GPR56 Functions Together with α3β1 Integrin in Regulating Cerebral Cortical Development

Loss of function mutations in GPR56, which encodes a G protein-coupled receptor, cause a specific human brain malformation called bilateral frontoparietal polymicrogyria (BFPP). Studies from BFPP postmortem brain tissue and Gpr56 knockout mice have previously showed that GPR56 deletion leads to breaches in the pial basement membrane (BM) and neuronal ectopias during cerebral cortical development. Since α3β1 integrin also plays a role in pial BM assembly and maintenance, we evaluated whether it functions together with GPR56 in regulating the same developmental process. We reveal that loss of α3 integrin enhances the cortical phenotype associated with Gpr56 deletion, and that neuronal overmigration through a breached pial BM occurs earlier in double knockout than in Gpr56 single knockout mice. These observations provide compelling evidence of the synergism of GPR56 and α3β1 integrin in regulating the development of cerebral cortex.


Introduction
The interaction between cells and their environment is essential to brain development. Dystroglycan and intergrins are two major cell surface receptors that mediate cell-extracellular matrix (ECM) interactions [1][2][3][4]. Recently, the family of adhesion G proteincoupled receptors (GPCRs) was identified as the third major category of ECM receptors [5,6]. Adhesion GPCRs are characterized by the presence of a large extracellular region and a G protein proteolytic site (GPS) domain that cleaves the receptor into N-and C-terminal fragments [7][8][9]. Mutations in one such adhesion GPCR, GPR56, cause a specific human brain malformation called bilateral frontoparietal polymicrogyria (BFPP) [10]. BFPP is a recessively inherited genetic disorder affecting human cerebral cortical development and characterized by disorganized cortical lamination that is most severe in the frontal and parietal lobes [10][11][12][13]. Histological analyses of postmortem human BFPP brain samples and Gpr56 knockout mice indicated that the histopathology of BFPP is a cobblestone-like brain malformation [14,15].
Cobblestone lissencephaly, also called type II lissencephaly, is defined as aberrant migration of cortical neurons through breaches in the pial basement membrane (BM), resulting in neuronal ectopias on the surface of the brain [16]. Cobblestone cortex is typically seen in three distinct human congenital muscular dystrophy syndromes: Muscle-eye-brain disease (MEB), Fukuyama-type muscular dystrophy (FCMD), and Walker-Warburg syndrome (WWS) [16]. These three disorders are autosomal recessive diseases that encompass congenital muscular dystrophy, ocular malformations, and cobblestone lissencephaly. MEB, FCMD, and some WWS cases are caused by aberrant glycosylation of a-dystroglycan, a receptor for laminin [17][18][19][20].
Defects in some members of the integrin family and their ligand, laminin, have been implicated in the pathogenesis of BM breakdown and neuronal ectopias [21][22][23][24]. Furthermore, our previous work demonstrated that loss of GPR56 diminishes the adhesion of rostrally derived granule cells to laminin [25]. Integrin a3 (gene symbol, Igta3) always associates with b1 integrin to form the a3b1 dimeric protein on the cell surface, serving as one of the major receptors for laminin [26,27]. We therefore investigated how GPR56 functionally interacts with a3b1 integrin in vivo by studying Gpr56 and Itga3 compound mutant mice. We demonstrate that loss of a3b1 integrin exacerbates the Gpr56-associated cortical phenotype in a dose dependent manner, indicating that the two receptors function synergistically during cortical development.

Loss of Itga3 Enhances the Cortical Phenotype Associated with Gpr56 Deletion
Since Itga3 2/2 mice die at birth with kidney and lung defects and moderate skin blistering, we crossed Gpr56 2/2 with Itga3 +/2 mice and intercrossed the F1 offspring to generate compound mutant mice [28]. Genotype for various mutant mice was confirmed by PCR as previously described [14,28]. The absence of protein expression was verified by immunohistochemistry (IHC) for a3 integrin ( Figure S1) and western blot analysis for GPR56 ( Figure S2). First, we evaluated the overall cortical lamination at P0 by Nissl staining of various coronal sections of compound mutant mouse brains. Consistent with our previous study, Gpr56 single knockout mice revealed neuronal ectopias (Table 1), while Gpr56 heterozygous mice appear phenotypically normal (data not shown) [14]. Loss of a3 integrin was previously shown to result in poor neuronal migration and the formation of heterotopia [29,30]. However, we did not observe any discernible lamination defect in Itga3 single knockout mouse brains, except mild neuronal overmigration in one out of 11 P0 Itga3 2/2/ Gpr56 +/2 brains ( Figure 1N and Table 1). Double heterozygous mice (Itga3 +/2/ Gpr56 +/2 ) were also lacking any obvious brain phenotype (Table 1).
Interestingly, Itga3 +/2/ Gpr56 2/2 and Itga3 2/2/ Gpr56 2/2 mice showed more severe cortical defects than what is observed in Gpr56 single knockout mouse brains ( Figure 1C, D, G and H). To further reveal the cortical lamination defects, we performed layer marker staining with Cux1 for layer II-IV, CTIP2 for layer V, and Tbr1 for layer II-III and layer VI neurons [31][32][33]. Cux1-, Tbr1-, as well as CTIP2-positive neurons were detected in the ectopic clusters, indicating that the developmental abnormalities affects to both superficial and deeper layer neurons ( Figure 1J-L and N-P).
To further quantify the severity of the cortical defects in various compound mutant mice, we performed semi-quantitative analyses of Nissl-stained brain sections at E16.5. We defined the cortical ectopias as small, medium, and large based on the width of the ectopic outgrowth (Figure 2A-C). Compared to Gpr56 single knockout mice, there was a significantly greater number of large size cortical ectopias in Itga3 +/2/ Gpr56 2/2 and Itga3 2/2/ Gpr56 2/ 2 mice, with Itga3 2/2/ Gpr56 2/2 mice being the most severely affected ( Figure 2D-G). Again, we did not identify any discernible cortical phenotype in Itga3 2/2/ Gpr56 +/+ , Itga3 2/2/ Gpr56 +/2 , and Itga3 +/2/ Gpr56 +/2 mice at E16.5 (Table 1). These data confirmed that cortical ectopias are only associated with Gpr56 mutation and that loss of a3b1 integrin enhances the Gpr56-associated cortical phenotype, with each additional allele loss corresponding to a more severe phenotype, suggesting a possible functional interaction between a3b1 integrin and GPR56 during cortical development.
Earlier Pial BM Breakdown with Associated Neuronal Overmigration was Observed in the Itga3 +/2/ Gpr56 2/2 and Itga3 2/2/ Gpr56 2/2 Neocortices We have previously showed that the pial BM was properly formed in Gpr56 single knockout mouse embryonic brains before E12.5 (10 am on the 12 th day of vaginal plugging), and regional pial BM breaches started to occur at E12.8 (6 pm on the 12 th day of vaginal plugging) [14]. Based on the fact that more severe cortical ectopias were observed in Itga3 +/2/ Gpr56 2/2 and Itga3 2/ 2/ Gpr56 +/2 mice, we hypothesized that pial BM breaches would occur earlier in these two mutant mice. To test this hypothesis, we performed a detailed time course study of the occurrence of the breached pial BM and overmigrated neurons in double mutant mice. Double IHC of Tuj1 and laminin revealed that the pial BM was well formed in E10.5 neocortices of double mutant mice (Table 1, Figure 3C, E, I and K). Regional breakdown of the pial BM with concurrent neuronal overmigration was observed at E11.5 ( Figure 3D, F, J, and L). Again, we observed migrating neurons piercing through a well formed pial BM at approximately E12.8 in Gpr56 single knockout mice ( Figure 3B and H) [14]. The notably earlier onset of BM breaching and neuronal overmigration seen in Gpr56 and Itga3 double mutant mice further supports the notion that a3 integrin has a cooperative function with GPR56 during cortical development. Although the onset of BM breaching and neuronal overmigration was observed at a similar developmental stage by immunostaining in both Itga3 +/2/ Gpr56 2/2 and Itga3 2/2/ Gpr56 2/2 mice, the number of larger ectopias was significantly higher in Itga3 2/2/ Gpr56 2/2 mice ( Figure 2D-G). The possible explanation could be that immunostaining is not sensitive enough to detect the difference in the onset of pial BM breakdown between both Itga3 +/2/ Gpr56 2/2 and Itga3 2/ 2/ Gpr56 2/2 mice.
To investigate the relationship of migrating neurons, radial glial endfeet, and pial BM, we performed triple IHC on E11.5 Itga3 2/ 2/ Gpr56 2/2 neocortices. Interestingly, both migrating neurons and radial glial endfeet were protruded through seemingly well formed pial BM ( Figure 4D and H). Taken together, our data suggests a possibility that migrating neurons and radial glial endfeet pierce through a previously well formed pial BM.

Loss of Itga3 Disrupted Collagen III-mediated Neuronal Migration Inhibition
On a cellular level, the binding of GPR56 and its ligand collagen III causes an inhibition of neuronal migration [34]. To elucidate the synergistic function of GPR56 and a3b1 integrin at the cellular level, we questioned whether loss of a3 integrin affects collagen III-mediated neuronal migration inhibition. Neuronal migration assays were conducted to determine whether deleting Itga3 would lead to a decrease in the inhibition of migration. Neurospheres established from E13.5 cortices of either wild type, Gpr56 2/2 , or Itga3 2/2 mice were cultured in neuron culture medium containing 84 nM of purified collagen III or control solution (acetic acid). After two days in culture, the neurospheres were assessed for being either positive or negative for migration, using criteria described previously [34]. Neurospheres derived from Itga3 2/2 cortices had a significantly diminished migration inhibition in comparison to wild type ( Figure 5E, F, and G).

Integrin a3b1 does not Bind Directly to Collagen III
Thus far, we have demonstrated that GPR56 functions together with a3b1 integrin in regulating cerebral cortical development. Mechanistically, there are two possibilities to account for this synergistic activity of the two receptors: (1) GPR56 and a3b1 integrin function in the same receptor complex and bind collagen III as a common ligand; (2) the two receptors act indirectly through an unknown mediator, which requires the activation of GPR56 pathway. To investigate the first possibility, we conducted a solid phase binding assay using recombinant a1b1 and a2b1 integrins, the natural receptors for collagen III, as the positive control [35,36]. Recombinant human a1b1, a2b1, a3b1, and a6b1 integrins were tested for their ability to bind to human collagen III and recombinant human laminin-511 coated on 96well plates in the presence of 1 mM Mn 2+ or 10 mM EDTA. Integrins a1b1 and a2b1 bound to collagen III as expected ( Figure 6A and B). However, a3b1 and a6b1 integrins did not bind to collagen III ( Figure 6C and D). The biological activities of a3b1 and a6b1 integrins were confirmed by their binding to laminin-511 ( Figure 6C and D). This result is consistent with our previous publication, in which loss of GPR56 reduced the adherence of granule cells to integrin ligands, laminin and fibronectin, without direct binding to these two substrates (further discussed below) [25].
Taken together, our data suggests that a3b1 integrin functions together with GPR56 in regulating cortical development, probably via an unknown mediator that requires the activation of GPR56 pathway (Figure 7).

Discussion
This study demonstrates synergistic activities of GPR56 and a3b1 integrin during cerebral cortical development. A more severe cortical phenotype was uncovered in compound mutants reflecting the fact that several receptors must function together in regulating cortical development. The requirement of both a3b1 integrin and GPR56 for the proper cortical development illustrates that the regulation of the pial BM integrity and cortical lamination are integrated processes and suggests a link with two different signaling pathways.
Integrins are heterodimeric transmembrane receptors composed of noncovalently associated a and b chains [38]. While some a chains bind to multiple b chains, a3 integrin is always associated with b1 to form a mature a3b1 integrin receptor on the cell surface [39]. Integrin a3b1 was described as a promiscuous receptor for many ligands, among which is laminin-5 [40]. However, a3b1 integrin does not bind directly to collagen III, which serves as the ligand of GPR56 ( Figure 6). This is consistent with our previous study on the developing cerebellum [25]. GPR56 is expressed specifically in the rostral granule cells perinatally and absence of GPR56 causes loss of adhesion of rostrally derived granule cells to laminin-1 and fibronectin, known to be the ligands of integrin. However, this cell-matrix adhesion is not through direct binding of GPR56 to the matrix proteins because neither addition of soluble GPR56 to the media in the granule cell adhesion assay nor overexpression of full-length mouse GPR56 in HEK 293T cells altered cell adhesion to laminin-1 and fibronectin [25]. Taken together, our previous and current study results suggest a functional link between integrin and GPR56, despite the lacking of a common binding partner.
Loss of a3 integrin was previously shown to result in poor neuronal migration and the formation of heterotopia [29,30]. We did not observe any obvious cortical defect in Itga3 single knockout mice, presumably due to differences in mouse genetic background (129/BL6 vs 129/BL6/FvB). However, Gpr56 and Itga3 double knockout mice developed a more severe cortical malformation manifested by neuronal ectopias and an earlier occurrence of pial BM breachment and neuronal overmigration, compared to Gpr56 single knockout mice. Our finding is consistent with other reports relevant to a3 integrin. For example, a3 and a6 integrins double knockout mice exhibit regional breakdown of pial BM with concurrent neuronal ectopias [22]. Studies of b1 integrin, as well as integrin associated downstream partners, integrin-linked kinase (ILK) and focal adhesion kinase (FAK), also support the notion that a3b1 integrin is essential to maintaining and reforming the pial basement membrane, with ILK, FAK, and b1 integrin knockout mice exhibiting neuronal ectopias on the brain surface with concurrent regional fragmentation in the basal lamina, features of cobblestone-like cortical malformation [41][42][43].
The leading pathology of cobblestone-like cortical malformation is thought to be a defective pial BM [16]. However, recent literature suggests that abnormal neuronal migration could be partly responsible [44,45]. We have previously showed that (1) a gradient expression of GPR56 in preplate neurons that matches the regional cortical defects associated with loss of GPR56, in spite of the fact that no such pattern is apparent in the radial glia [37]; (2) neurons have direct contact with the pial BM during early cortical development [34]; and (3) the interaction of GPR56 and collagen III inhibits neuronal migration [34]. In this study, we further demonstrated the expression of a3 integrin in the preplate neurons as well as an attenuated collagen III-mediated neuronal migration inhibition in Itga3 2/2 neural progenitor cells. Taken together, it is likely that GPR56 functions together with a3b1 integrin in mediating the interaction between the pial BM and preplate neurons as well as radial glial endfeet, thus defining the boundary between the neocortex and the meninges while providing a framework for the developing cortex (Figure 7).

Ethics Statement
Experiments were performed in accordance with National Institutes of Health guidelines for the care and use of laboratory

Mice
Gpr56 knockout mice were obtained from Genentech, maintained in a mixed genetic background of 129/BL6/FvB. Itga3 knockout mice were generated on a 129 background [28]. Crossing Gpr56 2/2 with Itga3 +/2 produced Gpr56/Itga3 double mutant mice in a mixed 129/BL6/FvB genetic background. Fetal stage was calculated from the day when a vaginal plug was observed (considered as E0.5).

Histology and Immunohistochemistry
Histological analysis was carried out as previously described [14,37]. Embryonic brains of E10.5 to E12.5 were fixed for frozen sectioning in 4% PFA for 2-3 hr at 4uC. E16.5 brains and P0.5 brains were fixed in 4% PFA at 4uC for 48 hrs or 24 hrs, respectively, cryoprotected by 30% sucrose in PBS at 4uC sinking, embedded in OCT compound (Tissue Tek, Sakura Finetek USA INC.), and stored at 280uC until sectioned. For double IHC of a3 integrin and GPR56 at E10.5 and E11.5, the sections were retrieved by boiling for 8 min followed by cooling them down at room temperature (RT) for 30 min. After washing the slides with PBS three times, they were incubated with 1% SDS for 5 min followed by washing three times again with PBS. The sections were incubated with rabbit anti-a3 integrin and mouse anti-GPR56 antibodies overnight at 4uC and washed with normal PBS one time, PBS containing 2.7% NaCl (instead of 0.8% NaCl for normal PBS) twice, and normal PBS once. Primary antibodies were visualized by goat anti-rabbit Alexa-546 and goat anti-mouse Alexa-488 secondary antibodies. Nuclei were stained with Hoechst 33342 (Invitrogen, 1:2000). Images were captured using a Nikon 80i upright microscope or a Olympus confocal FluoView Laser System (FluoView FV1000).
For double IHC of Tbr1 and CTIP2, TSA-TMR (Perkin Elmer) was used to amplify the Tbr1 signal in a 1:50 dilution followed by incubation of peroxidase-conjugated goat anti-rabbit IgG antibody (Sigma) for 30 minutes.

Semi-quantitative Measurements of Cortical Dysplasia
Neuronal ectopias in various compound mutant brains at E16.5 were captured by a Nikon 80i upright microscope and subjected to blind quantitative analysis. Sagittal cryostat sections (8 mm) were obtained serially from lateral to medial, starting from the point when the posterior horn of the lateral ventricle could be clearly viewed. Every six continuous sections were collected and stained with 0.1% cresyl violet/0.5% acetic acid for semi-quantitative analysis. For each brain, 13 sections in total were analyzed. The base of the neuronal cluster invading the marginal zone was measured and the ectopic cluster was defined as large (.100 mm), medium (50-100 mm), and small (,50 mm) (Figure 2 A-C). Data are presented as means 6 S.E.M. Statistical analysis was performed using t-test with P,0.05 considered significant.

Vector Construction
Expression vectors for recombinant human a3b1 and a6b1 integrins were prepared as described previously [47]. The cDNAs encoding the extracellular domains of human integrin a1 and a2 subunits were generated by reverse transcription-PCR using premade double-stranded cDNAs derived from human fetal tissues (Clontech). The primers for human integrin a1 subunit were 59-AAGGTACCAC-

Expression and Purification of Recombinant Proteins
The recombinant human integrins were designed as heterodimeric soluble proteins composed of the extracellular domains of integrins with ACID/BASE coiled-coil peptides and FLAG/ 66His tag sequences at their C-termini for solid-phase binding assays [46,47]. Recombinant integrins were produced using a FreeStyle TM 293 Expression System (Invitrogen). Briefly, 293F cells were simultaneously transfected with expression vectors for a and b subunits using 293fectin (Invitrogen), according to the manufacturer's instructions, and grown in serum-free FreeStyle TM 293 Expression medium for 72 h. The conditioned media were collected and clarified by centrifugation, and then subjected to affinity chromatography using an anti-FLAGH M2-agarose (Sigma) column. The column was washed with Tris-buffered saline (TBS) containing 1 mM MgCl 2 and 1 mM CaCl 2 , TBS (+). The bound proteins were eluted with TBS (+) containing 100 mg/ ml FLAG peptide (Sigma) and dialyzed against TBS. Recombinant human laminin-511 was produced using a FreeStyle TM 293 Expression System (Invitrogen) and purified from conditioned media as described previously [48]. The protein concentrations of all the recombinant products were determined using a BCA protein assay kit (Thermo Scientific) using bovine serum albumin  (BSA) as a standard. The expression vectors for recombinant human laminin-511 were constructed as described previously [48,49].

Solid-phase Binding Assays
Ninety six-well microtiter plates were coated with human type III collagen (BD Bioscience) and recombinant human laminin-511, and BSA (10 nM, 50 ml/well) overnight at 4uC, and then blocked with TBS containing 1% BSA and 0.02% Tween-20 for 1 h at RT. Next, the plates were washed with TBS containing 0.1% BSA, 0.02% Tween-20, and 1 mM MnCl 2 (Buffer A) or TBS containing 0.1% BSA, 0.02% Tween-20, and 10 mM EDTA (Buffer B). Diluted integrins (10 nM, 50 ml/well) were added to the plates and allowed to bind to the substrate-adsorbed ligand proteins in the presence of 1 mM MnCl 2 or 10 mM EDTA for 3 h at RT. The plates were washed three times with Buffer A or Buffer B, and the bound integrins were quantified by an enzyme-linked immunosorbent assay as reported previously [47]. Briefly, the wells were incubated with the biotinylated anti-Velcro antibody (1 mg/ ml, 50 ml/well) for 30 min at RT in Buffer A, washed three times with Buffer A and incubated with HRP-conjugated streptavidin (0.33 mg/ml, 50 ml/well) for 15 min. After three washes with Buffer A, the bound antibodies were quantified by measuring the absorbance at 490 nm after incubation with o-phenylenediamine.

Primary Neural Culture and Immunocytochemistry (ICC)
Primary neural culture was performed as previously described [37]. Briefly, cortical cells were harvested from E13.5 mouse cortices with the meninges removed. Dissociated cells were seeded on 10 ml tissue culture dish at 37uC for 10 min to deplete the fibroblasts. The cells (1610 5 /ml) were placed on cover glasses precoated with poly-D-lysine (100 mg/ml) and cultured in neural culture medium (neurobasal medium supplemented with B27, 1% penicillin/streptomycin, and 1% L-glutamate) for one day (1DIV).
To perform ICC, the cultured primary neuronal cells were fixed in cold 95% ethanol and 5% glacial acetic acid for double ICC of H11 and a3 integrin, followed by three washes with PBS. Cells were permeabilized with 0.1% Triton-X 100 in PBS for 10 min followed by three washes with PBS. After blocking with 10% goat serum, 1% BSA, and 0.1% Triton-X100 in PBS for 30 min, the primary antibodies were incubated at 4uC overnight and visualized by appropriate fluorophore-conjugated secondary anti-  bodies (Invitrogen, 1:1000). Nuclei were stained with Hoechst 33342 (Invitrogen, 1:2000). Images were captured using a Nikon 80i upright microscope.

Migration Assay
Neurosphere generation and migration assay were performed as previously described [34]. Briefly, E13.5 mouse cerebral cortex was separated with the overlying meninges removed. Individual embryonic mouse cortex was cultured in neuron culture media (DMEM supplemented with B27, 1% penicillin/streptomycin, 1% L-glutamate, and 10% FBS). After genotyping, the mouse cortices were pooled blindly into two groups (wild type and Itga3 knockout). The neurospheres were plated in a 48-well dish precoated with 100 mg/ml poly-D-lysine (PDL) and cultured in neuron culture medium overnight to allow the neurospheres adhere to the PDL substrate. The culture medium was then changed to the experimental mediums: neuron culture medium with 84 nM purified human collagen III (Abcam), or with carrier solution (acetic acid) as control. The neurospheres were imaged and the number of migrating neurospheres was quantified, as detailed previously. Figure S1 Integrin a3 immunostaining on Itga3 +/+ , Itga3 2/2 , and Itga3 2/2/ Gpr56 2/2 brain sections and progenitor cells.