Author's address (for correspondence): Romy Hribal, Leibniz-Institute for Zoo and Wildlife Research (IZW), Department for Reproduction Biology, PF 700430, 10324 Berlin, Germany. E-mail: email@example.com
Early embryos are characterized by an accurately controlled gene expression pattern that might be deregulated during in vitro culture (IVC). The expression pattern of the developmental genes may serve as markers for embryo quality. Here, we examined the temporal pattern of relative mRNA abundance of genes important for early embryonic development in embryos produced by different fertilization methods [in vitro fertilization (IVF) vs intracytoplasmic sperm cell injection (ICSI)] and sperm sources (fresh vs frozen–thawed) applying reverse transcriptase (RT) PCR. The temporal pattern of gene expression was found to be gene specific and similar in all four examined groups in a semi-quantitative assay. In morulae, higher relative mRNA levels were found in embryos generated with fresh sperm, whereas in blastocysts, mRNA abundance tended to be higher in embryos produced with cryopreserved sperm cells. This indicates an influence of sperm cryopreservation on the temporal gene expression pattern in early cat embryos. We also examined relative mRNA abundances by real-time quantitative RT-PCR in blastocysts. In this context, blastocysts produced with fresh semen tended to have lower DNA methyltransferase 3A (DNMT3A) but higher gap junction protein alpha 1 (GJA1) and octamer-binding transcription factor 4 (OCT4) mRNA levels compared with those derived with frozen–thawed semen. We conclude that assessing embryo quality by measuring gene expression pattern in early embryos is challenging because of a high variability between individual embryos.
In addition to morphology, cleavage rates and kinetics, the expression of specific genes important for embryo development is used as a tool to evaluate embryo quality. These methods, mainly applied in bovine embryos, were proven to be far more sensitive and revealed differences even between embryos that did not differ with respect to conventional quality markers (Wrenzycki 2007). Aberrations in gene expression pattern, associated with in vitro methods, such as in vitro maturation (IVM), IVF and IVC of embryos have been demonstrated (Rizos et al. 2002; Lonergan et al. 2003; Jones et al. 2008; Heinzmann et al. 2011). In cats, only few studies have been carried out to examine early embryonic gene expression pattern. OCT4 expression was reduced in in vitro-produced embryos compared with their in vivo counterparts (Filliers et al. 2012). Expression of genes involved in the maintenance of pluripotency, histone acetylation and DNA methylation was different in cloned embryos compared with in vitro-derived embryos (Imsoonthornruksa et al. 2010). OCT4 expression was decreased in cloned blastocysts produced with frozen–thawed donor fibroblast cells compared with fresh cells or IVF embryos (Gomez et al. 2008). OCT4 expression was furthermore influenced by the culture medium composition used for the production of cat embryos (Hribal et al. 2012). The impact of ICSI and sperm cryopreservation, both long-term applied methods in the domestic cat (Pope et al. 1998; Ringleb et al. 2011), on expression of specific genes at morula stage has been examined previously (Waurich et al. 2010). Morulae produced by IVF with fresh sperm tended to have higher mRNA expression levels compared with those produced by IVF with frozen–thawed semen as well as those generated by ICSI with fresh or frozen–thawed semen. In the present study, we aimed to examine the influence of both methods on the temporal pattern of mRNA abundance not only at the morula stage but throughout early cat embryo development. The expression of developmental important genes was examined by semi-quantitative reverse transcription polymerase chain reaction (RT-PCR). In addition, gene expression was examined by real-time reverse transcription quantitative PCR (RT-qPCR) in blastocysts produced either with fresh or frozen–thawed semen to confirm the pattern found by semi-quantitative PCR.
Two experiments were performed in this study
Experiment 1: Embryos were produced by IVF with fresh epididymal semen, by IVF with frozen–thawed semen, by ICSI with fresh semen as well as ICSI with frozen–thawed semen. Pools of 20 two-cell, 10 four-cell, five 8–16-cell embryos, two morulae (n = 6 replicates) as well as single blastocysts (n = 6 replicates for the two IVF groups, n = 2 for ICSI with fresh semen, n = 1 for ICSI with frozen–thawed semen) were used for RNA extraction in each experimental group. Gene expression was determined by semi-quantitative RT-PCR. In a previous paper (Waurich et al. 2010), we showed the data for the temporal pattern of relative mRNA abundances in embryos produced by IVF with fresh semen as well as relative mRNA abundances at the morulae stage for the four experimental groups. Here, we complemented the temporal pattern of gene expression for all four experimental groups.
Experiment 2: Blastocysts were produced by IVF with fresh and frozen–thawed epididymal semen (n = 6 replicates for IVF with fresh semen, n = 7 replicates for IVF with frozen–thawed semen) and used individually for RNA extraction. Gene expression was examined by RT-qPCR.
Materials and Methods
In vitro production of domestic cat embryos
All chemicals were purchased from Sigma Aldrich (Taufkirchen, Germany) unless stated otherwise. Ovaries and testes were obtained from cats neutered in a local veterinary clinic. Transport and storage of gonads occurred as reported previously (Waurich et al. 2010). Ovaries were processed immediately after arrival at the laboratory. Testes were stored for 24 h at 4°C.
For experiment 1, processing of ovaries, IVM, semen preparation, IVF or ICSI as well as embryo culture were performed as reported previously (Waurich et al. 2010; Ringleb et al. 2011).
For experiment 2, embryos were produced as described for experiment 1 with a single exception. Embryo culture was performed in Ham's F10 (Modified Ham's F10 Basal Medium; Irvine Scientific, Medical Technology Vertriebs GmbH, Bruckberg, Germany), supplemented with 5% FBS (v/v), 1 mm sodium pryruvate, 1 mm l-glutamine, 0.1 mg/ml streptomycin and 100 IU/ml penicillin (Comizzoli et al. 2003). Day-7 blastocysts produced for experiment 2 were washed three times in Dulbecco's PBS and stored in 5 μl of RNA later (Qiagen, Hilden, Germany) at −80°C.
Determination of relative mRNA abundances
For experiment 1, relative mRNA abundance of selected genes was determined by semi-quantitative endpoint RT-PCR as previously reported (Waurich et al. 2010). In brief, total RNA was isolated with RNeasy Micro Kit (Qiagen) after addition of 1 pg rabbit globin RNA (non-felid standard RNA) to each sample. Reverse transcription into cDNA was carried out with RevertAid™ First Strand cDNA Synthesis Kit (Fermentas GmbH, St. Leon-Rot, Germany) in a total volume of 40 μl.
Amplification occurred in a G-Storm GS1 Thermocycler (Gene technologies Ltd., Essex, UK) in a total volume of 25 μl using the Fast Start Taq Polymerase dNTP Pack (Roche Diagnostics GmbH, Mannheim, Germany). PCR products were stored at −20°C. The sequences of specific primers as well as annealing temperatures and fragment lengths have been described previously (Waurich et al. 2010) and are provided in Table S1.
Amplified specific PCR products were analysed genewise after gel electrophoresis. The relative mRNA amounts were calculated by dividing the specific fragment gel band intensity of each template by the intensity of the corresponding globin band. Subsequently, values were divided by the number of embryos pooled for extraction so that data could be shown as values per single embryo. For signal intensity measurement, imagej (National Institutes of Health, http://rsbweb.nih.gov/ij/) was used.
For experiment 2, relative mRNA abundance was measured by RT-qPCR after RNA extraction and cDNA synthesis from single day-7 blastocysts produced by IVF with fresh or frozen–thawed semen. Analysis was performed as described for experiment 1 followed by RT-qPCR in a CFX96 cycler (Bio-Rad Laboratories GmbH, München, Germany). The gene products were amplified using the SsoFast EvaGreen Supermix (Bio-Rad Laboratories GmbH). For each gene, 4 μl of template (diluted 1:2 in sterile, filtered Aqua bidest with exception for amplification of GJA1, here undiluted template cDNA was used) was added to 5 μl of Supermix and 0.5 μl of forward and reverse primer (final primer concentration 500 nm) each (Table S1). The PCR programme consisted of an initial heating step for 2 min at 98°C, 40 cycles of 8 s at 98°C, 8 s at primer-specific annealing temperatures and 5 s at 65°C. To confirm the presence of a single amplification product and the absence of primer dimers, melting curve analysis was performed from 65°C to 95°C in 0.5°C steps each lasting 5 s.
Amplification was performed in duplicate, Aqua bidest was used instead of cDNA in non-template controls to check for contaminations. To generate a standard curve for each gene (from 101 to 106 molecules per reaction), a recombinant linearized plasmid DNA derived from cloned cDNA was created. Normalization was performed by applying again rabbit globin as exogenous internal standard. Relative mRNA abundances were calculated by dividing the quantities of each transcript by the corresponding globin quantities.
Relative mRNA abundance in blastocysts produced by IVF with either fresh or frozen–thawed semen in both experiments was compared by Mann–Whitney U-test (PSAW Statistics 18.0; SPSS Inc., IBM Corporation, NY, USA).
The temporal pattern of relative mRNA abundance from the two-cell stage until blastocyst stage, measured by semi-quantitative PCR, is shown in Fig. 1. Compared were embryos produced by IVF or ICSI using fresh or frozen–thawed semen. Insulin-like growth factor 1 receptor (IGF1R) expression remained low over all developmental stages in all four examined groups (Fig. 1a). Insulin-like growth factor 2 receptor (IGF2R) mRNA abundance remained low during the development from the two-cell until the morula stage in all four groups (Fig. 1b). The blastocyst was characterized by a dramatic increase. The levels were higher in blastocysts produced with frozen–thawed semen than in those generated with fresh semen. DNA methyltransferase 1 and 3A (DNMT1 and DNMT3A) levels increased from the two-cell stage to the four-cell stage, then stagnated until the morula stage and increased in blastocysts (Fig. 1c, d). Levels were also higher in embryos produced by fertilization with frozen–thawed semen compared with those generated with fresh semen. Octamer-binding transcription factor 4 and beta-Actin (OCT4 and ACTB) showed a different pattern, with a first increase from two-cell stage to four-cell stage, followed by a doubling of mRNA amounts in morulae and another massive increase in blastocysts (Fig. 1e,f). Comparison between the groups revealed highest levels in morulae produced by IVF with fresh semen. In blastocysts, however, embryos generated by fertilization with frozen–thawed sperm displayed higher levels compared with those produced with fresh semen (both for IVF and ICSI, Fig. 1e,f). Gap junction protein alpha 1 (GJA1) levels increased slightly during development to morula stage and remained constant or even were decreased in the blastocyst (Fig. 1g). This transcript was not detected in 8–16-cell embryos produced by ICSI with frozen–thawed sperm. Differences between embryos produced by IVF with fresh or frozen–thawed sperm in the blastocyst stage were not significant for all examined genes.
Figure 2 shows relative mRNA abundance in blastocysts produced by IVF with either fresh or frozen–thawed semen measured by RT-qPCR. There were no significant differences between the two groups. DNMT3A expression tended to be higher in blastocysts produced by IVF with frozen–thawed semen. GJA1 and OCT4 mRNA abundances were lower in embryos after IVF with frozen–thawed semen compared with those produced with fresh semen.
Sperm cryopreservation and ICSI are both commonly applied methods, and despite their invasiveness, the influence on relative gene expression in early embryos was, to our knowledge, never tested before apart from our recent evaluation (Waurich et al. 2010). In that study, the impact of both methods on relative mRNA abundance was examined at morula stage, the stage where the developmental block occurs in cat embryos. The expression levels of most genes that were analysed was found to be intensified during or shortly after that specific time (Waurich et al. 2010). Although the examination of gene expression at typical developmental stages such as the morula or blastocyst is appropriate to compare embryos of different origins, the temporal pattern of relative mRNA abundance, typically underlying accurate regulation, could be also affected by in vitro methods. For all four experimental groups, we found a gene-specific expression pattern, which was characterized by an increase in relative mRNA abundances from the two-cell stage onwards (with exception of IGF2R). This early onset of embryonic gene expression, the minor genomic activation (Wrenzycki 2007), was not influenced by ICSI or sperm cryopreservation. Also for later stages of development, no significant differences between the experimental groups were found. Previously, a trend towards higher mRNA levels was detected in morulae produced with fresh semen (Waurich et al. 2010). In the present study, gene expression at blastocyst stage seemed also to be influenced by the sperm source. By the semi-quantitative mRNA availability assessment, blastocysts generated with frozen–thawed semen tended to have higher mRNA expressions (for DNMT1, DNMT3A, OCT4, IGF2R and ACTB) thus converting the results obtained in morulae. It might be suggested that frozen–thawed spermatozoa influence the temporal pattern of gene expression in early cat embryos with a delayed onset of gene activation.
We suspected that RT-qPCR, a far more sensitive method, will confirm the results obtained by the semi-quantitative PCR assay, and even more, should more faithfully demonstrate the differences between the experimental groups. However, the results obtained after RT-qPCR did not confirm the findings obtained with the semi-quantitative approach. We can only speculate about the reasons for those differing results, which include different measurement systems (semi-quantitative endpoint RT-PCR vs RT-qPCR), different chemicals and equipment and, last but not least, a different embryo culture medium applied within the second experiment. In opposite to the semi-quantitative PCR method, quantitative PCR allows the exact identification of the exponential amplification phase. Therefore, a measurement of expression levels at the exponential amplification phase is guaranteed, whereas a semi-quantitative PCR assay is always an endpoint measurement. For this reason, RT-qPCR is more accurate compared with semi-quantitative PCR. Although many studies have employed semi-quantitative measurements that were sensitive enough to detect differences in relative mRNA abundance (for example, Nowak-Imialek et al. 2008), RT-qPCR has become the method of choice also for very tiny sample sizes, such as single embryos (Filliers et al. 2012). However, the procedure of expression normalization varies and merits examination. In some studies, single housekeeping genes were used for normalization (Imsoonthornruksa et al. 2010; Gomez et al. 2011), although also housekeeping genes were expressed in a regulated pattern (Robert et al. 2002) and might be influenced by in vitro conditions. Filliers et al. (2012) validated a set of reference genes, whose expression was found to be stable. The disadvantage of that method is that much of the valuable sample material has to be used for reference gene amplification and fewer genes of interest can be analysed. For example, in our laboratory, cDNA from one single blastocyst is appropriate for the maximum amplification of nine or ten different genes by our RT-qPCR system. With at least four reference genes, only six genes of interest can be examined within the same sample. Given the rare availability of feline embryos, it definitely makes sense to examine as many developmental relevant transcripts as possible. We used rabbit globin mRNA as an exogenous internal control for normalization of RNA extraction, cDNA synthesis as well as amplification, because this has been proven as a reliable internal control for bovine embryos (Nowak-Imialek et al. 2008; Heinzmann et al. 2011). The greatest disadvantage of rabbit globin is the missing control for the number of cells within an embryo, which might have caused the high individual variations between single embryos. In future studies, mRNA abundance should be corrected for the number of blastomeres.
For the second experiment, we applied a different combination of culture media that proved to allow more efficient development of blastocysts. The alteration in the procedure could be responsible for the diverging results, as it was shown previously that the culture medium composition affects relative mRNA abundance in cat embryos (Hribal et al. 2012). It can be hypothesized that embryos produced in Ham's F10 are able to compensate negative effects caused by the sperm source better.
In conclusion, the examination of relative gene expression pattern in early embryos is a promising tool to evaluate environmental influences and to control for embryo quality. Among the challenging aspects that remain are high variations between individual embryos, demanding constant experimental conditions (media composition, sperm source). Our results indicate that different developmental genes follow different activation patterns and that early cat embryos express a broad plasticity to react to their environment.
We thank the animal shelter of Berlin for providing cat gonads and Stephan Karl for technical assistance.
Conflicts of interest
The authors declare that there are no conflicts of interest that could be perceived as prejudicing the impartiality of the research reported.
The project was funded by the Leibniz-Gemeinschaft (SAW-2011-IZW-2).
All four authors contributed to the study design. Romy Hribal and Jennifer Ringleb produced the embryos examined in this study. Expression analysis was done by Romy Hribal with help of Beate C. Braun. The paper was drafted by Romy Hribal and Katarina Jewgenow, with comments by Beate C. Braun and Jennifer Ringleb.