Autocrine Activation of the IGF-I Signaling Pathway in Mesangial Cells Isolated From Diabetic NOD Mice

Tissue culture plates were obtained from Nunclon (Nalge Nunc). Reagents for SDS-PAGE and immunoblotting were obtained from Novex (San Diego, CA). HEPES, phenylmethylsulfonyl fluoride (PMSF), aprotinin, leupeptin, benzamidine, EDTA, sodium pyrophosphate, sodium fluoride and sodium orthovanadate, insulin, Triton X-100, Tween 20, bovine serum albumin (BSA; Fraction V), glycerol, and NaCl were from Sigma Chemical (St. Louis, MO). PD98059 and LY294002 were purchased from Calbiochem (La Jolla, CA). Nitrocellulose membranes were from Amersham Pharmacia Biotech (Piscataway, NJ). Recombinant human IGF-I and a specific neutralizing monoclonal antibody for IGF-I were purchased from Upstate Biotechnology (Lake Placid, NY). All antibodies and the protein A agarose were obtained from Santa Cruz (Santa Cruz, CA).

Mesangial cells were isolated from NOD mice before and after the spontaneous onset of diabetes as previously described (7). Briefly, glomeruli were isolated from kidneys of 4- to 6-month-old D-NOD and ND-NOD mice. Diabetic mice had glycosuria for 4–8 weeks before sacrifice and were receiving two insulin injections each day. Nondiabetic mice had normal glucose tolerance, as determined by a glucose tolerance test before sacrifice. Several lines of mesangial cells were derived from each of several D-NOD and ND-NOD mice (7). In each experiment, two separate mesangial cell lines derived from different D-NOD mice and two separate mesangial cell lines from two different ND-NOD mice were used between passages 15 and 28.

Three days before the collection of protein, cells were plated in either 6-well plates or T75 cm² flasks in B medium containing 20% fetal bovine serum and 6 mmol/l glucose as previously described (7). Twenty-four hours before collection, the medium was replaced with B medium containing 0.1% BSA. Cell number was determined in duplicate wells at each experimental time point. Phosphorylation of IGF-I receptor and insulin receptor substrate (IRS)-1 and IRS-2 were examined after exposure to either IGF-I (50 ng/ml) or insulin (50 ng/ml) for 10 min. To block activation of the signaling pathway, a neutralizing IGF-I antibody (34 μg/ml), PD98059 or a specific inhibitor of mitogen-activated protein kinase (MAPK)/extracellular response kinase (ERK) kinase (MEK), an upstream kinase activator of ERK (20–40 μmol/l final concentration), or a specific phosphatidylinositol 3-kinase (PI3K) inhibitor, LY294002 (4 μmol/l final concentration), was added to the culture medium for 12 and 36 h.

Each well of a 6-well plate was washed twice with cold PBS and 100 μl cold lysis buffer (20 mmol/l Tris, 140 mmol/l NaCl, 3 mmol/l EDTA, 10 mmol/l NaF, 10 mmol/l Na pyrophosphate, 2 mmol/l NaVO₄, 10% glycerol, pH 7.4, 1% Triton X-100, aprotinin, leupeptin, and PMSF) was added for 2 min before harvesting the cell layer by scraping. Cell lysates were incubated for 45 min at 4°C, and the insoluble material was removed by centrifugation at 20,000g for 30 min at 4°C. Samples were analyzed by electrophoresis through 6% (IRS-1, IRS-2, and IGF-IRβ), 8% (PI3K-p85), and 10% (ERK-1/2 and phospho-ERKs) polyacrylamide gels and electrophoretically transferred to nitrocellulose membranes. After overnight incubation at 4°C in Tris-buffered saline, 5% milk, or 1% milk plus 1% BSA and 0.05% Tween-20, the blots were exposed to antibodies recognizing either IRS-1, IRS-2, IGF-IRβ, PI3K-p85, ERK-1/2, phospho-ERKs, and antiphosphotyrosine PY20 or PY99 for 1 h at room temperature. The primary antibodies were revealed using the corresponding goat or donkey peroxidase–conjugated secondary antibodies (1/2,000–1/4,000) for 1 h. Peroxidase activity was detected using the Santa Cruz chemiluminescence kit.

For immunoprecipitation experiments, 250 μg protein extract was incubated with either IGF-IRβ or IRS-1 antibody for 1 h at 4°C, followed by the addition of protein A agarose overnight. The resulting protein A antibody conjugate was centrifuged at 4°C and washed four times with PBS (pH 7.4). The final pellet was resuspended in PBS, sample buffer was added, and the mixture was boiled for 3 min before analysis as described above.

To determine whether blocking IGF-I activation increased MMP-2 in diabetic mesangial cells, medium was collected from cells exposed to PD98059 for 24 h or LY294002 for 36 h. Cell supernatants were electrophoresed and incubated for 1 h in 2.5% Triton X-100 and incubated overnight in collagenase as previously described (11).

Comparison between groups was performed using one-way analysis of variance and Tukey’s multiple comparison test. Changes in MMP-2 activity after exposure to PD98059 or LY294002 were analyzed by unpaired Student’s t test.

IGF-IRβ-chain expression in mesangial cells was assessed by Western immunoblot. Lysates from at least six separate experiments on each of two mesangial cell lines from two separate ND-NOD mice and two other mesangial cell lines from two D-NOD mice mesangial cell lines were examined. Serial dilutions of a large pool of mesangial cell lysates was run with each gel as an internal standard, which allowed a semiquantitative comparison between gels. Mesangial cells from D-NOD mice contained approximately twofold more IGF-I receptor protein than those from ND-NOD mice (P < 0.001) (Fig. 1).

IRS-1 and -2 are the most important substrates of IGF-I receptor kinase activity in many cell lines (12). Both proteins were expressed in mesangial cells isolated from NOD mice (Fig. 2). The expression of both IRS-1 and -2 was increased twofold in mesangial cells from D-NOD mice, as compared with ND-NOD mice (P < 0.001). This increase was similar to the increase in IGF-I receptor expression in the corresponding cells.

IGF-I receptor activation leads to tyrosine phosphorylation of its β-chain and binding of the regulatory (p85) subunit of PI3K (13). We examined these events by immunoprecipitating cell lysates with IGF-I receptor antibody followed by Western immunoblotting with antibodies against phosphorylated tyrosine residues (PY99) and the p85 subunit of PI3K. Two bands were identified (Fig. 3). Based on its molecular weight and the effect of added IGF-I, the lower band corresponded to the β-chain of the IGF-I receptor (Fig. 3, middle panel). We found that the IGF-I receptor was phosphorylated in the basal state in mesangial cells isolated from D-NOD mice but not in those from ND-NOD mice (Fig. 3, middle panel). In all cell lines, incubation for 5 min in the presence of 50 ng/ml IGF-I resulted in a prominent phosphorylation of the IGF-I receptor β-chain. Interestingly, whereas incubation with a similar dose of insulin had no effect on phosphorylation of the β-chain of the IGF-I receptor, it induced phosphorylation of a band migrating above the IGF-I receptor (Fig. 3, middle panel). We postulated that the protein phosphorylated by insulin may be the β-chain of insulin receptor, based on its homology with the β-chain of IGF-I receptor and its molecular weight (14).

The p85 subunit of PI3K was shown to spontaneously associate with the β-chain of the IGF-I receptor. The amount of p85 associated with IGF-I receptor was greater in mesangial cells from diabetic mice (Fig. 3, lower panel). The addition of IGF-I to media increased the association of the p85 subunit to the IGF-I receptor in both the ND-NOD and the D-NOD cell lines.

The effect of IGF-I on IRS-1 and -2 phosphorylation was studied by immunoprecipitating cell lysates using a mixture of IRS-1 and -2 antibodies. The precipitates were examined by Western immunoblots. Because the amount of IRS in mesangial cells from D-NOD mice was twofold greater, the quantity of extract loaded for the Western blot after immunoprecipitation was halved (Fig. 4, top panel). Under these conditions, the phosphorylated forms of both IRS-1 and -2 were detected. Examination of mesangial cells in the basal state revealed that phospho–IRS-1 was increased in those cells isolated from diabetic mice compared with those isolated from ND-NOD mice. A 10-min exposure to 50 ng/ml IGF-I resulted in a twofold increase in IRS-1 phosphorylation, whereas IRS-2 phosphorylation was unchanged. Stimulation with 50 ng/ml insulin had no effect, consistent with the fact that there are few insulin receptors on the surface of glomerular mesangial cells (2,3).

As has been reported, there was also spontaneous association of the p85 subunit with IRS-1/2 (Fig. 4, lower panel) (13). Incubation with IGF-I increased the association of the p85 subunit to IRS in all of the ND-NOD and D-NOD cell lines.

The MEK pathway can be activated by IGF-I in many cell types (15). ERK1 and -2 are activated by phosphorylation of their tyrosine residues (16). ERK1 and -2 were expressed at similar levels in all mesangial cell lines, irrespective of the diabetic status of the NOD mice (Fig. 5A). The phosphorylated (activated) forms of ERK1 and -2 in the basal state were examined by Western immunoblotting using an antibody that recognized the tyrosine-phosphorylated forms. Phosphorylated ERK2 was more abundant in the basal state of D-NOD than in ND-NOD mesangial cells (8.2-fold higher) (Fig. 5B). The addition of 50 ng/ml of IGF-I for 120 min (Fig. 5C) increased the amount of phospho-ERK1 and -ERK2 in all of the ND-NOD and the D-NOD cell lines. However, the increase of ERK1 phosphorylation was more pronounced in mesangial cells isolated from ND-NOD mice (1.7-fold increase).

Mesangial cells synthesize and secrete IGF-I (2,3,4,5). Therefore, an IGF-I–neutralizing antibody was added to cell culture medium to determine whether autocrine secretion was responsible for the activation of IGF-I pathway in mesangial cells isolated from D-NOD mice. IGF-I receptor β-chain and IRS-1 phosphorylation was assessed by immunoprecipitation with the corresponding antibodies, followed by Western immunoblotting with an anti–phosphorylated tyrosine antibody. The addition of the IGF-I–neutralizing antibody (Fig. 6A) blocked basal phosphorylation of the IGF-I receptor in mesangial cells from D-NOD mice. We confirmed by zymography that it also increased MMP-2 activity as previously reported (data not shown) (9). Furthermore, the addition of the neutralizing antibody reduced the amount of phospho–IRS-1 in D-NOD mesangial cells to a level that was similar to that of resting mesangial cells isolated from ND-NOD mice (Fig. 6B, lower panel). In contrast, the neutralizing antibody did not affect the amount of phospho–IRS-1 in mesangial cells from ND-NOD mice (Fig. 6B, lower panel).

After exposure to an IGF-I–neutralizing antibody, the amounts of phosphorylated ERK1 and -2 were only decreased in D-NOD mesangial cells, whereas they were unchanged in similarly treated ND-NOD (Fig. 7A, lower panel). PI3K activation is an important downstream event after IGF-I receptor activation (17). The increased amount of p85 subunit associated with the IGF-I receptor and IRS (Figs. 3 and 4) suggested that PI3K might be implicated in the downstream events of the intracellular IGF-I pathway, such as phosphorylation of MAPKs. This was investigated using LY294002, a specific PI3K inhibitor (18). LY294002 did not alter the growth curve of mesangial cells isolated from either ND-NOD or D-NOD mice. After the addition of LY294002 for 12 h, the basal levels of phosphorylated ERK1 and -2 decreased in D-NOD mesangial cells (57.9 ± 5.93% and 63.3 ± 4.34%, respectively) but not in ND-NOD mesangial cells (Fig. 7B). This decrease was similar to that obtained with the IGF-I–neutralizing antibody (51.6 ± 7.77 and 56.9 ± 2.26%, respectively). No further inhibition was observed with the addition of both the PI3K inhibitor and the IGF-I antibody (data not shown). When D-NOD cells were exposed to the specific MEK inhibitor PD98059 for 12 h, phosphorylated ERK1 and -2 decreased by 54 and 80% of baseline, respectively (data not shown).

To determine whether the inhibition of IGF-I signaling resulted in increased MMP-2 activity, D-NOD and ND-NOD mesangial cells were treated with LY294002 for 36 h or PD98059 for 16 h. Medium was collected, and zymography was performed to assess MMP-2 activity. MMP-2 activity was increased only in D-NOD cells after treatment with LY294002 (134 ± 0.94%, P < 0.0001) (Fig. 8A) or PD98059 (136 ± 9.3%, P < 0.05) (Fig. 8B).

The purpose of the current study was to determine whether the phenotypic change in IGF-I expression in mesangial cells isolated from D-NOD mice resulted from specific alterations in its signaling pathways. We found that the IGF-I signaling pathway was activated in mesangial cells isolated from D-NOD mice due to the synthesis of excess IGF-I. For instance, the β-subunit of the IGF-I receptor was significantly increased in mesangial cells isolated from D-NOD mice when compared with those from nondiabetic littermates. These data are consistent with our previous report that diabetic cells had more surface IGF-I receptors when autocrine IGF-I was blocked with a neutralizing antibody to IGF-I (7). In addition, the increase in IRS-1 and -2 levels suggested that the number of receptors and their proximate substrate(s) were increased in a coordinate manner. Similarly, there was increased phosphorylation of tyrosine residues and an increased association of the P85 subunit to the receptor in mesangial cells isolated from diabetic mice. In addition, whereas ERK1 and -2 levels were found to be similar in all cell lines from diabetic and nondiabetic mice, activated (phosphorylated) ERK2 was increased only in the basal state of cells isolated from diabetic mice. Although the addition of exogenous IGF-I further increased levels of phosphorylated ERKs in D-NOD cells, the increase was less prominent than in ND-NOD cells.

The IGF-I signaling cascade has been investigated in many cell types, including vascular smooth muscle cells (19) and fibroblasts (20). Oemar et al. (8) reported the induction of tyrosyl phosphorylation of nuclear proteins by IGF-I in mesangial cells isolated from a model of type 2 diabetes (db/db). However, to our knowledge, there are no reports on the activation of this pathway in mesangial cells isolated from a model of type 1 diabetes, such as the NOD mouse.

Because cells isolated from D-NOD mice synthesized more IGF-I at baseline (7), we added a neutralizing antibody to IGF-I to determine whether there were also abnormalites in the IGF-I signaling pathway. We found that the addition of this antibody prevented phosphorylation of the IGF receptor β-subunit, reduced the amount of phospho–IRS-1, and decreased the amounts of phosphorylated ERK1 and -2 in the cells isolated from diabetic mice.

Finally, because the amount of p85 subunit associated with the IGF receptor was increased in cells isolated from diabetic mice, we used an inhibitor (LY294002) to determine whether downstream signaling events were affected. Importantly, after the addition of LY294002, we found decreased levels of phosphorylated ERK1 and -2 in D-NOD cells, suggesting that the IGF-I–dependent signal transduction from PI3K to ERKs was inhibited. Although it was previously reported in MCF-7 breast cancer cells (21), this pathway has not been explored in glomerular mesangial cells.

Thus, the IGF-I signaling pathway is intact in mesangial cells isolated from both D-NOD and ND-NOD mice. However, the levels of activated ERK were relatively increased in the D-NOD cells. Stimulation with IGF-I, either autocrine or exogenous, activated signal transduction in both D-NOD and ND-NOD mesangial cells, although the magnitude of the response differed markedly. Removal of autocrine IGF-I with a neutralizing antibody decreased the activation in the D-NOD mesangial cells but had little effect on mesangial cells isolated from ND-NOD mice, even though they synthesized and secreted small amounts of IGF-I under basal conditions.

We have previously shown that excess IGF-I synthesized by mesangial cells isolated from D-NOD mice resulted in decreased MMP-2 activity. Conversely, removal of the autocrine IGF-I caused an increase in MMP-2 activity (9). These data suggest that overexpression of IGF-I may result in decreased degradation of ECM and lead to an accumulation of ECM, a characteristic feature of diabetic nephropathy. A recent report suggests that constitutive activation of MEK (MEK1, upstream from ERK) leads to activation of MMP-2 in a rat fibroblast cell line (22). In contrast, we found that activation of ERK1 and -2 is associated with decreased MMP-2 activity in diabetic mesangial cells, since blocking IGF-I activation through either the PI3K or MAPK pathway increased MMP-2 activity. This suggests that MAPK regulation of MMP-2 activity may be cell- or tissue-specific.

In conclusion, the constitutive overexpression of IGF-I in mesangial cells isolated from D-NOD mice leads to autoactivation of the IGF-I signaling pathways, including increased basal expression and phosphorylation of signaling components (ERK1 and -2). This may lead to decreased ECM degradation and appears to be part of the phenotypic changes induced after the onset of diabetes.