Plasticity of the response of fetal mouse fibroblast to lactationhormones

Sen Wang, Jingpu Zhang*, Yanyan Geng, Shaoxia Zhu

Institute of Genetics and Developmental Biology, Chinese Academy of Science, Beijing 100080, China

Received 12 December 2002; revised 31 May 2003; accepted 8 July 2003

Abstract

In the work described in this article, mouse fetal fibroblasts were treated with three lactation hormones (insulin, progesterone andoxytocin) and the cellular changes were analyzed by RT–PCR–Southern hybridization. A gene-expression pattern characteristic ofmammary epithelioid cells was induced by the hormones, as indicated by expression of the marker genes �-casein and �-casein. Twomammary epithelial cell-specific gene expression vectors were constructed with bovine �-s1-casein or ovine �-casein gene promotersdirecting an EGFP reporter gene. Transient expression of the EGFP gene was observed in cells treated by the hormones but not incontrol cells. Cell morphology also changed after insulin and oxytocin treatments; the cells resembled epithelial cells rather thanfibroblasts. Our results suggest that mouse fetus fibroblasts can be partially induced by lactation hormones to resemble mammaryepithelial cells. This procedure might help to increase the efficiency of gene targeting in studies of mammary gland bioreactors.� 2003 Published by Elsevier Ltd.

Keywords: �-casein; �-casein; Fetus fibroblast; Induction; Insulin; Progesterone; Oxytocin; Mammary epithelioid cell; Transgene; Enhanced greenfluorescent protein

1. Introduction

Mammal fibroblasts have considerable potential ascell platforms for homologous gene recombination intransgenic animals, but so far the success rate has beenlow. If fibroblasts could be induced, even transiently, toadopt characteristics of the target cells during exogenousgene integration, the rate should improve. A problem intransgenics is the random integration of transgenic con-structs into host chromosomes, which causes low or zeroexpression of the target genes and instability in heredity.In principle, homologous recombination can overcomethe obstacle. However, though this technique has beenused widely for over a decade for modifying the genomein cell platforms (Baguisi et al., 1999; Cibelli et al., 1998;McCreath et al., 2000; Schnieke et al., 1997), extensiverandom integration of transgene constructs into hostchromosomes still impairs the efficiency with which

transgenes are site-directed in cell transfection or electrointroduction of embryo stem cells or somatic cellplatforms. Pronucleus microinjection techniques are theobvious comparison. If transgenes could be placedprecisely into host cell chromosomes, thus (for instance)changing silent gene loci into active ones, the chances offunctional integration would be much improved.

A cell plateau offers an operational and screeningplatform for gene targeting but cannot resolve theproblem of accurate gene construct docking at targetsites. However, exogenous gene integration is known tooccur preferentially at hypersensitive or active gene sitesrather than silent ones, so a cell’s current gene expres-sion pattern causes some potential integration sites forexogenous genes to be favoured over others. A goodexample has been reported by Kind’s group (McCreathet al., 2000); ovine �1(I) procollagen, a highly expressedgene in fibroblasts, was used for gene-targeting andexact site–directed integration was achieved in ovinefetal fibroblasts. Fibroblasts proliferate actively, areeasily cultured, have the potential to develop and differ-entiate, and are convenient for use in somatic nucleustransfer and gene-targeting plateau studies. However,

* Corresponding author. Fax: +86-10-62551951E-mail address: jpzhang@genetics.ac.cn (J. Zhang).Abbreviations: RT–PCR, Reverse transcription–polymerase chain

reaction; EGFP, Enhanced green fluorescent protein; MEF, Mouseembryo fibroblast.

Cell Biology International 27 (2003) 755–760

CellBiologyInternational

www.elsevier.com/locate/cellbi

1065-6995/03/$ - see front matter � 2003 Published by Elsevier Ltd.doi:10.1016/S1065-6995(03)00160-4

fibroblasts do not seem suitable for mammary bio-reactor studies, because their milk protein genes aresilent and therefore difficult to target for integration ofexogenous genes. In developmental terms, fibroblastsare remote from mammary epithelial cells.

Mammogenic hormones are the primary regulatorsof mammary gland development, so the potential forhormonally induced conversion of embryo fibroblasts tomammary epithelioid cells merits consideration. If silentmilk protein genes can be activated by hormone stimu-lation in fetal fibroblasts, active gene-integration sitescould become available for more accurate homologousgene recombination. In the present work, we treatedmouse fetal fibroblasts with the lactation hormonesinsulin, progesterone and oxytocin. The fibroblastsexhibited a plastic response and the results suggest thatthe gene expression patterns of the fibroblasts can beconverted toward a pattern more typical of mammaryepithelioid cells.

So far as we are aware this is the first report thatlactating hormones can directly affect embryonic fibro-blasts. The results suggest a technique for increasing theefficiency of gene targeting in transgenic studies, andfor producing cells with desired characteristics fortissue engineering and directed differentiation studies bycytokine and/or hormone induction.

2. Materials and methods

2.1. Reagents and kits

Prime-a-Gene labeling system was obtained fromPromega Co, USA. �-P32-dATP was purchased fromBeijing Yahui Bio Medical Engineering Co, Beijing,China. The PCR kit, reverse transcriptase (RAV-2) and

ribonuclease inhibitor were obtained from TaKaRaBiotechnology (DaLian) Co. Ltd, Japan. Nylon mem-branes were purchased from Amersham-Phamacia, Co.,Sweden. D-MEM and trizol reagent were obtained fromGiBcol/ BRL, USA. Lipofectamine reagent was anInvitrogen product, USA; the DNA purification kit waspurchased from BioDev Scientific & Technical Co. Ltd,Beijing, China. Oxytocin injection (10 units/ml) wasobtained from Nanking Xintian Co. Ltd of Bio-chemistry Pharmaceutics, Nanking, China; progesteroneinjection (20 mg/ ml) from Shanghai TongyongPharmaceutical Co. Ltd, Shanghai, China, and insulininjection (400 units/10 ml) from XuZhou Wanbang Co.Ltd of Biochemistry Pharmaceutics, China. The otherreagents were of analytical grade and were made inChina.

2.2. Instrument

We used a fluorescence inversion microscope,Olympus CK40/U-RFLT50.

2.3. Animals

KunMing mice were used.

2.4. Plasmids and gene constructs

Gene cloning and recombination were performed bystandard methods (Sambrook et al., 1989). pEGFP-N1(Clontech Laboratories, Inc., USA) was kindly providedby Dr Claus Matthaei. Mouse �-actin cDNA was usedas an internal control in RT–PCR–Southern hybridiz-ation. Ovine �-casein promoter was cloned in ourlaboratory (unpublished data). Two gene constructs for

Fig. 1. Structures of p�-casein-EGFP and p�-casein-EGFP. A: p�-casein-EGFP constructed by substituting CMV promoter with bovine �-s1-caseinpromoter in pEGFP-N1. B: p�-casein-EGFP constructed by substituting CMV promoter with ovine �-casein promoter in pEGFP-N1.

S. Wang et al. / Cell Biology International 27 (2003) 755–760756

mammary epithelia cell-specific expression (p�-casein-EGFP and p�-casein-EGFP, see Fig. 1) were made bysubstituting the CMV promoter for bovine �-s1-caseinpromoter (2.2 kb) (Chen et al., 1992) or ovine �-caseinpromoter (1.7 kb, two copies in tandem) in pEGFP-N1.

2.5. Cell culture and hormone induction

Mouse embryo fibroblasts (MEF) were obtainedfrom 12–13 day embryos and cultured in DMEM plus10% bovine fetus blood serum, penicillin and strepto-mycin (complete DMEM) as described by David et al.(1998), Espejel and Blasco (2002), Hogan et al. (1986).After five passages of the MEF, the cells were cultivatedin 25 ml culture flasks seeded with 105 cells/flask andgrown to 80–90% confluence. The medium was replacedwith fresh complete DMEM plus insulin (final concen-tration 4 units/ml), progesterone (final concentration10 µg/ml), oxytocin (final concentration 1 unit/ml), or amixture of all three hormones at these same concen-trations. For controls, hormones were omitted. The cellswere then cultivated for at least 3 days.

2.6. RT–PCR–Southern hybridization

2.6.1. Isolation of total RNAsThe hormone-treated cells were harvested with

trypsin and suspended in 500 µl of Trizol Reagent. TotalRNAs were isolated in accordance with the TrizolReagent instructions. Reverse transcription of mRNAwas performed for the first chain as described bySambrook et al. (1989). The PCR primers (synthesizedby Shanghai Sangon Co. Ltd, Shanghai, China) formouse �-casein cDNA are 5#-CTCATCCTCACCTGCCTCG-3# (forward) and 5#-GTGCCTGATCCACTACACTCAT-3# (reverse). For mouse �-casein cDNA theyare 5#-GGTGAATCTCATGGGACA-3# (forward) and5#-GAAGGGTGCTACTTGCTG-3# (reverse), and formouse �-actin cDNA as internal mRNA control theyare 5#-AGGGAAATCGTGGGTGACATCAAA-3# (for-ward) and 5#-ACTCATCGTACTCCTGCTTGCTGA-3#(reverse) (Han et al., 1995). The PCR templates were thefirst strands of cDNAs to the mRNAs obtained fromcells with and without hormone induction. cDNA frag-ments of mouse �-casein and �-casein served as positivecontrols. Each of the forward and reverse primers(see above) was used at 30 pM concentration for eachPCR. Each reaction mixture (50 µl) comprised 10�PCRbuffer (5 µl), 4�10 mM dNTP (1 µl), 25 mM Mg++

(3 µl), Taq DNA polymerase (3 units) and cDNA1st

template (w0.1 µg). The parameters for the first PCRcycle were 95 (C 240 s, 46 (C 60 s, 72 (C 60 s. For thesecond to 35th cycles they were 94 (C 60 s, 52 (C 60 s,72 (C 60 s, with a 10 min incubation after the cycles.The sizes of the PCR products were 769 bp, 460 bp and478 bp for �-casein, �-casein and �-actin respectively.PCR products (each 5 µl) were separated on a 0.8%

agarose gel, transferred to Hybond Nylon membranesby capillary adsorption on a paper tower, and hybrid-ized at 65 (C using standard methods (Sambrook et al.,1989). The probes were the PCR products from mouse�-s-casein and �-casein labeled with �-P32-dATP usingthe Prime-a-Gene labeling system.

2.7. Transfection of fibroblasts

After five passages, fibroblasts were passagedto 6-well plates at 3�104 cells/well and grown tow90% confluence before transfection with p�-casein-EGFP or p�-casein-EGFP (w1 µg DNA/well), usingLipofectamine Reagent following the kit instructions.After 5 h the reagent was replaced with fresh completeDMEM with or without insulin, progesterone, oxytocinand the hormone mixture, as described above. The cellswere cultured in a 37 (C, 5% CO2 incubator for 16 h orlonger and examined under a fluorescence inversionmicroscope.

3. Results

3.1. Expression of endogenous casein genes in thefibroblasts

Expression of endogenous �-Casein and �-Caseingenes in mouse fetus fibroblasts incubated with thedifferent hormones was detected by RT–PCR–Southernhybridization. The �-Casein gene was strongly inducedby insulin and the three-hormone mixture and weaklyinduced by oxytocin. It was not expressed inprogesterone-treated cells or controls. The �-Casein genewas strongly induced by insulin, progesterone, oxytocinand the three-hormone mixture, but there was noexpression in control cells (Fig. 2). These results sug-gest that lactation hormones differentially stimulatefibroblasts to express milk protein genes in vitro. Thedifferential effect might result from different signal path-ways or receptor activities; gene promoters might alsorespond differentially to the same signal.

Normally, endogenous casein gene expression isstrictly tissue-specific and is confined to mammary epi-thelial cells. Lactation and mammary gland develop-ment require many hormones including insulin,estrogens and progestogen (Ashktorab et al., 1999;Guyette et al., 1979; Kanai et al., 1993; Schmidhauseret al., 1990; Topper and Freeman, 1980). However, ourresults suggest that these hormones can also induce invitro expression of endogenous casein genes in MEF.

3.2. Transient expression of mammary epithelialcell-specific transgene constructs in mouse fetalfibroblasts

Fibroblasts were transfected with two mammaryepithelial cell-specific transgene constructs, a bovine

S. Wang et al. / Cell Biology International 27 (2003) 755–760 757

�-s1-casein promoter directing the EGFP gene (p�-casein-EGFP) and an ovine �-casein promoter directingthe same EGFP gene (p�-casein-EGFP) and were theninduced with insulin, progesterone, oxytocin or thehormone mixture. Green fluorescence appeared in cellsinduced by insulin, oxytocin or the hormone mixtureafter p�-Casein-EGFP transfection, and in cells inducedby progesterone (Fig. 3) or the hormone mixture (datanot shown) after p�-casein-EGFP transfection. Nogreen fluorescent was seen in controls after transfectionwith either of the gene constructs. These results furthersupport the conclusion that gene-expression patterns are

regulated by cell physiological states or tissue-specificinternal environment. The gene-expression pattern offetal fibroblasts can be changed into one resemblingmammary epithelioid cells by induction with lactationhormones. This implies that milk protein gene pro-moters can be activated by hormone stimulation ofdevelopmentally remote cells, and the hormones actdifferentially on the promoters.

Transfection efficiency is very low for the primarycultured cells and this might have affected the results.Negative results for transfection and hormone inductionare therefore difficult to interpret.

3.3. Morphological changes of the fibroblasts respondingto lactating hormones

Fetal fibroblasts were incubated with insulin, proges-terone, oxytocin or the three hormone mixture, orwithout hormones, for at least 3 days. After the thirdday, cells became rounder and less filament shape afterproliferation in insulin- or oxytocin-supplementedDMEM, and finally they assumed epithelioid shapes.However, the cells induced with progesterone or thethree hormone mixture, retained a typical fibroblastmorphology, even though the shapes of the control cellschanged more or less as cell density increased (Fig. 4).

Insulin or oxytocin could independently effect a mor-phological change from fibrous toward epithelioidcharacter. After progesterone treatment, the cellsretained fibroblast form. Progesterone could be the“decision maker” among the hormones in respect ofmorphological change, as implied by the three hormoneresult. The relationship between cell shape and milkprotein-gene expression remains to be investigated.

4. Discussion

Expression of milk protein genes during lactationin mammary epithelial tissue is strictly controlled bycertain hormones. Generally these hormones havedistinct signal pathways and functions. Insulin hasnumerous biological functions and has been deployed

Fig. 2. Expression of endogenous �-Casein and �-Casein genes de-tected by RT–PCR–Southern hybridization in mouse fetus fibroblasts.A: mouse �-Casein cDNA as the probe; B: mouse �-Casein cDNA asthe probe; C: mouse �-actin RT–PCR as the internal control. P:positive controls are �-Casein (A) or �-Casein (B) cDNA fragmentrespectively. M: DNA marker, Lambda DNA cut by HindIII andEcoRI; 1: uninduced control; 2: insulin induction; 3: progesteroneinduction; 4: oxytocin induction; 5: induction by the three-hormonemixture (see Materials and methods section for details).

Fig. 3. Transient expression of mammary epithelial cell-specific transgene constructs in mouse fetal fibroblasts (50�). A: p�-casein-EGFP by insulininduction; B: p�-casein-EGFP by oxytocin induction; C: p�-casein-EGFP by the hormone mixture induction; D: p�-casein-EGFP by progesteroneinduction.

S. Wang et al. / Cell Biology International 27 (2003) 755–760758

in transgenic studies of mammary gland bioreactors.Casein gene expression requires stimulation by lacto-genic hormones such as insulin, hydrocortisoneand prolactin both in vivo and in vitro (Ashktorabet al., 1999; Guyette et al., 1979; Kanai et al., 1993;Schmidhauser et al., 1990; Topper and Freeman, 1980).Progesterone and oxytocin play important roles duringpregnancy and parturition and during lactating (Berget al., 2002; Blyth et al., 1997; Delongeas et al., 1997;Kimura et al., 1998; Thomas et al., 1995). The mammaryepithelium has target sites for oxytocin (Kimura et al.,1998), which are regulated partly by progesterone (Blythet al., 1997; Thomas et al., 1995). However, induction oftransdifferentiation of embryonic fibroblasts by thesehormones has not been reported till now.

In the present work, we have demonstrated thatmouse fetal fibroblasts differentially express the mam-mary epithelial tissue-specific genes �-casein and�-casein after treatment with the lactation hormonesinsulin, progesterone and oxytocin. Our results showthat the three hormones, particularly insulin and oxy-tocin, can act on the fetal fibroblast and change bothgene expression pattern and cell morphology. Both�-casein and �-casein promoters in the fibroblasts(endogenous and exogenous) were activated by thehormones. We infer that there are probably effectors,perhaps activators of casein-gene promoters, which areactivated or produced in embryo fibroblasts in responseto these hormones or their analogues. In other words,hormones that usually act on terminally differentiatedmature tissues may have unexplored effects on embryofibroblasts. The results also imply that embryo fibro-blasts can differentiate along various pathways aftertreatment with appropriate inductive factors, so theymight have wide-ranging applications in embryodevelopment studies, biomedicine, and biotechnology.

In summary, mouse fetal fibroblasts respond to lac-tation hormones, which convert their gene expressionpatterns into one resembling mammary epithelioid cells.Hormone induction can activate lactation gene loci thatare suitable sites for gene targeting in mammary glandreactors. This suggests that more active gene integrationsites could become available for gene targeting in fibro-blast platforms stimulated by appropriate hormones,cytokines or other factors.

Acknowledgement

This work was supported by a grant from the ChineseAcademy of Science KSCX 2-3-08.

References

Ashktorab H, Reed WA, Thonabulsombat C, White KL. Bovine�-casein gene promoter activity and hormonal induction of itsexpression in a mammary epithelial cell line. Transgenics 1999;3:23–9.

Baguisi A, Behboodi E, Melican DT, Pollock JS, Destrempes MM,Cammuso C et al. Production of goat by somatic cell nucleartransfer. Nat Biotechnol 1999;17:456–61.

Berg MN, Dharmarajan AM, Waddell BJ. Glucocorticoids and pro-gesterone prevent apoptosis in the lactating rat mammary gland.Endocrinology 2002;143:222–7.

Blyth BJ, Hollingshead DJ, Amico JA. Time course of induction ofoxytocin messenger ribonucleic acid levels in the hypothalamicparaventricular nucleus of ovariectomized rats following gonadalsteroid administration. Life Sci 1997;60:2427–33.

Chen RH, Wang B, Zhang YZ, Liu W, Zhang JP, Lao WD. Cloningand sequencing of 5#-flanking region of bovine a-s1 casein gene.Chin J Biotechnol 1992;8:218–26.

Cibelli JB, Stice SL, Golueke PJ, Kane JJ, Jerry J, Blackwell C et al.Cloned transgenic calves produced from nonquiescent fetalfibroblasts. Science 1998;280:1256–8.

Fig. 4. Morphological changes in mouse fetus fibroblasts responding to the lactation hormones. The mouse fetus fibroblasts were cultivated withouthormones (control, A), with insulin (B), progesterone (C), oxytocin (D) or the three-hormone mixture (E) for 6 days (50�).

S. Wang et al. / Cell Biology International 27 (2003) 755–760 759

David LS, Robert D, Goldman LA, Leinwand DL. Culture andbiochemical analysis of cells. Cells: a laboratory manual. NewYork: Cold Spring Harbor Laboratory Press, 1998.

Delongeas JL, Trabarel C, Guittin P. Easy procedure for milkcollection in lactating rats. Contemp Top Lab Anim Sci 1997;38:80–3.

Espejel S, Blasco MA. Identification of telomere-dependent“senescence-like” arrest in mouse embryonic fibroblasts. Exp CellRes 2002;276:242–8.

Guyette WA, Matusik RJ, Rosen JM. Prolactin mediated transcrip-tional and post-transcriptional control of casein gene expression.Cell 1979;17:1013–23.

Han CW, Imamura M, Hashino S, Zhu X, Tanaka J, Imai K et al.Differential effects of the immunosuppressants cyclosporin A,FK506 and KM2210 on cytokine gene expression. Bone MarrowTransplant 1995;15:733–9.

Hogan B, Costantuni F, Lacy E. Manipulating the mouse embryo, alaboratory manual, 1st edn New York: Cold Spring HarborLaboratory Press, 1986.

Kanai A, Nonomura N, Yoshimura M, Oka T. DNA-binding proteinsand their cis-acting sites controlling hormonal induction of amouse �-casein: CAT fusion protein in mammary epithelial cells.Gene 1993;126:195–201.

Kimura T, Ito Y, Einspanier A, Tohya K, Nobunaga T, Tokugawa Yet al. Expression and immunolocalization of the oxytocin receptorin human lactating and non-lactating mammary glands. HumReprod 1998;13:2645–53.

McCreath KJ, Howcroft J, Campbell KH, Colman A, Schnieke AE,Kind AJ. Production of gene-targeted sheep by nuclear transferfrom cultured somatic cells. Nature 2000;405:1066–9.

Sambrook J, Fritsch EF, Maniatis T. Molecular cloning: a laboratorymanual, 1st edn New York: Cold Spring Harbor Laboratory Press,1989.

Schmidhauser C, Bissel MJ, Myers CA, Casperson GF. Extracellularmatrix and hormones transcriptionally regulate bovine �-casein 5#sequences in stably transfected mouse mammary cells. Proc NatlAcad Sci U S A 1990;87:9118–22.

Schnieke AE, Kind AJ, Ritchie WA, Mycock K, Scott AR, Ritchie Met al. Human factor IX transgenic sheep produced by transfer ofnuclei from transfected fetal fibroblasts. Science 1997;278:2130–3.

Thomas A, Crowley RS, Amico JA. Effect of progesterone onhypothalamic oxytocin messenger ribonucleic acid levels in thelactating rat. Endocrinology 1995;136:4188–94.

Topper YJ, Freeman CS. Multiple hormone interactions in thedevelopmental biology of the mammary gland. Physiol Rev 1980;60:1049–106.

S. Wang et al. / Cell Biology International 27 (2003) 755–760760