Bi-tangential hybrid IMRT for sparing the shoulder in whole breast irradiation.

BACKGROUND AND PURPOSE: A bi-tangential technique is proposed to reduce undesired doses to the shoulder produced by standard tangential irradiation. PATIENTS AND METHODS: A total of 6 patients affected by shoulder pain and reduced functional capacity after whole-breast irradiation were retrospectively analysed. The standard tangential plan used for treatment was compared with (1) a single bi-tangential plan where, to spare the shoulder, the lateral open tangent was split into two half-beams at isocentre, with the superior portion rotated by 10-20° medially with respect to the standard lateral beam; (2) a double bi-tangential plan, where both the tangential open beams were split. The planning target volume (PTV) coverage and the dose to the portion of muscles and axilla included in the standard tangential beams were compared. RESULTS: PTV95 % of standard plan (91.9 ± 3.8) was not significantly different from single bi-tangential plan (91.8 ± 3.4); a small but significant (p < 0.01) decrease was observed with the double bi-tangential plan (90.1 ± 3.7). A marked dose reduction to the muscle was produced by the single bi-tangential plan around 30-40 Gy. The application of the double bi-tangential technique further reduced the volume receiving around 20 Gy, but did not markedly affect the higher doses. The dose to the axilla was reduced both in the single and the double bi-tangential plans. CONCLUSION: The single bi-tangential technique would have been able to reduce the dose to shoulder and axilla, without compromising target coverage. This simple technique is valuable for irradiation after axillary lymph node dissection or in patients without dissection due to negative or low-volume sentinel lymph node disease.

Liprin-α1 regulates breast cancer cell invasion by affecting cell motility, invadopodia and extracellular matrix degradation.

Migration of cells and degradation of the extracellular matrix (ECM) are required for efficient tumor cell invasion, but the underlying molecular mechanisms are only partially known. The PPFIA1 gene for liprin-α1 is frequently amplified in human breast cancers. We recently demonstrated that liprin-α1 is an important regulator of cell edge dynamics during motility. We show, herein, that the liprin-α1 protein is highly expressed in human breast tumors. Functional analysis shows that liprin-α1 is specifically required for the migration and invasion of highly invasive human breast cancer MDA-MB-231 cells. We used time-lapse analysis to demonstrate defects in the motility of liprin-α1-depleted cells that include a striking instability of the lamellipodia. Liprin-α1 levels altered by either RNA interference or overexpression affected also cell spreading and the number of invadopodia per cell, but not the density of invadopodia per unit of surface area. On the other hand, silencing of liprin-α1 inhibited the degradation of the ECM, whereas its overexpression enhanced degradation, resulting in significant negative or positive effects, respectively, on the area of degradation per invadopodium. Transfection of fluorescent-labeled cortactin revealed that depletion of liprin-α1 also affected the assembly and disassembly of invadopodia, with decrease of their lifetime. Our results strongly support a novel important role of liprin-α1 in the regulation of human tumor cell invasion.

Cell migration: GAPs between membrane traffic and the cytoskeleton.

During cell migration, coordination between membrane traffic, cell substrate adhesion and actin reorganization is required for protrusive activity to occur at the leading edge. Actin organization is regulated by Rho family GTPases and, with a contribution from the endocytic cycle, serves to extend the cell front. The details of the molecular mechanisms that direct membrane traffic at sites of adhesion and rearrange actin at the cell edge are still unknown. However, recent findings show that a number of multi-domain proteins characterized by an ArfGAP domain interact with both actin-regulating and integrin-binding proteins, as well as affecting Rac-mediated protrusive activity and cell migration. Some of these proteins have been shown to localize to endocytic compartments and to have a role in regulating endocytosis. Given the participation of Arf proteins in regulating membrane traffic, one appealing hypothesis is that the ArfGAPs act as molecular devices that coordinate membrane traffic and cytoskeletal reorganization during cell motility.

Molecular mechanisms regulating the subcellular localization of p95-APP1 between the endosomal recycling compartment and sites of actin organization at the cell surface.

Cell migration requires coordination between adhesion, actin organization and membrane traffic. Rac and ARF6 have been shown to cooperate for the organization of actin at the cell surface. Recently, the GIT family of ARF-GAPs has been identified, which includes proteins that can functionally interact with both ARF and Rac GTPases. The p95-APP1 protein is a member of this family, isolated as part of a multi-molecular complex interacting with GTP-Rac. Our previous work has indicated that this protein may be part of the machinery redirecting membrane recycling towards sites of protrusion during locomotion. By analyzing the distribution and the effects of truncated forms of p95-APP1, we show here that the lack of the ARF-GAP domain of p95-APP1 dramatically shifts its localization to large vesicles. The use of several markers of the endocytic pathway has revealed that the truncated p95-APP1 localizes specifically to a Rab11-, transferrin receptor-positive compartment. Other markers are excluded from the p95-APP1-positive vesicles, while known components of the multi-molecular complex colocalize with truncated p95-APP1 in this compartment. Coexpression of a constitutively active form of Rac induces the redistribution of the truncated constructs and of the associated PIX, PAK, and paxillin to peripheral sites of Rac-mediated actin organization, and the disassembly of the large Rab11-positive vesicles. Together, the data presented indicate that p95-APP1 is part of a complex that shuttles between the plasma membrane and the endocytic recycling compartment, and suggest that the dynamic redistribution of the p95-APP1-containing complex is mediated both by the ARF-GAP domain, and by the recruitment of the complex at the cell surface at sites of Rac activation.

p95-APP1 links membrane transport to Rac-mediated reorganization of actin.

Motility requires protrusive activity at the cellular edge, where Rho family members regulate actin dynamics. Here we show that p95-APP1 (ArfGAP-putative, Pix-interacting, paxillin-interacting protein 1), a member of the GIT1/PKL family, is part of a complex that interacts with Rac. Wild-type and truncated p95-APP1 induce actin-rich protrusions mediated by Rac and ADP-ribosylation factor 6 (Arf6). Distinct p95-APP1-derived polypeptides have different distributions, indicating that p95-APP1 cycles between the cell surface and endosomes. Our results show that p95-APP1 functionally interacts with Rac and localizes to endosomal compartments, thus identifying p95-APP1 as a molecular link between actin organization, adhesion, and membrane transport during cell motility.

Estradiol induces differential neuronal phenotypes by activating estrogen receptor alpha or beta.

Estrogens are female sex steroids that have a plethora of effects on a wide range of tissues. These effects are mediated through two well characterized intracellular receptors: estrogen receptor alpha and beta (ERalpha and ERbeta, respectively). Because of their high structural homology, it has been argued whether these two receptors may elicit differential biochemical events in estrogen target cells. Here we examine the effect of 17beta-estradiol-dependent activation of ERalpha and ERbeta on neurite sprouting, a well known consequence of this sex hormone action in neural cells. In SK-N-BE neuroblastoma cells transfected with ERalpha or ERbeta, 17beta-estradiol induces two distinct morphological phenotypes. ERalpha activation results in increased length and number of neurites, whereas ERbeta activation modulates only neurite elongation. By the use of chimeric receptors we demonstrate that the presence of both transcription activation functions located in the NH2-terminus and COOH-terminus of the two ER proteins are necessary for maintaining the differential biological activity reported. ERalpha-dependent, but not ERbeta-dependent, morphological changes are observed only in the presence of the active form of the small G protein Rac1B. Our data provide the first clear evidence that, in a given target cell, ERalpha and ERbeta may play distinct biological roles and support the hypothesis that 17beta-estradiol activates selected intracellular signaling pathways depending on the receptor subtype bound.

Rac GTPases localize at sites of actin reorganization during dynamic remodeling of the cytoskeleton of normal embryonic fibroblasts.

Rac GTP-binding proteins are implicated in the dynamic organization of the actin cytoskeleton, and the mechanisms utilized for this purpose are not understood yet. In this paper we have analysed the effects of the expression of Rac proteins on the organization of the cytoskeleton, and their subcellular distribution in chicken embryo fibroblasts. In these cells, overexpression of wild-type Rac GTPases induces disassembly of stress fibers, and production of long, highly branched actin-rich protrusions, with consequent dramatic changes in cell morphology. The formation of these protrusions is mediated by adhesion to the substrate, and is prevented by incubation with anti-(beta)1 function-blocking antibodies. Rac-mediated cell shape changes require a wild-type GTPase, since expression of constitutively active V12-Rac proteins affects actin organization differently in these cells, without causing alterations in their morphology. Localization studies performed on ventral plasma membranes from fibroblasts transfected with wild-type or mutant GTPases show codistribution of Rac along stress fibers, before their disassembly and the formation of the actin-rich protrusions. These data show a link between Rac protein distribution, and their effects on the actin cytoskeleton. Altogether, our results are indicative of an active role of Rac proteins in stress fiber disassembly, and show that Rac, which can cycle its bound nucleotide, produces unique dynamic effects on actin organization.

A cell-free system to study regulation of focal adhesions and of the connected actin cytoskeleton.

Assembly and modulation of focal adhesions during dynamic adhesive processes are poorly understood. We describe here the use of ventral plasma membranes from adherent fibroblasts to explore mechanisms regulating integrin distribution and function in a system that preserves the integration of these receptors into the plasma membrane. We find that partial disruption of the cellular organization responsible for the maintenance of organized adhesive sites allows modulation of integrin distribution by divalent cations. High Ca2+ concentrations induce quasi-reversible diffusion of beta1 integrins out of focal adhesions, whereas low Ca2+ concentrations induce irreversible recruitment of beta1 receptors along extracellular matrix fibrils, as shown by immunofluorescence and electron microscopy. Both effects are independent from the presence of actin stress fibers in this system. Experiments with cells expressing truncated beta1 receptors show that the cytoplasmic portion of beta1 is required for low Ca2+-induced recruitment of the receptors to matrix fibrils. Analysis with function-modulating antibodies indicates that divalent cation-mediated receptor distribution within the membrane correlates with changes in the functional state of the receptors. Moreover, reconstitution experiments show that purified alpha-actinin colocalizes and redistributes with beta1 receptors on ventral plasma membranes depleted of actin, implicating binding of alpha-actinin to the receptors. Finally, we found that recruitment of exogenous actin is specifically restricted to focal adhesions under conditions in which new actin polymerization is inhibited. Our data show that the described system can be exploited to investigate the mechanisms of integrin function in an experimental setup that permits receptor redistribution. The possibility to uncouple, under cell-free conditions, events involved in focal adhesion and actin cytoskeleton assembly should facilitate the comprehension of the underlying molecular mechanisms.

Overexpression of a neural-specific rho family GTPase, cRac1B, selectively induces enhanced neuritogenesis and neurite branching in primary neurons.

Rho family GTPases have been implicated in cytoskeletal reorganization during neuritogenesis. We have recently identified a new gene of this family, cRac1B, specifically expressed in the chicken developing nervous system. This GTPase was overexpressed in primary neurons to study the role of cRac1B in the development of the neuronal phenotype. Overexpression of cRac1B induced an increment in the number of neurites per neuron, and dramatically increased neurite branching, whereas overexpression of the highly related and ubiquitous cRac1A GTPase did not evidently affect neuronal morphology. Furthermore, expression of an inactive form of cRac1B strikingly inhibited neurite formation. The specificity of cRac1B action observed in neurons was not observed in fibroblasts, where both GTPases produced similar effects on cell morphology and actin organization, indicating the existence of a cell type-dependent specificity of cRac1B function. Molecular dissection of cRac1B function by analysis of the effects of chimeric cRac1A/cRac1B proteins showed that the COOH-terminal portion of cRac1B is essential to induce increased neuritogenesis and neurite branching. Considering the distinctive regulation of cRac1B expression during neural development, our data strongly support an important role of cRac1B during neuritogenesis, and they uncover new mechanisms underlying the functional specificity of distinct Rho family GTPases.

Differential expression of distinct members of Rho family GTP-binding proteins during neuronal development: identification of Rac1B, a new neural-specific member of the family.

Previous studies on small GTP-binding proteins of the Rho family have revealed their involvement in the organization of cell actin cytoskeleton. The function of these GTPases during vertebrate development is not known. With the aim of understanding the possible role of these proteins during neuronal development, we have cloned and sequenced five members expressed in developing chick neural retinal cells. We have identified four chicken genes, cRhoA, cRhoB, cRhoC, and cRac1A, homologous to known human genes, and a novel Rac gene, cRac1B. Analysis of the distribution of four of the identified transcripts in chicken embryos shows for the first time high levels of expression of Rho family genes in the vertebrate developing nervous system, with distinct patterns of distribution for the different transcripts. In particular, cRhoA and cRac1A gene expression appeared ubiquitous in the whole embryo, and the cRhoB transcript was more prominent in populations of neurons actively extending neurites, whereas the newly identified cRac1B gene was homogeneously expressed only in the developing nervous system. Temporal analysis of the expression of the five genes suggests a correlation with the morphogenetic events occurring within the developing retina and the retinotectal pathway. Expression of an epitope-tagged cRac1B in retinal neurons showed a diffuse distribution of the protein in the cell body and along neurites. Taken as a whole, our results suggest important roles for ubiquitous and neural-specific members of the Rho family in the acquisition of the mature neuronal phenotype.

Preferential localization of tyrosine-phosphorylated paxillin in focal adhesions.

Focal adhesions are sites for integrin-mediated attachment of cultured cells to the extracellular matrix. Localization studies have shown that focal adhesions can be stained by antiphosphotyrosine antibodies, but the role of tyrosine-phosphorylated proteins in focal adhesions is not known. By using ventral plasma membranes prepared from chicken embryo fibroblasts spread on the substrate, we present evidence for the preferential localization of a minor pool of tyrosine-phosphorylated paxillin in focal adhesions. Ventral plasma membranes showed an enrichment in beta 1-integrins, and in several tyrosine-phosphorylated polypeptides, while focal adhesion proteins like vinculin and paxillin, although localized to focal adhesions in ventral plasma membranes, were not particularly enriched in these preparations compared to whole cell lysates. Biochemical and morphological analysis of ventral plasma membranes showed a dramatic increase in the level of tyrosine-phosphorylation of the pool of paxillin localized to the adhesive sites, when compared to the paxillin present in whole cell lysates. The observed preferential localization of tyrosine-phosphorylated paxillin to focal adhesions may represent a general mechanism to compartmentalize focal adhesion components from large non-phosphorylated, cytosolic pools.

Integrin-mediated tyrosine phosphorylation and redistribution of paxillin during neuronal adhesion.

Integrins are important receptors for neuronal adhesion to laminin, which is one of the best promoters of neurite outgrowth. The present study was carried out to understand some of the intracellular mechanisms which allow integrin-mediated neurite extension on laminin. In chicken retinal neurons, integrin-mediated adhesion to laminin and antibody-induced integrin clustering caused an increase in tyrosine phosphorylation of paxillin and focal adhesion kinase. The kinetics of phosphorylation and dephosphorylation of these proteins were different in neurons plated on laminin, compared to neurons in which the receptors were clustered with anti-integrin antibodies. Analysis of sucrose velocity gradients could not show any association of paxillin and focal adhesion kinase with the integrin receptors. On the other hand, by using digitonin and milder extraction conditions, we found an enrichment of the tyrosine-phosphorylated polypeptides in the cytoskeletal, digitonin-insoluble fraction. Furthermore, neuronal adhesion induced a dramatic increase in the fraction of tyrosine-phosphorylated paxillin recovered with the digitonin-insoluble fraction, suggesting redistribution of this protein following adhesion of neurons to laminin. Localization studies on the detergent-insoluble fraction showed codistribution of both paxillin and focal adhesion kinase with integrins. We also found that paxillin tyrosine phosphorylation, but not paxillin expression, is developmentally regulated in the retina. Our results show that integrin-mediated neuronal adhesion leads to the accumulation of a pool of highly phosphorylated proteins at adhesion sites. There they may be responsible for the reorganization of the cytoskeleton, which underlies the process of neurite extension.

Tyrosine phosphorylation induced by integrin-mediated adhesion of retinal neurons to laminin.

Integrin alpha 6 beta 1 is a laminin receptor involved in adhesion and neurite extension of retinal neurons on laminin. The present study was carried out to understand some of the intracellular mechanisms which allow integrin-mediated neurite extension on laminin in primary neuronal cultures. Both integrin-mediated adhesion to laminin and antibody-induced integrin clustering resulted in the increased tyrosine phosphorylation of a 120 kDa polypeptide which was identified as the focal adhesion kinase. The kinetics of phosphorylation and dephosphorylation of this kinase were dramatically different in neurons plated on laminin, than in neurons in which the receptors were clustered with anti-integrin antibodies. To look at possible interactions of the focal adhesion kinase with integrins, we made use of sucrose velocity gradients, which have allowed the identification of a large complex containing the alpha 6 beta 1 laminin receptor. Analysis of the gradients showed that the focal adhesion kinase was not associated with the integrin receptors under these experimental conditions, while about 26% of the c-Src kinase codistributed with the integrin receptor complex, and showed a molecular size and a distribution similar to that of a 59 kDa phosphoprotein comigrating with the alpha 6 beta 1 receptor. Our results suggest that integrin-induced tyrosine phosphorylation is an early intracellular event during neuronal adhesion, and that the integrin-mediated increase in tyrosine phosphorylation of the focal adhesion kinase is not sufficient per se for the induction of neurite outgrowth. Furthermore, our data indicate that Src kinase may be involved in integrin-mediated neuronal interactions with laminin.

Cytoplasmic topography of focal contacts.

To investigate the structure of focal contacts, the cytoplasmic faces of fibroblast membranes were examined in solution by scanning force and immunofluorescence microscopy. Focal contacts were identified in scanning force topographs by correlation with fluorescence images. Finer details were resolved in topographs of the focal contacts than in fluorescence micrographs. Increased separation of ventral plasma membranes from the substrate correlated with the duration of cell culture. The cytoplasmic projections of the focal contacts also increased with the cell culture period. These changes accompanied lateral spreading of fibroblasts during a period of several hours after seeding cells in culture medium.

Differential distribution of two cytoplasmic variants of the alpha 6 beta 1 integrin laminin receptor in the ventral plasma membrane of embryonic fibroblasts.

The integrin alpha 6 beta 1 is a receptor involved in the adhesion of several cell types to laminin. By using function-blocking antibodies, we have shown that alpha 6 beta 1 is a functional laminin receptor in chick embryo fibroblasts. We also found that these cells express two variants of the alpha 6 subunit, alpha 6A and alpha 6B, characterized by different cytoplasmic domains. By using indirect immunofluorescence with isoform-specific polyclonal antibodies, we showed that the two isoforms of the alpha 6 subunit distribute differently on the ventral plasma membrane of these cells cultured on laminin-coated substrates. In fact, while the alpha 6A subunit was found codistributing with vinculin in focal contacts, the alpha 6B subunit showed a homogeneously distributed punctate pattern. This difference was particularly evident when preparations of ventral plasma membranes were used for the immunolocalization. Furthermore, when cells were cultured on fibronectin, a substrate not recognized by the alpha 6 beta 1 laminin receptor, the distribution of the two alpha 6 isoforms was similar to that observed on laminin, with alpha 6A still colocalizing with vinculin in focal adhesions. Our results indicate that two forms of the alpha 6 beta 1 laminin receptor coexpressed in the same cells show distinctive distributions, and suggest that receptor occupancy by laminin is not essential for the accumulation of the alpha 6A beta 1 integrin in adhesion plaques.

Expression and in vitro function of beta 1-integrin laminin receptors in the developing avian ciliary ganglion.

In chick development, ciliary ganglion (CG) neurons go through a period of axon extension from approximately embryonic day (E)4 to E8, followed by a period of synaptogenesis and neuronal cell death. By examining the immunohistochemical localization of laminin, in conjunction with Dil labeling of the ciliary nerve projection, we have determined that the pathway taken by these neurons is rich in laminin expression. Therefore, laminins are good candidate molecules for mediating outgrowth of these neurons in vivo. In vitro, the ability of CG neurons to extend neurites on laminin-1 (EHS laminin, alpha 1 beta 1 gamma 1) is maximal up to E8, then declines dramatically. CG neuron outgrowth on laminin-1 requires the activity of beta 1-class integrins. We have used subunit-specific antibodies to determine which of the five beta 1-containing heterodimers known to be laminin receptors (alpha 1 beta 1, alpha 2 beta 1, alpha 6 beta 1, alpha 7 beta 1) are expressed, and which mediate neurite outgrowth. While we could not detect expression of alpha 2 or alpha 7, we have found that alpha 1, alpha 3 beta 1, and alpha 6 beta 1 are expressed on the surface of ciliary ganglion neuron cell bodies and axons, both in vitro and in vivo. Furthermore, antibodies against alpha 3 and alpha 6, but not alpha 1, interfered with CG neurite outgrowth on laminin-1 in vitro. Taken together, these data suggest that interactions of cell surface alpha 3 beta 1 and alpha 6 beta 1 integrins with laminin-1 are likely to mediate growth of CG neurons during pathfinding in vivo.

Integrin receptor alpha 6 beta 1 is localized at specific sites of cell-to-cell contact in rat seminiferous epithelium.

With the aim of investigating the presence and role of integrin receptors in cell-to-cell interactions during spermatogenesis, we have immunolocalized alpha 6A integrin chain in the rat seminiferous epithelium. In both prepubertal and adult seminiferous epithelium, the antigen was found to be restricted to definite sites of intercellular contact, following a stage-specific distribution almost invariably identical to that known for beta 1 chain. In the adult seminiferous epithelium, positivity for both antigens was found exclusively around the profiles of elongating and maturing spermatids and, in most but not all stages, at a characteristic suprabasal position. In the prepubertal rat, the antigen is localized along a very regular suprabasal line of intercellular contacts. In immunoprecipitation experiments on both seminiferous epithelium explants and Sertoli cell cultures from 3-wk-old rats, anti-alpha 6A antibody coprecipitates a polypeptide of 118 kDa, presumably corresponding to beta 1 chain. These data strongly suggest that the integrin heterodimer alpha 6A beta 1 is expressed at sites of intercellular contact in the rat seminiferous epithelium. The stage-specific and restricted pattern observed by immunofluorescence suggests that this integrin is involved in regulatory interactions between Sertoli cells and germ cells during spermatogenesis.

Identification of a large complex containing the integrin alpha 6 beta 1 laminin receptor in neural retinal cells.

Integrin alpha 6 beta 1 is a laminin receptor involved in adhesion and neurite extension of retinal neurons on laminin. The present study was carried out to identify potential interactions between the alpha 6 beta 1 receptor and cellular proteins that may be involved in integrin signaling and function. For this purpose we have used a biochemical approach based on the solubilization of retinal neurons cultured on laminin with nonionic detergents, followed by centrifugation on sucrose velocity gradients. Analysis of the distribution of the alpha 6 and beta 1 integrin subunits in the gradients showed that they migrate as a large complex after extraction of cells with octylglucoside, but not after Triton X-100 extraction. Cytoskeletal proteins known to localize in adhesion plaques did not comigrate with alpha 6 beta 1 in octylglucoside gradients, while a set of polypeptides whose tyrosine phosphorylation was enhanced by culture on laminin colocalized with alpha 6 beta 1 on the gradients after octylglucoside solubilization. Culture of retinal neurons on bovine serum albumin, a nonadhesive substratum, partially affected the gradient distribution of the receptor after octylglucoside extraction. Furthermore, analysis of the gradient distribution of two alternatively spliced isoforms of the alpha 6 subunit, alpha 6-cytoA and alpha 6-cytoB, showed that the effect of non-adhesion on the sedimentation properties of the two integrin alpha 6 isoforms was more dramatic for alpha 6-cytoB than alpha 6-cytoA. These differences in the sedimentation behaviour indicate distinct biochemical properties of the two alpha 6 isoforms that, together with previous observations on their differential distribution in the developing retina, may reflect functional specificities.

Function and spatial distribution in developing chick retina of the laminin receptor alpha 6 beta 1 and its isoforms.

We have recently shown that the laminin-binding integrin receptor, alpha 6 beta 1, is prominently expressed in the developing chick retina, and its expression and activity are regulated during development on both retinal ganglion cells and other neural retinal cells. In the present study, we show that antibodies specific for the extracellular portion of the chick alpha 6 subunit dramatically inhibit interactions in vitro between embryonic day 6 neural retinal cells and laminin, showing that alpha 6 beta 1 functions as an important laminin receptor on developing retinal neurons. In previous work, we showed that alpha 6 mRNA levels on retinal ganglion cells decrease dramatically after E6 during the period that RGC axons innervate the optic tectum. In the present study, we show decreases in alpha 6 mRNA are not prevented by ablation of the optic tectum, indicating that tectal contact is not the major cause of this decrease. Within the embryonic retina, the alpha 6 subunit is codistributed, in part, with laminin, suggesting that it functions as a laminin receptor during retina development in vivo. Furthermore, two isoforms of the alpha 6 protein with distinct cytoplasmic domains generated by differential splicing have quite different distribution patterns in the retina, suggesting that these two isoforms may have different functions during retinal development.