Truth in science: experimental design and the legacy of John D Biggers, PhD., DSc

Abstract

The current article presents a brief historical perspective on Professor John D Biggers, PhD, DSc. who died on 7 April, 2018. His interests covered reproductive physiology, embryo culture, cryobiology, sperm preservation, statistics and experimental design, and the history and ethics of human reproductive biology. Emphasis is placed on John Biggers’ approach to the development of media for the culture of mammalian preimplantation embryos and to correct several minor misconceptions that have arisen in recent years regarding some of his studies. Much can be learned from his detailed approach to scientific investigation and experimental design. His scientific accomplishments and seminal contributions are important, but the tapestry of his life and legacy continue to be woven through the many students, fellows, and collaborators with whom he worked with over many years. The present article builds on a previous conversation that Michael Summers and Catherine Racowsky had with John Biggers that was published in 2008 [1].

Introduction

“Numbers do not lie, but they have the propensity to tell the truth with intent to deceive. The human being is prone to seeing patterns and will often see patterns where only random noise exists.”

Eric Temple Bell, American Mathematician, Numerolgy, 1933.

On July 25, 1978 at 11:47 p.m., a baby girl, Louise Brown, was delivered by primary Caesarean section at Oldham General Hospital to a woman without functional fallopian tubes following fertilisation in vitro of an ovum from the patient by her husband’s sperm and replacement of the resultant embryo into her uterus. July 25, 2018 marked the 40th anniversary of the first “test tube” baby.

I was kindly invited by Daniel Brison to speak on 40 years of human embryo culture media at a meeting organised by the Society for Reproduction and Fertility to celebrate the 40th anniversary of the birth of Louise Brown entitled, ‘Edwards, Steptoe… and Dr Kershaw: an SRF symposium to mark the 40th anniversary of IVF.’ The meeting also provided an opportunity to reflect, albeit briefly, on some of the seminal contributions of Professor John D. Biggers who sadly died on 7 April 2018. I have therefore taken the opportunity to write a brief historical perspective on John’s approach to media development, among other things, and correct some minor misconceptions that have arisen over time regarding his work. The commentary builds on a conversation Catherine Racowsky and I had with John Biggers in 2008 that was subsequently published with the kind assistance of Henry Leese in Human Fertility [1]. It is to be hoped that the aforementioned will also fill in some gaps for those unfamiliar with John’s important contributions over many years to the field of reproductive physiology.

Historical perspective

John’s interest in chemically designed media started in the 1950s. He recognised early on that the traditional approach to media development of changing the concentration of each component one at a time was inefficient and would not detect the interactions that occur between the different components. To quote John [2]:

“The problem of optimizing a mixture, however, is not simple, and cannot be done by using the intuitive approach of varying the components one at a time, keeping the concentrations of the other components constant. An understanding of the problems involved needs consideration of a concentration-response surface…”

This was treated theoretically by Biggers et al [3] in 1957 who proposed that the totality of responses to the mixture of compounds in a medium can be represented by a concentration-response surface, although his interest at the time was focused on the development of chemically defined media for the culture of embryonic chick bones, not preimplantation embryos. As commented [1]:

“I did not start looking at the design of media for culturing preimplantation embryos until I moved to the Wistar Institute and the University of Pennsylvania in Philadelphia in 1959. Wes Whitten was working on this problem, and I did not want to usurp his field”

Work at the King Ranch Laboratory of Reproductive Physiology, University of Pennsylvania, Philadelphia, PA (1961–1966)

John became King Ranch Research Professor, School of Veterinary Medicine, University of Pennsylvania in 1961. Ralph Gwatkin joined as an Assistant Professor. He was also joined by Ralph Brinster and David Whittingham. Ralph Brinster was a Pennsylvania Plan Scholar, Department of Physiology, School of Medicine, University of Pennsylvania and his studies leading to a doctoral thesis were supervised by John Biggers. Parenthetically, John did not include his name as senior author on Brinster’s early papers for the reason outlined below [1].

“….in the UK a PhD thesis had to contain original work, and mentors did not append their name to graduate students’ papers emanating from their thesis work. The thesis had to demonstrate originality. It was natural for me to do this when I became a faculty member in the USA. That is why I did not add my name to Ralph Brinster’s papers on culturing mouse preimplantation embryos. I soon found out that the practice was not observed in the USA and that it would reduce my chances of securing research grants for my laboratory.”

David Whittingham initially joined John’s laboratory as a Fullbright Scholar, and later as a Pennsylvania Plan Fellow.

Brinster’s studies included mouse embryo development in organ cultures of the Fallopian tube and detailed investigations of the role of energy substrates and amino acids on early development of preimplantation mouse embryos. These studies first showed the beneficial effects of adding amino acids to simple chemically defined media used for the culture of preimplantation mouse embryos and afforded an opportunity for John to investigate the interactions of components, particularly energy substrates, such as pyruvate, lactate and glucose using factorial design experiments. Gwatkin investigated several different areas, including the culture and study of embryonic chick long bones, mouse embryo development in organ cultures of the Fallopian tube and the effects and role of different amino acids on blastocyst implantation and the development of blastocyst outgrowth in vitro. Biggers previously reviewed the studies of Brinster and Gwatkin on early mouse preimplantation embryo development [5]. Remarkably, these studies were undertaken 50–60 years ago and are still relevant today, particularly the role of amino acids on early embryo development.

The method of culturing mouse embryos in microdroplets under oil was also developed in John’s laboratory. The technique was described in a paper authored by Brinster [4]. However, adoption of the method is less well known but important from an historical perspective. As noted by John [1]:

“Culture of embryos in drops of medium under mineral oil was later adopted, from the field of virology by Ralph Gwatkin, in my laboratory at the University of Pennsylvania, although the method was first published in a paper by Ralph Brinster. I am sorry that Ralph Gwatkin never received credit for the method”

John previously noted [5]

“Whitten cultured early mouse embryos in small test tubes. This method is not particularly convenient for the frequent observation of developing embryos. A more suitable method, the microdroplet method, was introduced by Ralph Gwatkin,…….The method was then used routinely for the culture of early mouse embryos (Brinster, 1963).”

Gwatkin’s manuscript [6] was submitted on July 1, 1963 and Brinster’s [4] on August 5, 1963. Both papers demonstrated the successful culture of mouse embryos to the hatching/hatched blastocyst. John’s comments above had nothing to do with Brinster’s publication of the microdroplet method, after all he was Brinster’s supervisor and would have signed-off on any manuscript prior to submission. John was more interested in historical accuracy, attribution and provenance. Moreover, John probably never imagined that a simple embryo culture technique developed in his research laboratory in Philadelphia in the early 1960s would be embraced many years later by most clinical embryology laboratories involved in human in vitro fertilisation (IVF). Even with the recent introduction of time-lapse incubators, the basic method of embryo culture is still a microdroplet of culture medium under oil as are most embryo manipulations.

It is very clear that John recognised immediately the potential importance of culturing embryos in microdroplets under oil not only as a matter of convenience and ease of use compared to gassed test tubes, but also from the standpoint of experimental reproducibility. He owned a copy of Fisher’s 1935 book entitled, “The Design of Experiments”, and was fully aware of his classic studies in crop variation in the early 1920s at the Rothamsted Agricultural Experimental Station, Harpenden in England [7]. Under Fisher’s guidance, fields were subdivided into small plots that in turn were divided into rows and each row was given a different treatment. From these studies, randomised controlled experiments, factorial design experiments, analysis of variance and covariance and degrees of freedom were described for the first time. For John, the idea of a row of microdroplets of different media of the same volume containing an identical number of embryos in a controlled environment must have appeared irresistible.

John, as the lead author, published an important paper with Brinster [8] on experimental design that was based in part on the latter’s PhD research. The arguments presented in the paper were prompted by improved methods of in vitro development of preimplantation mouse embryos. In John’s estimation, the methods were considered sufficiently reliable for use in quantitative studies, and in particular, the interrogation of concentration-response surfaces as a means to understand the interaction of effects produced by different components of a medium, i.e., factorial design experiments. Philosophically, one can consider three basic tasks to set up an embryo culture experiment as described above: (1) the measured response, (2) improve the response, and (3) understand the response. Step (1) is based on established knowledge, whereas (2) and (3) are subdivisions of statistics known as “experimental design”. Ideally, all three must be thoroughly reviewed before starting an experiment. The use of experimental designs is a prescription for successful application of the scientific method

Professor, Department of Population Dynamics. The Johns Hopkins University School of Hygiene and Public Health, Baltimore, MD (1966–1971)

John moved to Johns Hopkins University in 1961. He was joined by David Whittingham. Much of John’s work at Johns Hopkins was focused on nutritional requirements and metabolism of the oocyte and preimplantation embryo (see for example, Biggers [9, 10]). He was also joined by Roger Donahue, a PhD student who studied oocyte and meiotic maturation in the mouse. Several research studies during this period standout and require comment due to their influence on the development of media used for human IVF.

Donahue’s studies soon demonstrated that maturation of mouse oocytes to the metaphase II stage in simple chemically defined media required the presence of either pyruvate or oxaloacetate and led to the view that the pattern of energy metabolism in the zygote is restricted and determined by the oocyte before fertilisation. Previous studies by Brinster showed that the late two-cell stage can use lactate, pyruvate, phosphenolpyruvate (PEP) and oxaloacetate to support cleavage. By contrast, the eight-cell stage can use glucose and other carbon sources, such as malate and α-ketoglutarate [11]. These findings were summarised in an important paper by Biggers, Whittingham and Donahue in 1967 [12]. There is thus evidence based on studies using simple chemically defined media that the metabolic requirements and capabilities of early mammalian preimplantation embryos change with development.

Kennedy and Donahue [13] subsequently showed that human oocytes could complete meiosis in several chemically defined media from complex, such as Ham’s F10, and simple, including a modified, pyruvate-supplemented Krebs-Ringer bicarbonate. Edwards et al. [14] also commented on their early work as follows,

“Our experiences with human ova indicate that they tolerate a wide variety of culture media. The most suitable and simple medium for fertilisation in vitro is, perhaps, Earle’s solution with the addition of pyruvate (1.1 mg/ml) and inactivated human serum (between 5% and 10%, v/v)”

Pyruvate has been included in all media used for fertilisation in vitro of human oocytes [15].

The second area of interest is related to animal studies, particularly the mouse, that have impacted the development of media for the culture of human preimplantation embryos. Briefly, two approaches have been used to determine the concentrations of components used in a medium: (1) “let the embryo choose” principle or empirical optimisation and (2) “back to nature” principle. In the first approach, a bioassay is done to measure a response to different concentrations of a component. The results are then used to construct a dose-response line and a measured maximum response. The second approach uses the concentration of a substance which is present in the natural environment of the embryo. The “back to nature” approach depends on prior knowledge of the compositions of the female genital tract fluids in the oviduct and the uterus. The subject has been critically reviewed and will not be pursued further in this article [2, 16, 17]. The choice of media for human preimplantation embryo culture has typically focused on a comparison of two-step sequential media protocols based on the “back to nature” principle versus a continuous single step culture medium based on empirical optimisation [2, 16–21]. The increased use of time-lapse incubators in human IVF laboratories has however resulted in increased interest in single-step culture media, largely for practical reasons. If the truth be known, there are no media currently available for the culture of preimplantation human embryos that are based solely on either the “back to nature” or “let the embryo choose” principle.

Laboratory of Human Reproduction and Reproductive Biology (LHRRB), Harvard Medical School

In 1971, John became a fully tenured Professor of Physiology, Department of Cellular Physiology (later Cell Biology), Harvard Medical School and Member, Laboratory of Human Reproduction and Reproductive Biology (LHRRB). Many aspects of John’s research and collaborations at Harvard were previously outlined by his colleagues and friends in an earlier issue of the Journal of Assisted Reproduction and Genetics celebrating his 90th birthday [22].

Much of John’s later work on culture media was supported by the National Cooperative Program on Non-Human In Vitro Fertilization and Preimplantation Embryo Development from 1986 through 1997. Some of the other members of the group included John Eppig, Robert Foote, Barry Bavister and Richard Schultz. A steering committee made up of the principle investigators met on a regular basis to review different research proposals and collaborations (see below). One of the challenges to early mammalian embryo research was that several species undergo blocks in development in vitro. In the mouse, the block occurs at the two-cell stage. John was convinced that blocks to development in vitro were caused by adverse culture conditions, such as imbalances in the concentration of specific components in the medium. He previously described trying to explain to the other members of the National Cooperative Program on Non-Human In Vitro Fertilisation and Preimplantation Embryo Development the use of sequential simplex optimisation methods as a means to locally “hill climb” on a concentration-response surface in an attempt to overcome the two-cell block [1].

“It was a tough sell. In the end they shook their heads and said, ‘Well, you go ahead and do it if you want to.’”

Simplex optimisation

Dantzig’s development of the simplex method in the late 1940s marks the beginning of the method of optimisation [23]. A concentration-response surface was previously proposed by Box and Wilson in 1951 [24] as a means of increasing industrial productivity, called evolutionary operation (EVOP). Box suggested changing the experimental conditions very slightly and by making small changes the responses that did occur could be monitored and reveal information about how the system could be changed to improve the overall response [25, 26]. Spendley et al. [27] published a paper that described how EVOP might be made automatic. The basic design of the technique is the regular simplex in k dimensions, k being the number of variables under study. A simplex is a geometric figure having a number of vertexes equal to one more than the number of dimensions of the factor space. A two-factor simplex experimental design has 3 vertexes (k + 1) and is a triangle. The simplex represents a tangential planar approximation to the response surface in the region of the design. The use of simplex EVOP has advantages over factorial EVOP. Firstly, the number of experiments in the initial simplex design is k + 1, not 2^k using a factorial design. The other significant advantage is that simplex requires fewer experiments to move into an adjacent region on the concentration-response surface. In 1965, Nelder and Mead [28] made two basic changes to the original simplex of Spendley et al. [27]. The modifications allow the simplex to expand in directions that are favourable and to contract in directions that are unfavourable and allows a more rapid identification of the region of the optimum.

The development of commercial software packages for simplex optimisation allowed Lawitts and Biggers [29] to develop culture media that overcame the two-cell block to mouse embryo development. Ten components commonly used in simple chemically defined media at the time were used to create 11 media: base medium, GENMED and 10 different START media. The concentrations of the START components were based on previous observations and what was already known about mouse embryo culture media. The optimisation strategy produced SOM (simplex optimised medium) after 20 cycles of sequential optimisation that was subsequently modified to produce medium mSOM and finally KSOM for the culture of mouse preimplantation embryos [30]. This represented the end of the use of simplex optimisation for mouse embryo culture media.

The components in a simplex optimisation process are referred to as variables. John was not interested in studying single variables, but their interactions. Many researchers are not aware of factor interactions and would be mistaken to assume that it is possible to optimise a multifactor system by carrying out a series of single-factor studies since the approach is highly limited [19, 31]. There is a simplex method to study a single variable, univariate simplex, but the approach does not provide information on the interactions of variables. The original 10 media in the START simplex each contained one of the 10 variables at a relatively high concentration. This was intentional. The medium with the worst response is identified experimentally. A new medium is then created by taking the geometric reflection of the co-ordinates of the simplex through the centroid, which is calculated from the co-ordinates of the remaining media. The procedure is repeated sequentially until the optimum is reached. Four components were rapidly identified as detrimental to embryo development: high sodium (hNaCl), phosphate (hKH₂PO₄), pyruvate (hPYR) and glucose (hGLUC). A geometric reflection of the worst medium (hNaCl) would be computed to generate a new medium (aNaCl) and new simplex. The new medium, aNaCl, includes a change in the concentration of all components under the rules of simplex. Thus, every component in each adjusted medium is at a different concentration. Ultimately, simplex optimisation does not select the variables of interest. The selection process is dependent on experimentation and the observed biological response.

As commented by Lawitts and Biggers [29]:

“The simplex optimization procedure does not select the compounds whose concentrations are to be varied. The selection process must depend on biological information. A strength of the simplex optimization method, however, is that it causes media to be tested that have compositions which may be considered extreme, according to previous information. The benefit from testing these media is that they create unexpected physiological questions”

Biggers was not originally driven by an interest in either mouse or human blastocyst development when using sequential simplex optimisation to produce SOM; this is a matter of public record [1]. He was mainly interested in overcoming the two-cell block to mouse embryo development that required a relatively simple scoring system. To attempt to assess blastocyst development without knowing if it was possible to overcome the two-cell block would have been unwise. In addition, the simplex method becomes less efficient if there is high variance in the measured response of interest. John previously commented at his surprise at the level of blastocyst development after overcoming the two-cell block which prompted further studies on blastocyst development [2].

“A bonus from the work was that SOM also favoured the subsequent development of the embryos into blastocysts, although the response used to optimise the medium was passage through the two-cell block.”

An optimisation strategy does not mean a culture medium is optimised since a multi-dimensional concentration-response surface can have multiple optima. The fact that there are many different types of culture media for the in vitro culture and development of mouse and human embryos suggests that there are multiple optima. This however is a subject for another day. The purpose of the simplex is to move rapidly into the region of a local optimum, and once located, other methods are then used to define the optimum, such as factorial design experiments. This is precisely what John did in several papers following the development of SOM [32–34]. It was only at this stage that he looked more closely at mouse blastocyst development. Increases in the level of both sodium and potassium (KSOM) were based on electron probe microanalysis of the intracellular concentrations of Na⁺ and K+ in blastomeres of two-cell mouse embryos exposed to SOM which showed a very low K⁺/Na⁺ ratio. Changing the K⁺/Na⁺ ratio by increasing the concentration of K⁺ to produce KSOM was further based on information available at the time on the impact of K+ and Na+ on protein synthesis in the early mouse embryo. KSOM increased the proportion of two-cell mouse embryos passing through the two-cell block that developed to the blastocyst [34].

The development of KSOM also permitted a detailed analysis of the joint effects of glucose and KH₂PO₄ and their interaction on the development in vitro of mouse zygotes to the blastocyst using a 4 × 4 factorial design [35]. The range of concentrations used for glucose and phosphate spans the measured concentrations found in the mouse and sheep oviduct, respectively. No significant interactions were detected between the effects of glucose and KH₂PO₄ indicating that the effects of both variables are independent. The addition of glucose to KSOM had no inhibitory effect on development, and KH₂PO₄ had a slightly inhibitory effect and the concentration-response surface was essentially a horizontal plane. In this setting, it is recommended that the in vivo measured “back-to-nature” concentration of glucose should be used [16, 33]. This is a reasonable recommendation for any variable when no local optimum is found by experimentation.

Shortly after the development of KSOM, it was shown that a single medium, modified KSOM (mKSOM_g), would support both fertilisation in vitro (IVF) of CF1mouse ova, complete preimplantation embryo development without the need for a change of medium to overcome the two-cell block and live foetuses following transfer to pseudopregnant foster mothers [36]. Parenthetically, the occurrence of the two-cell block to development was originally thought to be strain-dependent being observed in random-bred and inbred strains, although more recent studies have shown that F1 hybrid embryos will also experience a two-cell block if either the osmolarity or glucose/phosphate levels are increased [37].

It is perhaps worth noting that over 60 years have now passed since John Biggers and Anne McLaren published the first report of the successful development and birth of live young following mouse embryo culture in vitro [38]. Interestingly, neither Anne McLaren nor John Biggers were particularly interested in human IVF when undertaking their embryo transfer studies, but more fundamental questions regarding nature and nurture. See Biggers for a review of this classic paper [39].

The addition of amino acids to KSOM (KSOM^AA) was the subject of a meeting of the Steering Committee. Richard Schultz and John Eppig had different reasons to investigate KSOM^AA (personal communication). Richard Schultz was mainly interested in learning more about the effects of amino acids on protein synthesis and gene expression profiles during early preimplantation embryo development, for example, Ho et al. [40], whereas John Eppig had hoped KSOM^AA would be beneficial for mouse oocyte maturation in vitro. John Eppig does not specifically recall John Biggers discussing the addition of amino acids to KSOM, but expressed the following sentiment (personal communication),

“JB undoubtedly had more sophisticated and systematic plans than ours”

Indeed, John returned to a question first asked when Ralph Brinster [41] was a graduate student in his laboratory, viz., the role of a fixed nitrogen source for the in vitro development of preimplantation mouse embryos. These studies were published in two papers in 1997 [42] and 2000 [43], respectively. The latter is a masterclass in experimental design, data analysis and interpretation. This work, along with other studies subsequently formed the basis of the commercial product currently used in clinical IVF, Global® medium [2, 44].

The study and development of culture media for preimplantation embryo culture is extremely challenging, undeniably complex and requires a hearty dose of humility. To get it right is impossible, but to get preimplantation embryos to grow in culture is clearly possible. Ethical and moral issues inevitably arise when investigators study early mammalian development, particularly human development, but nowadays this seems to have been forgotten in the rush to push for improved IVF success rates. Any clinician, director of an embryology laboratory or clinical embryologist vaguely interested in culture media would do well to read Hannah Landecker’s article entitled, “It is what is eats: Chemically defined media and the history of surrounds” [45] and perhaps reflect for a moment on the issues raised within. To quote Erwin Chargaff [46]:

“But biology is limitless, and our experiments are only a drop out of an ocean that changes its shape with every rolling wave. Because our questions must skirt our fundamental ignorance of the nature of life, the answers we receive can be no more a travesty of truth; a truth, moreover that may be so much of a plural that we can never comprehend it. The manner in which questions are asked, i.e., experiments designed, is either completely random or conditioned by our ideas of a preestablished harmony, a harmony that we seldom recognise as a contract with God that he has never signed.”

John had previously expressed his concerns on the commercialisation of embryo culture media [47]. The same concerns persist to this day. Indeed, recent years have witnessed a further consolidation of this sector and marketing now trumps solid science. The few remaining companies seem to have carved out their market share for culture media and shown little interest in further innovation. One culture media company, for example, promotes a single-step medium that is optimised for use in a time-lapse incubator, but makes unclear what in fact is being optimised. We can surely do better.

Epilogue

Earlier I referenced a previous issue of the Journal of Assisted Reproduction and Genetics celebrating John’s 90th birthday as a good introduction to John’s myriad interests [22]. David Albertini and Lynda McGinnis commented,

“His accomplishments at the bench are far outweighed by the lasting legacy that lives on in his mentees and by his singular efforts to bring a scholarly interpretation of the evolution of ARTs to his peers, the public, and law makers”

Ryuzo Yanagimachi wrote,

"John Biggers’s many contributions including analyses of the metabolism of early preimplantation embryos, improvement of mammalian embryo culturing media, and his mentoring many superb young fellows remain unsurpassed. I cannot imagine how today’s field of experimental embryology in mammals would be if he were not in this field of study."

I have a copy of John’s curriculum vitae from 2008 that lists the different graduate students (9), postdoctoral fellows (39) and visitors (8) to his various research laboratories over the many years. The list does not include his colleagues and other research collaborators with independent laboratories with whom he worked with on different projects of mutual interest. Graduate students as noted in this article, include Ralph Brinster and David Whittingham. Ralph Brinster went on to make seminal contributions in mouse transgenics and male germ cell transplantation in rodents. He is Richard King Professor of Reproductive Physiology, School of Veterinary Medicine and a founding co-director of the Institute of Regenerative Medicine, University of Pennsylvania. He received the National Medicine of Science from President Obama in 2010. David Whittingham, DSc is Emeritus Professor, St George’s, University of London. In addition to his contributions to the development of embryo culture media, he also described successful cryopreservation of mammalian embryos and the role of calcium in oocyte activation. I still get to meet David on occasion due to his involvement in the Progress Education Trust (PET). He is a previous Chair of Trustees of PET whose mission is to educate and debate the responsible application of reproductive medicine and genetics. It was only later I learned that both David and I received the prestigious Beit Memorial Fellowship, Cambridge University, albeit at different times.

In the late 1994, possibly early 1995, John invited me to join him for dinner in Brookline, just outside Boston, MA. The third member of the group was Catherine Racowsky. Catherine previously worked in John’s laboratory as a post-doctoral fellow and was being considered for the position of Laboratory Director, IVF unit, Department of Obstetrics and Gynecology at Brigham and Women’s Hospital. The dinner went well, salmon I recall, possibly sea bass, but to put three English-speaking Brits together for an extended length of time can only end one way, a sedate version of Monty Python. Catherine got the position of Laboratory Director. She continued to work closely with John to the end of his life [48] and built up a superb body of work on early human embryo development. She is past Secretary, Vice President and current President of the American Society of Reproductive Medicine (ARSM).

Other important contributors who trained with John include Paul Burgoyne (MRC National Institute for Medical Research, Mill Hill, UK (sex chromosomes, infertility and germ line development), Jay Baltz (Professor, Department of Cellular and Molecular Medicine, Ottawa Hospital Research Institute, Canada, cell volume regulation, oocyte and early embryo development), Joel Lawitts (Director, Transgenic Facility, Beth Israel Deaconess Medical Centre, Boston, MA), Roger Perdersen (Professor of Regenerative Medicine, Department of Paediatrics, Cambridge University) and Paul Wassarman (Professor of Cell, Developmental and Regenerative Biology, Icahn School of Medicine, Mount Sinai (gametogenesis, embryogenesis and fertilisation).

Visitors included Raymond Wales, Henry Leese and later his graduate student, David Gardner. It was during this time that extensive studies were undertaken with Claude Lechene, Jay M. Baltz and R. M. Borland on nutritional requirements of early embryos, transport and homeostatic mechanisms and cell volume regulation, the dynamics of blastocoele formation, the composition of oviductal fluid and the development of ultra-microfluorometric methods to measure metabolites in single embryos.

Finally, it is perhaps worth commenting on John’s collaboration with the late Dr Melvin Taymor, Obstetrician/Gynecologist, Beth Israel Hospital, Boston, MA on clinical IVF/ET. The clinical embryologist was Dianne Smith, one of John’s graduate students. John described the next stage [1].

“I remember a meeting with Mel and the other medical members of the team one day before the procedure was used on the patient, which involved writing down a detailed checklist of steps to be taken, just as a pilot prepares for take- off. Also, attempts were made to identify problem areas and list alternative routes, if necessary. This meticulous document paved the way for the first attempt at human IVF/ET in Boston, which resulted in a baby being born from the first try.”

John’s statement is of historical interest given the time and reference to a detailed checklist. It can be argued that checklists as a cognitive aid probably go back centuries. The modern use of a checklist can be traced as noted by John to the field of aviation (http://www.atchistory.org/how-the-pilots-checklist-came-about-2/). However, the use a checklist in a medical sense has a different history, meaning and focus, and was not widely considered until 2004 when it was shown that the use of a checklist reduced the incidence of infections in central catheter lines [49]. It therefore seems remarkable that the necessary criteria for establishing a checklist using a dedicated team for IVF was being used in the early 1980s.