Induction of Dental Pulp Stem Cell Differentiation Into Odontoblasts by Electroporation-Mediated Gene Delivery of Growth/Differentiation Factor 11 (Gdf11)

Home , Dental papilla, Odontoblast

Gene Therapy (2002) 9, 814–818  2002 Nature Publishing Group All rights reserved 0969-7128/02 $25.00 www.nature.com/gt BRIEF COMMUNICATION Induction of dental pulp stem cell differentiation into odontoblasts by electroporation-mediated gene delivery of growth/differentiation factor 11 (Gdf11)

M Nakashima1, K Mizunuma1, T Murakami2 and A Akamine1 1Department of Clinical Oral Molecular Biology, Division of Oral Rehabilitation, Faculty of Dental Science, Kyushu University, Fukuoka, Japan; and 2Department of Orthodontics, Faculty of Dental Science, Kyushu University, Fukuoka, Japan

The long-term goal of dental treatment is to preserve teeth mesenchyme in organ culture. The Gdf11 cDNA plasmid and prolong their function. In dental caries an efﬁcient which was transferred into mesenchymal cells derived from method is to cap the exposed dental pulp and conserve the mouse dental papilla by electroporation, induced the pulp tissue with reparative dentin. We examined whether expression of Dsp. The in vivo transfer of Gdf11 by electro- growth/differentiation factor 11 (GDF11), a morphogen could poration stimulated the reparative dentin formation during enhance the healing potential of pulp tissue to induce differ- pulpal wound healing in canine teeth. These results provide entiation of pulp stem cells into odontoblasts by electropor- the scientiﬁc basis and rationale for gene therapy for endo- ation-mediated gene delivery. Recombinant human GDF11 dontic treatments in oral medicine and dentistry. induced the expression of dentin sialoprotein (Dsp), a differ- Gene Therapy (2002) 9, 814–818. DOI: 10.1038/sj/gt/3301692 entiation marker for odontoblasts, in mouse dental papilla

Keywords: dental pulp; odontoblasts; regeneration; gene therapy; electroporation; growth/differentiation factor 11

The emerging fields of tissue engineering and regenera- gene transfer in gene therapy. Although viral vectors pro- tive medicine seek to replace or repair lost or damaged vide gene transfer with high efficiency, attendant prob- tissues due to disease, trauma and tumors.1 Tissue engi- lems of cellular immunity due to adenoviruses or inser- neering and regeneration requires three key ingredients: tional mutagenesis due to retroviruses have been morphogenetic signals including growth and differen- recognized.13 On the other hand, plasmid-mediated gene tiation factors, responding stem cells and a scaffold of therapy, while minimizing immune responses, is inef- extracellular matrix.1 In general, regeneration and repair ficient. A potential method to overcome this dilemma is recapitulates embryonic development. Bone morphogen- electroporation using pulsed electric fields to deliver etic proteins (BMPs) are multifunctional cytokines and DNA.14–17 Many tissues respond to electroporation, and widely distributed both in skeletal and non-skeletal handling is relatively easy and rapid.18 Electroporation tissues and have a major role in organogenesis. BMPs also has been used to deliver genes to living animals.19–21 have actions beyond bone in neural, renal and cardiac However, electroporation yields only a transient gene development.1,2 BMPs also play a role in differentiation expression and not as efficient as viral vectors.18 of dentin3–8 in teeth. The recent progress in molecular In this investigation, we have optimized gene transfer developmental biology permits the delivery of BMPs of Gdf11 to pulp cells to initiate odontoblast differen- by gene therapy using optimal delivery.9 Growth/ tiation in vitro and reparative dentin formation in vivo by differentiation factor 11 (GDF11) is a novel member of electroporation for the endodontic treatment of pulp the BMP/TGF␤ family.10–12 It was expressed in terminally tissue regeneration and dentin repair. During terminal differentiating odontoblasts,11 implying a role in the dif- differentiation of odontoblasts, the expression of Gdf11 ferentiation of dental pulp stem cells into odontoblasts.11 mRNA by in situ hybridization in mouse tooth germ11 Therefore, we investigated whether the gene transfer of and by RT-PCR in the primary dental pulp cell culture Gdf11 might stimulate odontoblast differentiation and has been demonstrated. The human recombinant GDF11 reparative dentin formation in vitro and in vivo. The protein was used to explore the function of GDF11 in results revealed the potential utility of Gdf11 gene ther- dental pulp cells. Differentiation of dentin-forming odon- apy in endodontic treatment in dentistry. toblasts was monitored by three differentiation markers, Viral vectors and non-viral techniques can be used for dentin matrix protein1 (Dmp1), dentin sialoprotein (DSP) and osteocalcin. The expressions of these genes are known to increase during differentiation of pulp cells into odontoblast-like cells in pulp cell culture.22–24 In Correspondence: M Nakashima, Department of Clinical Oral Molecular Biology, Division of Oral Rehabilitation, Faculty of Dental Science, Kyu- murine developing molars, Dmp1 transcripts are shu University, Fukuoka 812-8582, Japan expressed at the late bud stage, while Dsp mRNA is Received 11 June 2001; accepted 5 February 2002 expressed later at the cap stage. Dmp1 expression is Gdf11 electrotransfection induces dentinogenesis M Nakashima et al 815 decreased in odontoblasts after the appearance of min- As Gdf11 was expressed during differentiation of pulp eral, while Dsp exhibits sustained high expression.25 In cells into odontoblasts, we next investigated the effect of the primary pulp cell culture, Gdf11 was expressed beads soaked in recombinant human GDF11 on the initially on day 0 (not shown), day 5 and very weakly on response of mouse dental papilla mesenchyme derived day 9. Then it declined and disappeared on day 14. It from 17.5 days post coitum (dpc) tooth germ. The reappeared with further differentiation on days 21 and expression of Dsp mRNA was detected in the tissue sur- 28 (Figure 1). The expressions of Dmp1 and Dsp rounding the beads with GDF11 (Figure 2). On the other decreased on day 9 and reappeared on day 21 in associ- ation with Gdf11 expression. Osteocalcin was expressed on day 28 (Figure 1), when mineralization was detected by Alizarin Red staining (data not shown). The expression pattern of Gdf11, Dmp1 and Dsp in vitro, together with the in situ expression results,11 suggest that Gdf11 plays a role in the terminal differentiation of odontoblasts.

Figure 2 Whole-mount RNA in situ hybridization analysis showing the Figure 1 RT-PCR analyses for Gdf11 (product size: 0.5 kb), Dmp1 (0.5 induction of the expression of Dsp mRNA by recombinant human GDF11 kb), Dsp (0.4 kb), Osteocalcin (0.3 kb) and ␤-actin (0.5 kb) in the primary in vitro organ culture. The dental papillae were isolated from enamel epi- dental pulp cell culture. The bovine dental pulp cells were isolated31 and thelium and odontoblastic layer in tooth germ at 17.5 dpc from ICR mouse inoculated at a density of 1 × 105 cells/ml, and 5, 9, 14, 21 and 28 days (CREA, Japan) and cultured on a Nucleopore Track-Etch membrane after cultivation, total RNA were isolated using Trizol (Life Technologies, (Whatman, Springfield Mill, UK) supported by metal grid in Fitton-Jack- Rockville, MD, USA). First-strand cDNA syntheses were performed by son modified BGJb medium (Life Technologies) supplemented with 2% reverse transcription using the SuperScript preamplification system (Life bovine calf serum (JRH Biosciences, Lenexa, KS, USA) and 50 ␮g/ml L- Technologies). The design of the oligonucleotide primers was based on ascorbic acid phosphate (Wako Pure Chemical Industries, Osaka, Japan). published cDNA sequences: Gdf11,11 Dmp1 (Hirst KL, Ibaraki K, Young Agarose beads of 5 ␮l of Affi-Gel Blue Gel (BioRad Laboratories, Hercules, MF, Dixon MJ. GenBank accession number U47636 (unpublished)), CA, USA) was soaked with 50 ␮g of recombinant GDF11 (kindly pro- Dsp,32 Osteocalcin33 and ␤-actin34 (Table 1). PCR amplifications were vided by Genetics Institute, Cambridge, MA, USA) (a), or 50 ␮g of bovine performed for 35 cycles (94°C for 30 s, 65°C for 1 min, 72°C for 1 min) serum albumin (Life Technologies) (b) overnight at 4°C. The beads were for Gdf11, 35 cycles (94°C for 30 s, 60°C for 1 min, 72°C for 1 min) for placed on the dental papillae tissue in the organ culture for 7 days. In situ Dmp1, 36 cycles (94°C for 30 s, 62°C for 1 min, 72°C for 1 min) for hybridization was performed using mouse DIG DSP probe as described DSP, 34 cycles (94°C for 30 s, 55°C for 1 min, 72°C for 1 min) for previously.35 For the mouse probe, the 1080 bp clone (9586–10665) was osteocalcin, and 34 cycles (94°C for 30 s, 55°C for 1 min, 72°C for 1 min) obtained by PCR from mouse genomic DNA (Clontech, Palo Alto, CA, for ␤-actin. The PCR conditions used were optimized and standardized. USA) using primers, DSP-5’-3 (5’-CGCGAATTCGACAGGAGA- Note the expressions of Dmp1 and Dsp, differentiation markers for dentin- GATGTGCAGACT-3’) and DSP-3’-4 (5’-TACGGATCCAGGA forming odontoblasts was associated with reappearance of Gdf11 on day GGTGAGCACCTGAGAA-3’), subcloned into pBlueScript II SK(-), lin- 21. The mineralization was confirmed by Alizarin Red staining. The earized with HindIII and transcribed by T3 polymerase for the antisense experiment was repeated three times and one representative experiment probe. The experiment was repeated five times and one representative is presented. experiment is presented.

Gene Therapy Gdf11 electrotransfection induces dentinogenesis M Nakashima et al 816

Figure 4 Whole-mount RNA in situ hybridization analysis showing the induction of the expression of Dsp mRNA, 7 days after electrotransfection with Gdf11 in vitro. The mouse dental papillae mesenchyme transfected with the pEGFP vector with TIMP promoter (a, c) or with the mouse Gdf11 driven by TIMP promoter-pEGFP vector (b, d). The plasmid pEGFPN3-TIMP-mGdf11 was constructed by ligating the 140 bp PCR product of TIMP and the 1.0 kb PCR product of mouse Gdf11.11 The plasmid was conﬁrmed by sequence. The puriﬁcation was performed by Wizard Plus Midi Prep (Promega, Madison, WI, USA) and the plasmids were adjusted to a concentration of 20 ␮g/␮l in phosphate buffer saline. The dental papillae were isolated in the same manner as described in Fig- ure 2, and 0.5 ␮l of the ice-cold plasmid was applied on the dental papillae. The condition of the electrotransfection was eight square-wave pulses at a frequency of 1 Hz, with a pulse length of 10 ms and 10 V. Note the much stronger expression of Dsp mRNA in the mesenchyme electro- transfected with Gdf11 (d) than that with GFP only (c). The experiment was repeated three times and one representative experiment is presented.

by electroporation in primary pulp cell culture and in organ cultures of tooth germ (Figure 3). A high transfection efficiency was observed using green fluorescent protein (GFP) as marker for gene transfer. GFP was not detected in the culture with the pEGFP vector without electrotransfection and in the cultures without the pEGFP vector with electrotransfection (data not shown). A regi- men of eight square-wave pulses were delivered at a frequency of 1 Hz, with a pulse length of 40 ms and 70 V Figure 3 The subconfluent primary pulp cells (a) and the tooth germ (b) in the multi-layered primary pulp cells, and with a pulse showing the high efficiency of electrotransfection by the pEGFP-N3 length of 10 ms and 100 V in the subconfluent primary (Clontech), 1 day after cultivation. Eight square-wave pulses at a fre- pulp cells (Figure 3a). In the organ cultures of mouse quency of 1 Hz, with a pulse length of 10 ms, and 100 V (a) or 10 V tooth germ, on the other hand, optimal delivery of Gdf11 (b). The optimal conditions for electrotransfection were examined in the subconfluent primary pulp cells using a comb-type electrode with a 5 mm plasmid was obtained by eight square-wave pulses deliv- distance (gap) between stainless steel wires, 30 mm in diameter, using a ered at a frequency of 1 Hz, with a pulse length of 10 electroporator, EDIT-TYPE CUY21 (Tokiwa Science, Fukuoka, Japan). ms and 10V (Figure 3b). Seven days after electroporation, The optimal conditions were determined by counting the number of GFP expression of Dsp mRNA was detected in the organ cul- particles, under a fluorescent dissection microscope (Leica, Heerbrugg, tures transfected with Gdf11 (Figure 4d). The organ Switzerland) with GFP fluorescent filter at 488 nm. For the electro- cultures transfected with pEGFP alone and those with transfection in the organ culture, a stainless steel electrode needle with a 1 mm gap (Tokiwa Science) was placed on the surface of the tissue. pEGFP-Gdf11 without electrotransfection expressed no Dsp mRNA (data not shown). The in vivo electroporation of Gdf11 in the amputated hand, as expected, no Dsp mRNA is seen around control pulp of canine teeth has shown that the pulp cells differ- BSA beads, suggesting that GDF11 stimulates differen- entiated into osteodentinoblasts and secreted osteodentin tiation of pulp stem cells into odontoblast lineage. matrix around them and formed osteodentin 1 month The current practice of endodontic treatment for pulp after surgery (Figure 5a, c). In the control teeth trans- involves use of calcium hydroxide for pulp capping. It fected with pEGFP only, the differentiation of the pulp was of interest to utilize Gdf11 gene delivery for endo- cells was not observed (Figure 5b). Surgical amputation dontic therapy. The plasmid, mouse Gdf11 driven in a of pulp and the use of a capping agent, such as calcium pEGFP vector with TIMP promoter site was transfected hydroxide stimulates pulp cells to differentiate into odon-

Gene Therapy Gdf11 electrotransfection induces dentinogenesis M Nakashima et al 817 Table 1 Primers used for RT-PCR ampliﬁcation on Gdf11 and differentiation markers for odontoblasts

Name 5Ј Sequence 3’

Human GDF11 Forward 5Ј-CAAGTCGCAGATCTTGAGCA-3Ј Reverse 5Ј-CACTTGCTTGAAGTCGATGC-3Ј Bovine Dmp1 Forward 5Ј-AGCCCAGAGTCCACTGAAGA-3Ј Reverse 5Ј-CTCCCATGGAGGGTGTTCTA-3Ј Human DSP Forward 5Ј-CAGGACCATGGGAAAGAAGATGATCA-3Ј Reverse 5Ј-TGTCTTGACATTGCCTTTGC-3Ј Bovine Osteocalcin Forward 5Ј-ATGAGAACCCCCATGCTGCTC-3Ј Reverse 5Ј-GTAGAAGCGCCGATAGGCTTCCT-3Ј Mouse ␤-actin Forward 5Ј-GTGGGCCGCTCTAGGCACCAA-3Ј Reverse 5Ј-CTCTTTGATGTCACGCACGATTTC-3Ј

a large amount of osteodentin formation was observed. The present result of GDF11 gene therapy in pulp has demonstrated osteodentin matrix formation as observed with BMP protein therapy.5 The osteodentin matrix pro- duction is a biological response to repair induced by Gdf11 gene therapy. Immediately adjacent to the electrode there was a lake of erythrocytes in a plasma clot possibly due to thermal effects (Figure 5d). It is possible to avoid the thermal effects of electroporation by the use of ultrasound-mediated gene delivery30 during endodontic treatment. In conclusion, transfer of Gdf11 gene by electroporation to pulp cells in an amputated tooth in vivo induced new reparative dentin. We are aware of the important implications for endodontic treatment by gene therapy in dentistry.

Acknowledgements The authors are grateful to K Hirakawa and T Yamauchi of Tokiwa Science for their help and to Genetic Institutes for their kind gift of recombinant human GDF11. This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports Figure 5 The in vivo electroporation of Gdf11 in the amputated pulp of and Culture, Japan, No. 11470406. dogs. Tissues were examined after 1 month. A total of 20 teeth from five young adult dogs weighting 15–18 kg were used. Surgical anesthesia was obtained by intravenous administration of 35 mg pentobarbital sodium References per kg body weight. The canine teeth were treated, and an exposure was made using a diamond round burr at high speed. After washing with a 1 Reddi AH. Role of morphogenetic proteins in skeletal tissue 5% solution of sodium hypochlorite and a 3% solution of H2O2, the ampu- engineering and regeneration. Nature Biotech 1998; 16: 247–252. tation was carried out by means of a round metal burr at low speed. 2 Kirker-Head CA. Potential applications and deliver strategies Twenty micrograms of pEGFP/Gdf11 plasmid (a, c, d) and pEGFP plas- for bone morphogenetic proteins. Adv Drug Del Rev 2000; 43: mid as a control (b) in phosphate buffer saline was applied by electropor- 65–92. ation in the amputated pulp tissue in a similar condition as described 3 Nakashima M. The induction of reparative dentine in the ampu- above; 10 square-wave pulses were delivered at a frequency of 1 Hz, with tated dental pulp of the dog by bone morphogenetic protein. a pulse length of 1 ms and 10 V. The cavity was filled with zinc phosphate Archs Oral Biol 1990; 35: 493–497. cement and composite resin. Note the formation of the thick osteodentin 4 Nakashima M. Induction of dentin formation on canine ampu- matrix (OD) (a, c) beneath the amputated site (arrows) in response to tated pulp by recombinant human bone morphogenetic protein Gdf11 gene therapy. The newly induced reparative dentin with intercellu- lar matrix contains only few dentinal tubules (c). The proliferation of pulp (BMP)-2 and -4. J Dent Res 1994; 73: 1515–1522. tissue in the cavity above the amputated site (a, b). No osteodentin matrix 5 Nakashima M. Induction of dentine in amputated pulp of the formation in the control (b). Pulp tissue (Pu). The plasma clot (Cl) dogs by recombinant human bone morphogenetic proteins-2 immediately adjacent to the electrode (d). and -4 with collagen matrix. Archs Oral Biol 1994; 39: 1085–1089. 6 Rutherford RB et al. Induction of reparative dentin formation in monkeys by recombinant human osteogenic protein-1. Arch Oral toblasts and produce a small amount of dentin matrix. Biol 1993; 38: 571–576. Some cells degenerate and leave a void in the spaces that 7 Rutherford RB et al. Time course of the induction of reparative 26 dentin formation in monkeys by recombinant human osteogenic they formerly occupied. This kind of reparative dentin protein-1. Arch Oral Biol 1994; 39: 833–838. containing few or no dentinal tubules is called osteoden- 8 Rutherford RB, Spangberg L, Tucker M, Charette M. Transdenti- 27–29 tin, distinguishing from tubular dentin. After appli- nal stimulation of reparative dentine formation by osteogenic cation of morphogens, such as recombinant human BMP2 protein-1 in monkeys. Arch Oral Biol 1995; 40: 681–683. and BMP4 proteins with collagen matrix,5 stimulation of 9 Winn SR, Hu Y, Sfeir C, Hollinger JO. Gene therapy approaches

Gene Therapy Gdf11 electrotransfection induces dentinogenesis M Nakashima et al 818 for modulating bone regeneration. Adv Drug Del Rev 2000; 42: proteins and differentiation of dental pulp cells. Dev Biol 1994; 121–138. 162:18–28. 10 McPherron AC, Lawler AM, Lee SJ. Regulation of 23 Yokose S et al. Establishment and characterization of a culture anterior/posterior patterning of the axial skeleton by system for enzymatically released rat dental pulp cells. Calcif growth/differentiation factor 11. Nature Genet 1999; 22: 260–264. Tissue Int 2000; 66: 139–144. 11 Nakashima M, Toyono T, Akamine A, Joyner A. Expression of 24 Couble ML et al. Odontoblast differentiation of human dental growth/differentiation factor 11, a new member of the pulp cells in explant cultures. Calcif Tissue Int 2000; 66: 129–138. BMP/TGF␤ superfamily during mouse embryogenesis. Mech 25 D’Souza RN et al. Gene expression patterns of murine dentin Dev 1999; 80: 185–189. matrix protein 1 (Dmp1) and dentin sialophosphoprotein 12 Gamer LW et al. A novel BMP expressed in developing mouse (DSPP) suggest distinct developmental functions in vivo. J Bone limb, spinal cord, and tail bud is a potent mesoderm inducer in Min Res 1997; 12: 2040–2049. Xenopus embryos. Dev Biol 1999; 208: 222–232. 26 Avery JK. Dentin. In: Bhaskar SN (ed.). Orban’s Oral Histology 13 Crystal RG. Transfer of genes to humans: early lessons and and Embryology. Mosby: St Louis, 1980, pp 107–140. obstacles to success. Science 1995; 270: 404–410. 27 Baume LJ. Dental pulp conditions in relation to carious lesions. 14 Neumann EM, Schaefer-Ridder YW, Hofschneider PH. Gene Int Dent J 1970; 20: 309–337. transfer into mouse lyoma cells by electroporation in high elec- 28 Takuma S, Yanagisawa T, Lin WL. Ultrastructural and micro- tric ﬁelds. EMBO J 1982; 1: 841–845. analytical aspects of developing osteodentin in rat incisors. Cal- 15 Fromm M, Taylor LP, Walbot V. Expression of genes transferred cif Tissue Res 1977; 24: 215–222. into monocot and dicot plant cells by electroporation. Proc Natl 29 Tziafas D, Smith AJ, Lesot H. Designing new treatment stra- Acad Sci USA 1985; 82: 5824–5828. tegies in vital pulp therapy. J Dent 2000; 28:77–92. 16 Andreason GL, Evans GA. Introduction and expression of DNA 30 Newman CM, Lawrie A, Brisken AF. Cumberland DC. Ultra- molecules in eukaryotic cells by electroporation. Biotechniques sound gene therapy: on the road from concept to reality. Echo- 1988; 6: 650–660. cardiography 2001; 18: 339–347. 17 Chowrira GM, Akella V, Fuerst PE, Lurquin PF. Transgenic 31 Nakashima M. Establishment of primary cultures of pulp cells grain legumes obtained by in planta electroporation-mediated from bovine permanent incisors. Archs Oral Biol 1991; 36:655– gene transfer. Mol Biotechnol 1996; 5:85–96. 663. 18 Muramatsu T, Nakamura A, Park HM. In vivo electroporation: 32 Gu K et al. Molecular cloning of a human dentin sialophospho- a powerful and convenient means of nonviral gene transfer to protein gene. Eur J Oral Sci 2000; 108:35–42. tissues of living animals. Int J Mol Med 1998; 1:55–62. 33 Kiefer MC, Saphire AC, Bauer DM, Barr PJ. The cDNA and 19 Weaver JC. Electroporation: a general phenomenon for manipu- derived amino acid sequences of human and bovine bone Gla rating cells and tissues. J Cell Biochem 1993; 51: 426–435. protein. Nucleic Acids Res 1990; 18: 1909. 20 Hofmann GA. Instrumentation. In: Nickoloff JA (ed.). Methods 34 Alonso S, Minty A, Bourlet Y, Buckingham M. Comparison of in Molecular Biology. Electroporation Protocols for Microorganisms. three actin-coding sequences in the mouse; evolutionary Humana Press: Totowa, NJ, 1995, pp 27–45. relationships between the actin genes of warm-blooded ver- 21 Dev SB, Hofmann GA. Electrochemotherapy – a novel method tebrates. J Mol Evol 1986; 23:11–22. of cancer treatment. Cancer Treat Rev 1994; 20: 105–115. 35 Platt KA, Michaud J, Joyner AL. Expression of the mouse Gli 22 Nakashima M, Nagasawa H, Yamada Y, Reddi AH. Regulatory and Ptc genes is adjacent to embryonic sources of hedgehog sig- role of transforming growth factor-␤, bone morphogenetic pro- nals a conservation of pathways between ﬂies and mice. Mech tein-2, and protein-4 on gene expression of extracellular matrix Dev 1997; 62: 121–135.

Gene Therapy