Transgenic isolation of skeletal muscle and kidney defects in lamininβ2 mutant mice: implications for Pierson syndrome

Basement membranes (BMs) are thin sheets of specialized extracellular matrix that underlie all epithelial and endothelial cells, surround all muscle fibers, fat cells and Schwann cell/axon units in peripheral nerves, and encase the entire central nervous system. BMs have been shown to influence cell proliferation, differentiation and migration. They are also involved in filtration and tissue compartmentalization and serve as barriers to cancer cells.

All BMs contain members of four protein families: type IV collagen,nidogen/entactin, sulfated proteoglycans and laminin (reviewed by Sasaki et al., 2004). The particular protein isoforms present in a given BM are presumed to impart unique functional properties important for behavior of the adjacent cells and for organ development and function. Perhaps the best evidence for this is the fact that mutations in genes encoding BM proteins cause human disease, despite substitution by related isoforms in some cases. For example, Alport syndrome is a type IV collagen disease of kidney and inner ear(Kashtan, 2004); Knobloch syndrome is a collagen XVIII (heparan sulfate proteoglycan) disease affecting the eye (Suzuki et al., 2002);and Herlitz's junctional epidermolysis bullosa and congenital muscular dystrophy type 1A are laminin diseases of the skin and neuromuscular system,respectively (Burgeson and Christiano,1997; Patton,2000).

Laminins exist in BMs as α-β-γ heterotrimers. In mammals there are five α, four β and three γ chains that assemble nonrandomly to form at least 15 different heterotrimers [see Aumailley et al.(Aumailley et al., 2005) for a new laminin nomenclature] (reviewed by Miner and Yurchenco, 2004). Evidence from zebrafish supports the existence of a fourth β chain(Parsons et al., 2002). Mutations in ten of the eleven known mouse laminin genes have been reported. These studies have shown that laminin is absolutely required for BM formation(Miner et al., 2004; Smyth et al., 1999) and that laminins play roles in diverse developmental and physiological processes(reviewed by Miner and Yurchenco,2004; Yurchenco et al.,2004).

Recently, mutations in the laminin β2 gene (LAMB2) were reported in a rare, lethal human disease called Pierson syndrome(Zenker et al., 2004a; Zenker et al., 2005). Pierson syndrome, also called microcoria-congenital nephrosis syndrome (OMIM 609049),is an autosomal recessive disease with congenital nephrotic syndrome/diffuse mesangial sclerosis, distinct ocular abnormalities including microcoria (small pupil), and impairment of vision and neurodevelopment(Zenker et al., 2004b). These features are consistent with the phenotype of mice lacking laminin β2,which stop gaining weight at 1 week of age and die at ∼3 weeks of age. Lamb2^-/- mice exhibit heavy proteinuria/congenital nephrotic syndrome with podocyte foot process effacement(Noakes et al., 1995b),aberrantly formed and functionally impaired neuromuscular junctions(Knight et al., 2003; Nishimune et al., 2004; Noakes et al., 1995a; Patton et al., 1998), and both structural and functional abnormalities in the retina(Libby et al., 1999). These phenotypes reflect the fact that laminin β2 has been reported to be concentrated in the kidney glomerulus and at the skeletal neuromuscular junction (Hunter et al., 1989)and to be deposited in the retina (Libby et al., 2000; Libby et al.,1997).

Although the eye abnormalities would not be expected to impair the growth and longevity of Lamb2^-/- mice, the glomerular and neuromuscular defects probably contribute to the severity of the failure to thrive phenotype. However, it has been difficult to determine the extent to which each of these is involved. Moreover, laminin β2 is expressed at other sites quite widely throughout the body(Sasaki et al., 2002), so there may exist defects in other tissues that are being masked by the severity of the kidney and muscle defects. To investigate these issues, which have relevance in terms of understanding both basic laminin biology and Pierson syndrome, we have used tissue-specific laminin β2 transgenes to isolate the neuromuscular and kidney defects, rescuing each one individually and both together by mating the transgenes onto the Lamb2^-/-background. A similar approach was previously used to isolate skeletal muscle defects from those in peripheral nerve in Lama2 mutant mice(Kuang et al., 1998). Our results suggest that the neuromuscular defect, rather than the kidney defect,is responsible for the severe overt phenotype. Furthermore, a requirement for laminin β2 in tissues other than muscle and kidney to maintain normal viability could not be demonstrated.

To construct the MCK-B2 transgene, the full-length rat laminin β2 cDNA(Green et al., 1992) (a gift from Joshua R. Sanes, Harvard University, Cambridge, MA) was inserted between the 3.3 kb mouse muscle creatine kinase (MCK) enhancer/promoter and the SV40 large T-antigen transcription termination and polyadenylation signal sequences(Jaynes et al., 1988) (a gift from Jean Buskin and Stephen Hauschka, University of Washington, Seattle, WA). To construct the NEPH-B2 transgene, the rat β2 cDNA plus SV40 3′processing signals from MCK-B2 were flanked with loxP sites and inserted downstream of the 4.2 kb mouse nephrin promoter(Eremina et al., 2002) (a gift from Susan Quaggin, University of Toronto, Toronto, Canada).

Production of mice carrying the targeted Lamb2 mutation (gift from Joshua R. Sanes) has been previously described(Noakes et al., 1995a). Briefly, the MC1neo cassette was inserted into the AgeI site in the second exon of mouse Lamb2. Antibodies to epitopes encoded by mRNA sequences 3′ of this insertion did not stain homozygous mutants(Noakes et al., 1995b; Sasaki et al., 2002),indicating that the insertion generates a null mutation. Mice were genotyped by PCR of DNA extracted from tail clips using the following primer pairs: for the mutant allele (MC1neo-specific),5′-CGAATTCGAACACGCAGATGCAG-3′ and 5′-CCGGGCGCCCCTGCGCTGACAGC-3′; for the wild-type allele,5′-TGACCCACTGTCCTCAGTGCTG-3′ and 5′-GAGTGTAGGATAGGTACCTTAG-3′.

To produce transgenic mice, the transgenes were digested and gel-purified away from plasmid vector sequences and then microinjected individually into the pronuclei of single-celled B6CBAF2/J embryos by the Washington University Mouse Genetics Core. Founders and their transgenic offspring were identified by PCR of DNA extracted from tail clips using the following primers: for MCK-B2, 5′-CTGGCTAGTCACACCCTGTAGG-3′ (MCK forward) and 5′-CTGGATAGCAGCTTCCTCGAG-3′ (β2 reverse); for NEPH-B2,5′-GAAGCAGCAGAATGAGTTCACAC-3′ (nephrin forward) and 5′-ATACGAAGTTATTCGAAGTCGAG-3′ (vector/loxP sequences between the nephrin promoter and the β2 cDNA).

Mouse mAbs D5 and D7 that recognize the rat laminin β2 C-terminal coiled-coil domain (Hunter et al.,1989; Sanes et al.,1990) were gifts from Joshua R. Sanes and were also obtained from the Developmental Studies Hybridoma Bank (Iowa City, IA). Polyclonal antisera to the mouse laminin β2 LF domain (formerly called domain IV)(Sasaki et al., 2002), which crossreact with rat β2, were a gift from Takako Sasaki and the late Rupert Timpl (Max Planck Institute for Biochemistry, Martinsried, Germany). Rat mAb 5A2 to the mouse laminin β1 LN domain (formerly VI)(Abrahamson et al., 1989; St John et al., 2001) was a gift from Dale Abrahamson (University of Kansas Medical Center, Kansas City,KS). Rabbit antiserum 8948 to the mouse laminin α5 LEb and L4b domains(formerly IIIb and IVa) has been described(Miner et al., 1997). Alexa 488- and Cy3-conjugated secondary antibodies were obtained from Molecular Probes (Eugene, OR) and Chemicon (Temecula, CA), respectively.

For immunostaining, fresh tissues were immersed in OCT and quick-frozen in 2-methylbutane cooled in a dry ice-ethanol bath. Frozen sections were cut (7μm) in a cryostat and air dried on gelatin-coated slides. Antibodies were diluted in PBS with 1% BSA, applied to sections for 1 hour, and rinsed in PBS. Secondary antibodies were then applied in a similar fashion. TRITC-conjugatedα-bungarotoxin (Molecular Probes) was included to localize synaptic sites in skeletal muscle. After rinsing, the sections were mounted in 90%glycerol containing 0.1× PBS and 1 mg/ml p-phenylenediamine and viewed under epifluorescence with a Nikon Eclipse 800 compound microscope. Images were captured with a Spot 2 cooled color digital camera (Diagnostic Instruments, Sterling Heights, MI).

For conventional light microscopy, tissues were fixed in 10% buffered formalin, dehydrated through graded ethanols, embedded in paraffin, sectioned at 4 μm and stained with periodic-acid-Schiff (PAS). For electron microscopy, tissues were fixed, embedded in plastic, sectioned and stained as previously described (Kikkawa et al.,2003; Noakes et al.,1995b).

To assay the integrity of skeletal muscle fibers, Evans blue dye (Sigma, St Louis, MO) at 10 mg/ml in 0.9% NaCl was injected intraperitoneally at 10μl/g body weight. After 6 hours mice were perfused under anesthesia with 10 ml PBS and then 40 ml of 2% paraformaldehyde in PBS. The diaphragm with rib insertions was dissected out and post-fixed overnight with 3% paraformaldehyde and 1% glutaraldehyde in PBS at 4°C. The diaphragm was then detached from the ribs and photographed with a QImaging Micropublisher digital camera(QImaging, Burnaby, BC, Canada) attached to a stereomicroscope.

Urine was collected from mice at various ages either by manual restraint,by caging on raised wire floors (Ancare, Bellmore, NY) for several hours or by bladder puncture under anesthesia at the time of sacrifice. Urine (1 μl)was run on precast 4 to 20% SDS polyacrylamide gels (Invitrogen/Novex,Carlsbad, CA), and gels were stained with Coomassie Blue and destained by standard methods. For quantitation, urinary protein and creatinine concentrations were measured with a Cobas Mira Plus analyzer (Roche,Somerville, NJ).

We wished to isolate the muscle and kidney defects observed in Lamb2^-/- mice so that they could be studied independently of one another. The failure to thrive phenotype of Lamb2^-/- mice is so severe that indirect effects of the kidney and neuromuscular diseases on each other, and perhaps on other tissues,could not be ruled out. One approach to accomplish this would be to make cell type-specific knockouts using Cre/loxP technology. However, we chose to attempt to rescue selectively each defect individually using cell type-specific laminin β2 transgenes expressed on the Lamb2^-/- background.

The reported developmental defects in Lamb2^-/- muscles are restricted to the synapse. Consistent with this, laminin β2 is concentrated in the synaptic region of the myofiber BM; it also accumulates at myotendinous junctions (MTJs) (Patton et al., 1997) and can be detected extrasynaptically with some antibodies (Hunter et al.,1989; Sasaki et al.,2002; Wewer et al.,1997). Previous studies have shown that β2 is expressed by skeletal muscle in vivo (Moscoso et al.,1995) and is concentrated at synapse-like sites that form on cultured myotubes in vitro even in the absence of neurons(Martin et al., 1995). Together, these data suggest that skeletal muscle is capable of concentrating laminin β2 at synaptic sites independent of motoneuron expression ofβ2, should there be any. We therefore constructed a β2 transgene(MCK-B2; Fig. 1A) that should be expressed at high levels in skeletal muscle, using the well-characterized mouse MCK enhancer/promoter. This element has also been shown to drive expression in cardiac muscle with ∼100-fold less activity, but,importantly, it is not significantly active in kidney(Johnson et al., 1989). We chose to use the rat β2 cDNA because of the availability of mouse mAbs that recognize rat but not mouse β2(Sanes et al., 1990). These mAbs facilitated specific detection of transgene-derived β2.

Seven MCK-B2 transgenic founders were generated. Because mouse tail contains multiple tissue types, including some skeletal muscle, transgene expression was assayed by staining frozen sections of a tail biopsy from each founder with anti-rat β2 monoclonal antibodies. Muscle synaptic sites were identified by co-staining with TRITC-conjugated α-bungarotoxin,which binds to acetylcholine receptors. Two of the seven founders expressed rat β2, and it was concentrated at synapses. These two were mated to Lamb2^+/- mice to generate Lamb2^+/-;MCK-B2 lines.

Analysis of offspring showed that only one of the two lines deposited the transgene-derived β2 consistently at all synapses, and this line was used for subsequent studies. An immunohistochemical survey of tissues was performed to determine whether expression was muscle specific. Rat β2 was detected at skeletal muscle synapses (Fig. 2C,D), but not in extrasynaptic regions of the myofiber BM. This is in contrast to the endogenous mouse β2, which is found in both synaptic and extrasynaptic regions (Fig. 2) (Sasaki et al.,2002). Expression was also observed in cardiac muscle and in some visceral smooth muscle, but not in nerve, kidney, lung parenchyma, skin,liver, retina, intestinal mucosa or brain(Fig. 2E-H and data not shown). Mosaic expression was observed in the vascular smooth muscle of arteries (data not shown). Based on these results, we conclude that the MCK-B2 transgene behaves in an appropriate tissue-specific fashion and that the presumed expression of the transgene throughout the skeletal muscle fiber nevertheless leads to concentration of β2 at synapses, as previously shown in vitro(Martin et al., 1995).

The known defects in Lamb2^-/- kidney are restricted to the glomerular filter, consistent with the fact that β2 is highly concentrated in the glomerular BM (GBM)(Hunter et al., 1989). The GBM is synthesized by two cells: the podocytes that are present in the urinary space and lie on the outer aspect of the GBM, and the endothelial cells that line the glomerular capillaries on the inner aspect of the GBM(St John and Abrahamson,2001). In order to restrict rat β2 transgene expression to the glomerulus, we chose to use a podocyte-specific promoter element isolated from the Nphs1 gene (Eremina et al., 2002), which encodes nephrin(Kestila et al., 1998), to make the NEPH-B2 transgene (Fig. 1B). For added flexibility in future experiments, the rat β2 cDNA and the adjacent SV40 sequences were flanked by loxP sites so that transgene expression could be halted by Cre-mediated recombination.

Three NEPH-B2 transgenic founders were obtained, and each was mated to a Lamb2^+/- mouse to generate three independent lines. Transgene expression was assayed in offspring by immunostaining kidney sections for rat β2, which was never detected in non-transgenic controls(Fig. 3A). All three transgenes were expressed, and rat β2 was only detected in the GBM(Fig. 3B). However, deposition was segmental and weak in two of the three lines, so only the third was used for subsequent experiments. Rat β2 deposition was not detected in any other tissues examined, including skeletal muscle, heart, intestine and lung(Fig. 3C,D; data not shown).

To determine whether muscle-derived laminin β2 was sufficient to rescue the Lamb2^-/- neuromuscular junction differentiation defects, and to attempt to isolate the glomerular filtration defect, Lamb2^+/- mice were crossed with Lamb2^+/-; MCK-B2 mice to generate Lamb2^-/-; MCK-B2 mice. Lamb2^-/-;MCK-B2 mice were significantly more healthy than Lamb2^-/-mice and were usually indistinguishable from littermate controls with respect to overall appearance and body weight during the pre-weaning period(Fig. 4A). However, they routinely died at 1 month of age, only ∼10 days later than Lamb2^-/- mice. SDS-PAGE analysis of urine showed massive proteinuria before and after weaning, with albumin being the major urinary protein (Fig. 4B; data not shown). Quantitation of urinary protein to creatinine ratios showed that proteinuria in Lamb2^-/-; MCK-B2 mice was comparable with that observed in Lamb2^-/- mice. Consistent with the heavy proteinuria, Lamb2^-/-; MCK-B2 mice were edematous near the time of death.

Immunohistochemical analyses showed that in skeletal muscle, rat β2 was concentrated at synaptic sites in Lamb2^-/-; MCK-B2 mice (Fig. 5). Interestingly,as discussed in detail in the Discussion, even with the more sensitive polyclonal antibody, little β2 was detectable along the extra-synaptic regions of the myofiber. (One exception was the MTJ, as shown below.)Expression was also detected in cardiac muscle but not in kidney or in other non-muscle tissues, consistent with the data in Fig. 2.

We have previously shown that loss of β2 from synapses in Lamb2^-/- muscle leads both to loss of laminin α5,the partner of laminin β2 in the synaptic laminin α5β2γ1(laminin-11 or Lm-521), and to increased synaptic β1(Patton et al., 1997), which presumably compensates for the absence of β2. Here, restoration ofβ2 to the synaptic cleft in Lamb2^-/-; MCK-B2 mice led to the reappearance of α5 and the disappearance of β1(Fig. 5C-F). Not only was the molecular composition of the synaptic BM normalized, but the endplate also acquired the normal pretzel-like shape(Fig. 5B) that Lamb2^-/- synapses lack(Noakes et al., 1995a). Ultrastructural analyses further confirmed the rescue of synaptic organization: the muscle endplate established junctional folds, and Schwann cell processes did not aberrantly extend into the synaptic cleft(Fig. 6A-C). We conclude that muscle-derived β2 is deposited into the synaptic cleft BM and is sufficient to rescue synaptic structural and compositional defects in the Lamb2 mutant. The overall improvement in health of the animal suggests that synaptic function is also improved.

As mentioned previously, laminin β2 is also normally concentrated at MTJs, but developmental defects in Lamb2^-/- MTJs have not previously been reported. Injection of Evans blue, a dye that accumulates only in damaged and regenerating skeletal and cardiac muscle fibers(Straub et al., 1997), into Lamb2^-/- mice resulted in labeling of muscle fibers in the diaphragm, but primarily at the ends, near MTJs(Fig. 7A,B). (This pattern contrasts with the labeling observed in typical dystrophic muscle, which occurs along the entire length of muscle fibers.) Furthermore, transmission electron microscopy showed that the BM associated with the MTJ was somewhat discontinuous and less compact compared with control, and the folding at the end of mutant muscle fibers exhibited reduced complexity(Fig. 6E,F). Together, these data suggest that in the absence of laminin β2, MTJs are structurally and functionally defective.

In Lamb2^-/-; MCK-B2 mice, β2 was detectable at most if not all MTJs (Fig. 7F),and this correlated with improved ultrastructure in some fibers(Fig. 6G). But surprisingly,muscle fibers in Lamb2^-/-; MCK-B2 mice were still labeled by Evans blue near MTJs (Fig. 7C). A possible explanation stems from our analysis of Lamb2^+/-; MCK-B2 MTJs, which should contain both endogenous mouse and transgene-derived rat β2. However, double staining for total and for transgene-derived rat β2 revealed that many MTJs do not contain rat β2, though they contain the endogenous mouse β2(Fig. 7G,H). This shows that transgene-derived protein accumulates at MTJs inefficiently compared to endogenous protein. It then follows that transgene-derived β2 should also accumulate inefficiently at MTJs in Lamb2^-/-; MCK-B2 mice(though it does accumulate, as shown in Fig. 7F), which could lead to a mild injury to the developing muscle and the observed Evans blue uptake in the highly active diaphragm.

To determine whether podocyte-derived laminin β2 is sufficient to restore the glomerular filtration barrier, and to attempt to isolate the neuromuscular defects, Lamb2^+/-; NEPH-B2 mice were crossed with Lamb2^+/- mice to generate Lamb2^-/-; NEPH-B2 mice. These mice were overtly indistinguishable from and had growth curves similar to Lamb2^-/- mice (Fig. 4A), and they died at 3-4 weeks of age, suggesting little improvement in overall health. However, immunohistological analysis revealed efficient deposition of rat β2 into the GBM(Fig. 8), and urinalysis revealed that there was no proteinuria in Lamb2^-/-;NEPH-B2 mice (data not shown). The abrogation of proteinuria led to no obvious improvement in synaptic architecture in the absence of synaptic β2(Fig. 6D), demonstrating for the first time that synaptic defects in Lamb2^-/- mice are not secondary to ill health caused by proteinuria. Based on these results, we conclude that the severe phenotype in Lamb2^-/- mice stems primarily from the neuromuscular junction defects.

We have previously shown that there is a laminin β1 to β2 developmental transition in the forming GBM(Miner and Sanes, 1994). In Lamb2^-/- mice, β1 remains in the GBM rather than being eliminated, thus serving a structural role to maintain GBM integrity,but nevertheless failing to maintain the glomerular barrier to protein(Noakes et al., 1995b). Here,restoration of podocyte-derived β2 to the GBM was accompanied by proper elimination of β1 as glomeruli matured, though it was still detectable in the mesangial matrix (Fig. 8). Ultrastructural analysis showed normal podocyte foot process architecture in Lamb2^-/-; NEPH-B2 glomeruli at all stages, even at a time near death (Fig. 6H,K; data not shown). By contrast, analysis of Lamb2^-/- and Lamb2^-/-; MCK-B2 glomeruli revealed widespread foot process effacement (Fig. 6H-J),consistent with the heavy proteinuria (Fig. 4B).

We next generated Lamb2^-/-; MCK-B2; NEPH-B2 mice through appropriate crosses. Except for occasional mild proteinuria, these mice have manifested no obvious defects and can live for well over 1 year,some to 18 months. Both males and females have exhibited normal fertility and are overtly indistinguishable from control littermates. Light microscopic analysis of PAS-stained paraffin sections revealed no significant pathology in either skeletal muscle or kidney in older mice(Fig. 9), and ultrastructural analysis of glomeruli and neuromuscular junctions showed no significant abnormalities (data not shown). Interestingly, Evans blue only occasionally labeled the ends of muscle fibers in the diaphragms of older doubly rescued mice (Fig. 7D and data not shown), and all MTJs contained β2(Fig. 7I-L), suggesting that subsequent to any initial injury at developmental stages, as in Fig. 7C, MTJs are apparently repaired. Furthermore, no histopathology was observed in other tissues reported to be rich in β2, including lung, retina, pancreas or gut (data not shown). Based on these results, a requirement for β2 in BMs other than those associated with muscle and the kidney glomerulus can not be demonstrated, at least in the context of the controlled environment of a mouse cage.

Accumulating evidence shows that laminins play crucial roles in the development and function of many organs, as well as in such fundamental developmental processes as gastrulation, notochord formation and neural tube closure (Miner and Yurchenco,2004). The growth of the laminin family during the evolution of multicellular organisms has allowed a segregation of laminins into two distinct groups in vertebrates: one group containing laminins absolutely required during embryogenesis for basic developmental processes, and the second containing laminins with more restricted, specialized functions in postnatal animals.

Here, we have focused on the laminin β2 chain, which falls squarely into the latter group. Although laminin β2 has no obvious function during development in utero, and Lamb2^-/- pups are indistinguishable from control littermates during the first postnatal week,homozygotes die at ∼3 weeks of age with severe defects in both the glomerular filter and the neuromuscular junction(Noakes et al., 1995a; Noakes et al., 1995b). Whether there might be defects in other tissues (besides retina)(Libby et al., 1999), and how the kidney and muscle defects each contribute to the lethal failure to thrive phenotype, have been important unanswered questions. The recent finding of mutations in LAMB2 associated with Pierson syndrome(Zenker et al., 2004a; Zenker et al., 2005)underscores the importance of these issues with regard to human development and disease.

To investigate whether the kidney or the synapse defect is most detrimental to the health of the animals, we used tissue-specific laminin β2 transgenes to rescue each defect individually. We have clearly shown that the defects in synaptic development and maturation are much more devastating:rescuing the GBM defects with the podocyte-specific β2 transgene(NEPH-B2) did not improve the overt phenotype of the Lamb2 mutant,whereas rescuing the structure and presumably the function of the neuromuscular junction with the muscle-specific β2 transgene (MCK-B2)greatly improved weight gain and extended lifespan by ∼50%. The latter mice must have died from nephrotic syndrome, because preventing it through simple addition of the NEPH-B2 transgene allowed for a very long life (well over 1 year) and normal fertility in Lamb2^-/-; MCK-B2;NEPH-B2 mice. Furthermore, we can conclude from this result that for normal longevity of caged mice, laminin β2 is only required in muscle and glomerular BMs, the only sites where we could detect β2 in Lamb2^-/-; MCK-B2; NEPH-B2 mice. Such an important conclusion would not have been possible had we chosen to use the more conventional Cre/loxP conditional knockout approach to mutate Lamb2specifically in skeletal muscle or podocytes, because all tissues not expressing Cre would contain their normal complement of β2.

Because the MCK regulatory element is active in both skeletal and cardiac muscle and, to a much lesser extent, in smooth muscle(Jaynes et al., 1988; Johnson et al., 1989), we cannot be absolutely certain that the basis for improvement in the health of Lamb2^-/-; MCK-B2 mice is solely due to restored synaptic architecture and function. But the fact that Lamb2^-/-hearts exhibit no obvious morphological or histopathological defects, such as ventricular dilation or cardiomyocyte necrosis or fibrosis (J.H.M., B.L.P. and G.J., unpublished), suggests that the improved health of Lamb2^-/-; MCK-B2 mice is not due to transgene expression in the heart. Moreover, that cardiac defects are not generally associated with Pierson syndrome (Zenker et al.,2004b) is also consistent with the lack of a crucial role for laminin β2 in the heart.

So why do defects in synaptic structure and function kill Lamb2^-/- mice? The mice usually die at ∼21 days of age, which is also the time of weaning. Together with the highly significant 50% decrease in mean quantal content at mutant nerve terminals(Knight et al., 2003), the most likely explanation for death is that the mice are simply unable to use their muscles properly to obtain nutrition. Before weaning, this involves suckling for many hours each day, in addition to competing with littermates for access to teats; after weaning, it involves obtaining and chewing solid food and reaching for water. Consistent with this explanation, we have found that hand feeding Lamb2^-/- mice can extend their life to up to 40 days of age, though they remain very small and weak. A lack of proper nutrition is fully consistent with the observed failure to thrive phenotype.

Although laminin β2 is concentrated at skeletal muscle synapses, it can normally also be detected extrasynaptically all along the muscle fiber with certain antibodies (Sasaki et al.,2002; Wewer et al.,1997), suggesting that the muscle deposits β2 along its entire length. Here, however, little MCK-B2 transgene-derived β2 was detected in the muscle fiber BM, other than at synapses and MTJs. Because there is no reason to suspect that the MCK regulatory element is not active in nuclei throughout the entire muscle fiber, these data suggest that skeletal muscle fibers are programmed to deposit β2 only in synaptic and MTJ BMs. This was predicted in part by in vitro data demonstrating that the lamininβ2 protein contains an inherent signal that targets it to synapse-like sites on cultured myotubes (Martin et al.,1995). Another cell type - perhaps the interstitial fibroblast -is likely to be responsible for deposition of laminin β2 along the extra-junctional regions of the muscle fiber.

During kidney glomerular development, distinct BMs synthesized by the epithelial podocytes and the invading endothelium fuse to form the immature GBM (Abrahamson and Perry,1986). The laminin component of the maturing GBM continues to be synthesized by both the podocytes and the endothelial cells(St John and Abrahamson,2001). Our approach to restore laminin β2 to the mutant GBM by expressing it exclusively in podocytes via the nephrin promoter was successful in several respects, including localization to the GBM, restoration of the glomerular barrier to protein and restriction of laminin β1 to the mesangium. This indicates that though GBM laminin β2 is normally synthesized by both podocytes and endothelial cells, expression is podocytes is sufficient for normal GBM structure and function.

Our results have important implications for Pierson syndrome, which is caused by mutations in human LAMB2(Zenker et al., 2004a; Zenker et al., 2005). Individuals with this disease present with congenital nephrotic syndrome,diffuse mesangial sclerosis and microcoria (fixed narrowing of the pupils)associated with absent dilator pupillae muscles in the iris. Other eye abnormalities include an almost absent ciliary body, an enlarged cornea, an abnormal lens shape, maldevelopment of the choroid and arrested retinal development. No obvious structural abnormalities of the brain have been found,but muscular hypotonia and abnormal movements have been reported(Zenker et al., 2004b). Affected individuals usually die within weeks of birth from kidney failure,but some have lived for over 2 years. The kidney and retinal defects seem to correlate quite well with defects in the Lamb2 mutant mouse, and the muscular hypotonia may stem from similar defects in development and function of the neuromuscular junction. However, cornea, lens and pupillary defects have not been found in the mouse.

Because of the availability of renal replacement therapy, through either dialysis or transplantation, the congenital nephrosis is the most treatable aspect of Pierson syndrome. However, the results of our transgenic rescue studies show that ameliorating the nephrotic syndrome on its own is not significantly beneficial in extending the life of the Lamb2^-/- mouse. This calls into question whether renal replacement therapy alone is warranted in individuals with Pierson syndrome. Indeed, one individual who received chronic dialysis treatment and lived for 1.8 years exhibited significant neurological and developmental deficits including severe hypotonia, psychomotor delay, and blindness(Zenker et al., 2004a). Recently, a missense mutation in LAMB2 was reported in a patient exhibiting congenital nephrotic syndrome but not the other features of Pierson syndrome (Hasselbacher et al.,2005), demonstrating that there is heterogeneity in the severity of disease resulting from LAMB2 mutations in humans. As more individuals with LAMB2 mutations are discovered and genotype-phenotype correlations are discerned, a better understanding of the neurological and muscular defects will hopefully emerge and lead to the development of extra-renal treatment paradigms.

We are grateful to Joshua R. Sanes for advice and for laminin β2 plasmid, antibodies and mutant mice; to Jean Buskin, Stephen Hauschka and Susan Quaggin for plasmids; to Dale Abrahamson, Takako Sasaki and the late Rupert Timpl for antibodies; to Cong Li, Jennifer Richardson, Dan Martin and Marilyn Levy for technical assistance; and to the Mouse Genetics Core for production and care of mice. This work was supported by an Established Investigator Award from the American Heart Association and grants from the NIH(R01DK064687 and R01GM060432) to J.H.M., by a grant from the NIH (R01NS040759)to B.L.P., and by a Research Grant from the National Kidney Foundation of Eastern Missouri and Metro East to G.J.