Inactivation of the microRNA<em>-183/96/182</em> cluster results in syndromic retinal degeneration

miR-183/96/182 Cluster Resides in the Intron of a Potentially Protein-Coding Gene.

The genomic features around miR-183/96/182 include two CpG islands ∼3.4–6.5 kb 5′ of premiR-183, the most 5′ miRNA of the cluster (Fig. 1A; and SI Appendix, Fig. S1 and Sequence (Seq.) S1) (http://genome.UCSC.edu/) (8). A cDNA clone, {"type":"entrez-nucleotide","attrs":{"text":"AK044220","term_id":"26090242","term_text":"AK044220"}}AK044220, extends ∼3.2–4.6 kb 5′ to premiR-183, encompassing the second CpG island (Fig. 1A and SI Appendix, Fig. S1). Multiple expressed sequences detected in gene-trap clones (27), including D016D06 (28, 29) [deposited by the German GeneTrap Consortium (GGTC) (http://tikus.gsf.de)], colocalize in {"type":"entrez-nucleotide","attrs":{"text":"AK044220","term_id":"26090242","term_text":"AK044220"}}AK044220 (Fig. 1A and SI Appendix, Fig. S1), suggesting a 5′ exon of the miR-183/96/182 gene. The region 3′ of miR-183/96/182 contains an EST clone, {"type":"entrez-nucleotide","attrs":{"text":"BB709579","term_id":"16062748","term_text":"BB709579"}}BB709579, ∼6 kb 3′ of premiR-182 (Fig. 1A), suggesting the miR-183/96/182 gene has a downstream exon. To confirm this model, we designed multiple primers complementary to the predicted exons to perform 5′ and 3′ RACE and RT-PCR using total RNA of the mouse retina and D016D06 ESCs followed by sequencing (see SI Appendix, Figs. S2 and S3 for details).

Fig. 1.

Genomic structures of the miR-183/96/182 gene and the miR-183C ^{allele of the gene-trap cell line D016D06. (}A) Genomic features around the miR-183/96/182 gene compiled from the University of California, Santa Cruz Genome Browser (Mouse July 2007 Assembly). The blue arrow points to the gene-trap clone, D016D06. The open arrowhead points to the mouse cDNA clone, {"type":"entrez-nucleotide","attrs":{"text":"AK044220","term_id":"26090242","term_text":"AK044220"}}AK044220. tel, telomeric; cent, centromeric. (B) Genomic structure of the miR-183/96/182 gene. Blue boxes represent exons. The dotted blue lines indicate splicing events. pA, polyA signal. One-headed arrows mark locations and directions of exonic primers 2F, 2.1F, 3F, 3R, 5R, 5Rnest, and 6R. The numbers above the double-headed arrows denote the distances (in bp) between the vertical lines. The wavy end of Ex2 indicates that the 3′ end of this exon not yet fully defined. Alternative splicing events result in two forms of transcripts, transcripts I and II. Transcript II has four putative ORFs. In phase I of the sequence, two ORFs consist of 98 (P98-1) and 70 (P70-1) codons, respectively. Phases 2 and 3 each have an ORF of 63 codons (P63-2 and P63-3). (C) Genomic structure of the miR-183C ^{allele of the gene-trap clone D016D06. The gene-trap construct was inserted in intron 1 of the} miR-183/96/182 gene. Location and directions of primers, 4Fintron1, 5Fintron1, 353, 354, and 361 are labeled. LTR, long tandem repeat.

Our results showed that the miR-183/96/182 gene spans more than 15 kb, with at least two alternatively spliced transcripts (Fig. 1B). Both transcripts start from the nucleotide (nt) G, 3,943 nt 5′ to premiR-183 (−3943G) and use the same exon 1 (Ex1) (Fig. 1B and SI Appendix, Seq. S1). Transcript I is at least 979 nt and comprises Ex1 and exon 2 (Ex2), separated by a 3,213-bp intron. However, the 3′ RACE product detecting Ex2 stopped 197 nt 5′ to premiR-183 and did not extend beyond miR-183/96/182 (Fig. 1B and SI Appendix, Seq. S1). Thus, the 3′ end of transcript I is not yet fully determined. Transcript II is about 1,432 nt and consists of Ex1 and an alternatively spliced exon 2′ (Ex2′), which is 14 kb downstream of Ex1. Ex2′ is 986 bp in size with a polyA signal (Fig. 1B and SI Appendix, Seq. S1). The structure of transcript II clearly positions miR-183/96/182 in the intron of the miR-183/96/182 gene (Fig. 1B and SI Appendix, Seq. S1).

Transcript I has no ORF >30 codons, whereas transcript II has at least four ORFs, all of which reside in Ex 2′ (Fig. 1B and SI Appendix, Seqs. S2 and S3). In phase I of the sequence, two ORFs consist of 98 (P98-1) and 70 (P70-1) codons, respectively. Phases 2 and 3 each have an ORF of 63 codons (P63-2 and P63-3). None of the putative peptides contain known conserved functional domains; however, P98-1 encodes motifs found in proteins from multiple species (SI Appendix, Fig. S4). Thus, there is a possibility that the miR-183/96/182 gene encodes one or more peptides.

Characterization of the Gene-Trap Allele, miR-183C^{, and Production of} miR-183C ^Mice.

Gene-trap clone D016D06 was derived from insertion of a retroviral construct, rFlpROSAβgeo, in 129S2 ESCs (Fig. 1 A and C) (http://tikus.gsf.de) (28, 29). The rFlpROSAβgeo construct carries a promoterless reporter gene, the β-geo cassette [an in-frame fusion of the β-galactosidase and neomycin resistance genes (30)], with a splicing acceptor (SA) immediately upstream and a polyA signal downstream of the β-geo cassette (Fig. 1C). To determine the location of the insertion of the construct in the “gene-trapped” allele, designated as “miR-183C^{,” we sequenced the product of genomic PCR using primers in intron 1 and in the gene-trap construct (}Fig. 1C and SI Appendix, Seq. S1). Our results revealed that the gene-trap construct was inserted 1,848 bp downstream of Ex1 in intron 1 of the miR-183/96/182 gene. In the trapped allele, Ex1 of miR-183/96/182 gene spliced to β-geo, creating a fusion transcript (Fig. 1C) that stops after the polyA signal of the β-geo cassette, preventing transcription of miR-183/96/182. This result predicts that miR-183C ^{mice would have complete loss of function of the} miR-183/96/182 gene, whereas the β-geo cassette under the control of native regulatory elements should report endogenous expression pattern of the gene.

To test this model, we generated chimeric mice using the D016D06 ESCs (SI Appendix, Figs. S5–S7 and Table S1). Germline-transmitting F1 and F2 mice were produced. Genotypes of the F2 mice followed the expected Mendelian inheritance pattern (SI Appendix, Fig. S8). The miR-183C ^{mice generally had decreased body weight (by ∼24%) and size (}SI Appendix, Fig. S8); however, their overall survival and life span were not significantly different from their wild-type and heterozygous littermates.

miR-183/96/182 Gene Is Inactivated in the Retina and All Sensory Organs of miR-183C ^Mice.

Using RT-PCR (Fig. 2 A and C) and Northern blot analyses (Fig. 2B), we showed that both alternatively spliced transcripts (Fig. 2A) and three mature miRNAs of the miR-183/96/182 gene were undetectable in the retina of miR-183C ^{mice (}Fig. 2 B and C), whereas the Ex1/β-geo fusion transcript was present only in miR-183C ^{but not in the wild-type retina (}Fig. 2A). Similarly, all 3p miRNAs of the cluster also were abolished in the retina (SI Appendix, Fig. S9B).

Fig. 2.

The miR-183/96/182 gene is inactivated in the retina and other sensory organs of miR-183C ^{mice. (}A) RT-PCR analysis of the transcripts of the miR-183/96/182 gene and the fusion transcript of the miR-183C ^{allele. Transcript I was detected by RT-PCR using primers 2F/3R followed by nested PCR using 2.1F/3R-1. Transcript II was detected by RT-PCR using 2F/5R and nested PCR using 2.1F/5Rnest. The Ex1/β-geo fusion transcript of the} miR-183C ^{allele was detected by RT-PCR with primers 2F/254 and nested PCR with 2.1F/361. (}B) Northern blot analysis. (C–G) qRT-PCR analysis of the expression of members of the miR-183/96/182 cluster in the retina (C), IE (D), tongue epithelia (E), DRG (F), and OE (G). *P < 0.05; **P < 0.01; ***P < 0.001. miR-183C^: n = 4; miR-183C^: n = 4 (6 for DRG); miR-183C^: n = 3 (4 for DRG).

To test whether the expression of miR-183/96/182 was affected in other sensory organs, we performed quantitative RT-PCR (qRT-PCR) analysis using RNA isolated from the inner ear (IE), tongue epithelia, dorsal root ganglions (DRGs), and olfactory epithelia (OE) (Fig. 2 D–G) and found no detectable expression of miR-183/96/182 in any major sensory organs of miR-183C ^mice.

We conclude that the miR-183/96/182 gene was completely inactivated in miR-183C ^mice.

β-Geo Cassette Reports Endogenous Expression Patterns of the miR-183/96/182 Gene.

X-Gal staining in the adult miR-183C ^{retina showed specific expression of β-gal in all photoreceptors in the outer nuclear layer (ONL) and in a subgroup of ganglion cells (}Fig. 3 A–C), similar to the previously described expression pattern of miR-183/96/182 (5, 8–10). We also observed staining in the outer portion of the inner nuclear layer (INL); the punctate pattern suggested that the X-gal positive signals in the INL possibly were derived from synapses of the photoreceptors (SI Appendix, Fig. S10).

Fig. 3.

Expression of the β-geo cassette reflects the endogenous expression patterns of the miR-183/96/182 gene in the retina. (A–C) X-Gal staining of the retina. Red arrowheads in A and C show the expression in a subgroup of ganglion cells. OS, outer segment; IS, inner segment; IPL, inner plexiform layer; GCL, ganglion cell layer. (D) Expression of the β-geo cassette in photoreceptors of miR-183C ^{mice has a circadian rhythm. X-Gal staining of retinal sections of} miR-183C ^{mice harvested every 4 h in a circadian cycle. (Scale bar: 20 μm.) Red dotted lines indicate unit squares used to measure optic density shown in} E. (E) Relative optic density of X-Gal signals in the inner segment of photoreceptors in a unit square (illustrated in D). n = 3 at all time points. *P < 0.05; **P < 0.01 compared with CT1. (F) X-Gal staining of the retina of E14.5, E16.5, E18.5, P1, P2, P5, P10, and adult miR-183C ^{mice. Arrows in P2 indicate positive signals in the outer neuroblast layer. The OPL started to form at P5. At this stage, ∼40% of photoreceptors are still in the INL (}115) and are positive to X-Gal staining. INBL, inner neuroblast layer.

X-Gal staining of retina harvested every 4 h from adult miR-183C ^{mice maintained in constant darkness showed that the intensity of X-Gal staining in photoreceptors followed a circadian rhythm, with peak intensity at Circadian Time (CT)1 and trough at CT13 (}Fig. 3 D and E). This rhythm is in an approximately opposite phase to that we detected for miR-183/96/182 at the RNA level (8), likely reflecting the delay from transcription to the production of functional β-gal protein. Similar delays from transcription to protein expression have been reported in many other rhythmic genes (31–33).

To demonstrate the developmental expression pattern, we performed X-Gal staining in the retinas of embryonic day (E)10.5, E14.5, E16.5, and E18.5 and postnatal day (P)1, P2, P5, and P10 heterozygous mice. The result showed that, similar to our previous qRT-PCR analysis on the mature miRNAs of the cluster (8), β-gal was not detectable in the embryonic stages or at P1 but started to express in the outer neuroblast layer (ONBL) at P2, was drastically up-regulated at P5, and was increased further at P10 and the adult stage (Fig. 3F). We conclude that the β-geo cassette of the miR-183C ^{allele reports the endogenous expression pattern of the} miR-183/96/182 gene in the retina developmentally, spatially, and temporally.

Also consistent with our previous results at the RNA level, X-Gal staining in other sensory organs of miR-183C ^{mice showed specific expression in all sensory epithelia in the IE, OE, and the pheromone-sensing vomeronasal organ (VNO), papillae of tongue epithelia, and DRGs (}Fig. 4; and SI Appendix, Figs. S11–S13). We also observed abundant expression in the olfactory bulbs (Fig. 4 C, a and SI Appendix, Fig. S14), possibly as a result of the expression in the axons and synaptic terminals of receptor neurons in the OE and VNO. Like many other retina-specific genes (34, 35), the miR-183/96/182 gene also is expressed in the pineal body (Fig. 4 C, b and SI Appendix, Fig. S14).

Fig. 4.

Expression of the β-geo cassette reflects the endogenous expression pattern of the miR-183/96/182 cluster in other sensory and related organs. X-Gal staining of the IE (A a–c), OE and VNO (B), OB (C a), pineal body (C b), papillae of the tongue (D a–d), and DRG (E a and b) of adult miR-183C ^mice.

Inactivation of the miR-183/96/182 Gene Results in Functional Defects in the Retina.

To assess retinal function, we performed scotopic (Fig. 5 A and B) and photopic (Fig. 5 C and D) electroretinograms (ERGs) on 5-wk-, 6-mo-, and 1-y-old miR-183C ^{mice. The most prominent defect, decreased} b-wave amplitude in miR-183C ^{mice as compared with their age-matched, wild-type controls, was common to scotopic (}Fig. 5 B, g–i) and photopic ERGs (Fig. 5D) and was detectable in mice as early as 5 wk of age, at a time when the histology of the retina was normal (see below). The ERG abnormalities progressed with age, and photopic b-wave amplitudes were more severely affected than scotopic ERGs (Fig. 5 and SI Appendix, Fig. S15 and Table S2). At 1 y of age, the photopic b-wave amplitude of miR-183C ^{mice was <10% that of wild-type littermate controls (}Fig. 5D). The ERG b wave is derived mostly from activities of depolarizing bipolar cells (36–39); defects in bipolar cells and photoreceptor synaptic transmission can produce defective b waves. Because miR-183/96/182 is expressed predominantly in photoreceptors, it is likely that the reduction in b-wave amplitude in miR-183C ^{mice reflects defective synaptic transmission in rods and cones.}

Fig. 5.

Inactivation of the miR-183/96/182 gene resulted in progressive ERG defects. (A and C) Representative scotopic (A) and photopic (C) ERG recordings of 5-wk-, 6-mo-, and 1-y-old mice. In A, blue, black, red, green, and violet lines represent recordings at 0, 10, 100, 1 k, and 10 k mcd⋅s⋅m^{, respectively. (}B) Comparison of scotopic a- and b-wave amplitudes, latency, and implicit time of miR-183C ^{and age-matched wild-type control mice.} miR-183C^: n = 12, 4, and 4 for 5-wk- (a, d, g, j), 6-mo- (b, e, h, k), and 1-y-old (c, f, i, l) mice, respectively; miR-183C^: n = 10, 3, and 4 for 5-wk-, 6-mo-, and 1-y-old mice, respectively. *P < 0.05; **P < 0.01; ***P < 0.001 between miR-183C ^{and age-matched wild-type controls. (}D) Comparison of photopic a- and b-wave amplitudes, latency and implicit time of miR-183C ^{and age-matched wild-type control mice.} miR-183C^: n = 3 at all ages; miR-183C^: n = 4, 3, and 4 for 5-wk-, 6-mo-, and 1-y-old mice, respectively. Error bars indicate SEM. ^P < 0.05; ^P < 0.01; ^P < 0.001 between two age groups.

The photopic and scotopic a-wave amplitudes were normal in 5-wk-old miR-183C ^{mice but decreased with age, and, as with the} b wave, the photopic amplitude decreased more quickly than the scotopic amplitude (Fig. 5 B, a–c and D): At 6 mo of age the photopic a-wave amplitude was reduced to 40% of control, and by 1 y it was <10% of control. The scotopic a-wave amplitude was normal until age 1 y, when it was reduced to 43–65% of control (Fig. 5D and SI Appendix, Table S2). The ERG a wave reflects the electrical response of photoreceptors to light (40–42). These results suggest that in miR-183C ^{mice the phototransduction pathways of rods and cones became significantly compromised with time, cones more so than rods.}

ERG latency, the time from the onset of stimulus to the beginning of the a wave, and implicit time, the time from the onset of light stimulus to the peak of the b wave, of both scotopic and photopic ERGs were increased in miR-183C ^{mice from age 6 mo, with more severe and persistent changes in photopic ERGs (}Fig. 5 B and D and SI Appendix, Table S2), suggesting more severe synaptic transmission defects in the cone pathway.

ERG tests in heterozygous miR-183C ^{mice at all ages showed no significant defects compared with their age-matched wild-type controls.}

Inactivation of the miR-183/96/182 Gene Results in Progressive Retinal Degeneration.

At 5 wk of age, despite significant ERG abnormalities, the retinas of miR-183C ^{mice were histologically indistinguishable from those of littermate controls (}Fig. 6A). By age 6 mo, however, miR-183C ^{mice showed retinal degeneration with decreased thickness of the ONL (}Fig. 6B), which had progressed at 1 y of age (Fig. 6C). Intriguingly, the superior retina was more affected than the inferior retina. At age 6 mo, the number of nuclear layers of the ONL of the superior retina was decreased by ∼30%, whereas no significant changes were detected in the inferior retina (Fig. 6B). At age 1 y, ONL thickness was decreased in the entire eye, but the superior retina continued to be affected more severely (decreased by 40–52%) than the inferior retina (decreased by 30–43%) (Fig. 6C).

Fig. 6.

Inactivation of the miR-183/96/182 gene resulted in progressive retinal degeneration and increased susceptibility to light damage. (A–C) (Left) H&amp;E staining of sagittal sections of the retina of 5-wk- (A), 6-mo- (B), and 1-y-old (C) miR-183C ^{and wild-type littermate control mice (}miR-183C^{). (}Right) Measurements of the thickness the ONL (Upper) and the number of nuclear layers of the ONL at different locations in reference to the optic nerve (Lower). Sup, superior; Inf, inferior. (D) H&amp;E staining (Left) and measurement (Right) of the retina of 6-mo-old miR-183C ^{and wild-type littermate control mice 2 wk after a 2-h exposure to bright light. P} < 0.06; *P < 0.05; **P < 0.01; ***P < 0.001.

Consistent with ERG results, heterozygous miR-183C ^{mice showed no significant retinal degeneration compared with wild-type littermates at all ages tested.}

Inactivation of the miR-183/96/182 Gene Results in Increased Susceptibility to Light Damage.

To determine if miR-183/96/182 influences the response of the retina to light damage, we exposed miR-183C ^{mice and wild-type littermates to 10,000-lx cool, white fluorescent light for 2 h followed by 2 wk in the dark. Retinas of} miR-183C ^{mice showed increased light toxicity as compared with wild-type littermates (}Fig. 6D and SI Appendix, Fig. S16). As with aging, the superior retina was more damaged than the inferior retina. The number of nuclear layers in ONL of the superior retina was decreased by 48–62% in miR-183C ^{mice compared with their wild-type littermates; this decrease was similar to or worse than the extent of degeneration in 1-y-old} miR-183C ^{mice without light exposure (}Fig. 6 C and D). The inferior retina had much less damage, suggesting that the superior retina is intrinsically more vulnerable to degenerative changes. This result is consistent with a recent report that knockdown of miR-183/96/182 in postmitotic photoreceptors resulted in increased susceptibility to light damage in the superior but not the inferior retina (26).

Inactivation of the miR-183/96/182 Gene Results in Photoreceptor Ribbon Synaptic Defects.

To identify target genes of miR-183/96/182 and to understand the mechanisms of functional defects in the retina of miR-183C ^{mice, we performed gene-expression profiling using the Affymetrix GeneChip Mouse Genome 430A_2.0 and total RNA from 5-wk-old} miR-183C ^{mice, a time when there is no histological evidence of degeneration in the retina (}SI Appendix, Fig. S17 and Tables S3–S9). A total of 1,341 genes showed significant differences in expression levels in miR-183C ^{mice as compared with wild-type littermates (}P < 0.05; fold change ≥1.2 folds) (SI Appendix, Tables S4–S6). Functional annotation analysis of the genes with altered expression showed significant enrichment for those involved in synaptogenesis (P = 3.2 × 10^{), synaptic contact (}P = 3.9 × 10^{), and transmission of nerve impulses (}P = 6.8 × 10^{) (}SI Appendix, Table S7 A and B). In addition, key members of the classical complement system and Class I MHC (MHCI) molecules also were significantly enriched in the genes with increased expression (SI Appendix, Table S8). Members of the classical complement cascade and MHCI are expressed in retina and the central nervous system and play important roles in synaptic remodeling and connectivity (43–49). These results are consistent with the ERG data indicating that inactivation of miR-183/96/182 gene results in abnormal photoreceptor synaptic transmission.

Neurotransmission of retinal photoreceptors is accomplished by a unique type of chemical synapse—the ribbon synapse—at their termini (50, 51). To evaluate the status of photoreceptor termini and ribbon synapses, we first performed immunofluorescence (IF) using synaptic markers. Ribeye/CtBP2, a major component of synaptic ribbons (52), normally has a horseshoe-shaped staining pattern in the wild-type photoreceptor termini (Fig. 7 A and C). However, Ribeye/CtBP2 staining became shorter and less defined, with dispersed immunoreactive puncta of variable size and less intensity in the miR-183C ^{retina as early as age 5 wk (}Fig. 7 A, B, E, and F), suggesting early-onset abnormalities of synaptic ribbons. This difference became more striking in 1-y-old animals (Fig. 7 C and D), suggesting the progression of synaptic defects. Consistently, peanut agglutinin (PNA) staining, which labels the inner and outer segment as well as pedicles of cone photoreceptors (53), showed that the number and size of cone pedicles apparently were decreased in 5-wk-old miR-183C ^{mice (}Fig. 7 E and F) and were reduced further in 1-y-old mice (SI Appendix, Fig. S18), suggesting defects in the maturation of photoreceptor termini.

Fig. 7.

Inactivation of the miR-183/96/182 gene resulted in morphological and molecular changes in the synaptic ribbons of photoreceptors in miR-183C ^{mice. (}A–D) Confocal images of Ribeye IF in the retinas of 5-wk- (A and B) and 1-y-old (C and D) miR-183C ^{mice (}B and D) and their wild-type littermate controls (A and C). (E and F) Coimmunostaining of Ribeye (green) and PNA (red) in retinas of 5-wk-old miR-183C ^{mice (}F) and their wild-type littermate controls (E). (G) An electron micrograph of a ribbon synapse of an miR-183C ^{mouse, showing a typical triad configuration of photoreceptor ribbon synapse. SR, synaptic ribbon; H, horizontal cell; B, bipolar cell. (}H) Example of a 3D reconstruction of synaptic ribbons in the OPL of miR-183C ^{mice (}n = 2) and their wild-type littermates (miR-183C ^mice, n = 2), respectively. (Scale cube, 500 nm^{.) (}I and J) Distribution of synaptic ribbons by size (I) and mean ribbon areas of miR-183C ^(red; n = 72) and miR-183C ^{littermate controls (black;} n = 49) (J). Data are shown as mean ± SEM; *P < 0.01.

To evaluate the photoreceptor synaptic structures directly, we performed EM of the retinas of 5-wk-old mice. Consistent with Ribeye/CtBP2 IF data, EM showed that, although synaptic ribbons were developed, ribbon size was decreased significantly in the miR-183C ^{mice as compared with their wild-type littermates (}Fig. 7 G–J).

Taken together, these observations support the notion that members of miR-183/96/182 modulate synaptic connections of photoreceptors to downstream neurons by directly and indirectly regulating a wide range of molecules involved in synaptogenesis and synaptic transmissions, contributing to retinal functional defects of miR-183C ^{mice (see} Fig. 9).

Fig. 9.

Hypothetical model of in vivo functions of the miR-183/96/182 gene in the retina. Members of miR-183/96/182 directly and indirectly regulate multiple genes involved in photoreceptor maturation and synaptic connections with downstream bipolar cells. When the miR-183/96/182 gene is inactivated, the homeostasis of the photoreceptors is disrupted, resulting in defects in photoreceptors and their synaptic transmission, leading to progressive retinal dysfunction and degeneration.

Inactivation of the miR-183/96/182 Gene Disrupts Normal Expression of Photoreceptor Genes and Postnatal Development of Photoreceptors, Contributing to Retinal Degeneration.

Among the genes with abnormal expression are some known to be important in the phototransduction pathway and to be responsible for retinal diseases (SI Appendix, Table S9). By qRT-PCR analysis, we confirmed reduced expression of M-cone opsin, Opn1mw, and cone-specific arrestin, Arr3, by 2.54- and 2.74-fold, respectively (SI Appendix, Table S9; see also SI Appendix, Table S11B). IF of whole-mount retina and retinal sections validated that Opn1mw was decreased in the retina of 5-wk-old miR-183C ^{mice (}Fig. 8 A and B). Quantification of M opsin- and PNA-positive cones in the whole-mount retina showed that both the total number and the percentage of M opsin-positive cones in total cones (PNA-positive) were decreased significantly in the superior retina, by 49% and 26%, respectively (Fig. 8A), whereas no significant changes were seen in the total number of cones or the number of M opsin–positive cones in the inferior retina. At age 1 y M opsin was almost completely lost in miR-183C ^{retina (}Fig. 8C). These data suggest that M opsin-positive cones were the most affected photoreceptors in miR-183C ^{mice and that this increased functional vulnerability of M opsin-positive cones leads to more severe and faster-progressing defects in photopic ERGs (}Fig. 5) and to the uneven distribution of retinal degeneration in miR-183C ^{mice (}Fig. 6).

Fig. 8.

Inactivation of the miR-183/96/182 gene resulted in a significant decrease of M opsin expression in the retina. PNA staining and co-IF of whole-mount retina (A) and retinal sections (B and C) of 5-wk- (A and B) and 1-y-old (C) miR-183C ^{mice and their wild-type (}miR-183C^{) littermate controls. The histograms at the bottom of} A show the number of M opsin–positive (M opsin+) cones, the total number of cones per square micrometer, and the percentage of total cones that were M opsin positive in the superior (sup.) and inferior (inf.) retina.

Inactivation of the miR-183/96/182 Gene Also Results in Vestibular Dysfunction.

In addition to retinal defects, miR-183C ^{mice exhibit circling behavior and an unstable gait (}Movies S1 and S2). These phenotypic features are typical of defects in vestibular organs (54) and appeared as early as 4 wk of age. Because miR-183/96/182 is highly expressed in the vestibular organs (Fig. 4 A and C), this phenotype likely indicates defects in the vestibular organ and possibly in other sensory organs. In the current work we have focused on the retina.

Gene-Trap Cell Line and Mouse Production.

Gene-trap ESC clone D016D06, was obtained from the German Gene Trap Consortium (http://tikus.gsf.de/). Standard protocols of ESC culture were followed as described previously (108). Chimeric mice were produced at the Transgenic Core Facility of the University of Illinois at Chicago using standard protocols (109). See SI Appendix for more details of ESC characterization and mouse production. Mice were kept in 12-h light (<100 lx)12-h dark cycles with light on at 7:00 AM [zeitgeber time (ZT) 0] and off at 7:00 PM (ZT12). Animal care and husbandry followed National Institute of Health (NIH) and Association for Research in Vision and Ophthalmology (ARVO) guidelines. All protocols were approved by the Institutional Animal Care and Use Committee of Rush University Medical Center.

RNA Isolation and Northern Blot Analysis.

Protocols for RNA isolation and Northern blot analysis as described previously (8) were followed with modifications. See SI Appendix for more details.

ERG.

ERGs were carried out using a portable hmsERG machine (OccuScience). For scotopic ERG, we followed a protocol similar to that we described previously (110), with modifications. For photopic ERGs, a protocol similar to that previously described (111) was followed. See SI Appendix for more details.

Preparation of Cryoprotected Sections of Adult and Embryonic Mouse Retina.

A protocol similar to that described previously was followed (108). To collect embryonic retinas, we checked vaginal plugs in the females in the breeding cages every morning. When a female was plugged, we separated her from the male, recorded the date of plugging, and registered the predicted age of the embryo as 0.5 d postconception (d.p.c). We harvested the eyes at E14.5 and E16.6 and at 18.5 d.p.c. For postnatal age, we designated the day of birth as P1.

H&amp;E Staining and Measurement of Retinal Layers.

Standard H&amp;E staining was performed as described previously (108). Retinal thickness was measured using NIS-Elements BR software on a Nikon Eclipse 80i microscope. Moving outwards from the optic nerve, measurements of the OS, ONL, OPL, INL, and IPL were taken every 500 μm on live color images; the number of nuclear layers across the ONL also was counted. Statistical analysis was performed using two-tailed t tests.

X-Gal Staining.

X-Gal staining was performed as described previously (108). See SI Appendix for more details.

Antibodies, IF, and PNA Staining.

Antibodies against rhodopsin (Ab5417; AbCam), M opsin (AB5405; Millipore), S-opsin (AB5407; Millipore), Ribeye/CtBP2 (612044; BD Bioscience), PSD95 (75-028; Antibody Inc.), and synaptophysin (MAB5258; Millipore) were used as primary antibodies. Alexa Fluor-conjugated secondary antibodies were purchased from Invitrogen. IF was performed as described previously (112). Fluorescein and rhodamine-conjugated PNA was purchased from Vector Laboratories. For PNA staining on retinal sections, PNA (200 μg/mL) was incubated with sections for 1 h at room temperature followed by a 10-min postfixation. For whole-mount retinal IF, a protocol similar to that previously described (113) was followed. For quantification of cones in whole-mount retina, a protocol previously described by Ng et al. (95) was followed. See SI Appendix for more details.

EM.

Eye cups were fixed in 3% paraformaldehyde, 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) for 2 h at room temperature and then were processed for conventional EM as previously described (114). See SI Appendix for more details.

Gene-Expression Profiling, Data Analysis, and Confirmation Study.

GeneChip Mouse Genome 430A_2 Array (Affymetrix) was used for gene-expression profiling. The probe labeling, GeneChip hybridization, and data collection were performed at the Core Facility of Research Resource Center at the University of Illinois, Chicago. See SI Appendix for more details.

Light-Damage Experiment.

Mice were dark adapted overnight before the light-damage experiment. Pupils were dilated using 1% tropicamide solution and 2.5% phenylephrine (Bausch &amp; Lomb) for 15 min. Then mice were exposed for 2 h to 10,000-lx cool, white fluorescent light in cages lined with aluminum foil. After light exposure mice were kept in the dark for 2 wk before being killed for preparation of retinal sections.

Supplementary Material

Author Summary

Previously, we identified a cluster of miRNAs—small, noncoding, regulatory RNAs—that appears to be evolutionarily conserved and specifically expressed in sensory organs (1). Expression of this miR-183/96/182 cluster (also referred to as “miR-183/96/182”) is minimal in embryonic retina but increases dramatically after birth and peaks in adult retina with specific expression in photoreceptors (1). Reducing the expression of miR-183/96/182 in rod photoreceptors increased light damage to the mouse retina (2), but no histological or functional defects of the retina were observed under normal lighting conditions (2). Thus, in vivo functions of miR-183/96/182 in the retina remain uncertain.

We first dissected the genomic structure of the gene encoding miR-183/96/182 and showed that the miR-183/96/182 gene produces at least two primary RNA molecules, the transcripts (Fig. P1A). The structure of the second transcript clearly positions miR-183/96/182 in the intron of the gene. The second transcript has four predicted protein-coding regions, suggesting that the miR-183/96/182 gene may encode one or more peptides.

Fig. P1.

(A) Genomic structures of the miR-183/96/182 gene. The dotted lines symbolize alternative splicing events. TSS, transcription start site. Ex, exon; pA, polyA signal. The numbers above the double-headed arrows denote sizes (in bp). (B and C) Functional defects detected by dark-adapted (B) and light-adapted (C) ERGs in 5-wk-old mutant (mut) mice. In B, blue, black, red, green, and violet lines represent recordings at 0, 10, 100, 1 k, and 10 k millicandela⋅s⋅m^{, respectively. (}D) Retinal degeneration in 1-y-old mutant mice. (E) Hypothetical model of in vivo functions of miR-183/96/182 in the retina. When miR-183/96/182 is inactivated, homeostasis of photoreceptors is disrupted, resulting in defects in photoreceptors and their synaptic transmission, leading to retinal dysfunction and degeneration.

To study the functions of miR-183/96/182, we generated a mouse model in which the miR-183/96/182 gene is inactivated. Functionally, inactivation of the miR-183/96/182 gene resulted in early-onset, progressive electroretinogram (ERG) abnormalities. The most prominent defect, decreased b-wave amplitude, was common to both dark-adapted and light-adapted ERGs in the mutant mice (Fig. P1 B and C) and was detectable as early as 5 wk of age. The ERG b wave is derived mostly from activities of secondary neuron-bipolar cells; defects in bipolar cells and photoreceptor synaptic transmission can produce defective b waves (3). Because miR-183/96/182 is expressed predominantly in photoreceptors, reduction in b-wave amplitude likely reflects synaptic transmission defects in photoreceptors of miR-183/96/182–mutant mice. The amplitude, latency, and implicit time of the a wave, which reflects the activities of photoreceptors, appeared normal in 5-wk-old mutant mice but became defective with age. In all parameters, light-adapted ERG was more affected than dark-adapted ERG, suggesting more severe defects in the cone pathways.

Histologically, despite significant ERG abnormalities, retinas of 5-wk-old mutant mice were indistinguishable from littermate controls. By 6 mo, however, the mutant mice showed significant retinal degeneration, which progressed at 1 y of age (Fig. P1D). Intriguingly, the superior retina, which is enriched with red/green cones, was more affected than the inferior retina, which is enriched with blue cones.

A light-damage experiment showed significantly increased degeneration in the retinas of the mutant mice as compared with wild-type littermates after a 2-h exposure to 10,000-lx cool, white fluorescent light. In this experiment, as with aging, the superior retina was more damaged than the inferior retina, consistent with a recent report in a miR-183/96/182 knockdown mouse model (2).

Gene-expression profiling revealed global gene-expression changes in retinas of 5-wk-old miR-183/96/182–mutant mice. Genes important for synaptogenesis, nerve-impulse transmission, and synaptic remodeling were enriched significantly in the dysregulated genes, consistent with ERG data indicating abnormal synaptic transmission of photoreceptors. Electron microscopy and immunostaining with synaptic markers confirmed that the sizes of synaptic ribbons were decreased significantly in the mutant mice as early as age 5 wk.

Many genes known to be involved in phototransduction, a process by which light signals are transformed into neuroelectrical impulses in photoreceptors, also were dysregulated. We confirmed that the expression of red/green opsin was decreased significantly. Quantification of red/green opsin-positive and total cones in the whole-mount retina showed that both the number and the percentage of red/green opsin-positive cones were decreased significantly in the superior retina, but in the inferior retina the total number of cones and the number of red/green opsin-positive cones did not change significantly, suggesting that red/green cones were the most affected photoreceptors and that the greater vulnerability of red/green cones contributed to more severe defects in photopic ERGs and more degeneration in the superior retina in miR-183/96/182–mutant mice.

In mouse, both rod and cone photoreceptor differentiation and synaptogenesis occur postnatally starting from postnatal day 4–5 (4), coincident with developmental up-regulation of the miR-183/96/182 gene (1). Our data suggest that miR-183/96/182 is required for postnatal differentiation of photoreceptors and establishment of fully functional synaptic connections with their downstream neurons by directly and indirectly regulating genes important for synaptic functions and the phototransduction pathway. Inactivation of the miR-183/96/182 gene disrupts the homeostasis and functions of photoreceptors, leading to progressive retinal dysfunction and degeneration (Fig. P1E).

In addition to retinal defects, miR-183/96/182–mutant mice exhibited a circling behavior typical of defects in vestibular organs, suggesting the presence of syndromic sensory defects and that the requirement for the miR-183/96/182 gene for normal functions of sensory organs other than the retina. Studies on genetic variations around miR-183/96/182 gene are warranted in patients with progressive, sensory syndromic retinal dysfunction and other neurological conditions.