The dawn of the future: 30 years from the first biopsy of a human embryo. The detailed history of an ongoing revolution

Abstract

Following early studies showing no adverse effects, cleavage stage biopsy by zona drilling using acid Tyrode’s solution, and removal of single blastomeres for preimplantation genetic testing (PGT) and identification of sex in couples at risk of X-linked disease, was performed by Handyside and colleagues in late 1989, and pregnancies reported in 1990. This method was later used for specific diagnosis of monogenic conditions, and a few years later also for chromosomal structural and/or numerical impairments, thereby establishing a valuable alternative option to prenatal diagnosis. This revolutionary approach in clinical embryology spread worldwide, and several other embryo biopsy strategies developed over three decades in a process that is still ongoing. The rationale of this narrative review is to outline the different biopsy approaches implemented across the years in the workflow of the IVF clinics that provided PGT: their establishment, the first clinical experiences, their downsides, evolution, improvement and standardization. The history ends with a glimpse of the future: minimally/non-invasive PGT and experimental embryo micromanipulation protocols. This grand theme review outlines a timeline of the evolution of embryo biopsy protocols, whose implementation is increasing worldwide together with the increasing application of PGT techniques in IVF. It represents a vade mecum especially for the past, present and upcoming operators and experts in this field to (re)live this history from its dawn to its most likely future.

Introduction

To perform a genetic test on a preimplantation embryo, preimplantation genetic testing (PGT; previously known also as preimplantation diagnosis, PID, or preimplantation genetic diagnosis, PGD) requires access to genomic DNA, if possible, with minimal immediate damage to the embryo or effects on implantation and later development following transfer to the uterus. This normally requires one or a small number of cells to be removed or ‘biopsied’ from the embryo. Embryo biopsy has its origins in attempts in the 1960s and early 1970s to sex the embryos of farm animals to selectively breed, for example, dairy cows. At the blastocyst stage, the bovine embryo can grow to several millimeters across enabling a section of the outer trophectoderm (TE) layer to be cut off by hand with a small scalpel blade, avoiding the embryonic disc, a method still in use today. Similarly, in 1968, Richard Gardner, one of the pioneers of micromanipulation of mammalian embryos and a student of Robert Edwards, used a holding pipette to position rabbit blastocysts and a pair of fine iridectomy scissors, normally used for eye surgery, to biopsy small pieces of TE aspirated into a fine pipette (Gardner and Edwards, 1968). The TE cells were then stained and examined microscopically for the presence of a dense Barr body on the periphery of the interphase nuclei indicating the heterochromatic inactive X chromosome in females. The male and female blastocysts were selectively transferred to different recipient females and the sex confirmed later in gestation. Interestingly, they found one abnormal anencephalic fetus, but this was later attributed to a high incidence in that strain of rabbits.

Gardner and Edwards speculated that a similar approach in humans could potentially be used in couples at risk of X-linked inherited diseases to avoid the transfer of affected male embryos (Gardner and Edwards, 1968). However, their prescient comments were largely overlooked for 20 years and only credited retrospectively. It would take another 10 years for the first IVF birth and a further 10 years before human embryo culture and development was good enough to consider embryo biopsy. Also, unlike farm animals, the human blastocyst only reaches a maximum size of 200–300 μm and a few hundred cells before implantation. Human embryo biopsy, therefore, requires micromanipulation, and sensitive methods are needed for genetic analysis of a single or few cells biopsied from the embryo, which was not possible before the PCR, which enables a million-fold amplification of target DNA fragments, was developed in the mid-1980s (Saiki et al., 1985, 1986). Wilton and Trounson, (1989) later demonstrated that, in the mouse, blastomeres removed from the embryo at the 4-cell stage could be cultured and in some cases divided into outgrowths of multiple cells. Also, a study in the human embryo demonstrated that isolated blastomeres co-cultured with the biopsied embryo underwent cell division at the same rate and then either arrested or formed miniature blastocysts (Geber et al., 1995). However, these miniature blastocysts could not be maintained in culture and did not significantly increase in cell number.

PGT was first achieved in the mouse, in a model of the human-inherited disease Lesch–Nyhan syndrome, a lethal X-linked disease caused by deficiency of the enzyme hypoxanthine phosphoribosyl transferase, using cleavage stage biopsy and a single cell biochemical assay (Hooper et al., 1987; Monk and Handyside, 1988). In the mouse, it was possible to biopsy 8-cell embryos simply by removing the zona pellucida (ZP), incubating briefly in Ca²⁺/Mg²⁺-free medium and pipetting gently to separate single cells from the rest of the embryo using a stereomicroscope. The embryos were returned to normal medium and allowed to develop to the blastocyst stage before transfer, while the single cells were used for testing. Single cells or blastomeres could be removed (or added) at this stage since they remain totipotent and capable of contributing to all embryonic and extraembryonic lineages. Zona-free cleavage stage embryos, however, do not survive in the uterus following transfer so it was essential to culture them to the blastocyst stage before transfer. Alternatively, mouse blastocysts could also be biopsied by hand, using a microneedle and first making a slit before pinching off the herniating TE cells after a further period in culture. However, without micromanipulation at higher magnification, it is difficult to ensure that the inner cell mass (ICM) is not included in the biopsy.

In the mid-1980s, human embryos were transferred no later than day 3 post-insemination and although embryos could be cultured to the blastocyst stage in media supplemented with maternal serum, pregnancies from later transfers were rarely successful. There was no alternative therefore but to consider cleavage stage biopsy at that time and the first crude attempts to biopsy human embryos by bisecting early cleavage stages for biochemical analysis were reported at the American Society for Human Genetics in 1985 by Verlinsky, Pergament and colleagues.

In this review, we describe the development and refinement of cleavage stage embryo biopsy in clinical practice, which was the predominant method of embryo biopsy for over 10 years, the parallel development of polar body (PB) biopsy to examine maternal genetic defects, and the eventual evolution towards TE biopsy at the blastocyst stage facilitated by the development of efficient cryopreservation of biopsied embryos using vitrification. Finally, with the discovery of cell-free DNA in the blastocoel fluid (BF) and culture medium, we briefly review the prospects for minimally−/non-invasive PGT.

Search method

A search in Scopus, Pubmed and Isi Web of Science for the peer-reviewed papers published in English from 1989 through April 2019 and concerning human embryo biopsy was conducted. The keywords were ‘IVF’ AND ‘biopsy’ AND ‘human’ AND (‘embryo’ OR ‘cleavage stage’ OR ‘blastomere’ OR ‘polar body’ OR ‘oocyte’ OR ‘blastocyst’ OR ‘trophectoderm’ OR ‘morula’ OR ‘blastocoel’ OR ‘blastocentesis’ OR ‘culture media’) AND (‘preimplantation genetic testing’ OR ‘preimplantation genetic diagnosis’ OR ‘preimplantation genetic screening’). This narrative review was mainly focused on the studies that introduced a relevant novelty or a significant improvement in the history of embryo biopsy.

The past

1989–1996: the dawn of human embryo biopsy approaches

Blastomere biopsy: from the first experiment to the first pregnancies

From 1953 to 1967, several crucial steps ultimately led Edwards and Gardner to sex rabbit blastocysts (Edwards and Gardner, 1967), namely the discovery of DNA (Watson and Crick, 1953), the definition of the human karyotype with its 46 chromosomes (Ford and Hamerton, 1956; Tijo and Levan, 1956) and the discovery that numeric imbalances in the karyotype were causative of conditions like the Down, Turner and Klinefelter syndromes (Lejeune et al., 1959; Ford et al., 1959; Jacobs and Strong, 1959). However, in the UK, it was the first publications demonstrating the power of PCR to amplify short fragments of DNA to perform a genetic diagnosis from as few as 100 cells (Saiki et al., 1985, 1986) that stimulated discussion on the possibility of conducting PGT on human embryos (Penketh and McLaren, 1987; Edwards and Hollands, 1988). The pioneers of IVF realized that a close tie exists between clinical embryology and genetics, and hence paved the way to the introduction of this game-changing tool in modern ART. A meeting was organized in London in November 1986, which brought together experts from diverse disciplines (Whittingham and Penketh, 1987). The consensus view, promoted by Robert Edwards based on the early rabbit work, was that blastocyst biopsy would be most likely to succeed, mainly because several cells could be biopsied potentially. However, Handyside working with Lord Winston, who directed a large IVF clinic at Hammersmith Hospital in London, realized this was not feasible at that time because of poor clinical outcomes and was committed to developing cleavage stage biopsy methods and single cell analysis.

Cleavage stage biopsy had a number of advantages. All of the normally fertilized embryos reaching the 6- to 10-cell stage early on the morning of the third day post-insemination could be biopsied in a single micromanipulation session. Biopsy of one or two single cells could be done with minimal physical damage and removed a controlled and reproducible proportion of the embryo. Critically at that time, the rest of the day was available for the single cells to be lysed, the DNA amplified and the products analyzed. Cryopreservation using slow freezing methods was possible then but often damaged some cells of the embryo, particularly after a slit or hole was made in the zona for biopsy, which compromises the protective effect of the zona. Hence, the challenge was to have the PGT result within 6–8 h.

The original biopsy protocol developed at Hammersmith involved transferring the embryo to a micromanipulation dish in a droplet of culture medium (supplemented with 10% maternal serum) under oil and holding it in position with a holding pipette (Supplementary Video). A fine pipette filled with acid Tyrode’s solution was then brought in on the opposite side and the acid carefully expelled onto the outer surface of the zona, which rapidly dissolved the zona proteins. This method of ‘zona drilling’ had been reported in the mouse as a possible approach to overcoming fertilization failure in oligospermia (Gordon and Talansky, 1986). Once the zona had been fully penetrated, the pipette was immediately removed from the drop and a larger sampling pipette brought in to aspirate a single blastomere. Compared with practice today, this process could be very slow (up to 10 min) as the cell was gently pulled away from adjacent cells to which it was adhering or ‘compacting’. In the mouse, Ca²⁺/Mg²⁺-free medium was routinely used to de-compact embryos at the morula stage, but prolonged exposure reduces viability after transfer. At the time, therefore, this was avoided. Fortunately, compaction was delayed in the medium used until late day 3 or day 4, and no temperature-controlled stage was used. Hence, the low-ambient room temperature is likely to have had a mildly de-compacting effect. Before clinical application, it was essential to demonstrate that the acid Tyrode’s drilling and biopsy had no adverse effect on the viability and development of the embryo to the blastocyst stage in vitro (Hardy et al., 1990). A series of embryos were biopsied, and one or two cells removed on day 3. Cell numbers in the ICM and TE of the biopsied embryos cultured to the blastocyst stage were then compared to controls. As expected, although cell numbers were reduced, they were only reduced in proportion to the fraction of cells removed. With the mouse embryo, removal of cells at cleavage stages also results in blastocysts with proportionately fewer total cells. However, after transfer to the uterus and implantation, cell division rate increased and the size of the offspring at birth was similar to intact embryo controls. Energy substrate uptake as a measure of metabolism was also not disproportionately affected. Similarly, promising data were reported also by Wilton and Trounson, who were trying to set up a blastomere biopsy protocol in Australia (Wilton and Trounson, 1989).

The first clinical cleavage stage biopsies and single cell PGT to identify the sex of the embryos were performed in October 1989, in a series of patients carrying X-linked inherited disorders, including adrenoleukodystrophy, X-linked mental retardation and Duchenne muscular dystrophy. Despite concerns that the cleavage stage embryo would escape through the relatively large hole in the protective zona after transfer and be destroyed or lost, five of eight women became pregnant, 10 of 22 (45%) embryos implanted, seven (32%) reached the fetal heart stage and the first female twins were born in July 1990 (Handyside et al., 1990). The authors were conscious of the enormous potential of this approach; however, they also acknowledged that several aspects still required deeper investigation, such as the potential effect of embryo manipulation at the cleavage stage on further development. Following this initial success, this cleavage stage biopsy protocol was replicated in clinics worldwide and continued to be used for many years. However, improvements in culture medium and the use of alternatives to maternal serum accelerated the development and compaction of embryos, eventually necessitating the use of Ca²⁺/Mg²⁺-free medium to reverse compaction for biopsy. This more aggressive approach is effective but has recently been shown to be detrimental to the subsequent development of the embryo (Kirkegaard et al., 2012; Bar-El et al., 2016).

Nowadays, PGT might be reliably adopted to test conditions such as monogenic diseases (PGT-M; previously also known as PGD), structural rearrangements (PGT-SR) and aneuploidies (PGT-A; previously also known as preimplantation genetic screening, PGS) in the human embryos produced during an IVF cycle (Zegers-Hochschild et al., 2017a, 2017b). In this review, each little milestone will be brought back to light on the pathway from the very first successful clinical biopsy of a human embryo down to the definition of the gold standard current approaches of a still ongoing revolution. Figure 1 represents a roadmap of the main events that outlined the past of human embryo biopsy from its dawn to its establishment down to the first concerns that arose in the scientific community.

TE biopsy: establishment and first refinements

Despite the astonishing results achieved with blastomere biopsy at the cleavage stage, it was soon hypothesized that several TE cells retrieved from a human blastocyst might represent an even more suitable specimen for genetic testing. Promising results were already described with mouse and marmoset blastocysts (Monk et al., 1988; Summers et al., 1988), and from a clinical perspective this approach would provide more material for the analysis from an extra-embryonic tissue (the ICM that gives origin to the fetus is kept untouched) and at a stage of preimplantation development when cryopreservation is even more successful. Therefore, based on these assumptions, Dokras and colleagues, inspired by Edwards and Gardner’s experiments in rabbit (Edwards and Gardner, 1967), reported the very first experience with the biopsy of a human blastocyst (Dokras et al., 1990). The procedure involved donated early blastocysts from IVF patients. Once the embryos reached the blastocyst stage, they were moved to hanging drops of Tyrode’s 6 (T6) medium in a chamber with liquid paraffin oil. The blastocysts underwent mechanical zona opening through a 1–2 min procedure, and ~18–24 h later the herniating TE cells were excised free-hand under a stereo dissecting microscope through a 1-min procedure. The method used to perform the TE biopsy involved a holding pipette and a needle in two different micromanipulators. The needle was introduced through the ZP in a site opposite the ICM and pierced back out, while the holding pipette was used to detach the blastocyst by friction on the needle with its lower edge. An opening between the two points pierced by the needle was thereby formed. Dokras and colleagues defined three parameters that must be set: the position of the slit (opposite the ICM to allow the herniation exclusively of TE cells), the day of preimplantation development (day 5–6 better than day 7) and the size of the slit (the bigger the slit, the more the cells that would herniate). Most of the biopsy procedures were conducted when ‘the size of the herniation was approximately the same as the blastocyst’; yet, the authors themselves disclosed some concerns dealing with this aspect since the removal of too many cells could affect the production of hCG and impair embryo implantation potential, as already suggested by Summers and colleagues previously (Summers et al., 1988). Therefore, 1 year later, Muggleton-Harris and Findlay developed a more elegant protocol to perform TE biopsy that did not entail assisted hatching (Muggleton-Harris and Findlay, 1991). The spare embryos obtained from the King’s College Hospital, London, underwent acid Tyrode’s solution-mediated ZP thinning and the mechanical removal of three TE cells, one by one. In both the studies, the blastocysts were cultured to day 14 to examine the outgrowths in vitro.

The early methods of blastocyst biopsy were highly aggressive and therefore were not tested in the clinical practice of IVF laboratories.

PB biopsy: an alternative to embryo biopsy to test genetic conditions

Verlinsky and colleagues are the pioneers of PB biopsy for PGT both for aneuploidy and monogenic diseases (Verlinsky et al., 1990). Initially, they suggested that preconception testing of the first PB alone (extruded by the mature oocyte) might be possible for maternal contributions to monogenic diseases. However, they soon realized that, because of recombination, analysis of both first and second polar bodies (extruded following fertilization) was necessary for accuracy. PB analysis had the advantage that 2 days were available to complete the genetic analysis before fresh transfer of unaffected embryos on day 3. Also, if either of the PBs failed to amplify a blastomere could be biopsied to verify the diagnosis. According to Verlinsky’s method, a sampling pipette was used to pierce the ZP and collect the PB, which was then analyzed by PCR with the probes designed to target the mutations under investigation. The whole procedure was suggested as minimally invasive in comparison to both blastomere and TE biopsies and therefore implemented clinically for several indications performed at their center during the following years.

The introduction of fluorescent in situ hybridization and whole genome amplification: at the roots of single cell analysis

Human reproduction is an inefficient process to the point that the maximum chance of pregnancy during natural conception even in young couples is estimated as ~30% (Slama et al., 2002). Most of the embryos produced either in vivo or in vitro arrest their development between fertilization and implantation, while some others might result in a miscarriage. Full chromosome meiotic aneuploidies are the main reason for these failures and, with some exceptions (viable trisomies and sex chromosome aneuploidies), do not result in a live birth. The main cause for chromosomal aneuploidies in humans is the impairment of meiotic segregation during oogenesis and, to a smaller extent, during spermatogenesis or the mitotic divisions post-fertilization (Capalbo et al., 2017; Hassold et al., 2007; Hassold and Hunt, 2001; Nagaoka et al., 2012). Maternal age is indeed the main variable associated with both lower oocyte quantity and quality that parallels an aneuploidy rate among the blastocysts produced, which dramatically increases from a ~30% baseline in women aged ≤35 years to >90% in women older than 42 years (Franasiak et al., 2014). Also, structural aneuploidies might impair human reproduction success, but they are more probably arising from de novo events in oogenesis and spermatogenesis as well as post-fertilization (Capalbo and Rienzi, 2017; Zhou et al., 2018; Babariya et al., 2017; Escriba et al., 2019). Therefore, it is not surprising that the implementation of aneuploidy testing among the analyses feasible on an embryo biopsy did not take long.

The first authors to introduce fluorescent in situ hybridization (FISH) for embryo sexing in the clinical practice were Griffin and Delhanty with their colleagues (Griffin et al., 1991, 1992), with some improvements implemented by Harper in 1994 (Harper et al., 1994). However, Munnè and colleagues were the first to adopt FISH on a single blastomere also to test autosomes (Munne et al., 1993). Specifically, they also included chromosomes 13, 18 and 21 in the analysis, chosen because their numeric imbalances globally represent ~95% of all the aneuploidies identifiable in the products of conceptions after a miscarriage in humans. In the following years, FISH was further refined until it could also detect chromosome structural rearrangements, such as translocations (Scriven et al., 1998).

In parallel with the introduction of FISH, the depth of single gene analysis was boosted through whole genome amplification (WGA) protocols aimed at scaling up the amount of template DNA submitted to PCR. In 1994, Kristjansson and colleagues adopted WGA in IVF and simultaneously tested multiple mutations of the dystrophin gene from a single blastomere (Kristjansson et al., 1994). They used a primer extension pre-amplification protocol targeting the chromosome X. The same group (Snabes et al., 1994), soon after, further refined the protocol to simultaneously analyze the main mutations causative of Tay–Sachs disease, cystic fibrosis, hemophilia A and Duchenne muscular dystrophy together with sexing of a single blastomere. The introduction of WGA in IVF was an outstanding idea, which was first clinically applied for accurate detection of an autosomal dominant defect predisposing to bowel cancer (Ao et al., 1998) and remains in widespread use today. DNA pre-amplification protocols have been indeed improved (e.g. multiple displacement amplification and advanced PCR-based methods) across the years to be adapted to comprehensive chromosome testing (CCT) techniques aimed at the analysis of the whole embryonic karyotype, such as array comparative genomic hybridization (array CGH) (Wells and Delhanty, 2000; Wells et al., 1999), single nucleotide polymorphisms array (SNP array) (Treff et al., 2010), quantitative real time PCR (qPCR) (Treff et al., 2012) and lately also next generation sequencing (NGS) (Fiorentino et al., 2014).

The first international PGT data monitoring: introducing the concepts of consistency and reproducibility

After the very first multicenter report published by Verlinsky and colleagues in 1994 (Verlinsky et al., 1994), it was Joyce Harper in 1996 who described the clinical experience with embryo biopsy and PGT of 14 IVF centers, who shared their data in view of a higher international standardization, reproducibility and efficiency (Harper, 1996). Overall, 197 cycles resulted in 171 transfers, 50 pregnancies, 28 deliveries and 34 live births. One year later, the ESHRE PGT Consortium was founded at the 13th ESHRE annual meeting held in Edinburgh. The duty of this Consortium still is to monitor the application, efficacy, clinical outcomes and long-term effect of PGT, as well as to provide the operators with guidelines for good laboratory practice. In 1999, they published their first data report (Geraedts et al., 1999), then it took a further 6 years for the definition of the best practice guidelines for biopsy (Thornhill et al., 2005). Those recommendations have been revised in 2011 (Harton et al., 2011) and a further update is currently in the pipeline.

1997–2008: perfecting the protocols and facing the first concerns

The second phase in the history of biopsy was the consequence of wider access worldwide to this procedure. Several centers implemented the three approaches outlined in the first phase and revised the protocols by adapting them and, in some cases, expediting the procedures. Similarly, the first concerns arose and shed light on the putative clinical, biological and technical pitfalls of PGT.

First more thorough tests to improve blastomere biopsy protocol

Some important studies were published shortly after the first report of the clinical biopsy of a human blastomere, which were critical in the evolution of such a protocol. For instance, Tarin and colleagues back in 1992 suggested that biopsies conducted at early stages of preimplantation development (e.g. day 2 post-insemination) can significantly affect further embryo development in terms of length of in vitro culture and number of cells constituting the deriving blastocyst (Tarin et al., 1992). Geber and colleagues instead tested a blastomere biopsy protocol involving the removal of one or two cells at the cleavage stage, which were then co-cultured with the embryo for 3 days more. These blastomeres kept dividing, up to more than four cells in most of the cases (Geber et al., 1995). Such a protocol represented an intriguing alternative to blastocyst biopsy to increase the number of cells that might be analyzed.

Expediting the protocols: introduction of two pipettes-based method, Ca²⁺/Mg²⁺-free medium, blastomere displacement method and laser

Chen and colleagues introduced the two-pipette blastomere biopsy technique and designed a randomized controlled trial (RCT) to compare it with the conventional three-pipette one (Chen et al., 1998). While the three-pipette approach entailed a drilling pipette containing acidified Tyrode’s solution to make a ~40 μm hole in the ZP plus a biopsy pipette to remove the selected blastomere, Chen and colleagues’ new protocol involved the use of a single pipette through which the acid solution was blown onto the ZP and then sucked back again before removing the blastomere with no need to change the pipette. This novel approach involved similar technical (successful biopsy, fixations and positive FISH signal rates) and biological (blastulation rate) outcomes, but in almost a quarter of the time required to perform the three-pipette method. Evidence from an independent study published the same year from a different group confirmed this protocol was reproducible (Inzunza et al., 1998). In parallel, Dumoulin and colleagues introduced the use of Ca²⁺-Mg²⁺ free HEPES-buffered medium for blastomere biopsy (Dumoulin et al., 1998). To this end, they conducted a paired-RCT, where sibling embryos were submitted to embryo biopsy in a HEPES-buffered medium either with or without Ca²⁺-Mg²⁺ and reported better biopsy outcomes (lower blastomere lysis rate) and a faster procedure (half the time required) with the latter medium. A last adaptation of the blastomere biopsy approach was published by Wang and colleagues in 2009 that described the ‘blastomere displacement’ approach (Wang et al., 2009) as an alternative to the ‘blastomere aspiration’ one. According to this protocol, after opening the ZP and blowing some media into it, a positive pressure was sufficient to push out one blastomere. The authors also claimed that the blastomere displacement protocol might be more operator friendly, less damaging and less time consuming than the blastomere aspiration one. Of note, Wang’s protocol entailed the use of a laser to open the ZP, whose introduction in 1994–1995 is probably one of the most important breakthroughs to expedite the biopsy procedure at any stage of preimplantation development (Germond et al., 1995; Obruca et al., 1994). In 1997, Veiga and colleagues reported the very first experience with laser-assisted ZP opening and TE biopsy from donated human blastocysts, then associated with FISH-based aneuploidy testing (Veiga et al., 1997). These authors outlined a protocol for laser-assisted TE biopsy, which is still currently adopted, known as the ‘day 5/6 laser assisted ZP opening and TE biopsy’ method. According to this protocol, fully expanded blastocysts are submitted to laser-assisted ZP drilling and further cultured until few TE cells herniate from that hole and are then excised with the help of some more laser shots targeted to the junctions between the cells. Several important pieces of evidence supporting the safety of non-contact lasers in IVF were then produced in the following years by Chatzimeletiou and colleagues, who tested its effects in comparison with acidified Tyrode’s solution (Chatzimeletiou et al., 2005; Chatzimeletiou et al., 2001).

First and second PB biopsy: the main alternative to embryo biopsy for aneuploidy testing

Verlinsky and colleagues further refined the PB biopsy approach and finally published in 1998 a protocol to conduct aneuploidy testing on both the first and the second PB (Verlinsky et al., 1998). Both the PBs were removed simultaneously after fertilization and submitted to FISH for chromosome 13, 18 and 21 analysis. The definition of the chromosomal constitution of both the PBs portrayed the meiotic segregation and allows the inference of the chromosomal constitution of the oocyte. Clearly, this workflow involved a limited inspection not including possible paternal meiotic and/or mitotic errors, which might occur post-fertilization. Nevertheless, maternal meiotic mis-segregations account for the most prevalent cause of embryonic aneuploidies (Capalbo et al., 2017) and covering at least them allowed a reasonable prediction. Verlinsky and colleagues demonstrated on donated embryos that the FISH-based analysis of both PBs allowed the prediction of the related blastomeres’ chromosomal constitution, and therefore then implemented this protocol clinically. In 1998, they reported the data from more than 500 cycles during which they tested more than 1600 oocytes and achieved 67 healthy live births by transferring 1208 embryos that were diagnosed non-trisomic for chromosome 13, 18 and 21 (Verlinsky et al., 1998).

The aneuploidy testing of both PBs approach suffers from several biological and technical limitations even when adopting CCT techniques (Forman et al., 2013; Handyside et al., 2012; Capalbo et al., 2013a). However, it still represents the only established alternative option to embryo biopsy for PGT. Another protocol exists to retrieve the two PBs, which entails the retrieval of the first PB from the oocyte on day 0 and the second PB after fertilization on day 1 by using the same opening in the ZP (Gianaroli, 2000). This protocol is known as ‘the sequential approach’, and it is essential for monogenic analyses, while representing a valuable alternative to the simultaneous retrieval for chromosomal testing.

Although PBs-based PGT is just marginally adopted at present, a multicenter international RCT funded by ESHRE has been recently published. In this study, ~400 women of advanced maternal age (AMA) underwent either array-CGH-based PGT-A conducted on both PBs or standard IVF. This study was known as the ESTEEM (ESHRE study into the evaluation of oocyte euploidy by microarray analysis) and was published in 2018 (Verpoest et al., 2018). Of note, the patients in the PGT-A arm achieved the same cumulative live birth rate per intention to treat as the standard care arm, but with less miscarriages, less transfers and less embryos cryopreserved. Moreover, some of the embryos undiagnosed after PBs testing were transferred in the absence of sibling euploid embryos and resulted in the same live birth rate as non-biopsied untested ones. In conclusion, the ESTEEM demonstrated that PBs removal might not impair embryo developmental/reproductive competence and that the array-CGH-based aneuploidy testing of these specimens will involve the same efficacy as untested embryo transfer (gamete/embryo intrinsic competence and couple’s chance to conceive are preserved), but a higher efficiency (lower risks for the patients with less transfers).

Ten years of PGT: first doubts and concerns

In August 2004, the three most active clinics providing PGT worldwide gathered the data produced during over a decade of clinical experience (Verlinsky et al., 2004). They reported promising results and very low risk of misdiagnosis from more than 750 live births achieved after either chromosomal testing or single gene analysis. David Hill published a commentary to their manuscript in which he questioned some aspects of PGT when adopted to conduct aneuploidy testing (Hill, 2004). Although giving credit to the higher prediction on embryo ‘implantability’ that ‘brings the goal of one embryo transferred, one healthy baby born closer’ along with lower multiple pregnancy and miscarriage rates, he was concerned about the putative effect of biopsy on both subsequent embryo culture and implantation; the issue of chromosomal mosaicism (i.e. the presence of cells with different chromosomal constitutions in the same embryo due to a mitotic mis-segregation post-fertilization); and the ~50% additional costs involved in PGT with respect to standard IVF. His concerns, in part still currently unresolved, anticipated the biennium 2010–2011: a watershed period in the history of PGT that set the end of its past and the beginning of its present.

2010–2011: a watershed period in the history of PGT

From the theory to the practice: the failure of ‘PGS 1.0’ and the quest for safer and more efficient strategies

In 2008 and again in 2010, the ESHRE PGT Consortium steering committee published two statements on the routine application of aneuploidy testing for indications such as AMA, repeated implantation failure (RIF), recurrent pregnancy loss (RPL) and severe male factor (SMF) (Harper et al., 2010, 2008). Particularly in 2010, this panel of experts acknowledged the absence of RCTs to investigate the utility of PGT-A to treat RIF, RPL and SMF and pointed out 11 RCTs comparing 9-chromosome FISH-based PGT-A to standard IVF in AMA patients that reported the absence of a statistically significant benefit, if not even an impact (Mastenbroek et al., 2007). Ten of these RCTs were conducted with a blastomere biopsy approach and only one with a TE biopsy approach. Based on this evidence, although some groups could still achieve reasonably good results (Rubio et al., 2013; Munne et al., 2010), the ESHRE PGT Consortium suggested that the FISH-based analysis of a blastomere might not represent an efficient workflow for PGT-A. They mainly blamed the limited screening potential of FISH and the putatively high prevalence of chromosomal mosaicism at this stage of embryo preimplantation development. Alongside this, the Consortium nominated an ESHRE PGS Task Force, which came forward with the idea of the ESTEEM using the polar bodies and also encouraged further investigations of TE biopsy as another promising approach. Their vision was in part shared by Mastenbroek and colleagues, who 1 year later published a systematic review and meta-analysis of nine RCTs and also claimed the inefficiency of PGT-A (Mastenbroek et al., 2011). Even though these authors did not question the theory on which PGT-A was based, they also firmly blamed the putative technical drawbacks undermining its clinical application and concluded that ‘new approaches in the application of PGS should be evaluated carefully before their introduction into clinical practice’. This study animated the already existing passionate discussion on the clinical value of PGT-A and its putative pitfalls (Munne et al., 2010), which is still ongoing. In particular, Mastenbroek’s paper set the end of what was later defined as ‘PGS 1.0’ and the beginning of a new era in the history of human embryo biopsy.

The present

2004–2019: integration of vitrification and TE biopsy protocols in a safe and efficient workflow

The present use of embryo biopsy in IVF is mainly based on TE biopsy and its reasoned recent integration in the workflow of a PGT cycle (Fig. 2 represents a roadmap). In 2004 and 2005, the Australian group of de Boer and McArthur published two important papers that detailed a blastocyst biopsy protocol, which is nowadays extensively used worldwide: the ‘day 3 ZP opening-based TE biopsy’ approach (de Boer et al., 2004; McArthur et al., 2005). They outlined a whole clinical workflow for the management of a PGT cycle: ZP opening on day 3, successful blastocyst culture (day 5–6), multicellular (5–6 cells) TE biopsy on the opposite side to the ICM, cryostorage and frozen-thawed unaffected/euploid elective single embryo transfer. The opening of the ZP might be performed through several approaches: mechanically, through acidified Tyrode’s solution, or through a few laser shots (Fig. 3). The protocol described by de Boer and colleagues exploits the laser, mainly because it is the most user-friendly of the three options. It was 2006 when Jones and colleagues provided clinical evidence in favor of such a tool: they conducted a RCT comparing embryos submitted to either acidified Tyrode’s solution-mediated or laser-assisted zona opening in terms of several technical and embryological outcomes (Jones et al., 2006). No difference was reported, thereby supporting a safe implementation of the laser approach. When dealing with the strategy for TE cells removal, two options exist: the ‘pulling’ and the ‘flicking’ approach (Fig. 3). Both of them were mentioned and detailed by McArthur and colleagues in 2005. Specifically, the pulling option was described as ideal for partially hatched blastocysts, while the flicking method was more appropriate for fully hatched ones (McArthur et al., 2005). The pulling option entails the stretching of the selected TE cells along with a few laser shots directed to the junctions between them and the rest of the blastocyst’s body; according to the flicking approach, after loosening the bonds between the selected TE cells and the rest of the blastocyst’s body through a sequence of laser shots, the embryo is released from the holding pipette and moved to its rim, then a vigorous downwards movement with the biopsy pipette finally detaches the chosen cells. Still, at present, no study has ever compared these two strategies for TE fragment removal; therefore, such studies are eagerly awaited. After that, the Australian group outlined a workflow to efficiently implement the blastocyst biopsy approach clinically by associating it with frozen-thawed single embryo transfer, and several groups started adopting it: a process that was even faster after 2013, when vitrification was declared no longer an experimental protocol (Practice Committees of American Society for Reproductive and Society for Assisted Reproductive, 2013) and the first evidence of its safety and efficiency was published (Rienzi et al., 2017).

Vitrification of biopsied blastocysts and biopsy of warmed blastocysts: first experiences

In 2007, Parriego and colleagues reported a live birth after the transfer of a vitrified-warmed biopsied blastocyst (Parriego et al., 2007), a clinical practice that is routinely used nowadays. Two years later, Lathi and Behr delineated a workflow entailing the thawing of previously frozen untested embryos, then submission to TE biopsy and same-day genetic analysis and transfer (Lathi and Behr, 2009). This is an option for centers that have an in-house genetic laboratory equipped to this end and that use diagnostic platforms characterized by a short turnaround time. It is indeed common that either a couple acquires an indication for PGT after completing an IVF cycle with supernumerary embryos, or an IVF center implements PGT and a couple owning previously cryopreserved embryos would like to benefit from their testing. Clearly, if it is impossible to do otherwise, the biopsied blastocysts must be re-cryopreserved. This implies exposing them to the risks related to several manipulations and demands a careful evaluation based also on features such as blastocyst morphology and day of development (Cimadomo et al., 2018a; Bradley et al., 2017; Taylor et al., 2014).

In general, embryo biopsy and cryopreservation are closely interrelated, especially when a TE biopsy-based CCT strategy is adopted. Therefore, an efficient blastocyst culture and vitrification programme is clearly essential to the implementation of PGT in an IVF center.

Two milestone studies in the history of PGT: TE biopsy is safer and more efficient than blastomere biopsy

In 2012 and 2013, Scott and colleagues published two milestone studies in the history of embryo biopsy for PGT purposes. They provided the scientific community with first class data that highlighted the superiority of TE biopsy over the blastomere biopsy approach. The first paper describes the results of a blinded non-selection study (Scott et al., 2012): either cleavage stage embryos or blastocysts were submitted to the removal of a single blastomere or of a TE fragment, respectively, and transferred. Only after defining the downstream clinical outcome (sustained implantation or not), the authors blindly analyzed the biopsy by SNP array. Such a study design is the only one possible to measure the positive and negative clinical predictive values, which are critical features strictly dependent on both the stage of biopsy and the diagnostic technique under investigation. The clinical positive predictive value is the rate of embryos that would have been diagnosed ‘euploid’ by SNP array and then implanted, while the negative predictive value is the rate of embryos that would have been diagnosed ‘aneuploid’ and then did not implant. Notably, while no difference was reported between blastomere and TE biopsy analysis for the latter outcome, the clinical positive predictive value of CCT performed at the blastocyst stage was almost 50% versus about 30% at the cleavage stage (P < 0.01). Nevertheless, it was the study that the same group published in 2013 that eventually highlighted a serious pitfall of blastomere biopsy (Scott et al., 2013). The authors adopted an unbiased, elegant randomized and paired design to define the putative impact of biopsy on embryo implantation potential. Specifically, the two best quality cleavage stage embryos produced by each enrolled couple from the same cohort of inseminated oocytes were randomized to be either submitted to the removal of a single blastomere or kept untouched; then both of them were transferred with a double embryo transfer approach. In case of either no implantation or a multiple pregnancy, they could easily infer that either both did not or did implant; conversely, in case of a singleton pregnancy they performed a fingerprinting of some DNA from the fetus/newborn and of the DNA isolated from the cell they removed from one of the two embryos at day 3 of preimplantation development, thereby defining whether the implanted embryo was the biopsied one or not. The same design was adopted for blastocysts. Cleavage stage biopsy per se involved a significant 20% absolute and 40% relative decrease in implantation, while blastocyst stage biopsy resulted in no significant impact. This paper is still the main evidence that blastocyst stage biopsy might be safer and therefore preferable for PGT purposes. Of note is that the efficiency and safety of blastomere biopsy had already been questioned by Kokkali et al., (2007): these authors reported lower diagnostic (75%) and implantation rates (27%) when a cleavage stage biopsy plus blastocyst stage transfer workflow was adopted for PGT-M purposes, rather than a blastocyst stage biopsy and transfer workflow (94% diagnostic and 48% implantation rates) (Kokkali et al., 2007). However, more studies are required with a larger sample size, less aggressive blastomere biopsy approaches, in different populations of patients and from other IVF centers to confirm whether blastomere biopsy is indeed detrimental for embryo developmental and/or reproductive competence. In fact, a recent RCT showed better outcomes in the PGT group with respect to the control groups when treating AMA women, even if a blastomere biopsy and array-CGH workflow was adopted (Rubio et al., 2017).

Perfecting and standardizing blastocyst biopsy approach to face the needs of a modern IVF center

In 2014, Capalbo and colleagues described the third and last blastocyst biopsy protocol published to date: the ‘day 5–7 sequential ZP opening and TE cells retrieval’ approach (Capalbo et al., 2014). Different from the other two protocols, which entailed ZP opening on day 3 (de Boer et al., 2004) (or even day 4) or on day 5–6 as soon as the embryos reach the fully expanded blastocyst stage (Veiga et al., 1997; Capalbo et al., 2016a), this third strategy did not involve any assisted hatching. The embryo might therefore be left undisturbed throughout its preimplantation development in vitro (Fig. 3). The blastocyst was secured to the holding pipette, the ZP opened via 2–3 laser shots at the opposite side with respect to the ICM and the TE was detached from the inner part of the ZP by gently blowing some media through the hole. The blastocyst started collapsing shortly afterwards; the operator introduced the biopsy pipette through the opening in the ZP, sucked about five cells into the pipette and started moving it backwards. When the selected TE cells were out of the ZP, the operator increased the time of exposure of the laser and fired, targeting the junctions between the cells. The pulling strategy was used to finally remove the cells.

In current IVF, the day 5–6 assisted hatching-based blastocyst biopsy method represents a good strategy for beginners with limited experience in embryo micromanipulation. However, it does not suit a busy IVF clinic that must conduct several biopsy procedures per working day. The operators must indeed rely on the time invested by the embryo to hatch from the opening in the ZP, the duration of which is unpredictable. When it comes instead to a comparison between the day 3 assisted hatching and the sequential zona opening and TE cells removal blastocyst biopsy protocols, some putative disadvantages of the former versus the latter might be outlined: the unnecessary source of stress at the cleavage stage; the time invested in the manipulation of embryos that might not reach blastocyst stage (a disadvantage shared also with PB and blastomere biopsy protocols); the risk of ICM herniation; the difficulties in scheduling the daily workload in the IVF laboratory to avoid the risk that the blastocysts hatch completely; and impairment of the expansion and hatching processes owing to presence of the hole in the ZP between cleavage and blastocyst stage. Yet, only a single RCT has been published to date that compared the main two TE biopsy protocols, where the authors reported similar technical and clinical outcomes but a higher blastulation rate when the ZP was not drilled on day 3 (Zhao et al., 2019). Nonetheless, the small sample size of this study and the need for reproducibility of this important outcome encourages more investigations in the future.

Several studies have been conducted over the last 5 years to standardize the blastocyst biopsy technique and its outcomes. Capalbo and colleagues investigated the technical and clinical outcomes of almost 2600 TE biopsies associated with qPCR analysis across seven operators from three IVF centers who underwent the same training (Capalbo et al., 2016b). In this paper, they reported consistent and reproducible results in terms of qPCR data (i.e. amplification failure, low quality and good quality), estimated number of cells retrieved, biochemical pregnancy loss, miscarriage and ongoing pregnancy rates, even when comparing the most experienced with the least experienced operator. In particular, they described an overall 1.2% DNA amplification failure and a 4.6% low-quality qPCR data rate when about seven cells are retrieved, in a range of 2–15. A similar approach was adopted when almost 9000 TE biopsy procedures conducted at six IVF centers were investigated to define the risk of re-biopsy and the reason behind an inconclusive qPCR result (Cimadomo et al., 2018b). Interestingly, the re-biopsy risk overall was 2.5% (similar to all previous reports on this topic (Zhang et al., 2014; Kaing et al., 2015; Lee et al., 2016)), of which 2% was related to DNA amplification failure and 0.5% to low-quality data. Of note, blastocyst quality was not associated with the risk of re-biopsy, while the day of biopsy and the IVF clinic showed significant associations. In particular, premature biopsy timing (i.e. less expanded blastocysts, as wells as less and bigger cells constituting the whole TE on day 5 versus day 6–7) and a lower expertise in the absence of strict pre-clinical training might double the risk of an inconclusive result (5%). Still, all the studies are concordant in claiming that a second biopsy and vitrification-warming procedure, when the data are corrected for blastocyst morphological quality and day of blastocyst development, are not detrimental in the hands of expert operators. The most recent paper that investigated the prevalence of inconclusive diagnoses and the outcomes after blastocyst re-biopsy was published by Neal and colleagues, and with an NGS-based protocol (Neal et al., 2019). Also in this report, the absence of a conclusive result characterized 2.5% of the 25 199 blastocysts biopsied and the ongoing pregnancy rate after re-biopsy and a further vitrification-warming cycle was 50% of 50 blastocysts transferred (Neal et al., 2019).

Although blastocyst biopsy did not show any detrimental effect on embryo implantation potential, Scott and colleagues’ study (Scott et al., 2013) did not define whether there is a threshold number of TE cells to retrieve. Thus, Neal and colleagues provided this information in 2017: they clustered the transferred blastocysts diagnosed as euploid after qPCR analysis in four quartiles according to their estimated number of biopsied cells (range 1–20). They reported an impact on implantation only in the 4th quartile (Neal et al., 2017); therefore, when the number of cells retrieved was kept lower than 15, the TE biopsy approach did not impact the clinical outcomes. More studies are, however, required especially to model such a number with respect to blastocyst expansion, the number of cells constituting the rest of the blastocyst and the morphological quality.

A study conducted at two large Italian IVF centers recently investigated the post-warming outcomes of vitrified blastocysts in terms of cryo-survival, re-expansion after 1.5 h and live birth rates with respect to different pre-vitrification workflows (i.e. TE biopsy, laser-assisted artificial shrinkage and control expanded blastocysts), corrected for critical features such as day of blastocyst development and morphological quality (Cimadomo et al., 2018a). TE biopsy showed no impact on blastocyst post-warming behavior, while laser-assisted artificial shrinkage (both per se in non-PGT cycles or associated with the removal of about five cells in PGT cycles) was even beneficial in limiting embryo degeneration. Indeed, the cryo-survival rate out of 2129 blastocyst warming procedures was 99%, higher for biopsied and shrunk (99.4%) and shrunk-only (99.3%) blastocysts but significantly lower for non-biopsied and non-shrunk blastocysts (97.2%). The ideal timings to perform TE biopsy have been detailed in a video-paper recently published by Maggiulli and colleagues (Maggiulli et al., 2019), where the authors also defined that preferably less than 30, but possibly no more than 90 minutes, should elapse between biopsy and vitrification. Such a time interval seems to prevent or limit the risk that a blastocyst re-expands before the cryopreservation procedure is initiated.

Another issue that has been addressed across the years is the definition of the population of blastocysts that can be submitted to a TE biopsy procedure. In 2011, Alfarawati and colleagues reported lower euploidy rates in embryos of a lower morphological quality (Alfarawati et al., 2011). A mild association of euploidy rate with embryo quality was then confirmed by Capalbo et al., (2014), although when euploid blastocysts were transferred, no significant difference was found in the implantation success in relation to morphology (Capalbo et al., 2014). This investigation has been extended in a larger and more recent study that focused on the differences between poor quality blastocysts (PQBs; <BB according to Gardner and Schoolcraft’s classification) and non-PQBs during PGT-A cycles (Cimadomo et al., 2019a). Although showing a lower euploidy rate (23% rather than 51%; mean maternal age: 40 years) and a lower live birth rate after euploid vitrified-warmed single embryo transfer (11% rather than 45%), clinical use of PQBs involved a general 2.5% relative increase in the live birth rate per started cycle. Such an increase might reach even 5% in women older than 42 years, where PQBs represent more than 30% of the population of blastocysts produced after IVF. Therefore, even if more studies are required to assess the reproducibility of this information, these embryos should not be disregarded for the clinical use, especially in poor prognosis patients. Dean Morbeck’s view, published in 2017, nicely summarizes this concept: ‘The answer to “Is this blastocyst good enough to use?” may depend on the patient’s circumstances, values, and resources, as much as on the embryo itself’ (Morbeck, 2017). A similar conclusion applies to embryos reaching the blastocyst stage on day 7 post-insemination. Specifically, Hammond and colleagues reviewed all the published reports up to 2018 that investigated the clinical outcomes after their clinical use: they represent just about 5% of the population of blastocysts produced and are characterized by lower but noteworthy euploidy (34%) and implantation rates (37%) after warming (cryo-survival: 95%). Therefore, it is certainly worth extending the culture one more day to aim at increasing the cumulative live birth rate per cycle, especially in poor prognosis patients (Hammond et al., 2018). More recently, Tiegs and colleagues published their clinical experience strongly supporting the use of these embryos during PGT-A cycles (Tiegs et al., 2019): even though day 7 blastocysts showed lower euploidy and implantation rates, they contributed to an increased chance of finding at least one transferable embryo and possibly achieving at least one live birth in that cycle.

A last important advance in the present history of blastocyst biopsy was introduced in 2016 and 2017 when Zimmerman and colleagues and Minasi and colleagues (Zimmerman et al., 2016; Minasi et al., 2017) described a qPCR and array-CGH protocol to conduct aneuploidy testing and monogenic disease diagnosis on a single TE biopsy, thereby preventing the need for the retrieval of multiple fragments in couples with an indication for both PGT-A and PGT-M.

Lastly, Natsuaki and Dimler recently published a meta-analysis and systematic review of all the reports investigating the neonatal outcomes after PGT with standard IVF, up to 9 years of age. They included 18 studies for childhood outcomes, which entailed anthropometric, psychomotor, cognitive, behavioral and family functioning data. Whatever biopsy stage and method was adopted, PGT resulted in no impact on the neonatal and post-natal outcomes, thereby re-assuring the scientific community of the safety of the approaches used to date (Natsuaki and Dimler, 2018). Recently, a further study has been published that was focused on 1721 children born from cryopreserved blastocysts, either biopsied or not. The authors mainly investigated gestational age, birthweight and macrosomia and concluded that ‘blastocyst biopsy may not add additional risk to neonatal outcomes when compared with a control group’ (He et al., 2019).

The future

Experimental alternative biopsy approaches: rationale, first reports and main issues

Blastocentesis

The BF has been suggested as an alternative source of DNA to conduct PGT (Leaver and Wells, 2020). A minimally invasive technique (known as blastocentesis) has been set up to retrieve the cell-free DNA possibly released into the inner cavity of the human blastocyst (Palini et al., 2013). Blastocentesis was performed by aspirating approximately 1 μl of the BF from expanded blastocysts with an ICSI needle, avoiding the aspiration of cellular material that could affect the results. Although a promising technique, the quantity and integrity of the DNA found in the BF remain uncharacterized, since it is derived, at least in part, from necrotic and/or apoptotic cells (Hammond et al., 2016; Rule et al., 2018). In fact, the amplification rates (56%–82%) as well as the PGT-A concordance rates with the respective TE (37%–97%) vary significantly among the studies published to date (Gianaroli et al., 2014; Tobler et al., 2015; Magli et al., 2016; Capalbo et al., 2018; Tsuiko et al., 2018): a variability that is also ascribable to the heterogeneous expertise in the performance of blastocentesis across different laboratories, (Magli et al., 2016).

To conclude, although blastocentesis has not yet shown sufficient reproducibility and accuracy to be considered clinically valuable for PGT, it still represents a promising source of molecular information that deserves future investigation for embryo selection purposes complementary to PGT (Magli et al., 2019) or independent from it (e.g. proteomic analyses) (Poli et al., 2015).

Morula biopsy

Morula biopsy was a theoretical possibility as it is technically similar to blastomere biopsy, but involving a relatively higher number of cells retrieved. In addition, as the biopsy is performed on day 4 post-insemination, it allows enough time to perform chromosomal testing and fresh embryo transfer. The first study claiming the feasibility of human morula biopsy and chromosomal testing was published in 2014 (Zakharova et al., 2014). Although they suggested that such a biopsy approach did not affect embryo competence in their setting, the same concerns associated with blastomere biopsy also apply to this method (e.g. the need for Ca⁺⁺/Mg⁺⁺-free buffer to loosen compaction). On top of that, there might be issues about the fate established for the cells retrieved from such a crucial and delicate stage of preimplantation development. Therefore, this biopsy protocol should still be considered experimental, especially owing to a lack of data from other IVF centers confirming its reproducibility and safety. In 2018, Irani and colleagues suggested that those embryos getting to the morula stage on day 6 post-insemination might be biopsied, with the aim of rescuing some of them for clinical use during PGT-A cycles (Irani et al., 2018). Indeed, although subject to higher aneuploidy and lower live birth rates, they might still be clinically valuable. Nevertheless, it is unclear whether such an outcome might not be equally achieved by extending their culture to day 7 post-insemination and waiting for them to reach blastocyst stage (Hammond et al., 2018; Tiegs et al., 2019). Beyond the clinical considerations related to this novel putative option, it is astonishing how the human embryo can de-compact and recompact at such a delicate stage of preimplantation development (Coticchio et al., 2019) and still form a viable blastocyst. All the molecular mechanisms coordinating these events certainly deserve thorough investigation in the future.

Biopsy approaches to investigate mosaicism: the link between embryology and molecular biology

Mosaicism is defined as the presence of two or more genetically distinct cell lines in a given organism caused by defective mitotic segregation or DNA mutagenesis in the post-zygotic developmental stages. Inherently, every collected biopsy sample from mosaic embryos is subject to potential sampling error and, by definition, its representativeness of an embryo’s genetic constitution is always imperfect. Accordingly, all diagnostic procedures performed after sampling should acknowledge this imperfect representativeness and the potential for discordancy of the results between the test sample and the diagnostic target. In particular, defining a trustworthy incidence and prevalence of mosaicism and its clinical implications still represent a great challenge for the field, especially when the diagnosis is based on a single randomly selected TE biopsy. Although true chromosomal mosaicism is diagnosed in <1% of prenatal samples (Malvestiti et al., 2015; MRC 1991; Smidt-Jensen et al., 1993) and estimated in <0.2% of live births (Hansteen et al., 1982), evaluation of its incidence in the preimplantation stages has produced inconsistent results, implausibly ranging between 2% and 90%.

At present, an unbiased approach to revealing the true incidence of mosaicism in human embryos does not exist. All approaches to estimating mosaicism can be affected at multiple levels owing to technical and biological limitations, and diverse classifications can result from the same data depending on the stringency of the criteria used for their analysis (Treff and Franasiak, 2016; Capalbo et al., 2016; Capalbo and Rienzi, 2017).

The first evidence of mosaicism in human embryos was produced using FISH (Delhanty et al., 1993). In this case, several blastomeres from a cleavage stage embryo were analyzed and discordant numbers of sex chromosomes were identified. The main limitation of the FISH approach is the assignment of an unequivocal mosaicism diagnosis to embryos showing even a single aneuploid signal in one of the tested blastomeres, thus failing to recognize and control for the impact of technical variations leading to false positive calls. With the concomitant development of CCT techniques for PGT and blastocyst culture and biopsy, direct single-cell analysis was gradually replaced by multicell biopsy at the blastocyst stage. Using array-CGH on a single TE biopsy, profile plots of putative mosaic blastocysts were reported by Greco and colleagues (Greco et al., 2015), whose analytical resolution and diagnostic sensitivity were subsequently increased by NGS (Munne and Wells, 2017) (Fig. 4). Regardless of its higher accuracy, NGS-based copy number analysis is not expected to be a bias-free technology for mosaicism assessment. As in every analytical measurement, sampling is prone to technical and experimental variability since the thresholds adopted for a mosaic call are not standardized or accurate. A subset of results generated from reference samples will inevitably overlap with different categories used for classification (i.e. euploid, mosaic and aneuploid) as a result of the technical variability of the method. Hence, it should be acknowledged that a minority of normal samples can produce data above or under the thresholds used to define normal profiles and will therefore be reported as mosaic, just as some aneuploid samples can produce intermediate copy number values lying in the mosaic range (Popovic et al., 2019; Goodrich et al., 2016). In practical terms, attempting to optimize sensitivity for mosaicism detection (true positive calls) is possible by decreasing the stringency on mosaic call thresholds, inevitably resulting in less specificity and increased false positive rates. Furthermore, mosaicism thresholds generated in experimental settings are based on defined mixes of isolated cells from cell lines, which provide only an ‘ideal’ and stable in vitro experimental model of mosaicism, while real TE biopsies are characterized by high variability in cellular quality and quantity. Thus, the application of experimentally derived thresholds to the clinical situation is itself challenging and prone to inherent errors. Eric Forman in a recent commentary referring to blastocysts that were reported mosaic based on a single TE biopsy named them ‘embryos with a PGT-A result falling in the mosaic range’ (Forman, 2019), which is a more suitable definition of their status.

A more reliable report of mosaicism entails the observation of reciprocal aneuploidies in multiple samples, or at least when the same ‘mosaic plot’ is observed in additional biopsies (Fig. 4). However, this approach is hardly applicable in the routine of an IVF clinic, and even if the ‘mosaic plot’ was not confirmed in a second biopsy, a putative mosaic status for the blastocyst still cannot be excluded.

From a basic research perspective, isolation of the ICM from human blastocysts is not only critical for several biotechnological applications (e.g. human embryonic stem cell derivation, regenerative medicine) but also provides the most valuable tool to study mosaicism (Fig. 4). Few ICM isolation strategies have been described so far (Strom et al., 2007; Hovatta, 2006) and most of them either are inefficient in terms of recovery rate or are associated with high TE cell contamination, and thus are not effective methods. The efficiency of most of these methods has been usually assessed by testing the ability to generate human embryonic stem cells and through evaluation of the number of ICM cells obtained and/or TE cell contamination rate. For instance, immuno-surgery and chemical dissolution of the TE layer by acidified Tyrode’s solution have been commonly used for selective cellular lysis and ablation of TE cells from the ICM (Thomson et al., 1998; Reubinoff et al., 2000; Cowan et al., 2004). However, other than modifying the physiological state of cells, all these methods result in TE dissolution, therefore excluding the possibility of comparative molecular analysis between ICM and TE tissues, and they also do not avoid potential contamination from TE cells. To circumvent the use of immuno-surgery, both mechanical and laser-assisted dissection methods were explored for isolating a pure ICM (Hovatta, 2006; Amit and Itskovitz-Eldor, 2002). However, the common limitation of these methods is the risk of blastocyst collapse during the biopsy procedure and the consequent loss of ICM visualization, thereby hampering several possible downstream applications and the validity of their results. Lastly, all the ICM isolation methods mentioned above have been proposed only for use on blastocysts of excellent quality. In 2013, an efficient and reliable method to isolate pure ICM samples was established using a micromanipulation approach in xeno-free media. The method showed low TE cell-derived contamination in most of the samples, with a mean contamination rate of 2% (Capalbo et al., 2013b). Nevertheless, this approach was subsequently employed only by a minority of investigators attempting to generate mosaicism data from human blastocysts. In general, most of these studies in fact lack of a clear description of the methodology used for ICM isolation.

In the future, it will be extremely valuable to dissect human blastocysts down to single cells and assess the chromosomal status of each of them, thereby overcoming the current analytical limitation of mosaicism estimates from multicellular biopsies based on intermediate chromosome copy number values. Unfortunately, at present, there is no effective method to achieve this end. Even so, should a successful methodology be developed, the assessment of technical errors intrinsic to single-cell analysis will be required and evaluated on single-cell samples of a known karyotype. Ultimately, specific and reliable scoring criteria must be developed in order to minimize the potential for overcalling aneuploidies from single cells.

A brief glimpse of the future

Non-invasive PGT: dream or reality?

Biopsy-based approaches for PGT entail technical skills and expertise, as well as dedicated and costly equipment. Both these features represent a limitation to their implementation worldwide. To limit all the challenges associated with embryo biopsy, several groups have started investigating non-invasive strategies to analyse embryo-derived cell-free DNA in the spent culture media after IVF (Leaver and Wells, 2020). While the diagnosis of monogenic conditions from cell-free DNA (Assou et al., 2014; Wu et al., 2015) seems implausible in the future for an IVF clinic (Capalbo et al., 2018), aneuploidy screening data are more promising. The first two independent studies were published in 2016 (Xu et al., 2016; Shamonki et al., 2016), and in the following years several groups compared the PGT results obtained from TE biopsies with their related spent culture media, reporting good DNA amplification rates from the latter that ranged from 80% to 100%, but highly variable concordance rates that ranged from 30% to 80% (Vera-Rodriguez et al., 2018; Ho et al., 2018; Rubio et al., 2019; Huang et al., 2019; Yeung et al., 2019). Rubio and colleagues set up a protocol that restricted intervention by the operators in the IVF clinic solely to a media change-over to a clean 10 μl drop on day 4 post-insemination. Such a truly non-invasive protocol elicited an amplification rate as high as 94% and a TE biopsy concordance rate as high as 84% when the embryos were cultured for at least 48 h in the final drop of media (Rubio et al., 2019). Further improvements in the accuracy and reliability of non-invasive PGT-A must include both an increased quantity and quality of the data produced, in turn requiring strategies to increase the embryo-derived cell-free DNA and decrease the impact of all the different putative sources of maternal and/or exogenous DNA contamination (Hammond et al., 2016). Similarly, the molecular methods should be adapted to boost their sensitivity towards the detection of cell-free DNA with minimal effect on their specificity (Belandres et al., 2019; Hammond et al., 2017; Vera-Rodriguez et al., 2018). In parallel, research is pivotal to identify the reasons for, and the molecular/cellular mechanisms underlying, the release of cell-free DNA into the culture media during IVF.

More recently, three independent groups have set up protocols for artificial blastocyst shrinkage in the same drop of medium as the embryo cultured. Such a smart expedient allows the collection and analysis of a specimen resulting from the combination of BF and spent blastocyst media (SBM) (Kuznyetsov et al., 2018; Li et al., 2018; Jiao et al., 2019). The results achieved with this study design were compared to both TE biopsies and whole corresponding blastocysts, eliciting very promising, although contrasting, evidence. Combined BF and SBM analysis surely represents an interesting future perspective; still, however, it involves the need for a micromanipulator and is to be considered minimally invasive. Conversely, if SBM itself would be sufficient to provide clinically valuable prognostic information, its revolutionary perspective is definitely broader. The move towards non-invasive PGT is therefore overwhelming since, by minimizing laboratory and personnel expenses, it holds potential for extending the applicability of aneuploidy screening to a larger number of clinics and increasing its accessibility to a wider population of patients. Therefore, careful validation, standardization and reproducibility are critical prerequisites for the implementation of non-invasive PGT in the clinical practice.

In 2017, Feichtinger and colleagues highlighted an intriguing perspective for countries like Austria, where embryo biopsy is not allowed by law: namely that the SBM might be exploited to provide additional non-invasive information on the embryo that could complement polar bodies biopsy-derived data, by definition limited to the detection of only oocyte meiotic chromosomal impairments (Feichtinger et al., 2017). The idea of exploiting the SBM as a source of additional information to complement conventional PGT-A (including TE biopsy-derived data) is indeed a valuable alternative for the future, in case SBM analysis would not be sufficiently reliable by itself, independent of TE analysis. For instance, the transfer of blastocysts reported euploid from both the TE biopsy and the SBM analyses might result in better outcomes than blastocysts reported euploid from the TE biopsy but aneuploid from the SBM. Such a scenario has indeed been suggested by Rubio and colleagues in the preliminary clinical outcomes of their study published in 2019 (Rubio et al., 2019), and certainly encourages future investigations also from this viewpoint. Lastly, as already suggested for the BF, the SBM might also be submitted to ‘omic’ analyses to produce novel additional molecular information that in the future might be used to improve embryo selection (e.g. (Capalbo et al., 2016d; Noli et al., 2016) (Cimadomo et al., 2019b; Stigliani et al., 2019; Sanchez et al., 2017; Pallinger et al., 2017; Giacomini et al., 2017; Siristatidis et al., 2018)).

PGT using in vivo blastocysts: is it a reasonable future practice to consider?

Recently, Munnè and colleagues published a study claiming that PGT might be conducted on embryos recovered directly from the uterus after IUI via uterine lavage. These embryos were characterized by having the same euploidy rate as blastocysts produced in vitro but a poorer morphological quality (Munne et al., 2020). This paper was elegantly discussed by Galia Oron, an Associate Editor of Human Reproduction, in a manuscript entitled ‘How far should we go in the name of science?’ (Oron, 2020). Her concerns dealt with four precepts of Medicine that might have been violated by Munnè’s study design: autonomy, justice, beneficence and non-maleficence. In particular, regarding the precept non-maleficence, Oron stated: ‘It seems, the relatively low embryo recovery rate (42%), the need for a double embryo biopsy in the majority of cases (90%) and the significant side effect of a persistent positive beta-hCG warranting methotrexate treatment may, at present, outweigh the current scientific value of the study’. Future studies are definitely warranted to determine whether PGT using in vivo blastocysts could ever be considered a reasonable approach in IVF.

The ethics of embryo biopsy and PGT at a glance

The rapid progress of science and technology makes it difficult to issue laws and regulations in a short space of time to cope with them. The topics faced by ART involve a radical reformulation of our idea of humanity and require a parallel constant renewal of the legal system. When it comes to embryo manipulations (such as biopsy) and PGT, several aspects are involved, all related to the nucleus of human rights and freedom issues. This entails an interdisciplinary perspective encompassing science, ethics and law, all legitimised by the nature of the object, i.e. a technique that implies choices about life. Embryo biopsy and PGT raise ethical issues spanning various concepts such as the difference between embryos and individuals and the fear of eugenics. When does life begin? When should we recognize ‘full human rights’? These issues encounter ‘paradoxes and logical traps’ that must be prevented through the equal involvement of science, ethics and law. In particular, the debate on the nature of the embryo arises from the hypothesis that it might become an individual, although in terms of probabilities most of the embryos do not even result in a pregnancy. This ‘ambiguity’ raised ethical questions about the value of a person, the respect for his/her dignity and the right to own self-determination, all being addressed to IVF-related practice. The PGT-related concerns were mainly 2-fold: the risk that the biopsy might be detrimental for the embryo; and the risk of eugenics. However, to date, there is no solid evidence to support an adverse impact of the biopsy (especially PB and TE biopsy approaches) on the embryo. Dealing with eugenics instead, at present, it is technically implausible per se (Karavani et al., 2019). Even so, PGT is not aimed at selecting any favourable characteristic, it is rather a neutral analysis aimed at ascertaining the health status of the embryo to limit the transmission of serious heritable diseases and/or to reduce the risk of miscarriage or chromosomal syndromes in the newborns. Nevertheless, embryo biopsy and PGT are still forbidden in some countries. Those countries that approved embryo biopsy and PGT mainly did so in line with their constitutions, aimed at protecting a woman’s right to be informed and make a conscious, responsible and private choice on the destiny of affected/aneuploid embryos. Although uncertainties persist on PGT from a legislative perspective regarding such a destiny, the would-be mother is allowed to take the decision, in accordance with her psychophysical health, and moral, ethical and religious beliefs, without any legal imposition or interference (Dondorp and de Wert, 2019). Notably though, more than 80% of the world’s population practice a religion and look for answers to questions of life and death in their beliefs. However, the reactions of the different religions to ART have been diverse, ranging from total rejection (e.g. Catholic) to total acceptance (e.g. Judaism, Hinduism or Buddhism). These controversies clearly affect couples’ attitudes to ART, from its simplest procedures, such as IUI or even semen collection, to the most debated practices such as PGT and sex selection (for a comprehensive review see (Sallam and Sallam, 2016)).

Conclusion

In the early days of embryo biopsy, clinical embryologists, who had been trained to minimize disturbance of embryos in culture, were understandably reluctant to perform invasive micromanipulation procedures to remove cells, which could cause additional damage to the biopsied embryo. However, when it was explained that, in the context of a couple trying to avoid a pregnancy affected by an inherited disease, an embryo could not be transferred unless it was biopsied and tested as unaffected, the risks to the embryo were balanced by the benefits of the testing. With care, even the early protocols used for cleavage stage biopsy gave pregnancy rates that were comparable to IVF generally. However, biopsying more than a single cell from 6- to 10-cell embryos reduced pregnancy rates. This is similar to experiments with mouse embryos in which progressively removing more cells from embryos at the 8-cell stage, particularly more than half of the embryo, first reduces the rate of post-implantation development of the fetus and then implantation after transfer.

The more serious and unanswered question at the beginning was: would the removal of cells at these stages cause fetal abnormalities after transfer? However, human embryos were often fragmented or partially damaged, particularly after cryopreservation using slow freezing methods. If these were the best embryos available for transfer, they had been routinely transferred for some years with no reports of an increase in fetal abnormalities or at birth. Therefore, it was assumed that, like mouse embryos, the early human embryo could regulate for the loss of cells and that the fate of individual cells had not been predetermined at those stages. Follow-up of pregnancies and live births following biopsy and PGT became an early priority of the ESHRE PGT Consortium, and after a number of years several publications reviewing large series of cases were generally reassuring that there was no increase in fetal or congenital abnormalities. Nevertheless, it would be incorrect to ascribe a putative consequence to the biopsy procedure itself, rather than to any other manipulation required to conduct PGT, which is sometimes adopted also in standard IVF cycles (e.g. assisted hatching or cryopreservation etc.

Using embryo biopsy more widely for PGT-A to improve implantation rates and reduce miscarriage rates in certain categories of infertile patients is more problematical. Indeed, some have argued that the reduced live birth rates reported by Mastenbroek and colleagues in a RCT where FISH was used in patients of AMA were caused partly by the use of cleavage stage embryo biopsy. As described here, this has led to the development of blastocyst biopsy methods, which allow the sampling of multiple cells from the extraembryonic TE layer, and which may be less detrimental to the implantation and later development of the embryo.

The 30-year story of the embryo biopsy revolution continues to evolve. It was speculated at the beginning that embryos might be tested by examining macromolecules in the culture medium, but it was not anticipated that there would be sufficient DNA of embryo origin to perform genetic testing. This has the potential to completely revolutionize clinical embryology by eliminating labor intensive and, in some cases, potentially damaging micromanipulation procedures. How effective these non-invasive methods are in terms of efficiency and accuracy is still under investigation. So, it may still be the case, as it was in the beginning, that invasive biopsy methods continue to be required for the highest accuracy and any effects on embryo development are secondary, where prevention of a serious inherited disease is the aim.

Supplementary data

Supplementary data are available at Human Reproduction Update online.

Authors’ roles

D.C., F.I., F.M.U. and L.R. drafted ‘The past’ and ‘The present’. A.C., C.R. and C.M.G.P. drafted ‘The future’. AH drafted ‘Introduction’ and ‘Conclusion’. All authors reviewed all sections and provided comments and active discussion.

Funding

No external funding was either sought or obtained for this study.

Conflict of interest

None.

References