Differential expression and functionally co-operative roles for the retinoblastoma family of proteins in epidermal differentiation

Di�erential expression and functionally co-operative roles for theretinoblastoma family of proteins in epidermal di�erentiation

Jesu s M Paramio1,2, Sonia Laõ n2, Carmen Segrelles1, E Birgit Lane2 and Jose L Jorcano1

1Cell and Molecular Biology Program, CIEMAT, E-28040 Madrid, Spain; 2CRC Cell Structure Group, DepartmentAnatomy and Physiology, Medical Sciences Institute, University of Dundee, Dundee DD1 4HN, UK

Terminal di�erentiation requires cell cycle withdrawal,suggesting the involvement of negative cell cyclecontrollers in the process. We have analysed theinvolvement of the retinoblastoma family of proteins(pRb, p107 and p130) in epidermal proliferation anddi�erentiation. These proteins play key roles asinhibitors of cell cycle progression and are involved inmuscle and neuron di�erentiation. We found that duringin vitro di�erentiation of human HaCaT keratinocytes,pRb, p107 and p130 are sequentially expressed, incontrast to the co-expression observed during cell cycleprogression in the same cells. Immuno¯uorescencestudies on skin sections revealed the presence of pRband p107 in basal and suprabasal cell layers, whilst p130is restricted to cells already committed to di�erentiationin the suprabasal compartments. To explore thefunctional signi®cance of the di�erential expression ofthese proteins, transfection experiments were performedin HaCaT keratinocytes. We observed that the forcedover-expression of pRb, p107 or p130 individually did notinduce di�erentiation of the transfected cells. However,the co-transfection of pRb and p107 induced theexpression of early di�erentiation markers (keratink10) and triple transfectants pRb+p107+p130 ex-pressed markers representative of later stages ofepidermal di�erentiation (involucrin). Finally, we ob-served that these three proteins repress keratinocyteproliferation, although to a di�erent extent(p1074pRb5p130). These results indicate that themembers of the pRb family play speci®c, yet co-ordinated roles during epidermal di�erentiation, and thatthe ordered progression along the di�erent stages of thisprocess results from the e�ects of di�erent combinationsof these proteins.

Keywords: cell cycle; epidermal di�erentiation; retino-blastoma family

Introduction

The epidermis is the strati®ed epithelium covering thebody surface and plays an essential role as a protectivebarrier against the environment. In this tissue, cell lossis constitutively high and physical and chemical traumaare frequent, demanding an e�cient and robustrenewal system incorporating tightly controlled cell

division and di�erentiation mechanisms. Proliferationin the epidermis takes place predominantly in the basalcell layer, where keratinocytes synthesize cytoskeletalkeratins k5 and k14. On commitment to di�erentiation,they switch their keratin synthesis to k1 and k10,down-regulate integrin expression and migrate up outof the basal layer. Di�erentiating keratinocytes thenmove progressively upwards synthesizing new proteinssuch as involucrin, loricrin and ®laggrin towards thestratum corneum, from where they are ultimately shedafter a functional period as corni®ed anucleatesquames (reviewed in Fuchs, 1990; Leigh et al.,1994). The clear distinction between proliferative anddi�erentiated compartments, and the existence ofmarkers allowing the identi®cation of cells at di�erentstages, make the epidermis an attractive model inwhich to study the control of proliferation anddi�erentiation. Finally, the very high and increasingincidence of non-melanoma skin cancers in humanpopulations has also increased the general interest inanalysing the mechanisms controlling these processesand their possible alterations in pathological situations.Terminal di�erentiation is preceded by an arrest of

cell proliferation. Therefore, it is conceivable thatmolecules that regulate cell cycle progression are alsoinvolved in the regulation of cell di�erentiation. Theretinoblastoma protein (pRb) is the product of atumor suppressor gene which plays an essential roleregulating cell proliferation (reviewed in Weinberg,1995; Picksley and Lane, 1994; Wang et al., 1994).This function is associated with the ability of thismolecule to bind and inhibit members of the E2Ffamily of transcription factors (LaThangue, 1994;Nevins, 1992) and is regulated during the cell cycleby phosphorylation processes mediated by cyclin-dependent kinases (Weinberg, 1995). p107 and p130(Ewen et al., 1991; Hannon et al., 1993; Li et al.,1993) are two proteins structurally and functionallyrelated to pRb and also able to bind and speci®callyinhibit E2F members in a phosphorylation-dependentmanner (Cobrinik et al., 1993; Sardet et al., 1995;Vairo et al., 1995; Zhu et al., 1993, 1995).pRb has also been implicated in processes of cell

di�erentiation. In this context, the lethality observedin mice lacking functional pRb has been attributed todevelopmental alterations in neuronal and haemo-poyetic di�erentiation rather than to unbalancedproliferation or tumorigenesis (Lee et al., 1992; Jackset al., 1992; Clarke et al., 1992). Moreover, duringmouse development pRb is expressed in the differ-entiating compartments of a number of tissues(Szekely et al., 1992). Finally, pRb, and its relativesp107 and p130, have also been associated with theprocesses of myogenesis and muscle di�erentiation

Correspondence: JM ParamioReceived 23 October 1997; revised 3 April 1998; accepted 6 April1998

Oncogene (1998) 17, 949 ± 957 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00

http://www.stockton-press.co.uk/onc

(Gu et al., 1993; Schneider et al., 1994; Shin et al.,1995; Throrburn et al., 1993; Kiess et al., 1995).Few data are available regarding the possible

involvement of these proteins in keratinocyte prolif-eration and di�erentiation. However, the reducedability to di�erentiate displayed by human keratino-cytes immortalized by the forced expression of viralproteins that bind and inactivate the pRb family, suchas T antigen (Taylor-Papadimitriou et al., 1982;Banks-Schlegel and Howley, 1983), HPV E7(McCance et al., 1988; Hudson et al., 1990) andE1A (Barrandon et al., 1989), together with therequirement of the HPV E7 protein in the etiologyof certain human skin pathologies (Zur Hausen,1991), strongly suggest the involvement of pRbmembers in epidermal proliferation and differentia-tion. This is reinforced by the alterations observed intransgenic mice in which the expression of these viralproteins (Missero et al., 1993; Arbeit et al., 1994;Auewarakul et al., 1994) or cyclin D1 (Robles et al.,1996) is targeted to the epidermis.In this study, we show that pRb, p107 and p130 are

sequentially expressed during in vitro di�erentiation ofhuman HaCaT epidermal keratinocytes, and that theyare distinctly localized on in vivo human skin.Transfection experiments demonstrate that the threeproteins have a di�erent capacity to inhibit keratino-cyte proliferation and that particular combinations,but not individual proteins, trigger speci®c stages ofepidermal di�erentiation.

Results

Expression of pRb and related proteins during in vitrodi�erentiation

To study the possible involvement of the pRb andpRb-related proteins in epidermal di�erentiation, weanalysed their expression during in vitro di�erentiationof human keratinocytes using HaCaT cells (humanimmortalized non-tumorigenic keratinocytes) as amodel (Boukamp et al., 1988, Paramio and Jorcano,1997). To avoid the use of chemical treatments and tobe able to discriminate cell cycle arrest and differentia-tion, we developed a method in which the differentia-tion process was induced by culturing the cells at highdensity in the absence of serum (see Materials andmethods). From a morphological point of view, after 4days of serum starvation, the cells had a normalappearance (Figure 1a), although they had undergonea complete cell cycle withdrawal, as demonstrated bytheir inability to incorporate base analogs into DNA(results not shown; see also Geng and Weinberg, 1993;Fahraeus et al., 1996). After 1 week, the progressiveformation of strati®ed domes was detected. Thesedomes were easily visible and occupied approximately5 ± 10% of the surface of the culture after 20 days ofserum deprivation (Figure 1b, arrows), at which timethese experiments were arrested. At this moment, theclonogenic capacity of the cultures had declined from80 ± 90% 4 days after serum withdrawal, to 10 ± 20%.Markers of epidermal di�erentiation, such as k10, werevery prominent at these domes, although they werealso detected in other areas of the culture showing lessprominent strati®cation (results not shown). Protein

extracts obtained from these cultures after di�erentperiods of serum starvation were used to analyse theexpression of keratin k10 and involucrin, as markers ofearly or more delayed epidermal di�erentiation,respectively (Rice and Green, 1979; Banks-Schlegeland Green, 1981). We observed (Figure 1c) a clearinduction of both markers during the course of theexperiments, although involucrin appeared later thank10 (Figure 1c). On the other hand, k14 expression,which is characteristic of basal undi�erentiatedkeratinocytes and used as a control, did not change.These data indicate that this culture system recapitu-lates many aspects of epidermal di�erentiation and canbe used to study this process temporally andbiochemically. The pattern of pRb, p107 and p130expression was analysed in HaCaT cells prior to serumremoval and at di�erent times during di�erentiation(Figure 1d). We observed that pRb level decreasesupon serum removal, increasing by day 8 anddecreasing thereafter to a basal level maintainedthroughout the di�erentiation period. On the otherhand, p107 followed a similar kinetics except that itwas induced by day 12, reached its maximum at day 16and decreased again at day 20. Finally, p130 expressionremained low until day 12 increasing thereafter, andremained at high level at day 16 and 20. These resultsdemonstrate that, during the in vitro di�erentiation ofHaCaT cells, there is a di�erential and sequentialinduction of the pRb family of proteins. Although, asexplained above (see also Geng and Weinberg, 1993;Fahraeus et al., 1996), in this system cell cycle arrestoccurs already 3 days after serum removal, well beforedetection of morphological or biochemical differentia-tion, in order to know if this di�erential induction ofthe pRb members is speci®c to the di�erentiationprocess, the expression of pRb, p107 and p130 wasanalysed during the entry of HaCaT cells into the cellcycle and the progression from G0/G1 to S, atransition in which these proteins play their pivotalfunctional role as cell cycle controllers. HaCaT cellswere synchronized in G0 by culturing them in theabsence of serum for 3 days, and forced to re-enter cellcycle by adding normal amounts of serum. Under theseconditions these cells enter S phase in a synchronizedmanner 16 ± 18 h after serum addition (Geng andWeinberg, 1993; Fahraeus et al., 1996). The results(Figure 1e) show that the three proteins are co-expressed throughout cell cycle progression, confirm-ing that the changes observed during in vitrodi�erentiation are speci®c to this process.

Expression of the pRb family of proteins in vivo

The above described results suggest that pRb, p107and p130 are di�erentially expressed during epidermaldi�erentiation. To con®rm this possibility in vivo, theexpression of these proteins was analysed by immuno-¯uorescence in human skin cryosections by confocalmicroscopy and compared with that of k10 (Figure 2).We observed pRb and p107 (Figure 2a and b)expressed in the basal layer, before k10 onset, andtheir expression remained through the spinous layers.On the other hand, p130 (Figure 2c) was detectedmostly in suprabasal cells expressing k10, which werealready committed in the terminal di�erentiationprogram. Similar results were obtained when the

pRb family in epidermal differentiationJM Paramio et al

950

expression of pRb, p107 and p130 was compared withthat of involucrin, although in this case, the onset ofp130 expression occurred before to that of this marker(results not shown). These immuno¯uorescence studieswere complicated by the observation that the stainingof pRb and related proteins was not homogenous,frequently giving irregular nuclear staining. Highermagni®cation confocal images demonstrated that, infact, all these proteins form nuclear aggregates (Figure2d, e and f). In the case of pRb, a similar stainingpattern has been observed in cultured ®broblasts usinga di�erent antibody, and this discrete localization hasbeen proposed to de®ne nuclear compartments relevantto the activity of this protein as a negative modulatorof the cell cycle (Lee et al., 1994). Our data indicatethat, in the case of keratinocytes, the three proteinsseem to be compartmentalized inside the nuclei.

Finally, to further con®rm the di�erent localizationof pRb family members in skin, we performedconventional double immuno¯uorescence analysis ofpRb and p130 (Figure 3). As can be seen, pRbexpression is predominant in the lower cell layers(basal and lower spinous cells). In contrast, p130expression is predominant in the upper spinous andgranular cell layers.

The forced expression of particular combinations of pRb,p107 and p130 induces di�erent stages of keratinocytedi�erentiation

The in vivo and in vitro di�erentiation studies implythat a di�erential and coordinated expression of themembers of the pRb family of proteins plays a directrole in epidermal di�erentiation. However, it is also

Figure 1 Sequential expression of pRb and pRb-related proteins during in vitro di�erentiation of HaCaT keratinocytes.Di�erentiation was triggered by culturing the cells in the absence of serum for di�erent periods of time. Note that after 4 days thecells showed a normal appearance (a), but areas of massive strati®cation appeared progressively and could be easily identi®ed after20 days (b, arrows). (c) Immunoblot analysis of protein extracts obtained from cultures at di�erent times (in days) after serumremoval, demonstrating the sequential induction of k10 (LH2 mAb) and involucrin (Sy5 mAb), whilst k14 (RCK 107 mAb)expression remained constant throughout the experiment. (d) Protein extracts as shown in (c) were used for immunoblot analysis ofthe expression of pRb, p107 and p130 using speci®c polyclonal antibodies. Note the sequential induction of pRb, p107 and p130during di�erentiation compared to their expression prior to serum removal (lane 0, prol.). (e) Analysis of the expression of pRb,p107 and p130 at di�erent times (in hours) after serum addition to cells synchronized in G0 by serum starvation, showing the co-expression of these three proteins during cell cycle progression from G0 to S

pRb family in epidermal differentiationJM Paramio et al

951

possible that this may be a consequence of the process.To distinguish the two alternatives, HaCaT keratino-cytes were transfected with di�erent combinations ofplasmids coding for members of the pRb family andthe estate of di�erentiation of the transfected cells wasanalysed as their ability to express k10 or involucrin.Two di�erent approaches were used to measure the

expression of these di�erentiation markers: immuno-histochemical staining, which allows the number ofpositive cells to be quanti®ed and Western blot, whichallows the levels of expression to be estimatedbiochemically in the transfected cell cultures. In thesetransfection experiments we included a plasmid codingfor b-Gal as a control for the transfection e�ciency

Figure 2 Localization of the pRb family of proteins in human skin. (a ± c) Double indirect immuno¯uorescences using monospeci®cantibodies against pRb (a), p107 (b) or p130 (c) (in red pseudocolor) together with keratin k10 (k8.60 mAb, in green pseudocolor)were visualized by confocal microscopy. Arrowheads in a and b denote basal keratinocytes expressing pRb and p107 respectively.The intranuclear localization of pRb (d), p107 (e) and p130 (f) (in red pseudocolor) indicated a compartmentalized, granulardistribution of these proteins (green pseudocolor depicts the distribution of lamins A/C stained with 6F8 mAb). Bars in a ± c =100 mm, in d ± f = 10 mm

pRb family in epidermal differentiationJM Paramio et al

952

and also to allow the identi®cation of the transfectedcells by X-Gal staining. Therefore, we could distinguishhistochemically undi�erentiated transfected cells (bluestained; b-Gal positive; Figure 4a and b), di�erentiatednon-transfected cells (pale brown stained; k10 orinvolucrin positive, Figure 4a), and di�erentiated,transfected cells (green-brown stained; arrowheads inFigure 4b). The number of this last type of cell wasnormalized to the number of transfected cells (b-Galpositive) and is represented in Figure 3c. In parallel,protein extracts from the transfected cultures were usedto analyse by Western blot the expression of k10(Figure 4d) and involucrin (Figure 4e). The b-Galexpression was also analysed by Western blot tocon®rm that in all the experiments the transfectione�ciency was similar (Figure 4f).The quantitative analysis of the transfection

experiments (Figure 4c) and the results from theWestern blot (Figure 4d, e) indicated that, althoughindividually transfected pRb, p107 and p130 were notable to drive the expression of k10 (Figure 4c, openbars, Figure 4d) or involucrin (Figure 4c, closed bars;Figure 4e), as compared with the b-Gal-transfectedcell control, certain combinations of the transfectedproteins modi®ed the expression of these markers. Inthis respect, the co-expression of pRb and p107increased signi®cantly k10 expression (Figure 4c openbars and d). On the other hand, pRb+p107+p130transfected cells expressed k10 at levels similar tothose found in control cells, and p130-transfected cellsexpressed k10 below these control values. Thissuggests that pRb+p107 induced k10 expressionwhereas, p130 appeared to inhibit it. The results

were di�erent when the expression of involucrin wasanalysed (Figure 4c, closed bars and e). In this case,p130 seemed to facilitate involucrin expression in thetransfected cells, in particular in the presence of pRband p107 (pRb+p107+p130 co-transfections).

HaCaT proliferation is di�erentially inhibited bypRb-related proteins

All the pRb family members have been shown toinhibit cell proliferation when transfected into culturedcells. However, each Rb member can display di�erente�ects in di�erent cell lines (Zhu et al., 1993, 1995).Since cell cycle arrest and di�erentiation are linkedprocesses, in view of the detected di�erential expressionof pRb, p107 and p130 during in vivo and in vitrokeratinocyte di�erentiation, as well as their distincte�ects on the expression of di�erentiation markersupon transfection, we have analysed the ability of theseproteins to inhibit HaCaT cell proliferation. To thisend, HaCaT keratinocytes were transfected with eitherpRb, p107 or p130 using the plasmids employed inFigure 4, which also bear the Neo resistance gene, orwith empty pcDNA3 vector plasmid as a control. Thenumber of colonies was scored after 2 weeks in G418-containing medium. The results (Table 1) demonstratedthat the overexpression of any Rb member inhibitedHaCaT keratinocyte proliferation. However, the extentof inhibition provided by each of these molecules wasdi�erent, suggesting that they may also play distinctroles in the control of keratinocyte proliferation. Tofurther con®rm this suggestion, we have analysed thesizes of the clones obtained in the above mentionedtransfections. We found (Figure 5) that colony size waslargest in Neo, signi®cantly smaller in p130, followedby pRb, which was larger than p107, thus indicatingthat this protein is the strongest inhibitor of HaCaTproliferation. In this regard, it is interesting to notethat the strongest inhibitors, pRb and p107, arealready expressed in basal skin keratinocytes (Figures2a, b and 3), whilst, the expression of p130, theweakest inhibitor, is shifted towards more suprabasal,postmitotic keratinocytes (Figures 2c and 3).

Discussion

Cell proliferation and di�erentiation are normallymutually exclusive processes and tissue renewalrequires a ®ne balance between them. The epidermisis an attractive model system in which to study theregulation of these processes, due to the compartmen-talization of proliferative and di�erentiated cells andthe existence of speci®c markers that allow theidenti®cation of di�erent stages of di�erentiation. Inthis study, we have analysed the expression ofmolecules which play a central role in the control ofcell proliferation: pRb, p107 and p130. The relevanceof these molecules to epidermal function is suggestedby the altered behavior of keratinocytes immortalizedby the forced expression of viral proteins, such as E1A,E7 or TAg, that bind and inactivate the pRb family ofproteins (Taylor-Papadimitriou et al., 1982; Banks-Schlegel and Howley 1983; McCance et al., 1988;Hudson et al., 1990; Barrandon et al., 1989). More-over, epidermal abnormalities were observed in

Figure 3 Di�erential distribution of pRb and p130 in humanepidermis. Double immuno¯uorescence against pRb (using IF8mAb, in red) and p130 (using rabbit polyclonal, in green)demonstrating that the former appears is predominantly expressedat the lower part of the epidermis, whereas p130 is mostprominent in the upper epidermal layers. Dashed line denotethe epidermal-dermal border; sb, stratum basal; ss, stratumspinous; sg stratum granulosum; sc, stratum corneum. Bar =50 mM

pRb family in epidermal differentiationJM Paramio et al

953

transgenic mice in which the expression of these viralproteins (Missero et al., 1993; Arbeit et al., 1994;Auewarakul et al., 1994) or cyclin D1 (Robles et al.,1996), which is involved in the functional incativationof Rb proteins (Weinberg, 1995), was targeted to theskin. Our results showing the di�erential expression ofthe Rb family of proteins both during in vitrodi�erentiation (Figure 1), and in epidermis in vivo(Figures 2 and 3), strongly suggest speci®c functionsfor these molecules during epidermal di�erentiation. Inagreement with this, we found that the overexpressionin HaCaT keratinocytes of certain combinations ofthese proteins (but none of them individually) triggersspeci®c stages of the epidermal di�erentiation process(Figure 4). The co-transfection of pRb and p107induces early di�erentiation (characterized by k10expression) whereas the presence of the three Rb

proteins is required to induce a later stage ofdi�erentiation (characterized by involucrin expres-sion). Moreover, the forced expression of thesemolecules inhibits HaCaT cell proliferation to adi�erent extent, p107 being the strongest and p130the weakest negative modulators (Table 1 and Figure5). Collectively, these results indicate the existence ofspeci®c, yet coordinated functional roles for pRb andRb-related proteins in the control of the proliferationand di�erentiation of human epidermal cells.Changes in the expression of the di�erent pRb

proteins during in vitro di�erentiation have beenreported in di�erent cell types. An increase in pRbduring muscle and neuronal cell di�erentiation hasbeen reported (Shin et al., 1995; Corbeil et al., 1995).Contradictory results have been obtained concerningp107 and p130 expression, probably due in part to the

Figure 4 Di�erent members of the Retinoblastoma family of proteins co-operate to trigger speci®c stages of epidermal di�erentiation.HaCaT keratinocytes were co-transfected with di�erent combinations of plasmids coding for pRb, p107 or p130 (all under the same CMVpromoter) together with a plasmid coding for b-Gal. Forty-eight hours after transfection cells were ®xed and stained with X-Gal to detectthe transfected cells (blue stained). The expression of keratin k10 (LH2 mAb) or involucrin (Sy5 mAb) was analysed by peroxidase stainingusing DAB as substrate (brown stained). Cells double positive for X-Gal and for keratin k10 or involucrin (green-brown stained) werescored. (a) Example of cells transfected only with b-Gal. (b) Example of b-Gal+pRb+p107+p130 co-transfection. Arrowheads in (b)denote transfected cells positive for involucrin expression (Sy5 mAb). (c) Summary of three independent experiments showing the percentageof cells positive for keratin k10 (open bars) or involucrin (closed bars) expression with respect to the total number of transfected, X-Galstained cells as a function of the di�erent combinations of plasmids used in the transfections. At least 500 transfected, b-gal positive cellswere scored in each experiment. (d) Immunoblot analysis of protein extracts from the same transfection experiments of (c) using LH2 mAbagainst keratin k10. (e) The same blot shown in (d) was stripped and probed with Sy5 mAb to analyse involucrin expression. (f) The sameblot shown in d and e was stripped and reprobed against b-Gal to con®rm that transfection e�ciency was similar in all the experiments

pRb family in epidermal differentiationJM Paramio et al

954

di�erent cells analysed. p107 increases during neuronal(Corbeil et al., 1995) and erythroid di�erentiation(Richon and Venta-Perez, 1996), and also duringC2C12 myoblast cell di�erentiation (Thorburn et al.,1993), but appears to decrease during L6 myoblastdi�erentiation (Kiess et al., 1995). Similarly, p130 hasbeen reported to increase during C2 and L6 muscledi�erentiation (Shin et al., 1995, Kiess et al., 1995),whilst its level remains constant during C2C12 musclecell and neuronal di�erentiation (Corbeil et al., 1995).However, to our knowledge, a sequential, coordinatedexpression of these three proteins as we have observedduring HaCaT keratinocyte di�erentiation has notpreviously been reported in other systems, suggestingthat epidermal di�erentiation may display speci®cfeatures not present in muscle or neuron celldi�erentiation.The biological reasons for the existence of multiple

Rb-like proteins have been a matter of debate. Strongevidence for functional di�erences among theseproteins comes from their capacity to bind andrepress di�erent members of the E2F family oftranscription factors during the cell cycle (Cobriniket al., 1993; Sardet et al., 1995; Vairo et al., 1995).The lethality observed in embryos lacking pRb (Lee etal., 1992; Jacks et al., 1992; Clarke et al., 1992), asopposed to the absence of apparent phenotypes inp107 and p130 null mice (Lee et al., 1996; Cobrinik etal., 1996) and the observation that MyoD expressiongenerates aberrant muscle di�erentiation in pRb-, butnot in p107- or p130-de®cient cells (Novitch et al.,1996) also suggests that these proteins play speci®croles during di�erentiation. In this context, ourtransfection experiments (Figure 4) indicate thatspeci®c combinations of Rb proteins are required toinduce progressive stages of terminal di�erentiation inepidermal cells (pRb+p107 to induce k10, andpRb+p107+p130 to induce involucrin). A similarcooperation between Rb proteins has also beendemonstrated by gene targeting experiments, sincep1077/7 or p1307/7 mice do not show overtabnormalities, but pRb7/7 p1077/7 double mu-tants display accelerated embryo lethality (Lee et al.,1996), and p1077/7 p1307/7 double mutant micesu�er deregulated chondrocyte growth, defective bonedevelopment and neonatal lethality (Cobrinik et al.,1996). These mutant animals will be essential tools tostudy the role of these molecules in skin, although it isnecessary to consider that species-speci®c di�erences

could make extrapolation of the results found in miceto humans di�cult. For example, it is worthremarking that whilst pRb appears to be expressedexclusively suprabasally in mouse epidermis (Szekelyet al., 1992, JMP, unpublished results), we foundstrong expression of this protein in the basal layer ofhuman epidermis (Figures 2a and 3).The expression pattern of pRb, p107 and p130

observed during in vitro di�erentiation of cell cycle-arrested HaCaT cells (Figure 1), the immunolocaliza-tion of these proteins in human skin in vivo (Figure 2)and the results of the transfection experiments (Figure4) correlate well. In addition, the more potentinhibitory e�ect of p107 on keratinocyte proliferationas compared with pRb and p130 (Table 1 and Figure5), suggests that this protein could be particularlyrelevant in achieving the cell cycle arrest that precedesthe onset of terminal di�erentiation. This is coherentwith its presence in basal cells in vivo (Figure 2b). Thiscell cycle arrest is provided in the in vitro differentia-tion experiments by serum starvation, explaining why,contrary to the in vivo situation, p107 is expressedlater than pRb. Therefore, the in vivo data correlatereasonably well and overall suggest a model implyingthe sequential and co-operative function of the Rbfamily of proteins in epidermal di�erentiation. In thismodel, the presence of p107 would always be required,it being essential to inhibit the cell proliferationrequired to trigger the process, and also co-operatingwith the other Rb members to the later events ofterminal di�erentiation. pRb seems to be fundamentalin co-operation with p107 in the onset of earlydi�erentiation (k10 induction) of cells, but itspresence is also needed to induce later di�erentiationevents. Finally, p130, in co-operation with p107 andpRb, appears to be involved mostly in terminaldi�erentiation processes (induction of involucrin andprobably later stages). These results, together with thereported predominant activity of p130 during G0

Table 1 Di�erential inhibition of HaCaT cell proliferation by pRBfamily

Number of clones Inhibition (%,Plasmid Exp 1 Exp 2 Exp 3 Exp 4 mean+s.d.)

pcDNA3CMV3pRbCMV3p107CMV3p130

118421937

26912537170

3212020

279411258

066+1191+656+18

HaCaT cells were transfected with the above-quoted plasmids. Forty-eight hours after DNA addition, cells were cultured in the presence ofG418 (0.5 mg/ml) for 2 weeks. Afterwards, colonies were ®xed,stained with Giemsa and counted. The inhibition produced by thedi�erent members of the pRb family was estimated comparing thenumber of clones obtained with the corresponding plasmids to thoseobtained with the empty vector pcDNA3

Figure 5 Size distribution of the colonies arising aftertransfection with the di�erent members of the pRb family orwith the empty vector (see Table 1)

pRb family in epidermal differentiationJM Paramio et al

955

(Cobrinik et al., 1993), suggest a progressive transitionfrom G1 to G0 as in vivo keratinocytes withdraw fromthe cell cycle and move towards terminal differentia-tion.

Materials and methods

Cell culture

Human keratinocyte HaCaT cells (Boukamp et al., 1988;kindly provided by Dr NE Fusenig) were cultured in DMEMcontaining 10% FCS (SeraLab, Westbury, NY) andantibiotics. Medium was changed every 3 ± 4 days. To allowin vitro di�erentiation, cells were seeded at 60 ± 80%con¯uence. After 16 ± 24 h, medium was replaced byDMEM/antibiotics without FCS. Cells were kept underthese conditions for di�erent times, changing the mediumevery 3 days. For experiments involving cell synchronization,a similar procedure was followed but, after 72 h in mediumwithout FCS, DMEM containing 10% FCS was added andsamples were taken at di�erent times afterwards (Geng andWeinberg, 1993; Fahraeus et al., 1996).

Plasmids and transfections

To construct pCMV3pRb, the full length human pRb cDNAwas obtained as a KpnI ±BamHI fragment from pBSKhupRb(kindly provided by Dr S Mittnacht), and subcloned intopcDNA3, which incorporates the Neo resistance gene(Invitrogen, Carlsbad, CA). To construct pCMV3p107, thefull length p107 cDNA was obtained as a BamHI ±BglIIfragment from pCMVp107 (Zhu et al., 1993; kindly providedby Dr E Harlow), and subcloned into pcDNA3. p130 fulllength cDNA was obtained as a BamHI ±AvrII fragmentfrom pBSKp130 (Hannon et al., 1993; kindly provided by DrD Beach) and subcloned into BamHI ±XbaI sites ofpcDNA3. pCMV bGal was kindly provided by Dr AAlonso. Transient transfections with these plasmids wereperformed using the poly-L-ornithine method (Nead andMcCance, 1995). Typically, 1 ± 10% of the cells in the culturewere transfected (b-Gal positive). Protein extracts wereobtained 48 h after DNA addition and used for immuno-bloting analysis or, alternatively, cells were ®xed with 5%glutaraldehyde in PBS containing 1 mM MgCl2 and thetransfected b-Gal expressing cells were detected by X-Galstaining. The di�erentiation state of transfected and non-transfected cells was analysed by immunocytochemistry.LH2 (neat supernatant) or Sy5 (Sigma, St. Louis, MO 1/500dilution) monoclonal antibodies were used to monitorkeratin k10 and involucrin expression, respectively, incombination with horseradish peroxidase-labeled secondaryanti-mouse antibodies (Jackson Immunoresearch Labora-tories, West Grove, PA 1/5000 diluted) and diaminobenci-dine as a substrate (DAB kit; Vector, Burlingame, CA).Permanent transfections were performed by the calciumphosphate method using 1 ± 56106 cells and 15 mg of theabove mentioned plasmids. On each independent experi-ment, equal number of cells were transfected with equalamounts of each plasmid. Cultures were split 48 h after DNAaddition and cultured in the presence of 0.5 mg/ml G418(Sigma) for at least 2 weeks, changing the medium every 4days. At this time colonies were ®xed, stained with Giemsa,counted and their size was measured.

Immuno¯uorescence

Fresh cryostat skin sections were ®xed for 10 min at 7208Cin methanol/acetone (1/1; v/v), air-dried and rehydrated byimmersion in PBS containing 0.1% Tween 20 (PBST). Thesections were blocked by incubation with preimmune serumor 5% BSA for 1 h. After washing, di�erent antibodies wereadded and the sections were incubated overnight at 48C.Rabbit polyclonal antibodies speci®c for pRb, p107 or p130(Sta. Cruz Biotech., Sta. Cruz, CA, 2 mg/ml), IF8 mAbagainst pRb (kindly provided by Dr DP Lane, Dundee, UK)K8.60 mAb against k10 (Sigma; 1/40 dilution), Sy5 mAb(Sigma; 1/100 dilution) against involucrin and 6F8 mAb (1/5dilution) against lamin A/C, were used as primaryantibodies. Secondary antibodies especially prepared formultiple labeling purposes (Jackson Immunoresearch) wereused as described elsewhere (Paramio and Jorcano, 1994).Specimens were analysed using a BioRad MRC600 confocalmicroscope or Zeiss Axiophot for conventional immuno-¯uorescences. Nuclear integrity was con®rmed routinely byDAPI (Boehringer Mannheim) staining. Controls wereperformed omitting primary antibodies.

Immunoblotting

Total protein extracts from cultured cells were obtained bylysis in RIPA bu�er (150 mM NaCl, 1.0% NP-40, 0.5%sodium deoxycholate, 0.1% SDS, 50 mM Tris HCl pH 8.0)containing protease inhibitors (2 mg/ml aprotinin; 2 mg/mlleupeptin; 100 mg/ml phenylmethylsulfonyl ¯uoride) for30 min at 48C. Lysates were precleared by centrifugationand supernatants were stored at7708C. Protein content wasdetermined colorimetrically (BioRad Protein Assay, BioR-ad, Richmond, CA). 75 mg of total protein was electrophor-esed in 8% SDS-polyacrylamide gels and transferred tonitrocellulose (Amersham, Cleveland, OH) using semi-dryequipment. Membranes were blocked by incubation in TBScontaining 0.1% Tween 20 and 5% non-fat dried milk(TBST-milk) and probed using IF8 mAb (neat supernatant)against pRb, or the above-quoted polyclonal antibodiesagainst pRb, p107 or p130 (diluted to 0.02 mg/ml in TBST-milk), or LH2 mAb against k10, RCK107 mAb against k14or Sy5 mAb (1/1000 diluted). Extracts from transfectionswere also probed using Gal40 mAb (1/5000 diluted; Sigma;St Louis, MO) against b-Gal to demonstrate that transfec-tion e�ciency was similar on each experiment. SecondaryHRP-labeled antibodies (Jackson Immunoresearch) wereused at 1/1000 dilution in TBST-milk. Blots were developedusing ECL (Amersham) following the manufacturer'srecommendations.

AbbreviationspRb, Retinoblastoma protein; mAb, monoclonal antibody;PBS, phosphate bu�ered saline; FCS, fetal calf serum.

AcknowledgementsThis work was supported by the CRC and partially fromgrant PB94-1230 from the DGCYT (Spain). SL wassupported by a postdoctoral fellowship from the Minister-io de EducacioÁ n y Ciencia (Spain). The generous gift ofmaterials from E Harlow, S Mittnacht, A Alonso and DBeach is acknowledged.

References

Arbeit JM, Munger K, Howley PM and Hanahan D. (1994).J. Virol., 68, 4358 ± 4368.

Auewarakul P, Gissmann L and Cid-Arregui A. (1994).Mol.Cell. Biol., 14, 8250 ± 8258.

Banks-Schlegel SP and Howley PM. (1983). J. Cell. Biol., 96,330 ± 337.

Banks-Schlegel S and Green H. (1981). J. Cell. Biol., 90,732 ± 737.

pRb family in epidermal differentiationJM Paramio et al

956

Barrandon Y, Morgan JR, Mulligan RC and Green H.(1989). Proc. Natl. Acad. Sci. USA, 86, 4102 ± 4106.

Boukamp P, Petruseva RT, Breitkreutz D, Hornung J,Markham A and Fusenig NE. (1988). J. Cell. Biol., 106,761 ± 771.

Clarke AR, Maandag ER, van Roon M, van der Lugt NMT,van der Valk M, Hooper ML, Berns A and te Riele H.(1992). Nature, 359, 328 ± 330.

Cobrinik D, Lee MH, Hannon G, Mulligan G, Bronson RT,Dyson N, Harlow E, Beach D, Weinberg RA and Jacks T.(1996). Genes. Develop., 10, 1633 ± 1644.

Cobrinik D, Whyte P, Peeper DS, Jacks T and Weinberg RA.(1993). Genes. Develop., 7, 2392 ± 2404.

Corbeil HB, Whyte P and Branton PE. (1995). Oncogene, 11,909 ± 920.

Ewen ME, Xing Y, Lawrence JB and Livingston DM. (1991).Cell, 66, 1155 ± 1164.

Fahraeus R, Paramio JM, Ball KL, Lain S and Lane DP.(1996). Curr. Biol., 6, 84 ± 91.

Fuchs EV. (1990). J. Cell. Biol., 111, 2807 ± 2814.Geng Y and Weinberg RA. (1993). Proc. Natl. Acad. Sci.

USA, 90, 10315 ± 10319.Gu W, Schneider JW, Condorelli G, Kaushai S, Mahdavi Vand Nadal Ginard B. (1993). Cell, 72, 309 ± 624.

Hannon GJ, Demetrick D and Beach D. (1993). Genes. Dev.,7, 2378 ± 2391.

Hudson JB, Bedell MA, McCance and Laimins LA. (1990).J. Virol., 64, 519 ± 526.

Jacks T, Fazeli A, Schmitt EM, Bronson RT, Goodell MAand Weinberg RA. (1992). Nature, 359, 295 ± 300.

Kiess M, Gill RM and Hamel PA. (1995). Cell. Growth.Di�eren., 6, 1287 ± 1298.

LaThange NB. (1994). Trends. Biochem. Sci., 19, 108 ± 114.Lee EYHP, Chang CY, Hu N, Wang Y-CJ, Lai CC, HerrupK, Lee W-H and Bradley A. (1992).Nature, 359, 288 ± 294.

Lee MH, Williams BO, Mulligan G, Mukai S, Bronson RT,Dyson N, Harlow E and Jacks T. (1996). Genes. Develop.,10, 1621 ± 1632.

Lee W-H, Xu Y, Hong F, Durfee T, Mancini MA, Ueng Y-C,Chen PL and Riley D. (1994). Cold. Spring. Harb. Symp.Quant. Biol., 59, 97 ± 107.

Leigh IM, Lane EB and Watt FM (eds). (1994). Thekeratinocyte handbook. Cambridge University Press:Cambridge.

Li Y, Graham C, Lacy S, Duncan AMV and Whyte P.(1993). Genes. Dev., 7, 2366 ± 2377.

McCance DJ, Kopan R, Fuchs E and Laimins LA. (1988).Proc. Natl. Acad. Sci. USA, 85, 7169 ± 7173.

Missero C, Serra C, Stenn K and Dotto GP. (1993). J. Cell.Biol., 121, 1109 ± 1120.

Nevins JR. (1992). Science, 258, 424 ± 429.Nead MA and McCance DJ. (1995). J. Invest. Dermatol.,105, 668 ± 671.

Novitch BG, Mulligan GJ, Jacks T and Lassar AB. (1996). J.Cell. Biol., 135, 441 ± 456.

Paramio JM and Jorcano JL. (1994). Exp. Cell. Res., 215,319 ± 331.

Paramio JM and Jorcano JL. (1997). Br. J. Dermatol., 137,44 ± 50.

Picksley SM and Lane DP. (1994). Curr. Op. Cell. Biol., 6,853 ± 858.

Rice RH and Green H. (1979). Cell, 18, 681 ± 694.Richon VM and Venta-Perez G. (1996). Cell. Growth. Di�.,7, 31 ± 42.

Robles AI, Larcher F, Whalin R, Murillas R, Richie E,Gimenez-Conti IB, Jorcano JL and Conti CJ. (1996).Proc. Natl. Acad. Sci. USA, 93, 7634 ± 7638.

Sardet C, Vidal M, Cobrinik D, Geng Y, Onufryk C, Chen Aand Weinberg RA. (1995). Proc. Natl. Acad. Sci. USA, 92,2403 ± 2407.

Schneider JW, Gu W, Zhu L, Mahdavi V and Nadal-GinardB. (1994). Science, 264, 1467 ± 1471

Shin EK, Shin A, Paulding C, Scha�hausen B and Yee AS.(1995). Mol. Cell. Biol., 15, 2252 ± 2262

Szekely L, Jiang WQ, Bulic-Jakus F, Rosen A, Ringertz Nand Wiman KG. (1992) Cell. Growth. Di�er., 3, 149 ± 156

Taylor-Papadimitriou J, Purkis P, Lane EB, McKay IA andChang SE. (1982). Cell. Di�er., 11, 169 ± 180

Thornburn AM, Walton PA and Feramisco JR. (1993) Mol.Biol. Cell., 4, 705 ± 713

Vairo G, Livingston DM and Ginsberg D. (1995). GenesDev., 9, 869 ± 881

Wang JKJ, Knudsen ES and Welch PJ. (1994). Adv. CancerRes., 64, 25 ± 85

Weinberg RA. (1995). Cell, 81, 323 ± 330Zhu L, Enders G, Lees JA, Beijesbergen RL, Bernards R andHarlow E. (1995). EMBO. J., 14, 1904 ± 1913

Zhu L, van der Heuvel S, Helin K, Fattaey A, Ewen ME,Livingston DM, Dyson N and Harlow E. (1993). Genes.Dev., 7, 1111 ± 1125

zur Hausen H. (1991). Science, 254, 1167 ± 1173.

pRb family in epidermal differentiationJM Paramio et al

957