Hum. Reprod. Advance Access published online on May 13, 2008
Human Reproduction, doi:10.1093/humrep/den123
Novel strategy with potential to identify developmentally competent IVF blastocysts
1 Monash Immunology and Stem Cell Laboratories (MISCL), Monash University, Level 3—STRIP Building 75, Wellington Road, Clayton, Victoria 3800, Australia 2 Monash IVF, Clayton, Victoria 3168, Australia 3 Centre for Human Reproduction, Genesis Athens Hospital, Athens 15232, Greece 4 Laboratory of Medical Genetics, Athens University, Athens 11527, Greece
5 Correspondence address. Tel: +61 3 9905 0778; Fax: +61 3 9905 0680; E-mail: david.cram{at}med.monash.edu.au
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
BACKGROUND: Currently there are no markers fully predictive of developmental competence of human IVF embryos. The present study investigated a novel strategy involving blastocyst biopsy and DNA fingerprinting to link developmental competence with gene expression patterns.
METHODS: Patient’s blastocysts were biopsied to remove 8–20 trophectoderm (TE) cells for molecular analysis prior to transfer. Biopsy samples were amplified and gene expression was evaluated using microarrays. Sibling TE biopsies and cells from resulting offspring were subjected to DNA fingerprinting to identify which blastocyst(s) in the transfer cohort developed to term.
RESULTS: Blastocyst biopsy did not appear to impair developmental competence. Comparative microarray analysis of cDNA from pooled ‘viable’ and ‘non-viable’ TE samples identified over 7000 transcripts expressed exclusively in ‘viable’ blastocysts. The most significant of these included transcripts involved in cell adhesion and cell communication, key processes that have been associated with mammalian implantation. DNA fingerprinting of three cohorts of sibling blastocysts identified those blastocyst(s) that produced term pregnancies.
CONCLUSIONS: The combination of blastocyst biopsy, microarray gene expression profiling and DNA fingerprinting is a powerful tool to identify diagnostic markers of competence to develop to term. This strategy may be used to develop a rapid diagnostic assay or for refining existing criteria for the selection of the single most viable blastocyst among a cohort developing in vitro.
Key words: blastocyst biopsy/DNA fingerprinting/microarrays/developmental competence/implantation
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Assisted reproductive technology (ART) developed over 30 years ago (Steptoe and Edwards, 1978
) is now a well-established treatment for human infertility. Despite numerous advances in the practice of ART in the past three decades, implantation rates following the transfer of in vitro produced embryos have remained <30% (Edwards and Beard, 1999). Elective single-embryo transfer has been employed in several IVF clinics worldwide, with implantation rates approximating 50%. However comparative trials between elective single- and elective double-embryo transfers have been conducted in selected patient groups, and the success of live-birth is only similar when transfer of frozen embryos in subsequent frozen–thaw cycles is considered (Bergh, 2005; Pandian et al., 2005). Therefore the identification of the single most viable embryo among a cohort available for transfer is the ultimate aim to achieve a live-birth within the fresh IVF cycle.
Implantation failure is believed to be due to a range of factors including chromosomal abnormalities which are common in the female (15–61% of oocytes and embryos generated for ART) (Munne et al., 1998; Clouston et al., 2002; Plachot, 2003) and typically further increased in women of advanced reproductive age (Munne et al., 1995; Kuliev et al., 2003; Pellestor et al., 2003; Rosenbusch, 2004), asynchrony between embryo development and uterine receptivity (Norwitz et al., 2001; Margalioth et al., 2006) and factors associated with treatment interventions including the potential adverse effects of in vitro culture conditions (Jones, 2000) and the technique of embryo transfer (Levi Setti et al., 2003; Margalioth et al., 2006). However, a large proportion of the failed implantations must be ascribed to the embryo, as it is not uncommon for a single implantation to result from the transfer of two or more embryos of good morphological quality.
Despite numerous informative correlations for morphological, metabolic or biochemical criteria, no accurate predictive markers of the embryo’s developmental capacity to term have been revealed. Morphological criteria are routinely used to select the ‘best’ embryos for transfer and these criteria have taken on greater importance, as the IVF community is moving towards a single embryo transfer policy (Bergh, 2005; Pandian et al., 2005). These criteria include pronuclear morphology and orientation at 17–19 h post-insemination (Scott and Smith, 1998; Tesarik and Greco, 1999), early cleavage at 25–27 h post-insemination (Shoukir et al., 1997; Sakkas et al., 2001), scoring systems for Days 2–3 embryos that incorporate embryo cleavage rates and synchrony of cleavage as well as the degree and pattern of fragmentation (Cummins et al., 1986; Puissant et al., 1987; Steer et al., 1992; Giorgetti et al., 1995; Alikani et al., 1999) and grading systems for blastocysts that incorporate the timing of blastocyst formation, the degree of expansion of the blastocyst and the appearance of the inner cell mass and trophectoderm (TE) cell lineages (Dokras et al., 1993; Schoolcraft et al., 1999). These criteria are currently the most predictive markers of embryo developmental competence and are widely used in many IVF laboratories. However these criteria are not fully predictive of an embryo's ability to develop to term in a receptive uterus, and combinations of the various scoring systems do not significantly improve the predictive value. Metabolic studies have found that both glucose uptake and glycolytic activity are unsuitable markers of blastocyst viability because the specific blastocysts that implant show a wide variance of metabolic activities (Jones et al., 2001). Biochemical markers of blastocyst viability such as CD30 ligand (Pera et al., 1997), LIF (Cullinan et al., 1996) and matrix metalloproteinases MMP-2 and MMP-9 (Turpeenniemi-Hujanen et al., 1995) have also been examined but these systems have lacked the required sensitivity to detect small quantities of these molecules in the spent embryo culture media. More recently, chromatographic assays of amino acid consumption (Brison et al., 2004), secretion of soluble human leukocyte antigen (HLA)-G molecules (Sher et al., 2004) and spectroscopic analysis of spent culture media (Katz-Jaffe et al., 2006; Brison et al., 2007) show some promise as a means of assessing embryo viability non-invasively but these methods are technically difficult and not sufficiently developed for routine application.
The power of all these observations has been frequently confounded by the practice of transferring multiple embryos which does not enable identification of the actual embryo responsible for the establishment of pregnancy. DNA fingerprinting is the ultimate form of biological identification but has not been used for the purposes of identifying the embryo(s) among a cohort transferred that is responsible for the pregnancy. Informative DNA fingerprints from low-template DNA samples can now be derived either by the analysis of highly polymorphic short tandem repeats (STRs) with high heterozygosity indices or using a large panel of single-nucleotide polymorphisms (SNPs). To aid pre-implantation genetic diagnosis (PGD) of single-gene disorders, linked STR markers have been used extensively to identify allele dropout (ADO), monitor external DNA contamination and indirectly track the inherited affected and non-affected parental alleles in biopsy samples from both cleavage and blastocyst stage embryos (Sermon et al., 2004; Kokkali et al., 2005). Furthermore, DNA fingerprinting of single blastomeres isolated from aneuploid cleavage-stage IVF embryos using a minimum of four STR markers has been shown to discriminate between sibling embryos (Katz et al., 2002). Hence, the application of DNA fingerprinting at the pre-implantation embryo stage and at birth could potentially identify the specific embryo(s), in a multiple transfer cohort, that is responsible for the establishment of pregnancy.
Pre-implantation development of mammalian embryos is governed by maternal transcripts until the time of zygotic genome activation. Recently, it has been demonstrated in the mouse that there are two independent and distinct waves of transcription. One which occurs at zygotic genome activation (one- to two-cell stage) and one which occurs at the morula-blastocyst transition (Hamatani et al., 2004; Wang et al., 2004). Therefore molecules important for implantation events may only be detectable at the blastocyst stage of development, precluding relevance of data from early cleavage stage embryos or blastomeres biopsied from these embryos. Molecular and immunohistochemical studies in humans, primates and laboratory animal species, together with data from experimental animal models such as gene knockout mice, have identified candidate growth factor ligands/receptors and adhesion molecules that are likely to play a key role in implantation and early placentation (Aplin, 1997). However, much of the data have been accumulated from studies of the endometrium, not the blastocyst, and the information regarding molecules involved in the early stages of blastocyst apposition and attachment for implantation is virtually non-existent (Simon et al., 2000). Even less information is available on the expression of molecules by human peri-implantation embryos. Further, although there are some published data on early gene expression through the molecular analysis of whole-human pre-implantation embryos (Adjaye et al., 1999, 2005; Holding et al., 2000; Dobson et al., 2004; Wells et al., 2005; Li et al., 2006), the data are difficult to relate to viability and implantation competence because the embryos were never transferred for confirmation.
The observed variance in viability among a cohort of a patient’s embryos suggests that some in vitro cultured embryos are not receiving adequate autocrine and paracrine signalling to achieve full implantation and developmental competence. We hypothesize that the reduced viability of otherwise morphologically good quality ART embryos is a result of a transcriptome that is not compatible with implantation and/or growth to term. The present study describes a novel strategy that utilizes a combination of blastocyst biopsy (Kokkali et al., 2005, 2007), DNA fingerprinting (Katz et al., 2002) and microarray mRNA analysis (Jones et al., 2007) to identify viable blastocysts among the cohorts transferred to patients, with the ultimate aim of deriving a gene expression profile associated with blastocyst developmental competence to a healthy baby at term. This is the ultimate objective of ART and the infertile couples seeking treatment for infertility.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
ART treatment
Patients undergoing infertility treatment at the Centre for Human Reproduction, Athens, Greece in 2003 whose ovarian ultrasound scan indicated the development of a large number of follicles in response to controlled ovarian hyperstimulation (Pantos et al., 1994
) were approached to provide informed consent to participate in this study. Patients consented to transfer at the blastocyst stage of development with TE biopsy prior to transfer. The study was approved by the First Regional Health and Welfare System of Attica (PESYAT), Department of Research and Development Ethics Committee. Molecular analysis of the RNA generated from biopsied material was approved by the Monash University Standing Committee on Ethics in Research Involving Humans, Clayton, Australia (SCERH 2005/272ED).
Following their retrieval, oocytes were cultured and fertilized as described previously (Jones et al., 1998). According to the policy of the Centre for Human Reproduction, if a patient has more than 15 zygotes, a proportion are frozen at the pronucleate stage using a slow cooling protocol (Testart et al., 1986) for use in a subsequent cycle if required. Zygotes were cultured to Day 3 in 10 µl microdrops of G1.3 culture medium (Vitrolife, Goteborg, Sweden) and then transferred to 20 µl microdrops of G2.3 (Vitrolife) for culture to the blastocyst stage (Jones et al., 1998) under standard incubation conditions.
TE biopsy
Early on Day 5 after oocyte retrieval, embryos were assessed for blastocyst development. TE biopsy of approximately 8–20 cells was performed on all blastocysts regardless of the degree of expansion as described previously (Kokkali et al., 2005). In brief, this involved making a small hole in the zona pellucida opposite the inner cell mass using a non-contact laser (Zilos-tk, Hamilton Thorne Biosciences, Beverley, MA, USA) and allowing the TE to herniate through the hole during a period of 1–4 h pre-incubation. Herniating TE was aspirated into an embryo biopsy pipette and TE was teased out of the zona opening using a combination of suction and gentle mechanical pulling. Once 8–20 TE cells had been aspirated into the biopsy pipette, the TE was dissected free of the blastocyst mass by laser ablation (ZILOS-tk). All TE biopsy samples were immediately transferred to PCR tubes containing lysis buffer (0.8% Igepal, MP Biomedicals, Co., Irvine, CA, USA; 5 mM DTT, Gibco BRL, Invitrogen Corporation; 1 U/µl rRNAsin, Promega Corporation, Madison, WI, USA) and frozen at –80°C to preserve the DNA/RNA fraction. Biopsied blastocysts were transferred to fresh microdrops of G2.3 and cultured singly overnight prior to transfer on Day 6 after oocyte retrieval. All biopsied blastocysts were re-assessed on Day 6 prior to transfer and 1 to 4 blastocysts were selected for transfer on the basis of the number of blastocysts available, the medical history and the patient’s age and wishes and in accordance with the assisted reproduction regulations in Greece in 2003. It should be noted that the number of blastocysts transferred was determined by consultation between doctor and patient and was not dictated by participation in the research study. Biopsied blastocysts were selected for transfer in preference to non-biopsied blastocysts. Blastocysts were transferred to the patient using a Wallace catheter and remaining blastocysts were frozen (Jones et al., 1998).
Clinical outcomes
Pregnancy was determined by serum βhCG levels 
50 IU/l on Day 16 after oocyte retrieval and confirmed by demonstration of a gestational sac by ultrasound scan performed 4 weeks after embryo transfer. Pregnancy was defined as the presence of a fetal heart beat beyond 7 weeks gestation. Pregnancies resulting from biopsied blastocysts and the normality of fetal development were confirmed in follow-up ultrasounds during the first and second trimesters. Where possible, cord blood or a cheek buccal cell sample was kept from all resultant offspring for the purposes of identifying which of the transferred sibling blastocysts developed to term.
Isolation of genomic DNA
Genomic DNA was purified from offspring cell samples using a Qiagen DNA Extraction Kit (Qiagen). Genomic DNA samples were sent to Melbourne on dry ice and stored at –80°C prior to molecular analysis.
Generation of TE cDNA libraries
TE biopsy lysates were shipped to Melbourne on dry ice and stored at –80°C prior to molecular processing. Exponential or linear amplification of the poly A+ fraction to generate T7-tagged cDNA was performed as described previously (Jones et al., 2007). For exponential amplification, single biopsy samples, or samples pooled to represent eight biopsied blastocysts, were subjected to first- and second-strand cDNA synthesis using the SMART III cDNA Library Construction Kit (Clontech, Mountain View, CA, USA) according to the manufacturer’s protocol. For linear amplification, an aliquot of 200 µl of Picopure extraction buffer (Arcturus Bioscience Inc., Mountain View, CA, USA) was added to a 40 µl pooled sample representing eight biopsied blastocysts and cDNA generated following two rounds of synthesis using the RiboAmp HS RNA Amplification Kit (Arcturus Bioscience Inc.).
Microarray analysis
Biotinylated cRNA for microarray analysis was generated from T7-tagged cDNA by in vitro transcription using the Codelink Expression Assay Reagent Kit (GE Healthcare Biosciences, Piscataway, NJ, USA) and purified using a QIAGEN RNeasy Mini Kit (Qiagen GmbH, Hilden, Germany). The quantity and purity of the cRNA were determined by UV spectrophotometry at 260 and 280 nm. A total of 10 µg of labelled cRNA was hybridized to CodeLinkTM Human Whole Genome Bioarrays printed with discovery probes for 54 840 transcripts (GE Healthcare Biosciences) (Jones et al., 2007). Following hybridization and washing, arrays were scanned on a GenePix Array Scanner (Molecular Devices Co., Sunnyvale, CA, USA) and images were captured as TIF files. Codelink Expression Analysis version 4.2 software (GE Healthcare Biosciences) was used to analyse images for each slide. Spots with intensities below that of the negative control (absence of an oligonucleotide probe) were excluded, as were those with irregular shapes or near-background intensity or oligonucleotides masked as part of the quality control process during manufacture. Spot quality and signal intensities were exported to Genespring compatible report format. The Codelink Expression Analysis output was loaded into Genespring GX 7.3.1 (Agilent Technologies, Santa Clara, CA, USA) and values below 0.01 were set to 0.01 with per chip normalization to the 50th percentile and per gene normalization to the median and these data were used for all comparisons and analyses. A principal components analysis using Genespring software based on all genes was applied to all microarrays generated from individual TE biopsies to identify any samples that showed significant variability and to determine discrimination between groups of blastocysts. Genespring GX 7.3.1 Bioscript Library 2.2 Biological Pathways analysis was used to identify the probes associated with particular GO biological processes represented on the microarrays that were significantly over-represented at the 5% level within the list of probes identified as being expressed uniquely by viable blastocysts.
DNA fingerprinting
Fluorescent PCR (F-PCR) based DNA fingerprinting was performed using either the TE cDNA library or a fraction of the TE cell biopsy lysate as a starting template. The DNA fingerprint from each sibling blastocyst was matched to the DNA fingerprint(s) generated from the cell sample of any resulting offspring to identify the specific blastocyst(s) that implanted and developed to term.
For cDNA fingerprinting, genes that were expressed by the TE of a single pool of eight blastocysts known to have implanted (classified as viable) and a single pool of eight blastocysts from young patients who had undergone few or no previous assisted reproduction cycles with tubal disease or male factor as the aetiology of infertility that failed to implant (classified as non-viable) were analysed to identify genes expressed in common. These genes were screened against the H-Invitational Database (www.h-invitational.jp) to identify cDNAs containing STR sequences suitable for analysis. A panel of 24 cDNAs that were expressed at high levels in both microarrays were identified as potential candidates for cDNA fingerprinting. Oligonucleotide primers spanning the polymorphic repeat regions of each gene (Table I) were designed specifically to exonic sequences so that they could amplify the equivalent region in genomic DNA from any resulting babies.
|
For genomic DNA fingerprinting, half the volume of the TE cell lysate was subjected to whole-genome amplification (WGA) using Genomiphi V2 DNA amplification kit (GE Healthcare). The volume of the TE cell lysate was reduced to 1 µl using a vacuum concentrator. The resulting lysate was mixed with 9 µl of Genomiphi sample buffer, denatured at 95°C for 3 min and then cooled on ice. Following the addition of 9 µl of reaction buffer and 1 µl of enzyme mix, the WGA reaction (20 µl) was subsequently incubated at 30°C for 2 h and then inactivated at 65°C for 10 min and stored at –20°C prior to F-PCR analysis. DNA fingerprinting of the WGA product was based on the analysis of a panel of chromosomal STR markers (Table I).
F-PCR and genescan analysis
F-PCR amplification was performed in 25 µl of PCR reaction mix containing 2.5 µl of 10x Taq PCR buffer (100 mM Tris–HCl, pH 8.3, 500 mM KCl and 25 mM MgCl2), 0.5 µl of 10 mM dNTP, 0.5 µl Taq polymerase (5 U/µl) (Roche Applied Science, Indianapolis, IN, USA), 18.5 µl molecular biology grade water (Scientifix P/L, Australia), 1 µl of each of 10 µM 6-FAM labelled forward primer and 10 µM unlabelled reverse primer (Sigma-Proligo, Australia) and either 2 µl of the TE cDNA library or 2 µl of the WGA product as the DNA template. A total of 35 thermal cycles were performed consisting of denaturation for 45 s at 94°C, annealing for 45 s at 50–60°C and extension for 45 s at 72°C. A 1 µl aliquot of the F-PCR reaction was mixed with 1 µl of formamide containing Genescan 500 ROX size standards (Applied Biosystems, Foster City, CA, USA). Samples were denatured at 95°C for 3 min, placed on ice and 1 µl subjected to capillary electrophoresis in POP4 resin on the ABI Prism 3100 DNA Sequencer. Fluorescent bands were analysed and sized by Genescan software (Applied Biosystems) and the number of allelic repeats calculated for each STR sequence.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Strategy to identify viable blastocysts
The strategy devised to identify viable human blastocysts and potentially reveal genes associated with viability is schematically represented in Fig. 1. Embryos produced for an ART cycle were grown to the blastocyst stage and a TE biopsy was performed to remove a small number of cells. The poly A+ mRNA fraction from the biopsied material was subsequently used to generate cDNA expression libraries for downstream molecular analyses. Biopsied blastocysts were then transferred to establish a pregnancy and once the baby was born, cord blood or a buccal cell sample was collected. Comparison of DNA fingerprints derived from the blastocyst biopsy samples and the baby’s cells therefore identifies which of the transferred blastocysts produced the pregnancy. This approach minimized maternal effects because sibling blastocysts transferred to the same uterine environment should have the same opportunity to implant. When all of the blastocyst(s) transferred implanted and produced term pregnancies, the blastocysts were classified as viable (all implanted, Group 1). When some of the blastocyst(s) transferred implanted and produced a term pregnancy, the origin of which could clearly be identified by DNA fingerprinting, these blastocysts were classified as viable and the ones that did not match by DNA fingerprinting were classified definitively as non-viable (some implanted, Group 2). When none of the blastocyst(s) transferred implanted, the blastocysts were classified as non-viable or the uterus was deemed to be non-receptive (none implanted, Group 3). The availability of cDNA libraries from the viable and definitively non-viable blastocysts provides a means to identify genes associated with implantation and developmental competence using microarray expression profiling.
|
Blastocyst biopsy and clinical outcomes
There were 48 patients recruited for blastocyst biopsy and all patients produced at least one blastocyst by Day 5. Approximately, 8–20 TE cells were biopsied from 153 blastocysts, some of which were cryopreserved for a later treatment cycle.
The biopsy technique did not appear to impact on the blastocyst’s ability to re-form a blastocoele cavity and continue to grow and hatch from the zona pellucida, as all blastocysts had a clearly visible blastocoele cavity within 1–2 h of biopsy and all blastocysts had a larger diameter on Day 6 compared with Day 5 (Fig. 2). In a small number of blastocysts, usually early blastocysts, few TE cells herniated from the hole in the zona pellucida following several hours incubation after laser-assisted hatching (Fig. 2B and C). At all instances, it was possible to apply suction around the opening and encourage a sufficient number of cells into the biopsy pipette using a combination of suction and gentle mechanical pulling. The ease of biopsy, however, was far greater for those blastocysts that had herniated a significant number of cells during the pre-incubation period (Fig. 2A). There were no blastocysts early on Day 5 that could not be biopsied later in the day and no blastocysts failed to re-form a blastocoele cavity following biopsy. No blastocyst showed major signs of morphological damage at the light microscopic level although occasional debris and individual TE cells could be identified in the perivitelline space and this material was always excluded from the re-sealed blastocyst and often remained within the empty zona pellucida when the blastocyst had completely hatched (data not shown). Transfer of biopsied blastocysts resulted in a similar pregnancy rate (52%) to that following transfer of non-biopsied blastocysts to patients not participating in the trial during the same time period (n = 50 patients; pregnancy rate 46%). Implantation rates were significantly higher following transfer of biopsied blastocysts (35 versus 19%; P < 0.004). However the validity of this comparison is questionable because patients were not randomly allocated to the biopsy or non-biopsy groups.
|
All 48 patients proceeded to blastocyst transfer of one or more blastocysts. Twenty-five patients became pregnant, with evidence of at least one fetal heart on ultrasound scan. The implantation outcome for all 48 patients is summarized in Table II. In seven patients (Group 1), all transferred blastocysts implanted indicating the blastocysts were all viable. In 18 patients (Group 2), some blastocysts implanted, whereas others did not, giving a powerful internal control for uterine receptivity. In the remaining 23 patients (Group 3), none of the blastocysts implanted, indicating that the blastocysts were either all non-viable or that the uterus was not receptive. A small number of patients from Group 2 undertook fetal reduction to reduce a multiple pregnancy, which meant that the blastocyst(s) resulting in the offspring could be identified by DNA fingerprinting but not all the blastocysts transferred could be definitively diagnosed as viable or non-viable. In total, 37 healthy babies were delivered.
|
Microarray analysis of cDNA libraries from biopsied TE
In order to identify candidate genes associated with viability, we have performed comparative microarray expression analysis using separate pools of TE biopsies from Group 1 and Group 3. Pools of eight TE biopsies were chosen to provide a representative cRNA template for microarray analysis and to reduce the effects of any heterogeneity between blastocysts. Group 1 blastocysts were those in which all blastocysts transferred implanted in the uterus and established ongoing pregnancies. Group 3 blastocysts were chosen from patients in which all blastocysts transferred failed to implant and also in whom implantation failure rather than endometrial deficiency was the more likely cause of infertility. Accordingly, blastocysts from young patients who had undergone few or no previous assisted reproduction cycles and had tubal disease or male factor as the aetiology of infertility were selected to generate pools of Group 3 blastocysts. Two pools of Group 1 blastocysts and two pools of Group 3 blastocysts were subsequently generated. One pool from each group was subjected to linear amplification and the other to exponential amplification and cRNA products were analysed on whole-genome arrays. Both types of amplification procedures were used to maximize the identification of genes expressed by such a small number of TE cells in the biopsy sample in the event that one of the procedures was less efficient or showed some degree of bias in the amplified cDNA. Comparison of Group 1 and Group 3 microarray profiles identified similarities and differences in gene expression profiles. A total of 44 947 gene transcripts were commonly expressed in TE. Comparison of Group 1 and Group 3 results identified 7317 and 1896 transcripts, respectively, which were unique. Of the 7317 transcripts identified to be expressed uniquely by viable embryos, 4208 (58%) were detected in the microarrays generated by both linear and exponential amplification techniques. Genespring analysis of the GO biological processes over-represented in these 7317 transcripts compared with the total number of probes on the array identified cell adhesion, cell communication, cellular metabolic processes and response to stimuli (Table III) as the most significant (P < 0.05).
|
In an attempt to refine the 7317 transcripts associated with viability, cDNA expression libraries were generated by exponential amplification of the TE biopsy RNA from eight individual blastocysts that implanted (Group 1) and 19 individual blastocysts that failed to implant (Group 3). Microarray analysis of these 27 TE cDNA libraries identified a significantly lower number of transcripts (10 222 ± 6614) compared with the four pooled samples (43 255 ± 10 009), indicating that the individual cDNA libraries derived from 8–20 TE cells were not completely representative of global gene expression. Nevertheless, the individual TE biopsy data were subjected to principle components analysis to identify any trends in the gene expression patterns that could discriminate between Group 1 and Group 3 blastocysts. Clustering of Group 1 and of Group 3 blastocysts on the basis of trends defined by principal components X (9.3% variance), Y (7.1% variance) and Z (5.5% variance) was indeed found, although there was a small degree of overlap between the two groups (Fig. 3).
|
DNA fingerprinting to define viable and non-viable blastocysts transferred to the same uterus (Group 2)
To unequivocally discriminate between viable and non-viable blastocysts, we first investigated the possibility of using the information provided by the cDNA libraries generated from the pooled Group 1 and Group 3 TE biopsy microarrays to define polymorphic cDNAs for DNA fingerprinting. The 44 947 transcripts expressed in common from Group 1 and Group 3 blastocysts were compared with the H-invitational database (http://www.jbirc.aist.go.jp/hinv/index.jsp) to identify transcripts with STR sequences in their exonic regions. A list of 24 polymorphic genes expressed at relatively high levels in both Group 1 and Group 3 TE was identified as initial candidate transcripts for fingerprint analysis. Analysis of the polymorphic nature of the STR region of these 24 transcripts across sibling blastocysts identified only four genes (LAMP2, SCEB3, BASP1 and HNRPC) with allelic heterozygosity indices of >0.5 that were suitable for unique discrimination. Eight cohorts of sibling blastocysts from Group 2 that had previously been fully processed to cDNA libraries together with the corresponding offspring cell sample were analysed. The resulting DNA fingerprints derived from the four genes listed earlier (data not shown) were not sufficiently informative in all cases to uniquely identify the embryonic origin of the baby.
As an alternative strategy, DNA fingerprinting of the genomic DNA fraction of the TE biopsy samples combined with WGA was investigated. For this purpose, we used TE biopsy lysates from an additional three sibling blastocyst cohorts from Group 2 that had not been previously processed to cDNA libraries. Half of the sibling blastocyst TE biopsy lysate was subjected to WGA, and amplified products together with the matching offspring cell sample(s) were subjected to DNA fingerprinting using the sexing amelogenin marker and seven STR markers which have heterozygosity indices of >0.8. The DNA fingerprinting outcomes for the three sibling blastocyst cohorts are summarized in Table IV, and the DNA fingerprint profiles of one of these sibling blastocyst families and the resultant offspring are depicted in Fig. 4.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
The hallmark of a successful ART program is a consistent and sustainable high pregnancy rate balanced by a low incidence of multiple gestations. Although several clinics have moved towards single-embryo transfers, a paradigm shift from multiple embryo to single-embryo transfers worldwide may be contingent on a reliable method to select the most viable embryo in the cohort with the ability to implant in a receptive uterus and develop to term. In the present study, we have developed and validated a new strategy on the basis of blastocyst biopsy and DNA fingerprinting that links blastocyst viability with gene expression profiles. The strategy was premised on previous observations that removal of 2–4 TE cells from early blastocysts for the purpose of PGD resulted in healthy babies unaffected with genetic disease (Kokkali et al., 2005
; McArthur et al., 2005), indicating that TE biopsy was safe. The present study demonstrates that removal of approximately 8–20 TE cells from Day 5 human blastocysts, representing 
10–20% of the blastocyst mass (Dokras et al., 1993; Fong and Bongso, 1999), using a non-contact mode laser is compatible with implantation and normal development to term following blastocyst transfer on Day 6. Further, it is possible using the methodology described to biopsy all blastocysts on Day 5 regardless of developmental stage.
Pre-implantation development is governed by two independent and distinct waves of transcription in mammals, the second of which occurs at the morula–blastocyst transition (Hamatani et al., 2004; Wang et al., 2004). Therefore molecules important to implantation events may be expressed only at the blastocyst stage of development, emphasizing the relevance of gene expression data derived from TE biopsies. Prior to implantation, adhesion of the blastocyst is mediated by cell–cell interactions involving cell surface molecules expressed by TE cells and luminal uterine epithelial cells (Giudice, 1999). Molecular pathology and immunohistochemical studies in humans, primates and rodent species, together with data from experimental animals models such as gene knockout mice, have identified candidate growth factor ligands/receptors and homo- and heterophilic adhesion molecules expressed by the endometrium that are likely to play a key role in implantation and placentation (Aplin, 1997). However, from the perspective of the pre-implantation primate blastocyst, our understanding of the molecules involved in apposition and attachment is limited (Simon et al., 2000). Bioinformatics analysis of the gene families over-represented in the 7317 transcripts uniquely expressed in blastocysts that implanted identified cell adhesion as the most significant, in particular homophilic cell adhesion and calcium dependent and independent adhesion interactions. Thus, cell adhesion mediated by a combination of these molecules may be a critical event for attachment of the human blastocyst to the endometrium. The incorporation of additional gene expression profiles from pooled TE biopsy samples from blastocysts identified to be viable by DNA fingerprinting with the data generated in the present study followed by bioinformatics analyses will assist in refining genes and cellular pathways that are important in the implantation process in the human. Experiments to generate additional known viable TE biopsy samples are underway, as insufficient samples were available to both develop the described technology and perform sufficient replicate microarrays to identify genes that might be differentially expressed between viable and non-viable blastocysts. Genes differentially expressed may be equally or more important than those genes identified to be uniquely expressed. Furthermore the genes identified to be uniquely expressed by viable blastocysts are not provided, as it is likely that this list will be refined to a smaller list once more replicate microarrays have been completed. When the gene data set is refined, it will be possible to develop real-time PCR assays for specific genes expressed in single TE biopsy samples to select the most viable blastocyst in the cohort prior to transfer. Further, the identification of genes critical for implantation and developmental competence should expand the potential targets for contraception and expand our understanding of the basis of some aetiologies of female infertility.
The identification of a large set of genes associated with viability from pooled TE biopsy samples and separation of individual blastocysts that implanted from those that did not implant by principle components analysis demonstrate for the first time that viable blastocysts do express a different transcriptome that is compatible with implantation competence. Microarray analysis of individual TE biopsy samples however revealed far fewer genes expressed than when TE biopsy samples were pooled and arrayed, which is typical of this methodology when the starting template of mRNA is low (Jones et al., 2007). Nevertheless principle components analysis separated viable from non-viable blastocysts even though the combined principle components explained only 22% of the variance. This highlights the need to accumulate and analyse additional TE biopsy samples where the implantation outcome is known and this is currently being undertaken. Although a large number of transcripts were identified in pooled samples to be expressed only by viable embryos, these data rely on single replicates amplified by two different amplification procedures, so statistical analysis of genes expressed in common between viable and non-viable blastocysts could not be undertaken. Nevertheless, the majority of genes found to be uniquely expressed by viable blastocysts were detected in microarrays generated using both linear and exponential amplification providing a further level of confidence in the viability gene list.
The reason for the difference in gene expression between viable and non-viable blastocysts is not clear given that the cohort from a patient’s ART cycle is grown in the same culture media under the same conditions in vitro. It has been proposed that chromosomal aneuploidy and/or the failure to properly activate the embryonic genome are two key inherent abnormalities that reduce the implantation potential of good quality IVF embryos (Braude et al., 1988; Margalioth et al. 2006). Meiosis I and II errors in the oocyte and mitotic errors in the first, second and third cleavage divisions of early embryos are common (Katz-Jaffe et al., 2004; Kuliev et al., 2003) and lead to the development of aneuploid and mosaic embryos, respectively. Depending on the chromosome(s) involved, these embryos either undergo early arrest or grow normally to the blastocyst stage (Magli et al., 2000). Therefore, altered gene dosage caused by either a monosomy or trisomy could potentially lead to an abnormal transcriptome culminating in implantation failure. However, the fingerprinting analysis for the three sibling cohorts of embryos using gene markers on chromosomes X, Y, 13, 18 and 21 did not reveal any evidence of aneuploidy even in blastocysts that failed to implant. On the other hand, the failure of euploid embryos to implant in a receptive uterus could be related to abnormalities in the oocytes from which they were derived. It is well known that within a cohort of mature oocytes retrieved from a standard stimulated cycle, only a proportion are developmentally competent. It has been proposed that during oocyte maturation, de-adenylation of specific transcripts required for early pre-implantation development in a time dependent manner is critical to maintain developmental competence (Brevini-Gandolfi et al., 1999; Zheng et al., 2005; Su et al., 2007). Thus, premature degradation of these transcripts could lead to a deficiency of developmental proteins that are required for proper activation of the embryonic genome and an altered transcriptome that is incompatible with implantation.
DNA fingerprinting has been a useful ART technique to assist in PGD of single-gene disorders (Sermon et al., 2004). Incorporation of polymorphic STR markers linked to specific mutations into the multiplex PCR reaction enables the identification of DNA contamination and ADO and thus assists with making a more accurate diagnosis. More recently, PGD cases have also been performed where disease-free embryos are also selected for an HLA match using STR markers across the HLA region on chromosome 6 (Fiorentino et al., 2004; Verlinsky et al., 2004), so that if the pregnancy reaches term, the cord blood from the donor child can be used for transplantation in an attempt to save the life of a sick sibling. Apart from these applications, DNA fingerprinting has not been used to track transferred embryos to term for the purpose of identifying developmentally competent embryos. In the present study, a novel strategy is described whereby a portion of the blastocyst TE biopsy sample was used to obtain unique sibling embryo DNA fingerprints to identify the specific blastocyst(s) in the transfer cohort that implanted and developed to term. Some of the embryonic DNA fingerprints derived from the TE lysate were incomplete due to sporadic instances of total amplification failure and ADO for one or more STR markers, indicting that the DNA template in half the TE biopsy samples (4–10 cells) was limiting. Nevertheless, the combined STR profiles established from all markers were sufficiently informative to confidently identify the sibling blastocyst(s) that produced the offspring in all three families examined.
Tracking of transferred blastocysts to term to reveal the viable sibling blastocyst among a cohort transferred represents a new approach for refining existing criteria for selecting viable blastocysts for transfer. This approach is particularly powerful when multiple blastocysts are transferred and a pregnancy results because the uterus must have been receptive. In contrast, single-blastocyst transfers where DNA fingerprinting is not required to identify the origins of the pregnancy remain confounding in cases where the blastocyst does not implant, as it is not known whether this is due to the transfer of a non-viable blastocyst or a non-receptive uterus. Using the method of DNA fingerprinting TE biopsy samples developed in the present study, it will now be possible to conduct controlled experiments whereby the test and control can be conducted within one patient’s cohort of embryos, i.e. the performance of two different culture media or two different culture conditions on pregnancy outcome can be analysed by transferring a biopsied blastocyst from each arm of the experiment, and the blastocyst(s) responsible for any resulting offspring can be identified by DNA fingerprinting. DNA fingerprinting will also assist in the re-assessment of classical morphological grading as a non-invasive tool for selecting the most viable blastocyst for transfer. For example, embryologists could record morphological observations of oocytes and resulting embryos up to the blastocyst stage and correlate the findings with those blastocysts that implanted or failed to implant in a receptive uterus. This approach may identify a set of morphological characteristics that is more highly predictive of blastocyst viability. DNA fingerprinting could also be used to improve the predictive value of other metabolic and biochemical non-invasive markers of viability. Furthermore, once additional TE biopsy samples have been accumulated and analysed, it will be possible to refine the set of viability genes to produce a smaller subset of genes more predictive of competence to develop to term. These molecular markers could then be measured using rapid PCR-based assays on TE biopsy samples to prospectively select the single most viable blastocyst among the cohort developing in vitro for transfer back to the patient.
If a simple non-invasive assay of blastocyst viability can be developed, it will permit the identification of the single most viable blastocyst among the cohort developing in vitro and enable single blastocyst transfers without a concomitant reduction in pregnancy rates. This would effectively reduce the incidence of multiple gestations, which is currently a priority in the field of ART medicine.
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
This research was supported by a grant from Monash IVF, Melbourne, Australia and the Centre for Human Reproduction, Athens, Greece.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
The authors wish to thank the embryology staff at the Centre for Human Reproduction, Athens, Greece for their kind assistance in the culture of human embryos for the present study. We would also like to thank Hamilton Thorne Biosciences for the donation of the ZILOS-tk laser optical system used for assisted hatching and biopsy throughout the present study.
|
Footnotes |
|---|
These authors contributed equally to this work. 
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Adjaye J, Bolton V, Monk M. Developmental expression of specific genes detected in high-quality cDNA libraries from single human preimplantation embryos. Gene (1999) 237:373–383.[CrossRef][ISI][Medline]
Adjaye J, Huntriss J, Herwig R, BenKahla A, Brink TC, Wierling C, Hultschig C, Groth D, Yaspo ML, Picton HM, et al. Primary differentiation in the human blastocyst: comparative molecular portraits of inner cell mass and trophectoderm cells. Stem Cells (2005) 23:1514–1525.
Alikani M, Cohen J, Tomkin G, Garrisi J, Mack C, Scott RT. Human embryo fragmentation in vitro and its implications for pregnancy and implantation. Fertil Steril (1999) 71:836–842.[CrossRef][ISI][Medline]
Aplin JD. Adhesion molecules in implantation. Rev Reprod (1997) 2:84–93.[Abstract]
Bergh C. Single embryo transfer: a mini-review. Hum Reprod (2005) 20:323–327.
Braude P, Bolton V, Moore S. Human gene expression first occurs between the four- and eight-cell stages of preimplantation development. Nature (1988) 332:459–461.[CrossRef][Medline]
Brevini-Gandolfi TA, Favetta LA, Mauri L, Luciano AM, Cillo F, Gandolfi F. Changes in poly(A) tail length of maternal transcripts during in vitro maturation of bovine oocytes and their relation with developmental competence. Mol Reprod Dev (1999) 52:427–433.[CrossRef][ISI][Medline]
Brison DR, Houghton FD, Falconer D, Roberts SA, Hawkhead J, Humpherson PG, Lieberman BA, Leese HJ. Identification of viable embryos in IVF by non-invasive measurement of amino acid turnover. Hum Reprod (2004) 19:2319–2324.
Brison DR, Hollywood K, Arnesen R, Goodacre R. Predicting human embryo viability: the road to non-invasive analysis of the secretome using metabolic footprinting. Reprod BioMed Online (2007) 15:296–302.[ISI][Medline]
Clouston HJ, Herbert M, Fenwick J, Murdoch AP, Wolstenholme J. Cytogenetic analysis of human blastocysts. Prenat Diagn (2002) 22:1143–1152.[CrossRef][ISI][Medline]
Cullinan EB, Abbondanzo SJ, Anderson PS, Pollard JW, Lessey BA, Stewart CL. Leukemia inhibitory factor (LIF) and LIF receptor expression in human endometrium suggests a potential autocrine/paracrine function in regulating embryo implantation. Proc Natl Acad Sci USA (1996) 93:3115–3120.
Cummins JM, Breen TM, Harrison KL, Shaw JM, Wilson LM, Hennessey JF. A formula for scoring human embryo growth rates in in vitro fertilization: its value in predicting pregnancy and in comparison with visual estimates of embryo quality. J In Vitro Fert Embryo Transf (1986) 3:284–295.[CrossRef][Medline]
Dobson AT, Raja R, Abeyta MJ, Taylor T, Shen S, Haqq C, Reijo Pera RA. The unique transcriptome through day 3 of human preimplantation development. Hum Mol Genet (2004) 13:1461–1470.
Dokras A, Sargent IL, Barlow DH. Human blastocyst grading: an indicator of developmental potential? Hum Reprod (1993) 8:2119–2127.
Edwards RG, Beard HK. Is the success of human IVF more a matter of genetics and evolution than growing blastocysts? Hum Reprod (1999) 14:1–4.
Fiorentino F, Biricik A, Karadayi H, Berkil H, Karlikaya G, Sertyel S, Podini D, Baldi M, Magli MC, Gianaroli L, et al. Development and clinical application of a strategy for preimplantation genetic diagnosis of single gene disorders combined with HLA matching. Mol Hum Reprod (2004) 10:445–460.
Fong CY, Bongso A. Comparison of human blastulation rates and total cell number in sequential culture media with or without co-culture. Hum Reprod (1999) 14:774–781.
Giorgetti C, Terriou P, Auquier P, Hans E, Spach J-L, Salzmann J, Roulier R. Embryo score to predict implantation after in-vitro fertilization: based on 957 single embryo transfers. Hum Reprod (1995) 10:2427–2431.
Guidice LC. Potential biochemical markers of uterine receptivity. Hum Reprod (1999) 14(Suppl. 2):3–16.[Medline]
Hamatani T, Carter MG, Sharov AA, Ko MS. Dynamics of global gene expression changes during mouse preimplantation development. Dev Cell (2004) 6:117–131.[CrossRef][ISI][Medline]
Holding C, Bolton V, Monk M. Detection of human novel preimplantation genes in cDNA derived from replicate individual preimplantation embryos. Mol Hum Reprod (2000) 6:801–809.
Jones GM. Growth and viability of human blastocysts in vitro. Reprod Med Rev (2000) 83:241–287.
Jones GM, Trounson AO, Gardner DK, Kausche A, Lolatgis N, Wood C. Evolution of a culture protocol for successful blastocyst development and pregnancy. Hum Reprod (1998) 13:169–177.
Jones GM, Trounson AO, Vella PJ, Thouas GA, Lolatgis N, Wood C. Glucose metabolism of human morula and blastocyst-stage embryos and its relationship to viability after transfer. Reprod Biomed Online (2001) 3:124–132.[Medline]
Jones GM, Song B, Cram DS, Trounson AO. Optimization of a microarray based approach for deriving representative gene expression profiles from human oocytes. Mol Reprod Dev (2007) 74:8–17.[CrossRef][ISI][Medline]
Katz MG, Trounson AO, Cram DS. DNA fingerprinting of sister blastomeres from human IVF embryos. Hum Reprod (2002) 17:752–759.
Katz-Jaffe MG, Trounson AO, Cram DS. Mitotic errors in chromosome 21 of human preimplantation embryos are associated with non-viability. Mol Hum Reprod (2004) 10:143–147.
Katz-Jaffe MG, Schoolcraft WB, Gardner DK. Analysis of protein expression (secretome) by human and mouse preimplantation embryos. Fertil Steril (2006) 86:678–685.[CrossRef][ISI][Medline]
Kokkali G, Vrettou C, Traeger-Synodinos J, Jones GM, Cram DS, Stavrou D, Trounson AO, Kanavakis E, Pantos K. Birth of a healthy infant following trophectoderm biopsy from blastocysts for PGD of beta-thalassaemia major. Hum Reprod (2005) 20:1855–1859.
Kokkali G, Traeger-Synodinos J, Vrettou C, Stavrou D, Jones GM, Cram DS, Makrakis E, Trounson AO, Kanavakis E, Pantos K. Blastocyst biopsy versus cleavage stage biopsy and blastocyst transfer for preimplantation genetic diagnosis of beta-thalassaemia: a pilot study. Hum Reprod (2007) 22:1443–1449.
Kuliev A, Cieslak J, Ilkevitch Y, Verlinsky Y. Chromosomal abnormalities in a series of 6733 human oocytes in preimplantation diagnosis for age-related aneuploidies. Reprod Biomed Online (2003) 6:54–59.[Medline]
Levi Setti PE, Albani E, Cavagna M, Bulletti C, Colombo GV, Negri L. The impact of embryo transfer on implantation—a review. Placenta (2003) 24(Suppl. B):S20–S26.[CrossRef][ISI][Medline]
Li SS-L, Liu Y-H, Tseng C-N, Singh S. Analysis of gene expression in single human oocytes and preimplantation embryos. Biochem Biophys Res Commun (2006) 340:48–53.[ISI][Medline]
Magli MC, Jones GM, Gras L, Gianaroli L, Korman I, Trounson AO. Chromosome mosaicism in day 3 aneuploid embryos that develop to morphologically normal blastocysts in vitro. Hum Reprod (2000) 15:1781–1786.
Margalioth EJ, Ben-Chetrit A, Gal M, Eldar-Geva T. Investigation and treatment of repeat implantation failure following IVF-ET. Hum Reprod (2006) 21:3036–3043.
McArthur SJ, Leigh D, Marshall JT, de Boer KA, Jansen RP. Pregnancies and live births after trophectoderm biopsy and preimplantation genetic testing of human blastocysts. Fertil Steril (2005) 84:1628–1636.[CrossRef][ISI][Medline]
Munne S, Alikani M, Tomkin G, Grifo J, Cohen J. Embryo morphology, developmental rates, and maternal age are correlated with chromosome abnormalities. Fertil Steril (1995) 64:382–391.[ISI][Medline]
Munne S, Marquez C, Reing A, Garrisi J, Alikani M. Chromosome abnormalities in embryos obtained after conventional in vitro fertilization and intracytoplasmic sperm injection. Fertil Steril (1998) 69:904–908.[CrossRef][ISI][Medline]
Norwitz ER, Schust DJ, Fisher SJ. Implantation and the survival of early pregnancy. N Engl J Med (2001) 345:1400–1408.
Pandian Z, Templeton A, Serour G, Bhattacharya S. Number of embryos for transfer after IVF and ICSI: a Cochrane review. Hum Reprod (2005) 20:2681–2687.
Pantos K, Meimeth-Damianaki T, Vaxevanoglou T, Kapetanakis E. Prospective study of a modified gonadotropin-releasing hormone agonist long protocol in an in vitro fertilization program. Fertil Steril (1994) 61:709–713.[ISI][Medline]
Pellestor F, Andreo B, Arnal F, Humeau C, Demaille J. Maternal aging and chromosomal abnormalities: new data drawn from in vitro unfertilized human oocytes. Hum Genet (2003) 112:195–203.[CrossRef][ISI][Medline]
Pera MF, Bennett W, Cerretti DP. Expression of CD30 and CD30 ligand in cultured cell lines from human germ-cell tumors. Lab Invest (1997) 76:497–504.[ISI][Medline]
Plachot M. Genetic analysis of the oocyte – a review. Placenta (2003) 24(Suppl. B):S66–S69.[CrossRef][ISI][Medline]
Puissant F, Van Rysselberge M, Barlow P, Deweze J, Leroy F. Embryo scoring as a prognostic tool in IVF treatment. Hum Reprod (1987) 2:705–708.
Rosenbusch B. The incidence of aneuploidy in human oocytes assessed by conventional cytogenetic analysis. Hereditas (2004) 141:97–105.[CrossRef][ISI][Medline]
Sakkas D, Percival G, D’Arcy Y, Sharif K, Afnan M. Assessment of early cleaving in vitro fertilized human embryos at the 2-cell stage before transfer improves embryo selection. Fertil Steril (2001) 76:1150–1156.[CrossRef][ISI][Medline]
Schoolcraft WB, Gardner DK, Lane M, Schlenker T, Hamilton F, Meldrum DR. Blastocyst culture and transfer: analysis of results and parameters affecting outcome in two in vitro fertilization programs. Fertil Steril (1999) 72:604–609.[CrossRef][ISI][Medline]
Scott LA, Smith S. The successful use of pronuclear embryo transfers the day following oocyte retrieval. Hum Reprod (1998) 13:1003–1013.
Sermon K, Van Steirteghem A, Liebaers I. Preimplantation genetic diagnosis. Lancet (2004) 363:1633–1641.