Integrin and dystroglycan compensate each other to mediate laminin-dependent basement membrane assembly and epiblast polarization

Abstract

During early embryogenesis, endodermal γ1-laminin expression is required for basement membrane (BM) assembly, promoting conversion of non-polar pluripotent cells into polarized epiblast. The influence of laminin-111 (Lm111) and its integrin and dystroglycan (DG) receptors on epiblast in embryoid bodies (EBs), a model for differentiation of the embryonic plate, was further investigated. Lm111 added to the medium of EBs initiated conversion of inner nonpolar cell to the polarized epiblast epithelium with an exterior-to-central basal-to-apical orientation. Microinjection of Lm111 into EB interiors resulted in an interior BM with complete inversion of cell polarity. Lm111 assembled a BM on integrin-β1 null EBs with induction of polarization at reduced efficiency. β-Integrin compensation was not detected in these nulls with integrin adaptor proteins failing to assemble. A dimer of laminin LG domains 4-5 (LZE3) engineered to strongly bind to α-dystroglycan almost completely inhibited laminin accumulation on integrin β1-null EBs, reducing BM and ablating cell polarization. When Lm111 was incubated with integrin-β1/dystroglycan double-knockout EBs, laminin failed to accumulate on the EBs, the EBs did not differentiate, and the EBs underwent apoptosis. Collectively the findings support the hypotheses that the locus of laminin cell surface assembly can determine the axis of epithelial polarity. This requires integrin- and/or dystroglycan-dependent binding to laminin LG domains with the highest efficiency achieved when both receptors are present. Finally, EBs that cannot assemble a matrix undergo apoptosis.

Keywords: cell polarity; embryonic stem cell; endoderm; epiblast.

Fig. 1

Surface assembly of basement membrane on endoderm-free embryoid bodies. EHS laminin-111 (0.1 mg/ml) was added to the culture media of 1-day endoderm free EBs which were then cultured for 7 days. The untreated (NT) and laminin-treated (+ Lm) EBs were fixed, sectioned and examined by phase and immunofluorescence microscopy. A BM containing laminin-111 (Lm), type IV collagen (Col-IV), nidogen-1 (Nd) and perlecan (Perl) was present on the outer surface of laminin-treated but not the untreated EBs. It consisted of a cell-associated component surrounded by a more loosely associated ECM that was easily detached and generally lost during the washing step, leaving behind the cellular BM. Co-treatment with collagenase removed the collagen but not the laminin-based ECM. Differentiation was detected by phase, apical F-actin (Fact), and central apoptosis (cleaved caspase-3) staining.

Fig. 2

Laminin alteration of lineage differentiation and polarity. A. One-day wild-type EB were treated or untreated with laminin-111 (0.1 mg/ml) prior to spontaneous formation of endoderm (day 2). Treatment with laminin resulted in the formation of a cell-associated ECM and epiblast differentiation despite selective suppression of endoderm formation. α-PF: α-fetoprotein. Arrowheads point to the accumulation of F-actin and the apical polarity protein MUPP1 on the apical side of epiblast. B. Immunoblots show that laminin treatment downregulated the pluripotent markers Nanog and Oct4. C. Immunoblots show that laminin treatment suppressed the expression of the endodermal marker cytokeratin-8 and induced the expression of the epithelial polarity proteins syntaxin-3 and CRB3. D. Compared with the 7-day spontaneously differentiated EBs, laminin treatment induced very little if any late epiblast markers (FGF-5, fibroblast growth factor-5; Bra: Brachyury; Snai1: Snail 1). E. Three-day EBs following endoderm formation were treated with 0.1 mg/ml laminin-111 (added on day 4) and cultured to 7 days. Late treatment with laminin resulted in assembly of an outer BM following assembly of a subendodermal BM. The endoderm sandwiched between the two BMs was multilayered with loss of apical ZO-1 and basal GM-130, and scattered intracellular expression of vimentin (Vm). Arrows point to vimentin-positive cells. E. Diagram of polarity transitions resulting from media treatments with laminin-111 (en, endoderm; epi, epiblast; Lm, laminin; TJ, tight junction; act, F-actin; Go, Golgi).

Fig. 3

Microinjection of laminin into endoderm-free embryoid bodies. Panel A. Laminin-111 was introduced into the interior of 4 day old endoderm-free EBs by microinjection followed by culturing for 4 days before harvesting. Untreated wild-type (WT), untreated endoderm-free and endoderm-free EBs treated with (external) laminin in medium are shown as controls. In laminin-treated control EBs, dystroglycan (DG) and α6-integrin, two basal components, are concentrated on the outer circumferential surface adjacent to laminin (Lmα1) while the F-actin belt, pericentrin and (for epiblast) GM130, three apical markers, are concentrated on the inner aspect of the EBs. Microinjection of laminin completely reversed the basal-apical orientation of these markers in the layer. Panel B. Diagram of polarity transitions resulting from external (media) vs. internal (microinjection) placement of laminin-111 (I/DG, α6-integrin/dystroglycan; pc, pericentrin).

Fig. 4

Ablation of both integrin β1 and dystroglycan inhibits endoderm differentiation and BM assembly. A. Immunoblots show that integrin β1, β-dystroglycan (β-DG) and α-dystroglycan (α-DG) are absent from integrin β1/dystroglycan double knockout (β1^−/−/DG^−/−, clone C6 and C7) EBs compared with the floxed control (β1^fl/fl/DG^fl/fl). B. Phase and immunofluorescence images of 2-day EBs show that β1^fl/fl/DG^fl/fl formed endoderm on the EB surface and an underlying BM. Endoderm was identified by disabled-2 (Dab2) immunofluorescent while the BM was stained with laminin α1 (Lm α1) and perlecan antibodies. C. One-day wild-type (WT), integrin β1-null (β1^−/−), dystroglycan-null (DG^−/−) and β1^−/−/DG^−/− EBs were treated with laminin-111 (0.1 mg/ml) for 24 h. Phase images of live cultures show extensive apoptosis occurring on the surface of β1^−/−/DG^−/− EBs (arrowheads). Immunostaining for laminin α1 (Lm α1) demonstrated that a dense laminin-containing ECM assembled on the surface of WT, β1^−/− and DG^−/− EBs but not that of β1^−/−/DG^−/− EBs. In addition, high levels of apoptosis were detected in the double knockout EBs. D. EBs were washed twice with PBS to remove the loosely attached ECM and then analyzed by immunoblotting for laminin α1. EB-associated laminin was reduced in β1^−/− and DG^−/− EBs and barely detectable in the double knockout EBs. Actin was used as a loading control.

Fig. 5

Contribution of integrins to epiblast polarization. A. One-day β1^−/− EBs were treated with laminin-111 (0.1 mg/ml) for 5 days and immunostained for the apical marker MUPP1 and basement membrane perlecan (perl). F-actin was visualized using rhodamine-phalloidin. Arrowheads indicate apical accumulation of MUPP1 and F-actin. The arrow indicates apoptotic cells. B. EBs with BM and polarized epiblast were counted and plotted as percentage of total EBs examined. N = 4 for each group with a total of 1379–1744 EBs. P < 0.01 vs WT. C. One-day EBs were treated with laminin-111 (0.1 mg/ml) for 24 h and immunostained for integrin αV, talin, integrin-linked kinase (ILK) and α-dystroglycan (α-DG). BM was identified with laminin γ1 or α1 antibody. Integrin αV was largely intracellular while α-DG (arrowheads) was recruited to the cell-ECM adhesion site in wild-type and integrin β1^−/− EBs. Talin and ILK were recruited to the BM zone in wild-type (arrowheads) but no integrin β1^−/− EBs. D. One-day β1^−/− EBs stably transfected with integrin αV shRNA (αV knockdown, αV KD) or the scrambled control (SC) were treated with laminin-111 (0.1 mg/ml) for 5 days. About 20% of αV KD EBs and the scrambled control formed BM and polarized epiblast. Arrowheads indicate apical accumulation of the polarity marker MUPP1. Arrows point to apoptotic debris. E. EBs with polarized epiblast were counted and plotted as percentage of total EBs examined. N = 4 for each group with a total of 325–327 EBs.

Fig. 6

Blockade of laminin binding to dystroglycan. A. A diagram shows the structure of laminin α1 and recombinant leucine zipper E3 (LZE3). LZE3 consists of an N-terminal leucine zipper sequence fused to the Lmα1 fragment corresponding to the elastase fragment E3. The leucine zipper mediates dimerization of the laminin fragment. LZE3 was predicted to bind to α-dystroglycan at higher affinity (inset). B. SDS-PAGE, reduced (10% acrylamide gel) of recombinant LZE3, Coomassie blue stained. C. Direct solid-phase binding of recombinant LZE3, miniagrin and Lmα1 LG4–5 to immobilized rabbit muscle α-dystroglycan (3 μg/ml coat). Ligands, incubated for 1 h in blocking buffer, were detected with HRP-conjugated anti-FLAG M2 antibody or anti-agrin rabbit antibody. Apparent K_D values determined from simple binding fit shown in parentheses. D. Solid-phase inhibition binding assay shows that LZE3 inhibits lamainin-111 binding to α-dystroglycan in a dose-dependent fashion. LZE3 moderately inhibited laminin binding to sulfatide but had no effect on the laminin-integrin interaction. Plates were separately coated (a) with WGA-purified muscle dystroglycan for αDG binding, (b) soluble α7β1 integrin, and (c) galactosyl-sulfatide for sulfatide binding. EHS Lm111 (14 nM, constant) was added to the wells mixed with serial two-fold dilutions of LZE3. The bound laminin was detected with anti-laminin E4 antibody (β1LN-LEa specificity). Average and s.d. shown for n = 3 wells. E–J. One-day wild-type (E,F), DG^−/− (G,H) and integrin β1^−/− (I, J) EBs were incubated with laminin-111 (0.1 mg/ml) in the presence or absence of LZE3 (0.05 mg/ml) for 24 h (for immunoblotting), or incubated for five additional days for detection of BM and determination of epiblast conversion by phase and immunofluorescence microscopy. For laminin immunoblotting, EBs were washed twice with PBS followed by SDS-PAGE (reducing conditions), transferred to membranes and the α1-subunit detected with laminin-specific antibody. Actin or E-cadherin (integrin-null EBs to increase sensitivity) was used as a loading control. The above EBs cultured to 5 days were analyzed by phase contrast (epiblast differentiation count) and laminin antibody/phalloidin fluorescence microscopy to detect BMs and confirm differentiation. The extent of BM (fraction of EBs with a linear laminin-staining pattern) and fraction of EBs showing epiblast differentiation were determined with corresponding plots shown in F, H and J; n = 3 culture in wells for each group with the following total EBs counted/group: 92 to 104 WT, 88 to 94 DG^−/−, and 182–248 integrin β1^−/− EBs. The laminin-treated EBs were compared to the laminin + LZE3 EBs by Student t-test. P values determined were 0.09 for WT BM, 0.11 for epiblast conversion, 0.87 for DG^−/− BM, 0.81 for DG^−/− epiblast conversion, <0.01 for integrin β1^−/− BM and <0.01 for integrin β1^−/− epiblast conversion. Only the integrin β1^−/− EBs exhibited significant BM and epiblast decreases.