Standard procedures established by the SooAm Biotech Research Foundation (Seoul, Korea) for the Accreditation of Laboratory Animal Care were followed strictly when caring for the dogs used in our study. All experiments and surgical procedures were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals published by the SooAm Biotech Research Foundation. Canines at the pro-estrus stage (mixed breed, 1–7 years old, 20–25 kg body weight) were obtained from a breeder and reared indoors in separated cages up until oocyte recovery. Following oocyte recovery, the canines were sent back to the breeder. None of the dogs were used repeatedly. The pregnant recipient canines were kept in the research facility while the others were returned to the breeder.

All chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA), unless otherwise stated.

Human IMR-32 neuroblastoma cells (Korean Cell Line Bank, Seoul, Korea) were maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing penicillin and streptomycin, and 10% fetal bovine serum (FBS) (all from Invitrogen, Carlsbad, CA, USA). Primary culture of canine adult and fetal fibroblasts was performed as described in a previous study (8). Briefly, fetal fibroblasts were obtained from artificially inseminated embryos on the 30th day of pregnance by trypsinization. Fibroblasts were cultured in DMEM containing 10% FBS.

Various regions of the Thy-1 promoter [from nucleotides −2.0 to +2.3 kilonucleotides (knts), +1 = the transcriptional start site] were isolated by a long-range polymerase chain reaction (PCR) (LA Taq; Takara Bio, Inc., Shiga, Japan) using genomic DNA from beagle fibroblasts as a template. Amplified fragments were ligated into the MluI site of the promoterless pGL3-Basic vector (Promega, Madison, WI, USA). Human APP cDNA was prepared by PCR using the full-length clone from the Mammalian Gene Collection (cat. no. 6152423; Invitrogen) as a template. The amplified APP gene was mutated by two primer sets containing Swedish mutation (KM to NL) or Indiana mutation (V to F) indicated in Table I using a Site-Directed Mutagenesis kit (Stratagene, Inc., La Jolla, CA, USA). The mutated APP gene were further ligated into the NcoI and XbaI sites of the recombinant pGL3-Basic vector containing the Thy-1 promoter. An enhanced green fluorescent protein (EGFP) gene from the internal ribosomal entry site (pIRES)2_EGFP (Clontech Laboratories, Inc., Madison, WI, USA) was ligated into the EcoRV and BamHI sites of the pIRES_Neo vector (Clontech Laboratories, Inc.). This selection cassette was inserted into the SalI site of the backbone plasmid containing the dog Thy-1 promoter region and mutant human APP gene. The final construct (pGL3-pThy1-mhAPP-pCMV-EGFP/Neo) was linearized by NotI digestion (Fig. 1). All amplified products obtained at each step of the construct production were analyzed by sequencing analysis (Genotech Co. Ltd., Daejeon, Korea).

To control for different transfection efficiencies of the various luciferase constructs, a Rous sarcoma virus (RSV)-lacZ plasmid was used to co-transfect the IMR-32 cells along with the Thy-1 promoter-luciferase construct. Briefly, 3×10⁵ IMR-32 cells were seeded in six-well tissue culture plates one day prior to transfection. Thy-1 promoter-luciferase constructs (3.5 μg) and 0.5 μg of the RSV-lacZ plasmid were used to transiently transfect the cells using Lipofectamine™ 2000 (Invitrogen) followed by incubation for an additional 48 h. Cellular lysates were prepared using 150 μl of reporter lysis buffer and assayed for luciferase activity using the Luciferase Assay System (both from Promega). Luminescence was measured using a GloMax 20/20 Luminometer (Promega) and β-galactosidase activity was measured using a β-galactosidase Enzyme Assay System (Promega). Promoter activity was expressed as a percentage of relative luciferase activity (RLA %; luciferase/β-galactosidase activity).

Mature canine oocytes were collected in vivo by laparotomy at 72–84 h after ovulation using standard procedures (8,9). The time of ovulation was determined by the serum progesterone concentration and vaginal cytology (8). All the operative and post-operative procedures and care were performed by licensed veterinarians. The reproductive tract was exposed during mid-ventral laparotomy under a general anesthesia. The fimbriated end of the oviduct was canalized using a six-gauge bulbed needle held in place by a surgical ligature. The oviductal lumen at the base of the utero-oviductal junction was cannulated using a 24-gauge hypodermic needle (Angiocath TM Plus; Becton-Dickinson, Franklin Lakes, NJ, USA). Approximately 10 ml of TCM 199 collection medium supplemented with HEPES (Sigma-Aldrich) was used to flush the hypodermic needle. The medium passed through the oviductal lumen and bulbed needle before being deposited into a sterile plastic Petri dish. After flushing both oviducts, each ovarian bursa was incised to pull the ovary out and the corpora lutea were counted. Finally, the abdominal incision was closed using a two-layer method followed by the application of a surgical adhesive along the skin incision.

SCNT was performed according to a previously described procedure (5,9) with some modifications. Briefly, enucleation of the stripped oocytes was performed under an inverted microscope with epifluorescence (TE2000-E; Nikon Corp., Tokyo, Japan). Using an injection pipette, a trypsinized fibroblast with a smooth surface was transferred into the perivitelline space of an enucleated oocyte. The couplets were fused with two DC pulses of 1.75 kV/cm for 15 μsec from a BTX Electro-Cell Manipulator 2001 (BTX Inc., San Diego, CA, USA) and transferred to naturally synchronized recipients The embryos were loaded into a Sovereign Tom Cat catheter (Sherwood Medical, St. Louis, MO, USA) with a minimal volume of medium and gently transferred into the 2/3 to deep distal portion of the oviduct without insufflating air.

Approximately 30 days after embryo transfer, the recipients were examined using a portable real-time ultrasonography machine with a 3.5-MHz curved transducer (SSD-900; Aloka Co. Ltd., Tokyo, Japan). To detect embryonic or fetal death, sizes and shapes of the chorionic cavities along with embryonic or fetal heartbeats were examined. After confirming pregnancy, this examination was performed every seven days to monitor fetal development until birth.

Canine genomic DNA was isolated using a G-DEX™ II Genomic DNA Extraction kit (Intron Biotechnology, Inc., Suwon, Korea). Genomic DNA (100 ng) or plasmid DNA (10 ng) was amplified in a 50-μl PCR reaction containing 1 U of LA™ or Ex™-Taq polymerase (Takara Bio, Inc.) and 10 pmol of specific primers. PCR was carried out for 35 cycles of denaturation for 30 sec at 94°C, and annealing and extension for 1–3 min at 68°C followed by a final extension for 30 min at 72°C. All primers are presented in Table I. PCR products were separated on an agarose gel, stained with ethidium bromide and photographed under UV illumination. The images were scanned using a Gel Doc EQ system (Bio-Rad Laboratories, Inc., Hercules, CA, USA).

Total RNA was isolated from the heart, pancreas, liver, lungs, kidneys, spleen and cerebrum of two puppies (AM144 and AM145) using TRIzol reagent (Invitrogen). DNA contamination was further eliminated by treatment with DNAase (Invitrogen). Total RNA (1 μg) was reverse transcribed into first-strand cDNA using M-MLV reverse transcriptase (Invitrogen) and random primers (9-mer; Takara Bio, Inc.). The GAPDH gene was amplified to evaluate RNA degradation and was used to control for variation in mRNA concentration in the RT-PCR reactions. GAPDH and human APP genes were quantified using 28 and 30 cycles, respectively. cDNA was amplified in a 20-μl PCR reaction containing 1 U TaqDNA polymerase (Intron Biotechnology, Inc.) and 10 pmol of the specific primers (Table I). The PCR procedure included denaturation at 95°C for 30 sec, annealing at 58°C for 30 sec and extension at 72°C for 30 sec. The PCR products (8 μl) were separated on a 2% agarose gel, stained with ethidium bromide and photographed under UV illumination. The images were scanned using a Gel Doc EQ system (Bio-Rad Laboratories, Inc.).

Proteins (40 μg) were separated by electrophoresis on SDS-PAGE and then transferred onto a PVDF membrane. The membrane was incubated with antibodies against the following proteins: APP and Aβ (1:1,000; cat. nos. 2450 and 2454; Cell Signaling Technology, Beverly, MA, USA) or β-actin (1:2,000; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). Immunoreactive proteins were visualized by exposure to X-ray film. Band intensities were quantified by image scanning using a Gel Doc EQ system (Bio-Rad Laboratories, Inc.), corrected by background subtraction, and normalized using β-actin as an internal control.

Brain tissues were embedded in paraffin. Sections (5-μm thick) were deparaffinized and hydrated in descending grades of ethanol. The sections were subsequently incubated at room temperature for 2 h with an anti-Aβ antibody (diluted 1:100, #2454; Cell Signaling Technology) and rabbit anti-ionized calcium-binding adapter molecule 1 (Iba1, diluted 1:500, #019–19741; Wako Pure Chemical Industries Ltd., Osaka, Japan) in 10% normal goat serum. The sections were reacted with biotinylated goat anti-rabbit IgG (Vector ABC Elite kit; Vector Laboratories, Inc., Burlingame, CA, USA) for 45 min at room temperature. Immunoreactivity was assessed using the avidin-biotin peroxidase complex (Vector ABC Elite kit) prepared according to the manufacturer’s instructions. The peroxidase reaction was developed using a diaminobenzidine substrate kit (SK-4100; Vector Laboratories, Inc.). The primary antibodies were omitted for a few test sections in each experiment as a control. The sections were counterstained with Harris’s hematoxylin before mounting.

We generated a human APP gene expressing a cassette containing two well-characterized mutations, Swedish (K670N and M671L) and Indiana FAD (V717F), which successfully induces AD-like pathological symptoms in mouse models (10–12). The mhAPP gene was controlled by a canine Thy-1 promoter for selective expression in neural tissues. Specific regions of the Thy-1 promoter were selected by a luciferase activity assay in human IMR-32 neuroblastoma cells (Fig. 2A and B). We deleted the upstream portion (−2.0 to −1.5 knts, +1 = transcriptional start) of the Thy-1 promoter as it possessed unknown inhibitory elements, and the neural element (between −2.0 to −2.3 knts) in intron A was added as it was essential for promoter activity in the IMR-32 cells. Finally, the Thy-1 promoter (−1.5 to +2.3 knts) was found to have maximum activities similar to the SV promoter in neural cells. The transgenic vector was transiently expressed in IMR-32 cells and induced expression levels of APP transcripts two-fold higher compared to the empty pGL3-Basic Vector (Fig. 2C). The transgenic vector also contained a selection cassette including EGFP and neomycin resistance genes linked by IRES, and were controlled by the CMV promoter to facilitate selection for further processes.

Canine fibroblasts were transfected with the linearized transgenic vector and cultured in medium containing 350 μg/ml of G418 to select a transgenic clone. The cells resistant to G418 for four weeks were further confirmed to express EGFP by fluorescence microscopy (Fig. 3A). The insertion of the transgenic vector in the transgenic cells was confirmed by PCR (Fig. 3B). The positive cells were isolated and frozen until SCNT.

SCNT was performed using the transgenic cell line and 332 oocytes matured in vivo from 29 donors. Six puppies (AM141 to 146) were born from four out of 17 surrogates. Five puppies (AM141, 142, 143, 144 and 146) were confirmed to be transgenic animals by PCR (Fig. 4A), and observed to express EGFP in their nails, toes and fur (Fig. 4B).

An EGFP-positive puppy (AM144) and EGFP-negative littermate (AM145) were sacrificed to measure mhAPP expression inserted by SCNT. To eliminate genomic DNA contamination, we treated DNase I on total RNAs before the reverse-transcriptase reaction. In addition, we designed specific primers to distinguish between human and canine APP transcripts on their mRNA sequences. The expression of the mhAPP gene was only observed in the organs of the EGFP-positive puppy (Fig. 4C). The amplified mhAPP fragments were confirmed not to be due to genomic DNA contamination (Fig. 4D). Endogenous canine APP mRNA was detected in the organs of both animals. Amplified mhAPP transcription was further confirmed by sequence analysis. Thus, the transgenic puppies we created were found to successfully express the exogenous mhAPP gene.

One of the mhAPP transgenic canines (AM146) was mated with a wild-type female dog (AF165) to produce more puppies for observing the development of AD-like symptoms and stabilizing of the mhAPP gene. This mating produced four male and three female puppies (SB16 to 22). All canines had the transgenic vector in their genomic DNA apart from SB21 (Fig. 5A). When the puppies were six months old, one transgenic puppy (SB17) presented recurrent behavioral disorders, such as tetanic convulsions. Although seizures are not a common phenotype of AD, recent reports have indicated that seizures may be one of the phenotypes of AD (13–17). The brain of SB17 was further examined by magnetic resonance imaging (MRI, 7.0 T). AD-like symptoms, including enlarged ventricles and an atrophied hippocampus, were observed (Fig. 5B). The symptoms are not common in canines; however, emerging evidence has indicated that ventricle enlargement and hippocampal atrophy are phenotypes of human AD (18–21). Although the other littermates did not develop the behavioral disorders observed in SB17, an MRI image of SB19 and SB20 showed results similar to those of SB17, indicating that these puppies also had enlarged ventricles and an atrophied hippocampus (Fig. 5C and D). Our results indicated that the transgenic puppies had developed human AD-like symptoms.

SB17 was sacrificed and pathological AD phenotypes in the entire brain were analyzed. The cerebral ventricles were greatly enlarged (Fig. 5E). The hippocampus had also atrophied and became degenerated in the transgenic puppy compared to the non-transgenic animal, as previously documented by an MRI. To further characterize this animal, we examined the molecular mechanisms underlying the AD-like symptoms. APP was highly expressed in the cerebrum and cerebellum of the transgenic puppy compared to the control animal (Fig. 6). However, we could not distinguish the expression of mhAPP from that of endogenous canine APP with the currently available anti-APP sera. Increased levels of Aβ, hallmarks of AD, were also observed in the transgenic puppy. This finding suggests that the overexpression of mhAPP in canines induces an accumulation of the hallmarks of AD (Aβ accumulation in the brain, resulting in AD-like symptoms such as enlarged ventricles, an atrophied hippocampus and abnormal behavior).

For additional histopathological characterization, Aβ accumulation in the hippocampal regions was further measured with an immunohistochemical assay. In the hippocampus, Aβ-positive immunoreactivity was rarely detected in non-transgenic canines (Fig. 7A and D), whereas the immunoreactivity was observed to a great extent in the transgenic canines (Fig. 7B and E). Additionally, microglial activation was observed in the hippocampus of the transgenic canines where Aβ accumulation was present (Fig. 7C and F). These findings suggest that Aβ accumulation in cells induces an immune response, as previously described (22).