). Both compounds used in our studies belong to this category. Notably, certain synthetic low-molecular-weight MMP inhibitors also inhibit related metalloproteinases that process membrane-bound cytokines and growth factors, such as TNF-α (McGeehan et al., 1994).
Synthetic MMP inhibitors already demonstrated significant benefits in various animal models for inflammatory diseases, such as arthritis (Conway et al., 1995), experimental allergic encephalomyelitis (Hewson et al., 1995), asthma (Kumagai et al., 1999), and glomerulonephritis, as already mentioned (Steinmann-Niggli et al., 1998). However, the precise mechanism on the cellular level beyond MMP inhibition remains to be fully determined.
Unopposed mesangial cell proliferation associated with accumulation of ECM proteins is an important cause of glomerular scarring with loss of kidney function. However, if mesangial proliferation resolves in a timely manner, glomerular structure and function may revert back to normal (Baker et al., 1994). We speculated that resolution of excess mesangial cell proliferation may be facilitated by the induction of cell cycle arrest with subsequent apoptosis. Apoptosis plays a central role in maintaining homeostasis in tissues and organs such as the kidney by the deletion of “unwanted” cells without induction of an inflammatory reaction (Baker et al., 1994). Increased apoptotic cells were detected in glomeruli of humans with IgA nephropathy (Tashiro et al., 1998). In animals, apoptosis was shown to play a prominent role in the resolution of anti-Thy1.1 nephritis, a rat model of mesangial proliferative glomerulonephritis (Baker et al., 1994). In this disease, anti-Thy1.1 antibodies induce a self-limited mesangial cell proliferation associated with marked increases in cyclin A and cyclin-dependent kinase 2 (Shankland et al., 1996). Apoptosis occurred approximately 10-fold more frequently in glomeruli of nephritic rats compared with healthy controls with clear morphological evidence of mesangial apoptosis leading to phagocytosis by neighboring mesangial cells (Baker et al., 1994).
Since MMP inhibitors can induce cell cycle arrest or apoptosis in malignant cells (Burke et al., 1997; Erba et al., 1999), we examined the effect of BB-1101 in anti-Thy1.1 nephritis. The process of mesangial cell clearance by apoptosis was greatly enhanced by MMP inhibitor treatment. TUNEL staining was used to discover glomerular cells with fragmented DNA, a hallmark of apoptosis. The therapy with BB-1101 led to a 4.5- to 6-fold increase in TUNEL-positive apoptotic glomerular cells 11 and 14 days after induction of nephritis, respectively. In anti-Thy1.1 nephritis, the overwhelming majority of proliferating cells that may be arrested in their cell cycle by any type of intervention are mesangial cells (Steinmann-Niggli et al., 1998). Accordingly, in our study apoptotic cells were almost exclusively limited to the mesangium, as described (Baker et al., 1994). Importantly, mesangial cell apoptosis was confirmed by electron microscopy as well as by TUNEL and α-smooth muscle actin double-staining.
Apoptosis remained negligible in the tubulo-interstitium and there were no signs of a concomitant inflammatory reaction as a result of an unlikely necrosis. Furthermore, the amounts of apoptotic cells, in relation to absolute numbers of glomerular cells, followed the time course of the proliferation rates of mesangial cells and were lower at day +14 than at day +11, despite the continuous application of the MMP inhibitor. In the case of nonspecific toxicity, one might have expected rather a more relentless rise in apoptosis.
Therefore, we selected the mesangial cells as a model to study the MMP inhibitor effect on the cellular level. RO111-3456 caused significant cell cycle arrest in G0/G1phase in cultured mesangial cells associated with increased expression of statin. Cell cycle arrest was followed by apoptosis. Increased induction of apoptosis was demonstrated and confirmed by various morphological and biochemical tests. Importantly, the induction of statin may well exclude a nonspecific, toxic effect of the MMP inhibitor treatment.
The MMP inhibitor concentrations used in vitro were much higher than its blood levels achieved in vivo. However, the application of BB-1101 in vivo occurred over a relatively prolonged period and achieved rates of apoptotic cells in the low range of pars pro mille. In contrast to these experiments, RO111-3456 was given in vitro over a very short time period and attained rates of apoptotic cells in the order of many percentages. The high amount of apoptosis in cultured mesangial cells facilitated its detection and confirmation by the various tests used.
Elimination of apoptotic cells is a rapid process with a clearance time of only a few hours (Baker et al., 1994). Therefore, even minute changes in absolute amounts of apoptotic cells over time may have significant effects on the cell content of a given tissue (Baker et al., 1994). In this respect, for future clinical studies, even lower amounts of MMP inhibitors may prove to be successful when given over a longer time period. To induce apoptosis in mesangial cells, MMP inhibitors may be particularly useful since these cells are believed to be very resistant to this type of cell death (Baker et al., 1994). Therefore, MMP inhibitors may be especially useful for the therapy of mesangial cell-mediated kidney inflammation.
To the best of our knowledge, this is the first report showing synthetic MMP inhibitor-induced cell cycle arrest followed by apoptosis in nontumor cells. Although MMP activity is inhibited by these compounds, it remains to be demonstrated to which extent precisely this effect is functionally linked to proapoptotic actions in mesangial cells.
Cell cycle arrest of mesangial cells as a result of RO111-3456 treatment is well explained by induction of p53 and p21 (Shaw, 1996;Amundson et al., 1998). Since captopril, a structurally very different agent inhibiting MMP activity, also influenced cell cycle of these cells to a similar extent, it is possible, although far from being proven, that inhibition of MMP in fact played a causative role. Future studies are needed to resolve this issue.
The subsequent apoptosis of mesangial cells is more complex, although well explained by the up-regulation of p53, also shown in vivo, and of bax (Shaw, 1996; Amundson et al., 1998). Furthermore, the observed down-regulation of c-myc, an important proliferation signal, probably is also the result of p53 induction (Amundson et al., 1998).
Hydroxamic acid-based MMP inhibitors, such as the compounds used in our study, were shown to inhibit the metalloproteinase responsible for processing of transmembrane FasL (Tanaka et al., 1998). Therefore, it might have been conceivable that RO111-3456 facilitated apoptosis by stabilization of the transmembrane form of FasL and hence prevented the down-regulation of FasL by its shedding (Tanaka et al., 1998). However, a significant participation of the Fas/FasL pathway, including the downstream caspases, was not detectable. Therefore, MMP inhibitor-related apoptosis appeared to follow a caspase-independent pathway (Brown et al., 2000). The investigation of the gene products responsible for p53-mediated apoptosis beyond bax is very extensive and involves a separate project.
The role of MMP and their inhibitors in the regulation of cell cycle and apoptosis is still emerging. MMPs themselves are involved in the regulation of cell proliferation (Turck et al., 1996; Steinmann-Niggli et al., 1997, 1998) and to a certain extent apoptosis (Vu et al., 1998). The influence of synthetic MMP inhibitors on cell cycle and apoptosis is well documented, especially for tumor cells. The MMP inhibitor batimastat (BB-94) was shown to enhance interferon-γ-induced apoptosis in mice with ovarian cancer (Burke et al., 1997) and to block ovarian cancer cells in the G0/G1 phase of the cell cycle unrelated to p53 expression (Erba et al., 1999). AG3340 promoted apoptosis in human prostate and colon carcinoma models (Shalinski et al., 1999). Furthermore, GM-6001 was shown to induce apoptosis in smooth muscle cells (Cowan et al., 2000).
In summary, MMP inhibitors induce cell cycle arrest with subsequent apoptosis in mesangial cells in vitro and in vivo. These features may explain the beneficial anti-inflammatory features, such as reduction of excess mesangial cell proliferation and of ECM accumulation. Therefore, it may be concluded that the concept of induction of cell cycle arrest with subsequent apoptosis may contribute to the development of new perspectives in the therapy of inflammation even beyond the scope of kidney diseases.
Acknowledgments
Professor B. Frey from our Division of Nephrology was of great help with the correction of the manuscript. We are also grateful to A. Kappeler (University of Bern) and in particular to G. Thomas (University of Edinburgh) for stainings of kidney sections and for the electron microscopy. We are also thankful to J. Caulfield (Roche Bioscience) and to A. Galloway (British Biotech Pharmaceuticals) for providing RO111-3456 and BB-1101, respectively, as gifts. Furthermore, we thank A. Marti from the Department of Clinical Research of the University of Bern for providing anti-caspase-3 antibody as a gift.
Footnotes
-
Send reprint requests to: Hans-Peter Marti, M.D., Division of Nephrology and Hypertension, Inselspital Bern, CH-3010 Bern, Switzerland. E-mail: hmarti{at}insel.ch
-
This work was supported by Grants 31-49765.96 and 31-55779.98 to H.-P.M. from the Swiss National Foundation for Scientific Research. Portions of the study were presented as poster at the 32nd Annual Meeting of the American Society of Nephrology, November 5–8, 1999, Miami Beach, FL. Daniel C, Duffield J, Thomas G, Ziswiler RA, Steinmann-Niggli K and Marti HP (1999) Matrix metalloproteinase inhibitor induces apoptosis in mesangial cells. J Am Soc Nephrol10:569A.
- Abbreviations:
- MMP
- matrix metalloproteinase
- ECM
- extracellular matrix
- PBS
- phosphate-buffered saline
- DMSO
- dimethyl sulfoxide
- TUNEL
- terminal deoxynucleotidyl transferase (TdT)-mediated deoxyuridine triphosphate (dUTP) biotin nick-end labeling
- FCS
- fetal calf serum
- LDH
- lactate dehydrogenase
- PAGE
- polyacrylamide gel electrophoresis
- AO/EB
- acridine orange/ethidium bromide
- FACS
- fluorescence-activated cell sorter
- PI
- propidium iodide
- TBS-T
- Tris-buffered saline-Tween 20
- DEVD-CHO
- acetyl-Asp-Glu-Val-Asp-aldehyde
- FITC
- fluorescein isothiocyanate
- TNF-α
- tumor necrosis factor-α
- IETD-AFC
- N-acetyl-Ile-Glu-Thr-Asp-7-amino-4-trifluoromethyl coumarin
- Received September 26, 2000.
- Accepted December 19, 2000.
- The American Society for Pharmacology and Experimental Therapeutics