High Glucose-Induced Hypertrophy of Mesangial Cells Requires p27^Kip1, an Inhibitor of Cyclin-Dependent Kinases

Abstract

Hypertrophy of mesangial cells is one of the earliest morphological alterations in the kidney after the onset of diabetes mellitus. We have previously shown that cultured mesangial cells exposed to high ambient glucose arrest in the G₁ phase of the cell cycle and that this is associated with an increased expression of inhibitors of the cyclin-dependent kinase (CDK)-inhibitors p21^Cip and p27^Kip1. To further investigate a potential role of p27^Kip1 in the development of glucose-induced hypertrophy, mesangial cells from p27^Kip1 wild-type (+/+) and knockout (−/−) mice were established. High glucose medium (450 mg/dl) increased p21^Cip1 protein in p27^Kip1+/+ and −/− mesangial cells, and increased p27^Kip1 protein levels in p27^Kip1+/+ cells. In contrast to high glucose increasing de novo protein synthesis in p27^Kip1+/+ cells, high glucose did not increase protein synthesis in p27^Kip1−/− cells. High glucose also reduced DNA synthesis and caused cell cycle arrest in p27^Kip1+/+ cells. In contrast, despite an increase in transforming growth factor (TGF)-β mRNA and protein expression, DNA synthesis and cell cycle progression were increased by high glucose in p27^Kip1−/− cells. Exogenous TGF-β comparably induced fibronectin mRNA in p27^Kip1+/+ and −/− cells suggesting intact TGF-β receptor transduction. In addition, high glucose failed to increase the total protein/cell number ratio in p27^Kip1−/− cells. However, in the presence of high glucose, reconstituting p27^Kip1 expression by transient or stable transfection in p27^Kip1−/− cells, using an inducible expression system, increased the de novo protein synthesis and restored G₁-phase arrest. These results show that p27^Kip1 is required for glucose-induced mesangial cell hypertrophy and cell cycle arrest.

Mesangial expansion, one of the earliest renal abnormalities observed after the onset of hyperglycemia in diabetes mellitus, is because of growth of mesangial cells, as well as an increase in glomerular extracellular matrix accumulation. 1-4 In vivo studies in different models of type I and II diabetes mellitus, and cell culture studies using mesangial cells exposed to high glucose, have shown a characterized biphasic growth response. First, there is an early and self-limited proliferation of mesangial cells, which is followed by cell cycle arrest in the G₁ phase of the cell cycle. This is followed by glomerular cell persistent and progressive hypertrophy. 5-7

Cell proliferation is regulated at the level of the cell cycle by cell cycle proteins, where activation of cyclin-dependent kinases (CDK) is required for progression through the cell cycle. In contrast, CDK inhibitors inactivate CDKs, and cause cell cycle arrest. There is a growing body of literature showing that CDK inhibitors may also be critical regulators of cell hypertrophy. 8-10 The present study was undertaken to determine the role of the CDK inhibitor p27^Kip1 in mediating glucose-induced mesangial cell hypertrophy. We show that in contrast to p27^Kip1+/+ mesangial cells, p27^Kip1−/− mesangial cells do not undergo glucose-induced hypertrophy. However, reconstituting p27 levels is necessary to induce hypertrophy in p27^Kip1−/− cells. Our results provide evidence for a role of p27^Kip1 in high glucose-induced hypertrophy of cultured mesangial cells.

Materials and Methods

Cell Culture

Mesangial cells from litter mate p27^Kip1+/+ and p27^Kip1−/− mice were isolated by differential sieving and characterized as previously described. 11 Cells were grown in Dulbecco’s modified Eagle’s medium (Gibco-BRL, Eggenstein, Germany) containing 100 mg/dl d-glucose (G100) supplemented with 10% fetal calf serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mmol/L glutamine. Mesangial cells were cultured at 37°C in 5% CO₂, and passaged every 4 to 5 days. Experiments were done using cells of passages 10 to 20.

Inducible p27^Kip1 Expression Construct and Transfections

A full-length mouse p27^Kip1 cDNA was constructed using reverse transcriptase-polymerase chain reaction (RT-PCR) techniques. Briefly, total RNA was isolated from murine mesangial cells 12 rested in serum-free medium. Ten μg of total RNA was reverse-transcribed using 0.7 μg of poly-d(T)primer (Pharmacia Diagnostics, Freiburg, Germany) in the presence of 500 U of Maloney murine leukemia virus reverse transcriptase diluted in 50 μl of a buffer containing 50 mmol/L Tris-HCl (pH 8.3), 75 mmol/L KCl, 3 mmol/L MgCl₂, 10 mmol/L dithiothreitol, and 500 μmol/L dNTP. After incubation for 2 hours at 37°C, 5 μl of the cDNA preparation was directly used for the PCR amplification with 5 μl of 10× amplification buffer, 25 mmol/L MgCl₂, 10 mmol/L dNTPs, and 1.5 μl of each primer (50 ng/μl), and 2.5 U Taq-polymerase (Promega, Madison, WI). The following primers specific for the murine p27^Kip1 were used: 5′GGGCCACCATGCAAACGTGAGAGTG3′, 5′GCTGTTTACGTCTGGCGTCGAAGG3′. 13 The 5′end primer contained an optimized Kozak sequence for efficient mRNA translation initiation. 14 A total of 40 amplification cycles (denaturing for 30 seconds at 94°C, annealing for 90 seconds at 50°C, and extension for 90 seconds at 72°C) were performed. Fifteen μl of the reaction product were separated in a 1.8% agarose gel containing 0.5 μg/ml ethidium bromide and a single band of the predicted 607 bp was isolated using a DEAE membrane.

The full-length p27^Kip1 cDNA was then cloned into the pIND TOPO vector (Invitrogen, Leek, The Netherlands) and the correct orientation was confirmed by sequencing. The expression plasmid pIND contains five modified ecdysone response elements upstream of a minimal heat-shock promoter allowing inducible expression. 15 For transient transfections, 10 5 cells were rested in serum-free medium for 24 hours and were co-transfected with 10 μg of pINDp27^Kip1 and 10 μg of the plasmid pVgRXR (Invitrogen) encoding subunits of the ecdysone receptor using lipofectin (Gibco-BRL) as previously described. 8 After 12 hours, cells were split into 24- or 96-well plates (for measurement of leucine and thymidine incorporation) or left in cell culture flasks (for Western blots). Cells were then treated with 0 to 5 μg/ml muristone (Invitrogen) to induce p27^Kip1 expression.

To establish stabile cell lines, p27^Kip1−/− mesangial cells were first transfected with pVgRXR and stable clones were selected in medium with 500 μg/ml Zeocin (Invitrogen) by several rounds of limited dilution. A stable cell line expressing pVgRXR was subsequently transfected with pINDp27^Kip1 and selection was then performed in medium with 500 μg/ml of Zeocin and 250 μg/ml of G418 (Sigma). Surviving cells were cloned by three rounds of limited dilution and several stable cell lines containing pVgRXR and pINDp27^Kip1 were established. These cell lines were carried normally in medium containing Zeocin and G418 to prevent back mutations.

Measuring the Expression of CDK Inhibitors

A total of 10 6 cells from the various cell lines and clones were incubated in serum-free G100 or G450 (450 mg/dl d-glucose) medium for 48 hours. To restore p27^Kip1 expression, p27^Kip1−/− cells were transient-transfected with pINDp27^Kip1 or permanent-transfected clones (clones 5.2, 6.2) were treated with 0 to 5 μg/ml of muristone for 24 hours. Cells were rinsed twice in ice-cold phosphate-buffered saline (PBS) at the end of the incubation period. After removing all PBS, monolayers were directly lysed in 150 μl of lysis buffer (2% sodium dodecyl sulfate, 60 mmol/L Tris-HCl, pH 6.8), and the protein content was measured in supernatants after centrifugation by a modification of the Lowry method that is insensitive to the used concentrations of sodium dodecyl sulfate. 8 Protein concentrations were adjusted to 80 μg/sample, and 100 mmol/L dithiothreitol, 5% glycerol, and 0.03% bromophenol blue were added and samples were boiled for 5 minutes. After centrifugation, supernatants were loaded onto a 12% sodium dodecyl sulfate-polyacrylamide gel. Low molecular rainbow markers (comprising 2,350 to 45,000 Daltons; Amersham, Braunschweig, Germany) served as the molecular weight standards.

After completion of electrophoresis, proteins were electroblotted onto a nitrocellulose membrane (Highbond-N, Amersham) in transfer buffer (50 mmol/L Tris-HCl, pH 7.0; 380 mmol/L glycine, 20% methanol). Filters were stained with Ponceau S to control for equal loading and transfer. Membranes were blocked for 1 hour at room temperature with 5% nonfat dry milk redissolved in PBS with 0.1% Tween 20. For the detection of p27^Kip1 protein, a 1:1,000 dilution of a mouse monoclonal anti-p27^Kip1 antibody (Transduction Laboratories, Lexington, MA) was used. This antibody reacts with murine p27^Kip1. p21^Cip1 protein expression was detected with a mouse monoclonal anti-human p21^Cip1 antibody exhibiting cross-reactivity with the murine protein at a 1:500 dilution (DAKO, Glostrup, Denmark). After incubation for another hour, membranes were washed in PBS with 0.1% Tween for 3 × 10 minutes, and a horseradish peroxidase-conjugated rabbit polyclonal anti-mouse antibody (Transduction Laboratories) was added at a 1:1,000 dilution. The luminescence detection of peroxidase activity was performed with the ECL system (Amersham) according to the manufacturer’s recommendations. To control for small variations in protein loading and transfer, membranes were washed and re-incubated with a mouse monoclonal anti-β-actin antibody (Sigma). Incubation with secondary antibody and detection was performed as described above. Exposed films were scanned with Fluor-S multi-imager (Bio-Rad Laboratories, Hercules, CA), and data were analyzed with the computer program MultiAnalyst from Bio-Rad. Western blots were independently performed three to four times with qualitatively similar results.

Northern Blot Hybridization for TGF-β and Fibronectin

p27^Kip1−/− or +/+ cells (10 7 cells) were made quiescent in serum-free G100 medium and were stimulated for 48 hours in either serum-free G100 or G450 medium. Some cells were also treated in G100 with 1 ng/ml of ultrapure human transforming growth factor (TGF)-β1 (Sigma). After washing in RNase-free PBS, cells were directly lysed with acid guanidinium thiocyanate, and total RNA was isolated. 5 Equal amounts of total RNA (25 μg per lane) were denatured in formamide-formaldehyde at 65°C and electrophoresed through a 1.2% agarose gel containing 2.2 mol/L formaldehyde. Blotting, hybridization, and washing conditions were exactly as previously described. 5 A 0.7-kb PvuII cDNA fragment encoding human fibronectin was used. For control hybridizations, a 2.0-kb cDNA insert of the plasmid pMCI encoding the murine ribosomal 18S band was used. Northern blots were repeated twice with qualitatively similar result.

Determination of TGF-β Protein in Culture Supernatants

To investigate whether p27^Kip1−/− and +/+ mouse mesangial cells increased the synthesis of TGF-β, 2 × 10 6 cells were plated in small culture flasks. After incubation for 12 hours in serum-free medium with normal glucose, cells were incubated for another 48 hours in either normal or high glucose medium. The supernatant was harvested and dried using a speed vac. Cell layers were lysed in 0.5 mol/L of NaOH and protein was determined with a modification of the Lowry method. TGF-β1 protein measurements were performed with a commercially available enzyme-linked immunosorbent assay (Predicta, Genzyme, Cambridge, MA). In brief, dried supernatants were reconstituted in 200 μl of sample diluent, activated by addition of 20 μl of 1 mol/L HCl, and neutralized by 15 μl of 1 mol/L NaOH. Measurements of TGF-β1 were done according to the manufacturer’s recommendations. Concentrations were calculated as pg TGF-β1 per μg protein. TGF-β1 measurements were independently repeated five times with qualitatively similar results.

Measuring Protein and DNA Synthesis

The incorporation of [³H]leucine was used to assess de novo protein synthesis. 5,8,9 Cells were plated (10 5 per well) in 24-well plates, and were made quiescent for 12 hours in normal glucose-containing medium. After an additional 12 hours, the medium was changed to normal glucose or high glucose for another 48 hours. Five μCi of [³H]leucine (142 Ci/mmol, Amersham) were included per well for the last 12 hours. At the end of the incubation period, cells were washed twice in ice-cold PBS and proteins were subsequently precipitated with ice-cold 10% trichloroacetic acid. After redissolving the precipitates in 0.5 mol/L NaOH containing 0.1% Triton X-100, 5 ml of scintillation cocktail (Roth, Karlsruhe, Germany) was added, and vials were measured by liquid scintillation spectroscopy. [³H]leucine incorporation experiments were repeated five times with duplicate measurements for each experiment.

The incorporation of [³H]thymidine into DNA was used to measure proliferation. Cells (10 4 cells per well) were transferred to a 96-well micrometer plate. After incubation for 12 hours in normal glucose medium, they were subsequently incubated for another 48 hours in either normal or high glucose. They were pulsed with 1 μCi [³H]thymidine (5 Ci/mmol, Amersham) during the last 6 hours of culture. At the end of the incubation period, MMCs were washed in PBS, trypsinized for 10 minutes at 37°C, and finally collected on glass-fiber paper with an automatic cell harvester. Radioactivity of dry filters was measured by liquid scintillation spectroscopy. [³H]thymidine experiments were independently performed four times with triplicate measurements.

A ratio of total protein content to cell number was determined as another parameter of cellular hypertrophy. For this experiment, 10 5 cells were seeded into each well of a 6-well plate and were made quiescent for 12 hours in normal glucose-containing medium. After incubation for another 48 hours as appropriate, cells were shortly trypsinized, scraped off the plate with a rubber policeman, and were washed twice in PBS. A small aliquot of cells was counted in a Fuchs-Rosenthal chamber after resuspension of cells in PBS. The remaining cells were lysed in 0.5 mol/L NaOH and total protein content was measured by a modified Lowry method. Total protein content was expressed as μg protein per 10 3 cells. These experiments were independently performed five times.

Statistical Analysis

All values are presented as means ± SEM. Statistical significance among multiple groups was tested with nonparametric Kruskal-Wallis test. Individual groups were then tested using the Wilcoxon-Mann-Whitney test. A P value of < 0.05 was considered significant.

Results

High Glucose Increases the Levels of CDK Inhibitors p21^Cip1 and p27^Kip1 in p27^Kip1+/+ Cells

Figure 1 ▶ shows that quiescent p27^Kip1+/+ mesangial cells express p27^Kip1 protein. However, the incubation of p27^Kip1+/+ mesangial cells in serum-free media with high glucose (450 mg/dl) for 48 to 72 hours increased the levels of p27^Kip1 protein (Figure 1) ▶ (G100 for 48 hours:1.0 ± 0.0; G450 for 48 hours: 4.8 ± 0.6*; G450 for 72 hours: 6.2 ± 0.8* relative changes in p27^Kip1 expression normalized to β-actin; *P < 0.05, n = 3.). As expected, p27^Kip1 protein expression was totally absent in p27^Kip1−/− mesangial cells and was not induced by high glucose (Figure 1) ▶ .

Figure 1.

Western blot of total cell lysates incubated with an antibody against p27^Kip1. p27^Kip1 expression was, as predicted, not induced in p27^Kip1−/− mesangial cells incubated in serum-free medium with high glucose (G450; 450 mg/dl) for up to 72 hours. However, a strong high glucose-mediated induction of p27^Kip1 was detectable in wild-type (p27^Kip1+/+) cells. The membrane was re-incubated with an antibody against β-actin to demonstrate that these changes were not because of unequal loading or transfer of proteins. This blot is representative of three independent experiments with qualitatively similar results.

High glucose induced an increase in levels for the CDK inhibitor p21^Cip1 in p27^Kip1+/+ mesangial cells, and a similar increase in p21^Cip1 expression was observed in p27^Kip1−/− cells in response to high glucose (Figure 2) ▶ (p27^Kip1−/− G100 for 48 hours: 1.0 ± 0.0; p27^Kip1−/− G450 for 48 hours: 8.1 ± 1.1*; p27^Kip1+/+ G100 for 48 hours: 1.0 ± 0.0, p27^Kip1+/+ G450 for 48 hours: 8.6 ± 0.9*, relative changes in p21^Cip expression normalized to β-actin; *P < 0.05, n = 3). To ensure that any differences in glucose-induced hypertrophy were not because of alterations in TGF-β expression, 5 TGF-β mRNA and protein were measured in p27^Kip1+/+ and p27^Kip1−/− mesangial cells grown in high glucose. Figure 3, A and B ▶ , shows that TGF-β mRNA as well as protein were increased similarly in both cell lines when incubated with high glucose medium (G450) for 48 hours. Moreover, exogenous TGF-β1 increased fibronectin mRNA expression in both p27^Kip1−/− and +/+ cells (Figure 3C) ▶ . These results convincingly show that the response to high glucose is not different in p27^Kip1+/+ and −/− mesangial cells, except for the expected lack of p27^Kip1 response in the p27^Kip1−/− cells. Furthermore, putative TGF-β1 receptors and signaling transduction pathways involved in fibronectin transcription are present in both cell lines.

Figure 2.

Western blot for p21^Cip1 expression. In contrast to p27^Kip1, high glucose for 48 hours increased p21^Cip1 protein abundance in p27^Kip1−/− and +/+ mesangial cells. This blot is representative of three independent experiments with qualitatively similar results.

Figure 3.

A: mRNA expression of TGF-β1. High glucose medium induces TGF-β transcripts independent of whether cells express p27^Kip1 or not. This blot is representative of three independent experiments with qualitatively similar results. B: High glucose also leads to an equal amount of TGF-β1 protein synthesis in p27^Kip1+/+ and −/− mesangial cells. *, P < 0.01, n = 5. C: Exogenous TGF-β1 induced fibronectin mRNA expression in p27^Kip1+/+ and −/− cells indicating that both cell lines express TGF-β receptors and exhibit appropriate signal transduction systems downstream of the receptors. This blot is representative of three independent experiments with qualitatively similar results

High Glucose-Induced Hypertrophy Occurs in p27^Kip1+/+ but Not p27^Kip1−/− Cells

Proliferation was measured by [³H]thymidine incorporation. There was no significant difference in DNA synthesis in p27^Kip1+/+ and −/− mesangial cells when grown in serum-free media with normal glucose (Table 1) ▶ . However, incubation of p27^Kip1+/+ mesangial cells in serum-free medium with high glucose (G450) for 48 hours significantly reduced proliferation, consistent with glucose-induced cell cycle arrest in p27^Kip1+/+ cells. In marked contrast, high glucose caused a significant increase in cell cycle progression and proliferation in p27^Kip1−/− cells.

Table 1.

Effect of High Glucose Medium on Proliferation and Protein Synthesis of Cultured Mesangial Cells from Wild-Type and p27^Kip1 Knockout Mice

	[3H]Thymidine (proliferation)	[3H]Leucine (de novo protein synthesis)
p27Kip1+/+ in 100 mg/dl glucose	4,646 ± 675	276 ± 16
p27Kip1+/+ in 450 mg/dl glucose	2,511 ± 277*	369 ± 20*
p27Kip1−/− in 100 mg/dl glucose	6,914 ± 1,126	878 ± 47*
p27Kip1−/− in 450 mg/dl glucose	11,696 ± 1,720†‡	900 ± 26†

cpm, incubation for 48 hours, n = 10–12.

*P < 0.05 versus p27^Kip1+/+ in normal glucose (100 mg/dl).

^†P < 0.05 versus p27^Kip1−/− in normal glucose (100 mg/dl).

^‡P < 0.05 versus p27^Kip1+/+ in high glucose (450 mg/dl).

The de novo synthesis of protein in response to high glucose was measured by [³H]leucine incorporation and the results are shown in Table 1 ▶ . High glucose stimulated de novo protein synthesis in p27^Kip1+/+ mesangial cells. In contrast, [³H]leucine incorporation did not increase in p27^Kip1−/− mesangial cells in response to high glucose (Table 1) ▶ .

In addition, total protein content and cell number were determined and a hypertrophy index (protein content divided by cell number) was calculated. As shown in Table 2 ▶ , high glucose significantly increased protein content/cell number only in p27^Kip1 cells. Taken together, these results show that high glucose induced hypertrophy in p27^Kip1+/+ mesangial cells, but not in p27^Kip1−/− cells.

Table 2.

Effect of High Glucose Medium on Hypertrophy Index (Total Protein Content/Cell Number) of Cultured Mesangial Cells from Wild-Type and p27^Kip1 Knockout Mice

	Protein content/cell number (μg/103 cells)
p27Kip1+/+ in 100 mg/dl glucose	1.63 ± 0.49
p27Kip1+/+ in 450 mg/dl glucose	4.27 ± 1.00*
p27Kip1−/− in 100 mg/dl glucose	1.71 ± 0.26
p27Kip1−/− in 450 mg/dl glucose	2.08 ± 0.31

Incubation for 48 hours, n = 5–6.

*P < 0.05 versus p27^Kip1+/+ in normal glucose (100 mg/dl).

Transient Reconstituting p27^Kip1 Rescues the Hypertrophic Phenotype in p27^Kip1−/− Cells

To confirm a role for p27^Kip1 in glucose-induced hypertrophy, we used an ecdysone-inducible gene expression system to reconstitute p27^Kip1 expression in p27^Kip1−/− mesangial cells. This system offers the advantage of a lower basal activity compared to tetracycline-based expression vectors and the absence of any toxicity of the inducer muristone, a synthetic analog of the insect molting hormone ecdysone. 15 Figure 4A ▶ shows that p27^Kip1−/− cells transiently transfected with pINDp27^Kip1/pVgRXR grown in normal glucose medium (100 mg/dl) showed no detectable p27^Kip1 protein expression in the absence of muristone. However, incubating cells with 1 and 5 μg/ml of muristone for 24 hours increased p27^Kip1 expression (0 μg/ml muristone, 1.0 ± 0.0; 1 μg/ml muristone, 2.1 ± 0.3*; 5 μg/ml muristone, 4.7 ± 0.9* relative changes in p27^Kip1 expression normalized to β-actin; P < 0.05, n = 3). A similar pattern was observed when cells were grown in high glucose medium (450 mg/dl; Figure 4B ▶ ; 0 μg/ml muristone: 1.0 ± 0.0, 1 μg/ml muristone: 3.1 ± 0.4*, 3 μg/ml muristone: 2.8 + 0.3*, 5 μg/ml muristone: 3.7 ± 0.5* relative changes in p27^Kip1 expression normalized to β-actin; *P < 0.05, n = 3). However, the levels of p27^Kip1 were not increased in p27^Kip1+/+ cells transiently transfected with pINDp27^Kip1/pVgRXR (Figure 4C) ▶ .

Figure 4.

A-C: Inducible p27^Kip1 expression in transient transfected mesangial cells. A: p27^Kip1−/− cells were transiently transfected with pINDp27^Kip1/pVgRXR and grown in normal glucose medium without serum. One and 5 μg/ml of the synthetic ecdysone analog muristone for 24 hours strongly induced p27^Kip1 expression as detected in this Western blot. B: However, no further induction of p27^Kip1 was obtained when grown in high glucose indicating that the expression of the transgene is not under control of glucose. C: No additional induction of p27^Kip1 was obtained with muristone in p27^Kip1+/+ mesangial cells grown in high glucose medium that were transiently transfected with pINDp27^Kip1/pVgRXR. These blots are representative of four independent experiments with qualitatively similar results.

The biological effect of transiently transfecting p27^Kip1−/− cells is shown in Table 3 ▶ . Restoring p27^Kip1 expression with muristone inhibited [³H]thymidine incorporation in p27^Kip1−/− cells, and induced cell cycle arrest (Table 3) ▶ . Moreover, in the presence of high glucose, muristone induced cellular hypertrophy in p27^Kip1−/− cells to levels comparable to p27^Kip1+/+ cells (Table 4) ▶ . Thus, transiently reconstituting p27^Kip1 was required to induced hypertrophy in p27^Kip1−/− cells independent of the medium glucose content (Table 4) ▶ .

Table 3.

Effect of the Inducer Muristone on Proliferation and Protein Synthesis of p27^Kip1−/− Transiently Transfected with pINDp27^Kip1/pVgRXR

	[3H]Thymidine (proliferation)	[3H]Leucine (de novo protein synthesis)
p27Kip1−/− cotransfected with pINDp27Kip1/pVgRXR in mg/dl glucose without muristone	1,061 ± 42	948 ± 73
p27Kip1−/− cotransfected with pINDp27Kip1/pVgRXR in 100 mg/dl glucose+ 2 μg/ml muristone	846 ± 33*	1,105 ± 198
p27Kip1−/− cotransfected with pINDp27Kip1/pVgRXR in 450 mg/dl glucose without muristone	1,027 ± 35	1,272 ± 123
p27Kip1−/− cotransfected with pINDp27Kip1/pVgRXR in 450 mg/dl glucose+ 2 μg/ml muristone	792 ± 36‡	1,787 ± 154†

cpm, incubation for 24 hours, n = 10–12.

*P < 0.01 versus cells in 100 mg/dl glucose without muristone.

^†P < 0.05 versus cells in 450 mg/dl glucose without muristone.

^‡P < 0.01 versus cells in 450 mg/dl glucose without muristone.

Table 4.

Effect of the Inducer Muristone on Hypertrophy Index (Total Protein Content/Cell Number) of p27^Kip1−/− Transiently Transfected with pINDp27^Kip1/pVgRXR

	Protein content/cell number (μg/103 cells)
p27Kip1−/− cotransfected with pINDp27Kip1/pVgRXR in 100 mg/dl glucose without muristone	1.33 ± 0.22
p27Kip1−/− cotransfected with pINDp27Kip1/pVgRXR in 100 mg/dl glucose+ 2 μg/ml muristone	2.04 ± 0.18*
p27Kip1−/− cotransfected with pINDp27Kip1/pVgRXR in 450 mg/dl glucose without muristone	1.93 ± 0.17
p27Kip1−/− cotransfected with pINDp27Kip1/pVgRXR in 450 mg/dl glucose+ 2 μg/ml muristone	2.50 ± 0.15†

Incubation for 24 hours, n = 5.

*P < 0.05 versus cells in 100 mg/dl glucose without muristone.

^†P < 0.05 versus cells in 450 mg/dl glucose without muristone.

p27^Kip1−/− Cells Stably Transfected with Inducible p27^Kip1 Undergo Glucose-Induced Hypertrophy

To circumvent some of the intrinsic problems using mass cultures of transiently transfected cells, stable cell lines were generated. Therefore, mesangial cells from p27^Kip1−/− mice were first transfected with pVgRXR and selected in Zeocin to permanently express the ecdysone receptor. After several rounds of limited dilutions, a cell line was obtained that showed muristone-inducible expression of p27^Kip1 after transient transfection with pINDp27^Kip1 (data not shown). This cell line was subsequently transfected with pINDp27^Kip1 and selection of surviving cells was performed in medium containing Zeocin and G418. Two cell lines (clones 5.2 and 6.2) were selected for further analysis.

As shown in Figure 5 ▶ , muristone induced p27^Kip1 expression in these two cell lines suggesting that the inducible expression systems works in permanent-transfected cell lines. However, in contrast to transient-transfected cells, clones 5.2 and 6.2 revealed some minimal basal p27^Kip1 expression even in the absence of the inducer muristone indicating that the suppression is somewhat leaky (Figure 5) ▶ . Similar to the transiently transfected cells, muristone-induced p27^Kip1 expression inhibited the proliferation of both cell clones in high glucose medium (Table 5) ▶ . In addition, muristone converted the proliferation into a hypertrophy phenotype only in the presence of high glucose (Table 5) ▶ .

Figure 5.

Western blot for p27^Kip1 in clones 5.2 and 6.2, two stable transfected clones. Muristone induced p27^Kip1 expression in both cell lines. However, in contrast to transient-transfected cells, both clones demonstrated a minimal basal expression in the absence of the inducer indicating that genomic integration had occurred distal to a minimal promoter. Cells were incubated in normal glucose medium. These blots are representative of four independent experiments with qualitatively similar results.

Table 5.

Effect of the Inducer Muristone on Proliferation and Protein Synthesis of Clones 5.2 and 6.2

	[3H]Thymidine (proliferation)	[3H]Leucine (de novo protein synthesis)
Clone 5.2 in 100 mg/dl glucose without muristone	1,970 ± 132	193 ± 18
Clone 5.2 in 100 mg/dl glucose+ μg/ml muristone	1,620 ± 153*	186 ± 10
Clone 5.2 in 450 mg/dl glucose without muristone	1,435 ± 151	184 ± 11
Clone 5.2 in 450 mg/dl glucose+ 2 μg/ml muristone	1,138 ± 146†	256 ± 30†
Clone 6.2 in 100 mg/dl glucose without muristone	709 ± 61	259 ± 12
Clone 6.2 in 100 mg/dl glucose+ μg/ml muristone	708 ± 67	260 ± 15
Clone 6.2 in 450 mg/dl glucose without muristone	893 ± 72	225 ± 8
Clone 6.2 in 450 mg/dl glucose+ 2 μg/ml muristone	684 ± 43†	316 ± 42†

cpm, incubation for 24 hours, n = 10–12.

*P < 0.05 versus cells in 100 mg/dl glucose without muristone.

^†P < 0.05 versus cells in 450 mg/dl glucose without muristone.

Discussion

Glomerular hypertrophy, defined as an increase in cell size is because of an increase in protein content without DNA replication and occurs very early after the onset of diabetes mellitus. 1,2-4 A better understanding of the potential molecular mechanisms causing renal hypertrophy in diabetes mellitus is necessary because these events precede the development of irreversible structural changes such as glomerulosclerosis and tubulointerstitial fibrosis. 1 Although hemodynamic changes such as glomerular hyperfiltration may contribute through mechanical alteration to growth of glomerular cells, 16 there is now evidence that the diabetic milieu itself is pivotal in the development of mesangial hypertrophy. 1 Yet, the nuclear mechanisms underlying this are not well established.

High glucose as well as advanced glycation end products induce in vitro and in vivo TGF-β in the kidney. 17-20 Furthermore, angiotensin II may additionally induce TGF-β synthesis, particularly in the setting of high ambient glucose. 21 Neutralization experiments have clearly demonstrated that TGF-β is a necessary prerequisite for the development of glomerular hypertrophy in streptozotocin-induced diabetic mice. 20

We and others have been interested in the role of specific cell cycle proteins in diabetic hypertrophy, because we have previously described that high glucose, in the absence of other factors, induces immediate early genes in mesangial cells and the early entry G₀ to G₁ phases. 5 However, after a very limited proliferation, mesangial cells exposed to high glucose are growth-arrested in the G₁ phase, and do not progress into the S phase of the cell cycle. 5 Moreover, high glucose-mediated expression of TGF-β is pivotal for this G₁-phase arrest because neutralizing anti-TGF-β antibodies convert the G₁-phase arrest into a proliferative phenotype. 5 Young and colleagues 7 showed similar results in vivo where an early and limited glomerular proliferation occurred in the streptozotocin model which preceded glomerular hypertrophy.

CDK inhibitors cause cell cycle arrest by inactivating specific cyclin-CDK complexes required for cell cycle progression. In the current study we focused on specific CDK inhibitors because we have previously shown p21^Cip1 and p27^Kip1 are increased in vitro and in vivo in response to high glucose. 8-10 The major finding in the current study was that an induction of p27^Kip1 is necessary for high glucose-induced mesangial cell hypertrophy. Our results show that in contrast to an increase in protein synthesis and decrease in DNA synthesis (measures of hypertrophy) in p27^Kip1+/+ cells, high glucose did not stimulate the de novo synthesis of proteins in p27^Kip1−/− cells. Moreover, in contrast to glucose-induced cell cycle arrest in p27^Kip1+/+ cells, glucose increased cell cycle progression in p27^Kip1−/− cells. Finally, we showed that in the presence of high glucose, reconstituting p27^Kip1 levels in p27^Kip1−/− mesangial cells by either a transient or stable transfection converted these cells from a proliferative to hypertrophic phenotype.

A second major finding in this study was that in accord with our previous in vitro and in vivo studies, 10 the CDK inhibitor p21^Cip1 was increased in both p27^Kip1+/+ and p27^Kip1−/− mesangial cells in response to high glucose. 22,23 Because TGF-β can mediate p21^Cip1 expression by p53-dependent and -independent pathways, 24-26 it is possible that high glucose-induced TGF-β is responsible for the increase in p21^Cip1 in the diabetic environment. This expression of p21^Cip1 is required for glomerular hypertrophy because p21^Cip1−/− mice made diabetic by streptozotocin did not develop glomerular hypertrophy as measured by computer image analysis of glomerular tufts. 27 This absence of glomerular hypertrophy in p21^Cip1−/− diabetic mice seemed to be protective of renal function because these animals did not develop proteinuria. 27

A role for specific CDK inhibitors in hypertrophy has been recently shown. We found that treatment of BBdp rats, a model of autoimmune diabetes mellitus type I, with angiotensin-converting enzyme inhibitors prevented glomerular expression of p16 and p27^Kip1, but not of p21^Cip1. 28 Increased kidney weight, a parameter of hypertrophy, was also abolished by angiotensin-converting enzyme inhibitor treatment. 28 Hence, a picture is emerging in which both CDK inhibitors, p27^Kip1 and p21^Cip1, are required for the development of high glucose-mediated hypertrophy, but interference with the expression of one of these proteins may attenuate hypertrophy.

Terada and co-workers 29 overexpressed p21^Cip1 and p27^Kip1 in tubular LLC-PK₁ cells using adenovirus vectors. Although overexpression of each of the CDK inhibitors alone was sufficient to stimulate de novo protein synthesis of tubular cells in the presence of the mitogen epidermal growth factor, these overexpressions surprisingly failed to cause an inhibition of proliferation. 29 On the other hand, overexpression of p21^Cip1 in vascular smooth muscle cells suppressed serum-induced proliferation and stimulated hypertrophy. 30 Although species differences and/or the particular cell type may all explain these somewhat inconsistent findings, a more likely explanation is the fact that different total amounts of CDK inhibitors may have been expressed in the various systems.

It has been suggested that CDK inhibitors, in particular those of the Cip/Kip family, are redundant in some of their cell-cycle regulatory functions. 31,32 The rationale for this assumption was the observation that mice with targeted disruption of the p27^Kip1 gene, although exhibiting more cells in several organs, nevertheless revealed normal cell cycle arrest of lymphocytes treated with TGF-β or rapamycin. 33,34 Moreover, despite a high frequency of pituitary tumors in p27^Kip1-deficient mice, these animals are not predisposed to a general increase in cancer frequency suggesting that p27^Kip1 is not a genuine tumor suppressor gene. 33,34

p27^Kip1 inhibits several cyclin/CDK complexes including cyclin E/CDK2, and cyclin D/CDK4,6. 31 However, these cyclin/CDK complexes may show different susceptibilities to inhibition by p27^Kip1. 35 For example, binding of p27^Kip1 does not necessarily inhibit cyclin D/CDK4 in proliferating cells, whereas it always inactivates cyclin E/CDK2 and cyclin A/CDK2. 31,35 Studies with inducible p27^Kip1 expression have shown that the amount of p27^Kip1 required for the inhibition of cyclin D/CDK4 is much larger than that required for the inhibition of cyclin A/CDK2. 35 TGF-β decreases the expression of CDK4 36 and also induces the CDK4-specific inhibitor p15^INK4b. 37 These changes will disrupt binding of p27^Kip1 to cyclin D/CDK4 complexes leading to a redistribution and binding of p27^Kip1 to cyclin E/CDK2 and cyclin A/CDK2 complexes with their inhibition. 36,37 Thus, our observation that induced p27^Kip1 expression leads to hypertrophy only in high glucose medium may be explained by the fact that TGF-β-mediated down-regulation of CDK4 expression and/or induction of p15^INK4b, which occurs only in high glucose medium, is necessary for liberation of p27^Kip1 from cyclin D/CDK4 complexes with consecutive inhibition of cyclin E/CDK2. In accordance with this theory, Huang and Preisig 38 have demonstrated that cyclin D/CDK2 is activated, but cyclin E/CDK2 kinase is inhibited during diabetic hypertrophy. 38 This effect is mediated by TGF-β. The switch from the initial hyperplastic to the hypertrophic response of tubular cells was mediated by a decrease in cyclin E activity and p21^Cip1, p27^Kip1, and p57 played a role in this inhibition 38 suggesting that all three CKI may be necessary in diabetic hypertrophy.

In summary, we would like to propose the following orderly molecular course of glucose-mediated hypertrophy of mesangial cells as shown in Figure 6 ▶ : high ambient glucose primarily induces dormant cells to re-enter the cell cycle. After completing one or two rounds, CDK inhibitors such as p21^Cip1 and p27^Kip1 are induced, 8-10 likely through TGF-β-dependent as well as -independent mechanisms. Direct phosphorylation of p27^Kip1 by high glucose-activated MAP kinases could additionally increase the protein stability. 39-41 These CDK inhibitors interact with cyclin/CDK complexes, inhibit their activities, and arrest cells in the G₁ phase. It seems that p27^Kip1 is the major CDK inhibitor being necessary for this arrest. In addition, p21^Cip1 and p27^Kip1 may counteract apoptosis so that the overall number of G₁-phase-arrested cells remain intact. 11,40 Arrested cells are undergoing cellular hypertrophy through stimulated protein synthesis, increases in extracellular matrix, reduction in protein and matrix turnover, and maybe cellular enlargement by additional osmotic changes. 42-44 It is likely that TGF-β plays a central role in several of these processes. 20,45 Hence, interference with any of these consecutive events would inevitably abolish mesangial hypertrophy. However, whether a prevention of glomerular hypertrophy would ultimately attenuate renal function and structure in diabetic nephropathy remains unclear. 46

Figure 6.

Proposed molecular events leading to high glucose-induced hypertrophy of mesangial cells. High glucose induces TGF-β. TGF-β in turn stimulates the expression of the CDK inhibitors p27^Kip1 and p21^Cip1. We have previously shown that the high glucose-mediated induction of p27^Kip1 is to some extent independent of TGF-β and is mediated by protein kinase C. 8 Both CDK inhibitors, likely in concert, mediate G₁-phase arrest by binding to and inhibiting G₁-phase CDK 2, 4-cyclin E complexes. However, for the full development of cellular hypertrophy of G₁-phase-arrested cells other high glucose-induced factors are necessary such as other hypertrophic growth factors and/or cell-cycle-independent effects of TGF-β including stimulated matrix synthesis. Moreover, it has been shown that TGF-β leads to a down-regulation of cyclin D and induces p16^Ink4 with a potential release of p27^Kip1 from cyclin D-containing complexes that could now bind to cyclin E and further reinforce the G₁-phase arrest. —➤ = induction; · · · ·➤ = inhibition.