**Semen Modulates Inflammation and Angiogenesis in the Reproductive Tract of Female Rabbits**

### **Jaume Gardela 1,2, Amaia Jauregi-Miguel 3, Cristina A. Martinez 1, Heriberto Rodríguez-Martinez 1, Manel López-Béjar 2,4 and Manuel Álvarez-Rodríguez 1,2,\***


Received: 19 October 2020; Accepted: 18 November 2020; Published: 25 November 2020

**Simple Summary:** In mammals, the expression of regulatory genes is modified by the interaction between semen and the female reproductive tract. This study intends to unveil how mating or insemination with sperm-free seminal plasma, as well as the presence of preimplantation embryos, affects inflammation and angiogenesis in different segments of the reproductive tract of female rabbits. Gene expression of anti-inflammatory cytokines and angiogenesis mediators was analyzed in segmented tracts (cervix to infundibulum) in response to mating and sperm-free seminal plasma infusion. Moreover, the gene expression at different times post-mating was also analyzed. Results showed that gene expression changes were mainly localized in the uterus in the natural mating group, describing a clear temporal variation, while limited to the oviduct in the sperm-free seminal plasma group. These changes suggest an early response in the uterus and late modulation in the oviduct, distinctly demonstrating that semen and seminal plasma, through their interaction with the female reproductive tract, can differentially modulate the expression of anti-inflammatory and angiogenesis mediators.

**Abstract:** The maternal environment modulates immune responses to facilitate embryo development and ensure pregnancy. Unraveling this modulation could improve the livestock breeding systems. Here it is hypothesized that the exposure of the female rabbit reproductive tract to semen, as well as to early embryos, modulates inflammation and angiogenesis among different tissue segments. qPCR analysis of the gene expression changes of the anti-inflammatory interleukin-10 (IL10) and transforming growth factor beta family (TGFβ1–3) and the angiogenesis mediator vascular endothelial growth factor (VEGF-A) were examined in response to mating or insemination with sperm-free seminal plasma (SP). Reproductive tract segment (cervix to infundibulum) samples were obtained in Experiment 1, 20 h after gonadotropin-releasing hormone (GnRH) stimulation (control), natural mating (NM) or vaginal infusion with sperm-free SP (SP-AI). Additionally, segmented samples were also obtained at 10, 24, 36, 68 or 72 h after GnRH-stimulation and natural mating (Experiment 2). The results of gene expression, analyzed by quantitative PCR, showed that NM effects were mainly localized in the uterine tissues, depicting clear temporal variation, while SP-AI effects were restricted to the oviduct. Changes in anti-inflammatory and angiogenesis mediators indicate an early response in the uterus and a late modulation in the oviduct either induced by semen or preimplantation embryos. This knowledge could be used in the implementation of physiological strategies in breeding systems to face the new challenges on rabbit productivity and sustainability.

**Keywords:** gene expression; endometrium; oviduct; spermatozoa; seminal plasma; inflammation; angiogenesis; rabbit

#### **1. Introduction**

Considered either laboratory animals, pets, invasive species or livestock, the rabbit (*Oryctolagus cuniculus*) is one of the most versatile and multipurpose animal species bred by humans [1–3]. Subjected today to intense production for meat and fur, does are affected by requirements on fertility and lifespan [4]. Contrary to other livestock species, rabbits require the generation of genital-somatosensory signals during mating to generate the preovulatory peak of gonadotropin-releasing hormone (GnRH) [5,6]. The consequent release of luteinizing hormone from the anterior pituitary induces ovulation [5]. Despite requiring hormonal stimulation if artificial insemination (AI) is used [7], this technique is extensively employed, giving similar or better results than natural mating [8] when using fresh or cooled semen [9]. The success of AI with diluted semen in several species suggest that seminal plasma (SP) components, excluding spermatozoa, are not required for pregnancy [10]. However, in some species, such as rodents or pigs, reproductive success and pregnancy can be jeopardized when the SP signaling is disrupted [10]. Several studies demonstrated that SP has multiple effects on the female reproductive tract essential for conceptus and pregnancy, improving reproductive performance [10].

In different species of mammals, the direct interaction between the female reproductive tract and semen, as well as the sensorial stimulation during the act of mating [11], induce molecular and cellular modifications into the female reproductive tract [12–14], modulating the immune system and optimizing the reproductive outcomes [15]. The immune response generated as a result of mating leads to a suitable environment for embryo survival, implantation success, optimal fetal and placental development and the overall reproductive process [16].

The first description of female post-mating inflammation responses was made by observing rabbits in 1952 [17]. The observations of McDonald and collaborators showed a leucocytic influx in the uterus in response to SP but not by sterile spermatozoa administration [17]. Later studies demonstrated that both SP and spermatozoa triggered a uterine leucocytic response in rabbits [18,19]. Contrary to the vagina, cervix and uterus, the oviduct seems to respond to insemination without a reduction of the spermatozoa population at the time of ovulation [20], providing the optimal conditions for spermatozoa survival until fertilization takes place [7]. The interaction between spermatozoa and oviductal epithelial cells could regulate the immunological environment of the oviduct-inducing anti-inflammatory cytokines [21], like transforming growth factor beta 1 (TGFβ1) or interleukin-10 (IL10), protecting spermatozoa from immune responses [13].

The transforming growth factor beta (TGFβ) family regulates many cellular activities, including cell growth, proliferation, differentiation, tissue remodeling, extracellular matrix formation, control of cell surface molecules, immunoregulation, angiogenesis and apoptosis [22]. The TGFβ belong to a large superfamily of cytokines composed of three 25 kDa homodimeric proteins (TGFβ1–3) whose biological effects are mediated by three transforming growth factor beta receptors (TGFβR1–3) [22,23]. The IL10, first described as cytokine synthesis inhibitory factor or CSIF [24], is a cytokine that inhibits inflammatory responses preventing the secretion of inflammatory cytokines and regulating the differentiation and proliferation of immune cells [25].

The vascular endothelial growth factor (VEGF, also referred to as VEGF-A), a heparin-binding homodimeric glycoprotein, is a mitogen for endothelial cells and a potent inducer of angiogenesis [26] that also promotes vascular permeability [27]. VEGF protein family includes other factors like VEGF-B, VEGF-C, VEGF-D, VEGF-E, VEGF-F and placental growth factor (P1GF) [28,29]. VEGF-A is expressed in the endometrium of several species [30–32], being proposed in the rabbit as a local signal between the implantation embryo and vascular structures in the receptive endometrium [33].

This study aimed to determine whether natural mating and/or sperm-free SP infusion modulates the expression of anti-inflammatory cytokines and angiogenesis mediator genes (*IL10*, *TGF*β*1–3* and *VEGF-A*) in the maternal environment of the doe at 20 h post-exposure. Moreover, we determine the expression of these mediators when early embryo development occurs along the reproductive tract of the doe up to 72 h post-mating. Additionally, we analyzed the spatial changes of these genes along the tubular reproductive tract of the doe.

#### **2. Materials and Methods**

#### *2.1. Ethics Statement*

Rabbits were handled according to the Spanish Royal Decree 1201/2005 (BOE, 2005: 252:34367-91) and the Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes (2010; 276:33–79). The Committee of Ethics and Animal Welfare of the Universitat Autónoma de Barcelona, Spain, approved this study (Expedient #517).

#### *2.2. Animals and Experimental Design*

Animals were housed in the nucleus colony at the farm of the Institut de Recerca i Tecnologia Agroalimentaries (IRTA-Torre Marimon, Caldes de Montbui, Barcelona, Spain) under a controlled photoperiod and temperature [34]. Single cage equipped with plastic footrests, a feeder (restricted to 180 g/day of an all-mash pellet) and nipple-drinkers (ad libitum access to water) were used for each rabbit.

Six New Zealand White (NZW) adult bucks (from seven to thirteen months old) were included in the study. At 4.5 months of age, the bucks were started to be trained with a homemade polyvinyl chloride artificial vagina at 50 ◦C to ensure 41–42 ◦C at the time of semen collection. One ejaculate was collected per male, discarding ejaculates that contained urine and calcium carbonate deposits on visual inspection.

Two separate experiments were performed in this study (Figure 1). In Experiment 1, 9 NZW adult does were randomly allocated into three experimental groups. Sequential segments of the right side of female reproductive tracts were retrieved after 20 h post-induction of the ovulation with 0.03 mg GnRH (Fertagyl®, Esteve Veterinaria, Barcelona, Spain) intramuscularly (im) (control of the ovulation; control, *n* = 3), 20 h post-induction of the ovulation with 0.03 mg GnRH im and sperm-free SP vaginal infusion (SP-AI, *n* = 3) and 20 h post-induction of the ovulation with 0.03 mg GnRH im and natural mating (NM, *n* = 3). In Experiment 2, 15 NZW adult does were sequentially euthanized at 10, 24, 36, 68 or 72 h post-induction of the ovulation with 0.03 mg GnRH im and natural mating (*n* = 3/collection time). The 10 h group was established as the reference group to compare the ovulatory moment (10 h post-mating) with some different time-point steps of the preimplantation embryo development: 2–4 cell embryo (24 h), 8-cell embryo (36 h), early morula (68 h) and morula (72 h).

**Figure 1.** Representation of the experimental design and tissue sections obtained from does. Sequential tissue segments derived from does were endocervix, distal uterus, proximal uterus, utero-tubal junction, distal isthmus, ampulla and infundibulum. Intramuscular injection of 0.03 mg gonadotropin-releasing hormone (GnRH) was used to induce ovulation in all groups of both experiments.

#### *2.3. Mating and Semen Collection*

Two randomly selected bucks were used to mate the does included in the mating group of Experiment 1 and Experiment 2 after the hormonal induction of the ovulation with 0.03 mg GnRH im. Additionally, semen was collected from the same rabbit bucks through an artificial vagina, as described above. The sperm-free SP was obtained after centrifugation at 2000× *g* for 10 min and checked for the absence of spermatozoa. The harvested sperm-free SP was immediately used for sperm-free SP vaginal infusions of Experiment 1.

#### *2.4. Tissue Sample Collection*

For each experimental condition, seven consecutive tissue sections from the female reproductive tracts (endocervix, distal uterus, proximal uterus, utero-tubal junction, ampulla, isthmus and infundibulum; Figure 1) were obtained after the euthanasia of the does [34]. The tissue segments were retrieved and stored in RNAlater solution (Thermo Fisher Scientific, Waltham, MA, USA) at −80 ◦C.

Before tissue segmentation of does included in experiment 2, the oviduct was isolated and flushed (phosphate-buffer saline supplemented with 5% fetal calf serum and 1% antibiotic-antimycotic solution) to collect the embryos, which were examined by number and developmental stage. The number of ovulated ovarian follicles (5.08 ± 2.06 follicles, mean ± standard deviation (SD)) and embryos (4.67 ± 3.14, mean ± SD) were counted for each side of the reproductive organs, as previously published elsewhere [34]. Briefly, at 24 h (2 and 4-cell), 36 h (8-cell stage), 68 h (early morula) and 72 h (morula), the embryos were collected by the method described above.

#### *2.5. Quantitative PCR Analyses*

Briefly, total RNA was extracted from the tissue segments following a TRIzol-based protocol [34]. The RNA concentration of the extracts was determined from the absorbance of 260 nm with Thermo Scientific NanoDropTM 2000 (Fisher Scientific, Gothenburg, Sweden). The quality of the RNA was determined by the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA), using the samples with an RNA integrity number higher than 8. High-Capacity RNA-to-cDNA™ Kit (Applied Biosystems™, Foster City, CA, USA) was used to synthesize the first-strand cDNA for quantitative polymerase chain reaction (qPCR) analyses (CFX96™; Bio-Rad Laboratories, Inc; Hercules, CA, USA). Following our previous qPCR protocol [34], the gene relative expression levels were quantified using the Pfaffl method [35]. The primer sequences, product sizes and efficiencies are shown in Table 1. For the β*-ACTIN* gene, commercial gene-specific PCR primers for rabbit samples were used (PrimePCR™SYBR® Green Assay: ACTB, Rabbit; Bio-Rad Laboratories, Inc., Hercules, CA, USA). Product sizes were confirmed by loading the amplicons in an agarose gel using a gel imaging system (ChemiDoc XRS+ System, BioRad Laboratories, Inc., Hercules, CA, USA).


**Table 1.** Primers used for the quantitative PCR analyses.

β*-ACTIN*: β-actin, *IL10*: interleukin 10, *TGF*β*1*: transforming growth factor β 1, *TGF*β*2:* transforming growth factor β 2, *TGF*β*3*: transforming growth factor β 3, *VEGF-A*: vascular endothelial growth factor A, F: forward, R: reverse, A: adenine, C: cytosine, G: guanine, T: thymine, bp: base pair.

#### *2.6. Statistical Analyses*

CFX Maestro™ 1.1 software version 4.1.2433.1219 (Bio-Rad Laboratories, Inc; Hercules, CA, USA) was used to export all data. Normal distribution and homoscedasticity of the data were analyzed using the Shapiro–Wilk normality test and Levene's test. Non-normal data distribution was restored using Log(x) transformation prior to analysis. R version 3.6.1. [36] was used to conduct the statistical analyses with *nlme* [37] to perform linear mixed-effects (LME) models and *multcomp* [38] to perform pairwise comparisons adjusted by Tukey's test. The threshold for significance was set at *p* < 0.05. Data are presented as median (minimum, maximum) unless otherwise stated.

A first LME model included the treatments of Experiment 1 (control, SP-AI, NM) as fixed effects and the females as the random part of the model. Pairwise comparisons were adjusted by Tukey's test. A second LME model included the different collection times of Experiment 2 (10, 24, 36, 68 and 72 h post-mating) as fixed effects and the females as the random part of the model. Post hoc comparisons were performed using Tukey's multiple comparisons test.

The differential expression changes in qPCR results among tissues in both Experiments were re-analyzed separately. The utero-tubal junction was used as an arbitrary anatomical compartment reference among all tissues examined to compare the gene expression changes per gene [34], issued both by control, NM or SP-AI (Experiment 1) or by different times post-mating (Experiment 2). The LME model was followed by Tukey's multiple comparison test to analyze the differences among each anatomical region of the female reproductive tract. Data are presented as median (minimum, maximum) unless otherwise stated. Data on the differential expression among tissues are presented as Row Z-scores.

#### **3. Results**

#### *3.1. Natural Mating and Sperm-Free Seminal Plasma Infusion Di*ff*erentially Modulated Anti-Inflammatory Cytokines and VEGF-A Genes*

Differences in *IL10*, *TGF*β*1*, *TGF*β*2*, *TGF*β*3* and *VEGF-A* expression among groups included in Experiment 1 are displayed in Figure 2 (and Figure S1). The NM treatment downregulated the *IL10* expression in the utero-tubal junction (*p* < 0.05); upregulated the *TGF*β*1* and *TGF*β*3* expression in the distal (*p* < 0.01) and proximal uterus (*p* < 0.001); upregulated the *TGF*β*1* and *TGF*β*2* expression in the infundibulum (*p* < 0.01); and upregulated the *VEGF-A* expression in the distal uterus (*p* < 0.05). The SP-AI treatment upregulated the *TGF*β*1* expression in the proximal uterus (*p* < 0.01), upregulated the *TGF*β*1* and *TGF*β*3* expression in the infundibulum (*p* < 0.01), and upregulated the *VEFG-A* expression in the infundibulum (*p* < 0.01).

**Figure 2.** Results of Experiment 1. Gene expression of (**a**) *IL10*, (**b**) *TGF*β*1*, (**c**) *TGF*β*2*, (**d**) *TGF*β*3* and (**e**) *VEGF-A* in differentially expressed segments (*p*<0.05) (DistUt: distal uterus, ProxUt: proximal uterus, UTJ: utero-tubal junction, Amp: ampulla and Inf: infundibulum) of the does internal reproductive tract after 20 h of the induction of the ovulation with 0.03 mg gonadotropin-releasing hormone (GnRH) intramuscularly (20 h\_C, *n* = 3), 20 h post-GnRH-stimulation and seminal plasma vaginal infusion (20h\_SP-AI, *n* = 3) and 20 h post-GnRH-stimulation and natural mating (20 h\_NM, *n* = 3). Fold changes relative to the control of the ovulation group are shown. Different letters (a,b) represent statistical differences between groups (*p* < 0.05). Median (minimum, maximum).

#### *3.2. Temporal Gene Expression of Anti-Inflammatory cytokines and VEGF-A at 10 h up to 72 h Post-Mating*

Differences in *IL10*, *TGF*β*1*, *TGF*β*2*, *TGF*β*3* and *VEGF-A* expression in does sampled at different times post-mating are displayed in Figure 3 (and Figure S2), with 10 h post-mating as the control group. The *IL10* expression was downregulated at 36, 68 and 72 h post-mating in the endocervix and distal uterus (*p* < 0.001); downregulated at 24, 36, 68 and 72 h post-mating in the utero-tubal junction and infundibulum (*p* < 0.001); and upregulated at 68 and 72 h post-mating in the distal isthmus (*p* < 0.05). The *TGF*β*1* expression was downregulated at 36, 68 and 72 h post-mating in the distal uterus (*p* < 0.05); and downregulated at 24, 36, 68 and 72 h post-mating in the infundibulum (*p* < 0.001). The *TGF*β*2* expression was downregulated at 24, 36, 68 and 72 h post-mating in the distal uterus and infundibulum (*p* < 0.001); downregulated at 24 h post-mating in the proximal uterus (*p* < 0.01); and downregulated at 24 and 72 h post-mating in the utero-tubal junction (*p* < 0.05). The *TGF*β*3* expression was downregulated

at 36 h post-mating in the endocervix (*p* < 0.05); downregulated at 36, 68 and 72 h post-mating in the distal isthmus (*p* < 0.05); and downregulated at 24, 36, 68 and 72 h post-mating in the infundibulum (*p* < 0.001). The *VEGF-A* expression was upregulated at 24 h post-mating in the utero-tubal junction (*p* < 0.05); downregulated at 24 and 72 h post-mating in the ampulla (*p* < 0.05); and downregulated at 24, 36, 68 and 72 h post-mating in the infundibulum (*p* < 0.001).

**Figure 3.** Results of Experiment 2. Gene expression of (**a**) *IL10*, (**b**) *TGF*β*1*, (**c**) *TGF*β*2*, (**d**) *TGF*β*3* and (**e**) *VEGF-A* was statistically significant in rabbit endocervix (Cvx), distal uterus (DistUt), proximal uterus (ProxUt), utero-tubal junction (UTJ), distal isthmus (Isth), ampulla (Amp) and infundibulum (Inf) after 10, 24, 36, 68 or 72 h post-induction of the ovulation (intramuscular injection of 0.03 mg gonadotropin-releasing hormone) and natural mating (*n* = 3/collection time). Fold changes relative to 10 h post-mating group are shown. Different letters (a–c) represent statistical differences between groups (*p* < 0.05). Median (minimum, maximum).

#### *3.3. Spatial Gene Expression Triggered by Natural Mating and Sperm-Free Seminal Plasma Infusion*

The tissue analysis expression showed significant differences in *IL10*, *TGF*β*1*–3 and *VEGF-A* expression among tissues in the different groups included in Experiment 1 (Figure 4a and Figure S3). The infundibulum presented the highest *IL10* expression in the SP-AI and control groups (*p* < 0.05), whereas the cervix presented the highest *IL10* expression in the NM group (*p* < 0.05). In the case of *TGF*β*1* expression, the endocervix presented the highest expression in the NM group (*p* < 0.05). The distal isthmus presented the highest *TGF*β*2* expression in all groups included in Experiment 1

(*p* < 0.05). Similarly, the oviductal tissues presented the highest *TGF*β*3* expression in the SP-AI and control groups (*p* < 0.05). However, the proximal uterus presented the highest *TGF*β*3* expression in the NM group (*p* < 0.01). The oviductal tissues presented the highest *VEGF-A* expression in the SP-AI and control groups (*p* < 0.05), whereas the uterine tissues presented the highest *VEGF-A* expression in the NM group (*p* < 0.05).

**Figure 4.** Heatmap composition of the gene expression changes among tissues in the different groups included in the study. Changes in *IL10*, *TGF*β*1*, *TGF*β*2*, *TFG*β*3* and *VEGF-A* expression among tissues (endocervix, Cvx; distal uterus, DistUt; proximal uterus, ProxUt; utero-tubal junction, UTJ; distal isthmus, Isth; ampulla, Amp; and infundibulum, Inf) (**a**) at 20 h after the induction of ovulation with 0.03 mg of gonadotropin-releasing hormone (GnRH) intramuscularly, as the control of ovulation (20 h\_C); 20 h post-GnRH stimulation and sperm-free seminal plasma vaginal infusion (20 h\_SP-AI); and 20 h post-GnRH stimulation and natural mating (20 h\_NM); (**b**) at 10, 24, 36, 68 and 72 h post-GnRH-stimulation and natural mating. Row Z-scores of the mean fold change relative to the reference group (UTJ) are shown. Red indicates upregulation, and green indicates downregulation.

#### *3.4. Spatial Gene Expression at 10 h up to 72 h Post-Mating*

The tissue analysis expression showed significant differences in *IL10*, *TGF*β*1–3* and *VEGF-A* expression among tissues in the different groups included in Experiment 2 (Figure 4b and Figure S4). At 10 h post-mating (time of ovulation), the infundibulum presented the highest *IL10* expression

(*p* < 0.01). Conversely, at 24 h post-mating, the endocervix presented the highest *IL10* expression (*p* < 0.05). At 36, 68 and 72 h post-mating, the oviductal segments presented higher *IL10* expression compared to the uterine tissues (*p* < 0.05).

At 10 h post-mating, the infundibulum presented the highest *TGF*β*1–3* expressions (*p* < 0.05). The *TGF*β*1* expression was higher in the uterine tissues compared to the oviduct segments (*p* < 0.05), whereas the *TGF*β*2* expression was higher in the oviduct segments compared to the uterine tissues (*p* < 0.05). Both uterine and oviductal segments presented the highest *TGF*β*3* expressions (*p* < 0.05). The uterine tissues presented higher *VEGF-A* expressions (*p* < 0.05) compared to oviduct segments. However, at 10 and 36 h post-mating, the infundibulum presents the highest *VEGF-A* expression (*p* < 0.05).

#### **4. Discussion**

The results showed a differential expression triggered by sperm-free SP infusion and natural mating on anti-inflammatory cytokines and angiogenesis-related genes, as well as the differential expression produced at different times post-mating in the rabbit, an induced ovulatory species. Additionally, the study provides the differential expression on anti-inflammatory cytokines and angiogenesis-related genes along the doe reproductive tract triggered by natural mating and sperm-free SP infusion, as well as by preimplantation embryo at different times post-mating.

The SP interacts with the female reproductive tract-inducing female immune adaptation processes required to tolerate male antigens suppressing inflammation and immune rejection responses [10]. This maternal immune tolerance is essential for ongoing pregnancy success [39]. Additionally, the SP induces molecular and cellular modifications in the uterus and in the highest reproductive tract promoting embryo development and implantation competence [10]. In this regard, different soluble components present in the SP play key roles in these processes, such as the TGFβs, which can drive immune cells to tolerogenic phenotypes [40]. This study revealed differences between sperm-free SP and natural mating (that also contains SP) in the *IL10*, *TGF*β*1*, *TGFB3* and *VEGF-A* expression. The data suggest that natural mating, at 20 h post-exposure, downregulated the *IL10* expression in the utero-tubal junction, probably related to the spermatozoa selection or clearance in this region. Conversely, the *TGF*β*1*, *TGF*β*3* and *VEGF-A* expressions were upregulated in the uterine tissues in the natural mating group, whereas upregulated in the infundibulum in the sperm-free SP infusion group.

Even though the SP appears to not reach the upper segments of the oviduct after mating, it seems that the signaling originated in lower segments travel to upper segments in response to SP [10]. The range of effects produced by SP in different species includes clearance of microorganisms in the uterus after mating, sperm selection, induction of ovulation, the formation of corpus luteum and supporting the development of the preimplantation embryo, among others [10]. Our observations indicate that the sperm-free SP infusion increased the expression of *TGF*β*1*, *TGF*β*3* and *VEGF-A* in the infundibulum, whereas the expression of these genes in the natural mating group, also containing SP, remained unaltered compared to the control group. Moreover, the natural mating group increased *VEGF-A* expression in the distal uterus, but not in the oviduct, suggesting that the rabbit uterus reacts to post-mating inflammation [15], likely to process seminal material and recovery for uterine tissue homeostasis after mating. Our data suggest that the responsiveness of the doe reproductive tract differs between the sperm-free SP and sperm-containing SP treatments. The signaling originated by the sperm-free SP infusion was able to travel to upper segments of the oviduct, as previously demonstrated in several species [10], whereas the signal generated by the natural mating group remained in the uterine tissues. This pattern was also observed in the temporal tracking experiment of the study, suggesting that the rabbit uterus may react to the early stages of embryo development before its arrival from the oviduct, but the mechanism involved is yet not fully understood.

The TGFβs isoforms present a diverse range of functions in reproductive tissues, including gonad development, gamete production, embryo implantation, fetal and placental development, development of secondary sex organs and immune tolerance of gametes and conceptus antigens [22]. The effects of TGFβs are mediated by antigen-presenting cells such as macrophages and dendritic cells, creating and maintaining an immunotolerant environment for the conceptus antigens if an anti-inflammatory dominance is established [22]. In this regard, our results revealed an upregulation of *TGF*β*s* expression at 20 h post-mating in the endometrium, suggesting that mating induces suppression of anti-sperm immune response [39]. Similar results were found at 10 h post-mating in oviductal segments, probably related to the ovulation that takes place 10 h after coitus or GnRH stimulation [7] or the suppression of anti-sperm immune response induced by the transport of spermatozoa through the oviduct before fertilization [41,42]. However, the upregulation of *TGF*β*s* expression in the oviduct was not present in the rest of the times after mating compared to 10 h post-mating, when the early embryo stages of development are present along the oviduct [43]. Thus, the effect produced by semen on the *TGF*β*s* expression changes seemed to be stronger than the effect of the embryo transport through the oviduct at different stages of development, presumably due to the different number of cells in contact with the female reproductive tract (300 million spermatozoa [44] vs. 4–12 embryos [45]). Although the effects of spermatozoa and the early embryo stages of development on *TGF*β*s* expression seems to differ in magnitude, tissue analyses revealed a *TGF*β*2* and *TGF*β*3* expression pattern in the oviduct, increasing the expression of those genes. Moreover, the *TGF*β*1* expression pattern in the uterus in the presence of spermatozoa and the early stages of embryo development may indicate that both antigenic compounds induced an anti-inflammatory response creating and maintaining an immunotolerance environment in the oviduct and uterus, essential for the pregnancy success [39]. Similarly, this immunotolerant environment to spermatozoa may be observed by the upregulation of the *IL10* expression in the endocervix and distal uterus at 10 and 24 h post-mating.

The tissue and temporal analyses revealed an interesting pattern along the oviduct during the preimplantation embryo transport. When fertilization occurs in the ampulla [46], embryos are delayed at the ampulla–isthmus junction until 48 h post-mating and then pass through the isthmus portion after 70 h post-mating to finally reach the utero-tubal junction and enter the uterus [43]. Our observations showed an upregulation of *TGF*β*s* expression in the infundibulum at the time of induced ovulation (10 h post-mating), similarly, as the pattern observed in the *IL10* expression at the same time, might relate with the inflammation process associated with ovulation [47]. At 24 h post-mating, the endocervix displayed the highest expression of *IL10, TGF*β*1, TGF*β*3* and *VEGF-A*, whereas the *TGF*β*2* expression was higher in the isthmus compared to the endocervix. Although fertilization takes place at 12–13 h post-mating [46], this upregulation of anti-inflammatory cytokines and angiogenesis mediators in the endocervix may be related to preserving spermatozoa viability and fertilization competence [13], in accordance with previous research that proposed the cervix as the first anatomical barrier for spermatozoa reservoir in rabbits [20,48,49].

The coordination between the preimplantation embryos and the differentiation of the uterus to a receptive state must be closely synchronized for successful embryo implantation [50]. Following the stromal and epithelial cell changes, the endometrial vascular bed undergoes a marked expansion before implantation [50,51]. This vascular expansion, in size and number, is markedly enhanced during implantation and placental structure development [51]. In the rabbit, the vascular supply during pregnancy is produced by the addition of new vessels and the expansion of the existing ones [52]. Our results revealed that *VEGF-A* expression was upregulated at 10 h post-mating in the infundibulum and ampulla, probably related to the ongoing processes like ovulation [7] and the imminent fertilization [46]. The *VEGF-A* expression displayed higher relative expression values at 20 h post-mating in the endometrium compared to the oviduct, perhaps related to preparing the receptive endometrium for the embryo implantation. However, this effect was not observable at different times post-mating, suggesting that additional factors, other than the analyzed in our experimental setup, are involved in the differentiation of the uterus to a receptive state before embryos reach the uterus.

#### **5. Conclusions**

Our results confirmed a differential modulation of the expression of anti-inflammatory and angiogenesis-related-genes by sperm-free SP and natural mating. Moreover, the gene expression analysis up to 72 h after induction of the ovulation might suggest a concerted effect of the presence of sperm and preimplantation embryos in the differential expression of these genes along the reproductive tract of the doe. Our findings warrant further research both to discriminate the effects of mating, spermatozoa, SP and preimplantation embryos on the reproductive tract of the doe and to fully describe the mechanism involved in such differential gene expression. Thus, mimicking the gene modulations induced by natural mating and sperm-free SP using physiological strategies in the current rabbit breeding systems may enhance the efficiency, productivity and sustainability of such systems.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-2615/10/12/2207/s1, Figure S1: Changes in (a) *IL10*, (b) *TGF*β*1*, (c) *TGF*β*2*, (d) *TGF*β*1* and (e) *VEGF-A* expression among different treatments at 20 h post-treatment: 20 h post-induction of the ovulation, control (20 h\_C); 20 h post-seminal plasma infusion, 20 h\_SP; and 20 h post-natural mating, 20 h\_NM. Tissue anatomical regions of the rabbit female reproductive tract (endocervix, Cvx; distal uterus, DistUt; proximal uterus, ProxUt; utero-tubal junction, UTJ; distal isthmus, Isth; ampulla, Amp; and infundibulum, Inf).-fold changes relative to the reference group (20 h\_C) are shown. Different letters (a–c) represent statistical differences between tissues (*p* < 0.05). Median (minimum, maximum); Figure S2: Changes in (a) *IL10*, (b) *TGF*β*1*, (c) *TGF*β*2*, (d) *TGF*β*1* and (e) *VEGF-A* expression at different times (from 10 to 72 h) post-mating: 10, 24, 36, 68 and 72 h post-natural mating. Tissue anatomical regions of the female rabbit reproductive tract (endocervix, Cvx; distal uterus, DistUt; proximal uterus, ProxUt; utero-tubal junction, UTJ; distal isthmus, Isth; ampulla, Amp; and infundibulum, Inf).-fold changes relative to the reference group (10 h post-mating) are shown. Different letters (a–c) represent statistical differences between tissues (*p* < 0.05). Median [minimum, maximum]; Figure S3: Changes in (a) *IL10*, (b) *TGF*β*1*, (c) *TGF*β*2*, (d) *TGF*β*1* and (e) *VEGF-A* expression among different tissues at 20 h post-treatment. 20 h post-induction of the ovulation, control (20 h\_C); 20 h post-seminal plasma infusion, 20 h\_SP; and 20 h post-natural mating, 20 h\_NM. Tissue anatomical regions of the female rabbit reproductive tract (endocervix, Cvx; distal uterus, DistUt; proximal uterus, ProxUt; utero-tubal junction, UTJ; distal isthmus, Isth; ampulla, Amp; and infundibulum, Inf).-fold changes relative to the reference group (UTJ) are shown. Different letters (a–d) represent statistical differences between tissues (*p* < 0.05). Median [minimum, maximum]; Figure S4: Changes in (a) *IL10*, (b) *TGF*β*1*, (c) *TGF*β*2*, (d) *TGF*β*1* and (e) *VEGF-A* expression among different tissues in the period 10–72 h (10, 24, 36, 68 and 72 h post-natural mating). Tissue anatomical regions of the female rabbit reproductive tract (endocervix, Cvx; distal uterus, DistUt; proximal uterus, ProxUt; utero-tubal junction, UTJ; distal isthmus, Isth; ampulla, Amp; and infundibulum, Inf).-fold changes relative to the reference group (UTJ) are shown. Different letters (a–d) represent statistical differences between tissues (*p* < 0.05). Median [minimum, maximum].

**Author Contributions:** Conceptualization, H.R.-M. and M.Á.-R.; methodology, M.L.-B.; software, J.G., A.J.-M. and M.Á.-R.; validation, A.J.-M.; formal analysis, J.G., A.J.-M. and M.Á.-R.; investigation, J.G. and C.A.M.; resources, H.R.-M., M.L.-B. and M.Á.-R.; data curation, J.G.; writing—original draft preparation, J.G.; writing—review and editing, A.J.-M., C.A.M., H.R.-M., M.L.-B. and M.Á.-R.; visualization, J.G.; supervision, M.L.M and M.Á.-R.; project administration, M.Á.-R.; funding acquisition, H.R.-M., M.L.-B. and M.Á.-R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Research Council FORMAS, Stockholm (Project 2017-00946 and Project 2019-00288), The Swedish Research Council (Vetenskaprådet, VR; project 2015-05919) and Juan de la Cierva Incorporación Postdoctoral Research Program (MICINN; IJDC-2015-24380). J.G. is supported by the Generalitat de Catalunya, Agency for Management of University and Research Grants co-financed with the Eropean Social Found (grants for the recruitment of new research staff 2018 FI\_B 00236).

**Acknowledgments:** We thank Annette Molbaek and Asa Schippert, from the Genomics Core Facility at LiU for expert assistance when running the bioanalyzer. We appreciate the kind support provided by Míriam Piles and Oscar Perucho and the technical staff from Torre Marimon—Institut de Recerca i Tecnologia Agroalimentàries (IRTA, Caldes de Montbui, Barcelona, Spain) and Annaïs Carbajal, Sergi Olvera-Maneu and Mateo Ruiz-Conca for their kind assistance in sample handling.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript or in the decision to publish the results.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Antibacterial Activity of Some Molecules Added to Rabbit Semen Extender as Alternative to Antibiotics**

**María Pilar Viudes-de-Castro 1,\*, Francisco Marco-Jimenez 2, José S. Vicente <sup>2</sup> and Clara Marin <sup>3</sup>**


**Simple Summary:** This study was conducted to evaluate the antibacterial activity of two aminopeptidase inhibitors and chitosan-based nanoparticles in liquid-stored rabbit semen. This study reports that the aminopeptidase inhibitors used to prevent bacterial growth could be used in semen extender as a suitable alternative to antibiotics.

**Abstract:** Although great attention is paid to hygiene during semen collection and processing, bacteria are commonly found in the semen of healthy fertile males of different species. As the storage of extended semen might facilitate bacterial growth, extenders are commonly supplemented with antibiotics. This study aimed to evaluate the antibacterial activity of ethylenediaminetetraacetic acid (EDTA), bestatin and chitosan-based nanoparticles added to rabbit semen extender and their effect on reproductive performance under field conditions. Four different extenders were tested, supplemented with antibiotics (TCG+AB), with EDTA and bestatin (EB), with EDTA, bestatin and chitosan-based nanoparticles (QEB) or without antibiotics (TCG-AB). Extended semen was cooled at 15 ◦C for three days. Cooled samples were examined for bacterial growth and semen quality every 24 h for 3 days. The enterobacteria count increased considerably during storage at 72 h in semen extended with TCG+AB and TCG-AB, while extenders EB and QEB showed a bacteriostatic effect over time. After 24, 48 and 72 h, quality characteristics were retained in all groups, with no significant motility differences, either in acrosome integrity, membrane functionality or the viability of spermatozoa. Additionally, bacterial concentration present in fresh semen did not affect reproductive performance. In conclusion, EDTA and bestatin exerted a potent bacteriostatic effect over time and could be used as an alternative to conventional antibiotics in rabbit semen extenders.

**Keywords:** artificial insemination; antibiotics; antibacterial activity; semen quality; reproductive performance

#### **1. Introduction**

Artificial insemination (AI) is a highly efficient assisted reproductive technology used worldwide in animal breeding. Semen from healthy fertile males from different species contains bacteria stemming from natural colonisation in the male tract and the environment, despite the application of strict hygienic measures during collection and manipulation processes [1], so international directives stipulate the addition of antibiotics to semen extenders to prevent bacterial growth. However, as a consequence of excessive antibiotics use in different fields such as human medicine, veterinary medicine, livestock and fish production, agriculture and food technology, microbial resistance has emerged as one of the main concerns worldwide [2,3].

**Citation:** Viudes-de-Castro, M.P.; Marco-Jimenez, F.; Vicente, J.S.; Marin, C. Antibacterial Activity of Some Molecules Added to Rabbit Semen Extender as Alternative to Antibiotics. *Animals* **2021**, *11*, 1178. https://doi.org/10.3390/ani11041178

Academic Editors: Rosa María García-García and Maria Arias Alvarez

Received: 21 March 2021 Accepted: 17 April 2021 Published: 20 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Semen contaminated with bacteria such as *Enterobacteriaceae* might adversely affect the quality of semen used for AI [4], in addition it is an important hygiene indicator [5–8]. Even though most of them are non-pathogenic bacteria, they can negatively influence sperm quality and longevity if present in high concentrations [6,7,9,10]. Additionally, several studies have shown that the seminal plasma of different species contains aminopeptidases [11–16]. Aminopeptidase activity promotes the proliferation of many bacteria that act as virulence factors, essential for the survival and maintenance of many microbial pathogens [17,18].

Nowadays, bacterial resistance to commonly used antibiotics and the global spread of resistance genes has become a serious health problem. Bacteria could counteract the actions of antimicrobials through different mechanisms, such as enzyme modification, alteration of the target binding sites, active efflux pumps or decreased permeability of bacterial membrane [19]. This resistance may occur through spontaneous mutations or by the horizontal transfer of mobile genetic elements from other bacteria, phages and/or the transmission of resistance genes from the environment [19,20]. In this sense, the transmission of resistance genes plays an important role in the spread of antimicrobial resistance among strains [19,20] and efforts need to be made to replace conventional antibiotics in the animal production industry. Some alternatives to conventional antibiotics in semen extender, such as colloid centrifugation [21–24] or removal of seminal plasma [25], have been investigated in different species. However, these techniques involve an increase in the processing time of semen. Others approaches for overcoming bacterial growth include the use of active molecules such as EDTA, [26], chitosan [27,28], nanoparticles [29], peptides [30,31] and aminopeptidase inhibitors [18,32], etc., with recognised antibacterial activity. EDTA is known to increase the outer cell wall permeability of Gram-negative bacteria, allowing other molecules easier access into the cell, facilitating an alteration or inhibition of its metabolism [33–35]. Chitosan, a biocompatible biodegradable and non-toxic polycationic copolymer extensively used as material for encapsulation and controlled release of chemicals [36], interacts with the bacterial cell membrane and causes cell lysis [37,38]. Furthermore, the encapsulation of GnRH in chitosan-based nanoparticles added to the extender supplemented with EDTA and bestatin can overcome the poor stability of the GnRH analogues in the presence of aminopeptidases [16,39] and allows to reduce the hormonal concentration used without affecting the reproductive performance of female rabbits [40]. This method of inducing ovulation in rabbit AI was developed to increase the welfare of rabbit insemination procedures and reduce the concentration of GnRH analogues added to semen extenders.

Against this background, this study aimed to evaluate the antibacterial activity of EDTA, bestatin and chitosan-based nanoparticles added to rabbit semen extender and their effect on the reproductive performance under field conditions, with a view to improving the sustainability of the rabbit production system.

#### **2. Materials and Methods**

The chemicals used in this study were purchased from Sigma-Aldrich (Merck Life Science S.L.U. Madrid, Spain). Animal housing and the protocols for semen collection and AI were approved by the Animal Care and Use Committee of Centro de Tecnología Animal, Instituto Valenciano de Investigaciones Agrarias. All animals were handled according to the European regulations for the care and use of animals for scientific purposes (European Commission Directive 2010/63/European Union).

#### *2.1. Extenders Composition*

Four different extenders were tested (Table 1). The solution used as a carrier for the molecules tested was Tris-citric acid-glucose (TCG extender; [41]). Chitosan and alginate were dissolved (0.05%) in the TCG supplemented with EDTA (20 mM) and bestatin (10 mM) according to Casares-Crespo et al. [40]. The nanoparticles were formed spontaneously in the coacervation process, directly after mixing the solutions of chitosan and alginate (4:1) through magnetic stirring (~600 rpm) for 30 min at room temperature.

**Table 1.** Semen extenders composition.


#### *2.2. Experimental Design*

2.2.1. Experiment 1: In Vitro Evaluation

#### Animals

Twelve males of New Zealand White origin were kept individually under similar conditions to those described by Viudes de Castro et al. [16].

#### Semen Collection

Semen from males was collected twice a week in three replicates. Strict attention was paid to the hygiene of collection equipment and semen samples were collected into sterile tubes. Semen collection and evaluation were conducted in accordance with Casares Crespo et al. [39]. Finally, all the ejaculates were pooled.

The pool was split into four equal fractions and diluted with the appropriate extender (dilution 1:10; *v*:*v*). The pools used in the experiment presented an average sperm concentration of 385 spermatozoa mL-1. Diluted samples were cooled at 15 ◦C for three days. Cooled samples were examined for bacterial growth, total motility, percentage of live sperm and membrane status every 24 h for 3 days.

#### Microbiological Analysis

Using Enterobacteriaceae as the sentinel, bacterial growth was evaluated by enumeration. Tenfold dilution series were performed in each extender to 10−6, and 1000 μL of each dilution was then plated onto Violet Red Bile Dextrose Agar (VRBD agar, Scharlab®, Barcelona, Spain) per duplicate; after homogenisation of the plate, 10 mL of VRBD were added to seal the plate. Plates were incubated for 24–48 h at 37 ± 1 ◦C. Typical colonies were counted, and the least dilute pair of plates that contained an average of between 30 and 150 colonies was used to calculate the number of bacteria (CFU/mL).

#### Seminal Quality Evaluation

Percentage of total motile sperm was evaluated using a computer-assisted sperm analysis system (ISAS Proiser, Valencia, Spain) as described by Viudes-de-Castro et al. [13]. Briefly, ten microlitres of each sample was placed into a 10 mm deep Makler counting chamber (Sefi Medical Instruments, Haifa, Israel). Sperm motility was assessed at 37 ◦C by negative phase contrast objective at a magnification of X100 (NIKON E-400 microscope, Izasa Scientific, Barcelona, Spain). Six microscopic fields were captured for each sample. A minimum of 400 sperm were evaluated using the same criteria described by Casares Crespo et al. [39].

Flow cytometric analyses to assess viability (integrity of sperm membrane) and acrosome integrity were performed with a CytoFLEX Flow Cytometer (Beckman Coulter, S.L.U., Barcelona, Spain) equipped with red (638 nm), blue (488 nm) and violet (405 nm) lasers and operated by the CytExpert Software v.2.3 (Beckman Coulter, S.L.U., Barcelona, Spain). The cytometer was calibrated daily using specific calibration beads provided by the manufacturer. Data were collected from 10,000 events. Gating the spermatozoa population after Hoechst 33,342 staining eliminated non-sperm events. Doublets and clumps were further excluded by using a plot of side scatter area and side scatter height followed by a gate of simple events. A compensation overlap was performed before each experiment. A

FITC-PNA/PI/Hoechst triple staining method, validated for rabbit semen in our laboratory, was used to determine viability and acrosomal status. To this end, 100 μL of semen at <sup>30</sup> × <sup>10</sup><sup>6</sup> sperm/mL were stained with 0.5 <sup>μ</sup>L Hoechst 33,342 (0.5 mg/mL) for 20 min at 37 ◦C without light. Subsequently, 1.5 μL FITC-PNA (1 mg/mL) and 0.5 μL PI (1 mg/mL) were added to the sample and incubated 10 min at 37 ◦C without light. Then 400 μL of TCG extender [36] were added to obtain a final concentration of 6 × <sup>10</sup><sup>6</sup> sperm/mL. PI-negative sperm were considered viable. The normal apical ridge (NAR) percentage was calculated as the proportion of acrosome intact sperm.

Membrane functionality analysis was assessed by hypo-osmotic swelling test (HOST). An aliquot of 100 μL of diluted semen was added to 1 mL of warmed 150 mOsm hypoosmotic swelling solution containing sodium citrate (25 mmol/L) and fructose (75 mmol/L) and incubated for 30 min at 37 ◦C. Subsequently, 10 μL of each sample were placed on a clean glass slide with a coverslip, and sperm swelling was assessed under phase-contrast microscopy. For each sample, a total of 200 spermatozoa were examined.

#### 2.2.2. Experiment 2: In Vivo Evaluation

Eight hundred and ninety-seven crossbreed females from a commercial farm (Altura, Castellón, Spain) were inseminated using fresh semen from 50 adult males belonging to a paternal rabbit line (Line R, [42]). Animal housing and seminal evaluation were similar to the previous experiment. All the ejaculates were pooled. The pool was split into four aliquots and diluted 1:10 with the four experimental extenders. After diluting the semen in the four experimental extenders, the insemination was initiated immediately. Each female was randomly assigned to one of the four experimental groups and was inseminated with 0.5 mL of semen using standard curved cannulas (24 cm). About 2 h elapsed between the first and the last inseminated female. At birth, pregnancy rate (number of kindlings/number of inseminated does) and prolificacy (total number of kits born) were evaluated.

#### *2.3. Statistical Analysis*

To analyse the effect of extender on Enterobacteriaceae growth, motility, viability, acrosome integrity and membrane functionality, a general linear model was used. The extender, refrigeration time and their interaction were taken as fixed effects and, in the case of seminal parameters, the corresponding parameter of the pool was introduced as a covariate in the analysis. A chi-square test was used to test differences in pregnancy rate at birth between groups. For the total number of kits born per litter, an ANOVA was performed, including as fixed effect the extender group and pool as covariate. All analyses were performed with the SPSS 26.0 software package (SPSS Inc., Chicago, IL, USA). Values were considered statistically different at *p* < 0.05.

#### **3. Results**

#### *3.1. Experiment 1: In Vitro Evaluation*

Results for bacterial growth are shown in Figure 1. The enterobacteria count increased considerably during storage at 72 h (Figure 1) in TCG+AB and TCG-AB groups. At 24 h, a significant increase of bacterial growth was observed in semen extended in TCG-AB (extender without antibiotics) compared to the rest of the groups. Additionally, at this time point, a significant decrease in bacterial growth was observed in the EB group versus the TCG+AB (extender with antibiotics) or QEB groups. At 48 h, it was found that there were no significant differences between EB and QEB, but there was significantly lower bacterial growth than TCG+AB and TCG-AB group. At 72 h, the trend was similar, with no significant differences between EB and QEB groups and showing a significantly lower bacterial growth than TCG+AB and TCG-AB groups. From 24 h, bacterial growth in the TCG+AB group increased over time, being three times higher at 72 h than that observed in groups EB and QEB, which maintained the same number of CFUs over time.

**Figure 1.** Bacterial growth (Log CFU/mL; mean ± SEM) in sperm samples from the rabbit where there was storage at 15 ◦C for 72 h in four extenders: EB: TCG supplemented with EDTA (20 mM) and bestatin (10 mM). QEB: TCG supplemented with EDTA (20 mM), bestatin (10 mM) and nanoparticles of chitosan-alginate (0.05%). TCG+AB: TCG extender supplemented with 100 IU/mL penicillin + 100 μg/mL streptomycin. TCG-AB: TCG extender without antibiotics. Different superscripts (a, b and c) indicate differences in values at the same time point (*p* < 0.05).

Seminal quality parameters of samples from the experimental extenders are shown in Table 2. There was no interaction between extender and refrigeration time. The presence of EDTA, bestatin and chitosan nanoparticles had no effect on total motility, acrosome integrity, membrane functionality or the viability of spermatozoa. Following 24, 48 and 72 h, quality characteristics were retained in all groups, with no significant differences in motility, acrosome integrity, membrane functionality or the viability of spermatozoa.


**Table 2.** Sperm quality in stored rabbit spermatozoa in four extenders (means ± standard deviation).

UTCG+AB: TCG extender supplemented with 100 IU/mL penicillin + 100 μg/mL streptomycin. EB: TCG supplemented with EDTA (20 mM) and bestatin (10 mM). QEB: TCG supplemented with EDTA (20 mM), bestatin (10 mM) and nanoparticles of chitosan-alginate (0.05%). TCG-AB: TCG extender without antibiotics. N: number of seminal pools; Total Mot: total motility; HOST: hypo-osmotic swelling test; NAR: acrosome normality.

#### *3.2. Experiment 2: In Vivo Evaluation*

Pregnancy rate at birth and the total number of kits born are presented in Table 3. Neither pregnancy rate at birth nor prolificacy were affected by the experimental group, both parameters being similar between groups.


**Table 3.** Fertility (%) and prolificacy (means ± standard deviation) obtained from insemination of 897 females.

TCG+AB: TCG extender supplemented with 100 IU/mL penicillin + 100 μg/mL streptomycin. EB: TCG supplemented with EDTA (20 mM) and bestatin (10 mM). QEB: TCG supplemented with EDTA (20 mM), bestatin (10 mM) and nanoparticles of chitosan-alginate (0.05%). TCG-AB: TCG extender without antibiotics. N: number of inseminated does.

#### **4. Discussion**

Bacterial contamination is of particular relevance in rabbit AI, where most inseminations are carried out with liquid semen storage at 15 ◦C [43]. Even though semen collection protocols in the livestock industry are very strict, semen collection is not a sterile process, and the addition of antibiotics to extenders to control contaminating bacterial populations is a routine fact at farm level. The efficacy of different antibiotics added to semen extenders in livestock has been widely demonstrated [44]. However, how quickly bacteria acquire tolerance and/or resistance to antibiotics is essential. Numerous studies show the critical resistance patterns found in semen samples of different species such as boars [45], humans [46] and bovine [47]. However, it can be suggested that the antibiotics currently used in routine practices in livestock, such as AI, may need to be modified to avoid future complications arising from bacterial resistance.

Previous study demonstrated that sperm microbiota diversity is influenced by host genetics [9]. The rabbit semen samples were contaminated with bacteria, especially those that belong to the *Enterobacteriaceae* family [44], ranging from 27.6% for Line V to 50.9% for Line R of semen samples analysed [9]. Bacteria contamination, such as *Enterobacteriaceae* family might adversely affect the quality of semen used for AI [4] and is an important hygiene indicator [5], although the majority of these bacterial strains are not currently considered pathogens [7]. The results of this study prove that the replacement of antibiotics in the current extenders by EDTA and bestatin prevent bacterial growth through 72 h in the rabbit doses. Therefore, the present study results validate the bacteriostatic effect of aminopeptidase inhibitors, such as bestatin and EDTA, and highlight the role of protease inhibitors in the control of seminal bacterial growth. This is in agreement with several studies in which an inhibitory activity of EDTA against Gram-negative bacteria, Grampositive bacteria (staphylococci) and fungi (*Candida* spp.) was observed [33–35]. Likewise, some studies indicated that EDTA, alone or in combination, is an effective antibiofilm agent with a spectrum covering both Gram-positive and Gram-negative bacteria [35,48–52]. In addition, bacterial proteases participate in important metabolic pathways and have key roles in cell viability, stress response and pathogenicity [53]. On the other hand, despite the antimicrobial activity demonstrated by chitosan against several pathogens [54–56], in the present study, the presence of chitosan-alginate nanoparticles in the extender did not show a synergistic action with aminopeptidase inhibitors on the microbial growth, with both extenders showing similar results. A possible explanation for this is that the presence of alginate can interfere in the inhibition of bacterial growth by chitosan. As the presence of cationic charge situated in the amino group of chitosan is essential for exhibiting high antimicrobial properties [57], the ionic linkages between functional groups of the oppositely charged chitosan and alginate would result in low availability of unreacted positive amino groups of chitosan when nanoparticles were formed, which reduces the chances of interaction with negatively charged components of microbial cell membranes. Our results suggest that both aminopeptidase inhibitors (EDTA and bestatin), alone or in combination with nanoparticles of chitosan-alginate, maintained total motility, viability, acrosome status and functional integrity of the sperm plasma membrane for at least three days. Moreover, the use of both aminopeptidase inhibitors preserved the fertility and prolificacy under field

conditions. Undoubtedly, this result could be used to improve the sustainability of the rabbit production system. In the present study, the classic combination of penicillin and streptomycin contributed to the diluent's efficacy in controlling the growth of *Enterobacteriaceae* only for up to 24 h. Nevertheless, from this moment on, this antibiotic cocktail cannot prevent the bacterial growth, which was probably due to an increase in tolerance of antibiotics by *Enterobacteriaceae* [58,59]. The common use of antibiotics in extenders is an important concern: apart from being prophylactic and non-therapeutic, and therefore going against the recommendations for prudent use of antimicrobials, they can cause increase antibiotic resistance in the bacteria commonly found in semen [24]. As expected, in the extender without antibiotics (TCG-AB), no bacteriostatic effect was observed throughout the entire cooling period, showing an increasing number of enterobacteria over time, highlighting the need to supplement AI extenders with substances that control bacterial growth. However, there were no effects of enterobacteria contamination on in vitro quality sperm in long-term stored samples. Although the presence of microorganisms in semen may reduce semen quality and fertilising capacity during preservation time, our results indicate that bacterial concentration present in fresh semen (0 to 4 h from semen collection) has no effect on reproductive performance, which is in agreement with Jäkel et al. [60] in pig, who reported similar reproductive performance at 24 h between groups inseminated with semen diluted in extender with or without antibiotics. Extenders and storage temperature are important factors to preserve the fertilising capacity of rabbit semen. As regards the sperm quality variables, all extenders used in the present study preserved the quality of rabbit semen throughout the cooling period. Several studies to evaluate stored rabbit semen have been carried out under different experimental conditions. On the one hand, several authors had observed that motility decreased when semen was stored at 5 ◦C during 48 h, irrespective of the extender used [61–63]. On the other hand, other authors have shown that rabbit semen stored at 15 ◦C up to 48 h retains fertility capacity [43,64]. However, further studies are needed to verify the reproductive performance of rabbit semen stored for 72 h in extenders supplemented with EDTA and bestatin.

#### **5. Conclusions**

In conclusion, we demonstrated in this study that the addition of EDTA and bestatin to semen extender exerted a potent bacteriostatic effect over time, effectively inhibiting the growth of Enterobacteriaceae, which suggests that EDTA and bestatin could be used as an alternative to conventional antibiotics in rabbit semen extenders.

**Author Contributions:** Conceptualisation M.P.V.-d.-C., C.M.; Methodology and data curation M.P.V.-d.-C., C.M., F.M.-J. and J.S.V.; writing—original draft preparation, M.P.V.-d.-C.; writing and analysis review and editing, M.P.V.-d.-C., C.M., F.M.-J. and J.S.V. All authors have read and agreed to the published version of the manuscript.

**Funding:** Funding from the Ministry of Economy, Industry and Competitiveness (Research project: AGL2017-85162-C2-1-R) is acknowledged.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Animal Care and Use Committee of Centro de Tecnología Animal, Instituto Valenciano de Investigaciones Agrarias (Protocol #2018/VSC/PEA/0116).

**Acknowledgments:** The authors thank Manuel Sierra and Felipe Lavara for offering their animal facility for conducting this study under field conditions.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**

The following abbreviations are used in this manuscript:


#### **References**


#### *Article*

## **Influence of Different Regimes of Moderate Maternal Feed Restriction during Pregnancy of Primiparous Rabbit Does on Long-Term Metabolic Energy Homeostasis, Productive Performance and Welfare**

**Carlota Fernández-Pacheco 1, Pilar Millán 1, María Rodríguez 2,†, Nora Formoso-Rafferty 2, Ana Sánchez-Rodríguez 1,‡, Pedro L. Lorenzo 1, María Arias-Álvarez 3, Rosa M. García-García <sup>1</sup> and Pilar G. Rebollar 2,\***


**Simple Summary:** In rabbit farms, the main production costs come directly from food supplies. Although the reproductive outcomes in this species are acceptable, the results are worse when it comes to primiparous rabbits, so it is recommended that insemination be carried out post-weaning. By avoiding the overlap of the second gestation and the first lactation, better fertility results are expected. Still, despite this, the rabbits garner more adipose tissue than desired, and the productive efficiency deteriorates in the long term. The purpose of this study was to evaluate the influence of different periods of moderate feed restriction (one, two or three weeks) applied during the second pregnancy of primiparous does. We studied fetoplacental development, productive parameters, metabolism and possible stress indicators. Results showed that the voluntary feed intake of dams increased right after feed restriction. No permanent alterations were found in reproductive outcome, metabolism or welfare of does, meaning that this feeding strategy could be successfully applied in rabbit farms.

**Abstract:** In this study, a maternal feed restriction (MFR; 105 g/d) in primiparous rabbit does was applied from day 0 to 7 post artificial insemination (AI) (R07, n = 96), from day 7 to 21 post AI (R721, n = 92), from day 0 to 21 post AI (R021, n = 94) or fed ad libitum during whole pregnancy (Control, n= 92). Feed intake (FI) was measured after MFR was over. On day 28 of gestation, fetoplacental development was evaluated (n = 11/group) and the productive parameters of the remaining dams were analyzed. Plasma free tri-iodothyronine (T3) and thyroxine, glucose, insulin, non-esterified fatty acids (NEFA), and corticosterone were analyzed during gestation and lactation (n = 5/group). After MFR, all groups significantly increased their voluntary FI. The longer MFR was, the lower the weight and length of the fetuses, but no long-term effects over litter performance were observed. R021 groups had the lowest T3 and the highest NEFA concentrations during pregnancy and showed insulin resistance at the end of gestation, but during lactation, energy homeostasis was balanced in all groups. MFR did not affect corticosterone concentrations. In conclusion, the ration setting applied slightly involved the energy homeostasis and metabolism of the animals, but their overall metabolic condition, productive performance and welfare were not compromised.

**Citation:** Fernández-Pacheco, C.; Millán, P.; Rodríguez, M.; Formoso-Rafferty, N.; Sánchez-Rodríguez, A.; Lorenzo, P.L.; Arias-Álvarez, M.; García-García, R.M.; Rebollar, P.G. Influence of Different Regimes of Moderate Maternal Feed Restriction during Pregnancy of Primiparous Rabbit Does on Long-Term Metabolic Energy Homeostasis, Productive Performance and Welfare. *Animals* **2021**, *11*, 2736. https://doi.org/ 10.3390/ani11092736

Academic Editor: Víctor Hugo Parraguez Gamboa

Received: 1 September 2021 Accepted: 16 September 2021 Published: 19 September 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Keywords:** rabbit; feed intake; free tri-iodothyronine; thyroxine; insulin; glucose; corticosterone; NEFA; fetus; placenta

#### **1. Introduction**

In rabbit farms, dam feeding represents roughly a third of the total feed costs (3.7% and 31.7% for replacement and reproductive does, respectively) [1]. Moreover, the number of kits born alive and the feed conversion rate during fattening are of high economic importance. These traits depend directly on maternal nutrition during pregnancy and lactation. Artificial insemination (AI) of does is usually applied on Day 4 or 11 post-partum, while females are lactating [2]. Alternative breeding strategies can be applied to alleviate the negative energy balance that befalls after first parturition with concurrent gestation and lactation of primiparous does, which are still completing their body growth [3]. One of them is delaying the interval parturition-AI until weaning (Day 30 post-partum), applying a reproductive extensive rhythm [4]. This improves reproductive outcome [5], but considering the first weeks of pregnancy do not involve a very high energy expenditure and it is considered an anabolic period [6,7], the important risk of fattening exists due to ad libitum feeding of post-weaning inseminated does. Nonetheless, it is still a critical phase that can influence the rate of embryonic implantation and the development of the placenta, which are limiting factors for good fetal growth. Moreover, the mentioned fattening risk increases if these dams do not become pregnant (pregnancy diagnosis is usually performed on Day 10–14 post AI). Thus, feeding strategies as a subject of study in these animals include applying different levels of maternal feed restriction (MFR) during diverse periods of pregnancy.

MFR has had diverse targets to avoid fattening and high mortality around parturition [8], to determine the competition for materno-fetal resource partitioning [9], to increase voluntary feed intake at the beginning of lactation, or to allow a longer productive life of rabbit does [10]. López-Tello et al. [9,11] verified that undernourishment in dams (50% MFR) during the entire pregnancy or during preimplantation period (1st week) has adverse effects on the fetoplacental unit (reduced fetal crown-rump lengths, asymmetrical growth and high apoptotic rates at the decidua and labyrinth zone). In this sense, it has also been described that if the daily amount of food provided increases to 60% of the dam's total voluntary intake, negative effects are observed only in the second half of gestation. In contrast, if MFR is applied in the first 15 days of pregnancy, no adverse effects are observed on the viability or weight of the kits at birth [12]. Moreover, ad libitum refeeding in the last third of gestation has been shown to improve the energetic status of the mothers before parturition [13], precisely when fetal needs are most significant. It also reduces the negative effects of subsequent lactation on their reproductive function (poor ovulatory response and increased delivery-fertile insemination interval) and improves the weight and viability of the kits.

What we can introduce about the application of MFR is largely based upon our previous experimental studies [14,15], in which we have described the metabolic consequences observed in mothers and offspring subjected to a 60% MFR during three weeks of pregnancy (from day 0 to 21). This MFR led to a compensatory feed intake during the last week of pregnancy that offset the effects in the live body weight (LBW) and body reserves of the dams at parturition and the fetal body weight and phenotype at day 28 of pregnancy. These studies on MFR applied during three weeks of pregnancy have been limited to specific comparisons of the metabolic profiles of mothers and fetuses at a particular moment of gestation (day 28).

Thus, in order to gain further understanding of the metabolic changes and energetic modulations of primiparous does subjected to MFR, the novelty of this study has been to apply shorter and more specific periods of restriction during gestation studying the long-term consequences on additional hormones and metabolites. Among others, thyroid hormones are key to the regulation of metabolism and adaptation to fasting. They contribute to both mandatory and adaptive thermogenesis, regulating appetite and energy expenditure [16]. On the other hand, glucose is central to energy consumption. Carbohydrates, lipids, and proteins all ultimately break down into glucose, which serves as the main metabolic fuel of mammals' cells and the prevalent fuel of the fetus [17]. Likewise, insulin is a primary anabolic hormone secreted in response to increased blood glucose and amino acids following feeding. The major action of insulin is to stimulate and control glucose consumption and exertion. Like other hormones, insulin performs its activities by binding to specific receptors present on many cells distributed all through the body [18]. Additionally, non-esterified fatty acids (NEFA) concentrations reflect the mobilization of body reserves since they are generated from lipolysis in fatty tissue and are an indicators of negative energy balance [19]. Briefly, the main interest of this study was to measure these metabolic pathways and hormones since they are known to have different behaviors when adapting to the specific needs of gestation, fetal development and milk production [20]. Furthermore, and considering animal welfare standards, food or water deprivation is a procedure that can cause pain or distress [21], making the analysis of stress-related hormones mandatory when a feed restriction is applied. It should be noted that in rabbits and rodents, unlike other mammals, the main glucocorticoid analyzed is usually corticosterone rather than cortisol [22].

Therefore, the main goal of this study was to determine in primiparous lactating does the comparative effect of MFR applied during one, two or three weeks of pregnancy followed by ad libitum refeeding on the feed intake of the dams, feto-placental development on day 28 of pregnancy, productive performance, the weekly plasma concentrations of thyroid hormones and corticosterone, and the glucose and lipid metabolism during pregnancy and consequent lactation.

#### **2. Materials and Methods**

#### *2.1. Experimental Design*

The animals were housed at the animal facilities of the Technical University of Madrid (Spain), which meet the local, national and European requirements for Scientific Procedure Establishments (RD 53/2013; PROEX 302/15). The experimental design is displayed in Figure 1. New Zealand × California rabbits (18 weeks old), were fed ad libitum during their first pregnancy (length 31 days) and lactation (length 30 days) with a diet containing 16% crude protein, 37% crude fiber, 3.7% fat and 2400 kcal/kg of digestible energy (NANTA, Madrid, Spain). During the first pregnancy, the feed intake of each doe was recorded daily and established in 175 g per animal and day.

After parturition and weaning (day 30 post-partum) these does started a second cycle and were artificially inseminated (AI) with fresh diluted semen (commercial extender, MA 24, Ovejero, León, Spain). Each dose contained at least 20 million spermatozoa in 0.5 mL of semen diluent. Ovulation was induced with gonadoreline at the time of AI (20 μg/doe, i.m.; Inducel-GnRH, Ovejero, León, Spain). At this time (day 0), a total of 374 does were randomly allocated into four groups:


MFR was set at 60% of the measured voluntary feed intake of the first pregnancy (105 g/day), in order to maintain their basal metabolic needs. In the fourth week of pregnancy, all animals were fed ad libitum until the end of the study. Pregnancy was diagnosed at day 11 post-AI by means of abdominal palpation and fertility ([no. of does pregnant at day 11 after AI/no. of does inseminated] × 100) was calculated. Feed intake was weekly calculated exclusively in the pregnant does from day 0 until the day of parturition.

#### *2.2. Fetoplacental Study*

On day 28 of gestation, 44 pregnant does (n = 11 from each group) were weighed and euthanized with an overdose barbiturate (Dolethal, Lab. Vetoquinol, Madrid, Spain) to study fetoplacental development. A ventral laparotomy was performed to count the corpora lutea present in the ovaries and calculate the ovulation rate per animal. Both uterine horns were opened to extract the fetuses and placentas, calculating the rate of viable structures [(number of viable fetuses/number total of fetuses) × 100]. In morphologically viable fetuses, crown–rump length (CRL), biparietal (BPD), occipito-nasal (OND), and thoracic (TD) diameters were measured. Whole fetuses were weighed, and after decapitation, heads and trunks were weighed separately. Fetal organs' (brain, liver, heart, lungs, digestive tract and kidneys) weights and their ratios (organ weight/fetus weight) were calculated to assess fetal growth patterns. Placentas were weighed both as a whole and separately for the maternal (decidua) from the fetal (labyrinth) part. The length, width and thickness of each of the pieces were recorded. Placental efficiency was calculated by dividing the weight of the fetus between the weight of the whole placenta.

#### *2.3. Productive Outcome*

The gestation of the remaining rabbit does was carried out, and prolificacy (no. of kits born alive and stillborn per doe) and litter weights at birth were recorded. Litter size was subsequently standardized to 8–12 kits by removing or adding kits within each experimental group.

During lactation, a total number of 50 litters from each experimental group were weighed at day 7, 14 and 21 post-partum. Milk production was estimated using the regression equation developed by Helfenstein et al. [23] as follows: milk production (kg) = 0.75 ± 1.75 LW21 (kg) where LW21 corresponds to litter weight at 21 days of lactation. The day of weaning (day 30 post-partum) the number of weaned kits per doe, litter weight and mortality during lactation period were assessed. At the end of the study, the percentage of culled does (abortions or sudden death) in each experimental group was calculated.

#### *2.4. Long-Term Maternal Metabolic and Hormonal Study*

In addition, the evaluation of metabolic and hormonal parameters was carried out on blood samples obtained from the marginal ear vein of randomly chosen females. An initial blood sample was taken at the moment of AI (day 0 of pregnancy) from 10 dams before being allocated into the experimental groups. After pregnancy diagnosis, five pregnant animals per group were sampled at three different moments of gestation: day 14 (n = 20), day 21 (n = 20) and day 28 (n = 20) of gestation. Following parturition, samples were

taken from lactating females (five per group) at three moments of lactation: day 7 (n = 20) and day 14 (n = 20). On day 14 post-partum, all does were re-inseminated as previously described. Then, on day 30 post-partum, weaning was performed and blood samples were taken from pregnant (n = 20) and non-pregnant (n = 20) does from the third AI. Blood samples were placed in tubes with EDTA as an anticoagulant and centrifuged for 15 min at 1200 g to obtain plasma.

Moreover, the evaluation of energetic homeostasis was performed by measuring free triiodothyronine (T3) and thyroxine (T4) in plasma samples using commercial immunoassays (Demeditec Diagnostics GmbH, Kiel, Germany). The assay sensibility was 0.05 pg/mL for both T3 and T4, and the inter-assay and intra-assay variation coefficients were 9.8% and 3.6% for T3 and 4.9% and 3.3% for T4. Glycemic metabolism was assessed by the determination of insulin and glucose levels. Insulin was determined using a commercial immunoassay (Mercodia Ultrasensitive Insulin ELISA, Mercodia AB, Uppsala, Sweden) with an assay sensitivity of 0.071 mU/L and inter-assay and intra-assay variation coefficients of 6.2% and 4.2%, respectively. Insulin sensitivity was calculated by glucose-to-insulin ratio and the homeostasis model assessment for insulin resistance (HOMA-IR) using the following equation: [insulin (mU/L) × (glucose (mg/dL)/18)]/22.5 [23]. Low HOMA-IR values (<1.96) indicate high insulin sensitivity, whereas high HOMA-IR values (>3) indicate low insulin sensitivity (insulin resistance). Non-esterified fatty acids (NEFA) were selected as the main indicator of lipid metabolism and were measured using a clinical biochemistry assay (Biolabo SAS, Maizy, France), with detection in the range of 0.01–3.0 mmol/L and assay sensitivity 0.050 mmol/L. Corticosterone was used as a stress indicator [22] and was determined using a commercial immunoassay (Demeditec Diagnostics GmbH, Kiel, Germany). The assay sensitivity was 1.63 nmol/L and the inter-assay intra-assay variation coefficients were 6.5% and 2.8%.

#### *2.5. Statistical Analysis*

SAS software (Statistical Analysis System Institute Inc.; Cary, NC, USA, 2001) was used for the statistical analysis of the data. In all analyses, the doe was considered the experimental unit. Differences between groups in the total feed intake were analyzed by a one-way ANOVA (Proc GLM) with the feed regime as the main effect. The effect of MFR on weekly feed intake during gestation was analyzed by repeated measure analysis (MIXED), with feed regime (Control, R07, R721 and R021), time (weeks 1, 2, 3 and 4), and their interaction as main effects. Ovulation rate, live body weight of does, number, weights and measurements of fetuses and placentas, and placental efficiency on day 28 of pregnancy were analyzed (Proc GLM) with feed regime as main effect and considering the litter size as a covariate. With the same procedure, prolificacy and milk production were analyzed with feed regime as the main effect. A chi squared test (Proc CATMOD) was used to evaluate the effect of MFR on fertility and percentage of culled does. Litter weights during the lactation period were analyzed by repeated measure analysis (Proc MIXED), considering the MFR, the time (birth, 7, 14, 21, and 30 days post-partum) and their interaction in the model. For hormonal and metabolic parameters, due to the blood sampling being carried out in independent randomly chosen rabbit does, a GLM procedure was performed, considering as main effects the treatment (MFR), the time (pregnancy: day 14, 21 and 28 of gestation; and lactation day 7, 14 and 30 post-partum), and the interaction between both. If significant main effects were detected, a Tukey test (for parametric variables) or Kruskal–Wallis (non-parametric variables) were used to compare means among groups, considering the existence of significant differences for a *p* value of less than 0.05. All data are presented as least squares means.

#### **3. Results**

#### *3.1. Feed Intake during Pregnancy*

As expected, whole feed intake during pregnancy was different between groups (*p* < 0.0001). The Control and R07 groups consumed 5854 ± 124.3 g and 5529 ± 109 g, respectively, followed by R721 does (4518 ± 112 g) and the R021 group (4007 ± 106.3 g). The weekly evolution of daily feed intake is shown in Figure 2. Interestingly, the week after MFR, we observed that all groups increased their voluntary feed intake, surpassing the control group, like group R07 during the second week, and groups R721 and R021 during the fourth week of pregnancy.

**Figure 2.** Weekly evolution of the daily feed intake (g/day) in pregnant rabbit does fed ad libitum (Control), restricted to 60% one week from day 0 to 7 (R07), two weeks from day 7 to 21 (R721) and three weeks from day 0 to 21 (R021) of gestation. (a–c): Different letters represent significant differences between experimental groups (*p* < 0.0001).

#### *3.2. Fetoplacental Study*

Results obtained on day 28 of pregnancy are shown in Table 1. This table is quite revealing in several ways. First, regarding body reserves of does at the end of pregnancy, does restricted for two weeks (R721) presented a lower LBW than Control and R07 groups (*p* = 0.0254), whereas does restricted for three weeks (R021) had an intermediate body weight.

From a reproductive point of view, the number of corpora lutea, total fetuses per doe, implantation rate and fetal viability were high and similar in all groups. In R07 does, there is a tendency (p = 0.0824) to have more fetuses but with lower viability.

Feed restriction significantly affected (*p* < 0.01) the total, head and trunk weights and CRL of fetuses, being lower in those from R721 and R021 does than in the control group, and intermediate in R07 females. Regarding fetal cephalic diameters, BPD was similar in all groups (*p* > 0.05), however OND was higher in Control and R07 does than in R021, being intermediate in R721 (*p* = 0.0032). Fetal thoracic diameter was greater in R07 compared to the other three groups (*p* = 0.0034). Regarding fetal organs, only liver, lung and kidney weights were different between groups (*p* = 0.0097, *p* = 0.0166, *p* = 0.0001, respectively), with those of the Control group being heavier except for the liver.

Weights and dimensions of labyrinth and decidua from euthanized does at day 28 of pregnancy are shown in Table 2. Regarding these data, it is interesting to indicate that the whole placenta of R07 does weighed more than those from the R721 and R021, and the Control group had an intermediate value (*p* < 0.0001). Furthermore, the labyrinth of R07 does tended to be the heaviest (*p* = 0.0809), whilst that of R021 does was the shortest (*p* < 0.0001) and that from R721 the thinnest (*p* < 0.0053). In does subjected to MFR for two weeks (R721), decidua had lower weight (*p* < 0.0001) than the other three groups. The longest decidua was observed in Control does (*p* < 0.0001) and the thinnest was observed in R721 (*p* = 0.0069). Finally, the lowest placental efficiency was observed in R07 does, the highest in R721, and intermediate in Control and R021 does (*p* = 0.0162).


**Table 1.** Number and biometric parameters of fetuses of 28 days of gestational age in rabbits fed ad libitum (Control), restricted to 60% one week from day 0 to 7 (R07), two weeks from day 7 to 21 (R721) and three weeks from day 0 to 21 (R021) of gestation.

LBW: live body weight. <sup>1</sup> (number of fetuses/number of corpora lutea) × 100; <sup>2</sup> (number of viable fetuses/total number of fetuses) × 100; CRL: crown–rump length; TD: thoracic diameter; BPD: biparietal diameter; OND: occipito-nasal diameter. Data are shown as least squares means. Ratio: organ weight/fetus weight. Values in the same row with different letters are significantly different. SEM: standard error of mean.

**Table 2.** Weight and dimensions of placentas from pregnant rabbits on day 28, fed ad libitum (Control), restricted to 60% one week from day 0 to 7 (R07), two weeks from day 7 to 21 (R721) and three weeks from day 0 to 21 (R021) of gestation.


<sup>1</sup> Fetus weight/whole placenta weight. Data are shown as least squares means. Values in the same row with different letters are significantly different. SEM: standard error of mean.

#### *3.3. Productive Outcome*

Productive outcome is shown in Table 3.

**Table 3.** Productive parameters of rabbits fed ad libitum (Control), or restricted to 60% of their needs, one week from day 0 to 7 (R07), two weeks from day 7 to 21 (R721) and three weeks from day 0 to 21 (R021) in its second gestation.


<sup>1</sup> Number of does pregnant after pregnancy diagnosis at 11 days post AI/total does inseminated × 100. <sup>2</sup> This parameter was obtained from 50 litters per experimental group. <sup>3</sup> Milk production = 0.75 ± 1.75 × LW21 (kg) where LW21 corresponds to litter weight at 21 days of lactation. <sup>4</sup> Mortality: [100−(kits weaned/no. of kits after litter adjustment to 8–12 kits) × 100]. Data are shown as least squares means. RMSE: root mean square error. Values in the same row with different letters are significantly different. pp: post-partum.

Fertility, kits born alive and stillborn, and litter weight at birth and during lactation were similar between groups (*p* > 0.05). However, after litter adjustment, the effect of MFR on the number of weaned kits and mortality of kits during lactation was significantly different between groups (*p* = 0.0004 and *p* = 0.0325, respectively). Control and R07 does weaned more kits than R721 does, whereas R021 does had an intermediate result. The highest and lowest mortalities during lactation were observed in R021 and R07 litters, respectively.

At the end of the study, the percentage of culled does was 7.6, 4.2, 5.4 and 6.4% in Control, R07, R721 and R021 groups, respectively (*p* > 0.05).

#### *3.4. Long-Term Maternal Metabolic and Hormonal Study*

Endocrine and metabolic parameters obtained in the experimental groups during pregnancy are shown in Table 4.

Regarding thyroid hormones, the animals subjected to the most extended restriction treatment (R021) had lower plasma T3 concentrations than Control and R07 does (*p* = 0.0196) but similar to R721 ones. Plasma T3 concentrations decreased significantly on days 14 and 21 of pregnancy in relation to day 0 and returned to intermediate values on day 28 (*p* = 0.0026). Plasma T4 concentrations on Days 14 and 21 of pregnancy were also significantly lower than on day 28 (*p* = 0.0001) but similar to that on day 0. MFR did not affect T4 concentrations, T3 to T4 ratio, plasma glucose and HOMA-IR (*p* > 0.05).


**Table 4.** Effect of maternal feed restriction (MFR) and time on endocrine and metabolic parameters during pregnancy of rabbits fed ad libitum (Control), restricted to 60% one week from day 0 to 7 (R07), two weeks from day 7 to 21 (R721) and three weeks from day 0 to 21 (R021) of gestation.

NEFAS: no esterified fatty acids. RMSE: root mean square error. HOMA-IR: homeostasis model assessment for insulin resistance [Insulin × (glucose/18)/22.5]. Values in the same row with different letters (a, b, c) are significantly different.

> However, insulinemia was lower, and glucose-to-insulin ratio was higher in R721 does than in Control and R07 does, but similar to R021 ones (*p* = 0.0225 and *p* = 0.0482, respectively). In addition, plasmatic glucose increased dramatically at the end of pregnancy in all groups (*p* < 0.0001).

> The highest and lowest NEFA concentrations were obtained in R721 and R07 does, respectively, and Control and R021 does showed intermediate values (*p* = 0.0158). No significant differences in NEFA levels throughout the pregnancy period were observed (*p* > 0.05).

> Finally, plasma corticosterone concentration was only affected by time. The lowest concentrations of this stress indicator were observed at 21 and 28 days of pregnancy (*p* < 0.0001).

> MFR and time had significant effects on plasma concentrations of T3 (*p* = 0.0353), insulin (*p* = 0.0055), NEFA (*p* = 0.0100) and HOMA index (*p* = 0.0049) (Figure 3A–D, respectively). On day 21 of pregnancy, R021 does showed a dramatic reduction in plasma concentrations of T3 compared to R07 groups, but it was similar to Control and R721 does. The lowest plasma insulin concentrations were observed on day 21 of pregnancy in the rabbit females restricted at this time (R721 and R021). Control and R07 does had the lowest NEFA concentrations on day 14 of pregnancy. During the first half of pregnancy, HOMA-IR remained lower than 3 and by day 21 of gestation its values rose above the limit established to consider insulin resistance in Control and R07 groups. On day 28 of pregnancy, all groups showed HOMA-IR values over 3 except for group R721.

> As shown in Table 5, MFR did not affect any maternal endocrine or metabolic parameters in the experimental groups during lactation (*p* > 0.05). As the lactation period progressed, T3 and glucose concentrations remained at similar values during the first two weeks of lactation, but insulin and HOMA-IR decreased on day 14 (*p* = 0.0006) compared to day 7. On weaning day, non-pregnant does had lower T3 concentrations than pregnant ones (*p* < 0.0056), but HOMA-IR, glucose and insulin concentrations were similar in both (*p* > 0.05).

> Energy mobilization measured by means of plasma NEFA concentrations was similar in all groups and during all lactation. Plasma corticosterone concentrations as welfare indicators were not affected by MFR. However, significant increases in this hormone were observed as lactation progressed. No significant interactions between MFR and time in any of the studied variables were detected in this period.

**Figure 3.** Effect of Maternal Feed Restriction and Time on (**A**) plasma triiodothyronine (T3), (**B**) insulin, (**C**) NEFA concentrations and (**D**) HOMA-IR during pregnancy of rabbits fed ad libitum (Control), restricted to 60% one week from day 0 to 7 (R07), two weeks from day 7 to 21 (R721) and three weeks from day 0 to 21 (R021) of gestation. (a–c): different letters indicate differences between experimental groups on each day of pregnancy (*p* < 0.05). Each bar on day 14, 21 and 28 shows the mean of five does, and on day 0 is the mean of 10 does. The dashed line in (D) represents the value above which HOMA-IR indicates the presence of insulin resistance.

**Table 5.** Effect of Maternal Feed Restriction (MFR) and Time on endocrine and metabolic parameters during lactation of rabbits fed ad libitum (Control), restricted to 60% one week from day 0 to 7 (R07), two weeks from day 7 to 21 (R721) and three weeks from day 0 to 21 (R021) of gestation.


HOMA-IR: homeostasis model assessment for insulin resistance. NEFA: no esterified fatty acids. <sup>1</sup> 3rd artificial insemination was carried out on day 14 of lactation. <sup>2</sup> Weaning day. P: pregnant. NP: non-pregnant. RMSE: root mean square error. Values in the same row with different letters are significantly different.

#### **4. Discussion**

This work deals with the potential consequences of moderate MFR in different periods (1, 2 or 3 weeks) throughout gestation on reproductive outcome, including fetoplacental development and post-natal growth and survival of litters from primiparous rabbit does, and introduces further insight involving endocrine and metabolic parameters during pregnancy and the following lactation.

After different MFR periods, when fed ad libitum, all pregnant does significantly increased their feed intake. These results match with those observed in our previous studies [9,14,15]. Dams from R721 and R021 groups altogether consumed 1.3 kg and 1.8 kg less feed than the Control group during gestation, respectively. Those in the R07 group reduced their feed intake by about 300 g compared to the Control group. Considering the feed intake of Control females during the second and third weeks as the average regular consumption (227 ± 6.0 g/d and 228.3 ± 6.1 g/d), R721 and R021 does ingested 46% of the food that the Controls did. Consequently, LBW of these does was impaired, showing around a 5.2% decrease in LBW compared to Control or R07.

As expected, the number of corpora lutea assessed in the rabbits on day 28 of gestation was similar in all groups, as MFR was applied right after AI, with no impact on the ovulation rate. Similar results have been described in our previous studies where only Control and R021 treatments were compared [14,15]. Nonetheless, R07 does tended to have a higher number of fetuses, although with worse viability. In rabbits, the implantation and placentation events occur after day 6–7 post-AI [24]. In the current study, at that time R07 females experienced a significant compensatory increase in feed intake, surpassing the Control group. The slight increase in implanted fetuses observed in this group could lead to greater competition between them for uterine space and nutrients, ultimately leading to lower viability [25].

The dimensions of the fetuses were affected by the restriction, so that the longer it was, the more weight and length of the fetuses decreased, although the TD and the OND did not seem to have the same response. MFR usually causes vital organs, such as the brain or the liver, to increase their relative weight to the detriment of other organs [9,26]. This trend was not observed in this study, and other organs such as the kidneys and the liver reduced their weight in the fetuses of more restricted females. Several studies [12,27,28] assert that the longer and greater the feed deprivation is, the greater the probability of finding impaired fetal growth and development. Cappon et al. [29] performed an experiment evaluating MFR levels from slight to severe (110 g/d down to 15 g/d) and from days 7 to 19 of gestation (a period that registers from implantation to closure of the hard palate), recording the effect on fetal development. These authors showed that in severe feed deprivation, the size and weight of the fetuses were reduced, and the incidence of abortions and inadequate ossification increased. These results are in accordance with other studies [27,30] that evaluated the incidence of abortions and impairment of fetal development/growth due to feed restriction, where authors agreed that the main factors to be considered are the degree of intake restriction, its length, and more critically the period of gestation when it is applied [20], highlighting the second half of gestation as that where the energy intake of the mothers is most critical. The findings regarding fetal development comparing organ sizes of the fetuses between experimental groups at day 28 of gestation, suggest that the ration setting and periods of feed restriction applied have no long-term effects over the fetuses and later litter performance. In fact, in these experimental conditions, the brain/liver ratio was the same in all groups, and no critical variations in the relative weights of the fetal organs were observed due to the feed restriction, which supports the idea that the feed restriction applied is to be considered moderate compared to the existing records and does not induce severe impairment in fetal growth nor organogenesis.

In addition, MFR for 1, 2 or 3 weeks affected the growth of the placental structures. The largest placentas were those of R07 does, but this group was one of the least efficient, meaning that even though fetal growth directly depends on the placenta [31], in this group placentas were bigger than their fetuses. However, R721 does showed the smallest placenta because the mothers were restricted in the period when this tissue is developed. These results suggest that although placental efficiency was low in R07 and high in R721, the amount of energy provided during the period when does were fed ad libitum was enough to maintain gestation adequately, since does of these two groups were able to achieve a similar litter size at parturition, a high number of weaned kits (after adjustment), and the lowest lactation mortality rates. In this sense, Rommers et al. [28] described that a moderate decrease in energy intake during the first week of gestation, provided it is above maintenance needs, does not affect implantation rate, while in the fourth week of pregnancy, energy deprivation can generate more problems in the long-term.

Furthermore, during lactation and in the juvenile stage, our previous studies [14,15] that focused on the likely consequences on progeny of a three-week MFR (equal to the current R021 group) evidenced that the surviving offspring did not manifest any outstanding alterations in growth, serum or metabolic parameters nor feed intake during juvenile phase until puberty.

Regarding the effects of MFR on fertility and prolificacy of does, there were no adverse consequences. Although litter weight and milk production were similar between groups during lactation, a significant increase in kit mortality was observed in both the Control and R021 groups. It may be that litter adjustment performed after birth affected these results, but it is difficult to explain. Interestingly, does in group R07, despite previous lower fetal viability detected during pregnancy, had a better weaning outcome and lower kit mortality, outpacing the results of the other groups.

Concerning the mortality rate of the dams during the study (culled due to experimental needs or for other reasons such as low prolificacy, abortions, sudden deaths or mastitis), the data collected did not show significant differences between groups. They were consistent with the experimental farm's historical data, so these casualties can be considered incidental rather than related to the experimental conditions.

As to the effects of MFR on energy metabolism, thyroid hormones were determined to evaluate metabolic adaptation to energy intake deprivation. In the current study, the mothers that underwent a longer MFR presented lower T3 plasma concentrations. During the first three weeks of gestation, a decrease in T3 and T4 was also noted, along with an increase at the end when the females were fed ad libitum. The significant treatment × time interaction observed in T3 concentrations was due to the R021 group having the lowest T3 values on day 21 of pregnancy (after three continuous weeks of moderate restriction), but by the end of gestation all groups had similar concentrations. The decrease in free T3 during periods of feed restriction usually reduces the basal metabolic rate, resulting in energy savings for the animal [32], but of all metabolism-involved hormones, T3 is the fastest to recover [33]. Thus, in the 4th week of pregnancy and during lactation, when all groups were already fed ad libitum, energy homeostasis was balanced even in the case of T3. T4 plasma concentrations during lactation rose significantly more than those observed during gestation, indicating a total recovery of the energy metabolism after MFR.

The studied relation between insulin and glucose profiles suggests that females in the first half of pregnancy, being subjected to MFR, manage the diet's energy and body reserves. Insulin is an intermediate metabolic hormone, involved in many metabolic pathways to maintain energy homeostasis by coordinating the different axes and systems [34,35]. Furthermore, it should not be forgotten that insulin undergoes postprandial alterations and at day 14 of pregnancy the Control and R07 groups are both being fed ad libitum since day 0 and 7 of pregnancy, respectively. Afterwards, on day 21 of pregnancy, insulin increases and insulin resistance manifests with the rise in HOMA-IR in these groups, which Menchetti et al. [32] described as an adaptation to the energy demand produced by the fetuses. Following the same dynamic, decreased insulin plasma concentrations on day 21 of pregnancy in groups R721 and R021 reinforces the role of insulin in the gear of metabolism, since Buczkowska and Jarosz-Chobot [36] reported, in the same way as other authors [35], that reduced insulin plasma concentrations promotes lipolysis, which makes sense since by day 21 of pregnancy these dams had been restricted for two and three weeks, respectively, and needed to obtain energy from their body reserves. On the other hand, Fortun-Lamothe [37] described that the fetuses request high amounts of glucose at the end of gestation. Therefore, as MFR was over in R721 and R021 groups by the 4th week of gestation, these dams were obtaining glucose directly from the diet and their voluntary feed intake was considerably increased to compensate for the time of restriction, which would explain the glucose rise detected at this point, particularly in group R021, which underwent the longer MFR.

Moreover, HOMA-IR results showed significant differences between groups by the end of gestation (day 28). Particularly, R021 does displayed a critical rise in HOMA-IR values, while other groups still did not exceed the limit to be considered insulin resistant (R07 and R721). This could be due to the continuous consumption of glucose by the fetuses that forces an increase in the mother's hepatic production of glucose by the liver during fasting to maintain plasmatic glucose sufficient to meet the needs of glucose-dependent tissues [38], which would also explain the slight but not significant rise of glucose in group R021 on day 28. Since circulating NEFA concentrations are not high enough in group R021 at this point to consider body reserve mobilization, it can be assumed that the main source of glucose was energy obtained from the diet, which at this time was being provided ad libitum for all groups.

In humans and other species, the second half of pregnancy is defined by a progressive increase in resistance to the action of insulin, so that in the third-trimester insulin sensitivity is about a third of normal and insulin levels are increased by about four times [39,40], which in our study would be equivalent to the 3rd and 4th weeks of gestation. Other authors have also obtained these same results and appreciations in rabbits [41], which concur with the data obtained in this study.

As shown by the data, MFR during gestation affected the overall energetic balance of the does and high NEFA plasma concentrations are a manifestation of body reserves mobilization. In this study, MFR affected plasma NEFA concentrations, and the highest values were observed on day 14 in the mothers with a more prolonged period of restriction at this time. This is in accordance with other studies where authors determined higher NEFA plasma concentrations in does undergoing slight to moderate MFR during similar gestation periods [32,42]. Additionally, lower circulating NEFA concentrations were observed on day 14 of gestation in Control and R07 groups, which supports the previous statement and suggests a redirection of metabolism towards a reduction of lipolysis due to higher energy intake in these groups.

Conductive to the lactation period, the main difference between groups occurred with T3 concentrations in pregnant and non-pregnant females, with lower levels in the latter. Since the third AI was performed on day 14 after parturition, increased T3 in lactating-pregnant dams may be explained by its role in tissue synthesis, particularly in the first half of pregnancy [43]. Moreover, Bober et al. [44] reported that thyroid hormones are considered necessary for cellular metabolism of the mammary gland and energy utilization, which may be an important factor in fetal development during pregnancy and milk biosynthesis, which is relevant for our study since pregnant does are overlapping lactation and 3rd pregnancy. Highlighting the slight insulin and HOMA-IR decrease that occurred on day 14 compared to day 7 post-partum, this could be explained by the stressful event of AI, which implies manipulation of the animals (removing the females from their cages, holding them by the tail, inserting the insemination catheter and injecting GnRH). When stressful situations occur, insulin levels fall, glucagon and epinephrine (adrenaline) levels rise and more glucose is released from the liver [45].

Surprisingly, no differences due to MFR between experimental groups were observed during pregnancy and lactation, something unexpected since corticosterone was assessed as a hormonal measure of stress in the animals. Some authors contend that plasma corticosterone concentrations correspond more with metabolic stress than with behavioral stress [33], which would support the assertion that in this study, MFR is only moderate and implies no metabolic depletion of the dams, since corticosterone was lower by the end of gestation. On the other hand, Menchetti et al. [32] measured cortisol levels during the pregnancies of does subjected to MFR and determined that this parameter was not altered during gestation, which can be attributed to changes in the body's regulation during pregnancy that favor the maintenance of glucocorticoid levels, which is essential for the correct maturation of many fetal organs [46–49].

However, it is interesting to note that at the end of lactation, corticosterone levels are higher than those on day 7 of lactation. This could be explained due to the size of the kits towards the end of lactation. Kits begin to leave the nest at 14–16 days of age [50], and other authors have described that higher housing density could increase corticosterone circulating levels [51]. In these experimental conditions, by the end of lactation there are approximately 10 kits per doe weighing around 500 g, which leave the nest and occupy a lot of space in the cage, try to continue suckling and can increase the mother's level of discomfort and explain the rise in corticosterone. Nonetheless, further studies are needed to corroborate this hypothesis.

#### **5. Conclusions**

The present study states that a moderate MFR (60% of the estimated voluntary feed intake) applied during 1, 2 or 3 weeks of gestation followed by ad libitum refeeding confirms previous results regarding compensatory food intake by mothers, which helps to preserve maternal energy homeostasis (thyroid hormones, glucose and insulin) and lipid metabolism (NEFA). This allowed the necessary energy supply to be delivered to the fetuses to conduct fetal growth and development satisfactorily, confirming previously obtained results as no major long-term effects in the early life development of the offspring were observed. Moreover, the compensatory feed intake granted the maintenance of an adequate productive performance in the current pregnancy and showed no effect whatsoever in the amount of milk produced, the viability of litters or the mortality rate of the dams. In addition, the corticosterone concentrations observed throughout pregnancy and lactation seem to indicate that, in our experimental conditions, the moderate MFR applied seems not to induce additional stress in dams along the gestation and lactation periods studied. Moreover, it has proven that this strategy can be profitable for the farmer as overall feed intake was lower in all experimental groups compared to the Control group. In summary, these results reveal that as long as the feed supply covers maintenance needs, moderate MFR is a suitable strategy to reduce production costs, since it implies no adverse effects on the metabolic, energetic and welfare status of the dams, and on pre- and post-natal growth of fetuses and litters, respectively.

**Author Contributions:** Conceptualization, P.L.L., M.A.-Á., R.M.G.-G. and P.G.R.; data curation, M.R. and P.G.R.; formal analysis, C.F.-P., P.M., M.R. and P.G.R.; funding acquisition, P.L.L., M.A.-Á., R.M.G.-G. and P.G.R.; investigation, C.F.-P., P.M., M.R., N.F.-R., A.S.-R., M.A.-Á., R.M.G.-G. and P.G.R.; writing—original draft, C.F.-P., P.M. and P.G.R.; writing—review and editing, C.F.-P., P.M., M.R., N.F.-R., A.S.-R., P.L.L., M.A.-Á., R.M.G.-G. and P.G.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the SPANISH MINISTRY OF ECONOMY AND COMPETI-TIVENESS (Grant Number: AGL2015-65572 and RTI2018-094404-B).

**Institutional Review Board Statement:** The study was conducted according to the guidelines of The Declaration of Helsinki and approved by the Ethics Committee of the TECHNICAL UNIVERSITY OF MADRID (protocol code PROEX 302/15 approved in 2015).

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** We thank B. Velasco for her assistance with the animal husbandry and maintenance.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

