A MicroRNA, mir133b, suppresses melanopsin expression mediated by failure dopaminergic amacrine...

Cellular Signalling 24 (2012) 685–698

Contents lists available at SciVerse ScienceDirect

Cellular Signalling

j ourna l homepage: www.e lsev ie r .com/ locate /ce l l s ig

A MicroRNA, mir133b, suppresses melanopsin expression mediated by failuredopaminergic amacrine cells in RCS rats☆

Yaochen Li 1, Chunshi Li 1, Zhongshan Chen, Jianrong He, Zui Tao, Zheng Qin Yin ⁎Southwest Hospital/Southwest Eye Hospital, Third Military Medical University, Chong Qing 400038, China

Abbreviations: RP, retinitis pigmentosa; RCS, Royal Cmelanopsin-expressing retinal ganglion cells; TH, tyrosinINL, inner nuclear layer; Mir-133b, MicroRNA-133b; DAcells; pitx3, paired-like homeodomain 3.☆ This work were supported by Nature Science FounInternational (Regional) Joint Research Project 3091010lies Mark) and a grant from the Key Program of NationalChina to Zheng Qin Yin (No. 81130017).⁎ Correspongding author at: Southwest Hospital/Sout

itary Medical University, Chong Qing 400038, China. Tel.fax: +86 23 65460711.

E-mail address: [email protected] (Z.Q. Yin).1 Both authors contributed equally to this work.

0898-6568/$ – see front matter © 2011 Elsevier Inc. Alldoi:10.1016/j.cellsig.2011.10.017

a b s t r a c t
a r t i c l e i n f o
Article history:Received 10 October 2011Accepted 28 October 2011Available online 9 November 2011

Keywords:MicroRNAmir133bPitx3mRGCsMelanopsinDopaminergic amacrine cellsTHD2 receptor

The photopigment melanopsin and melanopsin-containing RGCs (mRGCs or ipRGCs) represent a brand-newand exciting direction in the field of visual field. Although the melanopsin is much less sensitive to light andhas far less spatial resolution, mRGCs have the unique ability to project to brain areas by the retinohypotha-lamic tract (RHT) and communicate directly with the brain. Unfortunately, melanopsin presents lower ex-pression levels in many acute and chronic retinal diseases. The molecular mechanisms underlyingmelanopsin expression are not yet really understood. MicroRNAs play important roles in the control of devel-opment. Most importantly, the link of microRNA biology to a diverse set of cellular processes, ranging fromproliferation, apoptosis and malignant transformation to neuronal development and fate specification isemerging. We employed Royal College of Surgeon (RCS) rats as animal model to investigate the underlyingmolecular mechanism regulating melanopsin expression using a panel of miRNA by quantitative real-time re-verse transcription polymerase chain reaction. We identified a microRNA, mir133b, that is specificallyexpressed in retinal dopaminergic amacrine cells as well as markedly increased expression at early stage dur-ing retinal degeneration in RCS rats. The overexpression of mir133b downregulates the important transcrip-tion factor Pitx3 expression in dopaminergic amacrine cells in RCS rats retinas and makes amacrine cellsstratification deficit in IPL. Furthermore, deficient dopaminergic amacrine cells presented decreased TH ex-pression and dopamine production, which lead to a failure to direct mRGCs dendrite to stratify and enterINL and lead to the reduced correct connections between amacrine cells and mRGCs. Our study suggestedthat overexpression of mir133b and downregulated Pitx3 suppress maturation and function of dopaminergicamacrine cells, and overexpression of mir133b decreased TH and D2 receptor expression as well as dopamineproduction, which finally resulted in reduced melanopsin expression.

© 2011 Elsevier Inc. All rights reserved.

1. Introduction

Photoreception in the mammalian retina is not restricted to rodsand cones. Recent evidence has demonstrated the existence of athird type of mammalian photoreceptor that differs greatly fromrods and cones and utilizes a different photopigment, melanopsin

ollege of Surgeon rats; mRGCs,e hydroxylase; DA, dopamine;ACs, dopaminergic amacrine

dation of China Grant, Major3913 (to Zheng Qin Yin & Gil-Natural Science Foundation of

hwest Eye Hospital, Third Mil-: +86 13808336957 (mobile);

rights reserved.

[1]. Although the melanopsin is much less sensitive to light and hasfar less spatial resolution, these photoreceptors have the unique abil-ity to project to brain areas by the retinohypothalamic tract (RHT)and communicate directly with the brain because they are a subsetof retinal ganglion cells [2]. The primary role of mRGCs is to signallight for unconscious visual reflexes, such as pupillary constriction,regulating a number of daily behavioral and physiological rhythms,collectively called circadian rhythms [3,4]. Thus, those blind peoplewho keep eyes can still maintain their normal circadian rhythm. Itseems that understanding the reasons and mechanisms involved inregulating melanopsin expression is essential for successful applica-tion of the third photoreception to improve some retinal diseasethat causes blindness such as retinitis pigmentosa (RP) in the future.However, they have long been controversial issues that whether thenumber of mRGCs has changed and whether melanopsin can con-stantly express under pathological conditions such as dystrophicrats. Anthony A. Vugler et al. reported melanopsin to be robustlyexpressed in a population of ganglion cells in the inner retina ofboth non-dystrophic and dystrophic rats [5]. The opposite reportswere also found that degeneration of the classical photoreceptors

http://dx.doi.org/10.1016/j.cellsig.2011.10.017

mailto:[email protected]

http://dx.doi.org/10.1016/j.cellsig.2011.10.017

http://www.sciencedirect.com/science/journal/08986568

Table 2Antibodies used in immunofluorescence staining and western blot.

Antibody Company Con. Speci Cat# Purpose

Melanopsin Abcam 1:200 Mouse ab19306 IF&WBD2 receptor Abcam 1:200 Goat Ab32349 IF&WBPitx3 lifespan 1:100 Rabbit LS-B994 IF&WBCalretinin Millipore 1:200 Mouse Mab-

1568IF

GAPDH Cwbio, China 1:2000 Rabbit CW0101 Internalcontrol

β-actin Cwbio, China 1:1000 Mouse CW0096 Internalcontrol

Secondary antibodiesAlexa-568 (mouse)

Invitrogen 1:300 Goat A-11004 IF

Secondary antibodiesAlexa-488 (rabbit)


Secondary antibodiesAlexa-568 (rabbit)


Secondary antibodiesAlexa-488 (mouse)


Secondary antibodiesCy3 (mouse)

Beyotime,China

1:500 Goat A0521 IF

Secondary antibodiesCy3 (rabbit)

Beyotime,China

1:500 Goat A0516 IF

Table 1Designing QT-PCR primers.

Genes(rat)

Forward primer Reverse primer

OPN4 ATCTGGTGATCACACGTCCA TAGTCCCAGGAGCAGGATGTThy1.1 GAGGGCGACTACATGTGTGA AGGAAGGAGAGGGAAAGCAGD2receptor

CTGGTGTGCATGGCTGTATC CTTCCGGAGGACGATGTAGA

GAD65 CGCACTGCCAAACAACTCTA CAGGGGCGATCTCATAGGTAPitx3 TTCCCGTTCGCCTTCAACTCG GAGCTGGGCGGTGAGAATACAGGPACAP ACAGCGTCTCCTGTTCACCT CCTGTCGGCTGGGTAGTAAATH CAGGGCTGCTGTCTTCCTAC GGGCTGTCCAGTACGTCAATNurr1 ACACAGCGGGTCGGTTTACTACAA AAAGGAGAAGAGTGAAAGGCGGGAGAPDH AGACAGCCGCATCTTCTTGT TGATGGCAACAATGTCCACT

686 Y. Li et al. / Cellular Signalling 24 (2012) 685–698

(rods and cones) in RCS N ⁄–rdy rat retina induces a significant(>90%) reduction in the levels of melanopsin mRNA and protein[6]. Furthermore, the melanopsin expression is also found to be re-duced in acute photoreceptors degeneration [7].

Unfortunately, although there are many studies that have reportedthe melanopsin expression level in several of retinal diseases, few re-ports have looked at the molecular mechanism regulating melanopsinexpression. It was not until recently that the expression and regulationof melanopsin was briefly recognized [8]. Here, we will present ourstudies on effect ofmir133b and its target gene pitx3 on the dopaminer-gic amacrine cells and effect of disrupted dopaminergic amacrine cells(DA ACs) on the expression of melanopsin in RCS rats.

2. Materials and methods

2.1. Animals and reagent

RCS-p+ rats (retinal dystrophic, pigmented RCS rats) were used asthe test group (abbreviated as RCS) and RCS-rdy+-p+ rats (non-retinaldystrophic, pigmented RCS rats,abbreviated as rdy+) as the controlgroup. RCS rats at postnatal day 15(p15), p21, p30, p60 and p90, andthe age-matched rdy+ controls were housed in a temperature-controlled room on a 12-hour light–dark cycle in the Laboratory AnimalUnit of Third Military Medical University. All the experimental and ani-mal handling procedures complied with the Association for Research inVision and Ophthalmology Statement and were also reviewed andapproved by the Faculty Committee on the Use of Live Animals inTeaching and Research, The Third Military Medical University. N-Propylnorapomorphine HCl (NPA) was purchased from Sigma(Cat#18426-20-5).

2.2. Retrograde labeling techniques and Imaging and morphometricmeasures

Five rats for each strain were anesthetized i.p. a mixture of keta-mine (70 mg/kg) and xylazine (7 mg/kg) during the experiments. An-imals were subsequently killed with an overdose of urethane (1 g/kg).After anesthesia the rat was fixed in stereotaxic apparatus and the cor-responding positions for the superior colliculus (SC) and lateral genic-ulate nucleus (LGN) for labeling RGCs, suprachiasmatic nucleus (SCN)for labeling mRGCs were defined according to the rat stereotaxic atlasof Paxinos and Watson [9]. Trephine holes were bored through theskull with an electric drill and 2.5 μl Fluorogold (80014, Biotium Com-pany, USA) was injected slowly into each location using microinjector(Gaoge Company, Shanghai, with the tip diameter 10 μm). The syringeremained in place for a further 5 min to avoid dye leakage. After theinjections the skin and subcutaneous tissuewas sutured and antibioticeye ointment applied. After a survival period of 5–7 days the eyeballswere enucleated either for quantitative and morphological analysis asabove, or for whole-cell recording.

Confocal microscope pictures were taken from superior, inferior,nasal and temporal quadrants of retina, and each quadrant was medi-ally divided into three zones. The pictures were analyzed by using thesoftware Image Pro Plus. According to the results of Rivera and ourlaboratory [10,11] the cells in GCL with nucleus diameter above6 μm and appeared round and ellipsoidal were regarded as ganglioncells. Bacilliform, ramous and broken nucleus were excluded

2.3. Quantity real-tim PCR

After enucleation, the retinas were isolated and frozen in liquid ni-trogen for subsequent RT-PCR analysis. In brief, the total RNAswere iso-lated with Trizol reagent, according to the manufacturer's instructions.Contaminated DNAs were removed by using TURBO DNA-freeTM Kit(ABI) and cDNAs were synthesized from 2 μg of total RNA using Prime-ScriptTM RT reagent Kit (TAKARA) in a 20 μl of reaction mixture

following the manufacturer's instructions. Primers were designed on-line with IDT Scitools and the sequences were listed in Table 1. Real-time PCR was carried out in the Bio-Rad 5-Color System. To confirmthe specificity of PCR products, melting curves were determined usingiCycler software, and samples were run on an agarose gel. The expres-sion change of a target gene in RCS rats relative to the control rats wascalculated as: fold change=2−(ΔCT,Tg-ΔCT,control). The following PCRscheme was used: 5 min at 94 °C, (30 s at 94 °C, 30 s at 63 °C, 30 s at72 °C)×35, 10 min at 72 °C and then 4 °C thereafter.

2.4. Immunofluorescence staining and western blot

The rats were anesthetized with intraperitoneal urethane andtranscardially perfused with 4% PFA. Eyes were enucleated. The eye-cups were immersed in 4% paraformaldehyde in 0.1 M phosphatebuffer (PB) at pH 7.4 at 4 °C for 2 h, transferred to 10%, then 30% su-crose in 0.1 M PB, and stored at 4 °C overnight. Next day the eyecupswere embedded in optimum cutting temperature (O.C.T.) compoundand cut into 10 μm thick radial sections on a cryostat (Leica, Germa-ny) at−20 °C, andmounted on glass slides. The sections were treatedfor 5 min at room temperature with 0.5% Triton X-100 followed by 5%goat serum in PBS for 1 h without washing, they were then immuno-stained for Melanopsin, Pitx3, TH, D2 receptor and Calretinin (Shownin Table 2). Sections were incubated overnight at 4 °C. Next day sec-tions were washed three times (15 min each) with PBS at room tem-perature, incubated for 2 h at room temperature in the dark insecondary antibodies, either goat anti-rabbit IgG-cy3 (Beyotime,

Melanopsin expression level

0

0.2

0.4

0.6

0.8

1

1.2

Rat

io o

f m

elan

opsi

n to

GA

PDH p+

rdy+

Comparison of mRNA level in retina between RCS and control rats

0

1

2

3

4

5

6

7

Rel

ativ

e fo

ld c

hang

es Controlp+ PACAPp+ Thy1.1p+ Melanopsin

A

C

BRCS, P+ Rdy+

Melanopsin

GAPDH

P15 p30 p60 p90 P15 p30 p60 p90

p15 p30 p60 p90

p15 p30 p60 p90

Fig. 1. The expression levels of melanopsin diminished tremendously as a result of retinaldegeneration in RCS rats. (A), Quantitative reverse transcriptase-polymerase chain reac-tion (qRT-PCR) show that both TH and melanopsin mRNA level are decreasing with pro-gressing retinal degeneration in RCS rats as compared to the control retinas. PACAPmRNA level does not change as a result of progressing retinal degeneration in RCS ratsas compared to the control retinas. The statistical analysis is made from three individualtimes real-time PCR. Data are presented asmean values±SEM; (B), Western blot showedthat reduced melanopsin expression levels in RCS rat retinas. Notably, RCS at P60, mela-nopsinmRNA levelswere greatly reduced (90%), andmRNA and protein levelswere near-ly undetectable until at P90. In contrast, themelanopsin protein levels of control ratswereincreased along with the growth of the age. (C), The relative protein expression of mela-nopsin to GAPDH. The values shown represent mean±standard error from threereplicates.

687Y. Li et al. / Cellular Signalling 24 (2012) 685–698

China) or goat anti-mouse IgG-cy3 (Beyotime, China). All secondaryantibodies were used at a 1:500 dilution in 0.01 M PBS. After incuba-tion for 5 min at room temperature with DAPI (Beyotime, China), thesections were washed in 0.01 M PBS and then coverslipped withwater soluble mounting liquid (Boster, China) and examined in thefluorescence microscope or confocal microscope. The positive areawas measured from confocal images of sections antibodies withimage-analysis software Image-Pro Plus 6.0 (IPP 6.0).

For the detection of the protein expression levels of Melanopsin,pitx3, Thy1.1 and D2 receptor, the RCS rats' retinas in each groupwere homogenized in ice-cold RIPA Lysis buffer (50 mM of Tris–HClbuffer (pH 7.4), 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycho-late, 0.1% SDS, sodium orthovanadate, sodium fluoride, EDTA, andleupeptin). The homogenate was then centrifuged at 12,000×g for5 min at 4 °C. The clear supernatants were stored at −80 °C untiluse. Protein concentration was determined with bicinchoninic acidKit (Beyotime Institute of Biotechnology, China). The samples (50 μgof protein/lane) were loaded and electrophoresed on a 12% SDS-polyacrylamide gel (SDS-PAGE) for 40 min at 120 V. The proteinswere transferred from the gel onto a nitrocellulose (NC) membranefor 70 min at 120 V. After transferring, the NC membranes wereblocked with block solution containg Tris-buffered saline, 0.1%Tween 20 (TBST) and 5% free-fat milk for 1 h at room temperature.The blots were then washed three times (5 min each time) withTBST, and then incubated with primary antibodies (Shown inTable 2) overnight at 4 °C. After washing, all membranes were againincubated with GAPDH antibody for 1 h at room temperature. After-wards, all membranes were incubated in different secondary anti-bodies by turns for 1 h at room temperature while shaking. Betweenincubations, the membranes were rinsed. Finally, the NC membraneswere scanned using the Odyssey infrared imaging system with theOdyssey Application software V1.2.15 for melanopsin, pitx3, Thy1.1,D2 receptor and GAPDH bands. The ratio of the densitometry wasobtained to semi-quantify the relative level of melanopsin, pitx3,Thy1.1, D2 receptor.

2.5. Delivery of D2 agonist with intravitreal injection

To exclude the possible function failure of mRGCs to express mel-anopsin, the involvement of dopamine and D2 receptors pathwaywas assessed by administering the agonist R(−)-propylnorapomor-phine hydrochloride (NPA, Sigma). The ligands were dissolved insterile water. Concentrations used 5 μg/μl in the injection volume of5 μl. The solution had a pH of ~7. Injections were made in the vitreoushumor. After 3 d of injection, the eyes were enucleated and sampleswere made and used in qRT-PCR and western blot analysis.

2.6. MicroRNA panel

For microRNA panel assay, the eyes of RCS and rdy+rats at differ-ent age were harvested and were extracted using miRCURY RNA Iso-lation kit (Exiqon, product# 300111). For purifying microRNAs,contaminated DNA was removed. The cDNA was synthesized usingmiRCURY LNA™ Universal RT cDNA synthesis kit (Exiqon, product#203300). miRCURY LNA™ Universal RT microRNA PCR, Rat panel Iand miRCURY LNA™ SYBR® Green master mix were used in Micro-RNA panel assay through ABI 7900HT machine. All manipulationswere performed with the recommended by manufacturer.

2.7. In situ hybridization

MirRCURY LNA mir133b RNA detection probes, rno-mir-133b(Prod. No. 38579-15), were labeled with 5′- and 3′-digoxigenin andprovided by EXIQON. Retinal sections from both RCS rats' retinasand control rat retinas were processed in parallel and incubated over-night at 60 °C with either antisense or sense probes. Hybridization

signal was detected using an anti-digoxigenin antibody, 5-bromo-4-chloroindolyl-phosphatase and nitroblue tetrazolium (Roche AppliedBiosciences).

2.8. Delivery of Lentiviru-shRNA/mir133b

Rno-mir-133b (Gene ID MIMAT0003126) interference Lentiviruswere made by designing the nucleotide sequence (TAGCTGGTT-GAAGGGGACCAAA) blocking mir-133b site and constructing pMagic4.1 expression vectors (sbo-bio) with hU6 promotor and EGFP mark-er. Meanwhile, the nucleotide sequence (TTCTCCGAACGTGTCACGT)was used to be negative control. The virus packaging plasmids (pMa-gic 4.1-mir133b-RNAi, pHelper1.0 and pHelper2.0) were cotrans-fected into 293 T cells. The Lentivirus-mir-133b RNAi was harvestedfrom 293 T cells. The integration unit was 1.05E+09 per ml. TheLentivirus-mir-133b RNAi were used to infect interneurons by

A B

C

020406080

100120140160

Num

ber

of m

RG

C (

/mm

2 ) ControlRCS

P21d P60d P90d

Fig. 2. Morphology of rat mRGCs at different postnatal ages. In panel A, (A) control rat (CTR) retinal whole-mounts are stained at p21, p60 and p90withmelanopsin (green) and PI (red).(B) RCS retinal whole-mounts are stained withmelanopsin (green) and PI (red) at the same ages as controls. Melanopsin positive signals could be seen in axons, dendritic processes andthe perinuclear. (C) Vertical sections through the retina of a control p90 retina stained with melanopsin (green) and PI (red) shows the position of an mRGC cell within the GCL. (D) Avertical section through the retina of a RCS p90 retina shows the ramification of cell processes and soma position of anmRGCwithin the GCL. GCL: ganglion cell layer; IPL: inner plexiformlayer; INL: inner nuclear layer; PRC: photoreceptor cells. Scale bar: 10 μm. In panel B, distribution and morphology of retrograde-labeled mRGCs. (A) Labeled RGCs are observed afterFluorogold injection in LGN and SC of a p21 RCS rat, and the amount and distribution of RGCs appears approximately normal at this stage. (B) In contrast, the distribution of cells labeledafter SCN injection appears sparse, but also distribute across the whole retina. (CE) The soma and processes of labeled cells after an SCN injection of flurogold are shown at low and highmagnification from a control rat. No extensive dendritic arbors are shown clearly and no obvious connections are observed with other cells. (DF) Labeled cells after an SCN injection offlurogold are shown at low and high magnification from a p90 RCS rat. Note that even thought there is considerable loss of RGCs these cells still maintain dendritic processes similarto that seen in control retinas. In panel C, the density and ratio of mRGCs/RGCs in RCS and control rat retinas. (A),The density of RGCs (/mm2) in RCS progressively decreased fromp21 to p90 compared to control rats (*: Pb0.05, **: Pb0.01 vs Control). (B),The density of mRGCs (/mm2) in RCS and control rats remain relatively stable at the ages sampled and arenot significantly different. (C),The ratio of mRGCs to RGCs significantly increase in RCS rats in line with the changes in RGC density. (*: Pb0.05, **: Pb0.01 vs Control).


Table 4The number change of mRGCs during degeneration in RCS rats retina by retrograde la-beling with fluorogold.

P21d P60d P90d

Control 129±12.5 131±8.1 128±7.4RCS 141±11 134±16 136±13

No significant difference between control and RCS rats retina.


intravitreal administration with a volume of 5 μl. After 7 d, 14 d and30 d of injection, the eyes were enucleated and samples were madeand used in frozen sections, qRT-PCR and western blot analysis.

2.9. Statistics

All data are presented as mean and standard error of the mean(mean±SEM). Data were evaluated by one-way ANOVA and LSD-qtest with post hoc statistical analysis in the same group, and t-testsfor comparing means at the same time point between the experimentand control group. Values of pb0.05 were considered to be significant.

3. Results

3.1. Decreased expression of melanopsin in RCS rats

Melanopsin is a photopigment found in about 2–3 percent of RGCsin normal rats. Our results showed that gradually reducedmelanopsinexpression levels were found by quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) and western blot during retinaldegeneration in RCS rats. Notably, melanopsin mRNA level in RCS ratretinas at P15 presented a higher level as compared to control rat ret-inas. However, melanopsin mRNA levels were greatly reduced (90%)at p30, and thereaftermelanopsinmRNAmaintained very lower levelsin RCS rat retinas. Untill P90, melanopsin were nearly undetectable(Fig. 1A). A similar expression pattern of melanopsin protein, progres-sing decrease with retinal degeneration, was observed in RCS rat ret-inas by western blotting. In contrast, the melanopsin protein levelsof control rats were increased along with the growth of the age(Fig. 1B&C). These data showed that melanopsin expression levelwas indeedly diminished in chronic retina degenerative diseases, atleast in RCS rat retinas.

3.2. Density and distribution changes ofmRGCs during retinal degenerationin RCS rats

Normally, a subset of retinal ganglion cells (RGCs) that send pro-jections to the SCN and express melanopsin is intrinsically photosen-sitive (ipRGCs or mRGCs). These cells distributed across the entireretina although cell density appeared higher in temporal quadrantin the whole-mount retina by melanopsin staining. To ascertainwhether the density and distribution of mRGCs changed in RCS ratsand to exclude possible effect of mRGCs number loss on melanopsinexpression level, the density and distribution of mRGCs during retinaldegeneration in RCS and control rats were detected respectively byFluoro-Gold Retrograde labeling techniques combined immunofluo-rescence staining. The density of RGCs and mRGCs in the GCL was cal-culated with software Image Pro Plus. With the immunofluorescencestaining of retinas and retrograde labeling with fluorogold (Fig. 2,panel A & panel B), the results showed that there were significant dif-ference on total RGC numbers between RCS and control rats. TheRGCs numbers of RCS retina presented progressively decrease, theywere 5673±171/mm2, 4263±224/mm2 and 3060±271/mm2 atp21, p60 and p90(Table 3) respectively. The differences betweenRCS and control retinas, especially at p60 and p90, were significant(Fig. 2, panel C). What is in difference with total RGC is that the den-sity of mRGCs remained relatively constant after p21, p60 and until

Table 3The number change of RGCs during degeneration in RCS rats retina.

P21d P60d P90d

Control 5735±227.3 5529±183.2 5413±206.8RCS 5673±170.9 4263±223.9⁎ 3060±271.2⁎⁎

⁎ Pb0.05 vs control.⁎⁎ Pb0.01 vs control.

p90 (141±11, 134±16 and 136±13/mm2 respectively) (Table 4)in RCS rats,and no significant difference was seen between RCS ratsand control rats (Fig. 2, panel C). Furthermore, the ratio of mRGCs/RGCs was about 2.5% in RCS and control rats at p21, but progressiveand significant increase trends were found at p60 and p90, theyreached 3.2% and 4.5% when compared with control rats (Fig. 2,panel C). At late stages of retinal degeneration, the retrograde trans-port of Fluorogold from the SC and LGN confirmed the decreased den-sity of RGCs and consistent density of mRGCs(Fig. 2, panel B). Inaddition, real time PCR assay that Thy1.1 (a general marker forRGCs) mRNA levels were significantly reduced supported total RGCnumbers progressively decrease during retinal degeneration in RCSrats. Previous studies have pointed that PACAP and melanopsinwere colocalized in certain RGCs and PACAP-containing nerve fibers[12,13]. In this study, PACAP mRNA levels were also shown to be sig-nificantly increase at the whole degeneration process in RCS rat ret-inas (Fig. 1C). Taken together, the numbers of mRGCs were normalin RCS rat retinas. Consequently, the down-regulated melanopsin ex-pression level should be due to the expression ability of mRGCs, butnot number variation of mRGCs in RCS rat retinas.

3.3. Decreased melanopsin expression should be due to the failure of do-paminergic amacrine cells in RCS rats

As well known, the PACAP is negatively regulated by dopamine D(2) receptors [14]. Therefore, in this study, D2 receptors expressionlevels were determined through qRT-PCR and immunofluorescencestaining analysis. Moreover, TH (Tyrosine hydroxylase, a marker forDA ACs) was used to evaluate the DA ACs status. The qRT-PCR showedthat both TH and D2 receptor mRNA levels were close to normallevels at p15, but largely decreased from P30 and thereafter (Fig. 3I)in RCS rat retinas. We also detected Gad65 markers for GABAergicamacrine cells and didn't find reduction of GABAergic amacrine cellsin RCS rat retinas (Fig. 3 I). Fig. 3. A~H showed the immnuofluores-cence staining against D2 receptor of RCS and control rats' retinas re-spectively at different time points. D2 receptor mainly expressed onmembranes of DA ACs in INL beneath the inner limiting membrane(white arrows in Fig. 3) and ganglion cells in the ganglion cell layer(GCL), and also presented on axons in OPL (outer plexiform layer)and IPL. The positive signals of D2 receptor in all four layers abovewere slightly reduced in RCS rat retinas as compared to control ret-inas at both p15 (Fig. 3. A&B) and markedly reduced in RCS rat retinasat p30 (Fig. 3. C&D), p60 (Fig. 3. E&F) and p90 (Fig. 3. G&H). Wethought that the reduction of D2 receptors were due to thedisrupted DA ACs during degeneration in RCS rat retinas. When D2agonist R(−)-N-Propylnorapomorphine (NPA) was delivered byintravitreal injection, Nurr1, melanopsin and D2 receptor mRNA ex-pression levels were upregulated when compared with those retinastreated with PBS(Fig. 3K). These studies taken together suggestedthat mRGCs remained the ability to express melanopsin during retinaldegeneration in RCS rat retina, but lower melanopsin level in RCS ratretinas should be due to the failure of DA ACs.

3.4. Overexpressed mir133b and diminished pitx3 in RCS rat retinas

To explore all possible reasons for function failure of DA ACs dur-ing retinal degeneration, MicroRNA panel assay was carried out.


Notably, the results showed that mir133b was enriched and upregu-lated more than 3 times at initiation of retinal degeneration, p15, inretina of RCS rats when compared with control retinas (Fig. 4A). Asa down stream molecule of mir133b, Pitx3 is an important transcrip-tion factor that activates genes involved in the development and dif-ferentiation of dopaminergic neurons. Western blot showed Pitx3

A

rdy, p15, X400 D2(green)

C

rdy, p30, X400 D2(green)

E

rdy+, p60, X400 D2(green)

G HH

rdy+, p90, X400 D2(green)

Fig. 3. The expression levels of TH and D2 receptors diminished tremendously during degenof RCS and control rats' retinas respectively at different time points. D2 receptors mainly exbrane and ganglion cells in the ganglion cell layer (GCL), and also presented on axons in OPabove are slightly reduced in RCS rat retinas as compared to control retinas at both p15(A&Bthe qRT-PCR shows that both TH mRNA level is close to normal levels at p15, but largely deMeanwhile, D2 receptor mRNA expression levels are diminished at all selected time pointsdoes not change as a result of progressing retinal degeneration in RCS rats as compared to thPCR. Data are presented as mean values±SEM; (J), a possible interaction model between mRcontained TH produce and release DA. The released DA may exert important roles through DR(−)-N-Propylnorapomorphine (NPA) was delivered by intravitreal injection, Nurr1, melanthose retinas treated with PBS. The statistical analysis is made from three individual times

expression was decreased at p15 and slightly restored at p30 in RCSrat retinas compared with that in control rat retinas. Subsequently,pitx3 expression was decreased with progressing retinal degenera-tion, at p60 and p90 (Fig. 4C&D). These results suggested that retinaldegeneration was accompanied by increased mir133b and dimin-ished pitx3 in RCS rat retinas.

B

P+, p15, X400 D2(green)

D

P+, p30, X400 D2(green)

F

P+, p60, X400 D2(green)

P+, p90, X400 D2(green)

eration in RCS rats. (A~H), show the immnuofluorescence staining against D2 receptorpress on membranes of DA ACs (white arrows) in INL beneath the inner limiting mem-L (outer plexiform layer) and IPL. The positive signals of D2 receptors in all four layers), and markedly reduced in RCS rat retinas at p30(C&D), p60 (E&F) and p90 (G&H). (I),crease from P30 and thereafter in RCS rat retinas when compared with control retinas.including p15, p30, p60 and p60. GAD65 mRNA level, a marker of GABAergic neurons,e control retinas. The statistical analysis is made from three individual times real-timeGCs and dopaminergic amacrine (DA) cells in retina. Dopaminergic amacrine (DA) cells2 receptors on mRGCs such as regulating melanopsin expression. (K), When D2 agonistopsin and D2 receptor mRNA expression levels are upregulated when compared withnumber counting. Data are presented as mean values±SEM.

Comparison of mRNA level in retina between control and RCS rats

0

0.5

1

1.5

2

2.5

3

3.5

4

Rel

ativ

e fo

ld c

hang

es

IControlp+ THp+ D2p+ GAD65

mRGC

J

K

0

0.5

1

1.5

2

2.5

3

Rel

ativ

e fo

ld c

hang

es

NaCl- IV 3dNPA-IV 3d

p15 p30 p60 p90

Thy1.1 GAD65 OPN4 Nurr1 D2

Fig. 3 (continued).


In order to test the effect of overexpressed mir-133b and de-creased pitx3 on DA ACs, the localization of microRNA 133b in retinaswas investigated in greater detail by in situ hybridization. Fixed-frozen vertical sections revealed the positive signals of mir133bwere observed in some cells with somas proximal to the IPL of retinas.To further determine the mir133b expression was within in cyto-plasm of DA ACs, the immunofluorescence staining against TH andfluorescent mir133b in situ hybridization were performed simulta-neously because DA ACs are main cells releasing dopamine in retina.We noted that the intensity of mir133b signals in RCS rat retinaswas stronger than that in control. We also observed TH positive sig-nals mainly express in some somatic proximal to the IPL and a de-crease in the TH-IR plexus throughout the S1 IPL sublamina. For theintensity of TH, immnuofluorescence staining results showed thatthere was not big difference between RCS and control rats beginningat early stage of retinal degeneration, for example at p15. And there-after, the very faint TH positive signals were observed in RCS rat ret-inas as compared to controls. The results showed that TH wasdramatically reduced in RCS rat retinas, which suggested that DAACs were disrupted and resulted in a diminished TH and dopamineproduction during retinal degeneration. The two positive signals,mir133b and TH, were co-localized in DA ACs of RCS rat retinas.According to the results of combined IF and fluorescent ISH, the stron-ger mir133b positive signals always accompanied the weaker TH pos-itive signals in RCS rat retinas. In contrast, very faint positive signalsof mir133b accompanied stronger TH positive signals were observedin control retinas (Fig. 4E). The results suggested that mir133b andTH positive signals were co-localized in dopaminergic amacrinecells, and overexpressed mir133b down-regulated TH expressionlevel in degenerative retinas. In addition, the double-labeling tech-niques were used with combine antibodies such as TH/pitx3(Fig. 5).The immunofluorescence staining of pitx3 showed that thepitx3 positive cells were located in INL and GCL of retina. Comparedwith control retinas, the pitx3 positive signals were weaker in eitherINL or IPL in RCS rat retinas at every time points we selected duringprogressing retinal degeneration (control and RCS retinas at p15

were shown in Fig. 5A&B, p30 were shown in Fig. 5C&D, p60 wereshown in Fig. 5E&F and p90 were shown in Fig. 5G&H). Importantly,the pitx3 positive signals and TH positive signals were co-localizedin DA ACs of RCS rat retinas. Taken together, just like dopamine neu-rons in midbrain, these data suggested that there might be a negativefeedback relationship between mir133b and pitx3 in DA ACs in retinaand that they involve in the development and function regulation ofDA ACs.

3.5. Downregulated pitx3 lead to dysfunction of dopaminergic amacrinecells

Previous studies have reported that mir-133b and pitx3 involve indevelopment and differentiation of dopaminergic neurons. Herein,based on calretinin and melanopsin staining, we used division ofthe IPL into five parallel sublaminae (S1–S5, S5 being closest to theganglion cell layer) for development and differentiation analysis inthis study as previously described [15,16,17] (Fig. 6). The tyrosinehydroxylase-positive DA ACs were predominantly stratify in the S1sublamina of the IPL and the localization of calretinin interneuronsprocesses is not disrupted in wild-type retinas. Furthermore, thecalretinin-positive amacrine somata were closed contacted or sur-rounded with melanopsin positive dendrites and calretinin positiveinterneurons form a discrete plexus with melanopsin cells. In anotherword, a close interaction patterns between DA ACs and mRGCs arbor-ization was observed in control retinas. In contrast, we identified de-fects in the stereotypic lamina-specific neurite arborization of DA ACsduring degeneration in RCS rats (arrowed) by con-focal microscope.Also, very different with control retinas, the melanopsin positiveneurite arborization presented uniformly discrete patterns in RCSrat retinas (light blue arrows Fig. 6) at p60 and p90. In addition,based on melanopsin staining, five parallel sublaminae were also di-vided in IPL for analysis. The chaotic sublaminae were noted at p15and p30 in RCS rats. Furthermore, the melanopsin positive signals ofIPL were very faint. The melanopsin positive neurite arborizationswere shorter and didn't reach into INL. Taken together, the data


reflected that downregulated pitx3 lead to developmental deficiencyof DA ACs.

3.6. Mir133b interference, rno-mir-133b RNAi in lentiviral rescuedmelanopsin expression in RCS rat retinas

Considering that overexpressed mir133b and downregulated pitx3can lead to deficient dopaminergic amacrine cells, the lentivirues ofRNA interfering against rno-mir-133bwere delivered by intravitreal in-jection. Meanwhile, PBS and Sham/RNAi lentivirus were used to be ascontrol. After 1w, 2w and 4w, the retina tissues were harvested anddetected. Green fluorescence proteins were significantly observedunder the fluorescence microscopes in mir133b/RNAi (Fig.7G) andSham/RNAi lentivirus (Fig. 7 F2) delivering groups as compared to thePBS delivering group (Fig. 7 F1). The result suggested that mir133b/RNAi lentivirues was delivered successfully by intravitreal injection.Real time PCR showed that both OPN4 and TH mRNA expression levelwere upregulated in those retinas treatedwithmir133b/RNAi lentivirus

dopaminergic neurons ↓

Rai

to o

f pi

tx3

to a

ctin

Dmir133b ↑

Pitx3↓

—+

B

Comparison of mir133b expression level duringretinal degeneration between RCS and control rats

00.5

11.5

22.5

33.5

44.5

p15 p30

Rel

ativ

e fo

ld c

hang

es

A C

rdy+p+

rdy+, p15, X400,mir133b(greem), TH(red)E

Fig. 4. The expression changes of mir133b and its target molecule pitx3 during retinal degendiminished pitx3 expression in INL of RCS rat retinas. (A), MicroRNA panel assay is employmir133b was enriched and upregulated more than 3 times at initiation of retinal degenermir133b expression level gradually falls back at p30. The values show represent mean±mir133b and pitx3. The increased mir-133b level and decreased Pitx3 level, a transcriptionated protein network. The similar relationships are putatively existed in dopaminergic amlevels during retinal degeneration. Pitx3 expression starts to decrease at p15 in RCS rat rewas decreased with progressing retinal degeneration, at p60 and p90. (D), The relative perror from three replicates. (E~L) show ISH using LNA mir133b probe. The mir133b hybinner nuclear layer (INL) at the border of the inner plexiform layer (IPL),or within the GCLupregulated in RCS-p+rat retinas at p15(E&F), p30(G&H), p60(I&J) and p90(K&L). To furthstaining against TH and ISH are combined to perform. TH-positive soma situated in the INL sproximal borders of the IPL in a vertical section. Especially, the mir133b signals and TH positi+rat retinas as compared with control retinas.

as compared to that treated with control virus (Fig. 7B). Western blotrevealed that protein expression levels of pitx3 (Fig. 7.A&E) and TH(Fig. 7.A&C)in retinas of RCS rats treatedwith rno-mir-133b/RNAi lenti-viral were markedly restored as compared to the control lentiviral-injected from 7 days to 30 days after injection. Furthermore, melanop-sin expression levels were markedly increased in retinas of RCS rats inagainst rno-mir-133b lentiviral-injected groups as compared to thecontrol lentiviral-injected and the noninjected groups (Fig. 7.A&D). Im-munofluorescence staining analysis showed that the histological devel-opment of calretinin positive neurite arborization was normalizedgradually at 7 days and 14 days after mir-133b was interfered by intra-vitreal lentivirus administration. We noted that very clear stereotypiclamina-specific neurite arborization of dopaminergic amacrine cellswere observed in retinal degenerative RCS rats delivered rno-mir-133b RNAi lentivirus bymicroscope (Fig. 7.P&Q). Real-time PCR re-sults showed that mRNA expression levels of TH, D2 receptor and mela-nopsin in retinas of RCS rats in rno-mir-133b/RNAi lentiviral-injectedgroups were higher than that in retinas of RCS rats in control RNAi

Comparison of pitx3 expression level

00.5

11.5

22.5

33.5

4

p15 p30 p60 p90

rdyp+

rdy +, Control P+, RCS

P15 p30 p60 90 P15 p30 p60 p90

Pitx3

GAPDH

p+, p15, X400,mir133b(greem), TH(red)F

eration in RCS rats. Overexpression of mir-133b at early stage in RCS rat retinas leads toed to detect mir 133b expression level. The relative quantitative analysis showed thatation, p15, in retina of RCS rats when compared with control retinas. Thereafter, thestandard error from three replicates. (B), a negative feedback circuit model betweenfactor that activates genes involved in dopaminergic differentiation, formed an associ-acrine cells in retinas. (C),Western blot showed that gradual reduced Pitx3 expressiontinas when compared with that in control rat retinas. Subsequently, pitx3 expressionrotein expression of pitx3 to GAPDH. The values shown represent mean±standard

ridization signals (green color) located either in large somatas (white arrows) in the. Compared with control retinas, mir133b positive hybridization signals were sharplyer determine mir133b locate in dopaminergic amacrine neurons, immunofluoresencetratifies at the border between the INL and the IPL, and the TH-positive processes are inve somata strictly co-localize to INL. Obviously, TH-positive signals are weaker in RCS-p

rdy+, p30, X400,mir133b(greem), TH(red) p+, p30, X400,mir133b(greem), TH(red) HG

rdy+, p60, X400,mir133b(greem), TH(red) p+, p60, X400,mir133b(greem), TH(red)

rdy+, p90, X400,mir133b(greem), TH(red) p+, p90, X400,mir133b(greem), TH(red) LK

JI

Fig. 4 (continued).


lentiviral-injected groups. These results showed that a block inmir-133bat early retinal degenerative stage in RCS rats can protect dopaminergicamacrine cells from developmental and differentiation failure and canmake sure dopaminergic amacrine cells to display TH releasing function.

4. Discussions

We have noted that the expression level of melanopsin was de-creased during retinal degeneration in RCS rat over time, and this fur-ther study shows that the total numbers of RGCs were decreased indegenerative retinas, but the numbers and distribution of mRGCsdidn't alter using retrograde labeled and immunofluorescence

staining technology. Meanwhile, this result is also supported by ourother result that increased PACAP mRNA level was noted in RCS ratretinas because other research has shown that PACAP and melanop-sin are colocalized in the certain RGCs and PACAP-containing nerve fi-bers, which means that PACAP can be as a specific marker of mRGCs.Apparently, the decreased melanopsin expression cannot be ascribedto the alteration of mRGCs numbers in RCS rats.

In our further study, we noted that the expressions of TH and D2receptors in interneuron and ganglion cells were markedly decreasedin RCS rats. Importantly, the melanopsin and D2 receptors expressionlevel is significantly recovered or even increased after intravitreal in-jection of a agonist of D2-like receptor (NPA) in RCS rats retina. The

D

B

C

A

p+,p30, X400 TH(red) and Pitx3(green) Rdy+,p30, X400 TH(red) and Pitx3(green)

p+,p15, X400 TH(red) and Pitx3(green) Rdy+,p15, X400 TH(red) and Pitx3(green)

FE p+,p60, X400 TH(red) and Pitx3(green) Rdy+,p60, X400 TH(red) and Pitx3(green)

p+,p90, X400 TH(red) and Pitx3(green) Rdy+,p90, X400 TH(red) and Pitx3(green) G H

Fig. 5. Double-labeled immunofluorescence staining against pitx3 and TH. TH (red color) presents a constant expression in control retinas and intense TH fluorescence staining isobserved at a various of time points(A, C, G and F show control retinas at p15, p30, p60 and p90 respectively). Unlike control, TH expression level presents progressing declinetrends and weaker TH positive signals are observed in RCS rat retinas with retinal degeneration (B, D, F and H show RCS rat retinas at p15, p30, p60 and p90 respectively). Similarly,same expression pattern is noted for pitx3 expression. The intense pitx3 fluorescence staining is observed at various time points in control retinas. On the contrary, the expressionlevel of pitx3 decreased tremendously in INL of RCS rat retinas at p15 and then weaker pitx3 positive signals are observed in RCS rat retinas with time over during retinal degen-eration Furthermore, double-labeled immunofluorescence staining reveal that there are colocalization positive signals between TH and pitx3 staining, and both TH and pitx3 pos-itive signals are progressing decreased in RCS rat retinas. Original magnification 400.


data suggested that the ability of mRGC expressing melanopsin is notdamaged during retinal degeneration. Once the D2 pathway is acti-vated or dopamine is produced enough, the melanopsin expressionlevel can be improved. Considering the following several issues, ①inthe retina, DA is exclusively contained in the amacrine cells and is re-leased by a unique set of amacrine cells, DA ACs; ②Dopamine acti-vates D1 and D2 dopamine receptors distributed throughout theretina [18]; ③The expression of D2 receptors in amacrine cells andmRGCs is dependent on dopamine release[19]. Therefore, we can sup-pose that the DA ACs might be disrupted in RCS rats, which shouldlead to diminished dopamine production and release in amacrinecells as well as reduced expression of D2 receptors in RCS rat retinasas compared to control rats.

Fig. 6. Disrupted dopaminergic amacrine cells. Based on calretinin staining, we used divisiolayer) for development and differentiation analysis in this study as previously described (2,3stratify in the S1 sublamina of the IPL and the localization of calretinin interneurons processomata were closed contacted or surrounded with melanopsin positive dendrites and calrword, a close interaction patterns between dopaminergic amacrine cells and mRGCs arborizalamina-specific neurite arborization of dopaminergic amacrine cells during degeneration inthe melanopsin positive neurite arborization presented uniformly discrete patterns in RCSstaining, five parallel sublaminae were also divided in IPL for analysis. The chaotic sublaminof IPL were very faint. The melanopsin positive neurite arborizations were shorter and didn

It is well known that amacrine cells constitute approximately 40%of all neurons in the inner nuclear layer (INL) [20]. They are impor-tant interneurons in the retina and function in retinal circuitry, medi-ating horizontal processing of neuronal signals delivered from bipolarto ganglion cells [21]. They are responsible for 70% of input to retinalganglion cells. Bipolar cells, which are responsible for the other 30% ofinput to retinal ganglia, are regulated by amacrine cells. The most ac-cessible dopaminergic neurons of the vertebrate CNS are the dopami-nergic amacrine cells (DA cells) of the retina. Histological studieshave demonstrated DA ACs and mRGC dendritic are co-localized inthe S1 sublaminae of retinas. DA ACs may direct RGC dendritic strat-ification [22,23,24] and provide specific cues mRGCs to form selectivesynaptic contacts. Therefore, it is possible also that those disrupted

n of the IPL into five parallel sublaminae (S1–S5, S5 being closest to the ganglion cell). The tyrosine hydroxylase-positive dopaminergic amacrine cells were predominantlyses is not disrupted in wild-type retinas. Furthermore, the calretinin-positive amacrineetinin positive interneurons form a discrete plexus with melanopsin cells. In anothertion was observed in control retinas. In contrast, we identified defects in the stereotypicRCS rats (arrowed) by con-focal microscope. Also, very different with control retinas,rat retinas (light blue arrows Fig. 4) at p60 and p90. In addition, based on melanopsinae were noted at p15 and p30 in RCS rats. Furthermore, the melanopsin positive signals't reach into INL.

E G

F Hp+, p60

rdy+, p90p+, p90

rdy+, p60

INL

IPL

GCL

S1S2S3S4S5

B

C D

p+, p15

rdy+, p30p+, p30

rdy+, p15

A

J

p+, p30, X40 K L

p+, p15, X40 I

Rdy+, p30, X40

Rdy+, p15, X40


1w 4w2w

TH

actin

Mir133b/RNAi Mir133b/RNAi Mir133b/RNAi control control control

Mela.

actin

pitx3

TH expression level

00.5

11.5

22.5

33.5

4

1w 2w 4w 1w 2w 4w 1w 2w 4w

The

rat

io o

f T

H to

act

in

mir133b/RNAiControl virus

Melanopsin expression level

00.20.40.60.8

11.21.4

Rat

io o

f m

elan

opsi

n to

act

inmir133b/RNAi

control virus

Pitx3 expression level after deliveryingmir133b/RNAi lentivirus

00.20.40.60.8

11.2

Rat

ion

of p

itx3

to a

ctin

mir133b/RNAi

control

OPN4 and TH mRNA expression level

00.20.40.60.8

11.21.41.61.8

1w 2w 4w

Rel

ativ

e fo

ld c

hang

es

A B

mir133b control virusmir133b/RNAi OPN4mir133b/RNAi TH

C D E

F2

2W control, OPN4(red color),X400

K

J

I

H

2W control, TH(red color),X400

G

F1

PBS control lentivirus

2W mir133b/RNAi, OPN4(red color),X400

Delivery mir133b/RNAi lentivirus by intravitreal injection

2W mir133b/RNAi, TH(red color),X400

Q

P

O

N

M

L

2W control, calretinin(red color),X400 2W control, pitx3(red color),X400 2W control, D2(red color),X400

2W mir133b/RNAi, pitx3(red color),X400 2W mir133b/RNAi, calretinin (red color),X400 2W mir133b/RNAi, D2(red color),X400



DA ACs lose their processes and hereby lose the majority of normalsynaptic contacts between mRGCs and DA ACs themselves so thatmRGCs cannot receive modulatory information from DA ACs duringretinal degeneration in RCS rats. In addition, retinal dopamine hasmultiple roles in vision, mediating visual processes such as adaptinglight/dark retinal circuits, serving as an output of the retinal circadianclock, modulating gap junction, coupling horizontal cell, influencingrod photoreceptor Na+, K+−ATPase activity and trophic processes,including photoreceptor survival and the degree of refractive errorsin myopia [18]. Taken together, disrupted DA ACs cause a great effec-tion to visual system, especially to melanopsin expression. Based onthe results and hypothesis above, our study further focused on DAACs.

These guesses were confirmed by our followed experiments. Undernormal condition, DA ACs are located in the proximal inner nuclearlayer (INL) and send long (0.5 mm) processes laterally in sublamina 1(S1) of the inner plexiform layer (IPL) and into the outer plexiformlayer. Our results revealed that DA ACs during retinal degeneration inRCS rats couldn't elaborate their processes and presented defectivelamina-specific neurite arborization. An ultra-structural study in micehas shown that melanopsin-expressing ganglion cells located in theganglion cell layer (GCL) and displaced to the inner nuclear layer(INL) have close contact with amacrine and bipolar cells of unknownphenotype and form a photosensitive network [25](Fig. 3J. Schematicmodel of correct connection between amacrine cells and mRGC). In ad-dition, it also remains plausible that communication between thesecells via secretion of transmitters influences their arborization becauseamacrine cells express neurotransmitters even before their cell bodiesreach the INL/IPL border [26,27]. Upon the whole, mRGCs photosensi-tive network may easier to be influenced by DA ACs.

We next asked what is the underlying mechanism that cause fail-ure of DA ACs function during retinal degeneration. MicroRNAs (miR-NAs) are about 22-nucleotide, short, noncoding RNAs that arethought to regulate gene expression through sequence-specific basepairing with target mRNAs. MicroRNAs play important roles in thecontrol of development. Most importantly, miRNAs are involved in adiverse set of cellular processes, ranging from proliferation, apoptosisand malignant transformation to neuronal development and fatespecification [28,29,30]. We took a subtractive approach and com-pared miRNA expression profiles of the RCS rat retinas with the pro-files of control rat retinas. Expression analyses were performed for apanel of miRNA by quantitative real-time reverse transcription poly-merase chain reaction (qPCR). The result showed that expression ofone of these precursor miRNAs, miR-133b, was specifically enrichedin the RCS rat retinas at p15. MiR-133b has been implicated in Parkin-son's disease (PD) by a mechanism that involves the regulation of theimportant transcription factor Pitx3 in dopaminergic neurons [31].Over-expression of miR-133b suppresses dopaminergic neurons mat-uration and function and led to decreased expression of Pitx3 througha negative feedback loop forming between mir133b and Pitx3 in mid-brain. In PD patients [32,33,34] the inadequate processing of visualinformation has been correlated with decreased retinal levels of do-pamine and its rate-limiting synthesizing enzyme, tyrosine hydroxy-lase (TH), in dopaminergic processes. We next verified that miR-133bdid over-express in DA ACs of RCS retinas by performing combinedimmunfluorescence staining with in situ hybridization, namelymir133b and TH were colocalized in DA ACs. Moreover, the

Fig. 7. The levels of TH, pitx3 and melanopsin restores tremendously as a result of blocked m1w, 2w and 4w after delivering mir133b/RNAi lentivirus compared to the RCS rat retinas dmelanopsin mRNA expression level in RCS rat retinas at 1w, 2w and 4w after delivering mirobserved by real time PCR. (C), The relative protein expression of TH to actin. (D), The relativto actin. The values shown represent mean±standard error from three replicates. (F&G), Dtively by immunofluorescence staining. (H&I), TH expression levels at 2w after delivering comelanopsin expression level at 2w after delivering control and mir133b/RNAi lentivirus redelivering control and mir133b/RNAi lentivirus respectively by immunofluorescence staininvirus respectively by immunofluorescence staining.

expression of Pitx3 was significantly down-regulated at early stageof retinal degeneration in RCS rat as compared to the age-matchedcontrol retinas. Meanwhile, miR-133b was slightly reduced at p30,which might be due to the loss of dopaminergic neurons numbers.Correspondingly, Pitx3 expression level in RCS retinas sustained alevel close to that in control retinas at p30. Therefore, a negative feed-back loop was also confirmed between mir133b and Pitx3 in retina.These results suggested that at early stage of degeneration overex-pressed miR-133b may directly target Pitx3 in DA ACs. Our resultsalso raise the possibility that the overexpressed mir133b at earlystage during RCS retinal degeneration might be involved in collapseof DA ACs and caused maturation and stratification deficiency of DAACs because the retinal dopaminergic system appears particularlyvulnerable to developmental insults [35]. During retina development,retinal ganglion cells and amacrine cells are the first cell classes to dif-ferentiate and form the earliest functional circuits in the IPL of the de-veloping retina. Shortly after, horizontal cells and photoreceptorsdifferentiate and contact each other in the outer retina, forming theOPL. The vertical networks in the inner and outer retina are laterinterconnected when bipolar cells are born and connections withganglion cells are established. This sequential pattern of retinal circuitdevelopment is common across species, ranging from a few hours inanimals such as the zebrafish to many days and weeks in mammals[36,37,38]. In this study, chaotic neurite arborizations were alsoseen in mRGCs and disorderedmRGC dendrites loss their parallel sub-laminae structures within the IPL of RCS retinas, which might be dueto a secondary consequence of the deficient DA ACs or a compensato-ry response of mRGC.

Obviously, when mir-133b was interfered at early retinal degenera-tive stage in RCS rats by delivering lentivirus by intravitreal injection,DA ACs can be protected from developmental and differentiationfailure. The interfered mir-133b can make sure DA ACs to displayTH releasing function. What's worth mentioning is that the GABAer-nic neurons, GAD65 were not affected. It's also suggested that theoverexpression of mir133b only suppresses DA ACs, but not GABAer-nic amacrine cells.

In summary, it's very important to elucidate the regulatory mech-anism of melanopsin expression for adequately exerting all mRGCs'powers after visual loss. The major finding in this study was thatmir133b and pitx3 may be key factors for the development and ma-ture DA ACs in retinal. There were five main interrelated findings inthis study. First, decreased melanopsin expression level may be notdue to the number loss of mRGCs, but due to development deficientof DA ACs which lead to an impaired DA/D2 receptor regulatory sig-nals. Second, decreased the amount of retinal TH and D2 receptormRNA and protein is main reason for decreased melanopsin expres-sion in RCS rat retinas. Third, increased mir133b and diminishedpitx3 in RCS rat retinas. Fourth, dysfunction of DA ACs may be dueto increased mir133b and diminished pitx3. Finally, melanopsin andTH expression level can be recovered by delivery of mir133b/RNAilentivirus.

Conflict of interest statement

The authors declare that they have no competing financialinterests.

ir133b. (A),Recovered TH, pitx3 and melanopsin protein expression in RCS rat retinas atelivering control lentivirus were observed by Western blotting. (B), Recovered TH and133b/RNAi lentivirus compared to the RCS rat retinas delivering control lentivirus weree protein expression of melanopsin to actin. (E), The relative protein expression of pitx32 expression levels at 2w after delivering control and mir133b/RNAi lentivirus respec-ntrol and mir133b/RNAi lentivirus respectively by immunofluorescence staining. (J&K),spectively by immunofluorescence staining. (L&M), pitx3 expression level at 2w afterg. (N&O), calretinin expression at 2w after delivering control and mir133b/RNAi lenti-


Author contributions

Yaochen Li designed and conducted the experiments and contrib-uted to manuscript preparation. Chunshi Li, Zhongshan Chen,Jianrong He and Zui Tao conducted the experiments. Zheng Qin Yincontributed to design and funding of the project, and contributed todiscussing and reviewing the manuscript.

Acknowledgments

Thanks to The Central Laboratory of Xi'nan Hospital for providinga good experimental platform for us. Thanks go to all laboratory staffalways present to help us.

References

[1] C. Sedwick, PLoS Biology 8 (2010) e1001003.[2] J.J. Gooley, J. Lu, D. Fischer, C.B. Saper, Journal of Neuroscience 23 (2003)

7093–7106.[3] H.J. Bailes, R.J. Lucas, Cellular and Molecular Life Sciences 67 (2010) 99–111.[4] S.N. Peirson, S. Halford, R.G. Foster, Philosophical Transactions of the Royal Society of

London. Series B, Biological Sciences 364 (2009) 2849–2865.[5] A.A. Vugler, P. Redgrave, M. Semo, J. Lawrence, J. Greenwood, P.J. Coffey, Experi-

mental Neurology 205 (2007) 26–35.[6] K. Sakamoto, C. Liu, M. Kasamatsu, N.V. Pozdeyev, P.M. Iuvone, G. Tosini, European

Journal of Neuroscience 22 (2005) 3129–3136.[7] D.L. Boudard, J. Mendoza, D. Hicks, European Journal of Neuroscience 30 (2009)

1527–1536.[8] K. Sakamoto, C. Liu, G. Tosini, Journal of Neuroscience 24 (2004) 9693–9697.[9] G. Paxinos, C. Watson, 2nd ed. Academic Press, San Diego, 1986.

[10] W. Sun, N. Li, S. He, Visual Neuroscience 19 (2002) 483–493.[11] Z.S. Chen, Z.Q. Yin, S. Chen, S.J. Wang, Neuroreport 16 (2005) 971–975.[12] J. Hannibal, P. Hindersson, J. Ostergaard, B. Georg, S. Heegaard, P.J. Larsen, J. Fahrenkrug,

Investigative Ophthalmology & Visual Science 45 (2004) 4202–4209.[13] J. Hannibal, International Review of Cytology 251 (2006) 1–39.[14] H. Kanasaki, T. Yonehara, Y. Yamada, K. Takahashi, K. Hata, R. Fujiwaki, H. Yamamoto,

Y. Takeuchi, K. Fukunaga, E. Miyamoto, K. Miyazaki, Biology of Reproduction 67(2002) 1218–1224.

[15] R.L. Matsuoka, K.T. Nguyen-Ba-Charvet, A. Parray, T.C. Badea, A. Chédotal, A.L.Kolodkin, Nature 470 (2011) 259–263.

[16] J.R. Sanes, S.L. Zipursky, Neuron 66 (2010) 15–36.[17] H. Wässle, Nature Reviews Neuroscience 5 (2004) 747–757.[18] P. Witkovsky, Documenta Ophthalmologica 108 (2004) 17–40.[19] J.M. Beaulieu, R.R. Gainetdinov, Pharmacological Reviews 63 (2011) 182–217.[20] E. Strettoi, R.H. Masland, Journal of Neuroscience 15 (1995) 875–888.[21] R.H. Masland, Nature Neuroscience 4 (2001) 877–886.[22] A.D. Huberman, T.R. Clandinin, H. Baier, Cold Spring Harbor Perspectives in Biol-

ogy 2 (2010) a001743.[23] R.C. Stacy, R.O. Wong, The Journal of Comparative Neurology 456 (2003)

154–166.[24] J.N. Kay, T. Roeser, J.S. Mumm, L. Godinho, A. Mrejeru, R.O. Wong, H. Baier, Devel-

opment 131 (2004) 1331–1342.[25] J. Østergaard, J. Hannibal, J. Fahrenkrug, Investigative Ophthalmology & Visual

Science 48 (2007) 3812–3820.[26] M. Feller, Webvision: The Organization of the Retina and Visual System [Internet],

University of Utah Health Sciences Center, Salt Lake City (UT), 1995.[27] J. Morgan, R. Wong, Webvision: The Organization of the Retina and Visual System

[Internet], University of Utah Health Sciences Center, Salt Lake City (UT), 1995.[28] D. Leaman, P.Y. Chen, J. Fak, A. Yalcin, M. Pearce, U. Unnerstall, D.S. Marks, C.

Sander, T. Tuschl, U. Gaul, Cell 121 (2005) 1097–1108.[29] Y. Li, F. Wang, J.A. Lee, F.B. Gao, Genes & Development 20 (2006) 2793–2805.[30] R. Fiore, G. Siegel, G. Schratt, Biochimica et Biophysica Acta 1779 (2008) 471–478.[31] L. de Mena, E. Coto, L.F. Cardo, M. Díaz, M. Blázquez, R. Ribacoba, C. Salvador, P.

Pastor, L. Samaranch, G. Moris, M. Menéndez, A.I. Corao, V. Alvarez, AmericanJournal of Medical Genetics. Part B, Neuropsychiatric Genetics 153B (2010)1234–1239.

[32] C. Harnois, T. Di Paolo, Investigative Ophthalmology & Visual Science 31 (1990)2473–2475.

[33] O. Dkhissi, N. Dalil-Thiney, C. Versaux-Botteri, E. Chanut, J. Repérant, J. Nguyen-Legros, Ophthalmic Research 25 (1993) 280–288.

[34] T. Vuvan, M. Geffard, P. Denis, A. Simon, J. Nguyen-Legros, Brain Research 614(1993) 57–64.

[35] D.C. Jones, G.W. Miller, Biochemical Pharmacology 76 (2008) 569–581.[36] J.W. Olney, Investigative Ophthalmology 7 (1968) 250–268.[37] J.C. Blanks, A.M. Adinolfi, R.N. Lolley, The Journal of Comparative Neurology 156

(1974) 81–93.[38] L.J. Fisher, The Journal of Comparative Neurology 187 (1979) 359–372.

A MicroRNA, mir133b, suppresses melanopsin expression mediated by failure dopaminergic amacrine...

Documents

Transcript of A MicroRNA, mir133b, suppresses melanopsin expression mediated by failure dopaminergic amacrine...