Derivation, Characterization, and In Vitro Differentiation of Canine Embryonic Stem Cells



Canine embryonic stem (cES) cell lines were generated to establish a large-animal preclinical model for testing the safety and efficacy of embryonic stem (ES) cell-derived tissue replacement therapy. Putative cES cell lines were initiated from canine blastocysts harvested from natural matings. Times of harvest were estimated as 12–16 days after the presumed surge in circulating levels of luteinizing hormone. Four lines established from blastocysts harvested at days 13–14 postsurge satisfied most of the criteria for embryonic stem cells, whereas lines established after day 14 did not. One line, Fred Hutchinson dog (FHDO)-7, has been maintained through 34 passages and is presented here. FHDO-7 cells are alkaline phosphatase-positive and express both message and protein for the Oct4 transcription factor. They also express message for Nanog and telomerase but do not express message for Cdx2, which is associated with trophectoderm. Furthermore, they express a cluster of pluripotency-associated microRNAs (miRs) (miR-302b, miR-302c, and miR-367) characteristic of human and mouse ES cells. The FHDO-7 cells grow on feeder layers of modified mouse embryonic fibroblasts as flat colonies that resemble ES cells from mink, a close phylogenetic relative of dog. When cultured in nonadherent plates without feeders, the cells form embryoid bodies (EBs). Under various culture conditions, the EBs give rise to ectoderm-derived neuronal cells expressing γ-enolase and β3-tubulin; mesoderm-derived cells producing collagen IIA1, cartilage, and bone; and endoderm-derived cells expressing α-fetoprotein or Clara cell-specific protein.

Disclosure of potential conflicts of interest is found at the end of this article.


Embryonic stem (ES) cells, derived from the inner cell mass of a mouse blastocyst, were first described more than 25 years ago [1, 2]. Since then, they have been used in vitro to study the molecular control of lineage commitment and differentiation and to generate transgenic mice for in vivo analysis of altered gene expression [3, 4]. Theoretically, the unlimited proliferative and developmental potential of ES cells offer considerable opportunity for applications in regenerative medicine and tissue engineering. However, there are several problems that must be addressed before ES cell-derived tissues can be used for treating human disease.

One major barrier is the inability to control, with high efficiency, the differentiation of ES cells into transplantable precursors of a specific lineage. High efficiency is required to reduce the possibility of ES cell-derived tumor development, whereas transplantable precursors are necessary to establish and then maintain the transplanted tissue. A second barrier is recipient-mediated rejection of the ES cell-derived tissue. Since transplanted ES cell-derived tissue will be, except in the case of syngeneic mice, an allograft, methods for establishing tolerance must be developed to prevent rejection. Finally, issues of long-term genetic or epigenetic stability of ES-derived tissue must be addressed. Ideally, all these issues should be addressed in vivo in a large-animal preclinical model in which proliferative demand on transplanted cells and the “at risk” lifespan of the transplanted tissue approximate those of humans.

There is a considerable body of evidence that has already established the dog as a reliable preclinical model for adult, marrow-derived stem cell transplantation [5]. Dogs are large, relatively long-lived, outbred animals that have many diseases similar to humans [6]. Protocols developed in the dog to establish tolerance and overcome allograft rejection have been successfully applied to humans [7, 8]. Moreover, the long-term outcomes of organ or hematopoietic transplantation in dogs have accurately predicted the outcomes in patients [9]. Given the value of the dog model for extrapolation to humans, we have pursued the derivation of canine ES cells. The long-term goal is to provide a preclinical model for developing ES cell-derived tissue replacement therapies.

Materials and Methods

All animal studies were approved by the Fred Hutchinson Cancer Research Center Institutional Animal Care and Use Committee.

Ovulation Timing

Serum progesterone levels were measured by an independent veterinary laboratory from blood serum samples collected daily from females in proestrus. A preovulatory luteinizing hormone (LH) surge with a duration of approximately 12 hours occurs in the canine at the time progesterone levels rise above 2 ng/ml. This time point was designated day 0. Ovulation was assumed to occur at day 2. Males were placed with females at the first signs of proestrus and mated naturally.

Cell Line Establishment

Blastocysts were flushed from the uterus following routine ovariohysterectomy at days 12–16 post-LH surge. The zona pellucida was mechanically removed, and blastocysts were placed on primary mouse embryonic fibroblasts (MEFs) that had been exposed to 40 Gy of γ-radiation. Growth medium consisted of Dulbecco's modified Eagle's medium/Ham's F-12 medium from HyClone (Logan, UT,; 15% fetal bovine serum (FBS) (HyClone); 0.1 mM nonessential amino acids (Invitrogen, Carlsbad, CA,; 1 mM l-glutamine (Invitrogen), 1 mM sodium pyruvate (Invitrogen); 1 U/ml, 1 μg/ml penicillin/streptomycin (Invitrogen). Human leukemia inhibitory factor (hLIF) (Millipore, Billerica, MA, was added to initial culture conditions at 20 ng/ml to supplement the fibroblast production of murine leukemia inhibitory factor (LIF). After three passages on primary MEFs, canine cells were transferred to SNL 76/7 feeders, a derivative of the STO (Sandoz inbred Swiss mouse, thioguanine-resistant, ouabain-resistant) cell line engineered to express LIF. The addition of exogenous hLIF was then discontinued. Subculture was accomplished by mechanically cutting out the central area of colonies during the initial culturing. When cells seemed established, they were passaged enzymatically with TrypLE select (Invitrogen). Subsequent studies suggested that single cell passage using TrypLE may contribute to aneuploidy; therefore, mechanical passage was reinstated.

EB Formation and Differentiation Conditions

EB formation was induced by transferring cells to Costar Ultra Low Attachment plates from Corning (Corning, NY, The cells remained suspended for at least 5 days, at which time they were plated on gelatin-coated tissue culture dishes at various densities and serum concentrations, with and without exogenous growth factors to induce differentiation. Induced differentiation conditions included growth to confluence and plating at a high cell density, which resulted in cells expressing collagen IIA1 that stained positive for bone and cartilage. Coculture during both the nonadherent and adherent periods with fetal canine lung tissue, separated from the Fred Hutchinson dog (FHDO)-7 cells using a Millicell-PCF (Millipore) transwell, led to expression of Clara-cell specific protein (CCSP), indicative of lung epithelial precursors. Culture in 10% knockout (KO) serum resulted in β3-tubulin and γ-enolase-positive cells. Culture in 0.5% serum (Invitrogen)/100 ng/ml Activin A (Peprotech, Rocky Hill, NJ, ng/ml bone morphogenetic protein 4 (BMP4) resulted in α-fetoprotein (AFP)-positive cells. Ten percent KO serum/100 ng/ml Activin A also led to CCSP-positive cells.

RNA Isolation and cDNA Synthesis

Cell cultures were washed once with phosphate-buffered saline (PBS) and lysed directly on the plate. Total RNA was extracted from cultured cells with the RNeasy Mini kit (Qiagen, Hilden, Germany, Tissues were collected from discarded biological material resulting from other protocols. All tissues were directly flash-frozen in liquid nitrogen and then homogenized in Trizol (Invitrogen) with a Tekmar Tissuemizer (Tekmar Instruments, Cincinnati, Manufacturer protocols were followed in both cases.

Complementary DNA was synthesized from total RNA using a modified Superscript II protocol (Invitrogen). Briefly, RNA was diluted in 7 μl of nuclease-free H2O. Four microliters of 0.5 μg/μl oligo(dT) 16–18-mer was added, and the solution was heated to 70°C for 10 minutes and then cooled on ice for 5 minutes. Four microliters of 5× first strand buffer, 2 μl of 0.1 M dithiothreitol, and 1 μl of RNaseOUT were combined and added to the initial 11 μl. Finally, 1 μl Superscript II reverse transcriptase or H2O for RT+ and RT−, respectively, was added. The whole 20-μl reaction was incubated at 42°C for 2 hours and then inactivated at 70°C for 20 minutes. The resulting cDNA was stored at −20°C.

Polymerase Chain Reaction

Platinum Taq polymerase (Invitrogen) was used for thermal cycling. Antibody was inactivated at 94°C for 5 minutes before cycling. DNA was amplified for 35–40 cycles by a melting step of 94°C for 10 seconds, annealing at 68°C for 30 seconds, and extension at 72°C for 1 minute. Oct4 and Nanog sequences were amplified with Advantage Taq by 36 cycles of melting at 94°C for 20 seconds and amplification at 68°C for 2 minutes. Primers are listed in Table 1. Products were separated and visualized on agarose gels. Canine β-actin reverse transcription (RT)-polymerase chain reaction (PCR) was performed at three-cycle intervals from cycle 25 to cycle 37 and found to be equivalent in all samples.

Table Table 1.. Canine-specific primer sequences
original image

MicroRNA Analysis: RNA Isolation and Quantitative Reverse Transcription-PCR

Total RNA isolation for microRNA quantitation was performed using the mirVana microRNA (miRNA) isolation kit (Ambion, Austin, TX, according to the manufacturer's protocol. H1 human ESC (hESC), FHDO-7, and MEF cell cultures were rinsed with ice-cold PBS prior to initial lysis using 600 μl of mirVana Lysis/Binding buffer (Ambion). H1 hESC lysates were kindly provided by Carol Ware and Angel Nelson (University of Washington, Seattle). Canine adult tissue samples were homogenized in mirVana Lysis/Binding buffer using a Qiagen TissueLyser device before continuing with the RNA isolation protocol. Human adult tissue RNAs were miRNA-certified FirstChoice Total RNAs purchased from Ambion. Total RNA quality was verified using an Agilent Bioanalyzer (Agilent Technologies, Foster City, CA,

Quantitation of miRNAs was carried out using TaqMan microRNA assays (Applied Biosystems, Foster City, CA, that are based on quantitative reverse transcription (qRT)-PCR. Briefly, 3.34 ng of template total RNA was reverse-transcribed in a 5-μl reaction using the TaqMan MicroRNA Reverse Transcription Kit and miRNA-specific stem-loop primers for hsa-miR-302b (catalog no. 4378071), hsa-miR-302c (catalog no. 4378072), hsa-miR-367 (catalog no. 4373034), and hsa-miR-16 (catalog no. 4373121) (Applied Biosystems). The reverse transcription reaction was diluted by adding 32.3 μl of H2O, following which 2.25 μl of the diluted material was introduced into a 5-μl PCR using TaqMan Universal PCR master mix without AmpErase UNG. Real-time PCR was carried out on an Applied Biosystems 7900HT Thermocycler at 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute.

Western Blotting

Cells were washed twice with PBS, and cell pellets were solubilized in RIPA buffer (10 mM Tris-HCl [pH 7.6], 150 mM NaCl, 1% [wt/vol] sodium dodecyl sulfate [SDS]) containing freshly added proteinase inhibitors (Complete Mini Protease inhibitor cocktail tablet, Roche Diagnostics, Mannheim, Germany). Lysates were centrifuged at 16,000g for 15 minutes at 4°C, and protein concentrations were determined by The Bradford Method with Biorad Protein Assay (Bio-Rad, Hercules, CA, Total cell lysates were mixed with Laemmli sample buffer (Bio-Rad), heated to 96°C for 5 minutes, run on a 4%–20% Tris-glycine gel (Invitrogen), and transferred to nitrocellulose (Invitrogen). Membranes were blocked with 5% milk in Tris-buffered saline (TBS) for 1 hour at room temperature and then incubated with primary antibody (2 μg/ml anti-Oct4 (R&D Systems Inc., Minneapolis, or 1 μg/ml anti-CCCTC-binding factor [Millipore]) overnight at 4°C. Blots were washed five times for 5 minutes each with TBS containing 0.005% Tween 20 (TBS-T). Blots were incubated with the appropriate secondary antibody (0.05 μg/ml rabbit anti-goat IgG/horseradish peroxidase [HRP] [Dako, Glostrup, Denmark,] or goat anti-rabbit-IgG [1:1,000] [Bio-Rad]) for 1 hour at room temperature. Blots were washed again with TBS-T, and antibody-reactive proteins were detected using enhanced chemiluminescence plus (ECL+) and hyperfilm-ECL (GE Healthcare, Piscataway, NJ). Antibodies were removed by incubating blots with stripping buffer (62.5 mM Tris-HCl, pH 6.7, 100 mM β-mercaptoethanol, 2% SDS) at 50°C for 30 minutes. Stripped immunoblots were then washed, blocked, and probed as described.

Alkaline Phosphatase Staining

Cell cultures were fixed with 10% formalin for 2 minutes and then incubated for 15 minutes with Fast Red Violet, Naphthol AS-BI solution, and water in a 2:1:1 ratio from an alkaline phosphatase detection kit (Millipore).

Alizarin Red Staining and Alcian Blue Staining

Cell cultures were fixed with 100% ethanol for 1 hour, rinsed with water, stained with 0.05% (wt/vol) alizarin red (Sigma-Aldrich, St. Louis, or 0.15% (wt/vol) Alcian Blue (Sigma-Aldrich) in 95% ethanol/5% acetic acid for 1 hour, and then rinsed with H2O.


Cell cultures were fixed in 10% neutral buffered formalin for 20 minutes and then washed twice for 5 minutes in PBS. Endogenous peroxidase activity was quenched with 3% H2O2 for 10 minutes. After washing with PBS-0.05% Tween (Sigma-Aldrich), cultures were blocked for 10 minutes with 10% serum of the same species as the secondary antibody, 1% bovine serum albumin (Sigma-Aldrich), 0.05% Tween, 0.1% Triton X-100 (Sigma-Aldrich). Primary antibodies were used at the manufacturers' recommended dilutions when available and diluted in blocking solution. Antibodies tested include anti-Oct3/4 (AF1759; R&D Systems), anti-stage-specific embryonic antigen (SSEA)-1, anti-SSEA-3 (MAB1434; R&D Systems), anti-SSEA-4 (MAB1435; R&D Systems), anti-TRA1–81 (MAB4381; Chemicon, Temecula, CA,, anti-TRA1–60 (MAB4360; Chemicon), anti-β3-tubulin (clone TUJ1; BabCo, Richmond, CA,, and anti-CCSP (07-623; Millipore). Cells were incubated for 30 minutes at room temperature with the primary antibodies and then washed twice with PBS-Tween. Primary antibodies used for control labeling were isotype-matched with the specific primary antibody. Secondary HRP-conjugated antibody (rabbit anti-goat [P0449; Dako]), goat anti-rabbit polymer (Dako K4010), or goat anti-mouse polymer (K4000; Dako) was applied for 25 minutes to both specific and control antibody-labeled cells, followed by two washes in PBS-Tween. Diaminobenzidine substrate (BioFX, Owings Mills, MD, was developed for 2–5 minutes. Cultures were counterstained with hematoxylin (S3301; Dako), followed by a rinse in 0.037 M ammonium hydroxide (Sigma-Aldrich).

Microscopy, Photography, and Image Processing

Photomicrographs were taken on two microscopes, a Nikon Eclipse TS100 with a Nikon CoolPix E995 camera (Fig. 3A), and a Nikon Diaphot SMZ1500 microscope equipped with a Nikon DS Fi1 5-megapixel charge-coupled device camera (Figs. 1, 2, 3, Figure 4., Figure 5.6) (Nikon, Tokyo, PCR amplification products were visualized by ethidium bromide staining of the agarose gels on a Bio-Rad GelDocXR imaging system with QuantityOne software. TIFF images were combined in Photoshop 7.0 (Adobe Systems Inc., Seattle,

Figure Figure 1..

Canine embryos at different developmental stages. (A): Day 12, ∼200-μm diameter. (B): Day 14 blastocyst, ∼500 μm-diameter. (C): Day 15 blastocyst, ∼900-μm diameter.

Figure Figure 2..

Undifferentiated canine embryonic stem (ES) cells. (A): Phase-contrast image of FHDO-7 colony on SNL 76/7 feeder cells; magnification, ×100. (B): Undifferentiated FHDO-7 stained for alkaline phosphatase; magnification, ×100. (C): Polymerase chain reaction (PCR) showing Oct4 and Nanog mRNA expression in FHDO-7 and testes. (D): Western blot showing Oct4 protein in undifferentiated FHDO-7 and mouse ES cell line AK-7. The blot was stripped and reprobed for the ubiquitous factor CTCF (lower panel), demonstrating equivalent protein loading. (E): PCR showing canine telomerase mRNA expression in FHDO-7 and fetal lung fibroblasts. β-Actin signals in all RT+ reactions were comparable (data not shown). (F): Goat anti-Oct4 IgG labeling revealed by horseradish peroxidase (brown stain) showing nuclear localization of Oct4 in cES cells growing on top of MEFs; magnification, ×100. (G): Goat IgG isotype control staining; magnification, ×100. Abbreviations: CTCF, CCCTC-binding factor; Diff, differentiated; FhDo, Fred Hutchinson dog; RT, reverse transcription.

Figure Figure 3..

Expression of pluripotency-associated miRs in FHDO-7 cells. The ESC-associated miRs miR-302b, miR-302c, and miR-367 were readily detected by quantitative reverse transcription-polymerase chain reaction in H1 hESC and in FHDO-7 cells but not in differentiated adult tissues derived from either species. The ubiquitously expressed miR-16 was easily detectable in all samples. MEFs were used routinely in the culture of FHDO-7 cells. Please note that greater values of cycling threshold (CT) correspond to lower expression level. Abbreviations: FHDO, Fred Hutchinson dog; hESC, human embryonic stem cells; MEF, mouse embryonic fibroblast; miR, microRNA.

Figure Figure 4..

Induced differentiation of canine embryonic stem (ES) cells after passage 20. (A): Embryoid body formation after 4 days of nonadherent culture; magnification, ×100. (B): Fred Hutchinson dog (FHDO)-7 cells plated at high density and grown to confluence. Shown are alizarin red-positive canals with structure similar to Haversian networks (left panel magnification, ×100) and Alcian Blue-positive fibers indicative of cartilage formation (right panel magnification, ×100). (C): Mouse anti-β3-tubulin IgG labeling revealed by horseradish peroxidase (HRP) showing a FHDO-7 cell with neuronal morphology representing spontaneous differentiation of a FHDO-7 cell grown on feeders; magnification, ×400. (D): Mouse IgG isotype control labeling; original magnification, ×400. (E): β3-Tubulin-positive cells detected after culturing FHDO-7 EBs on gelatin-coated plates for 4 weeks. (F): Mouse IgG isotype control for the β3-tubulin antibody. (G): Rabbit anti-Clara-cell specific protein IgG labeling revealed by HRP showing FHDO-7 cells cultured in the presence of canine fetal lung tissue separated from the canine ES cells by a transwell membrane (magnification, ×400). (H): Rabbit IgG isotype control; magnification, ×400.

Figure Figure 5..

Semiquantitative RT-polymerase chain reactions in Undiff. and Diff. FHDO-7. (A): β3-Tubulin with fetal brain as a positive control. (B):CCSP with fetal lung as a positive control. (C):AFP with fetal liver as a positive control. SNL 76/7 feeder cells did not express the transcripts, and Diff. FHDO-7 upregulated all five transcripts. (D):CollagenIIA1 with cartilage as a positive control. (E): γ-Enolase with fetal brain as a positive control. β-Actin levels were comparable in all canine tissues. Reactions were done with and without the addition of RT to the cDNA reactions (RT+ and RT−, respectively). Abbreviations: AFP, α-fetoprotein; CCSP, Clara-cell specific protein; Diff., differentiated; FHDO, Fred Hutchinson dog; RT, reverse transcription; Undiff., undifferentiated.

Figure Figure 6..

Fred Hutchinson dog-7 cells at passage 15 showing a normal canine 78 XY karyotype.

Karyotype Analysis

Cells were analyzed for the normal canine karyotype in the Clinical Pathology Laboratory of the Fred Hutchinson Cancer Research Center. Briefly, canine embryonic stem (cES) cells were treated with colchicine to arrest cells in metaphase, fixed in methanol:acetic acid, and dropped onto clean glass slides.


Embryo Collections

A total of 67 embryos were collected from 10 pregnant dogs at 12–16 days following the estimated LH surge, predicted by progesterone levels exceeding 2 ng/ml. All collections preceded implantation, which occurs around day 20. Day 12 embryos (n = 7) represented variable immature stages (Fig. 1A) and did not form outgrowths of cells when cultured on MEFs. Day 13–14 embryos (n = 26; 4 used for RNA) had a mature blastocyst morphology (Fig. 1B), and their cellular outgrowths were robust. Embryos harvested at day 15 or later (n = 34) had larger diameters and a very thin zona layer. Although outgrowths were robust, cell lines established from day 15 or older blastocysts (Fig. 1C) did not express Oct4 and were not pursued. Once the optimal time for harvest was determined, the efficiency for generating lines was 4 of 22 attempts.

ES Cell Derivation

The zona pellucida was mechanically removed from the blastocysts, and whole embryos were cultured on primary MEFs. After 6–12 days of culture, the initial outgrowths were cut into clumps of approximately 100–300 cells and passaged onto fresh MEFs. After three passages, cells were transferred onto the STO derivative SNL 76/7 cell line [10] and maintained on this immortalized mouse cell line for all subsequent passages. Passaged clumps of cells gave rise to flat colonies that differed in appearance from mouse and human ES cell colonies but were morphologically similar to that reported for mink ES cells [11, [12], [13]–14]. A phase-contrast image of a canine ES cell colony is shown in Figure 2A. The original cell lines were designated FHDO-1–FHDO-11 in order of their derivation; of these, FHDO-6, -7, -8, and -9 expressed markers consistent with ES cells.

Characterization of FHDO-7 (Passage Number >20)

As reported for all species of ES cells, FHDO-7 cells are positive for alkaline phosphatase (Fig. 2B). They also express mRNA for the nuclear transcription factor Oct4 (Fig. 2C), as well as Oct4 protein detected by Western blot (Fig. 2D) that is localized to the nucleus, as shown by immunohistochemistry (Fig. 2E, 2F). They do not express message for CDX2 (data not shown) but do express mRNA for telomerase and for the transcription factor Nanog. Unfortunately, commercially available antibodies specific for Nanog do not appear to cross-react with dog. We also failed to detect staining with antibodies specific for the SSEA tumor rejection antigens (TRA1–60, TRA1–81). In the case of antibodies specific for SSEA-1, also known as CD15, which is expressed on human granulocytes as well as human ES cells, there was no specific labeling of dog granulocytes or ES cells (data not shown). This differed from the two previous reports on dog ES cells [15, 16]; however, in the first study, no specificity controls for staining were included, and in the second, the data were not shown and the methods were not discussed.

In addition to the markers above, we examined FHDO-7 cells for expression of microRNAs that have been found to be characteristic of embryonic stem cells. MicroRNAs are small (typically approximately 22 nucleotides) non-protein-encoding RNA molecules that are believed to regulate gene expression at the post-transcriptional level by interacting physically with target messenger RNAs (mRNAs) [17]. Expression of microRNAs is frequently restricted in a developmental stage-specific and cell type-specific manner [18]. ESC-specific microRNAs have been identified in human and mouse ESC [19, [20]–21]. In humans, this corresponds to a group of five microRNAs (microRNA [miR]-302a, miR-302b, miR-302c, miR-302d, and miR-367) that are all tightly clustered in the genome and are cotranscribed as a polycistronic transcript that is apparently processed to generate the individual microRNAs [19]. Commercially available qRT-PCR assays can be used to measure the expression of these and other miRNAs [22]. On the basis of an examination of an alignment of dog and human genome sequences, we noted that three of the microRNAs (miR-302b, miR-302c, and miR-367) in the human cluster are predicted to be perfectly identical in their mature sequence between human and dog (data not shown).

We used TaqMan-based qRT-PCR assays to measure the expression of these three microRNAs in FHDO-7 cells, as well as in a set of differentiated adult canine tissues that would not be expected to express this pluripotency-associated microRNA cluster. As positive and negative controls, we examined RNA isolated from the H1 human ESC line (grown under feeder-free conditions) and a set of adult human tissues. We also measured the expression in all samples of miR-16, a ubiquitously expressed microRNA that is predicted to be perfectly identical in mature sequence between dog and human on the basis of a genomic sequence alignment (data not shown).

As expected, we found that the H1 human ESC line abundantly expresses all three pluripotency-associated RNAs, as well as miR-16, as manifested by low cycling threshold (CT) values for each of these miRNAs (Fig. 3). Human adult liver, pancreas, and spleen do not express miR-302b, miR-302c, or miR-367 at detectable levels (defined as a CT value of 35 or less) but do express miR-16 as expected. FHDO-7 cells show a pattern of microRNA expression that parallels that of H1 human hESC, in that miR-302b, miR-302c, and miR-367 are all expressed, along with miR-16 (Fig. 3). In keeping with the human tissue results, canine tissues corresponding to adult liver, pancreas, and spleen all do not express detectable levels of the pluripotency-associated microRNAs but do express miR-16. Given that FHDO-7 cells are grown on feeder cultures of MEFs, we also considered the possibility that contaminating MEFs might be contributing to some signal in qRT-PCRs. When we performed the same microRNA qRT-PCR assays on RNA isolated exclusively from MEFs, none of the pluripotency-associated miRNAs were detected, whereas miR-16 was detected (Fig. 3). However, miR-16 was detected in the FHDO-7 sample at a slightly lower CT value than in the MEFs, indicating that FHDO-7 samples express at least as much miR-16 as any contaminating MEFs (Fig. 3). Taken together, the data suggest that FHDO-7 cells express a cluster of microRNAs characteristic of embryonic stem cells, further supporting the contention that FHDO-7 represents an embryonic stem cell line.

In Vitro Differentiation of FHDO-7 Cells after Passage 20

The FHDO-7 cells are capable of differentiating, in vitro, into cells representing ectoderm, endoderm, and mesoderm. To induce differentiation, the FHDO-7 cells were passaged onto nonadherent six-well plates and cultured for at least 5 days to allow EB formation (Fig. 4A). The EBs were then replated on gelatin-coated plates under different conditions in an attempt to promote differentiation. Depending on the culture conditions, the resulting cultures contained a heterogeneous mixture of cells representing either endoderm, ectoderm, or mesoderm.

EBs plated at high density or allowed to grow to confluence in 15% FBS spontaneously differentiate to excrete both a hard calcified bone matrix positive for Alizarin red and an Alcian Blue cartilage matrix (Fig. 4B), indicative of the mesoderm lineage. CollagenIIA1 mRNA was also detected in these cultures.

Rare β3-tubulin-positive cells with neuronal morphology occurred spontaneously without specific induction and without an EB transition stage (Fig. 4C, 4D). This explains the faint but consistent detection of β3-tubulin mRNA in cultures of undifferentiated FHDO-7 cells. In contrast, EBs plated in low serum or KO serum gave rise to patches of cells that expressed β3-tubulin (Fig. 4E, 4F). In support of this, β3-tubulin message was also strongly upregulated by growth in KO serum alone (Fig. 5A), as was mRNA for γ-enolase, both β3-tubulin and γ-enolase markers for ectoderm-derived neuronal cells.

Transwell cocultures of EBs (lower chamber) and canine fetal lung tissue (upper chamber) resulted in cells staining positive for CCSP (Fig. 4E, 4F), an endodermal marker typical of very early lung cell development. EBs grown in KO serum and Activin A (Fig. 5A) without fetal lung tissue also upregulated CCSP mRNA, suggesting that the positive staining in the presence of lung tissue was not merely absorption by FHDO-7 cells of lung-secreted CCSP. A second endodermal marker, AFP, was induced most strongly by Activin A and BMP4 in the presence of 0.5% serum (Fig. 5B).

FHDO-7 has been carried through past passage 34 and has maintained expression of Oct4 and alkaline phosphatase, as well as the ability to differentiate. The karyotype has been verified as normal through passage 15 (Fig. 6); however, aneuploidy did result when FHDO-7 was passaged as single cells following treatment with TrypLE. The most common abnormality was trisomy 8, which appeared to confer a growth advantage.

Attempts to grow cES cell-derived teratomas in immune-compromised mice were unsuccessful. First, we injected 106 FHDO-6 or FHDO-7 cells into both testes of 5 NOD/SCID mice and followed these mice for 13 weeks. We then injected 106 cells under the kidney capsule of 8 NOD/SCID γ-null mice and followed these mice for >8 weeks. Next, we asked more experienced investigators at the University of Washington (Drs. Michael LaFlamme and Carol Ware) for assistance. Dr. LaFlamme injected 5 × 105 cells into the heart of three Rag 2−/− mice and followed these mice for 8 weeks. Finally, Dr. Ware, who directs the ESC laboratory, injected 3 × 106 cells subcutaneous into 6 SCID mice and followed the mice for 20 weeks. In all cases, there were no detectable teratomas. Of note, we also failed to detect canine hematopoiesis from CD34+ cells injected intravenously in NOD/SCID mice, whereas we detected consistent engraftment with human CD34+ cells. We conclude from this experience that it is particularly difficult to establish canine cells in immune-deficient mice. Evidence for the ability of FHDO-7 cells to differentiate into all lineages must await the further development of canine blastocyst injection and implantation into hormonally receptive dogs to produce chimeric animals. There are many steps to this procedure that require development.


Considerable effort is currently focused on understanding and controlling ES cell differentiation for the purpose of deriving cells for tissue replacement therapy. Studies using human ES cells have, by necessity, been restricted to in vitro or xenogenic in vivo models. Most in vivo analyses of ES cell-derived tissue grafts are performed with mouse ES cells in mouse recipients. One advantage of this model is that the ES cells and recipient mice can be genetically identical, thereby eliminating concerns of graft rejection. Unfortunately, until research permits the routine reprogramming of a patient's own somatic cells [23], human ES cell-derived products will not be matched with prospective recipients, so the problems of alloreactivity will have to be addressed.

Although the mouse model may provide proof of principle for ES cell-derived products, it has several limitations when extrapolating to humans. First, the small size of the mouse limits the proliferative demand placed on transplanted tissue. This might explain why gene-marked stem cells can fully reconstitute the marrow of a mouse but contribute only a small proportion of cells in the marrow of a dog or nonhuman primate. The short life span of the mouse also prevents long-term follow-up, which is critical for assessing genetic or epigenetic stability of the transplanted tissue. Clearly, a large-animal preclinical model is needed to test the safety and efficacy of ES cell-derived tissue prior to applying this technology to patients.

Both the nonhuman primate and canine models are known to accurately predict clinical outcomes in adult stem cell transplantation and are therefore likely to serve as accurate preclinical models for ES cell-derived therapies. Several nonhuman primate ES cell lines have been developed for this purpose [24, [25], [26], [27]–28]. Given our experience with the canine model and the extensive canine resource already established at our institute, we have focused on the dog for a preclinical ES cell model. To date, there have been two reports of cell lines derived from canine blastocysts: one in which the canine ES cells were lost after eight passages [15], and a second, more recent report in which passage 12 cells were shown to differentiate into mesoderm only [16]. Stable canine ES cells that give rise to cells representing all three germ layers have not been described.

The generation of canine ES cells has been challenging. Since in vitro oocyte maturation or in vitro fertilization procedures have not been defined, blastocysts were harvested after natural matings. However, dogs ovulate immature oocytes that require maturation in the oviduct prior to fertilization, making the exact fertilization time difficult to pinpoint. The most definitive time point in canine ovulation is the preovulatory rise in progesterone (>2 ng/ml) and concurrent LH surge, which we have designated as day 0. On the basis of this timing, we harvested blastocysts at days 12 through 16 post-LH surge. We found that the cell lines with ES cell potential were derived from day 14 blastocysts. These blastocysts appear to be of a developmental stage similar to that of blastocysts used to derive ES cells from other species.

Derivation of embryonic stem cells is a poorly understood event. This is evident from the inability to efficiently derive new lines and the differences among ES cell lines within and between species [29]. There is a wide variation in the duration of preimplantation development among species, with mice and humans having a relatively short preimplantation phase, whereas in ungulates and canines it is more extended. Differences in time of harvest and derivation techniques result in ES cell lines with different potentials. Differences exist among the human ES cell lines [30, [31]–32], as well as murine [33, 34], equine [35, 36], and porcine [37, [38], [39]–40] lines.


The quintessential proof of ES cell potential would be the contribution of ES cells to all lineages in a chimeric animal. Currently, this technology is not available in the canine model. However, like ES cell lines from other species, FHDO-7 cells strongly express Oct4, Nanog, alkaline phosphatase, telomerase, and the ESC-associated microRNAs miR-302b, miR-302c, and miR-367, making it reasonable to postulate that the FHDO-7 are ES cells. Finally, the fact that FHDO-7 cells can differentiate in vitro into cells representing all three germ layers further supports our conclusion.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.


This work was supported in part by Grants P30-DK56465, P01-CA015704, T32-CA80416, and T32-CA09515 from the NIH. We thank Drs. Carol Ware and Mike LaFlamme (University of Washington), Alex Travis (Cornell University), and Margaret Hough (University of Toronto) for constructive criticism; Dr. Eileen Bryant for the cytogenetic analysis; and Bonnie Larson and Helen Crawford for assistance with the preparation and editing of the manuscript. B.H., S.R.F., and A.R. contributed equally to this work. Authors' contributions were as follows: B.H. performed research, analyzed data, and wrote the paper; S.R.F. performed research and analyzed data; A.R. performed research and analyzed data; S.B. performed research; M.H. contributed new reagents and analyzed data; L.G. contributed new reagents and analyzed data; M.B. performed research and analyzed data; A.B. performed research and analyzed data; M.T. analyzed data; and B.T.-S. designed research, analyzed data, and wrote the paper.