Laminin-β2 (LAMB2) is a critical component of the glomerular basement membrane as content of LAMB2 in part determines glomerular barrier permeability. Previously, we reported that high concentrations of glucose reduce expression of this laminin subunit at the translational level. The present studies were undertaken to further define systems that control Lamb2 translation and the effect of high glucose on those systems. Complementary studies were performed using in vitro differentiation of cultured podocytes and mesangial cells exposed to normal and elevated concentrations of glucose, and tissues from control and diabetic rats. Together, these studies provide evidence for regulation of Lamb2 translation by IMP2, an RNA binding protein that targets Lamb2 mRNA to the actin cytoskeleton. Expression of Imp2 itself is regulated by the transcription factor HMGA2, which in turn is regulated by the microRNA let-7b. Elevated concentrations of glucose increase let-7b, which reduces HMGA2 expression, in turn reducing IMP2 and LAMB2. Correlative changes in kidney tissues from control and streptozotocin-induced diabetic rats suggest these control mechanisms are operative in vivo and may contribute to proteinuria in diabetic nephropathy. To our knowledge, this is the first time that translation of Lamb2 mRNA has been linked to the actin cytoskeleton, as well as to specific RNA-binding proteins. These translational control points may provide new targets for therapy in proteinuric disorders such as diabetic nephropathy where LAMB2 levels are reduced.
laminins are heterotrimeric (αβγ), extracellular glycoproteins in glomerular (GBM) and other basement membranes (16). The laminin-β2 (LAMB2) subunit is required for normal GBM structure and barrier function as human mutations causing loss of this protein lead to nephrotic syndrome and progressive kidney dysfunction (23, 28, 29, 33). Mice deficient in LAMB2 die soon after birth with severe proteinuria and abnormal neuromuscular junctions (32, 33). Thus, LAMB2 is a key subunit in the laminin family, providing critical functions in kidney and other tissues (40).
Mechanisms controlling expression of the Lamb2 gene as well as LAMB2 protein production, secretion, and degradation are not well-understood (17). We previously demonstrated that LAMB2 is not expressed in developing glomeruli of hyperglycemic animals, and it is reduced in kidneys of rats with streptozotocin-induced diabetes (2), as a result of glucose-mediated impairment of Lamb2 mRNA translation (39). Also, loss of LAMB2 results in impaired migratory responses of mesangial cells (MC), which may limit repair in the face mesangiolysis, a prominent feature of diabetic nephropathy (DN) (39). Glucose-mediated reductions in LAMB2 content in GBM may also contribute to microalbuminuria in DN (30).
In this report, we demonstrate that the RNA-binding protein (RNA-BP) IMP2 is required for LAMB2 translation, as 1) Lamb2 mRNA binds specifically to IMP2 and associates with actin during translation, 2) LAMB2 protein production is reduced if IMP2 is absent, 3) under conditions of high glucose, IMP2 levels are reduced resulting in loss of Lamb2 mRNA translation, and 4) both IMP2 and LAMB2 protein levels are reduced in kidneys from hyperglycemic rats. In addition, we demonstrate that let-7b microRNA is likely to play a role as it regulates the level of HMGA2, a transcription factor required for IMP2 transcription. To our knowledge, this is the first time that a molecular circuit that controls translation of LAMB2 has been described.
METHODS AND MATERIALS
The following antibodies were used: IMP2 (TA501269, 3F9, Origene, Rockville, MD) for immunobloting (IB) or RN008P (MBL, Woburn, MA) for RNA immunoprecipitation; LAMB2 [C4 (IB) or D5 immunofluorescence (IF)], a monoclonal anti-laminin β2 developed by Dr. J. Sanes obtained from the Developmental Studies Hybridoma Bank (University of Iowa, Iowa City, IA); rabbit antibody to human COOH-terminal filamin A (gift from Dr. R. Tyler Miller, Case Western Reserve University, Cleveland, OH); actin (04–1040, EP184E, Millipore, Billerica, MA); or tubulin (04–1117, EP1332Y, Millipore).
Animal and tissue procedures.
Six-week-old male Sprague-Dawley rats weighing 200 g were given 65 mg/kg streptozotocin intraperitoneally. Diabetes mellitus was confirmed and monitored, and kidneys were processed for IF microscopy, glomerular isolation, and protein isolation as described (1). Slides were viewed using a Zeiss microscope equipped for epi-illumination and photographed using a digital RT-color Spot camera or a Zeiss AxioCam MRM black and white camera. Total protein excretion was measured on 24-h urine specimens (2). All studies were approved by the Institutional Animal Care and Use Committee of the Veterans Affairs Puget Sound Health Care System.
Cloned rat glomerular MCs were prepared and cultured as described (3, 4, 26). Human podocytes were obtained from and cultured as described by Saleem et al. (37). The podocyte cell line tested negative for evidence of replication-competent virus. The control positive virus “AM-MLV” is a representative of the type of virus that might be produced by the PA317 retrovirus packaging cells used to construct the podocyte cell line.
MC were transfected with Lipofectamine 2000 (Invitrogen/Life Technologies, Grand Island, NY) per the manufacturer's protocol (reverse transfection) with siRNA (0.5–0.1 μM). RNA and protein measurements were made 24–72 h later. Si-RNAs were as follows: controls (no. 45–2001, no. 45–2002) and rat-specific si-Imp2 (no. RS357477, no. RS357478, no. RS357479; Invitrogen/Life Technologies).
Podocytes (4 × 105 cells/well) were incubated with the appropriate miRs/anti-miRs/controls for 24 h using iPort NeoFX transfection reagent (Ambion/Life Technologies, Grand Island, NY) (21, 24). Let-7b mirVana mimic, let-7b mirVana inhibitor, and pre-miR negative control #1 (random sequences) were used (30 nM). Let-7b mirVana mimic, Let-7b mirVana inhibitor, and miR negative control #1 were obtained from Ambion/Applied Biosystems/Life Technologies.
Total RNA was isolated using RNAqueous-4PCR kit (Ambion/Life Technologies). Quantitative real-time PCR (qPCR) was performed using primers and conditions as described (39). Primers were as follows: no. QT01567573 rat IMP-2, no. QT01670781 human IMP2, no. QT00366821 rat lamb2, and no. QT00050771 human LAMB2 (Qiagen, Valencia, CA).
Cells (2 × 107) were resuspended in 200 μl of lysis buffer and mixed (overnight at 4°C) with magnetic beads bound with either 5 μg of anti-IMP-2 (RN008P, MBL) or control IgG (Magna RIP kit no. 17–700, Millipore). Bound RNAs were precipitated and quantified by RT-PCR.
For whole cell lysates, cells were extracted in PBS containing 0.1% SDS, 0.5% Triton X-100, and protease inhibitors (Sigma, St. Louis, MO). Western blots used Tris-HEPES-SDS precasted 4–20% gels (Thermo-Scientific, Rockford, IL) (39). Isolated glomeruli were processed as described in Ref. 39. Equal amounts of protein (BCA protein assay, Pierce, Rockford, IL) were subjected to SDS-PAGE electrophoresis and transferred to nitrocellulose. Band intensities were quantified by scanning using Image J software (22).
mRNA translation was determined by assessing the proportion of mRNA associated with polysomes using ultracentrifugation on a 7–47% sucrose gradient (44). Six 2-ml fractions were collected (P1-P6), and mRNA was precipitated, reverse transcribed, and assayed by qPCR for Lamb2. Load and recovery were followed by the addition of carrier Escherichia coli mRNA that was also assayed by qPCR (39).
All samples were run in triplicate and repeated a minimum of three times on separate occasions except for isolated rat glomeruli (n = 3/group). Results expressed as the group means ± 1 SD were compared by one-way ANOVA with subgroup testing by contrasts. Two groups were compared by t-test. P < 0.05 was considered significant.
The goal of this study was to define the mechanism(s) by which Lamb2 translation is regulated using podocyte differentiation in vitro as a model of normal control of LAMB2 expression, and glucose effects on MC as a model of altered mechanisms of control of Lamb2 translation that could be relevant to DN.
Ribosome-bound Lamb2 mRNA coprecipitates with components of the cytoskeleton.
Previously, we noted that Lamb2 mRNA is associated only with high molecular weight (HMW) polysomes, suggesting that Lamb2 mRNA might be associated with actin (25). To test this possibility, we isolated polysome fractions by sucrose gradient ultracentrifugation from podocytes 4 days after in vitro differentiation (37), identified those fractions containing Lamb2 mRNA, and examined them for associated proteins. Lamb2 mRNA was primarily found in the HMW polysome fraction (no. 6) with actin (Fig. 1, A and B). This fraction lacked α-tubulin, suggesting that Lamb2 mRNA is associated with cytoskeleton rather than microtubular structures.
Translation of mRNAs on the actin cytoskeleton usually requires delivery of mRNA to actin by binding to an RNA-BP that interacts with actin. This led us to a search for potential candidates and known regulators of those binding relationships. The insulin-like growth factor 2-binding proteins (IGF2BPs/IMPs) are actin-targeting RNA-BPs known to be important during cell differentiation (14). Lamb2 mRNA immunoprecipitated with IMP2 in differentiated podocytes (Fig. 1C, left) and in MC (Fig. 1C, right). In contrast, other mRNA species [Lamb1, integrin β 8 (Itgb8), glyceraldehyde 3-phosphate dehydrogenase (Gapdh)] were not found in the IMP2 complex (Fig. 1C, right). Anti-IMP2 immunoprecipitates from MC also contained actin, but not tubulin (Fig. 1D). The presence of filamin A (Fig. 1D), an actin cross-linking protein, is additional evidence that Lamb2 mRNA-IMP2 is complexed with cross-linked actin.
In summary, these data provide evidence that IMP2 targets Lamb2 mRNA to the cytoskeleton where translation is activated. Additional studies are needed to confirm this location of Lamb2 translation and to understand potential signal transduction cascades that activate it.
Control of LAMB2 translation.
To further assess the importance of IMP2 in Lamb2 translation, the expression of both proteins was evaluated before and after podocyte differentiation. LAMB2 expression increased on day 4 after induction of differentiation and was maintained through day 14 (Fig. 2A). The induction of IMP2 protein mimicked that of LAMB2 (Fig. 2B) and correlated with the timing of increased Imp2 mRNA levels (Fig. 2C). In contrast, mRNA transcript levels for other IMP family members, Imp1 and Imp3, were significantly reduced after podocyte differentiation (Fig. 2C), indicating they are not involved in regulation of Lamb2 mRNA translation. Of note, the majority of Imp2 mRNA (∼80%) was bound to polysomes at day 14 (Fig. 2D) further supporting its synthetic upregulation with podocyte differentiation. Taken together, these data implicate IMP2 as a RNA-BP that modulates Lamb2 translation in podocytes.
To determine whether IMP2 is required for Lamb2 mRNA translation, we used siRNA to silence IMP2 expression and assess the impact on LAMB2 expression. For these experiments, rat MC were used instead of podocytes because consistent transfection into MC is easier to achieve. Loss of IMP2 mRNA and protein was confirmed with three different siRNAs, but not controls. Imp2 mRNA and protein levels were quantified at 24, 48, and 72 h following transfection and were compared with scrambled siRNA controls. Loss of IMP2 protein was seen with all three siRNAs, but not the siRNA controls. In association with a reduction in IMP2 protein, LAMB2 levels also declined at 48 and 72 h. Figure 3A is a representative immunoblot showing the absence of IMP2 from cells and a marked reduction in LAMB2 protein at 72 h. Quantification of IMP2 (Fig. 3B) and LAMB2 (Fig. 3C) shows significant reductions in both proteins, indicating that IMP2 is required for LAMB2 protein production.
To further understand control of LAMB2 protein expression, we explored upstream modulators of IMP2. It is known that transcriptional regulation of the human IMP2 gene is dependent on the transcription factor high-mobility group AT-hook 2 (HMGA2) (11, 15). To test its role in LAMB2 synthesis, we followed HMGA2 expression as a function of podocyte differentiation. Consistent with our expectations, HMGA2 protein levels increased with podocyte differentiation, mirroring responses of IMP2 (Fig. 4). Indeed, 4 days of podocyte differentiation resulted in nearly a fivefold enhancement of HMGA2 protein levels. Although correlative in nature, these results are consistent with a role for HMGA2 in modulating IMP2 expression in kidney cells.
Let-7b is a candidate miR modulating our system because Hgma2 is known to be a let-7 target gene (10, 27). We used cultured podocytes to test whether alterations of let-7b levels modulate levels of HMGA2 and IMP2. As shown in Fig. 5, levels of HMGA2 and IMP2 were altered as expected. First, inhibition of let-7b using a single-strand RNA-based oligonucleotide that binds to and inhibits the endogenous let-7b resulted in increased expression of HMGA2 and IMP2 compared with controls (Fig. 5A). In differentiated podocytes (Fig. 5B), overexpression of let-7b using a small double-strand RNA that mimics endogenous let-7b resulted in decreased expression of HMGA2 and IMP2 compared with controls. Thus, we suggest that let-7b is part of the molecular circuitry modulating HMGA2 and IMP2, which in turn modulates LAMB2 protein levels.
Glucose modulation of LAMB2 translation.
Based on data obtained above, we turned to the model of glucose-mediated changes in Lamb2 translation using MC. This allowed us to establish that the putative relationships defined above existed in a system in which we previously established that Lamb2 translation was impaired by exposure to high concentrations of glucose (HG; 25 mM) (39). MC treated with HG for 7 days were assessed for levels of LAMB2, IMP2, HMGA2, and let-7b. As shown in Fig. 6, protein levels for LAMB2 (A), IMP2 (B), and HMGA2 (C) were markedly reduced by HG. In contrast, let-7b levels (Fig. 6D) were elevated by HG compared with low glucose (LG; 5 mM). These data suggest that glucose levels modulate Lamb2 translation through effects on IMP2 that are associated with the expected changes in Let7b and HMGA2.
To explore the in vivo relevance of these observations, kidney sections from control and diabetic rats were immunostained for LAMB2, IMP2, and HMGA2 just following the onset of proteinuria in diabetic rats (Fig. 7A, inset). As expected from the in vitro studies, we found decreased immunostaining for HMGA2, IMP2, and LAMB2 in diabetic kidney tissue (Fig. 7). LAMB2 staining in diabetic rats was discontinuous along the GBM. This suggests that as GBM turnover proceeds in the presence of hyperglycemia, newly synthesized GBM that lacks LAMB2 replaces GBM that formed before the induction of diabetes. Total glomerular LAMB2 protein detected by Western blotting was reduced by 30% (Fig. 7B). Given the relatively slow rate of GBM turnover, which has been suggested to be further slowed by diabetes (6, 7, 36), only modest reductions in total LAMB2 content would be expected. The temporal correlation between loss of LAMB2 and onset of proteinuria suggests that this could be an important contributor to changes in protein filtration as established by Miner et al. (30). Our findings do not exclude the possibility that other abnormalities in diabetes also contribute to abnormalities in the filtration barrier.
Several conditions have been reported to express different LAMB2 protein levels, including diabetes, cancer, and asthma (2, 5, 19, 40). Yet, few transcriptional stimuli have been identified and their effects on Lamb2 transcription have been modest (8). Our previous studies agree with modest regulation of Lamb2 at the transcriptional level (39). Here, we confirm modest changes in Lamb2 mRNA in differentiating podocytes and in MC exposed to HG; yet, significant changes in LAMB2 protein expression were observed. In studies designed to understand control of LAMB2 translation, we found Lamb2 mRNA in HMW polysomes bound to actin, which suggests that Lamb2 mRNA may be translated on the cytoskeleton. Translation was regulated by a circuitry involving the RNA-BP, IMP2, the transcription factor, HMGA2, and the miR let-7b.
IMP2 belongs to a family of three RNA binding proteins (IMP1–3), which are primarily expressed during development (14). IMP2 is singular in that it is also abundantly expressed in several adult organs including kidney. IMPs have NH2-terminal RNA recognition motifs and several COOH-terminal heterogeneous nuclear ribonucleoprotein-K homology domains; thus, they facilitate nuclear export of specific transcripts (31, 34). In the cytoplasm they form large ribonuclear protein complexes, controlling mRNA stability and translation potential (14). IMP2 is one RNA-BP that has an actin-binding domain, which is consistent with our observation that polysome-bound Lamb2 mRNA was found in association with actin. Correlations between IMP2 levels and LAMB2 in differentiating podocytes, glucose-treated MC, and glomeruli from diabetic rats suggest an important role for IMP2 in mediating LAMB2 translation, which was confirmed in vitro using siRNA-mediated IMP2 knockdown. Although correlative in nature, HMGA2 is a candidate control point for LAMB2 via transcriptional regulation of IMP2 expression, as it is already known to modulate IMP2 levels (11, 15). Let-7b miR is known to regulate expression of HMGA2 (10, 27); thus, we hypothesized that it might indirectly modulate IMP2 levels. This was confirmed by inhibition or overexpression let-7b with the expected alterations in HMGA2 and IMP2. It is also possible that IMP2 is a direct target of let-7b. Finally, we must confirm that manipulations in let-7b expression correspond to changes in LAMB2, a finding that will require stable expression of the mimic or inhibitor to allow changes in Lamb2 translation and turnover that can establish the consequence of changes in let-7b. Nonetheless, let-7b is a candidate in the circuitry modulating LAMB2 expression.
Based on our data, we propose that LAMB2 protein levels are regulated by IMP2, HMGA2, and let-7b as illustrated in Fig. 8. During podocyte differentiation, levels of let-7b are reduced, allowing expression of HMGA2 to promote the transcription of Imp2. IMP2 is then available to bind Lamb2 mRNA, targeting Lamb2 mRNA to the cytoskeleton where it is translated. There is a growing recognition that the cytoskeleton binds to and provides scaffolding for translational components (25, 35, 38). This “supramolecular organization” (25) is thought to determine subcellular localization of protein synthesis and to regulate mRNA stability (12, 41). Our work showing that HMW polysome fractions contain a complex of IMP2, Lamb2 mRNA, actin, and filamin A provides another example whereby a physical link has been established between intracellular cytoskeleton and translational components. This raises the possibility that signal transduction cascades that are activated by perturbations of the cytoskeleton may also regulate activation of LAMB2 synthesis locally.
Under HG conditions, let-7b levels are increased which reduces HMGA2 and, in turn, IMP2 levels. This results in the reduction of LAMB2 synthesis. Because levels of LAMB2 protein are critical to podocyte function (13, 23, 28, 30, 43, 45), understating mechanisms that modulate LAMB2 protein production provides new, potential therapeutic targets. A primary goal of our work is to apply such work to DN. Correlations between in vitro studies and findings in diabetic rats and that miRs are now recognized as potential therapeutic targets whose selective modulation may alter the course of disease (18, 20, 42) support further studies to test the role of let-7b in DN.
This material is based on work supported (in part) by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Biomedical Laboratory Research and Development, and it was supported with resources and the use of facilities at the Veterans Affairs Puget Sound Health Care System, Seattle, WA. These studies were also supported by a Merit Review award from the Medical Research Service of the Department of Veterans Affairs (to C. Abrass) and by National Institutes of Health Grants (R01DK068539: to C. K. Abrass; DP3 DK094311 and P01 HL092969: to R. C. LeBoeuf; P&F award from the Diabetic and Endocrine Research Center P30DK017047: to V. Schaeffer). V. Schaeffer is also supported by a P&F award from the Core Center of Excellence in Hematology (P30DK56465).
No conflicts of interest, financial or otherwise, are declared by the author(s). The contents of this report do not represent the views of the Department of Veterans Affairs or the United States Government.
Author contributions: V.S., D.R.M., and C.K.A. conception and design of research; V.S., K.M.H., and C.K.A. performed experiments; V.S., D.R.M., and C.K.A. analyzed data; V.S., D.R.M., and C.K.A. interpreted results of experiments; V.S., K.M.H., and C.K.A. prepared figures; V.S., R.C.L., and C.K.A. drafted manuscript; V.S., K.M.H., D.R.M., R.C.L., and C.K.A. edited and revised manuscript; V.S., D.R.M., R.C.L., and C.K.A. approved final version of manuscript.
The authors gratefully acknowledge the technical assistance of Dr. Dusty Miller for testing the podocyte cell line for retroviral infection.