Generation and Characterization of DSPP-Cerulean/DMP1-Cherry reporter mice.

Abstract

To gain a better understanding of the progression of progenitor cells in the odontoblast lineage, we have examined and characterized the expression of a series of GFP reporters during odontoblast differentiation. However, previously reported GFP reporters (pOBCol2.3-GFP, pOBCol3.6-GFP, and DMP1-GFP), similar to the endogenous proteins, are also expressed by bone-forming cells, which made it difficult to delineate the two cell types in various in vivo and in vitro studies. To overcome these difficulties we generated DSPP-Cerulean/DMP1-Cherry transgenic mice using a bacterial recombination strategy with the mouse BAC clone RP24-258g7. We have analyzed the temporal and spatial expression of both transgenes in tooth and bone in vivo and in vitro. This transgenic animal enabled us to visualize the interactions between odontoblasts and surrounding tissues including dental pulp, ameloblasts and cementoblasts. Our studies showed that DMP1-Cherry, similar to Dmp1, was expressed in functional and fully differentiated odontoblasts as well as osteoblasts and cementoblasts. Expression of DSPP-Cerulean transgene was limited to functional and fully differentiated odontoblasts and correlated with the expression of Dspp. This transgenic animal can help in the identification and isolation of odontoblasts at later stages of differentiation and help in better understanding of developmental disorders in dentin and odontoblasts.

1. Introduction:

Dentinogenesis is regulated by a single layer of highly differentiated post-mitotic odontoblasts originating from the neural crest-derived cells of the dental papilla (Arana-Chavez & Massa, 2004; Kawashima & Okiji, 2016). The differentiation of odontoblasts from dental papilla cells involves several intermediate steps that are dependent on and regulated by epithelial signals (Lesot et al., 2001; Ruch, Lesot, & Begue-Kirn, 1995; Thesleff, Keranen, & Jernvall, 2001). During this process, dental papilla in close proximity to the epithelial-mesenchymal interface first forms pre-odontoblasts that gradually differentiate into functional/secretory odontoblasts and eventually fully differentiated odontoblasts (Lesot et al., 2001; Ruch et al., 1995; Thesleff et al., 2001). In mice, the steps between the formation of pre-odontoblasts and highly differentiated odontoblasts are completed within 6–10 hours (Lesot et al., 2001; Ruch et al., 1995; Thesleff et al., 2001).

The formation of the mineralized matrix in dentin is similar to bone and requires Type I Collagen and non-collagenous proteins (NCPs). The SIBLING (small integrin-binding ligand N-linked glycoprotein) family of proteins represents the most abundant group of NCPs in bone and dentin. The SIBLING family includes osteopontin (OPN), bone sialoprotein (BSP), dentin matrix protein 1 (DMP1), dentin sialophosphoprotein (DSPP), and matrix extracellular phosphoglycoprotein (MEPE) (Fisher & Fedarko, 2003; Staines, MacRae, & Farquharson, 2012).

Although low levels of DSPP have been detected in other cells including bone-forming cells, (Prasad, Butler, & Qin, 2010), high levels of DSPP and DSP expression are the hallmark of odontoblast differentiation and are routinely used to distinguish differentiated odontoblasts from other cell types including osteoblasts (Prasad et al., 2010; Suzuki, Haruyama, Nishimura, & Kulkarni, 2012).

The DSPP gene encodes a single large precursor protein DSPP, which undergoes proteolytic cleavage forming dentin sialoprotein (DSP) and dentin phosphoprotein (DPP), representing the N-terminus and C-terminus of DSPP, respectively (MacDougall et al., 1997; Suzuki et al., 2012). These proteins have distinct roles in dentin mineralization. DSP regulates the initiation of dentin mineralization, and DPP plays a role in the maturation of mineralized dentin (Suzuki et al., 2012). The essential functions of DSPP in dentinogenesis were highlighted by the discovery of the linkage between multiple point mutations in the human DSPP gene with human dentinogenesis imperfecta (DGI) and dentin dysplasia (DD) (Prasad et al., 2010; Suzuki et al., 2012). Mice lacking DSPP and/or its processed fragments (DSP and DPP) also displayed severe dentin defects including a widened predentin along with a narrow and hypomineralized dentin

In teeth, Dspp/DSPP is first expressed at low levels in secretory/functional odontoblasts and then at high levels in terminally differentiated odontoblasts (Prasad et al., 2010; Suzuki et al., 2012).

DMP1 is another member of the SIBLING family with essential roles in the mineralization of both dentin and bone (Fisher & Fedarko, 2003; Staines et al., 2012). Unlike DSPP, which is predominantly expressed at high levels in odontoblasts, DMP1 is also highly expressed by osteocytes, chondrocytes, and preosteoblasts (Qin, D’Souza, & Feng, 2007; Suzuki et al., 2012). In developing teeth, the expression of Dmp1 is detected earlier than that of Dspp, and Dmp1 null mice displayed abnormalities in dentin quite similar to those seen in Dspp null mice, though less severe (Ye et al., 2004). The levels of Dspp in Dmp1 null mice were decreased as compared to wild-type (Ye et al., 2004), and the transgenic expression of Dspp rescued the tooth and alveolar bone defects of Dmp1 KO mice (Qin et al., 2007). These observations together with the identification of DMP1-response elements in the Dspp gene suggested uggests that the DMP1 transcription factor regulates DSPP expression in odontoblasts (Narayanan, Gajjeraman, Ramachandran, Hao, & George, 2006). Taken together, these studies clearly demonstrate the importance of both Dspp and Dmp1 in the regulation of odontoblast differentiation and formation of mineralized dentin matrix.

To gain a better understanding of the progression of progenitor cells in the odontoblast lineage, we have examined and characterized the expression of a series of GFP reporters during odontoblast differentiation. These studies together have shown the stage-specific activation of these transgenes during odontoblast differentiation in vivo and in vitro. pOBCol2.3-GFP and pOBCol3.6-GFP transgenes were activated at early stages of odontoblast differentiation (i.e. polarizing odontoblasts and prior to the expression of Dmp1 and Dspp) (Balic, Aguila, & Mina, 2010). DMP1-GFP was first activated in functional/secretory odontoblasts (cells expressing Dmp1 and low levels of Dspp) (Balic & Mina, 2011). All three transgenes (pOBCol2.3-GFP, pOBCol3.6-GFP, and DMP1-GFP) were also expressed at high levels in fully differentiated/mature odontoblasts, and their temporal and spatial patterns of expression mimicked those of endogenous transcripts and proteins (Balic et al., 2010; Balic & Mina, 2011).

However, these transgenes, similar to the endogenous proteins are also expressed by bone-forming cells (Kalajzic et al., 2004; Kalajzic et al., 2002). In bone, the pOBCol3.6-GFP transgene was expressed by pre- and early-osteoblasts, whereas pOBCol2.3-GFP and DMP1-GFP transgenes were expressed in mature osteoblasts and osteocytes (Kalajzic et al., 2004; Kalajzic et al., 2002).

The expression of these transgenes in both tissues makes it difficult to delineate the two cell types in various in vivo and in vitro studies. To overcome these difficulties and to be able to distinguish between dentin and bone forming cells, we generated DSPP-Cerulean/DMP1-Cherry transgenic mice using a bacterial recombination strategy with the mouse BAC clone RP24-258g7. In this study, we have analyzed the temporal and spatial expression of both transgenes in the developing teeth and bone in vivo, and during in vitro mineralization in primary pulp cultures.

2. RESULTS:

2.1. Expression of DSPP-Cerulean and DMP1-Cherry transgenes in developing teeth and alveolar bones in vivo:

The patterns of expression of both transgenes were examined in three independent lines in developing teeth and alveolar bones. DMP1-Cherry and DSPP-Cerulean transgenes were not detected in the dental tissues during the initiation, bud, cap and early bell stages of tooth development (E10–18) (data not shown). At the late bell stage (E19) DMP1-Cherry was expressed at low intensity in odontoblasts at the tip of the cusp of the first mandibular molar (data not shown). At the secretory stage of crown formation (P1- P6), DMP1-Cherry and DSPP-Cerulean transgenes were expressed at high intensity in the entire layer of odontoblasts covering the dental pulp in the unerupted molars (Figures 1 and Supplemental Figure 1 A & B). DMP1-Cherry expression was also detected at high levels in osteoblasts and osteocytes within the alveolar bone (Figures 1A, Figure 2A&D). DSPP-Cerulean was also detected transiently in the pre-secretory ameloblasts (Figure 1Af,i).

Figure 1: Expression of DSPP- Cerulean and DMP1-Cherry transgenes in developing teeth and alveolar bone in vivo.

A. Representative images of sagittal sections through lower jaw at P6. Images are from the same section visualized under Bright field (a, b) and epifluorescent microscopy with filter cubes to detect Cherry (c-e) and Cerulean signals (f-h). Higher magnifications of areas containing first molars are shown in b, d, e, g, h, k, l. DMP1-Cherry is detected in the odontoblasts lining the dental pulp of molars and incisors; as well as in the osteoblasts and osteocytes of the surrounding alveolar bone (c-e). DSPP- Cerulean is expressed in the odontoblasts of molars and incisors but not in the alveolar bone (d-f). Note the low levels of DSPP- Cerulean in ameloblasts in the cervical loop region (indicated by arrows in h and i). Also note the co-expression of DMP1-Cherry and DSPP- Cerulean in odontoblasts in incisor and molars (g-i).

B. Representative images of sagittal sections through the first molar and surrounding alveolar bone at P1. Images are from the same section visualized under epifluorescent microscopy with filter cubes to detect Cherry and Cerulean. Section in c was stained with anti-GFP antibody and visualized with filter cube for GFP. Note that DMP1-Cherry is detected in the odontoblasts and surrounding alveolar bone (a). DSPP-Cerulean is expressed in the odontoblasts of molars but not in the alveolar bone (b). The expression of GFP (c) is similar to DSPP-Cerulean (b). Note that GFP expression with high intensity is also detected in areas where DSPP-Cerulean expression is detected at low intensity (indicated by arrows in b & c). Also note the lack of GFP expression in the surrounding alveolar bone conforming the specificity of the antibody to Cerulean variant. Scale Bars in B (a-d) = 200μm.

In all images, the pulp chambers are denoted by dashed lines. I, incisor; M1, first molar; Ab, alveolar bone. Scale Bars in A (a, d, g) = 500μm; and in A (b-c, e-f, h-i) and B (a-d) = 200μm.

Figure 2: Detailed confocal-microscopy analysis of DMP1-Cherry/DSPP-Cerulean in adult 8 weeks old animals.

A. Sagittal section through the mandibular arch. Note the co-expression of DMP1-Cherry and DSPP-Cerulean in odontoblasts covering the dental pulp in incisor and molars. Note the increased expression of transgenes in apical-to-incisal direction. Also note the expression of DMP1-Cherry in osteoblasts and osteocytes of the alveolar bones. Scale bar = 500 μm

B. Odontoblast processes in incisor pre-dentin and dentin during the early secretory stage. A higher magnification of odontoblasts showing cells expressing both transgenes (white) (a), only DMP1-Cherry (red) and only DSPP-cerulean (blue). Note the elaborate branched odontoblast processes from all cells in pre-dentin and dentin (b and b’). Scale Bar in a, b = 50 μm and in a’, b’ - b’” = 20 μm

C. Odontoblast processes in molars. Sagittal section through the adult (8 weeks old) mouse first molar showing co-expression of DMP1-Cherry and DSPP-Cerulean in odontoblasts and odontoblasts process (a). Inset a’ shows the detailed image of the odontoblast layer and odontoblast processes in the region close to the cemento-enamel junction. Scale Bar in C(a) = 500 μm and in C(a’)=30um

D. Cross section through the apical part of the root from adult mouse molar (8 weeks old). Note the co-expression of DMP1-Cherry and DSPP-Cerulean in odontoblasts (a). Also note the expression of DMP1-Cherry in cementoblasts and alveolar bone. Inset a’ shows expression of DMP1-Cherry in alveolar osteocytes and cementoblasts. Scale bars D(a) = 500 μm, and in D(a’) = 30 μm.

E. Odontoblast processes and their interactions with pulp cells. (a-c) High magnifications of branched processes extending from basal end of the odontoblasts into dental pulp (all indicated by arrows) at different depths. Scale Bar in E(a)= 50 μm. Scale bar in E(a’-c)= 20 μm,

F. Odontoblast processes and their interactions with surrounding mineralized cells. Root dentin is structurally interconnected with surrounding cementoblasts via dentinal tubules and its terminal branching region (a). Separated channels (Cherry/Cerulean) are shown in panels a, a’. Dentino-cemental border and multiple contacts between odontoblasts and cementoblasts of the apical molar root is shown on tilted cross section (b). In early developing odontoblasts, before the dentin synthesis is initiated, odontoblast processes contacting neighboring ameloblasts (c and c’). Scale bars: a,a’ = 30 μm and b, c, c’ = 50 μm.

LiCL, lingual cervical loop, LaCL, labial cervical loop, cb, cementoblasts, pl, periodontal ligament, Ab, alveolar bone, and M, molar

In the erupted molars DMP1-Cherry and DSPP-Cerulean transgenes were co-expressed at high intensity in the entire layer of odontoblasts lining the crown and root (Figure 2C). DMP1-Cherry was expressed in osteocytes of the alveolar bones as well as cementoblasts (Figure 2C&D)

Developing incisors at all stages of development showed an apical to incisal gradient of expression of both transgenes. DMP1-Cherry and DSPP-Cerulean expression were not detected in the apical ends of the incisors, and were detected in odontoblasts at the midpoint and incisor tips (Figures 1A and Figure 2A).

To ensure that all DSPP-Cerulean⁺ cells (including possible low expressing cells) are visualized adequately, adjacent sections were also processed for immunohistochemistry using anti-GFP antibody (Figure 1B). The anti-GFP antibody was raised against the N-terminal of GFP as an antigen and does not detect RFP proteins, dsRED or Cherry (Cell Biolabs, Inc. San Diego, CA). The absence of GFP staining in DMP1-Cherry⁺ osteoblasts/osteocytes (Figure 1Bc) confirmed the specificity of the antibody to Cerulean. The pattern of staining with anti-GFP antibody (GFP⁺) cells was very similar to the expression pattern of DSPP-Cerulean⁺ cells in vivo (Figure 1Bb–c). The patterns of expression of these transgenes in developing teeth and alveolar bone were similar between the three lines (Supplemental Figure 2).

2.2. Interactionsof DMP1-Cherry/DSPP-cerulean expressing cells with surrounding tissues:

High resolution analysis of the adult (8 weeks old) mandibles with confocal microscopy revealed the morphological features of DMP1-Cherry/DSPP-Cerulean⁺ cells in incisors and molars (Figure 2). These studies showed the heterogeneity of the odontoblast layer containing DMP1-Cherry⁺, DSPP-Cerulean⁺ and DMP1-Cherry⁺/DSPP-Cerulean⁺ cells (Figure 2B). All three types of cells extended long and branched processes from their apical ends into predentin and dentin (Figure 2B&C).

Confocal microscopy also revealed many fine structural details of odontoblast processes and contacts between the terminal branches of odontoblast process and surrounding tissues (Figure 2E & F). Odontoblasts extended many long and short branched processes from their basal ends into the subodontoblast region of dental pulp (Figure 2E). In addition, odontoblast processes from the apical ends extended towards cementum making close contacts with cementoblasts (Figure 2F). Prior to deposition of mineralized matrices in the incisor, short odontoblast processes penetrated the ameloblast layer (Figure 2F).

Thus, these studies provide structural details of odontoblast processes emanating from both apical and basal side of the odontoblasts. The thin processes that come in close proximity to ameloblasts prior to matrix deposition and the additional processes extending from the basal side of odontoblasts into dental pulp are consistent with results reported in PLP-CreERT2/R26YFP mouse (Khatibi Shahidi et al., 2015). These together provide additional support for the highly integrated cellular network of communication of odontoblasts with the pulp and other neighboring cells in teeth (Bleicher, 2014; Farahani, Simonian, & Hunter, 2011; Kawashima & Okiji, 2016; Khatibi Shahidi et al., 2015).

2.3. Expression of DSPP-Cerulean and DMP1-Cherry transgenes during the mineralization of primary dental pulp cultures in vitro:

The rapid transition and the close proximity of cells at different stages of differentiation in the developing teeth in vivo make it difficult to fully appreciate the stage-specific activation of these transgenes during odontoblast differentiation. To gain insight into differences in the stage of activation of these transgenes, the temporal and spatial expression of both, the DMP1-Cherry and DSPP-Cerulean transgenes was examined during in vitro mineralization in primary pulp cultures. In these experiments, expression of the two transgenes were examined and correlated with the onset and subsequent growth of the Calcein Green-mineralized nodules in real time in the same live cultures over time.

In cultures, DMP1-Cherry expression was detected in a few cells at day 7 (Figure 3). With the onset and subsequent growth of CG-stained mineralized matrix, there were continuous increases in the mean intensity of DMP1-Cherry (Figure 3A, 4A & B). DMP1-Cherry⁺ cells were detected within CG-stained nodules and in areas between nodules (Figure 3A, overlay).

Figure 3: Expression of DMP1-Cherry and DSPP-Cerulean transgenes during the in vitro mineralization of primary pulp cultures:

A. Expression of DMP1-Cherry transgene during in vitro mineralization of primary cultures derived from dental pulp. Representative images of pulp cultures at different time points analyzed under phase contrast, epifluorescent light using appropriate filters for detection of DMP1-Cherry and Calcein green. DMP1-Cherry is expressed in a few cells at day 7. Between days 10-14, DMP1-Cherry is expressed at high intensity in Calcein green-stained mineralized nodules. At these time points DMP1-Cherry expression is also detected in area between mineralized nodules. Scale bar: 200μm.

B. Representative images of pulp cultures at different time points analyzed under epifluorescent light using appropriate filters. These cultures were also stained with anti-GFP antibody. DSPP-Cerulean⁺ and GFP⁺ cells are first detected at day 10 with increases thereafter. Note that anti-GFP antibody labeled cells that express DSPP-Cerulean at high and low levels. Also note the co-expression of GFP⁺ and DMP1-Cherry⁺ (indicated by arrows) in only fraction of cells in dental pulp cultures (row labeled Cherry-GFP).

C: Bright field and epifluorescent images of a culture at day 21 shows that labeled cells are detected only within mineralized areas (outside the dashed lines). Scale bar: 100μm.

Figure 4: Changes in the percentages of cells expressing transgenes.

A. Histogram showing changes in the intensity of Calcein green staining (absolute values) of primary cultures. Note the progressive increase in the intensity of Calcein green after day 10.

B. Histograms showing progressive increase in the intensity of DMP1-Cherry expression during mineralization of primary pulp cultures.

C.The percentages of DMP1-Cherry⁺ cells were examined in unsorted primary pulp cultures and FACS-sorted DMP1-Cherry⁻ cultures. Cultures were processed for FACS analysis at various time points. The percentages of DSPP-Cerulean⁺ and DSPP-GFP⁺ cells were examined in unsorted primary pulp cultures by FACS analysis and immunostaining with anti-GFP antibody, respectively. Approximately 20,000-30,000 Hoechst⁺ cells were counted from 10-40 different representative areas of the culture. Negative controls included primary BMSC cultures derived from the DSPP-Cerulean littermates stained with anti-GFP antibody and primary dental pulp cultures derived from the transgenic littermates without addition of anti-GFP antibody. The gating strategy for FACS analysis of DMP1-Cherry⁺ and DSPP-Cerulean⁺ cells are shown in supplemental Figure 4 was used. Results represent mean ± SEM of at least three independent experiments; ND, not detected.

DSPP-Cerulean⁺ cells were examined using filter cube for detection of Cerulean and after immunocytochemistry with anti-GFP antibodies (Figure 3B). DSPP-Cerulean⁺ and DSPP-GFP⁺ cells were not detected at early time point, and were first detected at day 10 with subsequent increases thereafter. The cells expressing DSPP-Cerulean at low levels in vitro were more clearly identifiable with anti-GFP antibody staining (Figure 3B). A fraction of cells within the mineralized nodules co-expressed GFP⁺ and DMP1-Cherry⁺ (Figure 3B, overlay). DSPP-Cerulean⁺ and GFP⁺ cells were localized primarily within mineralized nodules (Figure 3C).

Changes in the percentage of DMP1-Cherry⁺ and DSPP-Cerulean⁺ cells in primary pulp cultures were examined by flow cytometry and immunocytochemistry, respectively (Figure 4C). Examination of unsorted cultures showed that freshly isolated pulp cultures contained about 14% DMP1-Cherry⁺ cells (Figure 4C and Supplemental Figure 1 C and D) but no DSPP-Cerulean⁺/GFP⁺cells (Figure 4C). The percentage of DMP1-Cherry⁺ cells in the unsorted population decreased at day 7 and followed by increases at days 10 and 14. DSPP-Cerulean⁺/GFP^+. cells appeared at day 10 and increased thereafter (Figure 4C).

The decrease in DMP1-Cherry⁺ cells at day 7 most likely reflects the non-proliferative nature or lack of survival of these post-mitotic cells both in vivo and in vitro (Balic & Mina, 2011). Examination of cultures from FACS-sorted DMP1-Cherry negative population showed the absence of DMP1-Cherry⁺ at day 0 and continuous increases thereafter (Figure 4C).

These studies showed continuous increases in the percentages of DMP1-Cherry⁺, DSPP-GFP⁺ and DMP1-Cherry⁺/DSPP-Cerulean⁺ populations during in vitro mineralization of pulp cells. (Figure 4C). Time-lapse confocal imaging over 36 hours showed the transition of DMP1-Cherry⁺ cells to DMP1-Cherry⁺/DSPP- Cerulean⁺ cells (Supplemental Figure 3).

The expression of these transgenes was also compared with the expression of endogenous Dmp1 and Dspp by RT-PCR using primers listed in Supplemental Figure 4A. The expression of transgenes was well correlated with the expression of Dmp1 and Dspp. Similar to endogenous expression of Dmp1, DMP1-Cherry was expressed at day 7 with further increases till day 21. Dspp and DSPP-Cerulean expression was detected later at day 10, with further increases till day 21 (Supplemental Figure 4B).

These in vitro studies showed differences in the stage of activation of the transgenes during odontoblast differentiation in that the DMP1-Cherry⁺ transgene is activated at an earlier developmental stage (i.e., functional odontoblasts) than DSPP-Cerulean transgene that is activated in fully differentiated odontoblasts.

2.4. Expression of DSPP-Cerulean and DMP1-Cherry transgenes in calvaria, long bone in vivo and in primary BMSC and COB cultures in vitro:

The specificity of the expression of these transgenes was also examined in bones and primary cultures derived from Bone marrow stromal cells (BMSC) and Calvarial Osteoblasts (COB) (Figure 5).

Figure 5: Expression of DMP1-Cherry and DSPP-Cerulean transgenes in calvaria and long bone in vivo and primary cultures from BMSCs and COBs in vitro.

A. Representative images of sections through calvaria (a-c) and long bone (d-f) at P6. Images are from the same section visualized under epifluorescent microscopy with filter cubes to detect Cherry and Cerulean signals. DMP1-Cherry is expressed in the osteoblasts and osteocytes of calvaria and long bone (b, e). Note the lack of DSPP- Cerulean expression in these tissues (c, f). Scale bar: 200μm

B. Representative images of cultures at various time points from bone marrow stromal cells (BMSCs). Note formation of Calcein green-stained nodules at day 10. Also note increases in the Calcein green-stained areas at days 14 and 17. Expression of DMP1-Cherry in the mineralized nodules was detected at Day 10 which progressively increased thereafter. Note the lack of expression of DSPP-Cerulean transgene in these cultures. Scale bar: 200μm

C. Representative images of cultures at various time points from calvaria osteoblasts (COB) showing Calcein-green stained mineralized nodules at days 14 and 21. DMP1-Cherry expression is detected in the mineralized nodules at days 14 and 21. Note the lack of expression of DSPP-Cerulean in COB. Scale bar: 200μm

In calvaria and long bones isolated at P6, strong expression of DMP1-Cherry was detected in osteoblasts and osteocytes. DSPP-Cerulean was not detected in these tissues (Figure 5A).

In BMSC cultures, DMP1-Cherry⁺ cells were first detected around day 10 in discrete areas of CG-stained mineralization nodules (Figure 5B). The number of cells expressing the transgene increased with more advanced stages of differentiation at days 14 and 21.

In primary COB cultures, expression of DMP1-Cherry was first detected around day 14 within the areas of CG-stained nodules (Figure 5C). The number of cells expressing the transgene increased further at day 21. DSPP-Cerulean expression was not detected in BMSC and COB cultures (Figure 5B–C). These findings confirm that DSPP-Cerulean is exclusively expressed by odontoblasts.

3. DISCUSSION:

In our laboratory, we have characterized a series of transgenic mice, which express GFP under the control of tissue- and stage-specific promoters. However, the available animal models and different transgenes (pOBCol2.3-GFP, pOBCol3.6-GFP, and DMP1-GFP), like the endogenous proteins, are expressed by bone-forming cells (Kalajzic et al., 2004; Kalajzic et al., 2002), which in turn make it difficult to delineate the two cell types in various in vivo and in vitro studies. To overcome these difficulties and to be able to distinguish between dentin and bone forming cells we have generated DSPP-Cerulean/DMP1-Cherry transgenic mice.

Our in vivo and in vitro studies demonstrated that the DMP1-Cherry transgene was expressed by functional and fully differentiated odontoblasts (Figure 6), cementoblasts, osteoblasts and osteocytes. The patterns of expression of this transgene is the same as the reported expression of endogenous Dmp1 in various cell types (Qin et al., 2007) and DMP1-GFP (Balic & Mina, 2011). On the other hand, the DSPP-Cerulean transgene was expressed exclusively by odontoblasts identical to the endogenous expression of Dspp. These studies indicate that transgenes contain many of the cis-regulatory elements necessary for regulating faithful expression of DMP1 and DSPP in odontoblasts, cementum, and osteoblasts and osteocytes.

Figure 6: Schematic representation of proposed stages of activation of transgenes during odontoblast differentiation in vivo and in vitro.

A. The expression of 3.6-GFP, 2.3-GFP, and DMP1-GFP transgenes during odontoblast differentiation in vivo are based on results reported by (Balic et al., 2010; Balic & Mina, 2011). DMP1-Cherry expression during odontoblast differentiation in vivo is similar to previously observed DMP1-GFP expression in functional odontoblasts with a lower intensity of expression in fully differentiated odontoblasts. DSPP-Cerulean is first detected at low levels in late functional odontoblasts with increases in fully differentiated odontoblasts.

B. Schematic representation of proposed stages of activation during the in vitro mineralization of dental pulp cultures. In these cultures pre-odontoblasts and odontoblasts are derived from mesenchymal stem cells.

The stage-specific activation of these transgenes with the previously reported transgenic reporter animals (Balic et al., 2010; Balic & Mina, 2011), provide a panel of reporters that can help identify and isolate cells in odontogenic lineage at various stages of differentiation. This ability will bring a wealth of knowledge to help understand novel gene regulatory network involved in specification and progression of cells into odontoblasts lineage. In addition, these studies can lead to deciphering the underlying causes of dental congenital abnormalities and meaningful targeted therapies for affected patients

4. MATERIALS AND METHODS:

Generation of DSPP-Cerulean/DMP1-Cherry transgenic mice and analysis of the expression of transgenes in vivo:

DSPP-Cerulean/DMP1-Cherry transgenic mice were generated using a bacterial recombination strategy with the mouse BAC clone RP24-258g7 as described previously (Gong, Yang, Li, & Heintz, 2002; Maye et al., 2009) and outlined in Supplemental Figures 4 and 5. All mice were maintained in CD1 background.

Mandibular arches, long bones and calvaria at different stages were isolated from DSPP-Cerulean/DMP1-Cherry transgenic mice. Tissues were fixed overnight in 4% paraformaldehyde, decalcified and processed for cryosectioning (Dyment et al., 2016).

Sections were examined and imaged using Zeiss Axio Observer Z1 inverted microscope using filter cubes optimized for the detection of Cherry Red Fluorescent Protein (HQ577/20 ex, HQ640/40 em, Q595lp beam splitter), Cerulean (ECFP, enhanced Cyan Fluorescent Protein) (D436/20 ex, D480/40 em, Q455dclp beam splitter), Green Fluorescent Protein (GFP) (HQ525/50ex, HQ 470/40 em, Q495lp beam splitter), and DAPI variants (AT350/50ex, ET460/50 em, T400lp beam splitter). The full-size images were obtained by scanning at high power followed by stitching the scans into a composite. Exposure times were adjusted for optimum imaging, and kept consistent throughout the various time points.

Cell cultures, Digital Imaging and Epifluorescence Analysis:

Primary cultures from dental pulp, bone marrow stromal cells (BMSC) and calvarial osteoblasts (COB) were prepared from 5- to 7-day old hemizygous DSPP-Cerulean/DMP1-Cherry and non-transgenic mice (Sagomonyants, Kalajzic, Maye, & Mina, 2015; Sagomonyants & Mina, 2014).

Live cultures were first imaged for detection of various Fluorescent Proteins, and then processed for staining with anti-GFP Alexa Fluor 488 conjugated antibody (1:1000 dilution, Molecular Probes, Invitrogen) (Sagomonyants and Mina 2014). The nuclei were stained with 1.0 μg/ml Hoechst 33342 dye (Invitrogen). Mineralization in these cultures was examined by Calcein Green staining (3 μg/ml in 2% NaHCO_3, pH = 7.4) (Wang, Liu, Maye, & Rowe, 2006). The mean fluorescence intensity of Calcein Green (CG) and DMP1-Cherry were measured using a multi-detection monochromator microplate reader (Safire², Tecan, Research Triangle Park, N.C., USA) (Sagomonyants et al., 2015; Sagomonyants & Mina, 2014). Calcein green staining was measured at 475/515-nm wavelength (excitation/emission) and at a gain of 40. Background fluorescence was measured with cultures grown in the absence of mineralization inducing reagents and these values were subtracted from respective CG measurements.

DMP1-Cherry intensity was measured at 580/610-nm wavelength (excitation/emission) and at a gain of 65. Background fluorescence for Cherry was measured with cultures from the non-transgenic littermates that lack fluorescent reporter expression, and these values were subtracted from respective Cherry measurements.

Quantification of DSPP-Cerulean⁺ cells was performed by calculating the ratio of cells stained with anti-GFP antibody (DSPP-Cerulean⁺ cells) to the total number of Hoechst⁺ cells (Sagomonyants et al., 2015; Sagomonyants & Mina, 2014).

Flow cytometric analysis and sorting (FACS):

Cultures derived from DSPP-Cerulean/DMP1-Cherry transgenic and non-transgenic animals were processed for flow cytometric analysis at various days as described before (Sagomonyants and Mina, 2014; Sagomonyants et al., 2015). Cultures were also processed for FACS sorting based on Cherry expression. Upon separation, reanalysis confirmed that the purity of isolated DMP1-Cherry⁺ and DMP1-Cherry⁻ populations were >98%. Live Cherry⁺ and Cherry⁻ cells were collected and re-plated at the same density as the primary cultures (8.75 × 104 cells/cm2). Cultures were grown for an additional 14 days and processed for FACS analyses, as described for unsorted cultures. The gating strategy for FACS analysis of DMP1-Cherry⁺ and DSPP-Cerulean⁺ cells as shown in Supplemental Figure 4 was used.

RNA extraction and analyses:

Total RNA was isolated with TRIzol reagent (Invitrogen), and processed for gene expression analysis by real-time polymerase chain reaction analysis with primers specific for GAPDH, DSPP, DMP1 and different variants of GFP (Supplemental Figure 3A).

Time-lapse live imaging:

Primary pulp cultures were analyzed by time-lapse confocal imaging to visualize the conversion of DMP1-Cherry⁺ cells to DSPP-Cerulean⁺ cells. Six to eight different positions of the well already expressing Cherry⁺ cells were selected, followed by imaging over 36 h using Zeiss 780 LSM equipped with a humidified imaging chamber maintained at 37C and 6% CO2. 20x images were taken using 561nm laser for Cherry and 440nm laser for Cerulean every 30 mins over 36 hours. The images were converted to .mov video files and processed using ImageJ and Metamorph.

Statistical analysis of data:

was performed by GraphPad Prism 7 software using unpaired two-tailed Student t-test. Values in all experiments represented mean ± SEM of at least three independent experiments, and a *p-value ≤ 0.05 was considered statistically significant.

Supplementary Material

Supp Fig S6

Supplemental Figure 6. Generation of DSPP-Cerulean/DMP1-Cherry transgenic mice and analysis of the expression of transgenes in vivo: BAC Clone RP24-258G7 was obtained from Children’s Hospital Research Institute (CHORI). pLD53-SC2 and PSV1.RecA recombination vectors were generously provided by Shiaoching Gong (Gong et al. 2002). pUni-V5 HisB vector and Pfx DNA polymerase were obtained from Invitrogen. Homology arms were amplified using an C1000 Thermal Cycler (Biorad).

Homology arms for Dspp and Dmp1 were cloned into pLD53.SC2-Cerulean and pUni-V5-mCherry vectors, respectively. For the Dspp homology arm, a 271bp homology arm was amplified and cloned into the Not1/BamH1 site of pLD53.SC2-Cerulean using primers: 5’-tctcgcggccgcagagccacgagctatgtactct-3’ (sense), 5’-tctcggatcctgagtagtaatcaatgtctgt-3’ (antisense). For the Dmp1 homology arm, a 615bp homology was amplified and cloned into the Not1/Bamh1 site of pUni-V5-mCherry using primers: 5’-tctcgcggccgcgagttggtggagagatac-3’ (sense), 5’-tctcggatccggatgcgattcctctacctgtaatgaaag-3’ (antisense). Amplified products were run through a PCR clean up column (Qiagen), restriction digested and gel purified (Zymogen). Plasmids were digested and then briefly treated with Calf Intestinal Alkaline Phosphatase (New England Biolabs) followed by phenol chloroform extraction and precipitation. Vector and inserted were mixed, ligated at room temperature for 1 hour and electroporated into PIR2 cells. Transformants were selected on LB plates with ampicillin for pLD53.SC2-Cerulean and kanamycin for pUni-V5-mCherry.

Recombinase A was introduced into RP24-258G7 host bacteria by transformation with pSV1.RecA vector (100 ng) and selected for on chloramphenicol (12.5 ug/ml)/tetracycline (10ug/ml) LB agar plates. RP24-258G7 host bacteria containing RecA were then transformed by electroporation with 1ug (1-2ul) of the pUni-V5--mCherry containing the Dmp1 homology arm. SOC medium (1 ml) was added and transformed bacteria were incubated with shaking at 200 rpm for one hour at 30°C. Recombinants were first selected for by adding 5 ml of LB medium containing chloramphenicol (12.5 lg/ml), kanamycin (50 ug/ml), and tetracycline (10 ug/ml) to bacteria and grown overnight at 30°C with shaking at 200 rpm. Further selection for recombinant clones was carried out by plating 100 ul of overnight culture on to chloramphenicol (12.5 ug/ml), kanamycin (50 ug/ml) LB agar plates and incubated overnight at 42°C. Chloramphenicol and kanamycin resistant colonies were screened for recombinants by colony PCR. Following the identity of RP24-258G7-DMP1-mCherry positive clones, pLD53.SC2-Cerulean containing the Dspp homology arm was recombined into RP24-258G7-DMP1-mCherry using the same approach with the exception that ampicillin (50ug/ml) and kanamycin (50ug/ml) were used to select for the integration of pLD53.SC2 and pUni-V5 reporters, respectively. The final assembled dual BAC reporter clones were verified by diagnostic restriction digestion and run on a field inversion gel. BAC DNA was prepared for pronuclear injection as described previously (Maye et al., 2009). Transgenic mice were created at the Center for Mouse Genome Innovation (UConn Health). All mice were maintained in CD1 background.

Supp FigS1

Supplemental Figure 1: Expression of DMP1-Cherry in dental pulp and flow cytometric analysis of dental pulp cultures

A and B. Sagittal section through the first molar from P6. Showing expression of DMP1-Cherry in dental pulp (indicated by arrow heads). Scale bar: 2000μm.

C. Legend denoting identity of each gate analyzed by flow cytometry.

D. Gating strategy used for FACS analysis. (Top row) Gate P1 includes all cells exclusive of debris. P2 includes the population of single cells, while P3 includes live cells. Dead cells were excluded by staining with Fixable Viability Dye. Further analysis of Cherry and Cerulean transgenes was performed in the P3 population of live cells. Pulp cells from non-transgenic littermates served as a negative control for Cerulean and Cherry transgene expression in all experiments (shown in the Middle row). Positive controls for Cerulean and Cherry transgenes included pulp cells from 5-7-day-old Col3.6-Cerulean (P4), Osx-Cherry (P5), and Osx-Cherry/Col3.6-Cerulean (P6). Bottom row shows FACS analysis performed in DMP1-Cherry/DSPP-Cerulean cultures at Days 0, 7, 10 and 14 of culture. Note that fresh dental pulp isolates at Day 0 of culture show DMP1-Cherry⁺ cells. Cultures analyzed from D7 through 14 show progressively increasing percentages of DMP1-Cherry⁺ and DMP1-Cherry/DSPP-Cerulean+ populations. Note the lack of DSPP-Cerulean+ cells is not co-expressing DMP1-Cherry. % of DMP1-Cherry+ cells is noted in red while DSPP-Cerulean/DMP1-Cherry+ cells are indicated in blue. Values represent mean ± S.E from at least three independent experiments.

Supp FigS2

Supplemental Figure 2: Comparison of transgene expression in molar and incisor of Lines 1, 2 and 3 in vivo.

Representative images of sagittal sections through developing lower jaw at P6 captured using Brightfield and epifluorescent microscopy with appropriate filter cubes to detect Cherry and ECFP in Line 1 (A-E), Line 2 (F-J) and Line 3 (K-O). All 3 lines show similar in vivo expression of DMP1-Cherry (B,G,L) in alveolar bone as well as odontoblasts, and DSPP-ECFP (C,H,M) expression in odontoblasts. In all images, the pulp chambers are denoted by dashed lines. Ab, alveolar bone; I, incisor; M1, first molar. Scale bar: 50μm.

Supp FigS3

Supplemental Figure 3: Expression of Dmp1, Dspp, and transgenes during the in vitro mineralization of primary pulp cultures.

A. Primer sequences and conditions used for RT-PCR analysis

B. RT-PCR analysis of RNA extracted from primary pulp cultures isolated at various time points using primers shown in Table B. Note the similarities in the temporal expression of endogenous Dmp1 with Cherry and endogenous Dspp with GFP/Cyan and transgenes that are first detected at day 7 and 10, respectively with increases thereafter. Gapdh was used as an internal control.

Supp FigS5

Supplemental Figure 5. DNA Construct Map for Dspp-Cerulean DMP1-mCherry Reporter Mice. BAC clone RP24-258g7 with a size of 164368bp was used to generate Dspp-Cerulean DMP1-mcherry reporter mice. To generate Dspp-Cerulean, a cerulean fluorescent protein reporter gene was inserted just upstream of the Dspp translation start site by homologous recombination into the BAC clone using a 271bp homology arm. The Dspp translation start site (underlined sequence) was 87,894bp downstream from the 5’ end and 76,474bp upstream of the 3’ end of the BAC clone. To generate Dmp1-mCherry, a mCherry fluorescent protein reporter gene was inserted just upstream of the Dmp1 translational start site (underlined sequence) also by homologous recombination into the BAC clone using a 615bp homology arm. The Dmp1 translation start site was 120,964bp downstream from the 5’ end and 43,404bp upstream of the 3’ end of the BAC clone.