Reproductive BioMedicine Online (2014) 28, 624– 637
www.sciencedirect.com www.rbmonline.com
ARTICLE
Telomere length in human blastocysts Anastasia Mania a,b,*, Anna Mantzouratou a,b, Joy DA Delhanty a, Gianluca Baio c, Paul Serhal b, Sioban B Sengupta a a UCL Centre for PGD, Institute for Women’s Health, University College London, 86–96 Chenies Mews, London WC1E 6HX, United Kingdom; b Centre for Reproductive and Genetic Health (CRGH), The New Wing Eastman Dental Hospital, 256 Gray’s Inn Road, London WC1X 8LD, United Kingdom; c Department of Statistical Science, University College London, Gower Street, London WC1E 6BT, United Kingdom
* Corresponding author. E-mail addresses: [email protected], [email protected] (A Mania). Anastasia Mania is a senior clinical embryologist with an interest in embryo genetic research. She has been working as a clinical embryologist since 2007. She completed her clinical embryology training in 2010 at IVF Hammersmith (Imperial College Healthcare NHS Trust) and is registered as a clinical scientist with the Health and Care Professions Council. She is currently working in a private London IVF unit. Anastasia worked in preimplantation embryo genetics as a molecular geneticist in the Centre for PGD at University College London from 2004 to 2007 and completed her MSc in prenatal genetics and fetal medicine in 2007 at University College London.
Abstract This is a retrospective study aiming to assess telomere length in human embryos 4 days post fertilization and to determine
whether it is correlated to chromosomal ploidy, embryo developmental rate and patient age. Embryos were donated from patients undergoing treatment in the assisted conception unit. Seven couples took part, generating 35 embryos consisting of 1130 cells. Quantitative fluorescent in-situ hybridization (FISH) measured the telomere length of every cell using a pan-telomeric probe. Conventional FISH on six chromosomes was used to assess aneuploidy in the same cells. Maternal and paternal age, referral reason, embryo developmental rate and type of chromosomal error were taken into account. Chromosomally abnormal cells were associated with shorter telomeres than normal cells for embryos that were developmentally slow. Cells produced by women of advanced maternal age and those with a history of repeated miscarriage tended to have substantially shorter telomeres. There was no significant difference in telomere length with respect to the rate of embryo development 5 days post fertilization. Telomeres play an important role in cell division and shorter telomeres may affect embryonic ploidy. Reduced telomere length was associated with aneuploid cells and embryos from women of advanced maternal age. RBMOnline ª 2014, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved. KEYWORDS: aneuploidy, blastocyst, human, preimplantation embryo, quantitative FISH, telomere length
Introduction Telomeres are six base-pair repeats (TTAGGG) found at the ends of chromosomes to protect them from degradation
(Chan and Blackburn, 2002). Telomeres participate in processes of chromosomal repair, prevent nonspecific chromosomal recombination and help the chromosome bind to the nuclear matrix (Counter et al., 1992; De Lange, 1992). Prior
http://dx.doi.org/10.1016/j.rbmo.2013.12.010 1472-6483/ª 2014, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved.
Telomere length in human blastocysts to mitotic cell division, cellular DNA is duplicated by the action of a DNA polymerase. This enzyme does not replicate the ends of chromosomes and so telomeres get shorter after each cell division. In normal somatic cells, this shortening contributes to cellular senescence (Blasco et al., 1997). Telomere shortening limits the regenerative capacity of cells and is correlated with the onset of cancer, ageing and chronic diseases (Djojosubroto et al., 2003; Rudolph et al., 1999; Satyanarayana et al., 2003). Some specialized cells maintain their capacity to divide and regenerate their telomeres due to the action of an enzyme called telomerase. Telomerase is a reverse transcriptase that elongates telomeres in some tissues (i.e. germ and stem cells; Flores et al., 2006). Telomerase activity declines during late oogenesis and early preimplantation development, so maximum telomere length is thought to be established during early oogenesis (Wright et al., 1996). The telomere theory of reproductive senescence supports the idea that telomere shortening in late exit from the fetal production line and a long interval before ovulation in the adult thus causes reproductive ageing in women (Blasco et al., 1999). For example, in null mice for the telomerase gene, as telomeres shorten across generations, oocyte and embryo development deteriorates (Hassold and Hunt, 2001). Chromosomes frequently malsegregate during female meiosis leading to aneuploidy, failed implantation and miscarriage (Harper et al., 2004; Keefe et al., 2006; Munne ´ et al., 1995). The mechanism by which telomere shortening causes this misalignment is not well understood (Blasco et al., 1999), but could result from improper pairing of homologous chromosomes and defective formation of synapses and bouquet formation during early meiosis (Rudolph et al., 1999; Siderakis and Tarsounas, 2007). Meiotic progression and chromosome alignment at the metaphase stage in oocytes of first-generation null mice for the telomerase gene were comparable to that of the oocytes of wild-type mice; however, fourth-generation null mice showed chromosome misalignment, disruption of meiotic spindles and very short telomeres (Liu et al., 2002). In humans, telomere length of replicating somatic cells is inversely related to donor age (Vaziri et al., 1993), highly variable among donors of the same age (Slagboom et al., 1994), highly heritable (Jeanclos et al., 2000) and greater in women than in men (Benetos et al., 2001). In contrast to the considerable information known about telomere length in human extrauterine life, little is known about telomere length during embryonic development. Experiments on the telomere length at different stages of bovine and mouse early development have shown that telomeres at the morula–blastocyst transition reset their length to a specific set point (Schaetzlein et al., 2004), supporting the theory of telomere restoration in mammalian preimplantation development. Xu and Yang (2000) demonstrated that telomerase activity is detected at fertilization in bovine embryos created by IVF and reaches its highest level at the blastocyst stage. Recent data show that telomere length is critically related to fragmentation in IVF embryos in humans but further studies are needed (Keefe and Parry, 2005). This study measured the telomere length in embryos at 5 days post fertilization and assessed whether there is an association between the chromosomal complement of the embryo and its developmental rate using quantitative and
625 conventional fluorescence in-situ hybridization (FISH). The aim of this study was to start unravelling the correlation between telomere length and chromosome aneuploidy in human preimplantation embryos, a relationship that has not been studied before in such detail.
Materials and methods Experiments were performed in preimplantation embryos 5 days post fertilization from patients undergoing fertility treatment with chromosomal screening. Seven couples donated embryos that were unsuitable for transfer or cryopreservation after being diagnosed as abnormal by aneuploidy screening on the day 3 of embryo development. The data comprised 35 embryos consisting of 1130 blastomeres. Two control samples were used in each experiment along with the embryos: (i) maternal and paternal lymphocyte suspensions; and (ii) a human cell line of immortalized keratinocytes called HaCaT (German Cancer Research Centre, DKFZ, Heidelberg, Germany). The parental lymphocyte suspensions were also used to measure the efficiency of the quantitative and interphase FISH techniques. HaCaT cells were used as a positive control since they are modified so as to have standard telomere length of 4 kb (Boukamp et al., 1988). The HaCaT cell line allowed the conversion of the fluorescence intensity in each embryo cell to kilobases since fluorescence intensity has been found to be proportional to telomere length (Lansdorp, 1996).
Patient details and ethical consent Patients who undergo chromosomal screening are referred for three main reasons: (i) those of advanced maternal age (AMA; females 40 years of age); (ii) couples with history of repeated miscarriage (RM, more than three times); and (iii) those with repeated failed IVF cycles (repeated implantation failure, RIF, more than three times). Research on all embryos donated to this study was carried out under licenses from the Human Fertilisation and Embryology Authority (HFEA) of the UK. Informed research consent was obtained from all couples. Approval was obtained by the Research Ethics Committee (REC3), North London, UK (reference no. 10/H0709/26, approved 8 June 2010).
Lymphocyte culture and counts Lymphocyte culture from parental blood samples were carried out by standard methods and suspensions were prepared to be used in each experiment.
Embryo spreading Embryos showing aneuploidy following a clinical preimplantation genetic screening (PGS) cycle and donated for research were spread on slides using the method previously described by Harper et al. (1994). The position of every blastomere was mapped so that it could be identified in each round of FISH. The method has been validated extensively and approved by the HFEA to be used as a diagnostic test for genetic screening. The overall efficiency (percentage of cells
626 with diploid signals) of FISH in lymphocyte samples ranged from 89% to 96%, and no polymorphisms were found in any of the couples. There has been no misdiagnosis from the follow-up analysis of abnormal embryos.
Quantitative FISH on telomere repeats, slide scanning and cell analysis Telomere length was measured using quantitative FISH on interphase nuclei (embryos, lymphocytes and keratinocytes) with a first round of FISH (Figures 1–3, respectively). The quantitative FISH telomere kit, a FISH Kit/Cy3 conjugated with a peptide nucleic acid (PNA) probe, was provided by DAKO Scientific (Denmark), according to the manufacturer’s instructions with a few modifications. The lymphocyte suspensions and embryo spreads were immersed in Tris-buffered saline (TBS) for 2 min, washed in 3.7% formaldehyde (Sigma, USA) in TBS for 2 min and then washed twice in TBS for 5 min each. Lymphocyte suspensions were immersed in pretreatment solution (40 ll of 2000· proteinase K solution diluted in 50 ml TBS) for 15 min; embryo spreads were only incubated for 10 min (both at room temperature). This was because embryo suspensions were prepared through a more effective protocol to remove all cytoplasmic membranes (see section on embryo spreading) than lymphocyte nuclei suspensions that were prepared with hypotonic solution treatments resulting in a less effective cytoplasmic dispersion. Then, the slides were washed twice in TBS for 5 min. A series of dehydration washes followed (70%, 90% and 100% ethanol). The telomere PNA probe was added to each slide (3 ll on each 13-mm coverslip) and they were codenatured at 80C for 5 min. The slides were left to hybridize at room temperature in a dark chamber for 30 min. In a dark room, the samples were washed in rinse solution (1:50 Rinse solution, double-distilled water; DAKO) for 1 min at room temperature to remove the coverslips, immersed for 5 min in wash solution (1:50 wash solution, double-distilled water; DAKO) prewarmed to 65C and dehydrated as before. When slides were dry, 10 ll of DAPI/Vectarshield solution was added to a coverslip (22 · 50 mm), which was then placed on the slide. The slides were stored in the dark for at least 15 min at 4C. Slide scanning, cell identification, intensity measurement and quantification of telomere length were performed by fluorescence microscopy (Olympus BX 40) attached to a cooled charge-coupled device camera and equipped with capturing software (SmartCapture; Vysis, UK). DAPI (blue) and Cy3 (orange) fluorescence (telomere) signals were captured on interphase nuclei at ·1000 magnification. Exposure at the orange filter was manually set at 1.00 and gain at 2.00 and exposure at the blue filter was set at 0.1 and gain at 1.0. The maximum time of exposure was set at 5 s. After capturing, DAPI fluorescence was manually removed from the SmartCapture image produced and the photo was exported to IPLab software (Spectra Services) to be quantified. Telomere signals were manually selected in the focal plane that gave most in-focus hits in each nucleus. IPLab was set so that it did not normalize the results. The selected signals were then averaged giving one quantified result. Background noise variation was minimized by excluding any noise that appeared as a halo. This approach was limiting as some telomere signals had to be excluded. Unfortunately, this is a
A Mania et al. limitation that could not be overcome since embryonic cells are rarely found in their metaphase stage. Capturing was performed in the same light conditions (semi-darkness) in every experiment. Up to 50 interphase nuclei were captured from each parental lymphocyte suspension and HaCaT cell line and every blastomere was captured and mapped from day-5 embryos. Each experiment was analysed within 24 h to avoid fluorescence decay through time. By capturing the images of fluorescently labelled nuclei at the same time interval after hybridization in all experiments, it was ensured that the results were comparable. Changes in the luminosity of the light source of the microscope have also been found to play an important role in reproducibility of results (Narath et al., 2005). Using fluorescently labelled beads at the beginning and end of each experiment monitored this. To normalize between experiments for the quantitation of telomere length, a cell line of standard telomere length was used as previously described. Up to 50 HaCaT cells were captured in each experiment and the fluorescence intensity of the orange filter was quantified. The mean value corresponded to 4 kb and therefore every other fluorescence intensity could be translated into kilobases for the cells analysed in the same experiment. The normalization of the fluorescence intensity values for single cells (measured in intensity units, IU) was calculated by dividing the fluorescence value of the pan-telomeric signals by the fluorescence intensity of the reference probe and increased the reproducibility of the results. This approach allows the comparison of different experiments, the measurement of slides at different times after hybridization and the compensation for variations of the light source. Parental lymphocyte controls had consistently high FISH efficiency of more than 85% for the chromosomes tested. The FISH protocol used to test for the six chromosomes in the parental lymphocytes and embryos has been developed for clinical purposes and its efficiency has been optimized to achieve best results. It was used instead of hybridizing a FISH probe together with the telomere probe as an internal control (Narath et al., 2005), a combination that needed optimization and potentially could compromise the hybridization efficiency of the telomere probe and therefore the quantitation. Capturing and scoring criteria to ensure a comparable quantitative approach between experiments for telomere length measurement were set so as to avoid variation between experiments seen in previous studies (de Pauw et al., 1998). The criteria were: (i) exclusion of nuclei with artefacts such as background noise; (ii) exclusion of nuclei with less than five signals; (iii) focusing of nuclei before capturing to include most signals in one planel; (iv) calibration of the microscope with fluorescent beads before and after each experiment so as to make sure there is no decay in the UV light source; (v) FISH and capturing performed under the same dim light conditions; and (vi) capturing was completed within 24 h from the completion of the FISH procedure.
Conventional FISH for aneuploidy screening After washing off the telomere PNA probe, conventional FISH analysis of all embryos and their controls was performed as previously described by Mantzouratou et al. (2007) using probes for chromosomes 13, 18 and 21 in one round of hybridization and 15, 16 and 22 in the second round.
Telomere length in human blastocysts
627
Figure 1 Telomere quantitative FISH signals (red) with (right) and without (left) DAPI stain on embryonic cells. Bars = 1 lm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Embryo categories according to chromosomal abnormalities Embryos were grouped according to the chromosomal constitution of their cells (Table 1). Embryos with meiotic errors had the same chromosomal error(s) in all cells and
were classed as uniformly aneuploid. Embryos with post-zygotic errors were grouped according to the number of different cell lines that were present in the embryo and the proportion of the embryo that was aneuploid. When there were two cell lines in the embryo, it was classed as a mosaic embryo; when there were several different cell
628
A Mania et al.
Figure 2 Telomere quantitative FISH signals (red) with (right) and without (left) DAPI stain on lymphocyte cells. Bars = 1 lm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
lines in the embryo, it was classed as chaotic; when 50% of the embryo had abnormal cells it was defined as a major mosaic or a major chaotic embryo.
Data and description of the statistical model The data were summarized using descriptive statistics and applied a more advanced modelling based on hierarchical specification (German and Hill, 2006). Hierarchical models are particularly effective in dealing with data grouped or nested in different layers. In this work, cells were grouped within embryos, which were grouped within couples. It was plausible to assume that data on cells coming from the same embryo would tend to be correlated (i.e. to share some common features) and that they would tend to be somehow different from other observations relating to
other embryos. Similarly, embryos donated by the same couple would share some common characteristics and therefore would tend to have some correlation and to present differences compared with embryos of other couples. Analysing the data using procedures like t-tests or standard regression (which are based on the assumption of independence among the observations) does not recognize these different forms of correlation, therefore producing biased estimations. On the contrary, hierarchical models (sometimes referred to as ‘multilevel’) proceed by defining a structure for each layer in the data, specifying a suitable form of correlation among data in the same group. The information for a given unit is used to inform all units in the same group (a phenomenon known as ‘borrowing strength’), which accounts for the underlying uncertainty in a more appropriate way. The objective of this statistical analysis was to investigate the relationship between telomere length and the occurrence of chromosomal
Telomere length in human blastocysts
629
Figure 3 Telomere quantitative FISH signals (red) with DAPI stain (blue) on keratinocyte cells. Bars = 1 lm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Table 1
Embryo characterization: proportion of euploid and aneuploid cells in each embryo group.
Embryo characterization
Uniformly abnormal Minor mosaic Major mosaic Minor chaotic Major chaotic
Euploid cells
0 >50 50
www.sciencedirect.com www.rbmonline.com
ARTICLE
Telomere length in human blastocysts Anastasia Mania a,b,*, Anna Mantzouratou a,b, Joy DA Delhanty a, Gianluca Baio c, Paul Serhal b, Sioban B Sengupta a a UCL Centre for PGD, Institute for Women’s Health, University College London, 86–96 Chenies Mews, London WC1E 6HX, United Kingdom; b Centre for Reproductive and Genetic Health (CRGH), The New Wing Eastman Dental Hospital, 256 Gray’s Inn Road, London WC1X 8LD, United Kingdom; c Department of Statistical Science, University College London, Gower Street, London WC1E 6BT, United Kingdom
* Corresponding author. E-mail addresses: [email protected], [email protected] (A Mania). Anastasia Mania is a senior clinical embryologist with an interest in embryo genetic research. She has been working as a clinical embryologist since 2007. She completed her clinical embryology training in 2010 at IVF Hammersmith (Imperial College Healthcare NHS Trust) and is registered as a clinical scientist with the Health and Care Professions Council. She is currently working in a private London IVF unit. Anastasia worked in preimplantation embryo genetics as a molecular geneticist in the Centre for PGD at University College London from 2004 to 2007 and completed her MSc in prenatal genetics and fetal medicine in 2007 at University College London.
Abstract This is a retrospective study aiming to assess telomere length in human embryos 4 days post fertilization and to determine
whether it is correlated to chromosomal ploidy, embryo developmental rate and patient age. Embryos were donated from patients undergoing treatment in the assisted conception unit. Seven couples took part, generating 35 embryos consisting of 1130 cells. Quantitative fluorescent in-situ hybridization (FISH) measured the telomere length of every cell using a pan-telomeric probe. Conventional FISH on six chromosomes was used to assess aneuploidy in the same cells. Maternal and paternal age, referral reason, embryo developmental rate and type of chromosomal error were taken into account. Chromosomally abnormal cells were associated with shorter telomeres than normal cells for embryos that were developmentally slow. Cells produced by women of advanced maternal age and those with a history of repeated miscarriage tended to have substantially shorter telomeres. There was no significant difference in telomere length with respect to the rate of embryo development 5 days post fertilization. Telomeres play an important role in cell division and shorter telomeres may affect embryonic ploidy. Reduced telomere length was associated with aneuploid cells and embryos from women of advanced maternal age. RBMOnline ª 2014, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved. KEYWORDS: aneuploidy, blastocyst, human, preimplantation embryo, quantitative FISH, telomere length
Introduction Telomeres are six base-pair repeats (TTAGGG) found at the ends of chromosomes to protect them from degradation
(Chan and Blackburn, 2002). Telomeres participate in processes of chromosomal repair, prevent nonspecific chromosomal recombination and help the chromosome bind to the nuclear matrix (Counter et al., 1992; De Lange, 1992). Prior
http://dx.doi.org/10.1016/j.rbmo.2013.12.010 1472-6483/ª 2014, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved.
Telomere length in human blastocysts to mitotic cell division, cellular DNA is duplicated by the action of a DNA polymerase. This enzyme does not replicate the ends of chromosomes and so telomeres get shorter after each cell division. In normal somatic cells, this shortening contributes to cellular senescence (Blasco et al., 1997). Telomere shortening limits the regenerative capacity of cells and is correlated with the onset of cancer, ageing and chronic diseases (Djojosubroto et al., 2003; Rudolph et al., 1999; Satyanarayana et al., 2003). Some specialized cells maintain their capacity to divide and regenerate their telomeres due to the action of an enzyme called telomerase. Telomerase is a reverse transcriptase that elongates telomeres in some tissues (i.e. germ and stem cells; Flores et al., 2006). Telomerase activity declines during late oogenesis and early preimplantation development, so maximum telomere length is thought to be established during early oogenesis (Wright et al., 1996). The telomere theory of reproductive senescence supports the idea that telomere shortening in late exit from the fetal production line and a long interval before ovulation in the adult thus causes reproductive ageing in women (Blasco et al., 1999). For example, in null mice for the telomerase gene, as telomeres shorten across generations, oocyte and embryo development deteriorates (Hassold and Hunt, 2001). Chromosomes frequently malsegregate during female meiosis leading to aneuploidy, failed implantation and miscarriage (Harper et al., 2004; Keefe et al., 2006; Munne ´ et al., 1995). The mechanism by which telomere shortening causes this misalignment is not well understood (Blasco et al., 1999), but could result from improper pairing of homologous chromosomes and defective formation of synapses and bouquet formation during early meiosis (Rudolph et al., 1999; Siderakis and Tarsounas, 2007). Meiotic progression and chromosome alignment at the metaphase stage in oocytes of first-generation null mice for the telomerase gene were comparable to that of the oocytes of wild-type mice; however, fourth-generation null mice showed chromosome misalignment, disruption of meiotic spindles and very short telomeres (Liu et al., 2002). In humans, telomere length of replicating somatic cells is inversely related to donor age (Vaziri et al., 1993), highly variable among donors of the same age (Slagboom et al., 1994), highly heritable (Jeanclos et al., 2000) and greater in women than in men (Benetos et al., 2001). In contrast to the considerable information known about telomere length in human extrauterine life, little is known about telomere length during embryonic development. Experiments on the telomere length at different stages of bovine and mouse early development have shown that telomeres at the morula–blastocyst transition reset their length to a specific set point (Schaetzlein et al., 2004), supporting the theory of telomere restoration in mammalian preimplantation development. Xu and Yang (2000) demonstrated that telomerase activity is detected at fertilization in bovine embryos created by IVF and reaches its highest level at the blastocyst stage. Recent data show that telomere length is critically related to fragmentation in IVF embryos in humans but further studies are needed (Keefe and Parry, 2005). This study measured the telomere length in embryos at 5 days post fertilization and assessed whether there is an association between the chromosomal complement of the embryo and its developmental rate using quantitative and
625 conventional fluorescence in-situ hybridization (FISH). The aim of this study was to start unravelling the correlation between telomere length and chromosome aneuploidy in human preimplantation embryos, a relationship that has not been studied before in such detail.
Materials and methods Experiments were performed in preimplantation embryos 5 days post fertilization from patients undergoing fertility treatment with chromosomal screening. Seven couples donated embryos that were unsuitable for transfer or cryopreservation after being diagnosed as abnormal by aneuploidy screening on the day 3 of embryo development. The data comprised 35 embryos consisting of 1130 blastomeres. Two control samples were used in each experiment along with the embryos: (i) maternal and paternal lymphocyte suspensions; and (ii) a human cell line of immortalized keratinocytes called HaCaT (German Cancer Research Centre, DKFZ, Heidelberg, Germany). The parental lymphocyte suspensions were also used to measure the efficiency of the quantitative and interphase FISH techniques. HaCaT cells were used as a positive control since they are modified so as to have standard telomere length of 4 kb (Boukamp et al., 1988). The HaCaT cell line allowed the conversion of the fluorescence intensity in each embryo cell to kilobases since fluorescence intensity has been found to be proportional to telomere length (Lansdorp, 1996).
Patient details and ethical consent Patients who undergo chromosomal screening are referred for three main reasons: (i) those of advanced maternal age (AMA; females 40 years of age); (ii) couples with history of repeated miscarriage (RM, more than three times); and (iii) those with repeated failed IVF cycles (repeated implantation failure, RIF, more than three times). Research on all embryos donated to this study was carried out under licenses from the Human Fertilisation and Embryology Authority (HFEA) of the UK. Informed research consent was obtained from all couples. Approval was obtained by the Research Ethics Committee (REC3), North London, UK (reference no. 10/H0709/26, approved 8 June 2010).
Lymphocyte culture and counts Lymphocyte culture from parental blood samples were carried out by standard methods and suspensions were prepared to be used in each experiment.
Embryo spreading Embryos showing aneuploidy following a clinical preimplantation genetic screening (PGS) cycle and donated for research were spread on slides using the method previously described by Harper et al. (1994). The position of every blastomere was mapped so that it could be identified in each round of FISH. The method has been validated extensively and approved by the HFEA to be used as a diagnostic test for genetic screening. The overall efficiency (percentage of cells
626 with diploid signals) of FISH in lymphocyte samples ranged from 89% to 96%, and no polymorphisms were found in any of the couples. There has been no misdiagnosis from the follow-up analysis of abnormal embryos.
Quantitative FISH on telomere repeats, slide scanning and cell analysis Telomere length was measured using quantitative FISH on interphase nuclei (embryos, lymphocytes and keratinocytes) with a first round of FISH (Figures 1–3, respectively). The quantitative FISH telomere kit, a FISH Kit/Cy3 conjugated with a peptide nucleic acid (PNA) probe, was provided by DAKO Scientific (Denmark), according to the manufacturer’s instructions with a few modifications. The lymphocyte suspensions and embryo spreads were immersed in Tris-buffered saline (TBS) for 2 min, washed in 3.7% formaldehyde (Sigma, USA) in TBS for 2 min and then washed twice in TBS for 5 min each. Lymphocyte suspensions were immersed in pretreatment solution (40 ll of 2000· proteinase K solution diluted in 50 ml TBS) for 15 min; embryo spreads were only incubated for 10 min (both at room temperature). This was because embryo suspensions were prepared through a more effective protocol to remove all cytoplasmic membranes (see section on embryo spreading) than lymphocyte nuclei suspensions that were prepared with hypotonic solution treatments resulting in a less effective cytoplasmic dispersion. Then, the slides were washed twice in TBS for 5 min. A series of dehydration washes followed (70%, 90% and 100% ethanol). The telomere PNA probe was added to each slide (3 ll on each 13-mm coverslip) and they were codenatured at 80C for 5 min. The slides were left to hybridize at room temperature in a dark chamber for 30 min. In a dark room, the samples were washed in rinse solution (1:50 Rinse solution, double-distilled water; DAKO) for 1 min at room temperature to remove the coverslips, immersed for 5 min in wash solution (1:50 wash solution, double-distilled water; DAKO) prewarmed to 65C and dehydrated as before. When slides were dry, 10 ll of DAPI/Vectarshield solution was added to a coverslip (22 · 50 mm), which was then placed on the slide. The slides were stored in the dark for at least 15 min at 4C. Slide scanning, cell identification, intensity measurement and quantification of telomere length were performed by fluorescence microscopy (Olympus BX 40) attached to a cooled charge-coupled device camera and equipped with capturing software (SmartCapture; Vysis, UK). DAPI (blue) and Cy3 (orange) fluorescence (telomere) signals were captured on interphase nuclei at ·1000 magnification. Exposure at the orange filter was manually set at 1.00 and gain at 2.00 and exposure at the blue filter was set at 0.1 and gain at 1.0. The maximum time of exposure was set at 5 s. After capturing, DAPI fluorescence was manually removed from the SmartCapture image produced and the photo was exported to IPLab software (Spectra Services) to be quantified. Telomere signals were manually selected in the focal plane that gave most in-focus hits in each nucleus. IPLab was set so that it did not normalize the results. The selected signals were then averaged giving one quantified result. Background noise variation was minimized by excluding any noise that appeared as a halo. This approach was limiting as some telomere signals had to be excluded. Unfortunately, this is a
A Mania et al. limitation that could not be overcome since embryonic cells are rarely found in their metaphase stage. Capturing was performed in the same light conditions (semi-darkness) in every experiment. Up to 50 interphase nuclei were captured from each parental lymphocyte suspension and HaCaT cell line and every blastomere was captured and mapped from day-5 embryos. Each experiment was analysed within 24 h to avoid fluorescence decay through time. By capturing the images of fluorescently labelled nuclei at the same time interval after hybridization in all experiments, it was ensured that the results were comparable. Changes in the luminosity of the light source of the microscope have also been found to play an important role in reproducibility of results (Narath et al., 2005). Using fluorescently labelled beads at the beginning and end of each experiment monitored this. To normalize between experiments for the quantitation of telomere length, a cell line of standard telomere length was used as previously described. Up to 50 HaCaT cells were captured in each experiment and the fluorescence intensity of the orange filter was quantified. The mean value corresponded to 4 kb and therefore every other fluorescence intensity could be translated into kilobases for the cells analysed in the same experiment. The normalization of the fluorescence intensity values for single cells (measured in intensity units, IU) was calculated by dividing the fluorescence value of the pan-telomeric signals by the fluorescence intensity of the reference probe and increased the reproducibility of the results. This approach allows the comparison of different experiments, the measurement of slides at different times after hybridization and the compensation for variations of the light source. Parental lymphocyte controls had consistently high FISH efficiency of more than 85% for the chromosomes tested. The FISH protocol used to test for the six chromosomes in the parental lymphocytes and embryos has been developed for clinical purposes and its efficiency has been optimized to achieve best results. It was used instead of hybridizing a FISH probe together with the telomere probe as an internal control (Narath et al., 2005), a combination that needed optimization and potentially could compromise the hybridization efficiency of the telomere probe and therefore the quantitation. Capturing and scoring criteria to ensure a comparable quantitative approach between experiments for telomere length measurement were set so as to avoid variation between experiments seen in previous studies (de Pauw et al., 1998). The criteria were: (i) exclusion of nuclei with artefacts such as background noise; (ii) exclusion of nuclei with less than five signals; (iii) focusing of nuclei before capturing to include most signals in one planel; (iv) calibration of the microscope with fluorescent beads before and after each experiment so as to make sure there is no decay in the UV light source; (v) FISH and capturing performed under the same dim light conditions; and (vi) capturing was completed within 24 h from the completion of the FISH procedure.
Conventional FISH for aneuploidy screening After washing off the telomere PNA probe, conventional FISH analysis of all embryos and their controls was performed as previously described by Mantzouratou et al. (2007) using probes for chromosomes 13, 18 and 21 in one round of hybridization and 15, 16 and 22 in the second round.
Telomere length in human blastocysts
627
Figure 1 Telomere quantitative FISH signals (red) with (right) and without (left) DAPI stain on embryonic cells. Bars = 1 lm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Embryo categories according to chromosomal abnormalities Embryos were grouped according to the chromosomal constitution of their cells (Table 1). Embryos with meiotic errors had the same chromosomal error(s) in all cells and
were classed as uniformly aneuploid. Embryos with post-zygotic errors were grouped according to the number of different cell lines that were present in the embryo and the proportion of the embryo that was aneuploid. When there were two cell lines in the embryo, it was classed as a mosaic embryo; when there were several different cell
628
A Mania et al.
Figure 2 Telomere quantitative FISH signals (red) with (right) and without (left) DAPI stain on lymphocyte cells. Bars = 1 lm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
lines in the embryo, it was classed as chaotic; when 50% of the embryo had abnormal cells it was defined as a major mosaic or a major chaotic embryo.
Data and description of the statistical model The data were summarized using descriptive statistics and applied a more advanced modelling based on hierarchical specification (German and Hill, 2006). Hierarchical models are particularly effective in dealing with data grouped or nested in different layers. In this work, cells were grouped within embryos, which were grouped within couples. It was plausible to assume that data on cells coming from the same embryo would tend to be correlated (i.e. to share some common features) and that they would tend to be somehow different from other observations relating to
other embryos. Similarly, embryos donated by the same couple would share some common characteristics and therefore would tend to have some correlation and to present differences compared with embryos of other couples. Analysing the data using procedures like t-tests or standard regression (which are based on the assumption of independence among the observations) does not recognize these different forms of correlation, therefore producing biased estimations. On the contrary, hierarchical models (sometimes referred to as ‘multilevel’) proceed by defining a structure for each layer in the data, specifying a suitable form of correlation among data in the same group. The information for a given unit is used to inform all units in the same group (a phenomenon known as ‘borrowing strength’), which accounts for the underlying uncertainty in a more appropriate way. The objective of this statistical analysis was to investigate the relationship between telomere length and the occurrence of chromosomal
Telomere length in human blastocysts
629
Figure 3 Telomere quantitative FISH signals (red) with DAPI stain (blue) on keratinocyte cells. Bars = 1 lm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Table 1
Embryo characterization: proportion of euploid and aneuploid cells in each embryo group.
Embryo characterization
Uniformly abnormal Minor mosaic Major mosaic Minor chaotic Major chaotic
Euploid cells
0 >50 50
