identification of the bcl-2 family of genes in the rat retina

Identification of the bd-% Family of Genes in the Rat Retina

Leonard A. Levin, Cassandra L. Schlamp, Rachel L. Spieldoch, Kati M. Geszvain,and Robert W. Nickells

Purpose. Retinal ganglion cells die by apoptosis after axotomy, and this process may reflectaltered expression of cell-death genes. Several of these genes, including bcl-2, bcl-x, and bax,share homology at the amino acid level in the BH1 and BH2 domains, through which theyalso interact. To understand their role in the neuronal response to axotomy, the authorsstudied their expression in the adult rat retina and after optic nerve crush.

Methods. An initial survey was conducted with reverse transcription-polymerase chain reaction(RT-PCR), using oligonucleotides against identified members of this family and against theconserved BH1 and BH2 domains. Retinal bd-xL expression at the messenger RNA (mRNA)and protein level was studied by RT-PCR, Northern blotting, RNase protection analysis, insitu hybridization, Western blotting, and immunofluorescence staining. The effect of retinalganglion cell axotomy on the steady-state level of bd-x mRNA was investigated.

Results. RT-PCR results indicated that rat retinal cells predominandy express the long formof bd-x. Both clonal analysis and quantitative measurements using RNase protection assaysdemonstrated that bd-x^ message was at least 16 times more abundant than that of bd-2. Insitu hybridization and indirect immunofluorescence demonstrated that nearly all neuronalcells of the retina express bd-x. Northern and RNase protection analyses showed a moderatedecrease in bd-x^ message shortly after optic nerve crush.

Conclusions. These findings suggest that the antideadi gene bd-x^ is die predominant memberof the bd-2 family in the adult retina, and that its level decreases after optic nerve crush.Changes in bd-xL expression may correlate with increased retinal ganglion cell apoptosis afteraxotomy. Invest Ophthalmol Vis Sci. 1997; 38:2545-2553.

JN euronal apoptosis typically involves transcriptionalregulation of cell-death genes. One of the most stud-ied group of genes is the bcl-21 ~4 gene family, the mem-bers of which include bd-x,5 bax,6 bad,7 bak,8~10 andbik.u Many of these genes share nucleotide and aminoacid homology at two or more sites. Two initially de-scribed conserved homology domains are named BH1and BH2 (for Bcl-2 homology 1 and 2 domains), withinteractions between gene family members occurringat these sites.12 Control of cell death appears to be

From the Department of Ophthalmology and Visual Sciences, University ofWisconsin Medical School, Madison, Wisconsin.Supported in part by NIH KI1EY00340 (IAL), Retina Research Foundation(IAL), Research to Prevent Blindness Career Development Award (RWN), AmericanHealth Assistance Foundation (RWN and l^AL), The Wisconsin Lions Foundation,and an unrestricted departmental giant from Research to Prevent Blindness, Inc.1AL is an IU>B Dolly Green Scholar.Submitted for publication March 10, 1997; revised June 9, 1997; accepted July 9,1997.Proprietary interest category: N.Reprint requests: l^eonard A. Levin, Department of Ophthalmology and VisualSciences, University of Wisconsin Medical School, 600 Highland Avenue, Madison,Wl 53792. E-mail: [email protected]

regulated by these interactions and by constitutive ac-tivities of the various family members (e.g., blockadeof lipid membrane peroxidation by Bcl-2).13 Subse-quently, the BH3 and BH4 homology domains weredescribed14; the former is also a site for interactionbetween members of this extended family. Althoughbcl-2 is expressed in a variety of cell types, especiallystem cells and epithelia,15 bcl-x is expressed in a morerestricted subset of cells.5 Similarly, bcl-2 is widely ex-pressed in the developing nervous system and adultperipheral nervous system,16 but adult central neuronspreferentially express bcl-x.xl

In the retina, retinal ganglion cells respond toaxotomy by undergoing apoptosis,18"21 probably trig-gered by loss of retrograde transport of brain-derivedneurotrophic factor or other neurotrophins.22"25

Overexpression of bcl-2 in transgenic animals preventsretinal ganglion cell death after optic nerve injury,26'27

as well as death of sympathetic and other neuronalcell types after deprivation of nerve growth factor and

Investigative Ophthalmology & Visual Science, November 1997, Vol. 38, No. 12Copyright © Association for Research in Vision and Ophthalmology 2545

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933420/ on 04/11/2018

2546 Investigative Ophthalmology 8c Visual Science, November 1997, Vol. 38, No. 12

other neurotrophins.28 30 Although these and similarexperiments demonstrate the capacity for bcl-2 overex-pression to rescue dying neurons, the role of physio-logical expression of bcl-2 and similar genes in neuronssusceptible to neurotrophin deprivation injury is notwell understood. To determine the constitutive cell-death genes in the retina, especially the retinal gan-glion cell, we studied the expression of bcl-2 familymembers in normal and axotomized retinas using de-generate polymerase chain reaction (PCR) and RNAanalysis.

MATERIALS AND METHODS

Optic Nerve Crush

Animal procedures were in accordance with institu-tional, state, ARVO, and federal guidelines. All surgerywas performed on the right eye of adult Long-Evansrats. Animals were anesthetized with ketamine-xylaz-ine and a limited lateral canthotomy was performed.The conjunctiva was incised at the limbus and thesubtenon's space was bluntly dissected posteriorly.The muscle cone was then entered and the optic nerveexposed. The optic nerve was crushed with blunt for-ceps for 5 seconds 3 mm posterior to the globe underdirect visualization. The ocular fundus was observedophthalmoscopically, and animals with evidence ofcentral retinal artery occlusion were not used in subse-quent procedures. The skin was closed with absorba-ble sutures, topical bacitracin was applied, and theanimal was returned to the cage. After recoveringfrom anesthesia, animals were observed to feed anddrink normally. Typically, 6 to 10 animals were oper-ated on at each sitting.

RNA Isolation

Rats were killed with carbon dioxide and the eyes re-moved. For studies of axotomized retinas, animals werekilled 24 and 96 hours after optic nerve crush, and thecon trilateral eyes and eyes from unoperated animalswere used as controls. Retinas were dissected in sterileHank's balanced salt solution and flash-frozen on dryice. Total RNA was then isolated using the guanidiniumthiocyanate-phenol-chloroform method,31 treatedwith RNase-free DNase I, phenol-chloroform extracted,and precipitated with ethanol. This DNA-free RNA wasused for reverse transcription-PCR (RT-PCR), North-ern blotting, and RNase protection assays.

Degenerate PCR

Degenerate primers specific to the most conservedportions of the BH1 and BH2 domains were designedwith the aid of MacVector version 4.5 (Kodak Scien-tific Imaging, New Haven, CT) and a local primeranalysis program. The forward primer was 5'-AAC-

TGGGGKMGSATYGTGGC-3', the reverse primer 5'-CCCAKCCKCCGTTMTCCTGGATC-3'. These prim-ers were designed to maximize amplification of se-quences similar to the antiapoptosis genes bcl-2 andbcl-xL and were synthesized by Operon Technologies(Alameda, CA).

Total retinal RNA (10 //g) was treated with RNase-free DNase and reverse-transcribed in a 50-fA reactionwith MMLV-reverse transcriptase (Gibco, Gaithers-burg, MD) at 37°C. Approximately 1 ng ssDNA wasamplified in a PCR reaction containing 1 nmol dNTP(Boehringer-Manheim, Indianapolis, IN), 2.5 U Am-pliTaq (Perkin-Elmer, Norwalk, CT), and 50 pmol ofeach degenerate primer in 50 mM KC1, 10 mM Tris(pH 8.3), and 1.5 mM MgCl2 to a total volume of50 fi\. Polymerase chain reaction amplification wasperformed on a Perkin-Elmer 480 thermal cycler (Per-kin-Elmer) at 95°C for 5 minutes, 30 cycles of 95°Cfor 30 seconds, 55°C for 30 seconds, 55°C for 30 sec-onds, and 72°C for 60 seconds, followed by 72°C for5 minutes.

Cloning of PCR Products

Reaction products were electrophoretically separatedon 1.5% low-melting-point agarose gels. The gel slicecontaining the predicted 160-base pair (bp) band wasexcised with a new razor blade, digested with /?-agarase(New England Biolabs, Beverly, MA), and ethanol pre-cipitated. The PCR amplimer, containing a singleoverhanging A at each 3' end, was ligated into thepT7Blue TA-cloning vector (Novagen, Madison, WI)and used to transform supercompetent cells. The pres-ence of inserts was detected with blue-white selections.Standard procedures were used for overnight culturesand plasmid preparations.

Clone Analysis

To facilitate rapid analysis of clones, the presence of aninsert was detected by PCR, using primers for the T7and M13-40 flanking sites of the vector. To screen forclones containing bcl-2 or bd-x^ inserts, the followinginternal primers were designed using portions of thenucleotide sequence where there was minimal homol-ogy: bcl-^: (forward) 5'-AACTGGGGTCGATTGTG-3'and (reverse) 5'-GATCCAAGGCTCTAGGTGGTC-3';bcl-2: (forward) 5'-TTGAGTTCGGTGGGGTCATG-3'and (reverse) 5'-GATCCAGGTGTGCAGATGCC-3'.These were used in PCR performed directly after insertdetection, using standard methods. Manual dideoxy se-quencing was performed using the Sequenase 2.0 kit(Amersham, Arlington Heights, IL).

Cloning of Rat bcl-xL

Rat retinal complementary DNA (cDNA) was ampli-fied using bcl-x primers and conditions as describedpreviously.5 These primers amplify a 764-bp cDNA


Retinal Cell-Death Genes 2547

from the messenger RNA for bd-xL and an approxi-mately 600-bp cDNA from the alternately spliced bcl-x$. Reverse transcription-PCR using rat retinal cDNAas a template yielded only bd-xL in our experiments.For cloning, the 764-bp cDNA was gel purified, incu-bated with Klenow fragment in the absence of dNTPto remove overhanging 3' ends, and filled in withKlenow and dNTP. This blunt-ended amplimer wasthen ligated for 4 days at 4°C into the Smal site of pBK-CMV (Stratagene, Lajolla, CA) and used to transformDH5a; cells. Sequence analysis and PCR were used toconfirm both identity and orientation. This constructwas named pRbcl-x.53. To construct a smaller cDNAof bd'XL for RNase protection assays and in situ hybrid-ization experiments, pRbcl-x.53 was digested withKpnl, which cut both at an internal site and the 5'end of the polylinker, resulting in the excision of the5'-most 341 bp of bd-xL. The parent vector was reli-gated and cloned, resulting in pXEK, which containedthe 3'-most 423 bp of bd-xL between EcoRI and Kpnlsites.

Northern Blotting

Twenty micrograms of total RNA pooled from fourrats for each time point was separated with formalde-hyde-agarose electrophoresis and pressure blottedonto MagnaGraph nylon membranes (Micron Separa-tions, Westboro, MA). After ultraviolet light cross-link-ing, the blot was hybridized overnight at 42°C with arandom-primed (Stratagene) 32P-labeled probe pre-pared from pRbcl-x.53, containing the full-length ratbd-xL sequence, or a 1.9-kb murine bd-2 clone (gener-ously provided by S. Korsmeyer), washed twice in 2 XSSC-0.1% sodium dodecyl sulfate at 37°C, twice in0.1 X SSC-0.1% sodium dodecyl sulfate at 56°C, andexposed to XOMAT-AR (Kodak, Rochester, NY) filmat — 70°C using intensifying screens. Multiple expo-sures were used to take advantage of the linear rangeof the film. Quantification of autoradiograms was per-formed by densitometry and NIH Image version 1.59(National Institutes of Health, Bethesda, MD).

RNase Protection Assay

RNase protection analyses were carried out as de-scribed previously.32 Antisense RNA probes were syn-thesized from clone BH160.9 (encoding the 160 ntregion of bd-xL between the BH1 and BH2 domains),clone BH160.12 (encoding the homologous 160 ntregion of bd-2), and pXEK. All RNA probes for RNaseprotection analysis were labeled by incorporating a-[32P]-uridine triphosphate in the in vitro transcriptionreaction. After purification from 6 M urea, 5% poly-acrylamide gels, ~5 X 105 cpm of each RNA probewas coprecipitated with appropriate amounts of totalrat retinal RNA. For experiments comparing the levelsof bd-xt and bd-2 rnRNAs, RNA probe was precipitated

with an increasing amount of total retinal RNA rang-ing from 5 to 25 fig. For experiments on axctomizedretinas, RNA probe was precipitated with 6 //g totalRNA isolated from retinas at 0 days, 1 day, and 4 daysafter axotomy. In each case, nonspecific hybridizationof the RNA probes was controlled for by incubatingthe same amount of probe with an equal amount ofyeast tRNA. After precipitation, the samples were airdried and then resuspended in 20 /A hybridizationbuffer (80% formamide, 40 mM PIPES, 1 mM ethyl-enediaminetetraacetic acid, 200 mM sodium acetate,pH 6.8) and heated to 95°C for 5 minutes in a PerkinElmer 9600 Thermocycler followed by rapid coolingto 50°C, where the samples were incubated for 16 to18 hours. After hybridization, the samples were rapidlydiluted with 180 fA buffer containing 10 mM Tris (pH7.4), 5 mM ethylenediaminetetraacetic acid, and 200mM sodium acetate. The unhybridized probe in eachreaction was digested using 20 units of RNase ONE(Promega, Madison, WI) for 1 hour at 22°C. Protectedprobe was then precipitated and analyzed by denatur-ing gel electrophoresis followed by autoradiography.Quantification of protected molecules was made bycutting them out of a dried gel and counting them ina Beckman LS6500 liquid scintillation counter (Fuller-ton, CA).

In Situ Hybridization

In situ hybridization was carried out using a whole-mount procedure designed for vertebrate embryos33'34

and later modified for blocks of rat retina as previouslydescribed.32 Briefly, weaned rats were killed by carbondioxide inhalation and the eyes enucleated. An inci-sion was made at the midline of the globes, whichwere then placed in 4% paraformaldehyde in 100 mMphosphate buffer (pH 7.2) for 1 hour at 22°C. Eyeswere then placed in 0.4% paraformaldehyde andstored at 4°C for as long as 6 months. After fixation,the eyes were placed in 0.9% NaCl and the retinasdissected from the surrounding tissue, then furthercut into small blocks (about 1 mm).3 These smallblocks were used for the whole-mount hybridizationprotocol as described.32 Antisense RNA probes weresynthesized using the pXEK plasmid and digoxigenin-labeled uridine triphosphate (Boehringer-Mannheim,Indianapolis, IN). Tissue preparation and hybridiza-tion conditions were identical to those previously de-scribed, except that samples were hybridized with 500ng/ml antisense pXEK for 24 to 36 hours at 60°C.Control probes containing sense pXEK sequence andsequence specific for a gene expressed in photorecep-tor cells (rod opsin) were also generated to monitorfor nonspecific hybridization using these conditions.After hybridization, unbound probe was removed bywashing in 50% formamide at 60°C followed by incu-bation with RNase A, as described by Li et al.34 Hybrid-


2548 Investigative Ophthalmology & Visual Science, November 1997, Vol. 38, No. 12

ized probe was detected using Fab fragments of antidi-goxigenin antibodies conjugated to alkaline phospha-tase (Boehringer-Mannheim). Samples were stainedovernight in alkaline phosphatase buffer (100 raMTris, pH 9.5, 50 mM MgCl2, 100 raM NaCl, 0.1%Tween-20) containing 340 fJ,g/m\ nitro blue tetrazo-lium and 175 ng/m\ 5-bromo-4-chloro-3-indolyl-l-phosphate (Promega). After staining, samples werepostfixed in 4% formaldehyde and dehydrated to 90%methanol. This dehydration step enhances the signalby bleaching the surrounding tissue. All samples werethen embedded in glycomethacrylate (JB-4 Plus, Poly-sciences, Warrington, PA) and sectioned at 5 fim.

Immunofiuorescence Labeling

Rat eyes were harvested and fixed as described for insitu hybridization. For immunofiuorescence, the entireeye was equilibrated in 30% sucrose and embedded inHisto Prep (Fisher Scientific, Pittsburgh, PA). Cryosec-tions of 5 to 7 ^m were cut for immunolabeling. Immu-nolabeling was carried out as described32 with the follow-ing modifications. Sections were blocked in 100 mMphosphate buffer (pH 7.2) containing 4% bovine serumalbumin for 2 hours at 22°C. A polyclonal antibodyagainst Bcl-x, recognizing both the long and short pro-tein products of rat and human bcl-x (Santa Cruz Bio-technology, Santa Cruz, CA), was diluted 1:100 in phos-phate buffer and incubated with the sections overnightat 4°C. For detection, the sections were then incubatedwith goat antirabbit IgG conjugated to Texas Red (Jack-son ImmunoResearch, West Grove, PA) for 2 hours at22°C in the dark. Indirect immunofiuorescence wasmonitored and photographed using a Zeiss AxiophotPhotomicroscope (Thornwood, NY).

RESULTS

Amplification of the BH1-BH2 Region FromRetinal cDNARat retinal cDNA was amplified by PCR using degener-ate primers specific to BHl and BH2 homology do-mains shared by antideath bcl-2 family members. Agar-ose gel electrophoresis of the reaction product dem-onstrated a single 160-bp band, matching the lengthpredicted by the distance between the two homologydomains plus the length of the primers.

To identify which members of the bcl-2 familymight be present in the amplimers contained withinthe 160-bp band, amplified DNA was extracted fromlow-melting-point agarose and ligated into a TA-clon-ing vector, and clones of transformed bacteria wereisolated using ampicillin resistance and blue-white se-lection. White colonies were used for an initial screenby amplifying the insert within the vector's flankingsites. Of the 152 clones initially analyzed, 79 had in-

serts of the appropriate size. Clones with inserts werefurther analyzed by manual sequencing, PCR amplifi-cation using internal primers specific to bcl-2 or bcl-xL

sequences, or both methods.Analysis of all 79 of these clones revealed that 63

contained bcl-x^ sequence and 3 contained bcl-2 se-quence. The other 13 clones contained sequences lack-ing complete BHl and BH2 domains, presumably re-sulting from mispriming of the degenerate primers atnucleotide sequences weakly homologous to BHl andBH2. The identification of 21 of the 63 bcl-x^ clonesand all 3 of the bcl-2 clones was confirmed by directsequencing. The analysis of the other clones was per-formed by PCR amplification using bcl-x{- and bcl-2-spe-cific internal primers, generating a 141-bp amplimerwith bcl-xh primers and a 116-bp product with bcl-2 prim-ers. No clones were identified that contained completeBHl and BH2 domains, suggesting that expression ofother bcl-2 family members is relatively low in the retina.PCR of rat retinal cDNA using primers outside the BHl -BH2 area amplified 6c/-xL but not bcl-x$.

Relative Expression of Retinal bcl-2 and bcl-xL

Given that the vast majority of identified clones con-tained bcl-x^ sequence and relatively few contained bcl-2, we used RNase protection assays and Northern blot-ting to establish the relative abundance of mRNA forbcl-xL and bcl-2 in the retina. Total retinal RNA wasincubated with 32P-labeled 160-bp probes correspond-ing to the BH1-BH2 regions of rat bcl-xL and bcl-2.After RNase digestion, protected fragments were elec-trophoresed and auto radiograms prepared (Fig. 1).The bcl-xL probe protected 2 fragments of mRNA of160 and 140 nt. This reflects.a polymorphism of thebcl-x gene at nucleotide 546 of the coding sequence,where there is a C~»T transition in a degenerate isoleu-cine codon. This polymorphism is exactly 20 nucleo-tides from the end of the BH2 region, resulting in thetwo different protected fragments when derived fromRNA pooled from different outbred animals. No suchpolymorphism was evident for the bcl-2 probe, whichprotected a full-length 160 nt mRNA fragment. Therewas minimal bcl-2 mRNA expression, compared withbcl-xL. Direct counting of the labeled gel slices demon-strated that bcl-x^ mRNA was 16 times more abundantthan bcl-2 in the retina, which is similar to the 21-fold greater frequency of bcl-x^ clones than bcl-2 clonesfound in the clonal analysis. Similar findings were seenwhen Northern blots of total retinal RNA were probedin parallel for bcl-xXu and bcl-2 (data not shown).

Localization of bcl-xL in the Retina

Retinas were prepared for in situ hybridization usinga digoxigenin-labeled RNA probe corresponding tothe 423 nt of the 3' coding region of rat bcl-xXj. Noniso-topic detection revealed expression of bcl-xL in all cell


Retinal Cell-Death Genes

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 910

f 21Q nt

nt

B

FIGURE i. RNase protection assays demonstrating predomi-nantly bd-^ expression in the retina. Increasing amounts oftotal retinal RNA or yeast tRNA were incubated with a32P-labeled 210 nt RNA probe, which contained 160 nt of se-quence specific to the BH1-BH2 region of bcl-x^ or bcl-2. Apolymorphism in bd-x^ resulted in digestion to both 140 and160 nt protected fragments. (A) Lanes 1 to 5 correspond tobd-xL probe, lanes 6 to 10 to bd-2 probe. Lane order is asfollows: 1, 6: native probe; 2, 7: 10 fj,g yeast tRNA; 3, 8: 5 //gretinal total RNA; 4, 9: 10 fxg retinal total RNA; 5, 10: 25/j.g retinal total RNA. Two-hour exposure. (B) Same RNaseprotection assays as in A but a 6-hour exposure to demon-strate faint bd-2 protected fragment.

2549

prepared from eyes 1 and 4 days after optic nervecrush and amplified by PCR using degenerate primersspecific to the BH1 and BH2 regions. This resulted inthe previously detected 160-bp amplimer. Unexpect-edly, the intensity of this band was observed to de-crease over time. This finding was confirmed in threedifferent experiments using separate groups of experi-mental animals.

Previous experiments had demonstrated that thepredominant bcl-2 family member in the retina wasbcl-xL. To see whether the species decreasing in thedegenerate PCR amplifications corresponded to de-creases in retinal bd-xL after axotomy, we studied theexpression of the latter with Northern analysis andRNase protection assays. Blots of total RNA from axo-

layers of the retina, including the ganglion cell layer(Fig. 2). The cytoplasm and nuclei of many of the cellslabeled in the ganglion cell layer were large, which isa structure consistent with ganglion cells. However,the possibility that some displaced amacrine cells werelabeled could not be excluded. Hybridization withcontrol probes demonstrated no background staining.

The Bcl-x protein product was localized using apolyclonal primary antibody followed by an immuno-fluorescent second antibody. The specificity of the an-tibody was confirmed by its ability to detect a singleband of approximately 29 kDa on Western blots, con-sistent with rat Bcl-x (data not shown). Immunofluo-rescence studies revealed immunoreactive cells in thephotoreceptor inner segments, the outer plexiformlayer, the inner plexiform layer, and especially theretinal ganglion cell layer (Fig. 3). Again, many of thelabeled cells in the retinal ganglion cell layer had astructure suggestive of retinal ganglion cells.

4

GCL

Expression Decreases After RetinalGanglion Cell Axotomy

To understand the expression of bcl-2 family genesafter retinal ganglion cell axotomy, retinal cDNA was

ONL

INL

GCl

ONI

I N L

GCL

FIGURE 2. In situ hybridization of rat retina using a digoxi-genin-Iabeled RNA probe for the 3'-most 423 nt of bd-Xy^.Hybridized probe is detected as a dark reaction productafter incubation with substrate for alkaline phosphatase.There is moderate staining of all nuclear layers (A), as wellas focal staining adjacent to the outer limiting membraneof the photoreceptors (B) and in the ganglion cell layer(C). No staining was detected in retina incubated with asense probe (D). Retinas probed with a positive control op-sin probe stained in the inner segments and outer nuclearlayer (E). OS: outer segments; IS: inner segments; ONL:outer nuclear layer; OPL: outer plexiform layer; INL: innernuclear layer; IPL: inner plexiform layer; GCL: ganglion celllayer. Scale bar = 40 //m in A, D, and E and 20 /̂ m in Band C.



FIGURE ». Immunodetection of Bcl-x in the rat retina. A poly-clonal antibody against Bcl-x was followed by a Texas Redsecondary antibody. (A) There is strong staining in the pho-toreceptor inner segments, the outer plexiform layer, theinner plexiform layer, and the ganglion cell layer. Severallarge cells in the ganglion cell layer (B, C) are focally immu-noreactive for Bcl-x, suggestive of ganglion cells. Scale bar= 20 fxm in all panels.

tomized and control retinas were probed with a 32P-labeled full-length rat bcl-xL probe, and relative expres-sion was estimated by densitometry of multiple expo-sures of the autoradiograms. This demonstrated adecrease of approximately 30% at 4 days after opticnerve crush (Fig. 4). Similarly, total RNA from axo-tomized and control retinas incubated with a 32P-la-beled RNA probe for bd-xL, digested with RNase, andelectrophoretically separated demonstrated a declinein the intensity of the protected fragment, correspond-ing to a moderate decrease at 1 day and 32% at 4days, compared with the contralateral control (seeFig. 4).

DISCUSSION

The Lranslocation of a locus on chromosome 18 adja-cent to the immunoglobulin heavy-chain promoter onchromosome 14 in human follicular B-cell lymphomasis associated with inhibition of cell death. Subsequentstudies demonstrated that overexpression of the genederived from die breakpoint, B-cell lymphoma/leuke-mia 2 gene, or fo/-2,'~3 prevented cell deadi in a varietyof paradigms, including glucocorticoid and radiationtreatment of immature thymocytes,35 cytokine depriva-tion of hematopoietic cells,36 and neurotrophin depri-vation of culture neurons.28'2"1 Boise et al5 probed achicken lymphoid library with a murine bcl-2 probe

and identified a related gene bcl-x, which could bedifferentially spliced into two forms. The long form,or bd-xL> prevented apoptosis resulting from IL-3 dep-rivation of an IL-3-dependent prolymphocytic cellline; the short form, bd-Xs, inhibited the ability of bd-2 to block apoptosis induced by growth-factor with-drawal. Subsequently, several other genes sharing ho-mology to bd-2 have been identified through variousapproaches, including bax,6 bad,1 bak,8~w and biku

Some of these genes {bd-2 and bd-xL) appear to inhibitcell death; others (bd-x<,, bax, bad, bak) induce it. Inter-actions between members of this family may serve asa mechanism for regulating relative levels of celldeath,37 with binding occurring at £c/-2-related ho-mology domains BH1, BH2, and BH3.3738

Although bd-2 and bd-x$ are expressed predomi-nantly in populations of proliferating cells in adultanimals, bd-xL is frequently expressed in postmitotic

2.7kb

-28S-

-18S-

B

4 5

467

423 nt

C

FIGURE 4. Northern analysis and RNase protection assaysdemonstrate decreased bd-xL expression after optic nervecrush. (A) Northern blot of 20 /xg total RNA pooled fromfour retinas contralateral (lane 1) or ipsiiateral (lane 2) tooptic nerve crush, obtained 4 days after surgery, probed witha full-length bd-x^ probe. Arrow indicates expected 2.7-kbtranscript for bcl-xL. (B) Uniformity of gel loading deter-mined by ethidium bromide staining of 28S and 18S ribo-somal RNA from retinas contralateral (lane 1) or ipsilateral(lane 2) to optic nerve crush. (C) RNase protection assayusing a 467 nt 32P-labeled probe containing the 3r-most 423nt sequence of bd-X\,, resulting in a protected fragment of423 nt. Lane order: 1, native probe; 2, probe incubated with6 ^g yeast tRNA; 3, probe incubated with 6 yug RNA fromcontrol retinas; 4, probe incubated with 6 /zg RNA fromretinas 1 day after optic nerve crush; 5, probe incubatedwith 6 jug RNA from retinas 4 days after crush.



cells. In the central nervous system, embryonic neu-rons express both bd-2 and &cZ-xL

16'17'39 but adult neu-rons predominantly express bd-xL

17 The importanceof bd-x for development of the central nervous systemwas supported by studies of transgenic mice homozy-gous for disrupted bd-x genes.40 These mice have wide-spread apoptosis of central neurons during develop-ment, dying at embryonic day 13, in contrast totransgenic mice with disrupted bd-2 genes, which haveapparently normal central nervous systems41'42 and de-velop loss of facial motor nucleus, dorsal root gan-glion, and sympathetic ganglion neurons only afterthe period of developmental cell death.43

Our data demonstrate that bd-xL is the predomi-nant bd-2 family member expressed in the retina ofadult rats. This supports the concept that the neuralretina, a central nervous system structure, has a pat-tern of bd-2 family member expression similar to thatof other central neurons. The degenerate PCR ap-proach used in these studies makes unlikely (but doesnot exclude) the possibility that another moleculesharing the BH1 and BH2 homology domains is thepredominant member of the bd-2 family in the retina.Genes coding for molecules lacking these domainswould not have been detected by our methodology(e.g., that encoded by the death-inducing moleculebik, which lacks BH1 and BH2 and interacts with theprotein products of bd-2 and bd-x through its BH3domain).11 Significant expression of mRNA for thedeath-inducing molecule bd-xs, the protein productof which lacks BH1 and BH2 domains, was excludedby the absence of the appropriate size species whenprimers to the full-length bd-xL were used. Further-more, genes such as bax, which we have amplified byspecific RT-PCR from rat retina (Levin and Spieldoch,unpublished data), might not have been detected bythe primers used in this study because of their degen-erate design.

In addition, we found that bd-x^ expression ismodulated in response to retinal ganglion cell axo-tomy, decreasing after optic nerve crush. This suggeststhat bd-Xi. down regulation may be part of the pathwayby which retinal ganglion cells die by apoptosis afteraxotomy18~21 and is similar to the findings of Krajewskiet al,44 who found decreased expression of bd-xL afterischemic neuronal injury. The decrease in retinal gan-glion cell bd-xL is unlikely to be caused simply by neu-ronal loss, because levels decrease at 24 hours afteraxotomy, a time when ganglion cell loss is minimal,and there is a significant decrease in levels of bd-xL

mRNA at 4 days after axotomy, a time when there isrelatively little ganglion cell loss.18 More significantly,decreases in bd-xL levels are unlikely to reflect neu-ronal loss: we have demonstrated that another axo-tomy-induced gene, ceruloplasmin, increases in expres-

sion within retinal ganglion cells after optic nervecrush (Levin and Geszvain, unpublished data).

We detected minimal bd-2 expression in the nor-mal adult retina and observed no overt change 1 to 4days after optic nerve transection. Chen et al45 demon-strated increased retinal Bcl-2 expression 30 days afteroptic nerve transection in the adult rat, apparently inMuller cells, but did not look earlier. Our results sug-gest that neuronal expression of bd-2 in the retina isminimal, but we cannot exclude a small change inexpression early after ganglion cell axotomy that wasundetected by our methodology. The role of bd-2 inresponse to neuronal injury in other nervous systemareas is unclear because both increases and decreasesin bd-2 levels have been described after neuronal in-sults in a variety of paradigms, including ischemia andexcitotoxicity.44'46"50

Although the level of bd-xL mRNA is several timesthat of bd-2 in the adult retina, the latter retains capac-ity for inhibiting apoptosis of retinal ganglion cells.For example, retinal ganglion cells from mice overex-pressing bd-2 under the control of the neuron-specificenolase promoter are relatively resistant to axotomy-induced apoptosis.26'27'51 These mice appear to un-dergo less developmental cell death of the retinal gan-glion cell layer and facial nucleus,27'52 consistent witha role for bd-2 in preventing neuronal apoptosis in-duced by growth-factor deprivation. However, the rela-tive preservation of central neuronal numbers in bd-2-deficient mice43 indicates that other mechanismscan substitute for bd-2 expression in preventingapoptosis. A likely candidate is bd-xL. bd-xL can betterprevent apoptosis induced by IL-3 deprivation thanbd-2,5 the protein product of which may inhibit thedeath-inducing molecules Bax and Bak by binding bythe BH1, BH2, and BH3 domains,12'53 but has anti-death activity independent of Bax or Bak interac-tions.54 The mechanism by which the Bcl-x2 proteinproduct might modulate retinal neuronal survival iscontroversial and may include interactions with Bax,Bad, and Bak through BH1 and BH2 domains,12'53

blockage of apoptosis induced by free radicals,55 orsome other process.56 However, Bcl-xs, which lacks theBH1 and BH2 domains, induces cell death indepen-dent of heterodimerization with other bd-2 familymembers, suggesting a direct toxic effect.57 The recentsolution of the Bcl-xL crystal structure58 may help sug-gest structural similarities to other proteins associatedwith cell death and lead to a definitive role for Bcl-xL

and other family members in preventing neuronal celldeath.

Key Words

axotomy, bcl-2, bcl-x, optic nerve, retinal ganglion cell

References1. Bakhshi A, Jensen JP, Goldman P, et al. Cloning the



chromosomal breakpoint of t(14; 18) human lympho-mas: Clustering around JH on chromosome 14 andnear a transcriptional unit on 18. Cell. 1985; 41:899-906.

2. Cleary ML, SklarJ. Nucleotide sequence of a t(14;18)chromosomal breakpoint in follicular lymphoma anddemonstration of a breakpoint-cluster region near atranscriptionally active locus on chromosome 18. ProcNatAcadSci USA. 1985;82:7439-7443.

3. Tsujimoto Y, Gorham J, Cossman J, Jaffe E, Croce CM.The t(14;18) chromosome translocations involved inB-cell neoplasms result from mistakes in VDJ joining.Science. 1985; 229:1390-1393.

4. Tsujimoto Y, Cossman J, Jaffe E, Croce CM. Involve-ment of the bcl-2 gene in human follicular lymphoma.Science. 1985;228:1440-1443.

5. Boise LH, Gonzalez-Garcia M, Postema CE, et al. bcl-x, a &cZ-2-related gene that functions as a dominantregulator of apoptotic cell death. Cell. 1993; 74:597-608.

6. Oltvai ZN, Milliman CL, Korsmeyer SJ. Bcl-2 hetero-dimerizes in vivo with a conserved homolog, Bax, thataccelerates programmed cell death. Cell. 1993;74:609-619.

7. Yang E, Zha J, Jockel J, Boise LH, Thompson CB,Korsmeyer SJ. Bad, a heterodimeric partner for Bcl-XL and Bcl-2, displaces Bax and promotes cell death.Cell. 1995;80:285-291.

8. Chittenden T, Harrington EA, O'Connor R, et al. In-duction of apoptosis by the Bcl-2 homologue Bak. Na-ture. 1995; 374:733-736.

9. Farrow SN, White JH, Martinou I, et al. Cloning of abcl-2 homologue by interaction with adenovirus E1B19K. Nature. 1995;374-383.

10. Kiefer MC, Brauer MJ, Powers VC, et al. Modulationof apoptosis by the widely distributed Bcl-2 homo-logue Bak. Nature. 1995; 374:736-739.

11. Boyd JM, Gallo GJ, Elangovan B, et al. Bik, a noveldeath-inducing protein, shares a distinct sequencemotif with Bcl-2 family proteins and interacts with viraland cellular survival-promoting proteins. Oncogene.1995;11:1921-1928.

12. Yin XM, Oltvai ZN, Korsmeyer SJ. BH1 and BH2 do-mains of Bcl-2 are required for inhibition of apoptosisand heterodimerization with Bax. Nature. 1994;369:321-323.

13. Hockenbery DM, Oltvai ZN, Yin XM, Milliman CL,Korsmeyer SJ. Bcl-2 functions in an antioxidant path-way to prevent apoptosis. Cell. 1993; 75:241-251.

14. Zhai S, Yaar M, Doyle SM, Gilchrest BA. Nerve growthfactor rescues pigment cells from ultraviolet-inducedapoptosis by upregulating BCL-2 levels. Exp Cell Res.1996; 224:335-343.

15. Hockenbery DM, Zutter M, Hickey W, Nahm M, Kors-meyer SJ. BCL2 protein is topographically restrictedin tissues characterized by apoptotic cell death. ProcNat Acad Sci USA. 1991;88:6961-6965.

16. Merry DE, Veis DJ, Hickey WF, Korsmeyer SJ. bcb-2protein expression is widespread in the developingnervous system and retained in the adult PNS. Develop-ment. 1994; 120:301-311.

17. Gonzalez-Garcia M, Garcia I, Ding L, et al. bcl-x isexpressed in embryonic and postnatal neural tissuesand functions to prevent neuronal cell death. Proc NatAcad Sci USA. 1995;92:4304-4308.

18. Berkelaar M, Clarke DB, Wang YC, Bray GM, AguayoAJ. Axotomy results in delayed death and apoptosisof retinal ganglion cells in adult rats. / Neurosci.1994; 14:4368-4374.

19. Garcia-Valenzuela E, Gorczyca W, Darzynkiewicz Z,Sharma SC. Apoptosis in adult retinal ganglion cellsafter axotomy. / Neurobiol. 1994; 25:431-438.

20. Quigley HA, Nickells RW, Kerrigan LA, Pease ME,Thibault DJ, Zack DJ. Retinal ganglion cell death inexperimental glaucoma and after axotomy occurs byapoptosis. Invest Ophthalmol Vis Sci. 1995;36:774-786.

21. Levin LA, Louhab A. Retinal ganglion cell apoptosis inanterior ischemic optic neuropathy. Arch Ophthalmol.1996; 114:488-491.

22. Mey J, Thanos S. Intravitreal injections of neuro-trophic factors support the survival of axotomized reti-nal ganglion cells in adult rats in vivo. Brain Res.1993;602:304-3l7.

23. Castillo BJ, del CM, Breakefield XO, et al. Retinalganglion cell survival is promoted by genetically modi-fied astrocytes designed to secrete brain-derived neu-rotrophic factor (BDNF). Brain Res. 1994;647:30-36.

24. Mansour-Robaey S, Clarke DB, Wang YC, Bray GM,Aguayo AJ. Effects of ocular injury and administrationof brain-derived neurotrophic factor on survival andregrowth of axotomized retinal ganglion cells. ProcNatl Acad Sci USA. 1994;91:1632-1636.

25. Weibel D, Kreutzberg GW, Schwab ME. Brain-derivedneurotrophic factor (BDNF) prevents lesion-inducedaxonal die-back in young rat optic nerve. Brain Res.1995; 679:249-254.

26. Cenni MC, Bonfanti L, Martinou JC, Ratto GM, Stret-toi E, Maffei L. Long-term survival of retinal ganglioncells following optic nerve section in adult bcl-2transgenic mice. Eur J Neurosci. 1996;8:1735-1745.

27. Bonfanti L, Strettoi E, Chierzi S, et al. Protection ofretinal ganglion cells from natural and axotomy-in-duced cell death in neonatal transgenic mice overex-pressing bcl-2. J Neurosci. 1996; 16:4186-4194.

28. Garcia I, Martinou I, Tsujimoto Y, Martinou JC. Pre-vention of programmed cell death of sympathetic neu-rons by the bcl-2 proto-oncogene. Science. 1992;258:302-304.

29. Allsopp TE, Wyatt S, Paterson HF, Davies AM. Theproto-oncogene bcl-2 can selectively rescue neuro-trophic factor-dependent neurons from apoptosis.Cell. 1993; 73:295-307.

30. Batistatou A, Merry DE, Korsmeyer SJ, Greene LA.Bcl-2 affects survival but not neuronal differentiationof PC12 cells. J Neurosci. 1993; 13:4422-4428.

31. Chomczynski P, Sacchi N. Single-step method of RNAisolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem. 1987; 162:156-159.

32. Schlamp CL, Nickells RW. Light and dark cause a shiftin the spatial expression of a neuropeptide-processing



enzyme in the rat retina. / Neurosci. 1996; 16:2164-2171.

33. Harland RM. In situ hybridization: An improvedwhole-mount method for Xenopus embryos. MethodsCellBiol. 1991; 36:685-695.

34. Li HS, YangJM, Jacobson RD, Pasko D, Sundin O. Pax-6 is first expressed in a region of ectoderm anterior tothe early neural plate: Implications for the stepwisedetermination of the lens. Dev Biol. 1994; 162:181-94.

35. Sentman CL, Shutter JR, Hockenbery D, KanagawaO, Korsmeyer SJ. bcl-2 inhibits multiple forms ofapoptosis but not negative selection in thymocytes.Cell. 1991; 67:879-888.

36. Vaux DL, Cory S, Adams JM. Bcl-2 gene promoteshaemopoietic cell survival and cooperates with c-mycto immortalize pre-B cells. Nature. 1988; 335:440-442.

37. Sedlak-TW, Oltvai ZN, Yang E, et al. Multiple Bcl-2family members demonstrate selective dimerizationswith Bax. Proc Natl Acad Sci USA. 1995;92:7834-7838.

38. Zha H, Aime-Sempe C, Sato T, Reed JC. Proapoptoticprotein Bax heterodimerizes with Bcl-2 and homodi-merizes with Bax via a novel domain (BH3) distinctfrom BH1 and EH2.JBiol Chem. 1996;271:7440-7444.

39. LeBrun DP, Warnke RA, Cleary ML. Expression of bcl-2 in fetal tissues suggests a role in morphogenesis. AmJPath. 1993; 142:743-753.

40. Motoyama N, Wang F, Roth KA, et al. Massive celldeath of immature hematopoietic cells and neuronsin Bcl-x-deficient mice. Science. 1995;267:1506-1510.

41. Veis DJ, Sorenson CM, Shutter JR, Korsmeyer SJ. Bcl-2- deficient mice demonstrate fulminant lymphoidapoptosis, polycystic kidneys, and hypopigmentedhair. Cell. 1993;75:229-240.

42. Nakayama K, Nakayama K, Negishi I, Kuida K, SawaH, Loh DY. Targeted disruption of Bcl-2 alpha betain mice: Occurrence of gray hair, polycystic kidneydisease, and lymphocytopenia. Proc Natl Acad Sci USA.1994;91:3700-3704.

43. Michaelidis TM, Sendtner M, Cooper JD, et al. Inactiva-tion of bcl-2 results in progressive degeneration of moto-neurons, sympathetic and sensory neurons during earlypostnatal development. Neuron. 1996; 17:75-89.

44. Krajewski S, Mai JK, Krajewska M, Sikorska M, Mossa-kowski MJ, Reed JC. Upregulation of bax protein lev-els in neurons following cerebral ischemia. J Neurosci.1995; 15:6364-6376.

45. Chen ST, Garey LJ, Jen LS. Bcl-2 proto-oncogene pro-tein immunoreactivity in normally developing and ax-otomised rat retinas. Neurosci Lett. 1994; 172:11-14.

46. Montpied P, Weller M, Paul SM. N-methyl-D-aspartatereceptor agonists decrease protooncogene bcl-2mRNA expression in cultured rat cerebellar granuleneurons. Biochem Biophys Res Comm. 1993; 195:623-629.

47. Shimazaki K, Ishida A, Kawai N. Increase in bcl-2 onco-protein and the tolerance to ischemia-induced neu-ronal death in the gerbil hippocampus. Neurosci Res.1994; 20:95-99.

48. Chen J, Zhu RL, Nakayama M, et al. Expression ofthe apoptosis-effector gene, Bax, is up-regulated invulnerable hippocampal CA1 neurons followingglobal ischemia./Neurochem. 1996;67:64-7l.

49. Jaw SP, Su DD, Truong DD. Expression of Bcl-2 andBax in the frontoparietal cortex of the rat following-cardiac arrest. Brain Res Bull. 1995;38:577-580.

50. Gillardon F, Wickert H, Zimmermann M. Up-regula-tion of bax and down-regulation of bcl-2 is associatedwith kainate-induced apoptosis in mouse brain. Neu-rosci Lett. 1995; 192:85-88.

51. BurneJF, Staple JK, Raff MC. dial cells are increasedproportionally in transgenic optic nerves with in-creased numbers of axons. / Neurosci. 1996; 16:2064-2073.

52. Martinou JC, Dubois-Dauphin M, Staple JK, et al.Overexpression of BCL-2 in transgenic mice pro-tects neurons from naturally occurring cell deathand experimental ischemia. Neuron. 1994; 13:1017-1030.

53. Chittenden T, Flemington C, Houghton AB, et al. Aconserved domain in Bak, distinct from BH1 and BH2,mediates cell death and protein binding functions.EMBOJ. 1995; 14:5589-5596.

54. Cheng EH, Levine B, Boise LH, Thompson CB, Hard-wick JM. Bax-independent inhibition of apoptosis byBcl-XL. Nature. 1996;379:554-556.

55. Fang W, Rivard JJ, Ganser JA, et al. Bcl-xL rescuesWEHI 231 B lymphocytes from oxidant-mediateddeath following diverse apoptotic stimuli. J Immunol.1995; 155:66-75.

56. Shimizu S, Eguchi Y, Kosaka H, Kamiike W, MatsudaH, Tsujimoto Y. Prevention of hypoxia-induced celldeath by Bcl-2 and Bcl-xL. Nature. 1995;374:811-813.

57. Minn AJ, Boise LH, Thompson CB. Bcl-x(S) antago-nizes the protective effects of Bcl-x(L). J Biol Chem.1996;271:6306-6312.

58. Muchmore SW, Sattler M, Liang H, et al. X-ray andNMR structure of human Bcl-xL, an inhibitor of pro-grammed cell death. Nature. 1996;381:335-341.


identification of the bcl-2 family of genes in the rat retina

Documents