Transgenic Expression of Human LAMA5 Suppresses Murine Lama5 mRNA and Laminin α5 Protein Deposition

Laminin α5 is required for kidney glomerular basement membrane (GBM) assembly, and mice with targeted deletions of the Lama5 gene fail to form glomeruli. As a tool to begin to understand factors regulating the expression of the LAMA5 gene, we generated transgenic mice carrying the human LAMA5 locus in a bacterial artificial chromosome. These mice deposited human laminin α5 protein into basement membranes in heart, liver, spleen and kidney. Here, we characterized two lines of transgenics; Line 13 expressed ∼6 times more LAMA5 than Line 25. Mice from both lines were healthy, and kidney function and morphology were normal. Examination of developing glomeruli from fetal LAMA5 transgenics showed that the human transgene was expressed at the correct stage of glomerular development, and deposited into the nascent GBM simultaneously with mouse laminin α5. Expression of human LAMA5 did not affect the timing of the mouse laminin α1–α5 isoform switch, or that for mouse laminin β1–β2. Immunoelectron microscopy showed that human laminin α5 originated in both glomerular endothelial cells and podocytes, known to be origins for mouse laminin α5 normally. Notably, in neonatal transgenics expressing the highest levels of human LAMA5, there was a striking reduction of mouse laminin α5 protein in kidney basement membranes compared to wildtype, and significantly lower levels of mouse Lama5 mRNA. This suggests the presence in kidney of a laminin expression monitor, which may be important for regulating the overall production of basement membrane protein.


Introduction
The human kidney contains approximately one million individual nephrons, each beginning with a glomerulus, which is a unique capillary tuft that largely restricts the passage of serum albumin and larger proteins into the primary nephron filtrate. All three layers of the glomerular capillary wall, namely the glomerular endothelial cells, glomerular epithelial cells (podocytes), and an intervening glomerular basement membrane (GBM), are required for maintenance of normal filtration barrier properties [1][2][3]. For example, enzymatic degradation of glycosaminoglycans within the glomerular endothelial surface glycocalyx results in an increased fractional clearance for albumin [4]. Additionally, blockage of podocyte-derived VEGF signaling causes glomerular endothelial cell abnormalities in developing or mature kidneys and proteinuria [5,6].
A host of defects that affect the podocyte and its specialized intercellular junction, the epithelial slit diaphragm, also cause abnormal glomerular permeabilities [1][2][3]. These include mutations in the NPHS1 gene encoding the slit diaphragm component, nephrin, which causes congenital nephrotic syndrome of the Finnish type and results in massive proteinuria at birth [7]. Mutations to NPHS2, which encodes another slit diaphragm protein, podocin, also causes proteinuria in autosomal recessive steroid-resistant nephrotic syndrome, a disease often diagnosed in childhood [8]. Intracellularly, podocin and nephrin are both linked indirectly to the actin cytoskeleton through interaction with CD2-associated protein (CD2ap) [9,10]. Slit diaphragms are absent in mice that lack nephrin [11] or podocin [12] and these animals die perinatally with renal failure. Mice deficient in CD2ap also die from renal failure, but at 6-7 weeks of age [10].
The reasons why the GBM collagen IV and laminin composition changes during development are not fully understood, but evidence indicates that this is necessary for final glomerular maturation and full acquisition and maintenance of filtration barrier properties. Alport disease, which is a familial nephropathy marked by focal splitting, thinning, and regional thickening of the GBM and leads to renal failure, is caused by mutations in either the COL4A3, COL4A4, or COL4A5 genes encoding the collagen a3(IV), a4(IV), and a5(IV) protein chains, respectively [18,19]. Most Alport patients fail to assemble a stable network of collagen a3a4a5(IV) in the GBM, and there is retention of the infantile, collagen a1a2a1(IV) network. This isoform appears to be more  susceptible to proteolysis, which may explain why the GBMs of Alport patients ultimately deteriorate [19]. A model of Alport disease has been created in mice through the deletion of the Col4a3 gene [20][21][22], and these animals die of renal failure 2-4 months after birth with the same glomerular defects as those seen in Alport patients. The mouse Alport phenotype can be rescued when transgenic mice expressing human COL4A3-COL4A4 genes are crossed onto the mouse Col4a3 knockout background [23].
Failure to undergo laminin isoform transitioning from LN-111 to LN-521 also results in kidney malfunction in mice and in humans. Although normal glomerular development is seen in mice with laminin b2 deficiencies, they eventually exhibit podocyte foot process broadening, proteinuria, and die of renal failure [24]. Humans with mutations in the LAMB2 gene suffer from Pierson syndrome, which usually presents at birth as congenital nephrotic syndrome with severe neuromuscular junction abnormalities (owing to the presence of laminin b2 in the neuromuscular junction basement membrane as well) [25].
There are no human mutations described for LAMA5, but experiments in mice have shown its expression to be absolutely crucial for normal glomerular development and function. Mice with deletions of Lama5 die before birth with neural tube closure defects and placental dysmorphogenesis [26]. In kidney, a stable GBM fails to assemble, and endothelial cells do not form vascularized glomerular tufts [27]. This Lama5 knockout phenotype can be partially rescued when fetal kidneys from Lama5 mutants are grafted into newborn kidneys of normal, wildtype hosts [28]. In this case, host endothelial cells, which express laminin a5, migrate into the engrafted Lama5 null kidneys and vascularized glomeruli form within grafts. The host endothelial cell-derived laminin a5 does not project across the full width of these GBMs, however. This results in an unusual situation where there is retention of the infantile laminin a1 on the outer, subpodocyte layer of matrix and laminin a5 is present only on the inner, subendothelial layer. Additionally, these hybrid GBMs are abnormally wide and not as well condensed as normal GBM, and podocyte foot processes are absent [28]. In other experiments, deletion of Lama5 only in podocytes results in mild to severe proteinuria, and variable defects in GBM and podocyte ultrastructure [29]. In this same study, expression of a human LAMA5 transgene under control of a doxycyclin inducible, podocytespecific expression system rescues glomerular and tubular defects caused by a hypomorphic Lama5 mutation [29].
Taken together, these findings demonstrate that the timely expression of LN-521 is needed for glomerular endothelial cell and podocyte differentiation, and the appearance of collagen a3a4a5(IV) is required for long term GBM stability. However, very little is known at the gene level regarding activation of any of the mature GBM protein isoforms, what silences transcription of infantile chain genes at the appropriate developmental stages, and how the infantile collagen a1a2a1(IV) and LN-111 networks are removed from developing GBM. In addition, we do not understand what causes upregulation of Lama5 in Col4a3 knockout (Alport) mice, which may be an important contributor to fibrosis in that model [30].
To begin addressing some of these questions, we have developed bacterial artificial chromosome (BAC) transgenic mice expressing human LAMA5. These transgenics deposited apparently large amounts of human laminin a5 protein in basement membranes widely, and, specifically in glomeruli, at the appropriate developmental stage. Expression of human LAMA5 did not appear harmful and kidney functional tests and morphology were normal. The results suggest that the BAC used for transgenic injections contained all of the necessary regulatory information for proper LAMA5 expression. Of great interest, in kidneys from lines with the highest levels of human LAMA5 expression, there were significant decreases in native mouse Lama5 mRNA and mouse laminin a5 protein deposition.

Generation of human chromsome 20 BAC transgenics
All experiments with mice strictly followed policies and procedures established by the Animal Welfare Act and the Public Health Service Policy on the Humane Care and Use of Laboratory Animals. Human BAC clone RP11-1023E23, containing ,189 kB of human chromosome 20, was obtained from Empire Genomics (Buffalo, NY). Bacteria were grown in 15 mg/ml chloramphenicol overnight at 33uC, then BAC DNA was purified using a Nucleobond kit (Clontech, Mountain View, CA). The clone was verified by end-sequencing, polymerase chain reaction and restriction enzyme digestion followed by pulsed field gel electrophoresis. DNA was injected into FvB6C57Bl/6 F1 hybrid oocytes by personnel from the Transgenic and Gene Targeting Institutional Facility at the University of Kansas Medical Center. To screen founders, human-specific primers were designed to intronic regions of all five genes residing along the BAC clone: OSBPL2, ADRM1, LAMA5, RPS21 and C20orf151 using the Roche Universal Probe Library (Roche Applied Science, Indianapolis, IN). Primers were tested for specificity using mouse and human genomic DNA (Table S1). Genomic DNA from each founder was purified using Qiagen tissue kits (Valencia, CA). Founders that amplified all 5 human gene products were mated to wildtype C57Bl/6 mice, and several stable lines were established. Colonies from two of these lines, which expressed high and moderate levels of human LAMA5 (Lines 13 and 25, respectively), were maintained by mating BAC transgenic heterozygotes with wildtype C57Bl/6 mice.

Verification of human LAMA5 expression
Transgenic N1 BAC progeny, at various ages, were killed and heart, liver, spleen and kidney tissues were promptly harvested, surrounded by Tissue Tek O.C.T. Compound (Electron Microscopy Sciences, Fort Washington, PA), and frozen immediately in isopentane chilled in a dry ice-acetone bath. Cryostat sections, five mm thick, were labeled with mouse anti-human laminin a5 Figure 3. Characterization of two different human LAMA5 BAC transgenic lines. A-D: Fresh frozen kidney sections labeled with mouse antihuman laminin a5 antibody, followed by anti-mouse IgG conjugated to Alexa Fluor-594. Sections from Line 13 and Line 25 were imaged by routine immunofluorescence microscopy (A and B), and by scanning confocal microscopy (C and D), using the same exposure settings. Linear basement membrane labeling of anti-human laminin a5 is seen in both line 13 (A and C) and 25 (B and D), but fluorescent signal appears brighter in Line 13. Magnification (A and B): 2006; scale bar=50 mm. Magnification (C and D): 6306; scale bar=20 mm. E: Quantification of glomerular immunofluorescence intensities shows significantly higher expression of human laminin a5 in GBMs of Line 13 mice, *** p,0.0001. F: Relative transgene copy number estimates were made using cycle threshold from quantitative real time PCR of human LAMA5 genomic primers normalized to thresholds obtained with mouse hemoglobin A (Hba) primers. Compared to Line 25, Line 13 has more than 6 times as many copies of LAMA5, **p = 0.0013. G: Whole kidney total RNA from Line 13 or Line 25 (n = 3 samples per line) was amplified with human LAMA5 intron-spanning primers, relative to PPIA (cyclophilin). Considerably more LAMA5 mRNA is detected in Line 13, *** p,0.0001. doi:10.1371/journal.pone.0023926.g003 antibody (1:500 dilution, clone 4C7, Millipore, Billerica, MA) or rabbit anti-mouse laminin a5 antibody (1:200 dilution; antibody was a kind gift from Dr Jeffery Miner, Washington University, St Louis, MO). Appropriate, species specific secondary antibodies conjugated to Alexa Fluor dyes were used at a 1:200 dilution (Invitrogen/Molecular Probes, Eugene, OR). In some cases, sections were dually labeled with monoclonal rat anti-mouse laminin a1 or b1 IgGs (50 mg/ml) [31] and rabbit anti-mouse laminin a5 (1:200) or rabbit anti-mouse laminin b2 (1:2,000) (from Dr. Jeffery Miner). Slides were mounted using Prolong Gold plus DAPI (Molecular Probes). Sections were viewed and imaged by standard epifluorescence on a Leica SM5000B microscope (Bannockburn, IL). Slides were also examined with a Zeiss LSM 510 scanning laser confocal microscope (Thornwood, NY) and images were captured at an optical section thickness of 0.2 mm. qPCR At the time of sacrifice, a portion of kidney tissue was collected in RNAlater (Qiagen) and stored at 280uC. Total RNA was purified using a RNeasy Mini kit (Qiagen), incubated with primers designed to hybridize specifically with human LAMA5, mouse Lama5, Lamb1, and Lamc1 RNA (Table S1) and Quantitect SYBR Green RT-PCR reagents. Products were amplified and quantified in an iCycler (BioRad, Hercules, CA). Relative RNA abundance was determined using the comparative Ct method [32]. Normal human kidney total RNA served as a reference and was purchased from Clontech Laboratories, Inc. (Mountain View, CA).

Estimate of BAC LAMA5 copy number
A standard curve was prepared containing between 1 and 128 copies of human BAC clone RP11-1023E23 in 5 nanograms of wildtype mouse genomic DNA. Genomic DNA was isolated from transgenic liver tissue from both Line 13 and Line 25 using DNeasy Blood and Tissue kit (Qiagen). The standard curve and genomic DNA samples were amplified with human LAMA5 genomic primers and mouse hemoglobin alpha primers (Table S1) with the RT 2 SYBR Green/Fluorescein qPCR Master Mix (SABiosciences, Frederick, MD). Relative abundance was estimated using the comparative Ct method [32].

Blood and urine chemistries
Blood urea nitrogen was measured from serum samples using the QuantiChrom Urea Assay kit (BioAsssay Systems, Hayward, CA). Urinary albumin was measured using an enzyme-linked immunosorbent assay mouse Albuwell kit (Exocell, Philadelphia, PA). In some cases, urine was first resolved on polyacrylamide gels, which were then stained with Coomassie Blue.

Electron microscopy
For routine electron microscopy, kidneys were fixed and processed as described [33] and imaged on a JEOL JEM-1400 transmission electron microscope. For postfixation immunoelectron microscopy, 2-mm wedges of kidney cortices were fixed with 1% paraformaldehyde and 0.05% glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.3, for 1.5 hours on ice. Tissues were washed, equilibrated with 30% sucrose in buffer, and snap frozen in tissue freezing medium (Triangle Biomedical Sciences, Durham, NC) by using isopentane chilled in a dry ice-acetone bath. Frozen sections (30 mm thick) were collected on Thermanox coverslips (Miles Laboratories, Inc., Naperville, IL), and then air dried at room temperature. Sections were blocked for 30 minutes each in 0.5 M ammonium chloride in PBS and then with 5% goat serum and 0.1% bovine serum albumin in PBS. Sections were then immunolabeled with anti-human laminin a5 clone 4C7 ascites fluid (diluted 1:50 in PBS) for 1 hour, and washed with PBS. Sections were then incubated with rabbit antimouse IgG 2a -HRP (MP Biomedicals, Solon, OH; 50 mg/ml) for 1 hour, washed, refixed in Karnovsky's fixative, developed for peroxidase histochemistry, and processed for electron microscopy as described previously [17].

Immunoprecipitation and Western blotting
Kidneys were harvested from postnatal day 0-4 (P0-4) mice (human BAC LAMA5 Line 13 transgenic and wildtype littermate controls), frozen in liquid nitrogen, and stored at 280uC. Kidneys were dounce homogenized in lysis buffer (10 mM Tris, pH 7.5, 150 mM NaCl, 20 mM beta-glycerol phosphate, 5 mM ethlyenediaminetetraacetic acid, 10% glycerol, 2 mM sodium fluoride, 0.5% Igepal, 16 protease inhibitor cocktail [Sigma-Aldrich, St Louis, MO]), sheared five times by passage through a syringe fitted with a 21 gauge needle, extracted in lysis buffer for 2 hours at 4uC, and centrifuged 10 minutes, 14,000 g at 4uC. Protein content of the supernatant was determined using a DC Protein Assay Kit I (BioRad, Hercules, CA), and the concentration was adjusted to 2.2 mg/ml with lysis buffer. Ten microliters of anti-human laminin a5 mouse monoclonal 4C7 IgG was incubated overnight with the supernatant. To recover 4C7, Fastflow protein G sepharose beads (Protein G Sepharose Fastflow, GE Healthcare, Piscataway, NJ) were added and the solutions were gently rotated for 4 hours at 4uC. Beads were then washed three times with lysis buffer, and three times with tris-buffered saline. Beads were boiled in 26 SDS sample buffer (containing dithiothreitol) and stored at 280uC.

Identification of human LAMA5 BAC transgenics
Transgenic mice were established from 25 progeny born from zygotes injected with BAC DNA RP11-1023E23 (position 60273754-60463260 of human chromosome 20). Genomic DNA was screened with human-specific primers designed to intronic regions of the five genes found in this region of human chromosome 20 (Table S1). Here we characterize two stable transgenic lines that expressed all five human genes, indicating integration and germ line transmission of the complete BAC DNA.
To confirm transcription and translation of the human LAMA5 gene in the BAC transgenic mice, frozen sections of kidney from transgenic and wildtype siblings were incubated with monoclonal mouse anti-human laminin a5 IgG and fluorochrome-conjugated secondary antibody. As shown in Figure 1, glomerular and tubular basement membranes from the BAC transgenics were immunolabeled in a bright, linear pattern (Fig. 1A), whereas there was no labeling of sections from wildtype controls (Fig. 1B). In addition, sections of heart, liver, and spleen from BAC transgenics all demonstrated linear basement membrane labeling with antihuman laminin a5 ( Figs. 2A-C). Quantification of human LAMA5 expression Although basement membranes in multiple organs from BAC transgenics were positive for human laminin a5, the intensity of immunofluorescence signals varied between the two different transgenic lines. For example, kidney glomerular and tubular basement membrane immunofluorescence in Line 13 appeared to be more intense than in Line 25 (Figs. 3A-D). When glomerular immunolabeling profiles were quantified by pixel intensity measurements of confocal microscope images, Line 13 glomeruli were more than 6 fold brighter than those in Line 25 (Fig. 3E). We interpreted the increased immunofluorescence signals in Line 13 to signify more human laminin a5 protein deposition.
To estimate relative gene copy number in the human BAC transgenics, a standard curve was amplified from human clone RP11-1023E23 DNA in 2-fold serial dilution (representing from 1 to 128 copies) spiked into 5 nanograms of mouse genomic DNA. Genomic liver DNA was purified from N2 progeny of Lines 13 and 25 and amplified using human-specific LAMA5 primers, as well as mouse hemoglobin alpha (Hba) primers as a control (Table  S1). Line 13 was estimated to have 6 times as many copies as line 25 (Fig. 3F). Similarly, mRNA expression levels were dramatically different between line 13 and 25, with line 13 expressing approximately 18 fold greater LAMA5 message than Line 25 (Fig. 3G). These quantitative results suggest that the higher immunofluorescence signal seen in line 13 glomeruli was likely due to the higher LAMA5 copy number and gene expression.

Normal kidney function in human LAMA5 BAC transgenics
In both transgenic lines, integration of the human LAMA5 BAC into the mouse genome did not result in any overt phenotype for more than 8 months after birth. Heterozygous transgenic animals were active and fertile. Urine obtained from transgenic human LAMA5 BAC mice at 4, 8, 20 and 24 weeks of age was separated by polyacrylamide gel electrophoresis (PAGE) alongside albumin standards, and stained with Coomassie Blue. There was no evidence of urinary albumin by PAGE in any of these samples (not shown). For further verification, albumin was quantified in urine from a total of 8 wildtype and 12 Line 13 transgenic mice at 4 weeks of age using a mouse albumin enzyme-linked immunosorbent assay. There were no significant differences in urinary albumin (wildtype = 6.7 mg/mL, BAC transgenic = 11.7 mg/mL, p = 0.111). Albumin levels quantified by enzyme-linked immunosorbent assay were also within normal limits in urine from transgenic mice at 8 weeks and 20 weeks of age. There were also no differences in blood Consistent with the renal function data, normal kidney histology was seen in transgenic human LAMA5 BAC samples ranging in age from 3 days to 8 weeks old (Fig. 4A). The ultrastructure of glomeruli from 8 week old Line 13 transgenic mice also appeared entirely normal and displayed open capillary loops, fenestrated endothelium, GBMs of uniform thickness and density, and regular podocyte foot process registration (Figs. 4B and 4C). Similarly, there were no structural defects observed in kidney tubules or vasculature.
We next asked whether expression of human laminin a5 protein in developing glomeruli followed the same temporal and spatial patterns observed for the intrinsic mouse protein. During normal glomerulogenesis, laminins containing the a5 chain (LN-511 and LN-521) are first detected in the vascular clefts of comma-and Sshaped nephrons, where they replace the LN-111 isoform [14]. Subsequently, LN-521 is the only laminin isoform deposited into the GBM during glomerular maturation [13,14]. To determine if the normal sequence of laminin a5 synthesis was occurring in the human LAMA5 BAC transgenics, kidney sections from Line 13 newborn mice underwent double immunolabeling with antimouse laminin a1 and anti-human laminin a5 IgGs. As seen in Figure 5A, developing GBM within the vascular cleft of commashaped nephric figures contained mainly the endogenous mouse laminin a1, with only trace amounts of human laminin a5 of BAC origin (Figs. 5A-C arrowheads). At slightly later stages of glomerular development (S-shaped), more laminin a5 of both mouse and BAC origin was evident in the forming GBM (Figs. 5D-F, arrows). This signifies that the developmental expression of human LAMA5 paralleled that for its murine homolog, and that both mouse and human laminin a5 chains were deposited concurrently within the same basement membrane segments.
We also wondered whether the expression of human LAMA5 might have affected the temporal isoform switching schedule of endogenous mouse laminin [13][14]. As shown in Figures  To identify cellular sites of biosynthesis for human laminin a5, sections of lightly fixed BAC transgenic kidneys from 3 day old mice were treated with anti-human laminin a5 IgG and processed for immunoperoxidase electron microscopy. In glomeruli, peroxidase reaction product was seen intracellularly within endothelial cells and podocytes (Fig. 7A), which have previously been shown to be origins of mouse laminin a5 in developing glomeruli [16]. Similarly, intracellular vesicles within tubular epithelial cells were also labeled with anti-human laminin a5 in developing kidneys of BAC transgenics (Fig. 7B).
Human laminin a5 protein binds mouse b1 and c1 chains Extraction of proteins from intact basement membranes has been notoriously difficult, making biochemical analyses of these matrices challenging. However, because we detected human laminin a5 protein within the intracellular biosynthetic pathways of glomerular endothelial cells and podocytes, as well as tubule epithelial cells in developing kidney, we reasoned that we should be able to solubilize intracellular laminin and determine whether it contained a heterotrimer of human and mouse chains. For these experiments, anti-human laminin a5 IgG was used to immunoprecipitate laminin from cell lysates of 2 day old BAC transgenic and wildtype kidneys. When immunoprecipitates from BAC transgenics were probed on Western blots with anti-mouse laminin b1 and c1 chains, both were shown to be present (Fig. 8A). Although these immunoprecipitation experiments with whole kidney lysates could not define which nephron segment(s) contained human laminin a5-mouse b1-mouse c1 heterotrimers, earlier evidence shows that developing GBMs of immature glomeruli contain LN-511, and this same laminin isoform is particularly abundant in immature and mature tubular basement membranes [14,34]. In contrast to these findings from BAC transgenic lysates, kidney lysates from wildtype mice that eluted from anti-human laminin a5 were negative (Fig. 8A). Cell lysates from both wildtype and BAC transgenic kidneys that were not subjected to anti-human immunoprecipitation also contained laminin b1 and c1 chains, as expected (Fig. 8A). As mentioned earlier, the only laminin isoform in fully mature GBM is LN-521. To determine whether human laminin a5-mouse b2-mouse c1 heterotrimers were also present in developing transgenic kidney lysates, anti-human laminin a5 immunoprecipitates were probed on Western blots with anti-mouse laminin b2. The results from these experiments thus far were negative. However, glomeruli represent only a small fraction of total kidney mass, and our negative result could have been due to insufficient sample within the total kidney lysate contributed specifically by maturing glomeruli. Because we saw what appeared to be large amounts of human laminin a5 in developing and mature glomeruli (Figs. 1, 3, and 5) we think that it is probable that human laminin a5-mouse b2-mouse c1 heterotrimers are present within GBMs of the BAC transgenics. Further evidence in support of this came from double label immunofluorescence microscopy of adult transgenic glomeruli where, in most capillary loops, there was linear co-localization of anti-mouse laminin b2 and antihuman laminin a5 antibodies within GBMs (Figs. 8B-D).

Expression of human LAMA5 suppresses mouse Lama5
We next wondered whether the expression of apparently abundant human laminin a5 in Line 13 kidney affected expression of native mouse laminin. For these experiments, frozen sections from newborn wildtype and Line 13 BAC transgenic kidneys were labeled with anti-mouse laminin a5 and examined by immunofluorescence microscopy. In sections from wildtype kidney, bright linear immunolabeling for mouse laminin a5 was seen throughout glomerular and tubular basement membranes (Fig. 9A). By comparison, however, sections of human LAMA5 BAC transgenic kidney showed an obvious and marked reduction in intensity of immunolabel for mouse laminin a5 (Fig. 9B), and glomeruli and tubules were equally affected. An earlier study in which doxycycline-inducible human LAMA5 is expressed specifically in podocytes also noted what appeared to be a reduction of mouse laminin a5 within the GBM [29], but this observation was not pursued further.
To begin to investigate mechanisms accounting for the reduction in mouse laminin a5 protein, RNA was isolated from kidneys of 3 day old and 5 week old wildtype and Line 13 transgenic mice. Quantitative RT-PCR using mouse-specific primers showed that relative Lama5 mRNA levels were significantly higher in 3 day old kidneys than at 5 weeks of age, and this was true for both wildtype and human LAMA5 BAC transgenics (Fig. 9C). Similarly, the relative amount of LAMA5 mRNA in 3 day old BAC transgenics was significantly higher than at 5 weeks, and the fold decrease at 5 weeks was approximately the same as that for mouse Lama5 (Fig. 9D). We interpret these findings to reflect a burst of basement membrane assembly that accompanies the rapid induction and elongation of nephrons that occurs during kidney development. This would require more message for basement membrane proteins in kidneys from 3 day olds than at 5 weeks of age, when kidneys have reached their nearly mature size and basement membrane assembly has essentially concluded.
Our results from 3 day old, Line 13 LAMA5 BAC transgenic kidneys also showed less mouse Lama5 mRNA when compared to wildtype (Fig. 9C). Although the reduction in mouse Lama5 mRNA in transgenics was not as striking as what was seen at the protein level by immunofluorescence, it nevertheless was statistically significant (Fig. 9C). In addition, the loss was specific for Lama5 as qRT-PCR of Line 13 and wildtype kidneys showed no Figure 8. Human laminin a5 forms heterotrimers with mouse b1 and c1 chains and co-localizes with mouse laminin b2 in GBMs. A: Postnatal day 2 kidneys were harvested from wildtype (Wt) or human LAMA5 BAC transgenic littermates (Tg). Lysates were incubated with antihuman LAMA5 antibody 4C7, and recovered with protein G beads (+). Western blotting with chain-specific anti-laminin b1 (top blot) or anti-laminin c1 (bottom blot) shows that both wildtype and transgenic lysates contain laminin b1 and laminin c1. Lysates from wildtype kidneys immunoprecipitated with anti-human laminin a5 4C7 antibody do not contain mouse laminin b1 or c1, but both chains are present in immunoprecipitates from transgenic kidney. B-D: Double label immunofluorescence microscopy of fully mature LAMA5 transgenic glomeruli shows widespread co-localization of mouse laminin b2 and human laminin a5 in GBMs (arrows). Magnification: 6006; scale bar = 20 mm. doi:10.1371/journal.pone.0023926.g008 differences in expression of Lama1, Lamb1, Lamc1, or Col4a1 mRNAs. Moreover, there were no differences in Lama5 message between wildtype mice and Line 25 BAC transgenics, which expressed substantially lower amounts of human LAMA5 than Line 13, and the intensity of anti-mouse laminin a5 immunolabeling also appeared closely similar between Line 25 and wildtypes.
What mechanism(s) could account for the suppression of mouse Lama5 mRNA and reduced mouse laminin a5 protein deposition in the human LAMA5 BAC transgenics? Clearly, the transgenics transcribed both the human LAMA5 and mouse Lama5 genes, and, as shown in Figures 5D-F, laminin protein from both species was deposited in the same kidney basement membranes. Perhaps the reduction in mouse Lama5 message seen in the LAMA5 BAC transgenics was the result of a feedback mechanism in which the total amount of laminin a5 chain protein (mouse and human) somehow subdued the tempo of Lama5 gene transcription. Alternatively, there may have been competition between the human LAMA5 and mouse Lama5 genes for the same transcription factor(s). On the other hand, there may have been loss of Lama5 mRNA stability or other post-transcriptional changes in the presence of excess LAMA5 message.
Earlier work in cell culture has shown that the laminin a1 chain can be secreted separately from its b1 and c1 subunit partners, but that most of the secreted a chain monomer undergoes proteolytic cleavage [35]. In contrast, the laminin b1 and c1 subunits are not secreted in the absence of a1, and accumulate intracellularly [35]. As already mentioned, qRT-PCR of Line 13 and wildtype kidney samples showed that there were no differences in relative message Figure 9. Human LAMA5 BAC transgenics express less mouse laminin a5 protein and Lama5 mRNA. A-B: Wildtype (A) and Line 13 LAMA5 BAC transgenic (B) littermate kidneys were harvested and frozen sections were labeled with anti-mouse laminin a5. Digital images were captured using same exposure parameters. Compared to wildtype (A), note the marked reduction in immunolabeling for mouse laminin a5 in the transgenics (B). G: glomeruli; T: tubules. Magnification: 806; scale bar = 100 mm. C: Total kidney RNAs from 3 day-(3 d) and 5 week-old (5 wk) wildtype (blue) and Line 13 transgenic (red) mice were amplified using mouse Lama5 primers normalized to PPIA (cyclophilin). Significantly less mouse Lama5 mRNA expression was seen in 3-day old LAMA5 transgenic mice than in 3-day old wildtypes. Compared to 3-day olds, there was significantly less Lama5 expression at 5 weeks for both wildtypes and transgenics. ANOVA, * p,0.05, ** p,0.01, *** p,0.001. D: Total kidney RNA from 3 day-and 5 week-old Line 13 human LAMA5 BAC transgenics were amplified using human LAMA5 primers. Compared to the 3-day old time point, the fold reduction in expression of human LAMA5 mRNA at 5 weeks was similar to that seen for native mouse Lama5 (C). doi:10.1371/journal.pone.0023926.g009 abundance for Lamb1 or Lamc1, which encode laminin b1 and c1 chains, respectively, and which we showed were binding partners for human laminin a5 (Fig. 8). Perhaps the secretion of a stable human laminin a5-mouse b1-mouse c1 chimeric heterotrimer, and entirely mouse heterotrimers, was rate-limited by the amounts of laminin b and c chains available for heterotrimerization. Nevertheless, how the abundance of laminin a5 protein could autoregulate Lama5 gene transcription is not at all clear, but this could be an important control mechanism that becomes defective in fibrotic conditions where there is overproduction of basement membrane protein.
In summary, we have generated transgenic lines of mice that express human LAMA5 in temporally and spatially correct contexts within kidney, indicating that the appropriate genetic control elements are present. Unexpectedly, a transgenic line expressing the highest amounts of human laminin a5 suppressed mouse Lama5 mRNA and mouse protein deposition. These transgenics may prove useful for understanding regulation of laminin gene expression and provide new clues regarding mechanisms of basement membrane assembly.

Supporting Information
Table S1 Primers complementary to human (capitalized gene symbols) and mouse were designed using the indicated accession numbers as templates, and each pair was given a unique primer designation. Primer sequence and length in basepairs is also shown. (DOC)