Laparoscopic Oviductal Artificial Insemination Improves Pregnancy Success in Exogenous Gonadotropin-Treated Domestic Cats as a Model for Endangered Felids

Abstract

Artificial insemination (AI) in cats traditionally uses equine chorionic gonadotropin (eCG) and human chorionic gonadotropin (hCG) to induce follicular development and ovulation, with subsequent bilateral laparoscopic intrauterine insemination. However, long-acting hCG generates undesirable secondary ovulations in cats. Uterine AI also requires relatively high numbers of spermatozoa for fertilization (∼8 × 10⁶ sperm), and unfortunately, sperm recovery from felids is frequently poor. Using short-acting porcine luteinizing hormone (pLH) instead of hCG, and using the oviduct as the site of sperm deposition, could improve fertilization success while requiring fewer spermatozoa. Our objectives were to compare pregnancy and fertilization success between 1) uterine and oviductal inseminations and 2) eCG/hCG and eCG/pLH regimens in domestic cats. Sixteen females received either eCG (100 IU)/hCG (75 IU) or eCG (100 IU)/pLH (1000 IU). All females ovulated and were inseminated in one uterine horn and the contralateral oviduct using fresh semen (1 × 10⁶ motile sperm/site) from a different male for each site. Pregnant females (11/16; 69%) were spayed approximately 20 days post-AI, and fetal paternity was genetically determined. The number of corpora lutea (CL) at AI was similar between hormone regimens, but hCG increased the number of CL at 20 days post-AI. Numbers of pregnancies and normal fetuses were similar between regimens. Implantation abnormalities were observed in the hCG group only. Finally, oviductal AI produced more fetuses than uterine AI. In summary, laparoscopic oviductal AI with low sperm numbers in eCG/hCG- or eCG/pLH-treated females resulted in high pregnancy and fertilization percentages in domestic cats. Our subsequent successes with oviductal AI in eCG/pLH-treated nondomestic felids to produce healthy offspring supports cross-species applicability.

Introduction

In situ conservation of wildlife has been increasingly challenged by the growth of human populations and the escalating consumption of natural resources. Accordingly, ex situ conservation efforts have been receiving renewed attention as a potential safeguard for species facing uncertain futures in the wild. Assisted reproductive technologies (ARTs) have been incorporated into a few ex situ conservation programs to a limited degree and, in rare cases, have helped to bring critically endangered species back from the brink of extinction. One notable example of the latter is the black-footed ferret (Mustela nigripes), an imperiled species that benefitted substantially from application of artificial insemination (AI) as one component of a genetic management program (for review, see [1]).

Other endangered carnivore species, including felids, could potentially benefit from similar conservation efforts. Currently, 36–40 species of nondomestic felids have been identified—depending on the specific classification system used by taxonomists and geneticists [2] (for review, see [1])—and most of these felid populations are declining to some extent in the wild, mainly due to habitat loss and degradation [2]. Within zoological institutions, realization is growing that the effective application of ARTs, especially AI, will be a necessity for genetic management and long-term sustainability of endangered felids into the future [3–5]. AI represents a potentially valuable approach for creating gene flow, especially through the use of frozen-thawed spermatozoa. However, for the potential of AI to become a reality for genetic management of felids, substantial improvements in AI efficiency are needed for producing pregnancies and viable offspring in this diverse taxon.

For AI in felids, the standard hormone regimen for inducing ovarian stimulation and ovulation involves treatment with equine chorionic gonadotropin (eCG) followed 80–85 h later by human chorionic gonadotropin (hCG) [6] (for review, see [1]). However, these large exogenous glycoproteins have relatively long elimination half-lives (∼23 h) and remain in circulation for several days postinjection in domestic cats (at least 5 days for eCG and 4 days for hCG); this causes undesirable secondary follicular growth and ovulations in female cats [7, 8]. An alternative ovulatory signal, porcine luteinizing hormone (pLH; 1000 IU), following eCG treatment, has proven to be highly effective for inducing ovulation in domestic cats for embryo transfer procedures without significant formation of secondary ovarian structures [9]. To our knowledge, however, the effectiveness of this eCG/pLH regimen with AI procedures has never been assessed.

Traditionally, AI in nondomestic cats has involved using a minimally invasive laparoscopic method for bilateral sperm deposition into the uterine lumen (for review, see [1]). Although live offspring have been produced in eight nondomestic cat species to date using this approach, pregnancy percentages following AI in most species have been low (<10%). Further, uterine AI often requires relatively high numbers of spermatozoa (>10 million motile sperm) to achieve sperm transport through the uterotubal junction (UTJ) and fertilization within the oviduct. Nonsurgical AI methods including vaginal insemination and transcervical catheterization for intrauterine insemination also have resulted in pregnancies in domestic cats, but even higher sperm numbers (∼20–80 million sperm) typically are required [10, 11]. Sperm numbers and quality in ejaculates obtained from nondomestic cats frequently are poor, limiting the number of motile, morphologically normal spermatozoa available for AI [12, 13]. A laparoscopic oviductal (LO) AI approach could potentially improve pregnancy outcomes by allowing deposition of semen in closer proximity to the site of fertilization while using reduced numbers of motile spermatozoa for successful insemination [14, 15]. For example, small-sized cat species such as the black-footed cat (Felis nigripes), which average less than 30 million total sperm cells per ejaculate with fewer than 50% of spermatozoa classified as morphologically normal, could greatly benefit from this novel LO AI approach (for review, see [13]).

The goal of the present study was to determine if alterations in the standard AI protocol for domestic cats, including the use of short-acting pLH as the ovulatory signal and oviductal AI with low sperm numbers, could improve overall fertilization and pregnancy success for AI in felids. The specific objectives were 1) to compare ovarian responses and in vivo fertilization success in domestic cats treated with one of two gonadotropin regimens (eCG/hCG vs. eCG/pLH); 2) to assess relative fertility and pregnancy percentages following laparoscopic insemination at two distinct sperm deposition sites, the uterine horn and the contralateral oviduct, within individual females; and 3) to assess the capacity of eCG/pLH-treated females to carry offspring to term following AI.

Materials and Methods

Study 1: Intrauterine Versus Intraoviductal AI Following Ovarian Stimulation with eCG/hCG or eCG/pLH

Animals

A total of 18 domestic short-hair cats were used for study 1. Females (n = 16), ranging in age from 10 mo to 3 yr, were all nulliparous and naïve to exogenous gonadotropin treatment. Two proven breeder males were used as semen donors; one male (male 1) was the sire of the second male (male 2). Females were group-housed in a single large room (total area, 47 m³), whereas males were maintained individually or as pairs within pens in a separate room. All cats were maintained under a controlled environment (temperature, 22–25°C; humidity, 25%–29%; photoperiod, 14L:10D) based on federal animal welfare guidelines and were provided with commercial cat food (Iams; Procter & Gamble) daily and water ad libitum. Rooms were enriched with climbing structures (i.e., shelves, hammocks, and cat condos), beds and rugs, and multiple toys. All procedures were reviewed and approved by the Cincinnati Zoo and Botanical Garden Animal Care and Use Committee (protocols 11–102 and 11–103).

Gonadotropin treatment

The females used in study 1 were selected from a pool of candidates within our cat colony. Female cats were monitored every 1–3 days for signs of behavioral estrus (i.e., lordosis, rolling, pedaling of hindfeet, vocalizing, and sexual interest in other females) and excluded from gonadotropin treatment if estrus was detected. In the absence of behavioral signs of estrus, the female had a blood sample collected and recovered serum assessed for semiquantitative progesterone concentrations using a bench-top enzyme immunoassay kit (Target Rapid Feline Progesterone Kit; BioMetallics). Females that had ovulated spontaneously, as indicated by progesterone concentrations greater than 1 ng/ml, were temporarily excluded from gonadotropin treatment. Only nonestrual, nonluteal females were selected for gonadotropin injections and subsequent AI.

Lyophilized eCG (ProSpec-Tany TechnoGene Ltd.), hCG (Sigma-Aldrich), and pLH (Sioux Biochemical) were reconstituted in sterile water (Butler Schein Animal Health) to final concentrations of 400, 300, and 3000 IU/ml, respectively. Individual doses of eCG (100 IU), hCG (75 IU), and pLH (1000 IU) were aspirated into 1-ml syringes and kept frozen at −80°C until use (i.e., within 3 mo of freezing).

Each female was randomly assigned to treatment with either eCG/hCG or eCG/pLH in an alternating fashion (see below). All females received an initial injection of eCG (100 IU i.m.) followed 85 h later with an injection of either hCG (75 IU i.m.) or pLH (1000 IU i.m.), with AI conducted 31–33 h after the second gonadotropin injection.

Semen collection and processing

For semen collection, naturally estrual females that were not scheduled for AI were placed in an enclosed common holding area with the donor males. Each male was allowed to approach and mount each teaser female, and semen was collected using an artificial vagina [10]. Semen was collected from both males for each AI procedure.

Recovered semen samples were diluted 1:6 in Feline-Optimized Culture Medium with Hepes (FOCMH), and aliquots (5 μl) were used to assess percentage of sperm progressive motility, rate of progressive movement, concentration, morphology, and acrosome integrity [16]. Briefly, each sample was evaluated microscopically for percentage (0%–100%) of progressively motile sperm and rate of progressive movement (scale of 0–5). Sperm concentration was determined following seminal aliquot dilution (1:200) in tap water using a hemocytometer method. Remaining semen was centrifuged (600 × g, 10 min) and the resulting sperm pellet resuspended in 100–200 μl of FOCMH. Motility and concentration were reassessed and the sample diluted in FOCMH to 200 × 10⁶ motile spermatozoa/ml. A small aliquot (∼5 μl, 1 × 10⁶ motile spermatozoa) was used to inseminate each location (oviduct or uterus) within the female's reproductive tract.

Laparoscopic AI procedure

If the semen samples collected from both males were of adequate quality (progressive motility, >70%; rate of progressive movement, >3.0), the AI candidate was immediately anesthetized with an intramuscular injection of ketamine (9 mg/kg; Ketaset, Fort Dodge Animal Health) combined with acepromazine (0.09 mg/kg; PromAce, Fort Dodge Animal Health). Anesthesia was maintained throughout the procedure using 1%–3% isoflurane (Aerrane; Baxter Healthcare Corporation) with oxygen, delivered using a facemask.

Laparoscopy was performed as previously described [6] using a 7-mm laparoscope (Richard Wolf Medical Instruments Corporation). Before AI, the reproductive tract was thoroughly evaluated for any potential pathology (i.e., endometrial hyperplasia and ovarian or paraovarian cysts) or abnormalities (i.e., adhesions). Ovarian follicles (diameter, ≥2 mm) and corpora lutea (CL) were counted and measured, and the uterine horns were assessed for diameter, general tone, and presence of segmentation.

For laparoscopic uterine (LU) AI, atraumatic grasping forceps (diameter, 5 mm; Richard Wolf Medical Instruments Corporation) were inserted into the abdominal cavity through an accessory cannula and used to secure one uterine horn adjacent to the ventral abdominal wall [6]. An intravenous catheter (18G, 5 cm length, 20G stylet; Terumo Surflo Teflon IV Catheter; Terumo Medical Corporation) was inserted percutaneously through the abdominal wall directly into the uterine lumen near the midpoint of the uterine horn, then the catheter stylet was slightly withdrawn and the catheter advanced approximately 2 cm into the cranial uterus. Polyethylene tubing (PE10 tubing; Stoelting Co.), containing the semen dose at its distal end, was passed through the catheter, and concurrent with continuous forward advancement of the tubing, the catheter was completely withdrawn from the uterine horn. With the distal tip of the AI tubing in place near the UTJ, the semen was deposited into the uterine lumen using slight air pressure from an attached 1-ml syringe.

For LO AI, custom-made grasping forceps (diameter, 5 mm; MDS Incorporated) were inserted into the abdominal cavity through an accessory cannula and used to secure the craniomedial edge of the ovarian bursa. The bursa was extended and everted to allow visualization of the abdominal oviductal ostium. An intravenous catheter (18G, 32 mm length; Terumo Medical Corporation) was inserted percutaneously proximal to the ovarian bursa, and a modified 22G needle (i.e., a blunted stylet from a 20G IV catheter, 68 mm length; Terumo Medical Corporation), attached to a 1-ml syringe, was inserted through the catheter approximately 2 cm deep into the oviductal ostium. Semen (∼5 μl) was deposited into the oviductal lumen with slight air pressure.

Randomization of treatment groups

For each female, semen was deposited into one cranial uterine horn and into the contralateral oviduct, using spermatozoa from two different males, one for each location. To randomize the study design, a set criterion was used to select the site of semen deposition based on the number of CL observed on each ovary at laparoscopy before AI. For each AI, the ovary with the most CL was considered to be the “best” side. The specific semen donor to be used on each side was chosen based on the same criterion. At the initiation of the study, the gonadotropin regimen (eCG/hCG or eCG/pLH), AI location, and semen donor (for the “best” side) were randomly chosen for the first female. For subsequent females, gonadotropin regimen, AI location, and semen donor were alternated accordingly to ensure even distribution of variables among all AI procedures (Table 1).

Table. 1.

Randomization of gonadotropin treatment, AI method, and semen donor for AI procedures in domestic cats.

Cat	Hormone regimena	Best sideb		Opposite sidec
		AI technique	Male donor	AI technique	Male donor
Female 1	pLH	LU	Male 1	LO	Male 2
Female 2	hCG	LO	Male 2	LU	Male 1
Female 3	pLH	LO	Male 2	LU	Male 1
Female 4	hCG	LU	Male 1	LO	Male 2
Female 5	pLH	LU	Male 2	LO	Male 1
Female 6	hCG	LO	Male 1	LU	Male 2
Female 7	pLH	LO	Male 1	LU	Male 2
Female 8	hCG	LU	Male 2	LO	Male 1
Female 9	pLH	LU	Male 1	LO	Male 2
Female 10	hCG	LO	Male 2	LU	Male 1
Female 11	pLH	LO	Male 2	LU	Male 1
Female 12	hCG	LU	Male 1	LO	Male 2
Female 13	pLH	LU	Male 2	LO	Male 1
Female 14	hCG	LO	Male 1	LU	Male 2
Female 15	pLH	LO	Male 1	LU	Male 2
Female 16	hCG	LU	Male 2	LO	Male 1

^aAll females received eCG as a first injection and either hCG or pLH as the second injection.

^bBest side was the side of the reproductive tract (left or right) containing the ovary with the highest number of fresh CL.

^cOpposite side was the contralateral side of the reproductive tract.

The left side (oviduct or uterus) of the female's reproductive tract was always inseminated initially due to slightly easier accessibility and manipulation associated with insertion of the grasping forceps through the right abdominal wall. To maintain similar time intervals between semen collection and AI for both sides, the semen donor for the left side was always collected first.

Pregnancy detection and paternity of fetuses

At 20–22 days post-AI, female cats were examined via ultrasonography (7.5-mHz linear probe; Aloka Micrus; Aloka Co., Ltd.) for pregnancy assessment. The number of gestational vesicles, if any, were counted and recorded. Nonpregnant females were anesthetized within 24–48 h and evaluated via laparoscopy to assess ovarian structures as previously described. Pregnant females were anesthetized and subjected to ovariohysterectomy for fetal tissue collection and isolation of fetal genomic DNA.

After ovariohysterectomy, the ovaries were separated from the rest of the reproductive tract and weighed, and the number of ovarian structures (CL and follicles with a diameter, ≥2 mm) were counted. For the uterus, the number and dimensions (length and width) of individual gestational sacs were determined. Each gestational sac was carefully incised to prevent DNA cross-contamination, and the fetus and placenta were assessed grossly for morphology and development. Each fetus was measured for crown-rump length after positioning to straighten the natural body curvature, then weighed and bisected. Fetal tissue (half of each fetus) was transferred into a labeled cryovial containing 1 ml of RNA stabilization reagent (RNAlater; Qiagen, Inc.) and kept frozen (−80°C) until analysis. Venous blood samples from all pregnant females and the two donor males also were collected into ethylenediaminetetra-acetic acid tubes and maintained at 4°C for isolation of genomic DNA.

Extraction of genomic DNA was performed using the DNeasy Blood & Tissue kit (Qiagen, Inc.) according to manufacturer's specifications. From each fetal sample, approximately 10 mg of tissue were used for DNA extraction. Tissue samples were finely sliced and incubated for 16 h in tissue lysis buffer with proteinase K (Qiagen, Inc.). Following incubation, samples were isolated using the DNeasy Blood & Tissue kit following the manufacturer's animal tissue extraction protocol.

Thirty-eight short tandem repeats (STR) were used to determine parentage of the fetuses [17]. Ten nanograms of template DNA were amplified following previously described methods [17]. PCR amplicons were multiplexed and size separated on an ABI 3730 DNA Analyzer (Applied Biosystems), and sizes were determined using STRand software [18]. Sixteen additional STR markers (Supplemental Table S1, available online at www.biolreprod.org) were used to resolve parentage in two offspring using the reaction conditions described above.

Study 2: Offspring Production after AI of Females Treated with eCG/pLH

In study 2, three females were selected for eCG/pLH treatment followed by AI (LU and LO, on opposite sides) to assess the capacity of pregnant females to carry offspring to term. The three additional females (females 17–19) included two nulliparous and one multiparous queens; all had been treated previously one or more times with exogenous gonadotropins in earlier studies. AI was conducted as previously described, with ultrasonography performed 20 days post-AI to assess pregnancy status. Pregnant females were allowed to carry offspring to term and give birth naturally. Blood samples were collected from resulting kittens for parentage analysis as previously described. For the litter from one cat (female 17), an additional 16 STR markers were necessary to support parentage qualification.

Statistical Analysis

All statistical analyses were performed using SAS Software (Version 9.1; SAS Institute, Inc.). The effect of AI technique (LU or LO) on number of pregnancies was assessed using the McNemar test. Chi-square test was used to compare the number of females that became pregnant by AI following treatment with eCG/hCG versus the number following treatment with eCG/pLH. A mixed model with cat as a random effect was used to investigate the effect of the following fixed, independent variables on number of fetuses produced: 1) AI technique (LU or LO), 2) exogenous ovulation-inducing hormone (hCG or pLH), 3) sire (male 1 or male 2), and 4) side of the female's reproductive tract for semen deposition (left or right). The number of fetuses produced by AI technique (LU or LO) also was compared by sign rank test.

A paired t-test was used to compare the number of follicles (diameter, ≥2 mm) and CL, observed at AI as well as at 20 days post-AI, between females receiving eCG/hCG versus females receiving eCG/pLH. A Student t-test and the nonparametric sign and signed rank tests were used to compare the number of follicles or the number of CL between 0 and 20 days post-AI within each gonadotropin treatment group. Kruskal-Wallis nonparametric test was used to investigate the effect of hormone on the number of gestational sacs and the incidence of abnormalities in the gestational sacs. The influence of hormone on fetal dimensions (length and weight) was assessed via mixed model (with hormone, date of fetal harvest, and sire as fixed, independent variables and dam as a random effect). The correlations between fetal dimensions (length and weight) and date of fetus harvest or number of fetuses produced by female were analyzed using Pearson correlation.

Level of significance was set at 0.05, and P-values greater than 0.05 and up to 0.1 were considered to be approaching significance. Normality of data was assessed via normality plots and Shapiro-Wilk and Anderson-Darling tests, and homogeneity of variance was assessed via the Levene test and residual plots. Data are expressed as the mean ± SEM.

Results

Study 1: Intrauterine Versus Intraoviductal AI Following Ovarian Stimulation with eCG/hCG or eCG/pLH

Comparison of hCG versus pLH

All 16 cats within the gonadotropin treatment groups (eCG/hCG and eCG/pLH) developed multiple ovarian follicles and ovulated within 31–33 h of the second gonadotropin injection (Table 2). Follicle diameter ranged from 2 to 6 mm, as did CL diameter, with the exception of female 15, which had one 8-mm follicle at AI that grew to 10 mm 3 wk later and was possibly cystic. Diameter of each uterine horn was approximately 8 mm for most cats (n = 14), with the exception of two cats (females 15 and 16; diameter, 10 mm). Uterine tone varied from slight (n = 7) to moderate (n = 9), and uterine segmentation was absent in all cats except for one cat (female 7) whose horns were slightly segmented. The mean number of follicles observed at AI did not differ (P > 0.05) between eCG/hCG- and eCG/pLH-treated females, nor did the number of follicles 20 days post-AI (P > 0.05) (Table 3). At AI, eCG/hCG- and eCG/pLH-treated females had similar (P > 0.05) numbers of CL. However, at 20 days post-AI, eCG/hCG treatment tended to result in higher numbers of CL compared to eCG/pLH treatment (P ≅ 0.06). Within treatment groups, eCG/hCG resulted in an increased (P < 0.05) CL number at 20 days post-AI compared to at AI, whereas CL numbers for eCG/pLH did not differ (P > 0.05) between the two time points. Follicle numbers did not differ (P > 0.05) between the two time points in either gonadotropin treatment group. At AI, 15 (of 16) females had bilateral ovulations; the single exception (female 1, treated with eCG/pLH) had fresh CL on the right ovary only.

Table. 2.

Number of follicles and CL at AI and on Day 20 post-AI and number of fetuses by dam according to gonadotropin treatment and insemination site.

Treatment	LO side					LU side
	Follicles		CL		Fetusesa	Follicles		CL		Fetusesa
	AI	Day 20	AI	Day 20		AI	Day 20	AI	Day 20
eCG/pLH
Female 1	5	0	0	0	0	3	0	2	2	1
Female 3	0	0	11	18	4	3	0	8	26	0
Female 5	2	2	4	5	3	2	2	5	5	2
Female 7	0	1	12	8	0	0	3	12	14	0
Female 9	0	3	4	2	0	0	7	5	5	0
Female 11	1	0	3	4	1	4	0	2	5	0
Female 13	0	5	1	3	2	0	5	3	4	0
Female 15	1	1	8	10	0	0	3	7	6	0
Total	9	12	43	50	10	12	20	44	67	3
Mean	1.1	1.5	5.4	6.2		1.5	2.5	5.5	8.4
SEM	0.6	0.6	1.6	2.0		0.6	0.9	1.2	2.8
No. of fetuses (mean ± SEM)b								2.6 ± 0.6
eCG/hCG
Female 2	8	0	6	48	6	4	1	5	40	2
Female 4	0	0	4	22	8	0	0	7	17	0
Female 6	0	0	4	2	0	2	2	2	3	0
Female 8	0	7	5	13	7	0	9	10	18	3
Female 10	0	3	17	28	3	2	2	12	23	4
Female 12	10	2	9	22	0	9	0	10	24	1
Female 14	0	0	4	4	0	1	1	2	2	0
Female 16	2	1	6	10	2	4	1	8	17	0
Total	20	13	55	149	26	22	16	56	144	10
Mean	2.5	1.6	6.9	18.6		2.7	2.0	7.0	18.0
±SEM	1.4	0.9	1.6	5.3		1.0	1.0	1.3	4.3
Total no. of fetuses/AI site (LO or LU)					36					13
No. of fetuses (mean ± SEM)b								6.0 ± 1.0

^aTotal number of normal and abnormal fetuses.

^bMean number of fetuses for pregnant cats only.

Table. 3.

Number of follicles (diameter, ≥2 mm) and CL at AI and on Day 20 post-AI and number of fetuses by gonadotropin treatment group.^a

Gonadotropin regimen	Follicles		CL		Fetusesb
	AI	Day 20	AI	Day 20
eCG/hCG	5.2 ± 2.4	3.6 ± 1.8	13.8 ± 2.0	38.0 ± 5.3	4.5 ± 1.5
eCG/pLH	2.6 ± 1.0	4.0 ± 1.5	10.9 ± 1.4	14.6 ± 2.4	1.6 ± 0.7
P-value	>0.05	>0.05	>0.05	∼0.06	∼0.1

^aValues are mean ± SEM.

^bMean number of fetuses for all females (n = 8 per group).

Overall, 11 (of 16) females (69%) became pregnant following AI. The two gonadotropin treatment groups provided similar (P > 0.05) numbers of pregnancies: six of eight with eCG/hCG (75%) versus five of eight with eCG/pLH (63%). Forty-nine fetuses were harvested, with 36 produced in the eCG/hCG-treated group and 13 in the eCG/pLH-treated group. On average, females treated with eCG/hCG tended to produce more fetuses than the eCG/pLH-treated females (P ≅ 0.1) (Table 3).

Reproductive tracts were recovered from most (9/11) pregnant females at 20 or 21 days post-AI. Two cats (females 8 and 13) were ovariohysterectomized at 16 and 22 days post-AI, respectively. Date of fetal harvest (20–22 days post-AI) did not influence fetal dimensions (weight and length) (P > 0.05); thus, data from different days of harvest were combined except for the cat (female 8) spayed at 16 days post-AI, which was excluded from all the analyses regarding fetal dimensions. After ovariohysterectomy of pregnant females (n = 11), ovarian weight was assessed (0.53 ± 0.09 g; range, 0.18–1.20 g). For both gonadotropin treatment groups combined, 57 gestational sacs were produced in study 1, including normal (i.e., normal placenta and normal fetus) and abnormal sacs. The gestational sac length was 22.3 ± 0.5 mm (range, 11.1–30.1 mm), and the width was 20.2 mm ± 0.5 mm (range, 8.8–26.3 mm). The smallest and the largest of the normal gestational sacs were 14.1 × 16.2 mm (length × width) and 30.1 × 23.7 mm, respectively.

The number of gestational sacs observed in each hormone treatment group differed (P < 0.05) (Table 4), with 44 (of 57) gestational sacs (77%) observed in the eCG/hCG-treated group and 13 (23%) in the eCG/pLH-treated group. The number of normal gestational sacs (i.e., containing normal placenta and fetus) did not differ (P > 0.05) between hormone treatment groups, but the incidence of developmental abnormalities (i.e., absent, necrotic, or malformed placenta and/or fetus) did differ (P < 0.05) (Table 4): in the eCG/pLH group. No gross abnormalities were observed in any fetuses or placentae; however, in the eCG/hCG-treated group, 16 (of 44) gestational sacs (36%) exhibited various developmental abnormalities. The incidence of empty gestational sacs was influenced by the type of gonadotropin treatment (P < 0.05); eight empty gestational sacs were found in the eCG/hCG treatment group but none in the eCG/pLH group. Five (of the six) pregnant females in the eCG/hCG-treated group had at least one gestational sac that did not contain a fetus (Table 4).

Table. 4.

Number of gestational sacs (GS) and abnormalities in pregnant eCG/pLH- and eCG/hCG-treated cats following AI.

Treatment	Total GS	Normal GS (normal fetuses)a	Abnormal GS			Total abnormal GS
			Normal fetusb	Abnormal fetusc	Without fetusd
eCG/pLH
Female 1	1	1	—	—	—	—
Female 3	4	4	—	—	—	—
Female 5	5	5	—	—	—	—
Female 11	1	1	—	—	—	—
Female 13	2	2	—	—	—	—
Total	13	13	0	0	0	0
eCG/hCG
Female 2	11	8	—	—	3	3
Female 4	9	7	—	1	1	2
Female 8	11	6	3	1	1	5
Female 10	8	4	1	2	1	4
Female 12	3	1	—	—	2	2
Female 16	2	2	—	—	—	—
Total	44	28	4	4	8	16
P-value	0.05	>0.05	>0.05	<0.1	<0.05	<0.05

^aGS containing normal placenta and normal fetus.

^bAbnormal placenta.

^cThree (of four) had abnormal placenta, and one (of four) had a normal placenta.

^dFour (of eight) had normal placenta, and four (of eight) did not have a placenta.

Fetal weight did not differ (P > 0.05) between hormone treatment groups (eCG/pLH, 0.20 ± 0.07 g; eCG/hCG, 0.13 ± 0.06 g). However, females treated with eCG/pLH had larger (P < 0.05) fetuses, based on crown-rump length, than females treated with eCG/hCG (eCG/pLH, 17.42 ± 0.51 mm; eCG/hCG, 15.27 ± 0.62 mm).

Comparison of LO AI versus LU AI

Five (of 11 pregnant) cats (45%) became pregnant from LO AI only, two from LU AI only (18%), and four from both AI sites (36%). The percentage of cats that conceived as a result of either AI technique alone did not differ (P > 0.05). However, in comparing insemination sites, more fetuses (P < 0.05) resulted from LO AI (36/49; 73%) than from LU AI (13/49; 27%). For the eCG/pLH regimen, four of eight cats (50%) became pregnant by LO AI, producing 10 fetuses attributable to semen deposited within a single oviduct, from a total of nineteen fresh CL on the ipsilateral ovary at the time of AI (or 53% fertilization) (Table 2). The number of fetuses produced was not affected (P > 0.05) by insemination side (i.e., left or right side of the reproductive tract; left side, 31/49, 63%; right side, 18/49, 37%), nor was it affected (P > 0.05) by sire (male 1, 22/49, 45%; male 2, 27/49, 55%).

Across treatments, sire did not have an effect (P > 0.05) on fetal dimensions (weight or length), nor was there any correlation (P > 0.05) between fetal dimensions (length or weight) and the number of fetuses per female (−0.11 and −0.21, respectively) or the day (20–22 days post-AI) of fetal harvest (0.28 and 0.56, respectively).

Study 2: Offspring Production after AI of Females Treated with eCG/pLH

All three females inseminated in study 2 became pregnant and carried their offspring to term (Table 5). One of the cats (female 19) was confirmed as pregnant by ultrasound examination at 20 and 40 days post-AI and appeared to be pregnant at term based on abdominal girth and mammary development. However, no kittens were observed on the expected parturition dates, and no fetuses were seen upon subsequent ultrasound examination, suggesting possible cannibalism by the dam.

Table. 5.

Offspring production following LU AI and LO AI of eCG/pLH-treated females (n = 3).

Cat	No. of CL at AI	AI outcomea	Gestation period (days)	Kitten data
				AI	Birth weight (g)	Gender	Viability status
Female 17	22	Pregnant	60	LOc	88	Female	Live
				LUc	85	Female	Lived
Female 18	20	Pregnant	63	LU	82	Female	Live
				LO	104	Male	Live
				LO	104	Male	Lived
Female 19	4	Pregnantb	63–65?	—	—	—	—

^aPregnancy was confirmed via ultrasound examination at 20 and 40 days post-AI.

^bNo kittens were observed on the expected delivery days. Subsequent ultrasound examination of this female did not reveal any fetuses, suggesting that kittens were cannibalized soon after delivery.

^cPaternity could not be confirmed.

^dKittens were live at birth but died within the first day postpartum.

Parentage verification

All queens qualified as a potential parent in all 38 systems for their respective offspring (data not shown). The two males had diagnostic differences in 23 of the 38 systems. For each offspring from a given queen, parentage qualification was determined for the two potential sires, either qualifying or excluding based upon the STRs inherited. The qualifying sire and the numbers of systems as well as the excluded sire and the number of systems are provided in Table 6. Numbers of exclusions ranged from 2 to 14 per 23 informative markers. In all cases except two, the likely sire qualified in all systems examined. Additional markers for offspring 03QOC6-E and 03QOC6-SBF were used to resolve the single system exclusion in each offspring. For each kitten, 8 of 35 and 10 of 35 systems excluded one sire over the other (1/35), and the likely sire qualified in 34 of 35 informative systems.

Table. 6.

Paternity identification results showing the number of fetuses/kittens produced by each male via AI based on sire qualification/exclusion for STR markers.

Sire	Male 1	Male 2
Study 1
Qualifieda	22	27
Excluded	27	22
Study 2
Qualified	2	3
Excludedb	4	3

^aQualified or excluded as sire. No difference (P > 0.05) was found between the number of fetuses produced by male 1 versus male 2.

^bIn the case of female 17 in study 2, one male qualified as sire of one of the two kittens, whereas the other male qualified as sire of the other kitten. However, for unclear reasons, both males were excluded as sires of those two kittens.

Discussion

In the present study, the goal was to address two aspects of AI in felids that have confounded efforts to expand its usefulness for routine propagation within population management programs. Specifically, consistent AI success has been hindered by the requirement for relatively high sperm numbers for vaginal or uterine insemination and the necessity for treating anestrual or interestrual females with exogenous gonadotropins for induction of ovarian follicular growth and ovulation. By concurrently altering the site of insemination and the specific gonadotropin used as the ovulatory signal, we have shown that high pregnancy percentages and implantation numbers may be obtained and viable offspring produced in cats following AI using low numbers of spermatozoa. These findings may lead to more efficient application of AI for reproductive management of genetically valuable domestic cats and nondomestic felid populations.

Successful AI of domestic cats was first reported more than 40 years ago [10] following the vaginal deposition of freshly collected semen into naturally estrual, hCG-treated females. Although pregnancy percentages with a single insemination were relatively high (∼50%), this approach required high sperm numbers (10–50 million sperm/AI) for consistent conception, and the need to inseminate females during a natural estrous phase restricted the capacity for the timed AI and broader application across cat populations or species. Refinement of felid AI over the past four decades has included incorporation of frozen semen for vaginal or uterine AI [11, 19, 20] and development of transcervical and laparoscopic methods for intrauterine insemination [6, 11, 21, 22]. Other changes have included the use of exogenous gonadotropin regimens for both ovarian stimulation and ovulation induction, providing investigators with greater control over ovarian activity and AI timing [6, 23]. In nondomestic cats, laparoscopic intrauterine AI of gonadotropin-treated females has become the standard method used, with pregnancies produced in multiple species, although conception rates typically have been low (<10%) [6] (for review, see [1]).

In the present study, one objective was to compare pregnancy and fertilization results following laparoscopic AI of domestic cats treated with either eCG/hCG or eCG/pLH. The standard protocol for ovarian stimulation in felids consists of an intramuscular injection of eCG followed 80–84 h later by hCG (for review, see [1]). Due to their large molecular size, both exogenous chorionic gonadotropins have long elimination half-lives and remain in circulation for several days following administration, with eCG promoting follicular growth and hCG inducing ovulation within 30–48 h of injection. At the time of ovulation, eCG concentrations are minimal, and this gonadotropin, by itself, does not appear to adversely affect fertility following AI or natural breeding [23, 24]. However, the circulatory persistence of hCG, in combination with residual eCG, does stimulate development of ancillary follicles and CL that can disrupt the postovulatory endocrine environment [8, 25]. To our knowledge, an alternative ovulatory agent, pLH, has not previously been assessed in combination with eCG for AI in cats but has been shown to induce ovulation consistently for embryo transfer procedures, without considerable formation of secondary follicles and CL [9].

In the present study, both gonadotropin regimens were found to be 100% effective in inducing ovulation (i.e., all cats had one or more fresh CL at AI), with similar numbers of fresh CL observed in both treatment groups at the time of AI. However, eCG/hCG-treated females exhibited a higher number of CLs than the eCG/pLH-treated females at 20 days post-AI, indicating formation and ovulation of a high number of ancillary follicles. One could speculate that some oocytes released from the excessive ancillary follicles in the eCG/hCG-treated group might have been fertilized, possibly explaining the higher number of gestational sacs observed in that group compared to the eCG/pLH-treated group. Following AI, pregnancy percentages in both treatment groups were relatively high (63%–75%), and average fetal weight was similar between treatment groups. Although eCG/hCG-treated females tended to produce more implantation sites than eCG/pLH-treated females, many of these fetuses were stunted in growth, and more than one-third of their implantation sites contained abnormal placentae and/or fetuses. Most (five of six) pregnant females in the eCG/hCG-treated group also had at least one implantation site without a fetus. In contrast, no implantation sites in eCG/pLH-treated females lacked a fetus, and none contained abnormal placentae or fetuses. Because pregnant cats in study 1 were spayed before parturition, we cannot compare term development between groups. The pregnant females in the eCG/hCG-treated group did have relatively high numbers of morphologically normal fetuses that presumably had the capacity to develop to term. However, the abnormal placentae and/or fetuses in the eCG/hCG-treated cats are unlikely to have produced healthy offspring. This proposed reduction in term development due to empty gestational sacs and fetal/placental abnormalities would be consistent with findings from the earlier AI study in which litter sizes in eCG/hCG-treated females averaged just two kittens despite high ovulatory responses in the inseminated females [6].

The negative impact of hCG on normal implantation and fetal development could be attributable to a combination of physiological factors. One possibility is that endocrine alterations associated with ancillary follicle and CL formation occurring over the course of several days after initial ovulation could slow normal transport of embryos through the oviduct, retarding embryo cleavage and delaying entry into the uterus for implantation [7, 25]. Alternatively, ovulation of ancillary follicles and delayed fertilization of the released oocytes could produce embryos that are capable of implantation but with compromised fetal or placental development capacity. This possibility is supported by the results of a previous study of eCG/hCG-treated females that showed the presence of CL on the ovaries at the time of follicle aspiration compromised fertilization and development of recovered oocytes in vitro, possibly due to high progesterone concentrations during folliculogenesis [26]. With either alternative, implantation numbers could be increased due to the excessive recruitment and ovulation of ancillary follicles, but only a fraction of the resulting embryos and fetuses ultimately survive.

The other primary objective of the present study was to assess the feasibility of LO AI as an alternative to LU AI. Using an LO AI approach in cats, we hypothesized that low numbers of spermatozoa could be deposited closer to the site of fertilization, potentially producing higher pregnancy percentages than typically observed with vaginal or intrauterine insemination. Typically, vaginal AI requires 50–80 × 10⁶ spermatozoa for consistent conception [10, 27], compared to 7–8 × 10⁶ spermatozoa to obtain similar fertility with intrauterine AI [6, 28]. Oviductal AI has been conducted previously in domestic cats using laparotomy to gain access to the abdominal ostium of the oviduct [14]. In that study, naturally estrual queens treated with hCG at the time of insemination produced pregnancy percentages of 25% (two of eight) and 43% (three of seven) following bilateral oviductal AI with 2 × 10⁶ and 4 × 10⁶ spermatozoa, respectively [14]. However, studies in sheep and pigs have shown considerably higher pregnancy and oocyte fertilization percentages (∼60%–90%) with oviductal AI using 1 million spermatozoa or fewer for insemination [15, 29], and in pigs, a minimally invasive LO AI method has been developed as an alternative to laparotomy [15, 30]. A laparoscopic approach, though less traumatic than laparotomy, still allows direct visualization of the ovarian response to gonadotropin treatment and assessment of any reproductive pathology [31].

The present findings with domestic cats support the superiority of oviductal AI over uterine AI for insemination with low sperm numbers. Using a novel study design, each female cat was inseminated in one oviduct and one contralateral uterine horn with semen from two different males, and paternity analysis of resulting fetuses was used to assess fertilization success for each insemination site within each individual female. In study 1, nearly 70% of females became pregnant, with fetuses produced by LO AI in 9 of 16 females (56%) and by LU AI in 6 of 16 females (38%). Most strikingly, the results indicated that nearly 74% of all fetuses were produced by LO AI, despite the use of a deep uterine AI approach in the contralateral uterus to place highly concentrated, highly motile spermatozoa close to the UTJ. The UTJ in cats, as in other species, creates a formidable barrier to sperm penetration, with only a small fraction of total sperm numbers ejaculated into the vagina following natural breeding ever entering the oviduct [32]. With laparoscopic AI, anesthesia likely further compromises sperm transport through the UTJ by suppressing behavioral and physiological after-reactions associated with natural breeding (for review, see [1]). In the present study, high fertilization percentages were obtained by inseminating females within 1–2 h of ovulation and depositing the spermatozoa close to the oviductal ampulla (i.e., at the structural bend in the feline oviduct), albeit with oviductal sperm numbers that likely were far in excess of that seen with natural breeding.

Based on paternity testing of the 49 fetuses produced in the present study, identification of the likely sire, and hence of the insemination site that resulted in conception, for each fetus was unequivocally determined. Likewise, the alternative potential sire was excluded in multiple informative systems. Theoretically, a slight possibility existed that spermatozoa deposited near the UTJ of one uterine horn could traverse the entire length of both horns to fertilize oocytes in the opposite oviduct [33]. We cannot exclude that possibility completely. However, the deposition of low sperm numbers in a very small semen volume (∼5 μl) in the distal uterine horn and the overall low percentage of fetuses that resulted from uterine AI in the present study suggest that any such occurrence was negligible.

The overall pregnancy percentages (∼70%) in the present study were considerably higher than those reported previously using similar sperm numbers (∼2 million sperm/AI) for uterine AI and oviductal AI via laparotomy (12.5%–25% pregnancy success) [14, 28]. Using higher sperm numbers (∼7–8 million motile sperm/AI) for uterine AI, others have reported similar pregnancy percentages (50%–80%) via laparoscopy [6] or laparotomy [28]. Differences in pregnancy outcomes between studies may be partially attributable to other factors, such as timing of AI relative to ovulation and anesthesia, but the present results suggest that the use of LO AI was primarily responsible for the improved pregnancy percentages. Treatment with eCG/pLH as an alternative to eCG/hCG provided no advantage in terms of conception or number of implantations; in fact, it might represent a disadvantage in terms of total number of normal fetuses produced if most of those fetuses would be expected to continue development to produce healthy, live kittens. However, because pregnancies were not allowed to go to term in study 1, it is unknown whether all morphologically normal fetuses observed at approximately 21 days post-AI truly were developmentally competent. Our results did indicate that treatment with eCG/pLH appeared to produce implantations containing healthy placentae and fetuses more consistently than seen with eCG/hCG treatment. For eCG/pLH-treated cats, 50% (four of eight) became pregnant by LO AI of a single oviduct in each female, and based on the number of primary CL on the ipsilateral ovaries, more than 50% of ovulated oocytes were fertilized and implanted to form an average of 2.5 healthy fetuses per female. With bilateral oviductal insemination after eCG/pLH treatment and similar fertility, the production of four to five healthy fetuses per female may be expected. In study 1, the capacity of fetuses to develop to term in eCG/pLH-treated females was not assessed, but we did investigate this issue in a subsequent experiment.

In study 2, all three females ovulated following eCG/pLH treatment and conceived after AI. Including pregnancy results from study 1, approximately 74% of cats (14/19) became pregnant after LO AI and LU AI. Two (of three) females produced at least one viable kitten after a normal gestation length, whereas the third pregnant female possibly cannibalized her offspring. Although litter sizes and offspring viability were suboptimal with these initial term pregnancies, these births have demonstrated the capacity to produce viable offspring following eCG/pLH treatment and AI. In our subsequent studies (unpublished results), average litter sizes of approximately five kittens have been observed.

The cause of the uncertainty in determining the sire for two term kittens from one of the cats (female 17) is unknown. One marker (of 23) for each kitten excluded the likely sire and qualified the alternative sire. The possibility of PCR contamination or the mixing of samples for the marker was assessed by repeating the analyses multiple times with additional samples that had been collected from the kittens; consistent results were observed. Twelve additional informative markers further supported the likely sire, qualifying in 97% of systems with no further exclusions. A likely explanation for the observed results is a spontaneous mutation in the dam; thus, her type would not match her contribution to the offspring. Unfortunately, the dam was unavailable for resampling, and this explanation remains speculative.

In conclusion, the present study offers a novel approach to AI in felids by using laparoscopy to deposit semen into the oviduct. Our results have demonstrated that LO AI with low sperm numbers can produce high pregnancy percentages in gonadotropin-treated cats and that AI following eCG/pLH treatment, as an alternative to eCG/hCG treatment, produces normal implantations and healthy fetuses that are capable of developing to term. These findings indicate that oviductal insemination of eCG/pLH-treated females might be preferable to uterine AI in cats, especially if sperm numbers are low or possibly if sperm quality is compromised by factors such as teratospermia or cryopreservation-related damage [34]. This LO AI approach also may have value for reproduction and conservation management of nondomestic felid species. Following completion of the present study, we have had successes (unpublished results) in producing healthy offspring in the ocelot (Leopardus pardalis), Pallas cat (Otocolobus manul), and tiger (Panthera tigris) after LO AI of eCG/pLH-treated females.