Development of the paramedian lobule of the cerebellum in wild-type and tottering mice.

The mutant mouse tottering, (tg/tg), and the compound heterozygote mouse (tg/tg1a) exhibit three neurological disorders: ataxia, petit mal-like absence seizures and myoclonic intermittent movement disorders which are independent of the absence seizures. The tottering mouse carries an autosomal recessive single gene mutation on chromosome 8, and behavioral symptoms are first observed in the 3rd to 4th week of age. Using an additional genetic marker, Oligosyndactyly (Os), it is possible to distinguish tg/tg and tg/tg1a mice from wild-type mice at birth; nonaffected heterozygous littermates carry the Os mutation while tottering and compound heterozygous mice do not carry the Os gene. Similar to neurons found elsewhere in the brain, cerebellar Purkinje cells in both the wild-type and mutant mice were found to decrease in diameter with maturation. Forebrain weight, hindbrain weight, Purkinje cell dimensions and the thickness of the molecular layer in the paramedian lobule of the cerebellum in mutant mice were found to be reduced in mutants after, but not prior to the onset of behavioral symptoms.     

A novel gene mutation in the 60s loop of human coagulation factor VII - inhibition of interdomain crosstalk

A novel mutation in the factor VII gene resulting in procoagulant activity of 7.5% and antigen levels of 23% is presented. Single-stranded conformational polymorphism and DNA sequencing analysis revealed heterozygous shifts, and mutations were detected in exons 5, 7 and 8. The mutant L204P in exon 7 was novel, while the common polymorphisms, H115H and R353Q, were located in exons 5 and 8, respectively.The molecular effect of the L204P mutation was characterized using recombinant mammalian expression in Chinese hamster ovary cells. Low levels (4 ng/ml) of secreted mutant protein were found in transiently transfected cells compared to wild-type factor VII (83 ng/ml). Metabolic labeling demonstrated that the rate of mutant protein synthesis was similar to that of wild-type FVII, and the mutant protein accumulated intracellularly with no signs of increased degradation during a four-hour chase. No interaction between secreted P204 protein and immobilized soluble tissue factor was detected using surface plasmon reso-nance. The activation rate of recombinant mutant FVII protein was strongly reduced compared to wild-type FVII. A 9-fold reduction in the rate of FX activation was detected whereas Km was nearly the same for wild-type and the mutant. This slow rate was caused by a correspondingly lowered rate of P204 activation. A synthetic peptide sequence comprising amino acids 177-206 blocked binding of FVIIa to the TF-chip, and the subsequent factor X activation with an IC(50) value of 0.5 micro M in a chromogenic factor Xa assay. Additionally, evaluation of the peptide by surface plasmon resonance analysis resulted in inhibition of complex formation with an apparent K(I) of 7 micro M.     

A comparative study of Lrrk2 function in primary neuronal cultures

ObjectiveTo assess the contribution of wild-type, mutant and loss of leucine-rich repeat kinase-2 (LRRK2; Lrrk2) on dendritic neuronal arborization.BackgroundLRRK2 mutations are recognized as the major genetic determinant of susceptibility to Parkinson’s disease for which a cellular assay of Lrrk2 mutant function would facilitate the development of targeted molecular therapeutics.MethodsDendritic neuronal arborization (neurite length, branching and the number of processes per cell) was quantified in primary hippocampal and midbrain cultures derived from five lines of recombinant LRRK2 mice, including human BAC wild-type and mutant overexpressors (Y1699C and G2019S), murine knock-out and G2019S knock-in animals.ResultsNeuronal arborization in cultures from BAC Lrrk2 wild-type animals is comparable to non-transgenic littermate controls, despite high levels of human transgene expression. In contrast, primary neurons from both BAC mutant overexpressors presented with significantly reduced neuritic outgrowth and branching, although the total number of processes per cell remained comparable. The mutant-specific toxic gain-of-function observed in cultures from BAC mutant mice may be partially rescued by staurosporine treatment, a non-specific kinase inhibitor. In contrast, neuronal arborization is far more extensive in neuronal cultures derived from murine knock-out mice that lack endogenous Lrrk2 expression. In Lrrk2 G2019S knock-in mice, arguably the most physiologically relevant system, neuritic arborization is not impaired.ConclusionsImpairment of neuritic arborization is an exaggerated, albeit mutant specific, consequence of Lrrk2 over-expression in primary cultures. The phenotype and assay described provides a means to develop therapeutic agents that modulate the toxic gain-of-function conferred by mutant Lrrk2.     

Memory deficits of British dementia knock-in mice are prevented by Aβ-precursor protein haploinsufficiency

Familial British Dementia (FBD) is caused by an autosomal dominant mutation in the BRI2/ITM2B gene (Vidal et al., 1999). FBD(KI) mice are a model of FBD that is genetically congruous to the human disease, because they carry one mutant and one wild-type Bri2/Itm2b allele. Analysis of these mice has shown that the British mutation causes memory impairments due to loss of Bri2 function (Tamayev et al., 2010b). BRI2 is a physiologic inhibitor of processing of the Aβ-precursor protein (APP; Matsuda et al., 2008), a gene associated with Alzheimer's disease (Bertram et al., 2010). Here we show that APP haploinsufficiency prevents memory dysfunctions seen in FBD(KI) mice. This genetic suppression is consistent with a role for APP in the pathogenesis of memory deficits. Moreover, it provides compelling evidence that the memory dysfunctions caused by the British BRI2 mutant are dependent on endogenous APP and that BRI2 and APP functionally interact. This evidence establishes a mechanistic connection between Familial British and Alzheimer's dementias.     

Involvement of the ser-glu-pro motif in ligand species-dependent desensitisation of the rat gonadotrophin-releasing hormone receptor

There are two forms of gonadotrophin-releasing hormone (GnRH), GnRH-I and GnRH-II, in the vertebrate brain. Both GnRH-I and GnRH-II are thought to interact with the type-I GnRH receptor (GnRHR). The present study attempted to demonstrate whether GnRH-I and GnRH-II induce differential desensitisation of GnRHR and to identify the motif involved. Time course inositol phosphate (IP) accumulation assay reveals that, in cells expressing the wild-type rat GnRHR, GnRH-I induced continuous increase in IP production, whereas GnRH-II-induced IP production rate at later time points (30-120 min after ligand treatment) became attenuated. However, in cells expressing the mutant receptor in which the Ser-Glu-Pro (SEP) motif in extracellular loop 3 was replaced by Pro-Glu-Val (PEV), IP accumulation rates at later time points were more decreased by GnRH-I than GnRH-II. Ca2+ responses to repetitive GnRH applications reveal that GnRH-II desensitised the wild-type receptor faster than GnRH-I, whereas the opposite situation was observed in the PEV mutant. In addition, cell surface loss of GFP-tagged wild-type receptor was more facilitated by GnRH-II than GnRH-I, whereas that of the GFP-tagged PEV mutant receptor was more enhanced by GnRH-I than GnRH-II. The present study indicates that the SEP motif is potentially responsible for ligand species-dependent receptor desensitisation. Together, these results suggest that GnRH-I and GnRH-II may have different effects on mammalian type-I GnRHR via modulation of desensitisation rates.     

Presenilin-1 protects against neuronal apoptosis caused by its interacting protein PAG

Mutations in the presenilin-1 (PS-1) gene account for a significant fraction of familial Alzheimer's disease. The biological function of PS-1 is not well understood. We report here that the proliferation-associated gene (PAG) product, a protein of the thioredoxin peroxidase family, interacts with PS-1. Microinjection of a plasmid expressing PAG into superior cervical ganglion (SCG) sympathetic neurons in primary cultures led to apoptosis. Microinjection of plasmids expressing wild-type PS-1 or a PS-1 mutant with a deletion of exon 10 (PS1dE10) by themselves had no effect on the survival of primary SCG neurons. However, co-injection of wild-type PS-1 with PAG prevented neuronal death, whereas co-injection with the mutant PS-1 did not affect PAG-induced apoptosis. Furthermore, overexpression of PAG accelerated SCG neuronal death induced by nerve growth factor deprivation. This sensitizing effect was also blocked by wild-type PS-1, but not by PS1dE10. These results establish an assay for studying the function of PS-1 in primary neurons, reveal the neurotoxicity of a thioredoxin peroxidase, demonstrate a neuroprotective activity of the wild-type PS-1, and suggest possible involvement of defective neuroprotection by PS-1 mutants in neurodegeneration.     

Cytotoxicity of Mutant Huntingtin Fragment in Yeast Can Be Modulated by the Expression Level of Wild Type Huntingtin Fragment

相似文献   

Systemic overexpression of the alpha 1B-adrenergic receptor in mice: an animal model of epilepsy

PURPOSE: A lack of selective alpha1-adrenergic receptor (alpha1-ARs) agonists and antagonists has made it difficult to clarify the precise function of these receptors in the CNS. We recently generated transgenic mice that overexpress either wild-type or a constitutively active mutant alpha 1B-AR in tissues that normally express the receptor. Both wild-type and mutant mice showed an age-progressive neurodegeneration with locomotor impairment and probable stress-induced motor events, which can be partially reversed by alpha 1-AR antagonists. We hypothesized that the wild-type and mutant mice may exhibit spontaneous epileptogenicity as compared with normal (nontransgenic) mice. METHODS: Normal, wild-type, and mutant mice were studied. Twenty mice (1 year old) underwent prolonged video-EEG monitoring over a 4-week period. Raw EEG data were blindly analyzed by visual inspection for the presence of interictal and ictal epileptic activities. RESULTS: During the acute postoperative period (< or = 3 days), both wild-type (26.1 +/- 8.07 spikes/day) and mutant mice (116.87 +/- 55.13) exhibited more frequent interictal spikes than did normal mice (2.17 +/- 0.75; p value, <0.05), but all three groups showed EEG and clinical seizures. During the later monitoring periods (>3 days), wild-type and mutant mice showed more frequent interictal spikes (15.44 +/- 4.07; p < 0.01; and 6.05 +/- 2.46; p < 0.05, respectively) as compared with normal mice (0.41 +/- 0.41), but only mutant mice had spontaneous clinical seizures (means +/- SEM). CONCLUSIONS: The selective overexpression of the alpha 1B-AR is associated with increased in vivo spontaneous interictal epileptogenicity and EEG/behavioral seizures. These results suggest a possible role (direct or indirect) for the alpha 1B-ARs in the development and expression of epileptogenicity.     

Homocysteine induces oxidative cytotoxicity in Cu,Zn-superoxide dismutase mutant motor neuronal cell

Mutations in human Cu,Zn-superoxide dismutase (SOD1) cause approximately 20% of familial amyotrophic lateral sclerosis (FALS) cases. The mechanism of late-onset disease manifestation despite the innate mutation has no clear explanation. The relationship between homocysteine (HC) and amyotrophic lateral sclerosis (ALS) has not been investigated, in spite of the similarity in their pathogenesis. We investigated the effect of HC on the motor neuronal cell-line transfected with SOD1 of either wild-type or one of two mutant forms (G93A and A4V). In the MTT assay, HC induced significant cytotoxicity in A4V, but not in G93A, as compared with wild-type, even at the physiological concentration of 10 microM. This HC-induced cytotoxicity was inhibited by the antioxidant trolox and the Cu (I) chelator bathocuproinedisulfonate. Here we show that the vulnerability of the A4 V mutant involves the cytotoxic copper-mediated pathway, and that HC may be a lifelong precipitating factor in some forms of FALS, suggesting the possible treatment modality with vitamin supplements.     

Genetic modifications of voltage-sensitive sodium channels in Drosophila: gene dosage studies of the seizure locus

We have identified a genetic locus in Drosophila melanogaster whose product appears to have a structural role in the formation of functional voltage-sensitive sodium channels. This locus, designated seizure, is defined by two temperature-sensitive alleles (seits-1 and seits-2), each of which causes convulsive seizures followed by a rapid but reversible paralysis of adults at restrictive temperatures above 38 degrees C. Previous work had shown that seits-2 extracts display an altered pH dependence and an abnormally high Kd for [3H]-saxitoxin binding at high temperatures, suggesting that sodium channels in seits-2 mutants have an altered structure (Jackson F. R., S. D. Wilson, G. R. Strichartz, and L. M. Hall (1984) Nature 308: 189-191). These binding studies have now been extended to extracts of seits-1 which have a Kd not significantly different from wild-type at all assay temperatures. However, seits-1 extracts show a reduced number of saxitoxin binding sites (Bmax) relative to wild-type. This reduction is only 5 to 18% at 0 degree C but is 17 to 37% at 39 degrees C, suggesting that under certain conditions sodium channels in the seits-1 mutant are more labile than those of wild-type. Cytogenetic studies demonstrate that the seizure locus maps within region 60A to 60B8-10 on the second chromosome. Gene dosage analysis of approximately 99.7% of the genome, including this second chromosome region, failed to detect a wild-type locus whose dose affected saxitoxin-binding activity. Nevertheless, the mutant seits-2 allele has codominant and dose-dependent effects on paralytic behavior and saxitoxin-binding activity.(ABSTRACT TRUNCATED AT 250 WORDS)     

Dose-response analysis of locomotor activity and stereotypy in dopamine D3 receptor mutant mice following acute amphetamine

Accumulating evidence suggests that dopamine D3 receptor (D3R) stimulation is inhibitory to spontaneous and psychostimulant-induced locomotion through opposition of concurrent D1R and D2R-mediated signaling. To evaluate this model, we used homozygous D3R mutant mice and wild-type controls to investigate the role of the D3R in locomotor activity and stereotypy stimulated by acute amphetamine (AMPH) (0.2, 2.5, 5.0, 10.0 mg/kg). At the lowest dose tested (0.2 mg/kg), neither D3R mutant mice nor wild-type mice exhibited measurable change in locomotor activity or stereotypy relative to their respective saline-treated controls. D3R mutant mice exhibited a significantly greater increase in locomotor activity, but not stereotypy, relative to wild-type mice in response to treatment with AMPH 2.5 mg/kg. AMPH-induced locomotor activity and stereotypy were similar in both wild-type and D3R mutant mice at both the 5.0 and 10 mg/kg AMPH doses. These findings provide further support for an inhibitory role for the D3R in AMPH-induced locomotor activity, and demonstrate a more limited role for the D3R in modulating AMPH-induced stereotypy.     

Enhanced generation of intracellular Abeta42 amyloid peptide by mutation of presenilins PS1 and PS2

The accumulation of amyloid beta-peptide (Abeta) in the brain is a critical pathological process in Alzheimer's disease (AD). Recent studies have implicated intracellular Abeta in neurodegeneration in AD. To investigate the generation of intracellular Abeta, we established human neuroblastoma SH-SY5Y cells stably expressing wild-type amyloid precursor protein (APP), Swedish mutant APP, APP plus presenilin 1 (PS1) and presenilin 2 (PS2; wild-type or familial AD-associated mutant), and quantified intracellular Abeta40 and Abeta42 in formic acid extracts by sensitive Western blotting. Levels of both intracellular Abeta40 and Abeta42 were 2-3-fold higher in cells expressing Swedish APP, compared with those expressing wild-type APP. Intracellular Abeta42/Abeta40 ratios were approximately 0.5 in these cells. These ratios were increased markedly in cells expressing mutant PS1 or PS2 compared with those expressing their wild-type counterparts, consistent with the observed changes in secreted Abeta42/Abeta40 ratios. High total levels of intracellular Abeta were observed in cells expressing mutant PS2 because of a marked elevation of Abeta42. Immunofluorescence staining additionally revealed more intense Abeta42 immunoreactivity in mutant PS2-expressing cells than in wild-type cells, which was partially colocalized with immunoreactivity for the trans-Golgi network and endosomes. The data collectively indicate that PS mutations promote the accumulation of intracellular Abeta42, which appears to be localized in multiple subcellular compartments.     

Oxidized/misfolded superoxide dismutase‐1: the cause of all amyotrophic lateral sclerosis?

The identification in 1993 of superoxide dismutase-1 (SOD1) mutations as the cause of 10 to 20% of familial amyotrophic lateral sclerosis cases, which represents 1 to 2% of all amyotrophic lateral sclerosis (ALS) cases, prompted a substantial amount of research into the mechanisms of SOD1-mediated toxicity. Recent experiments have demonstrated that oxidation of wild-type SOD1 leads to its misfolding, causing it to gain many of the same toxic properties as mutant SOD1. In vitro studies of oxidized/misfolded SOD1 and in vivo studies of misfolded SOD1 have indicated that these protein species are selectively toxic to motor neurons, suggesting that oxidized/misfolded SOD1 could lead to ALS even in individuals who do not carry an SOD1 mutation. It has also been reported that glial cells secrete oxidized/misfolded mutant SOD1 to the extracellular environment, where it can trigger the selective death of motor neurons, offering a possible explanation for the noncell autonomous nature of mutant SOD1 toxicity and the rapid progression of disease once the first symptoms develop. Therefore, considering that sporadic (SALS) and familial ALS (FALS) cases are clinically indistinguishable, the toxic properties of mutated SOD1 are similar to that of oxidized/misfolded wild-type SOD1 (wtSOD1), and secreted/extracellular misfolded SOD1 is selectively toxic to motor neurons, we propose that oxidized/misfolded SOD1 is the cause of most forms of classic ALS and should be a prime target for the design of ALS treatments.     

Forebrain circumventricular organs mediate salt appetite induced by intravenous angiotensin II in rats

Mutations in the alpha-synuclein gene have been linked to rare cases of familial Parkinson's disease (PD). Alpha-synuclein is a major component of Lewy bodies (LB), a pathological hallmark of PD. Transgenic mice and Drosophila expressing either wild-type or mutant human alpha-synuclein develop motor deficits, LB-like inclusions in some neurons, and neuronal degeneration. However, the relationship between abnormal aggregates of alpha-synuclein and human dopamine (DA) neuron degeneration remains unclear. In this report, we have investigated the influence of alpha-synuclein expression on DA neurons in primary culture of embryonic human mesencephalon. Two days after culture, human DA cells were transduced with wild-type or mutant human (Ala(53)Thr) alpha-synuclein adenoviruses and maintained for 5 days. Overexpression of mutant and wild-type human alpha-synuclein resulted in 49% (P<0.01) and 27% (P<0.05) loss of DA neurons, respectively, while not affecting viability of other cells in the culture. Overexpression of rat alpha-synuclein or GFP (green fluorescent protein) had no effect on DA neuron survival. Cytoplasmic inclusions of alpha-synuclein were detected immunohistochemically in DA cells transduced with mutant human alpha-synuclein, but not wild-type alpha-synuclein. These results show that overexpression of human alpha-synuclein, particularly the mutant form, can cause human DA neuron death, suggesting that alpha-synuclein may have a primary role in the pathogenesis of PD.     

New DNA Diagnostic System for Detection of Factor V Leiden

Use was made of allele-specific PCR to develop a highly effective DNA diagnostic system for detection of the factor V Leiden mutation in exon 10 of human factor V gene. The allele-specific primers contain a 3′-OH end nucleotide, which matches a mutant or wild-type nucleotide of the template DNA (A-allele and G-allele, respectively) and also one mismatched nucleotide near the 3′-end. The universal primers have an internal mismatch with the mutant nucleotide of the template DNA and another mismatched nucleotide at 3′-OH end. The A-allele-specific primer enables the preferential amplification of both the homozygous and heterozygous mutant alleles. The extension of the G-allele-specific primer or the universal one is inhibited in the presence of the homozygous factor V Leiden. The developed assay system allowed us to detect five patients, who are heterozygous for factor V Leiden among the 48 patients with deep venous thrombosis and pulmonary thromboembolism.     

Development of serotonin-containing neurons in Drosophila mutants unable to synthesize serotonin

We have initiated a study of the CNS of mutant Drosophila melanogaster larvae carrying a genetic deletion of the gene Ddc that encodes the enzyme dopa decarboxylase (DDC). The two major objectives of this study were (1) to ascertain that the DDC encoded by the gene Ddc was the only decarboxylase utilized in serotonin (5HT)-containing neurons and (2) to determine the effect of DDC deficiency on the development of 5HT-immunoreactive neurons. CNSs of wild-type larvae and of larvae genetically deficient for the gene Ddc were processed for serotonin immunocytochemistry using a monoclonal antibody against 5HT. The pattern of 5HT immunoreactivity in the wild-type and the Ddc-deficient CNS is compared. In contrast to the wild-type, 5HT immunoreactivity is absent in the Ddc-deficient CNSs. The lack of immunocytochemically detectable 5HT in the mutant CNSs is consistent with the idea that the DDC encoded by the gene Ddc is utilized in 5HT-containing neurons. To study the development of neurons committed to the 5HT differentiation pathway in the absence of 5HT, we used a second biochemical property characteristic of 5HT-containing neurons, the ability to take up 5HT. CNSs from mutant animals were incubated in exogenous 5HT and the accumulated 5HT detected immunocytochemically. Neurons capable of selective 5HT uptake were present in the mutant CNSs in the same pattern as the 5HT-immunoreactive neurons in the wild-type CNS. This result suggests that the presumed inability to synthesize 5HT does not preclude differentiation of other normal biochemical properties of 5HT-containing neurons.     

Toward therapy for DYT1 dystonia: allele-specific silencing of mutant TorsinA

A three-nucleotide (GAG) deletion in the TOR1A gene is the most common cause of inherited dystonia, DYT1. Because the mutant protein, TorsinA (TA), is thought to act in a dominant manner to cause disease, inhibiting expression from the mutant gene represents a potentially powerful therapeutic strategy. In an effort to develop therapy for this disease, we tested whether small interfering RNA (siRNA) could selectively silence expression of mutant TA. Exploiting the three-base pair difference between wild-type and mutant alleles, we designed siRNAs to silence expression of mutant, wild-type, or both forms of TA. In transfected cells, siRNA successfully suppressed wild-type or mutant TA in an allele-specific manner: for example, mutant-specific siRNA reduced the levels of mutant TA to less than 1% of controls with minimal effect on wild-type TA expression. In cells expressing both alleles, thus simulating the heterozygous state, siRNA-mediated suppression remained robust and allele specific. Our siRNA studies demonstrate allele-specific targeting of a dominant neurogenetic disease gene and suggest the broad therapeutic potential of siRNA for DYT1 dystonia and other dominantly inherited neurological diseases.     

A dominant negative mutation of neuronal connexin 36 that blocks intercellular permeability

Rat connexin 36 (Cx36) was mutated by substituting serine for cysteine at residue 231 (C231S) and the mutant's effect on the subcellular localization of wild-type Cx36 and the intercellular permeability that it confers was determined in human HeLa and rat PC12 cells. Cells transfected with the mutant or wild-type Cx36 cDNA expressed the expected 36 kDa protein and Cx36 immunoreactivity. Co-immunoprecipitation experiments with monkey COS-7 cells transiently transfected with both mutant and wild-type Cx36 cDNAs demonstrated that the mutant protein bound to the wild-type. Double immunofluorescence microscopy of stably transfected HeLa cells demonstrated that mutant Cx36 blocked the transport of the wild-type Cx36 to the cell membrane, primarily by trapping it in the endoplasmic reticulum around the nucleus. Coexpression of the mutant Cx36 with the wild-type protein abolished the ability of the latter to permit dye transfer in both HeLa and PC12 cells. The findings are the first demonstration of a mutation of Cx36 that inhibits wild-type Cx36 function in mammalian cells.     

Lazy eyes zebrafish mutation affects Müller glial cells,compromising photoreceptor function and causing partial blindness

A behavioral assay based on the optokinetic reflex was used to screen chemically mutagenized zebrafish larvae for deficits in visual function. A homozygous recessive mutation, lazy eyes (lze), was isolated based on the observation that 5-day postfertilization (dpf) mutants displayed weaker and less frequent eye movements than wild-type fish in response to moving stripes. Electroretinographic (ERG) recordings revealed that mutants had severely reduced a- and b-wave amplitudes relative to wild-type fish, indicating outer retinal dysfunction. Retinal lamination and cellular differentiation were normal in the lze retina; however, mutant photoreceptor cells had small outer segments and pyknotic nuclei were occasionally observed in the outer retina and the marginal zone of lze. Cone, rod, amacrine, bipolar, and Müller cell marker analyses indicated that the typical lze retina contained fewer rod photoreceptors and fewer Müller cells than wild-type fish at 5 dpf. At 3 dpf, however, mutant retinas had normal numbers of rod photoreceptors and Müller cells, suggesting that the initial differentiation of these cell types occurred normally. Rod photoreceptor histology was normal at this early stage, but Müller cells were often hypertrophied, suggesting that they were unhealthy. Constant light rearing of mutant animals accelerated the Müller cell degeneration, severely worsened the visual deficit, but had no obvious affect on the photoreceptors. When ERG responses and Müller cell degeneration from the same mutant animals were analyzed, the extent of the Müller cell loss matched closely the degree to which ERG responses were reduced. In summary, the lze gene appears to be required for Müller cell viability and normal visual function. The lze mutant may be a model for the study of the involvement of Müller cells in photoreceptor development and function.