These studies suggest Wnt/β-catenin signaling might not be active and required for differentiation of postnatal photoreceptors.

Taurine

Another positive regulator of rod differentiation is taurine, a cystein derivative that is structurally similar to γ-aminobutyric acid (GABA) and glycine. It is highly expressed in the developing and mature retina. When omitted in the diet, developmental defects of brain and retina occur in kittens (for review, see [184]). Moreover, Altshuler et al. showed that taurine promotes rod differentiation in rat retinal cultures [185].

Interestingly, taurine can activate glycine and GABA receptors, and the action of taurine is dependent on the activity of these receptors [184]. In mouse retinal cultures, application of glycine or GABA receptor antagonists abolished the effects of taurine on rod differentiation. Importantly, inhibition experiments using small interfering RNA (siRNA) against glycine receptor a 2 (GlyRa 2) revealed that fewer photoreceptors develop in vivo, which is compensated by a higher number of late-born cell types such as Müller glia and bipolar cells [184]. These and other experiments suggest that taurine/ glycine receptor signaling negatively regulates RPC proliferation and stimulates rod differentiation. This effect appears to be specific for the photoreceptor lineage since forced expression for GlyRα2 can induce photoreceptor-specific genes in E16 retinal progenitors. Conversely, without taurine/glycine receptor activation, progenitor cells might keep dividing and differentiate eventually into late-born cell types due changes in the extracellular environment.

In agreement with the role of taurine and GlyRα2 during central nervous system development, disruption of the GlyRa 2 gene results in defects of electrophysiological and calcium imaging responses of cortical cells [186]. However, no abnormalities in development and function of retinae of GlyRa 2−/− could be observed; Nrl and rhodopsin expression and ERG analysis are normal without GlyRα2 [186]. Since expression of the other glycine receptor subunits is not changed, these results suggest that GABA receptor signaling may compensate or glycine receptor activation is not essential for rod development.

Ciliary Neurotrophic Factor/Leukemia Inhibitory Factor/Pleiotrophin/Signal Transducer and Activators of Transcription 3/SOCS

Ciliary neurotrophic factor (CNTF) is a member of the interleukin 6 (IL-6) cytokine family and activates a trimeric transmembrane receptor complex consisting of the shared β receptors gp130 and LIFRβ (leukemia inhibitory factor receptor β) and a CNTFRα- specific subunit. Ligand binding leads to activation of gp130 and subsequently of associated tyrosine kinases (Janus kinases; JAKs), which phosphorylate the β subunits and signal transducer and activators of transcription (STAT) 1 and 3. CNTF can activate three pathways: JAK/STAT, mitogen-activated protein (MAP) kinase, and Akt/PI3 K signaling [187–191].

CNTF, other CNTF-like ligands, and LIF (activation by dimerization of gp130 and LIRFβ) exert diverse effects on photoreceptors in vertebrates. In developing chick photo receptors, CNTF promotes expression and relocalization of opsin proteins in cultured photoreceptors, as shown by labeling with the rho4D2 antibody, which recognizes

rhodopsin and green cone opsin [192–194]. Although no such effect of CNTF was found on cones using peanut lectin as a marker [192], Adler’s laboratory later observed that CNTF specifically promotes expression of green cone opsin but not rhodopsin and, interestingly, can induce coexpression of green and red cone opsin [195, 196]. Thus, in the chick these experiments indicated that CNTF produced by Müller glia might function to increase production of green cone opsin specifically, which is supported by the expression pattern of CNTFRα in the photoreceptor layer [197–199]. However, loss-of-function experiments need to be performed to obtain further insight of a regulatory role of CNTF in developing chick photoreceptors in vivo.

In the adult mammalian retina, CNTF protects, most likely indirectly, photoreceptors in models of retinal degeneration and injury [200–207]. In the developing rodent retina, CNTF exerts a different role. Here, CNTFR signaling controls the timing of terminal differentiation and the number of rod photoreceptors by acting on photoreceptor precursors. These functions are generally in good agreement with the spatiotemporal expression pattern of LIF, CNTF, and CNTF-like ligands such as neuropoeitin, CLF/ CLC rodents [208–210], or possible downstream mediators such as pleiotrophin [211], receptor expression [212–216], and with the activation of the signal transduction machinery [215, 217–219].

First shown in dissociated or explant cultures, CNTF and LIF act to suppress recoverin and rhodopsin expression [194, 220–223]. Interestingly, CNTF treatment does not alter the morphogenesis of rod outer segments; it acts reversibly on rod progenitors [218, 223]. The CNTF effect is transient since photoreceptor precursors become less responsive to CNTF with increasing age, and one explanation could be that CNTF receptor expression in photoreceptors decreases with maturation [218, 219, 221, 223]. These observations are supported by the timing of expression and activation of CNTFR pathway components in vivo and suggest that CNTF acts as a negative regulator preand perinatally to delay terminal photoreceptor differentiation [194, 215]. It is possible that formation of outer segments and concomitant changes in the surrounding ECM might interfere with cell division in the apical layer necessary to produce bipolar and Müller glia [223, 224]. Thus, the long delay between birth of photoreceptor precursors and their terminal differentiation could function to enable proper maturation of late-born retinal cell types.

At higher CNTF concentrations (e.g., 20 ng/ml), more cells with bipolar-specific gene expression appear in vitro [220, 223, 225, 226]. Originally proposed as an effect on switch of cell fate, in vivo and in vitro studies now indicate that CNTF and LIF keep photoreceptor precursors in an immature state without affecting specification of other cell types [209, 220–223, 227]. For example, disruption of the CNTFRα gene leads to an increase or acceleration of rod differentiation in explant cultures but does not affect bipolar cells [220]. Very similar observations were made following in vivo injection of an inhibitor of the CNTF/LIF pathway [209]. A possible explanation is that CNTF might upregulate expression of its own receptor in vitro, suggesting that the promoting effect on bipolar cells depends on culture conditions and does not reflect a physiological role in vivo [223].

Although capable of activating different pathways, the CNTF effects on rod development are exclusively mediated through rapid phosphorylation and activation of STAT3, which is independent of EGF signaling, another inhibitor of rod differentiation