Role of INK4a locus in normal eye development and cataract genesis.

Abstract

The murine INK4a locus encodes the critical tumor suppressor proteins, p16(INK4a) and p19(ARF). Mice lacking both p16(INK4a) and p19(ARF) (INK4a-/-) in their FVB/NJ genetic backgrounds developed cataracts and microophthalmia. Histopathologically, INK4a-/- mice showed defects in the developmental regression of the hyaloid vascular system (HVS), retinal dysplasia, and cataracts with numerous vacuolations, closely resembling human persistent hyperplastic primary vitreous (PHPV). Ocular defects, such as retinal fold and abnormal migration of lens fiber cells, were observed as early as embryonic day (E) 15.5, thereby resulting in the abnormal differentiation of the lens. We also found that ectopic expression of p16(INK4a) resulted in the induction of gammaF-crystallin, suggesting an important role of INK4a locus during mouse eye development, and also providing insights into the potential genetic basis of human cataract genesis.

Full text

Page 1

Role of INK4a locus in normal eye development
and cataract genesis
Cheolho Cheonga,1, Young Hoon Sunga,1, Jaehoon Leea,1, Yoon Sik Choia,
Jaewhan Songb, Changwon Keec,**, Han-Woong Leea,*
aDepartment of Molecular Cell Biology, Samsung Biomedical Research Institute, Molecular Therapy Research Center,
Sungkyunkwan University School of Medicine, 300 Chonchon-Dong, Changan-Gu, Suwon 440-746, Republic of Korea
bFaculty of Life Sciences and Engineering, Sungkyunkwan University, Suwon 440-746, Republic of Korea
cDepartment of Ophthalmology, Samsung Medical Center, Sungkyunkwan University School of Medicine,
Seoul 135-710, Republic of Korea
Received 25 May 2005; received in revised form 1 February 2006; accepted 28 February 2006
Available online 18 April 2006
Abstract
ThemurineINK4alocusencodesthecriticaltumorsuppressorproteins,p16INK4aandp19ARF.Micelackingbothp16INK4aandp19ARF(INK4a?/?)
in their FVB/NJ genetic backgrounds developed cataracts and microophthalmia. Histopathologically, INK4a?/?mice showed defects in the
developmental regression of the hyaloid vascular system (HVS), retinal dysplasia, and cataracts with numerous vacuolations, closely resembling
humanpersistenthyperplasticprimaryvitreous(PHPV).Oculardefects,suchasretinalfoldandabnormalmigrationoflensfibercells,wereobserved
as early as embryonic day (E) 15.5, thereby resulting in the abnormal differentiation of the lens. We also found that ectopic expression of p16INK4a
resultedintheinductionofgF-crystallin,suggestinganimportantroleofINK4alocusduringmouseeyedevelopment,andalsoprovidinginsightsinto
the potential genetic basis of human cataract genesis.
# 2006 Elsevier Ireland Ltd. All rights reserved.
Keywords: p16INK4a; p19ARF; HVS; PHPV; Cataract; Crystallin
1. Introduction
Mammalian eye development requires a tight regulation of
cell cycle exit and differentiation pathways (Graw, 1999).
Cyclin-dependent kinases (Cdks) integrate proliferative signals
with the mechanical aspects of cell duplication. Orderly
progression through the cell cycle involves the coordinated
activationoftheCdks,which occursupontheirassociationwith
CDK-activating kinase and phosphorylates target substrates,
such as pRb, p107, and p130 which are members of the ‘pocket
protein’ family (Morgan, 1995). On the other hand, CDK
inhibitors (CKIs) act in opposition to the Cdks.
To date, there are two families of CKIs: the p21 family
(CIP1/Waf1) and the p16 family. The p21 family (p21,
p27KIP1, andp57KIP2) inhibitsvarious kinasesinvolvedinthe
G1/S transition, whereas the p16 family (p15, p16, p18, and
p19) specifically inhibits Cdk4 and Cdk6 (Harper and Elledge,
1996). Studies on the patterns of expression and biochemical
activities of CKIs during mammalian development suggest that
these proteins function as the primary effectors of signaling
pathways that control cell cycle exit, a critical event in the
process of differentiation. For example, both permanent
withdrawal from the cell division cycle and terminal
differentiation of many cell types, including the lens fiber
cells of the eye, are known to be induced by the CKI pathway
(Matsuoka et al., 1995).
Analyses of embryonic lens development have provided us
with many insights as to how cell cycle arrest and
differentiation are coordinated. The lens is composed of
differentiated lens fiber cells, capped on the anterior surface by
a layer of immature, mitotic epithelial cells (McAvoy, 1980;
Piatigorsky,1981).Inthisregard,loss ofpRb function inmouse
www.elsevier.com/locate/mechagedev
Mechanisms of Ageing and Development 127 (2006) 633–638
* Corresponding author. Present address: Department of Biochemistry,
Yonsei University, Seoul 120-749, Republic of Korea. Tel.: +82 2 2123 5698;
fax: +82 362 8096.
** Co-corresponding author. Fax: +82 2 3410 0074.
E-mail addresses: cwkee@smc.samsung.co.kr (C. Kee),
hwl@yonsei.ac.kr (H.-W. Lee).
1These authors contributed equally to this work.
0047-6374/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.mad.2006.02.010

Page 2

embryonic lens is associated with unchecked proliferation,
impaired expression of differentiation markers, and inappropri-
ate apoptosis in lens fiber cells (Morgenbesser et al., 1994).
The role of the p21 CKI families in mammalian lens
development is well-documented (Gao et al., 1999). Perturba-
tion in the expression of p57KIP1in the lens results in the failure
of fiber cell elongation and cataracts. Furthermore, the ocular
defects associated with the p57KIP1mutant lens are exacerbated
in the case of the p27-deficient genetic background, suggesting
that the inability of lens fiber cells to undergo cell cycle arrest
leads to defects in differentiation, most notably, elongation and
crystallin expression (Zhang et al., 1998). However, the
functions of p16 CKI families such as INK4a, INK4b, INK4c,
and INK4d during embryonic eye development remain fairly
nebulous. Recently, p19ARF, an alternative reading frame
proteinoftheINK4alocus,wasreportedtoplayacriticalrolein
HVSregression during the first postnatal week (McKeller etal.,
2002). In the absence of p19ARF, failed HVS regression appears
to cause a pathological process that resembles persistent
hyperplastic primary vitreous, a developmental human eye
disease thought to have a genetic basis.
Here, we report that mice which are doubly deficient for
p16INK4aand p19ARF(INK4a?/?) display similar postnatal eye
phenotypes of p19ARF-defcient mice, including failed HVS
regression, retinal dysplasia, persistent hyperplastic primary
vitreous (PHPV), and cataracts. Notably, these mice also
showed defects during embryonic eye development, which was
not observed in the mice deficient only for p19ARF. Likewise,
p16INK4abut not p19ARF, augmented the activity of the
luciferase reporter driven by the gF-crystallin promoter in the
transient transfection assay. Our results suggest that p16INK4a
together with p19ARFplays an important role in both normal
eye development and cataract genesis.
2. Materials and methods
2.1. Mice, ophthalmologic test, and blood chemistry
INK4a-deficient mice were previously described (Serrano et al., 1996). The
mouse cohorts were generated by the mating of INK4a+/?animals after back-
crossing more than 10 generations to FVB/NJ (Sung et al., 2005). All animals
were housed in the SPF facility and maintained on a 12-h light/12-h darkness
cycle. All experiments were carried out in an AAALAC-certified facility in
compliance with animal policies by Sungkyunkwan University School of
Medicine. Ocular photography and slit-lamp photography were conducted at
the Department of Ophthalmology of the Samsung Medical Center (SMC).
Briefly, micewerelightlyanesthetizedwithavertinand pupilsweredilated with
cyclomydril (Alcon) to facilitate examination of the posterior segments of the
eye. For theglucose tolerance test, micewere made to fast for 16 h, followed by
intraperitoneal glucose injections (2 g/kg body weight). Blood was withdrawn
from the tail, and glucose levels were determined using the One Touch Basic
Glucometer (Lifescan Canada Ltd.) 0, 30, 60, 90, and 120 min after the
administration of glucose.
2.2. Histology of adult and embryonic eyes
To evaluate ocular size and histopathology, eyes from adult mice and
embryos were removed with circular forceps immediately after the mice were
sacrificed. Extraocular muscles were further removed and fixed in 4% paraf-
ormaldehyde in PBS for 24 h, then embedded in paraffin by orientating lens
according to optic nerve side; 5- to 10-mm-thick sections from the blocks were
used for hematoxylin and eosin (H&E) staining, or for 40,60-diamidino-2-
phenylindole (DAPI) (Sigma) for the evaluation of nuclear density in the
vitreous. Images were obtained using a fluorescent microscope equipped with
a digital camera (Zeiss).
2.3. Reverse transcription (RT)-PCR analysis of gene expression, cell
culture, cDNA constructs, and luciferase assay
TotalRNAfrommouseembryonicfibroblasts(MEF)andE15.5lensofboth
wild type and INK4a?/?embryos was extracted using Trizol reagent (Life
Technology, MD), according to the manufacturer’s recommendations. cDNAs
were prepared using Superscript II RT (Life Technology, MD), using 1 mg of
total RNA. PCR was performed by utilizing primer sets specific for p16INK4a,
p19ARF, and hprt (Quelle et al., 1995; Nishiguchi et al., 1998). The PCR
products for p16INK4a(509 bp), p19ARF(566 bp), and hprt (702 bp) were run on
1% agarose gel and photographed. For the reporter assay, NIH3T3 cells were
cultured in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine
serum, and transfected with Effectin (Qiagen). Mouse cDNAs encoding
p16INK4aor p19ARFwere cloned into pcDNA3-HA (Invitrogen) and used for
transfection, coupled with the gF-crystallin promoter-luciferase construct (Tini
et al., 1993). The luciferase assay was performed between 36–48 h after
transfection with reagents purchased from Promega (Madison, WT) according
to manufacturer’s recommendations.
3. Results
3.1. INK4a?/?mice develop cataract
We observed that the eyes of most INK4a?/?mice were
noticeably smaller than those of their wild type or INK4a+/?
littermates and the ocular photograph further confirmed the
development of microopthalmia (Fig. 1A). Slit-lamp ocular
photograph revealed that the majority of INK4a?/?eyes
exhibited posterior cataracts (Fig. 1A). Ninety-four percent of
C. Cheong et al./Mechanisms of Ageing and Development 127 (2006) 633–638634
Fig. 1. INK4a-deficient mice develop cataracts. (A) Right (R) and left (L) eyes from 5–6-month-old wild type (WT) and INK4a?/?mice are photographed. Ocular
photograph reveals unilateral (#1) or bilateral cataracts (#2). Arrows indicate ocular opacity. Slit-lamp photograph further confirms that these mice (#1 and #2) have
cataracts. (B) Glucose tolerance test does not show any significant difference between wild type and INK4a?/?mice.

Page 3

INK4a?/?mice tested (65 out of 69 examined) exhibited either
unilateral (Fig. 1A, #1) or bilateral cataracts (Fig. 1A, #2)
withoutanysignificantdifferencebetweenbothsexes(Table1).
Incontrast,thewild typemice observedupto72-week-old (114
mice examined) did not show any noticeable ocular defects
(Table 1). Diabetes is known as a major systemic disorder
associated with secondary cataracts both human and mice (Lee
et al., 1995). Therefore, we then tested that diabetes is
associated with cataracts observed in the INK4a?/?mice.
However, glucose tolerance tests revealed no significant
differences between wild type and INK4a?/?mice (Fig. 1C),
implying that defects in the eyemight be directly attributable to
the genetic deficiency of INK4a locus.
3.2. Histopathological analysis of INK4a?/?eyes
In order to assess the eye histopathology, we stained ocular
sections from wild type, INK4a+/?, and INK4a?/?mice with
hematoxylin and eosin (H&E). In wild type mice, the lens
capsule encloses the lens epithelium at the periphery anteriorly
and the cortex and nucleus are located progressively towards
the center of the lens (Fig. 2A; Francis et al., 1999). The
epithelialcellsare mitotic, anddifferentiate terminallyinto lens
fibers, which make up the cortex and the nucleus. Thus, the
oldest lens fibers are found in the nucleus, whereas the more
recently formed lens fibers constitute the cortex (Francis et al.,
1999). The retrolental vitreous in the newborn mouse eye
contains elements of the hyaloid vascular system (HVS), and is
normally seen to regress during the first 2 weeks of postnatal
development (Fig. 2A, arrow), where the hyaloid cavity forms
(Ito and Yoshioka, 1999). On the other hand, histological
analysis revealed the presence of retrolental masses of cells in
the posterior vitreous cavity of INK4a?/?mice (Fig. 2F–I), but
not in wild type or heterozygous mutant mice (data not shown).
INK4a?/?mice tested at up to 12 months of age still retained
mature hyaloid artery (HA)-like structures, suggesting the
failure of HVS regression in these mice (data not shown). In
support, the deficiency of p19ARF, one of the two proteins
C. Cheong et al./Mechanisms of Ageing and Development 127 (2006) 633–638635
Table 1
Ocular defects (%) of mice
Age (weeks)
<1212–20 20–3232–4848–7272<
Total
Eye defect (%) (defective mice/total)
INK4a?/?
Wild type
93 (13/14)
0 (0/7)
88 (23/26)
0 (0/29)
100 (23/26)
0 (0/21)
100 (24/24)
0 (0/4)
100 (2/2)
0 (0/53)
(0/0)
13 (2/16)
94 (65/69)
1.5 (2/130)
INK4a?/?:formales,27micehadunilateral(59%,16/27)orbilateraldefects(41%,11/27);forfemales,38micehadunilateral(42%,16/38)orbilateraldefects(58%,
22/38). Wild type: for males, no mice had lateral defects; for females, two mice had unilateral (1.5%, 1/68) or bilateral defects (1.5%, 1/68).
Fig. 2. Eye defects in adult INK4a-deficient mice. (A–M) Eyes were surgically removed from 5–6-month-old wild type and INK4a?/?mice exhibiting cataracts, and
then subjected to H&E staining. Wild type mice exhibit intact lens architecture and fully regressed HVS, and display no remnants (A, arrow). The lens capsules are
intact from the equatorial to the posterior region (B–D, square bracket), and the structure of retinal layer shows normal (E). In contrast, the lenses’ architectures are
abnormal, showing tumbler-doll like shapes and cataractous vacuoles (F and I). The posterior region of INK4a?/?lens is fused with the retinal layers, and retinal
dysplasia is evident (F, Q, H, and I, arrow). (J–M); insets in (F–I).

Page 4

encoded in the INK4a locus, also resulted in the failure of HVS
regression in mice (McKeller et al., 2002). Majority of INK4a?/
?mice were noted with eyes exhibiting a tumbler-doll shape
and cataractous vacuoles (Fig. 2F and I). Furthermore, the
INK4a?/?neuroretinawasmarkedlyabnormal,asevidencedby
the presence of retinal folds (Fig. 2G and H), rosette-like
arrangements of dysplastic photoreceptor cells (Fig. 2K and L),
and the progressive physical attachment of the retrolental mass
to the neuroretina (Fig. 2F–I). Taken together, the eye
phenotype of adult INK4a?/?mice closely recapitulates the
phenotypes of p19ARF-deficient mice, which have been
reported to exhibit symptoms similar to the pathological
conditionsofhumanPHPVpatients(Goldberg,1997;McKeller
et al., 2002).
3.3. INK4a?/?mice have embryonic eye defects
Ocular abnormalities in our INK4a?/?mice were evident
when newborns opened their eyes, which prompted us to
examine their embryonic eye development. Histological
studies revealed that eyes of INK4a?/?mice were character-
ized by abnormal lens development. In normal eye develop-
ment, the primary vitreous (retrolenticular membrane)
disappears by E14 (Fig. 3A; Kim et al., 2002). However, the
primary vitreous membrane of INK4a?/?mice failed to
regress, and even developed, to fuse to the posterior capsule of
the lens (Fig. 3B, arrow), which was not observed in any of
control animals. Moreover, microophthalmia was apparent in
the INK4a?/?embryos (Fig. 3B). In order to further evaluate
the proliferation and differentiation status of lens cells, DAPI
stainingofeyesectionswasperformedinE15.5embryos.Most
of the lens cell nuclei in the wild type animals had migrated,
from the peripheral regions towards the center. In contrast, we
observed higher nucleus density near the posterior surface of
the lens and vitreous cavity, in the lenses of E15.5 INK4a?/?
embryos (Fig. 3C–E). These results suggest that the
proliferation and differentiation processes of the INK4a?/?
embryonic lens cells are profoundly impaired, which explains
the ocular defects observed in the adulthood of the animals.
Notably, embryonic eye defects were not reported in mice
deficient of p19ARF(McKeller et al., 2002). This led us to
compare the expression patterns of p16INK4aand p19ARFin
wild type embryos. When we assessed the mRNA levels of
p16INK4aandp19ARFintheembryoniclens(E15.5)byRT-PCR
analysis, interestingly p19ARFwas not detected in E15.5 as
previouslyreported (Zindy et al., 1997; McKeller et al., 2002).
However,p16INK4ahadbeenexpressed(Fig.3F),implyingthat
the observed phenotype of INK4a?/?eyes might be partially
attributable to the loss of p16INK4a.
3.4. The expression of gF-crystallin is regulated by
p16INK4a
The dysregulation or mutation of various crystallin genes,
which are the hallmark of normal lens differentiation alsocause
cataracts (Javitt et al., 1996). In this regard, p57KIP1-deficient
lenses exhibit reduced expression of b- and g-crystallins
(Zhang et al., 1998). Since we had already observed abnormal
lens development in the INK4a?/?mice (Fig. 3A–E), it was
postulated that the expression of the INK4a locus can induce
C. Cheong et al./Mechanisms of Ageing and Development 127 (2006) 633–638636
Fig. 3. Embryonic eye defects occurring in INK4a-deficient mice. E15.5 eyes from wild type (WT) (A) and INK4a?/?embryos (B). H&E staining reveals the failed
regressionoftheprimaryvitreous,retinalfolds,andabnormalnucleardensityintheposteriorregionsofINK4a?/?eyes.DAPIstainingtoassessthenucleardensityof
wild type (C) and INK4a?/?eyes (D) at E15.5. The neuroretina shows a prominent DAPI staining pattern compared with wild type (D, arrow). (E) Quantitative
representation of nuclear density in the dotted line region in (C and D). (F) Expression test of p16INK4aand p19ARFin the wild type and INK4a?/?MEFs at passage 5
and lenses at E15.5. RT-PCR analysis reveals the expression of p16INK4a, but not p19ARFin lenses of E15.5 p16 for p16INK4a; p19 for p19ARF; hprt for hypoxanthine-
guanine phosphoribosyl transferase.

Page 5

crystallins. To test this possibility, we examined the transcrip-
tional activity of gF-crystallin promoter-luciferase (Tini et al.,
1993). Notably, the ectopic expression of p16INK4aresulted in
the induction of gF-crystallin, whereas the ectopic expression
of p19ARFelicited no significant effects (Fig. 4), suggesting the
novel role of INK4a locus in the expression of crystallin.
Indeed, embryonic lens from INK4a?/?mice showed reduced
mRNAlevelofgF-crystallinwhenassayedbyRT-PCRanalysis
(Kim et al., 2002).
4. Discussion
Our study demonstrates that the mouse INK4a locus is
required for normal embryonic eye development, and
regression of the HVS during the postnatal maturation of
the primary vitreous of the eye. Previously, p19ARF, one of the
productsofthemouseINK4alocus,wasshowntoplayacrucial
role in preventing human PHPV based on pathological
condition of the eye with the following characteristics: (i)
the presence of retrolental tissues; (ii) the attachment of the
retrolentral tissues to the inner neuroretina; (iii) retrolental
tissue-induced traction on the neuroretina, which causes
neuroretinal detachment from the retinal pigment epithelium;
(iv)cellulardisorganizationandotherdysplasticchangesinthe
neuroretina; (v) posterior lens capsule destruction by the
retrolental mass; and (vi) cataractous degeneration of the lens
(McKeller et al., 2002). Our INK4a?/?
reproduced these phenotypes, further supporting the notion
of a crucial role of p19ARF(McKeller et al., 2002). Although
the expressions of both p16INK4aand p19ARFwere reported to
besuppressedduringembryogenesisbyNorthernblotanalysis
of total RNA obtained from the whole embryos, each cell type
or organ, especially the eye, was not fully tested for the
expression of these two gene products (Zindy et al., 1997).
Notably, apoptosis in the Rb-deficient lens was significantly
decreasedintheINK4a?/?background,clearlyunderliningthe
functionoftheINK4alocusduringembryogenesis(Pomerantz
et al., 1998). However, p19ARFalone did not influence
apoptosis or inappropriate S phase entry significantly in Rb
mutant lens (Rb?/?p19ARF?/?) (Tsai et al., 2002). Thus, the
loss of p16INK4amight have conferred a growth advantage to
these cells in the absence of Rb (Pomerantz et al., 1998),
further suggesting separable roles of two gene products in
INK4a locus during embryonic eye development. After we
submittedthismanuscript,wenotedthatthereisareportofthe
embryonic expression of p19ARFand its restricted expression
in the developing vitreous in mouse embryonic eye as early as
E11.5(Section5).Inaddition,thep19ARF-deficientembryonic
eyes manifested similar postnatal eye defects of p19ARF-
deficient mice (McKeller et al., 2002; Martin et al., 2004;
Section 5). However the embryonic eye of p19ARF-deficient
mice did not display alteration of the morphology of the
fibroblast-likecellsorabnormalityofthelensandneuroblastic
tissue (Section 5). Along these lines, we and others could not
detect the expression of p19ARFin the embryonic lens while
p16INK4acould be detected (Fig. 3F). Consistent with this, the
expression of crystallin, the hall marker of lens cell
differentiation is affected by INK4a-deficiency, which is
presumably caused by the deficiency of p16INK4anot by
p19ARF(Fig. 4).
Thus, this embryonic phenotype observed here might be due
to the lack of both p16INK4aand p19ARFduring embryonic eye
development (Fig. 3F). Therefore, our results together with
others further differentiate the role of p16INK4aduring
embryonic lens development, from the role of p19ARFin the
developing vitreous of mouse embryos.
In this regard, our p16INK4aand p19ARFdouble deficient
mouse model is not useful enough to fully dissect the role of
each gene product during the embryonic eye development.
Nevertheless, our results provide potential evidence that two
overlapping genes at INK4a locus interact with each other
genetically and the p16 CKI is also involved in normal lens
development, where perturbation of these factors results in the
formation of cataracts.
mice clearly
5. Note added in proof
After revision of our manuscript, we found a publication
regarding the embryonic eye phenotype of p19ARFknockout
mice (Silva et al., 2005).
C. Cheong et al./Mechanisms of Ageing and Development 127 (2006) 633–638 637
Fig. 4. p16INK4aregulates gF-crystallin. gF-crystallin promoter-luciferase
constructs were transfected with the indicated plasmids into NIH3T3, which
lacks both p16INK4aand p19ARF(A). The data represent the firefly luciferase
activity normalized to that of renilla luciferase. ASC-2 is a previously known
positive regulatorofgF-crystallin(Kimetal., 2002).Notethatp16INK4a,but not
p19ARF, induces gF-crystallin promoter activity. Error bars indicate ? S.D. (B)
Expression of gF-crystallin in the lens of E15.5. WT for wild type. RT-PCR
products from gF-crystallin (252 bp) and hprt, hypoxanthine-guanine phos-
phoribosyl transferase (702 bp).