The lens equator: A platform for molecular machinery that regulates the switch from cell proliferation to differentiation in the vertebrate lens

Authors

  • Toshiaki Mochizuki,

    1. Developmental Neurobiology Unit, Okinawa Institute of Science and Technology Graduate University, Onna, Okinawa, Japan
    Search for more papers by this author
  • Ichiro Masai

    Corresponding author
    1. Developmental Neurobiology Unit, Okinawa Institute of Science and Technology Graduate University, Onna, Okinawa, Japan
    • Author to whom all correspondence should be addressed. Email: masai@oist.jp

    Search for more papers by this author

Abstract

The vertebrate lens is a transparent, spheroidal tissue, located in the anterior region of the eye that focuses visual images on the retina. During development, surface ectoderm associated with the neural retina invaginates to form the lens vesicle. Cells in the posterior half of the lens vesicle differentiate into primary lens fiber cells, which form the lens fiber core, while cells in the anterior half maintain a proliferative state as a monolayer lens epithelium. After formation of the primary fiber core, lens epithelial cells start to differentiate into lens fiber cells at the interface between the lens epithelium and the primary lens fiber core, which is called the equator. Differentiating lens fiber cells elongate and cover the old lens fiber core, resulting in growth of the lens during development. Thus, lens fiber differentiation is spatially regulated and the equator functions as a platform that regulates the switch from cell proliferation to cell differentiation. Since the 1970s, the mechanism underlying lens fiber cell differentiation has been intensively studied, and several regulatory factors that regulate lens fiber cell differentiation have been identified. In this review, we focus on the lens equator, where these regulatory factors crosstalk and cooperate to regulate lens fiber differentiation. Normally, lens epithelial cells must pass through the equator to start lens fiber differentiation. However, there are reports that when the lens epithelium structure is collapsed, lens fiber cell differentiation occurs without passing the equator. We also discuss a possible mechanism that represses lens fiber cell differentiation in lens epithelium.

Overview of lens structure and development

The vertebrate lens is a transparent spheroidal ocular tissue that focuses visual images on photoreceptors in the retina. To obtain transparency, 90–95% of the lens consists of lens fiber cells, in which scattering of incident light is prevented by degrading their organelles, including nuclei, Golgi apparatus, and mitochondria, during lens fiber cell differentiation. In vertebrate animals, lens shape is diverse. For example, mammalian lens displays an oblate, spheroidal shape, whereas aquatic animals such as fish generally have a spherical lens (Nicol 1989; Kuszak & Costello 2004; Land & Nilsson 2012). These lens shapes are adapted to the environments inhabited by the organisms. Most mammals live on land and use the cornea to accommodate the focus, because the reflective index of the cornea is larger than that of air. In contrast, fish use the lens to accommodate the focus, because the reflective index of the cornea is close to that of water and the cornea plays little role in formation of the image in water (Greiling & Clark 2008).

In the early stage, the lens derives from the anterior surface ectoderm that overlies the neural retina. In mammalian lens, presumptive lens ectoderm thickens as the lens placode, when the neural retina evaginates from the forebrain and contacts the presumptive lens ectoderm. The lens placode detaches from the anterior ectoderm surface, invaginates into the pocket of the optic cup, and forms the lens vesicle (Fig. 1A). Cells in the posterior half of the lens vesicle, which is close to the neural retina, elongate to fill the cavity of the lens vesicle and differentiate into primary lens fiber cells, resulting in formation of the lens fiber core. On the other hand, cells in the anterior half of the lens vesicle, which is located underneath the cornea, maintain a proliferative state and form a lens epithelial monolayer. After formation of the primary lens fiber core, the equator becomes the boundary between lens epithelium and lens fiber core. At the equator, lens epithelial cells start to differentiate into lens fiber cells, which are called secondary fiber cells and cover the old lens fiber core. Zebrafish lens development resembles that of mammals, except that the lens placode invaginates as a cell mass without a cavity (Greiling & Clark 2009) (Fig. 1B). Thus, developmental lens differentiation is similar among vertebrates.

Figure 1.

Lens development in mammals and zebrafish. (A) Lens development in mammals. (Left) The lens placode detaches from surface ectoderm and forms the lens vesicle. (Middle) Cells in the posterior half of the lens vesicle elongate and differentiate into primary fiber cells, which form the lens fiber core. On the other hand, cells in the anterior half maintain a proliferative state and form mono-layered lens epithelium. (Right) Lens epithelial cells differentiate into lens fiber cells at the peripheral margin of the lens epithelium, which is called the equator. Newly differentiating lens fiber cells elongate to become secondary lens fibers and cover the old lens fiber core. (B) Lens development in zebrafish. In zebrafish, the lens placode invaginates as a cell mass. Primary fiber cells differentiate from the posterior side of the lens sphere and form the lens fiber core. Lens epithelium formation and secondary lens fiber differentiation proceed in the same way as in mammalian lens.

Research on mechanisms underlying lens development commenced in the early twentieth century. Hans Spemann showed that the lens failed to be specified in surface ectoderm when the presumptive retinal region was excised from the anterior neural plate in frog embryos (Spemann 1901). He concluded that an inductive signal or signals from presumptive retina promote lens differentiation, and this finding established the concept of induction in developmental biology. In the 1980s, Grainger and colleagues examined the lens induction process and proposed that it be classified into four discrete steps (Grainger 1992, 1996). In the mid-gastrula stage, all of the ectoderm in the animal cap is competent to generate the lens. However, in the second step, this competence becomes restricted to the presumptive lens ectoderm under the influence of signals from the anterior neural plate in the neurula stage. Third, the lens placode is specified when the presumptive retina is located close to the presumptive lens ectoderm. Fourth, lens cells start to differentiate into lens fiber cells (Grainger et al. 1997). Recently it has been reported that competence of lens formation in the head ectoderm is not only promoted by signals from the anterior neural plate, but is also restricted by transforming growth factor-β (TGF-β) and canonical Wnt signaling from neural crest cells (Grocott et al. 2010), suggesting that lens placode induction is cooperatively regulated by positive and negative signals.

After the formation of primary lens fiber core, lens epithelial cells start to differentiate into secondary lens fiber cells at the equator. Differentiating lens fiber cells become flat, and they extend apical and basal processes bi-directionally toward the anterior and posterior poles of the lens sphere, respectively. These apical and basal processes converge at both poles, forming sutures (Fig. 2A). Interestingly, the suture pattern is diverse among vertebrates; mammalian lens shows Y-shaped sutures, zebrafish and avian lenses show point sutures, and several fishes and amphibian lenses have linear sutures (Kuszak & Costello 2004) (Fig. 2B). Furthermore, Y-shaped sutures become more complex “Star” structures in primate lenses during development (Kuszak et al. 2004). These differences may contribute to the diversity of lens shapes, because the point and Y sutures are observed in spherical and ellipsoidal lenses, respectively. Furthermore, a correlation between suture architecture and optical quality has been reported (Kuszak et al. 1991). Mechanisms that regulate the diversity of lens suture patterns are interesting. Furthermore, lens fiber cells possess two types of protrusions: edge protrusions and ball-and-socket protrusions, which are formed along the short and long sides, respectively (Fig. 2C). Both protrusions contribute to interlocking attachments between lens fiber cells, and likely function as a platform for gap junctions (Blankenship et al. 2007; Bassnett et al. 2011). Gap junctions enhance lens fiber cell connectivity, stabilize lens morphology, and transport small molecules from lens epithelium to lens fiber cells and between lens fiber cells (Mathias et al. 2010). In addition to gap junctions, a macromolecule diffusion pathway between lens fiber cells has been suggested (Shestopalov & Bassnett 2000, 2003), which depends upon formation of a stratified syncytium. This is established by cell fusion mediated by a claudin-like protein, Lim2 (also known as MIP20) (Shi et al. 2009). Mechanisms regulating such sophisticated lens fiber morphologies and functions are also an interesting research topic.

Figure 2.

Lens fiber organization and structure. (A) The lens suture (red) is formed at the anterior and the posterior poles of the lens. (B) The suture pattern (red) is diverse: umbilical (point) sutures in birds and teleosts, including zebrafish (left), line sutures in amphibians and some fish species (middle), and Y-sutures in mammals (right). (C) Lens fiber cells possess two types of protrusions: edge protrusions and ball-and-socket processes, which are formed on the short and wide sides of lens fiber cells, respectively. Both types of protrusions interlock neighboring lens fiber cells and contain gap junctions.

Another important aspect of lens fiber differentiation is degradation of intracellular organelles, including the nucleus (Bassnett 2009). In lens fiber cell differentiation, the nucleus (Bassnett & Mataic 1997), mitochondria (Bassnett & Beebe 1992) and Golgi apparatus (Bassnett 1995) are degraded to form a transparent organelle-free zone (OFZ). Recently, it was reported that DNase II-like acid DNase (DLAD, DNase IIβ) promotes denucleation in the lens (Nishimoto et al. 2003). However, it remains to be elucidated how degradation of the nucleus and other organelles are coordinated to form the OFZ. Since DLAD knockout mice show cataracts, research on OFZ formation may contribute to our understanding of cataract pathology.

Classical studies on lens fiber differentiation in the 1970s

In the 1970s, using chick lenses, Coulombre showed that lens epithelial cells elongated and differentiated into lens fiber cells, after the lens was surgically rotated 180° within the optic cup so that the lens epithelium faced the neural retina (Coulombre & Coulombre 1963) (Fig. 3A). Similar results were obtained using mouse lenses (Yamamoto 1976). These results suggest that the whole lens epithelial region retains some competence to differentiate into lens fiber cells during later developmental stages, and that lens fiber differentiation is activated even without passing through the equator, provided that the lens epithelium faces the vitreous humor or the neural retina. This phenomenon is very interesting, because it suggests that the environment is different between the anterior and posterior sides of the lens (Fig. 3B). One possibility is that the posterior environment contains materials or houses mechanisms that direct lens fiber differentiation. Alternatively, the anterior environment may constitutively suppress lens fiber differentiation. Either possibility could explain why lens epithelial cells start to differentiate into lens fiber cells at the equator, which is the interface between anterior and posterior regions of the lens. In this context, the equator provides a platform of molecular machinery that regulates the switch from cell proliferation to differentiation in the vertebrate lens.

Figure 3.

Mechanism of differentiation from lens epithelium into lens fiber cells. (A) Classic experiments on lens inversion done by Coulombre & Coulombre (1963) and Yamamoto (1976). Lens epithelium (green) and lens fiber cells (white) are located in the anterior and posterior regions of the lens, respectively. After the lens was rotated upside down so that the lens epithelium faced the retina, inverted lens epithelial cells elongated and differentiated into new lens fiber cells. (B) The environments on the anterior and posterior sides of the lens suppress and promote lens fiber differentiation, respectively. The anterior side of the lens faces the aqueous humor or cornea, whereas the posterior side of the lens faces the vitreous humor and the retina. (C) Lens epithelium is subdivided into three regions: the anterior zone (grey), the germinative zone (red), and the transition zone (yellow).

In the 1970s, McAvoy examined proliferation in the developing lens and found that cell proliferation was observed only in lens epithelium, but not in lens fiber cells (McAvoy 1978; Zwaan & Kenyon 1984). McAvoy classified lens epithelium into three distinct regions, an “anterior zone”, a “germinative zone” and a “transition zone”, along the antero-posterior axis (McAvoy 1978) (Fig. 3C). The anterior zone indicates the most anterior region of lens epithelium, where cell proliferation is relatively low. The germinative zone is located just anterior to the equator. Cell proliferation is most active in this zone. The transition zone is located just posterior to the equator, where lens fiber cell differentiation occurs. These observations suggest that cell proliferation is not uniform throughout lens epithelium and that it is spatially regulated. BrdU (5-Bromo-2′-deoxyuridine) was highly incorporated into the region anterior to the equator in the zebrafish lens (Imai et al. 2010), suggesting that the germinative zone is common to all vertebrates. The biological significance of the germinative zone remains to be elucidated, although the high activation of cell cycle progression may play a role in the entry of lens fiber differentiation at the equator.

Intrinsic factors that regulate lens fiber cell differentiation

Genetic studies have revealed several transcription factors that regulate lens placode induction, lens vesicle formation, and lens fiber cell differentiation. We briefly summarize the role of these factors in lens development. A paired domain and homeobox domain-containing transcription factor, Pax6, is the master regulator of lens placode induction and maintenance (Hill et al. 1991; Glaser et al. 1992; Jordan et al. 1992; Matsuo et al. 1993; Halder et al. 1995; Altmann et al. 1997). Pax6 expression was detected in the head surface ectoderm and became restricted to the lens placode when the optic cup was formed (Grindley et al. 1995; Kamachi et al. 1998). Later, Pax6 expression was maintained only in lens epithelium (Kamachi et al. 1998), suggesting that Pax6 maintains an undifferentiated state of lens epithelial cells and the potency to generate lens fiber cells. Overexpression of Pax6 in lens fiber cells abnormally induced expression of a lens epithelial marker, α5/β1 integrin (Fig. 4), in lens fiber cells, which subsequently caused cataract formation (Duncan et al. 2000), suggesting that downregulation of Pax6 in lens fiber cells is important for lens fiber differentiation.

Figure 4.

Regulatory network of transcription factors in lens development. Pax6, Mab21l1, Six3, AP2α, Pitx3, Msx2 and Foxe3 promote maintenance of lens epithelial cells through expression of α5/ β1 integrin and E-cadherin, whereas Prox1, Sox1 and c-Maf promote lens fiber cell differentiation through expression of cdk inhibitors, crystallin, and Aquaporin 0. See the text for details.

Sry-related high-mobility group (HMG) box-containing transcriptional factors, Sox1, Sox2, and Sox3, are expressed in developing lens. In chicks and mice, Sox2 and Sox3 are initially expressed in the lens placode and downregulated in later stages, while Sox1 is initially expressed in the lens placode and is maintained in lens fiber cells (Collignon et al. 1996; Kamachi et al. 1998). Sox1 knockout mice displayed microphthalmia with lens cataracts (Nishiguchi et al. 1998). Sox1 mutant lens fiber cells abnormally expressed lens epithelial markers, Pax6 and α5 integrin (Donner et al. 2007), and did not express a lens fiber cell marker, γ-crystallin (Nishiguchi et al. 1998) (Fig. 4). They also failed to elongate, suggesting that Sox1 is required for differentiation of lens epithelial cells into lens fiber cells.

A vertebrate homologue of Drosophila sine oculus, Six3, is expressed in the lens placode and later restricted to lens epithelium, including the transition zone in mice and zebrafish (Oliver et al. 1995; Kobayashi et al. 1998). Pax6 knockdown reduced Six3 expression in the lens placode (Ashery-Padan et al. 2000). Overexpression of Six3 partially rescued lens defects in Pax6 haploinsufficiency, and Pax6 and Six3 mutually activate their expression (Goudreau et al. 2002) (Fig. 4). Together, these data suggest that Six3 promotes Pax6-mediated lens differentiation.

The forkhead/winged helix transcription factor, Foxe3 (Xlens1), is expressed in the lens placode and later restricted to the lens epithelium (Kenyon et al. 1999; Blixt et al. 2000; Brownell et al. 2000; Shi et al. 2006). Pax6 induces Foxe3 expression (Blixt et al. 2000; Brownell et al. 2000) (Fig. 4). In the Foxe3 mutant mouse, which was originally identified as Dysgenetic lens (dyl), cell proliferation was decreased in lens epithelium. Furthermore, expression of a marker of lens fiber differentiation, Prox1, was expanded anteriorly in the Foxe3 mutant (Fig. 4), although expression of a lens epithelial adhesion molecule, E-cadherin (cdh1), was not affected (Blixt et al. 2000), suggesting that Foxe3 maintains the lens epithelium and represses lens fiber differentiation, downstream of Pax6.

Mab21l1, a vertebrate homolog of C. elegans mab-21 (Baird et al. 1991), is initially expressed in the lens placode and later restricted to lens epithelium (Yamada et al. 2003). In Mab21l1 knockout mice, the lens placode did not invaginate completely, resulting in lens vesicle formation defects. Expression of a lens epithelium marker, Foxe3, was decreased in Mab21l1 knockout mice (Fig. 4), although Pax6 expression was not changed, suggesting that Mab21l1 functions downstream of Pax6 and upstream of Foxe3 to form the lens vesicle.

The AP-2 transcription factor, AP-2α, is initially expressed in the lens placode and later restricted to lens epithelium (Ohtaka-Maruyama et al. 1998). Knockdown of AP-2α caused a “lens stalk” phenotype, in which separation of the lens vesicle from the surface ectoderm is incomplete (West-Mays et al. 1999), suggesting roles of AP-2α in lens morphogenesis. In AP-2α mutant lens epithelium, expression of Pax6, Foxe3, and Pitx3 was not affected, but E-cadherin expression was reduced (Pontoriero et al. 2008). Ectopic expression of AP-2α in lens fiber cells abnormally induced E-cadherin expression in lens fiber cells and reduced a lens fiber cell marker Aquaporin 0 (also known as MIP26) (West-Mays et al. 2002) (Fig. 4). These data suggest that AP-2α regulates lens morphogenesis, probably through E-cadherin functions.

Msx2 is expressed in the lens placode and later in the entire lens. Msx2 null mice showed microphthalmia, accompanied by down-regulation of the lens epithelial marker, Foxe3, and up-regulation of lens fiber cell markers, Prox1 and crystallin (Fig. 4), although expression of Pax6 and AP-2α was not affected (Zhao et al. 2012). These data suggest that Msx2 is required for proper expression of transcription factors involved in lens development.

Bicoid-type homeobox transcription factor, Pitx3, is expressed in the lens vesicle and later restricted to lens epithelium (Shi et al. 2005; Ho et al. 2009). Pitx3 mutant mice known as Aphakia, developed cataracts, a lens stalk phenotype, and defects in primary lens fiber differentiation (Varnum & Stevens 1968; Zwaan 1975; Grimm et al. 1998). In Pitx3 knockout mice, lens epithelial cell markers, platelet-derived growth factor (PDGF) receptor α and Foxe3, were decreased (Fig. 4), but E-cadherin expression was not affected (Ho et al. 2009). In contrast, a differentiation marker for lens fiber cells, Prox1, and two cdk inhibitors p27Kip1 and p57Kip2, were expressed throughout the entire lens (Ho et al. 2009) (Fig. 4), suggesting that Pitx3 is required for maintenance of lens epithelial integrity.

Prox1, a vertebrate homologue of Drosophila prospero, is initially expressed in the lens vesicle and later restricted to cells passing over the equator and early differentiating lens fiber cells. Subcellular localization of Prox1 protein is unique. It occurs in cytoplasm in lens epithelium and in nuclei in differentiating lens fiber cells (Duncan et al. 2002). Prox1 null mice showed defects in elongation of primary lens fiber cells, and reduced expression of cdk inhibitors, p27Kip1 and p57Kip2 (Wigle et al. 1999) (Fig. 4). These data suggest that Prox1 is required for the inception of lens fiber cell differentiation.

bZIP transcription factor, MafA (also known as L-maf), is initially expressed in the lens placode, and later restricted to lens fiber cells in mouse and frog (Ogino & Yasuda 1998; Ishibashi & Yasuda 2001; Takeuchi et al. 2009). On the other hand, MafB is initially expressed in the lens placode and then restricted to lens epithelium (Takeuchi et al. 2009). Ectopic introduction of MafA and MafB was shown to induce ectopic lens fiber marker crystallins in chick and Xenopus embryos (Ogino & Yasuda 1998; Ishibashi & Yasuda 2001). However, neither MafA nor MafB knockout mice show defects in lens development (Takeuchi et al. 2009). Furthermore, MafA and MafB double knockout mice also show no defect in the lens (Takeuchi et al. 2009). Another Maf protein, c-Maf, is expressed in the lens placode and is restricted to lens fiber cells (Takeuchi et al. 2009). c-Maf mRNA expression is much more abundant than that of MafA and MafB, and c-Maf knockout mice do manifest impaired lens fiber differentiation and crystallin gene expression (Fig. 4), although expression of Pax6, Sox, and Prox1 was not affected (Kawauchi et al. 1999; Kim et al. 1999; Ring et al. 2000). These results suggest that c-Maf regulates terminal differentiation of lens fiber cells, probably downstream of Prox1.

Although the relationships of these transcription factors remain to be fully elucidated, their combinatory network plays a role in establishment and maintenance of lens epithelium and lens fiber cell area (Fig. 4). Next, we will discuss cell-extrinsic factors from a spatial regulation perspective.

Extrinsic factors that regulate lens fiber cell differentiation

As mentioned above, Coulombre (Coulombre & Coulombre 1963) and Yamamoto (Yamamoto 1976) showed that the lens epithelium initiated lens fiber differentiation when the lens was rotated 180° so that the lens epithelium faced the vitreous humor or the neural retina. This finding indicates that there is some material that promotes lens fiber differentiation on the posterior side of the lens. In the 1980s, using chick lens epithelial cell culture, Beebe and colleagues found that lens epithelial cells underwent lens fiber differentiation when they were cultured with vitreous humor (Beebe et al. 1980). He suggested that vitreous humor contains a substance, called “lentropin” that induces lens fiber cell differentiation, which was later identified as insulin-like growth factor-1 (IGF-1) (Beebe et al. 1987). Piatigorsky and colleagues independently found that insulin induced lens fiber cell differentiation in chick lens epithelial cell culture (Piatigorsky 1973; Piatigorsky et al. 1973). These observations suggest the existence of cell extrinsic factors that promote lens fiber cell differentiation.

Co-culture of rat lens explants with retina or a medium used for retinal explant culture, promoted lens fiber cell differentiation, suggesting that secretion molecules emanating from the retina are responsible for lens fiber differentiation (McAvoy 1980; McAvoy & Fernon 1984). The same group identified fibroblast growth factor (FGF) as a candidate (Chamberlain & McAvoy 1987, 1989). Western blot analyses revealed that FGF1 and FGF2 were present in bovine lens fiber cells and vitreous humor (Schulz et al. 1993) (Fig. 5A). FGF3, initially called Int2, is expressed in the early developing retina (Wilkinson et al. 1989). FGF8 is expressed in the early developing retina and later restricted to the inner nuclear layer of chick retinas (Vogel-Hopker et al. 2000). FGF9 is expressed in the mouse retina and pigmented epithelium (Colvin et al. 1996). FGF12 is expressed in retinal ganglion cells of the mouse retina (Hartung et al. 1997). FGF5 is expressed in photoreceptors, retinal ganglion cells, and retinal-pigmented epithelium of adult macaque retinas (Kitaoka et al. 1994). FGF receptors are also expressed in the lens. FGFR1 and a splice variant of FGFR2, FGFR2 IIIB, are expressed in lens epithelium in mice (Fig. 5A). Their expression formed a low to high gradient along the anterior–posterior axis, and the highest expression was observed near the equator of the rodent lens (Orr-Urtreger et al. 1993; de Iongh et al. 1996, 1997). Another splice variant of FGFR2, FGFR2 IIIc, is expressed in mouse lens epithelium (Orr-Urtreger et al. 1993; de Iongh et al. 1997) (Fig. 5B). FGFR3 is expressed in lens fiber cells (Peters et al. 1993) (Fig. 5B). FGFR4 is expressed in the germinative and transition zones of the chick lens (Marcelle et al. 1994) (Fig. 5B). These data suggest that lens epithelium expresses all FGF receptors, which may bind multiple FGF ligands expressed in ocular tissues.

Figure 5.

Extrinsic factors in lens fiber differentiation. (A–F) Expression patterns of FGF1/2 (A), FGFR1/2/3/4 (A, B), BMP4/7 (C), BMPRIa/II (C), TGF-β2 (D), TβRI/II (D), Wnt8a (E), Fz4 (E), Jagged1 (F), and Notch2 (F) in ocular tissues. (G) The interactions of extrinsic factors. See text for details.

Application of FGF to lens epithelium induced cell proliferation at low doses and lens fiber cell differentiation at high doses, suggesting that FGF regulates multiple steps of lens fiber differentiation in a dose-dependent manner (Chamberlain & McAvoy 1989). McAvoy suggested the possibility that an FGF dose gradient is formed from aqueous to vitreous humor and regulates step-wise lens fiber differentiation. However, FGFR1 mutant cells normally formed the lens even when they were transplanted into Pitx3 mutant mice, in which lens development is severely retarded (Zhao et al. 2006), suggesting that FGFR1 is dispensable for lens development. Knockdown of FGFR2 slightly compromised withdrawal of the cell-cycle exit of lens epithelial cells and increased cell death in both lens epithelial and fiber cells, resulting in small-sized lenses (Garcia et al. 2005). Knockdown of FGFR3 also did not affect lens development in mice (Deng et al. 1996). Interestingly, a lens-specific triple knockdown of FGFR1, 2 and 3 in mice severely inhibited primary lens fiber differentiation in the posterior half of the lens vesicle, resulting in microphthalmia (Zhao et al. 2008). These data suggest that FGF is necessary for lens fiber cell differentiation. Equarin is a novel secretion molecule, which binds FGFs in vitro, and facilitates FGF-dependent lens fiber differentiation (Song et al. 2010). Since equarin is highly expressed at the lens equator, it is likely that equarin activates FGF signaling at the equator. These data suggest that FGF signaling mediates the switch from epithelial cell proliferation to lens fiber cell differentiation (Fig. 5G).

Bone morphogenetic protein (BMP) is another candidate that regulates lens fiber cell differentiation. BMP4 and BMP7 are expressed in the ciliary epithelium, which is located adjacent to the germinative zone of the mouse lens (Zhao et al. 2002) (Fig. 5C). BMP type I receptor, BMPRIa (ALK3) and type II receptor, BMPRII, are expressed in lens epithelium, but both were decreased in lens fiber cells (de Iongh et al. 2004) (Fig. 5C). Application of a BMP antagonist, Noggin, and ectopic expression of dominant-negative BMPRIb (ALK6) inhibited lens fiber cell differentiation (Faber et al. 2002), suggesting that BMP signaling positively regulates lens fiber differentiation. In chick, BMP promoted FGF-dependent lens fiber cell differentiation in primary lens culture (Boswell et al. 2008), and induced expression of the FGF-positive regulator, equarin (Jarrin et al. 2012). These observations suggest that BMP cooperates with FGF to promote lens fiber cell differentiation (Fig. 5G).

In mice, all TGF-β family proteins, TGF-β 1, 2 and 3, are expressed in lens fiber cells, but not in lens epithelium (Pelton et al. 1991). However, TGF-β 2 mRNA was expressed in mouse lens epithelium and the ciliary body (Millan et al. 1991). TGF-β 2 was detected in human aqueous and vitreous humors (Connor et al. 1989; Jampel et al. 1990) (Fig. 5D). TGF-β type I receptor, ALK5, and type II receptor, TβRII, are expressed in lens epithelium and early differentiating fiber cells (de Iongh et al. 2001a) (Fig. 5D). These observations suggest roles for TGF-β signaling in lens fiber differentiation. Application of TGF-β to rat lens epithelial explants promoted FGF-mediated lens fiber differentiation, although TGF-β alone has little effect (Liu et al. 1994). Interestingly, application of TGF-β to whole rat lenses generated opaque plaques in lens epithelium, which consisted of abnormally aggregated cells expressing a mesenchyme marker, α-smooth muscle actin (α-SMA), extracellular matrix protein, laminin, and Type I collagen (Hales et al. 1995). This suggested that TGF-β transforms lens epithelial cells into myofibroblastic cells. Transgenic mice overexpressing TGF-β 1 in lens fiber cells also manifested similar opaque plaques, in which expression of lens epithelial markers such as Pax6 was reduced, but expression of α-SMA and lens fiber cell markers, α, β, γ-crystallins and Aquaporin 0 (MIP26), was induced, suggesting that TGF-β promotes lens epithelial cell differentiation into lens fiber cells (Lovicu et al. 2004). Overexpression of dominant negative TGF-β receptors inhibited lens fiber differentiation (de Iongh et al. 2001b). These data suggest that TGF-β promotes lens fiber differentiation, probably coupled with FGF (Fig. 5G).

Wnt signaling comprises three pathways, known as canonical Wnt, non-canonical Wnt/Planer cell polarity (PCP), and Wnt/Ca2+ signaling (Niehrs 2012). The canonical Wnt signaling pathway regulates cell proliferation, whereas the non-canonical Wnt/PCP pathway spatially coordinates cell migration and cell morphogenesis, resulting in tissue pattern formation. Wnt signaling molecules: ligands Wnt5a, 5b, 7a, 7b, 8a and 8b, receptors Frizzled 1-4, and mediators Dishevelled (Dev) 2 and 3, are expressed in lens epithelium, including the transition zone (Stump et al. 2003) (Fig. 5E). Knockdown of a co-receptor for canonical Wnt signaling, Lrp6, disrupted lens epithelial structure, resulting in extrusion of lens fibers into the space between the lens epithelium and the cornea (Stump et al. 2003). Knockdown of a canonical Wnt effector, β-catenin, reduced the expression of lens epithelial markers, Pax6 and E-cadherin, and impeded cell-cycle progression in lens epithelium, suggesting that Wnt signaling is required for lens epithelium formation and function. Furthermore, lens fiber differentiation was also abnormal in regard to apico-basal cell polarity and cadherin-actin co-localization (Cain et al. 2008). Overexpression of constitutive active β-catenin or knockdown of a Wnt negative regulator, adenomatous polyposis coli (APC) displayed opposite phenotypes: increased expression of lens epithelial markers, Pax6 and E-cadherin, and reduced expression of lens fiber cell markers, c-Maf and β-crystallin, in the posterior region of the lens, resulting in expansion of lens epithelium over the equator (Martinez et al. 2009). These in vivo data suggest that canonical Wnt signaling is required for maintenance of lens epithelium (Fig. 5G). However, an in vitro study using lens epithelial explants revealed that Wnt3a alone promotes cell-cycle progression; however, when explants were pre-treated with FGF, Wnt3a also promoted lens fiber differentiation (Lyu & Joo 2004). These data suggest that Wnt regulates lens epithelial proliferation and lens fiber differentiation in a context-dependent manner. It was reported that lens fiber elongation and morphogenesis requires Wnt/PCP pathway (Chen et al. 2008), which cooperate with FGF signaling (Dawes et al. 2013, 2014).

Notch signaling regulates the balance between proliferation and differentiation through a lateral inhibition mechanism (Bray 2006; Louvi & Artavanis-Tsakonas 2006). Notch2 is expressed in lens epithelium (Fig. 5F), and knockdown of Notch2 increased the fraction of differentiated lens fiber cells (Saravanamuthu et al. 2012). Knockdown of a Notch effector, CSL (CBF1/Su(H)/LAG-1), which was originally called Rbp-j in mammals, promoted a precocious lens fiber differentiation, whereas constitutive activation of Notch promotes cell proliferation in lens epithelium (Jia et al. 2007; Rowan et al. 2008). A Notch ligand, Jagged1, is expressed in early differentiating lens fiber cells, including the transition zone (Fig. 5F). Knockdown of Jagged1 in heterozygous mice, enhanced expression of a cdk inhibitor, p57Kip2, in the germinative zone, promoting precocious lens fiber cell differentiation (Le et al. 2009). These data suggest that the Jagged1-Notch2 signaling pathway regulates a proper balance between cell proliferation and cell differentiation around the lens equator (Fig. 5G). Although the relationship between Jagged1-Notch2 signaling and other factors such as FGF and Wnt remains to be elucidated, it is interesting that Wnt and Notch signaling pathways sequentially coordinate proliferation and differentiation in the ciliary marginal zone of the retina (Yamaguchi et al. 2005), which is spatially associated with the lens equator (Imai et al. 2010). Figure 5G summarizes the molecular network of extrinsic factors.

Suppression of lens fiber differentiation in the lens epithelium: Implication from research on cataract surgery

Classical experiments done by Coulombre and Yamamoto (Coulombre & Coulombre 1963; Yamamoto 1976) suggest the possibility that the posterior environment contains materials that direct lens fiber differentiation. Alternatively, the anterior environment may contain materials that constitutively suppress lens fiber differentiation. In the latter model, lens epithelial cells are prevented from entering lens fiber differentiation and the equator may release lens epithelial cells from this inhibition. It was reported that ectopic lens fiber differentiation occurred when lens epithelial structure was disrupted. Secondary cataracts, also known as posterior capsular opacifications (PCOs), illustrate this process (Wormstone 2002). PCOs are the most frequent complication of cataract surgery. Cataract surgery usually excises lens cells from the lens capsule and implants an intraocular lens for restoration of vision. It appears that cataract surgery induces a wound-healing response in the lens. Lens epithelial cells remaining in the lens capsule proliferate, migrate as in epithelial-mesenchymal transition (EMT) to become fibroblastic cells, and in some cases, undergo lens fiber regeneration, resulting in formation of Elschnig's pearl and Sommerring's ring (Saika 2004). These processes are the major contributors to PCO (Fig. 6).

Figure 6.

Posterior capsular opacification. Cataract surgery excises lens cells from the capsule and implants an intraocular lens (IOL). Lens epithelial cells left behind after surgery (grey) differentiate into myofibroblastic cells through a transforming growth factor TGF-β2/Smad3-dependent, epithelial-mesenchymal transition (EMT). These cells produce extracellular matrix such as collagens and fibronectin on the IOL. Residual lens epithelial cells also differentiate into lenticular fiber cells, expressing crystallins, resulting in formation of Sommerring's ring around the equator and Elschnig's pearl on the posterior capsule. These fibroblastic and lenticular fiber cells cause opacification.

Transforming growth factor-β plays a central role in fibrous type PCO. TGF-β 2 was detected in the aqueous and vitreous humors (Connor et al. 1989; Jampel et al. 1990) (Fig. 5D). Application of TGF-β to lens epithelium collapsed epithelial structure, induced expression of a mesenchymal marker, α-SMA, and EMT-like cell behavior, characteristic of fibrous PCO (Liu et al. 1994; Hales et al. 1995, 1997). It was reported that nuclear translocation of Smad3 and Smad4 occurs in lens epithelial cells during wound repair after injury (Saika et al. 2001, 2002), and that Smad3 mediates TGF-β-mediated, EMT-like phenotypes in lens epithelium after injury (Saika et al. 2004b). Furthermore, application of Smad2/3 antagonists, such as Smad7, BMP7, Id2, and Id3, suppressed injury-induced EMT in lens epithelium (Saika et al. 2004a, 2006). These data suggest that injury of lens epithelium or the anterior capsule causes fibrous PCO-like defects through activation of TGF-β. In the transition zone, TGF-β activates an E3-ubiqutin ligase, anaphase promoting complex (APC/C), which in turn degrades an inhibitory factor of lens fiber differentiation, SnoN, through activation of the ubiquitin proteasome system (Wu et al. 2007). Indeed, a proteasome inhibitor, lactacystin, inhibited TGF-β-dependent EMT in the lens (Hosler et al. 2006). We also reported that lens fiber cell differentiation was inhibited in zebrafish psmd6 mutants, in which ubiquitin-proteasome activity is reduced. Inhibition of APC/C also showed similar defects in lens fiber differentiation in zebrafish (Imai et al. 2010). These data suggest that spatial regulation of TGF-β signaling is a component of the molecular switch from proliferation to differentiation on the lens equator. SnoN expression in lens epithelium (Wu et al. 2007) suggests some mechanism that constitutively suppresses TGF-β signaling in lens epithelium.

Platelet-derived growth factor receptor α (PDGFα) is expressed in lens epithelium. Its ligand, PDGF-A, is expressed in the iris, the ciliary body and the cornea, whereas PDGF-B is expressed in the iris, the ciliary body, and vascular cells surrounding the lens (Reneker & Overbeek 1996). Ectopic expression of PDGF-A activated proliferation of lens epithelial cells through upregulation of expression of cell-cycle regulators, cyclin A and cyclin D2, resulting in multi-layer formation, reminiscent of lenticular defects in PCO. Interestingly, a marker of lens fiber cells, β-crystallin, was also expressed in multiple lens epithelial layers (Reneker & Overbeek 1996). This observation supports the possibility that disruption of lens epithelium causes ectopic lens fiber cell differentiation.

Integrin family proteins mediate interaction between cells and the extracellular matrix (ECM). They are classified into α and β subtypes, and function as heterodimers with α and β subtypes. The combination of α and β subtypes determines the specificity of interaction between the integrin complex and the ECM. Several integrin proteins are expressed in the lens (Walker & Menko 2009). β1 integrin is expressed in the lens, and conditional knockdown of β1 integrin in mice disorganized lens epithelium, with upregulation of α-SMA expression and induced ectopic expression of fiber cell markers, Prox1, c-Maf, β-crystallin (Simirskii et al. 2007). β1 integrin is required for structural integrity of lens epithelium, which likely prevents lens fiber cell differentiation. α5/ β1 integrin complexes with fibronectin. TGF-β2 caused α5 integrin redistribution from focal adhesions to a diffuse pattern of localization (Marcantonio & Reddan 2004), suggesting a link between TGF-β2-mediated PCO and dysfunction of α5/ β1 integrin. α3/ β1 and α6/ β1 integrin complexes interact with laminin. Neither single knockdown of α3 nor α6 integrin caused defects in lens development, but the double knockdown of both α3 and α6 integrin disrupted lens epithelium, suggesting that laminin-dependent cell adhesion is necessary for lens epithelium formation (De Arcangelis et al. 1999; Wederell & de Iongh 2006).

A cell adhesion molecule, E-cadherin (cdh1) is expressed exclusively in lens epithelium, while N-cadherin (cdh2) is expressed throughout the entire lens (Takeichi 1988; Xu et al. 2002; Pontoriero et al. 2009). N-cadherin-mediated adhesion is functional only in lens fiber cells (Leonard et al. 2011). Knockdown of E-cadherin caused disorganization of lens epithelium with up-regulation of α-SMA expression (Pontoriero et al. 2009). These data suggest that downregulation of E-cadherin causes EMT-like defects. All cases of E-cadherin, PDGF, and integrin indicate the importance of lens epithelial structural integrity in suppression of EMT and lens fiber differentiation. In the future, it is important to elucidate whether disruption of lens epithelium induces lens fiber cell differentiation through activation of TGF-β signaling.

Future directions of the lens research

Here we have reviewed research on intrinsic and extrinsic factors that provide insights into mechanisms that govern cell proliferation and cell differentiation at the equator. Although the molecular mechanisms remain to be elucidated, the next important direction is how the growth rate of the lens is regulated. The growth rate of the lens depends on how many lens epithelial cells pass over the equator and enter lens fiber differentiation in a given period. Since the lens focuses visual images on the retinal photoreceptor layer, the size of the lens must match that of the eye cup. It is reported that in spherical lenses of fish eyes, the focal length of the lens (lens center to retinal surface) is 2.55 times the lens radius, “Matthiessen's ratio” (Nicol 1989). Thus, there must be some mechanism that matches the growth rate of the lens and the retina in fish. It was reported that glucagon-expressing retinal neurons, called Bullwhip cells, are sparsely located in the inner layer of the neural retina and extend very thin processes to the retinal CMZ. Retinal electric activity linked to focus accommodation modulates the level of glucagon secreted from Bullwhip cells, subsequently suppressing proliferation of retinal CMZ cells (Fischer et al. 2005, 2008). However, it is still not clear whether growth of lens epithelial cells is modulated in response to defocus or hyper focus in the retina. Whatever it may be, what mechanism enables it? Understanding of molecular machinery at the lens equator will enable us to investigate the question more fully.

As mentioned above, lens morphology is diverse. The difference of the suture pattern may correlate with how ellipsoidal the lens is. However, it is unknown how the diversity of lens shapes is achieved. The four-eyed fish, Anableps, is an active fish that swims just below the water surface. The eye is divided by the water meniscus; the upper half is in the air and the lower half in the water. Two horizontal iris flaps cross the eye at the water level, forming two separate pupillary apertures. The dorsal retina receives images from the lower aquatic field and the ventral retina from the aerial field. From an evolutionary point of view, the four-eyed fish is interesting, because it has an oval lens with a sharper curvature in the water visual axis and a flatter curvature in the air visual axis (Nicol 1989; Land 2012). It is interesting to examine how differentiation rate of lens fiber cells is modified to achieve this oval shape. Once we understand the mechanism that coordinates lens epithelial cell proliferation and lens fiber differentiation in model animals such as zebrafish and mice, we will be able to answer this interesting question. Research on the switch machinery from cell proliferation to cell differentiation at the lens equator will enable us to study newly uncovered issues in the field of evolution, cell biology, and developmental biology.

Acknowledgement

We thank Dr Harukazu Nakamura and Dr Takashi Takeuchi for inviting us to write this review and Mr Steven D. Aird for critical reading of the manuscript.

Ancillary