|Molecular Vision 2004;
Received 30 January 2004 | Accepted 16 August 2004 | Published 23 August 2004
BMP and activin receptor expression in lens development
Robbert U. de Iongh,1
Maria I. Kokkinos,1
John W. McAvoy2,3,4,5
1Department of Anatomy & Cell Biology, The University of Melbourne, Parkville, Australia; 2The Save Sight Institute, 3Department of Clinical Ophthalmology & Eye Health, 4Department of Anatomy and Histology, and 5Institute for Biomedical Research (F13), The University of Sydney, New South Wales, Australia
Correspondence to: Dr. Robbert U. de Iongh, Department of Anatomy & Cell Biology, The University of Melbourne, Parkville, VIC 3010, Australia; Phone: +61 3 8344-5788; FAX: +61 3 9347-5219; email: r.deIongh@unimelb.edu.au
Purpose: Members of the TGFβ super-family have been shown to play important roles in lens development, including lens placode formation and fiber differentiation, and also induce changes characteristic of some forms of cataract. Previous studies demonstrated expression of TGFβ receptors during lens morphogenesis. However, the expression patterns of activin and BMP receptors or their signaling mediators, the Smad proteins, have not been well documented. In this study we examine the spatio-temporal expression patterns of activin receptors (ActRIIA, ActRIIB, ALK1, and ALK2), BMP receptors (BMPRII, ALK3, and ALK6), and the distribution of the phosphorylated forms of Smad1 and Smad2 during normal lens development (E12-P21) and aberrant development in transgenic mice that express dominant negative TGFβ receptors.
Methods: RT-PCR was used to identify receptor expression in total RNA isolated from P2 and P21 rat lenses. cDNAs were cloned and used for in situ hybridization analysis of spatio-temporal expression patterns in wild type and transgenic (OVE550 and OVE591) lenses. Expression of ALK3 was also examined by immunofluorescence and immunoblotting. Antibodies for phosphorylated forms of Smad1 and Smad2 were used to examine activation of BMP and activin signaling.
Results: RT-PCR of RNA from postnatal lenses showed distinct expression of ActRIIA, ActRIIB, BMPRII, and ALK3 but not ALK1, ALK2, or ALK 6. In situ hybridization with specific probes for BMPRII, ActRII, and ALK3 showed ubiquitous expression in ectoderm, lens pit, optic vesicle, and peri-optic mesenchyme during early lens formation at E12. During subsequent lens differentiation, from E14 onwards, expression of these receptors became increasingly restricted to the lens epithelium and to the equatorial region, including the germinative and transitional zones, where cells proliferate and commence differentiation, respectively. Expression for both receptors declined rapidly with fiber differentiation and maturation. Immunofluorescence with specific antibodies for phospho-Smad1 and phospho-Smad2 showed distinct localization of these signaling mediators in epithelial cells of the germinative zone and in fibers undergoing early differentiation in the transitional zone. Further investigation of the expression of these receptors in lenses of transgenic mice, which ectopically express a truncated TβRII, showed marked up regulation and aberrant expression of ALK3, but not BMPRII or ActRII.
Conclusions: These results indicate that multiple members of the TGFβ family have the potential to signal during lens fiber differentiation and suggest there may be cross-talk between different signaling pathways.
The TGFβ superfamily consists of three main groups of secreted polypeptide growth factors, TGFβs, activins, and BMPs. Members of the TGFβ family exert their biological actions by binding to and activating type I and type II receptors, which are transmembrane serine/threonine kinases. There are five type II (ActRIIA, ActRIIB, BMPRII, TβRII, and AMHRII) and seven type I receptors, known as activin-like kinases 1-7 (ALK1-ALK7). The prevailing view is that signal transduction involves binding of the ligand to the type II receptor, which then recruits a type I receptor. Ligand induced dimerization of the two receptors results in phosphorylation of serine and threonine residues in the intracellular glycine-serine rich (GS) domain of the type I receptor (reviewed in ). The activated type I receptor in turn recruits and phosphorylates specific members of the Smad family of cytoplasmic proteins, known as receptor regulated Smads (R-Smads). Activated R-Smads interact with the co-Smad (Smad4), and phosphorylation of Smad4 results in both of these cytoplasmic proteins being translocated to the nucleus to activate specific gene transcription (reviewed in ).
Biochemical and cell biological studies indicate that there are two distinct signaling pathways for TGFβ family members; one shared by TGFβ/activin and the other by BMPs. Smad2 and Smad3 transduce TGFβ/activin signals whereas Smad1, Smad5, and Smad8 transduce BMP type signals. However, there is evidence that different ligands can induce the formation of various heteromeric receptor complexes and, depending upon the receptor type in the complex, elicit activation of different signaling pathways and responses [1,2]. For example, Smad1 and Smad5 can transduce TGFβ signals via ALK1, in some cell types . TGFβ and activin show convergent signaling as both can activate signaling pathways through Smad2 and Smad3 by forming heterodimers with different receptor pairs (TβRII/ALK5 and ActRII/ALK4, respectively). Similarly, BMP2 and BMP4 bind BMPRII in concert with ALK3 or ALK6 to activate the Smad1 and Smad5 pathways. Divergent signaling has been shown for TGFβ as it can bind to TβRII/ALK1 to activate BMP signaling pathways via Smad1 and Smad5 . Similarly ALK2 in concert with ActRII can activate Smad1 and 5 .
Previous studies have provided evidence that members of the TGFβ growth factor super-family play important yet diverse roles in lens biology. TGFβ receptors (TβRII and ALK5) are expressed in lens fibers during development and become expressed in the epithelium during postnatal development . TGFβ1-3 induce cataractous changes in the postnatal (>P13) lens epithelium similar to those seen in forms of human subcapsular and after cataract [4-9]. Similarly, over expression of an active form of TGFβ1 in the lenses of transgenic mice results in the formation of opacities that show the characteristic features of subcapsular cataract, including expression of α-smooth muscle actin and extracellular matrix deposition . However, TGFβ expression in fibers and its modulation by bFGF  suggest roles in fiber differentiation. Inhibition of TGFβ signaling in lens fibers of transgenic mice by ectopic expression of truncated TGFβ receptors showed that TGFβ signaling plays a role in terminal fiber differentiation .
Targeted deletion of BMP4  and BMP7 [13-15] in mice resulted in failure of lens placode formation, indicating key roles for these molecules in lens induction. Inhibition of BMP signaling in the lens in both chick and mouse has recently implicated BMPs in lens fiber elongation [16,17]. While several studies have indicated a role for members of the TGFβ superfamily in lens biology, detailed knowledge about the expression of the various receptors is lacking or incomplete. Verscheuren et al.  investigated expression of the activin type I receptors ALK2 (ActRI) and ALK4 (ActRIB) mRNA in the mid-gestation (E12.5) mouse and documented ALK4 in the developing retina but neither receptor was apparent in lens. Similarly, Dewulf et al.  demonstrated expression of the BMP type I receptors, ALK3 (BMPR1A) and ALK6 (BMPRIB), in the neural retina at E12.5 but detected only ALK3 in the lens vesicle by in situ hybridization. ActRIIA mRNA has been detected in the chick embryo lens and cornea at stage 23 , but has not been reported in mouse or rat. Using monoclonal antibodies, Yoshikawa et al.  demonstrated reactivity for ALK2 in the developing eye in whole mounts of E9.5 and E10.5 mouse embryos. Similarly Obata et al.  used polyclonal antibodies to investigate expression of various TGFβ family receptors in adult rat eyes and detected expression of three Type II receptors (TβRII, ActRII, and BMPRII) and five type I receptors (ALK5, ALK4, ALK3, ALK6, and ALK1).
In this study we have screened for the expression of BMP and activin receptors in postnatal lenses by RT-PCR and used cDNA clones and antibodies of identified receptors to investigate their expression and activation during normal and aberrant lens differentiation.
All animal procedures were performed in accordance with the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research and the animal care guidelines published by the Institute for Laboratory Animal Research (Guide for the Care and Use of Laboratory Animals). All studies were approved by the Institutional Ethics Committees of the University of Sydney and the University of Melbourne.
Rabbit polyclonal antibodies to phosphorylated Smad2 were obtained from Upstate Biotechnology (catalog number 06-829; Lake Placid, NY) and from Abcam (catalog number Ab5490; Cambridge, UK). Both antibodies were raised to synthetic phospho-peptides of Smad2 including phospho-Ser465 and phospho-Ser467. Ab5490 has been shown to recognize phosphorylated Smad2 in western blots as a 55-60 kDA band and does not cross react with Smad3. An antibody to total Smad2 (catalog number 51-1300), raised against a 27 amino acid synthetic peptide derived from the MH1 domain of human Smad2 was obtained from Zymed (San Francisco, CA). This antibody has been confirmed to be specific by western blotting and recognizes a 58 kDA band for Smad 2. The phospho-Smad1 antibody (catalog number 06-702), raised to a synthetic peptide corresponding to amino acids 455-465 of human Smad1 protein was obtained from Upstate Biotechnology. It recognizes phosphorylated Smad1 (Ser463 and Ser465) with an apparent molecular weight of approximately 65-66 kDa. A goat polyclonal antibody to ALK3 (catalog number AF346) raised to human recombinant BMPR1A (ALK3) was obtained from R & D Systems (Minneapolis, MN).
Embryonic and ocular tissues were obtained from Wistar rats at the ages indicated and either used for extraction of RNA or fixed and processed for histology by routine protocols. Ocular tissues were also obtained from FVB wild type and transgenic mice that ectopically express a truncated TβRII in the lens fibers under the control of the αA-crystallin promoter . The lenses from these transgenic mice (OVE550 and OVE591) are characterized by defective terminal differentiation of fiber cells .
Total RNA was isolated from whole E14 rat embryos and from P3 and P21 lenses using TRI-Reagent (Sigma, Sydney, Australia). Total RNA was digested with RNAse free DNAseI (GE Healthcare, Castle Hill, NSW Australia) to remove any traces of contaminating genomic DNA and first strand cDNA was transcribed from 1-2 μg RNA using AMV reverse transcriptase (Promega, Sydney, Australia) according to the manufacturer's instructions. In control reactions AMV reverse transcriptase was omitted. Aliquots (2 μl) of first strand cDNA were used in PCR reactions to amplify cDNAs for ALK3, ALK6, BMPRII, ALK1, ALK2, ActRII, and the orphan receptor ALK7 using specific primers (Table 1).
PCR products for ALK3, ALK6, BMPR2, and ActRIIA (Table 1) were gel purified and cloned into pGEM-T transcription vector (Promega). The identity of the resulting clones was confirmed by restriction digest and by sequencing. Radiolabeled or digoxigenin (DIG) labeled riboprobes were transcribed from the linearized templates as described previously .
In Situ Hybridization
In situ hybridization was performed on neutral buffered, formalin fixed paraffin sections with either radiolabeled or DIG labeled riboprobes as described previously . Radiolabeled hybrids were detected by thin film autoradiography (NTB-2 emulsion; Kodak, Sydney, Australia) and DIG labeled hybrids reactivity was detected using an anti-digoxigenin, alkaline phosphatase linked antibody (Roche Diagnostics, Castle Hill, NSW, Australia) according to manufacturer's instructions.
Formalin fixed paraffin sections of rat ocular tissues were re-hydrated in PBS and incubated for 20 min (37 °C) in blocking reagent (3% goat serum or 3% horse serum in PBS with 0.1% BSA). Sections were incubated overnight at 4 °C with primary antibody to pSmad1, pSmad2, or ALK3 diluted to 5 μg/ml in blocking reagent. Reactivity was visualized with Alexa-488 conjugated goat anti-rabbit or rabbit anti-goat immunoglobulin (Molecular Probes, Eugene, OR) diluted 1:500 in PBS supplemented with 0.1% BSA. Control sections were incubated either with non-immune rabbit IgG or in the case of the goat antibody with normal goat serum. To visualize cell nuclei, sections were stained with 1μg/ml Hoechst stain (Calbiochem, La Jolla, CA).
Four to six lenses from FVB, OVE591 and OVE550 mice at P2-P3 were extracted in 100 μl lysis buffer (50 mM HEPES pH 7.5, 50 mM NaCl, 5 mM EDTA, 1% Triton-X, 50 mM sodium fluoride, and 10 mM sodium pyrophosphate, supplemented with 0.5% phosphatase inhibitor cocktail and 0.5% protease inhibitor cocktail mix [Sigma]), by about 5 s sonication on ice and stored in aliquots at -80 °C. Protein concentrations of extracts were assayed using a dye binding protein assay kit (PIERCE Coomassie Protein Assay Kit; Pierce, Rockford, IL), with bovine serum albumin (BSA) as a reference standard. Equivalent protein samples and molecular weight standards (ranging from 10-160 kDa; MBI Fermentas, Hanover, MD) were separated under reducing conditions on 10% SDS-polyacrylamide gels and transferred to nitrocellulose membranes. Efficient transfer of proteins was confirmed by Coomassie blue staining of gels. Nitrocellulose membranes were blocked overnight at 4 °C with either 5% skim milk powder (for ALK3 and Smad2 blots) or with 5% BSA (for pSmad1 and pSmad2 blots) in PBS with 0.1% Tween-20 (PBST). Blots were incubated for 1.5 h with agitation at RT with primary antibodies (1 μg/ml) for: ALK3 (AF346) in 3% skim milk in PBST, Smad2 (51-1300) in 3% skim milk in PBST, pSmad1 (06-702) in 2% BSA in PBST, or pSmad2 (Ab5490) in 2% BSA in PBST. Membranes were washed 3 times for 15 min each in PBST and incubated for 1.5 h with the appropriate horseradish peroxidase conjugated secondary antibody (Bio-Rad Laboratories, Hercules, CA) diluted 1:5,000 in 3% skim milk in PBST or in 2% BSA in PBST. Reactivity was detected using an enhanced chemiluminescence kit (Amersham-Pharmacia Biotech, Arlington Heights, IL).
Expression of BMP/activin receptor mRNA in lens
To examine the expression of BMP and activin receptors in the postnatal lens, RT-PCR was carried out using specific primers on RNA obtained from whole E14 rat embryos and from P3 and P21 rat lenses. Specific amplicons for all receptors were detected in RNA from whole E14 embryos but only amplicons for ALK3, BMPRII, ActRIIA, and ActRIIB were detected in RNA from P3 and P21 lenses (Table 1). The PCR fragments for these three receptors were cloned and sequenced to confirm their identity.
Expression of BMP receptors during lens morphogenesis
In situ hybridization with specific radiolabeled riboprobes was carried out to examine the spatio-temporal expression patterns of ALK3 and BMPRII during development of the lens from E12 to P21 (Figure 1). Experiments using the control sense riboprobes showed no hybridization signals (not shown).
At E12 (Figure 1A) the lens ectoderm has invaginated to form the lens pit or early lens vesicle (lv). Both ALK3 (Figure 1B) and BMPRII (Figure 1C) showed ubiquitous expression, including the lens vesicle (lv) and optic cup (oc), At E14, (Figure 1D) the lens vesicle is filled by the elongating primary fiber cells and there is a clearly defined anterior epithelium. Both ALK3 (Figure 1E) and BMPRII (Figure 1F) were expressed in the lens epithelium (le) and differentiating fibers (lf); although signal was reduced in the centrally situated (most mature) primary fibers (Figure 1E,F; asterisks). Transcripts for both receptors were detected in the optic cup (oc) and extra-ocular mesenchyme (m) with strong expression for ALK3 in the developing optic nerve (on) and for BMPRII along the inner optic cup (oc), where retinal ganglion cells start to differentiate (Figure 1F). At E18, ALK3 was uniformly expressed in most ocular tissues, including the lens, neural retina, cornea, iris, ciliary body, and the eyelids (Figure 1H). In the retina, expression was uniform throughout the retina including the neuroblast layer and the differentiated ganglion cell layer. In the lens, ALK3 was expressed in the anterior epithelium and in the equatorial region extending into the transitional zone. Little or no expression was detected in the cortical and mature fibers. BMPRII was also expressed in most ocular tissues but showed different levels of expression in some regions. For example, in the retina there was strong expression in the ganglion cell layers but weaker expression in the outer neuroblast layer (Figure 1I). In the lens, expression of BMPRII was similar to ALK3, being detectable in the anterior lens epithelium and in the transitional zone but not in the central differentiated fibers. Similarly, in lenses from weanling (P21) animals (Figure 1K-L), expression of ALK3 and BMPRII was detectable in the epithelium and in the transitional zone but little or no expression was detected in the mature fibers. Expression was also detected in the ciliary body, iris, and corneal epithelium and endothelium.
Expression of ActRIIA during lens morphogenesis
In situ hybridization with specific DIG labeled riboprobes was carried out to examine the spatio-temporal expression pattern of ActRIIA during lens development (E12 to P21). Experiments using the control sense riboprobes showed no hybridization signals (not shown). At E12 ActRIIA was ubiquitously expressed in most developing ocular tissues, including the lens vesicle (lv), optic cup (oc), and peri-ocular mesenchyme (m). Strongest expression appeared to be associated with the anterior cells of the lens vesicle (Figure 2A, arrowhead). From E14 to P21 expression of ActRIIA in the lens was similar, with expression detectable in the anterior epithelial cells and the early differentiating fiber cells in the transitional zone below the lens equator (Figure 2B-F). Little or no expression was detected in the mature fibers (Figure 2B-F).
In situ hybridization with radiolabeled riboprobes to ALK4 showed only very weak expression of this receptor in lens epithelium and fibers in E14 and E18 embryos (not shown).
Localization of phospho-Smads indicates activation of BMP/Activin signaling pathways
To investigate whether BMP or activin signaling pathways are activated in lens cells we immunolocalized phosphorylated Smad1 and Smad2 in formalin fixed, paraffin embedded sections. Smad1 is a specific receptor associated Smad for the BMP signaling pathway, whereas Smad2 is a specific receptor associated Smad for the activin/TGFβ signaling pathways. Reactivity for phosphorylated forms of Smad1 and Smad2 were predominantly detected in the nuclei of both lens epithelial and fiber cells at E18 (Figure 3). For phospho-Smad2 (pSmad2), similar patterns of reactivity were obtained with both antibodies used. Most prominent expression for pSmad1 and pSmad2 was detected in the nuclei of epithelial cells in the germinative and transitional zones. Weaker reactivity for both pSmads was also detected along the membranes of epithelial cells and fibers cells in the transitional zone (Figure 3). Similar results were obtained in lenses at E18 and P21 (not shown). No reactivity was detected when sections were incubated with non-immune IgG (Figure 3E).
The presence of these signaling mediators in the nuclei suggests activation of the respective BMP and activin signaling pathways in cells that are proliferating and undergoing early stages of lens fiber differentiation.
Altered expression of ALK3 in lenses with disrupted TGFβ signaling
Together with previous data, the results of the present study indicate that multiple members of the TGFβ family can potentially signal during lens development. Previous studies demonstrated that TGFβ receptors are expressed in the lens  in patterns similar to those shown here for BMP/activin receptors. Furthermore, signaling via TGFβ receptors is required during lens fiber differentiation  as expression of truncated TGFβ receptors in lenses of transgenic mice resulted in attenuated fiber differentiation, apoptosis, and fiber degeneration. As expected, comparison of levels of pSmad2 in wild type (FVB) and transgenic lenses (OVE550 and OVE591) revealed that, while total levels of Smad2 were similar, the levels of pSmad2 were markedly reduced in OVE550 and OVE591 lenses (Figure 4).
To investigate whether disruption of TGFβ signaling may impact on the expression of BMP and activin receptors, we examined the expression of ALK3, BMPRII, and ActRIIA in lenses from two lines of transgenic mice (OVE591 and OVE550). These mice express a truncated TGFβ type II receptor in the lens fibers, under the control of the αA-crystallin promoter. The different levels of transgene expression in these lines results in slightly different timing of phenotype onset . In P1 lenses of OVE591 mice, which have a lower level of transgene expression, the phenotype is not yet apparent (Figure 5). However, ALK3 was aberrantly expressed in the cortical fibers (Figure 5B; arrowheads) of these transgenic lenses. Expression of ALK3 in the lens epithelium and in other ocular tissues was similar to that shown in rat lenses (Figure 1) and in wild type mouse lenses (not shown). In P1 lenses of OVE550 mice, which have a higher level of transgene expression, the phenotype is clearly apparent with abnormal nuclear migration, pyknotic fiber nuclei and disrupted nuclear fibers (Figure 5C,E; asterisk). ALK3 was highly and aberrantly expressed in the cortical and nuclear fibers of these lenses (Figure 5D). At higher magnification strongest expression for ALK3 was detected in fibers about to undergo apoptosis and degeneration (Figure 5E,F; small arrowheads). Fibers that had undergone apoptosis and degeneration, no longer showed any signal for ALK3 (Figure 5E,F; asterisk). The anterior epithelium of these lenses, where the transgene is not expressed, showed slightly decreased expression of ALK3 compared with wild type lenses (Figure 5E,F; large arrowhead). Interestingly, there was also elevated expression of ALK3 in the presumptive ciliary body and iris (Figure 5E,F), which also do not express the transgene. By contrast, the expression of BMPRII (Figure 5H) and ActRIIA (not shown) were not altered significantly in the fibers of transgenic lenses compared to wild type. Similar to ALK3, regions of the anterior epithelium, where the transgene is not expressed, showed decreased expression of BMPRII (Figure 5H; arrowhead) compared with wild type lenses and with adjacent regions of epithelium. Expression of BMPRII was also detected in the tunica vasculosa lentis (Figure 5H; arrow).
Immunolocalization and western blotting experiments confirmed the expression patterns of ALK3 in the wild type lens and that ALK3 expression was elevated in transgenic lenses. By western blotting of whole lens extracts, similar levels of ALK3 were detected in wild type and OVE591 lenses but significantly increased levels were detected in OVE550 lenses (not shown). Similarly, immunofluorescence experiments showed distinct reactivity for ALK3 in the epithelium (Figure 6B; inset) and cortical fibers of wild type lenses but markedly increased staining of groups of cortical fibers (Figure 6B; arrows) and also the ciliary body of OVE550 eyes (Figure 6B). Controls using non-immune goat serum showed no specific reactivity (Figure 6C).
Previous studies have indicated that members of the TGFβ superfamily play important roles in lens development (see Introduction). In particular BMP signaling is thought to be involved in lens induction [12-15] and in lens fiber elongation during differentiation [16,17]. TGFβ signaling, on the other hand, has distinct and dichotomous roles; having been implicated in the induction of an epithelial-mesenchymal transition in the anterior lens epithelium [4-9] but being required for terminal differentiation of lens fibers .
In this study we have documented the expression of several TGFβ superfamily receptors (BMP and activin receptors) in the lens during rat development. RT-PCR experiments detected mRNA for ActRIIA, ActRIIB, BMPRII and ALK3 but not ALK1, ALK2, ALK6, or ALK7 in postnatal lenses. By in situ hybridization these receptors showed similar patterns of expression in lens, being expressed mainly in the lens epithelium and transitional zone fibers. These data are consistent with reports of immunoreactivity for BMPRII, ALK3, and ActRII in adult rat lens epithelia  and mRNA for ALK3 in embryonic lens . However, Obata et al.  additionally reported reactivity for ALK6, ALK2, and ALK4 in adult lens epithelium and ALK2 reactivity has also been reported in the pre-lens ectoderm at the lens placode stage . In contrast, while the primers used in our study successfully amplified transcripts from whole (E14), we did not detect expression of ALK6 (BMPRIB) or ALK2 by RT-PCR of neonatal and weanling rat lens RNA and in situ hybridization revealed only very weak expression of ALK4 in embryonic lenses. Consistent with this, previous in situ hybridization experiments did not detect transcripts for ALK6  or ALK4  in the embryonic lens. The finding of weak ALK6 and ALK2 reactivity in adult rat lens  may reflect cross reactivity or non-specific reactivity of the antibodies used in that study.
The results of this study show that lens cells can potentially respond to activin as well TGFβ and BMP signals. Ligand binding receptors (type II) for all three types of growth factors (activin, BMP, and TGFβ) are expressed and the phospho-Smad1 and Smad2 antibodies indicate that both BMP-like signals and activin/TGFβ-like signals are being transmitted in cells within the germinative and transitional zones. Expression and activation of these pathways in these regions of the lens suggest an involvement in epithelial proliferation and fiber differentiation. These results are consistent with findings from transgenic mice, which showed a requirement for BMP signaling for lens fiber elongation  and for TGFβ in lens fiber differentiation .
Activin subunits are expressed during development of the chick eye and activin α subunit expression has been reported in the chick lens vesicle . However, the roles of activin signaling in lens development have not yet been addressed. Targeted inactivation of activin βA or βB genes affects craniofacial development . Apart from failure of eyelid fusion in activin βB-/- mice [24,25] and a decrease in the number of rod photoreceptors in activin βA-/- mice  no other ocular defects have been reported. However, as α, βA, and βB subunits are expressed in the eye  there may be functional redundancy for these ligands. Indeed it has been reported that expression of activin βA in place of activin βB can rescue the craniofacial phenotype of activin βB-/- mice  suggesting that these factors can functionally compensate for each other. Null mutations of individual type II activin receptors have been generated but as yet few data are available about their effects on ocular development. ActRIIA-/- mice develop to term but exhibit defects of mandibles and testes and female sterility . ActRIIB-/- animals die postnatally and have defects of the axial skeleton and disturbance of left-right symmetry . Recent preliminary observations of double mutant mice (ActRIIA+/-/ActRIIB-/-) that survive until birth, suggest disturbances of ocular development (Christine Ferguson, Paul Sharpe, personal communication, January, 2002) . However, more detailed analyses of mice with these double mutations are needed to determine whether activin signaling plays a role in ocular development.
The putative role of BMP signaling in lens differentiation has been addressed recently by two separate studies. In experiments using noggin (a secreted antagonist of BMPs), either over expressed in chick embryo eyes or in vitro, Belecky-Adams et al.  showed that signaling via BMPs participates in the elongation and differentiation of lens fibers. (ALK6DN) driven by two different promoters (αA-crystallin and the Pax6 ectoderm enhancer) that inhibition of BMP signaling results in an asymmetric attenuation of lens differentiation in the ventro-nasal hemisphere of the lens. They suggested that asymmetry of the phenotype was related to the patterns of expression of BMP ligands in the eye (for example, BMP4 is predominantly expressed in the dorsal optic cup  during early fiber differentiation) and also to the relative affinities of the several possible receptor heterodimers (ALK2/BMPR2, ALK3/BMPR2) that may occur in the lens for these BMP ligands. It is of interest that the experiments performed by Faber et al.  used a dominant negative construct (ALK6DN) of ALK6, which our data shows not to be expressed during lens development. Heterodimers formed with this receptor are likely to have different affinities for BMP ligands than do heterodimers formed with ALK3 or ALK2 [2,16]. As different ligands (BMP4 and BMP7) are potentially available in different quadrants of the developing eye [12,16], the dominant negative effect of ALK6DN will be determined by the affinity of receptor heterodimers that it forms for the available BMP ligands. Hence it is possible that the asymmetric phenotype reflects the differential distribution of BMPs in the ocular environment and their affinity for ALK6DN receptor complexes and not the intrinsic capacity of lens cells to respond to those ligands. Another consideration is that dominant negative constructs of type I receptors are often less effective than those of type II receptors [11,30]. It would be of interest to determine whether dominant negative constructs of BMPRII or ALK3 result in a similar asymmetric lens phenotype.
The attenuated fiber cell phenotype caused by the truncated ALK6 receptor has similarities to the attenuated fiber phenotype seen in embryonic mice expressing the dominant negative TβRII receptor . The similarity of phenotypes between these two lines of mice suggests that there may be cross-talk between the TGFβ and BMP signaling pathways. The results of the present study may provide further evidence of such cross-talk as transgenic lenses, which express dominant negative TβRII (OVE550 and OVE591), have increased and aberrant patterns of ALK3 (but not BMPR2 or ActRIIA) expression in the cortical fibers. It is unclear from the current studies how or why ALK3 (a BMP receptor) is upregulated by dominant negative TβRII expression. The occurrence of the upregulated expression in fibers before they begin to degenerate is suggestive of a compensatory response by fibers to the loss of TGFβ signaling, such that TGFβ signaling directly or indirectly regulates ALK3 receptor expression and hence signaling. There is evidence from other systems that there is direct "cross-talk" intracellularly between BMP and TGFβ signaling pathways at the level of Smad4 and via the inhibitory Smad6 and Smad7 , though it remains to be shown that regulation of signaling at this level results in altered BMP receptor expression. An alternative explanation is that the high levels of transgene expression achieved by the αA-crystallin promoter result in the dominant negative TβRII dimerizing with ALK3, thus inhibiting BMP as well as TGFβ signaling. In this scenario the ALK3 upregulation might be due to feedback from lack of BMP signaling. However, while there is evidence for TβRII and ALK1 dimerizing to initiate BMP signals, there is currently no evidence that TβRII and ALK3 can form such a complex. In addition, this model does not account for the increases in ALK3 expression detected in the ciliary and iridial retina, which do not express the dominant negative TβRII transgene. Indeed the response in these non-transgene expressing tissues suggests another possible effect of the dominant negative TβRII may be at the level of inhibiting or sequestering ligands in the ocular environment. Clearly further studies will be required to elucidate the basis for these complex alterations in receptor expression in and around the lens.
In summary, this study has shown that the lens expresses receptors for activins and BMPs in patterns that suggest they play roles in lens differentiation. Furthermore the potential for cross-talk in signaling via these receptors suggest that members of the TGFβ superfamily play complex and complementary roles in lens development.
AcknowledgementsSupported by grants from the National Institutes of Health (EY03177) USA, the National Health and Medical Research Council (NHMRC) of Australia, and the Sydney Foundation for Medical Research. RdI was supported by a NHMRC Postdoctoral Fellowship and by a Sesqui grant from the University of Sydney. YC was supported by an Australian Research Council PhD Scholarship.
1. Miyazawa K, Shinozaki M, Hara T, Furuya T, Miyazono K. Two major Smad pathways in TGF-beta superfamily signalling. Genes Cells 2002; 7:1191-204.
2. Macias-Silva M, Hoodless PA, Tang SJ, Buchwald M, Wrana JL. Specific activation of Smad1 signaling pathways by the BMP7 type I receptor, ALK2. J Biol Chem 1998; 273:25628-36.
3. Oh SP, Seki T, Goss KA, Imamura T, Yi Y, Donahoe PK, Li L, Miyazono K, ten Dijke P, Kim S, Li E. Activin receptor-like kinase 1 modulates transforming growth factor-beta 1 signaling in the regulation of angiogenesis. Proc Natl Acad Sci U S A 2000; 97:2626-31.
4. de Iongh RU, Gordon-Thomson C, Chamberlain CG, Hales AM, McAvoy JW. Tgfbeta receptor expression in lens: implications for differentiation and cataractogenesis. Exp Eye Res 2001; 72:649-59.
5. Hales AM, Chamberlain CG, McAvoy JW. Cataract induction in lenses cultured with transforming growth factor-beta. Invest Ophthalmol Vis Sci 1995; 36:1709-13.
6. Hales AM, Chamberlain CG, McAvoy JW. Susceptibility to TGFbeta2-induced cataract increases with aging in the rat. Invest Ophthalmol Vis Sci 2000; 41:3544-51.
7. Hales AM, Schulz MW, Chamberlain CG, McAvoy JW. TGF-beta 1 induces lens cells to accumulate alpha-smooth muscle actin, a marker for subcapsular cataracts. Curr Eye Res 1994; 13:885-90.
8. Liu J, Hales AM, Chamberlain CG, McAvoy JW. Induction of cataract-like changes in rat lens epithelial explants by transforming growth factor beta. Invest Ophthalmol Vis Sci 1994; 35:388-401.
9. Gordon-Thomson C, de Iongh RU, Hales AM, Chamberlain CG, McAvoy JW. Differential cataractogenic potency of TGF-beta1, -beta2, and -beta3 and their expression in the postnatal rat eye. Invest Ophthalmol Vis Sci 1998; 39:1399-409.
10. Srinivasan Y, Lovicu FJ, Overbeek PA. Lens-specific expression of transforming growth factor beta1 in transgenic mice causes anterior subcapsular cataracts. J Clin Invest 1998; 101:625-34.
11. de Iongh RU, Lovicu FJ, Overbeek PA, Schneider MD, Joya J, Hardeman ED, McAvoy JW. Requirement for TGFbeta receptor signaling during terminal lens fiber differentiation. Development 2001; 128:3995-4010.
12. Furuta Y, Hogan BL. BMP4 is essential for lens induction in the mouse embryo. Genes Dev 1998; 12:3764-75.
13. Dudley AT, Lyons KM, Robertson EJ. A requirement for bone morphogenetic protein-7 during development of the mammalian kidney and eye. Genes Dev 1995; 9:2795-807.
14. Luo G, Hofmann C, Bronckers AL, Sohocki M, Bradley A, Karsenty G. BMP-7 is an inducer of nephrogenesis, and is also required for eye development and skeletal patterning. Genes Dev 1995; 9:2808-20.
15. Wawersik S, Purcell P, Rauchman M, Dudley AT, Robertson EJ, Maas R. BMP7 acts in murine lens placode development. Dev Biol 1999; 207:176-88.
16. Faber SC, Robinson ML, Makarenkova HP, Lang RA. Bmp signaling is required for development of primary lens fiber cells. Development 2002; 129:3727-37.
17. Belecky-Adams TL, Adler R, Beebe DC. Bone morphogenetic protein signaling and the initiation of lens fiber cell differentiation. Development 2002; 129:3795-802.
18. Verschueren K, Dewulf N, Goumans MJ, Lonnoy O, Feijen A, Grimsby S, Vandi Spiegle K, ten Dijke P, Moren A, Vanscheeuwijck P, Heldind C-H, Miyazonod K, Mummeryc C, Raaijc JVDE-V, Huylebroeck D. Expression of type I and type IB receptors for activin in midgestation mouse embryos suggests distinct functions in organogenesis. Mech Dev 1995; 52:109-23.
19. Dewulf N, Verschueren K, Lonnoy O, Moren A, Grimsby S, Vande Spiegle K, Miyazono K, Huylebroeck D, Ten Dijke P. Distinct spatial and temporal expression patterns of two type I receptors for bone morphogenetic proteins during mouse embryogenesis. Endocrinology 1995; 136:2652-63.
20. Ohuchi H, Noji S, Koyama E, Myokai F, Nishikawa K, Nohno T, Tashiro K, Shiokawa K, Matsuo N, Taniguchi S. Expression pattern of the activin receptor type IIA gene during differentiation of chick neural tissues, muscle and skin. FEBS Lett 1992; 303:185-9.
21. Yoshikawa SI, Aota S, Shirayoshi Y, Okazaki K. The ActR-I activin receptor protein is expressed in notochord, lens placode and pituitary primordium cells in the mouse embryo. Mech Dev 2000; 91:439-44.
22. Obata H, Kaji Y, Yamada H, Kato M, Tsuru T, Yamashita H. Expression of transforming growth factor-beta superfamily receptors in rat eyes. Acta Ophthalmol Scand 1999; 77:151-6.
23. Belecky-Adams TL, Scheurer D, Adler R. Activin family members in the developing chick retina: expression patterns, protein distribution, and in vitro effects. Dev Biol 1999; 210:107-23.
24. Matzuk MM, Kumar TR, Bradley A. Different phenotypes for mice deficient in either activins or activin receptor type II. Nature 1995; 374:356-60.
25. Vassalli A, Matzuk MM, Gardner HA, Lee KF, Jaenisch R. Activin/inhibin beta B subunit gene disruption leads to defects in eyelid development and female reproduction. Genes Dev 1994; 8:414-27.
26. Davis AA, Matzuk MM, Reh TA. Activin A promotes progenitor differentiation into photoreceptors in rodent retina. Mol Cell Neurosci 2000; 15:11-21.
27. Brown CW, Houston-Hawkins DE, Woodruff TK, Matzuk MM. Insertion of Inhbb into the Inhba locus rescues the Inhba-null phenotype and reveals new activin functions. Nat Genet 2000; 25:453-7.
28. Ferguson CA, Tucker AS, Heikinheimo K, Nomura M, Oh P, Li E, Sharpe PT. The role of effectors of the activin signalling pathway, activin receptors IIA and IIB, and Smad2, in patterning of tooth development. Development 2001; 128:4605-13.
29. Dudley AT, Robertson EJ. Overlapping expression domains of bone morphogenetic protein family members potentially account for limited tissue defects in BMP7 deficient embryos. Dev Dyn 1997; 208:349-62.
30. Brand T, Schneider MD. Inactive type II and type I receptors for TGF beta are dominant inhibitors of TGF beta-dependent transcription. J Biol Chem 1995; 270:8274-84.