Could Failure in Preimplantation Genetic Diagnosis Justify Editing the Human Embryo Genome?

Abstract

It has been argued that developing any approach involving genetic modifications of human embryos is unnecessary because couples at risk of transmitting a serious genetic disease already have a straightforward alternative to prenatal diagnosis and termination of pregnancy with preimplantation genetic diagnosis (PGD) (Cyranoski, 2015Cyranoski D. Ethics of embryo editing divides scientists.Nature. 2015; 519: 272Crossref PubMed Scopus (36) Google Scholar). However, this argument does not take into account the burden of PGD, and its low chance of success, as a large number of viable embryos are discarded because they are affected. In weighing the risks and benefits of gene editing technology in the embryo, one needs to consider the potential of genome editing to cure genetic diseases at an early stage of life, and the concept that we have a moral duty to cure affected human embryos instead of discarding them (National Academies of SciencesEngineeringand Medicine, 2017National Academies of Sciences, Engineering, and MedicineHuman Genome Editing: Science, Ethics, and Governance. The National Academies Press, 2017https://doi.org/10.17226/24623Google Scholar, Hirsch et al., 2017Hirsch F. Lévy Y. Chneiweiss H. CRISPR-Cas9: A European position on genome editing.Nature. 2017; 541: 30Crossref PubMed Scopus (15) Google Scholar). We elaborate on the results generated in the PGD Centre of Béclère-Necker hospitals (PCBN) in Paris during the last 5 years and consider if there is potential for responsible use of genome editing in the human embryo. In PCBN, PGD has been performed since 2000 and more than 1,000 couples have undergone this procedure (1,939 attempts). PGD is routinely performed on blastomeres, sampled from in-vitro fertilized human embryos, with the aim of only transferring unaffected embryos. Over the last 5 years (2011–2016), 358 couples at risk of transmitting a single-gene disorder have been enrolled, and 78% of initiated ovarian stimulation cycles have yielded oocyte retrievals (ORs) (Figure S1). There were 384 embryo transfers leading to 127 clinical pregnancies (fetal heart beat) and 95 deliveries (18% per OR, similar to the 20% reported by the ESHRE consortium in De Rycke et al., 2015De Rycke M. Belva F. Goossens V. Moutou C. SenGupta S.B. Traeger-Synodinos J. Coonen E. ESHRE PGD Consortium data collection XIII: cycles from January to December 2010 with pregnancy follow-up to October 2011.Hum. Reprod. 2015; 30: 1763-1789Crossref PubMed Scopus (113) Google Scholar). Due to multiple births common in in-vitro fertilization (IVF), 118 babies were born in total. Of note, the rate of success (18%, measured by the number of deliveries per OR) was lower for PGD couples compared to the rate for couples that underwent IVF for infertility alone during the same period (38% per OR in PCBN). In the PGD cohort, after several attempts, only 95 couples out of the 358 brought home a baby. In other words, PGD, with the goal of having a healthy baby, had a failure rate of 73%. Is it possible to do better? A significant number of cycles did not result in embryo transfer (n = 141, Table S1). To achieve embryo transfer on day 4 or 5, at least one embryo that is both unaffected and harboring viable hallmarks is needed. Maternal age is known to negatively impact the success of all assisted reproductive technologies but there was no significant difference in maternal age between the couples who achieved embryo transfer (mean age: 32.2 years, range 23–41) versus those who did not (mean age: 31.5 years, range 21–39). The chance of successful transfer correlated with the mean number of good quality embryos available to biopsy, meaning that the embryos harbored viable hallmarks (mean = 5.1 versus 3.2 in transferred and non-transferred patients, respectively). In addition to the presence of viable hallmarks, another factor predictive of achieving embryo transfer was the embryo’s genetic risk, depending on the model of inheritance (Table S1). The higher the risk of the embryo being affected with the genetic disease, which would be detected by PGD, the lower the chance to transfer. For cases at risk of dominant disorders, 30% of ORs did not reach the embryo transfer stage versus 20% in cases at risk of recessive disease. For embryos at risk of mitochondrial DNA disorders, there was a particularly high risk of having no embryo transfer (4/4 in this small series). But these numbers do not highlight the fact that in many cases the affected embryos expressed the morphology hallmarks of viable embryos (52 cycles, n = 43 couples). Our experience clearly shows that one major limitation of PGD is an insufficient number of viable unaffected embryos available for transfer. In this series, 959 affected embryos were discarded. While most of these embryos were viable and carried morphological criteria predictive of a good chance of implantation, they were discarded due to their genetic defect detected after PGD. Careful consideration of the use of genome editing could enable the rescue of viable embryos in the future, thereby increasing the number of embryos available for transfer, and improving the chance of a successful pregnancy. Clinical application of genome editing techniques on human embryos is however obviously premature for several reasons. Applying safe genome-editing therapies on embryos affected by a genetic disorder requires the correction of the deleterious mutation. This implies not only cutting DNA at the right place, without off-target effects, but also efficiently mobilizing the Homologous Directed Recombination (HDR) DNA-repair system to copy the wild-type sequence. Genome editing on preimplantation embryos has been performed in non-human species (Tan et al., 2016Tan W. Proudfoot C. Lillico S.G. Whitelaw C.B. Gene targeting, genome editing: from Dolly to editors.Transgenic Res. 2016; 25: 273-287Crossref PubMed Scopus (100) Google Scholar). The percentage of newborns carrying the genetic modification was less than 50% in most studies. Additionally, some animals showed partial correction only, with a mosaicism for the genetic modification, and in most cases, genes were edited through non-homologous end joining recombination (NHEJ), resulting in random insertions or deletions. Such results are clearly far from fulfilling criteria required for a translation to clinical application. In human embryos, CRISPR/Cas9 has been used to edit tri-pronuclear zygotes (Liang et al., 2015Liang P. Xu Y. Zhang X. Ding C. Huang R. Zhang Z. Lv J. Xie X. Chen Y. Li Y. et al.CRISPR/Cas9-mediated gene editing in human tripronuclear zygotes.Protein Cell. 2015; 6: 363-372Crossref PubMed Scopus (752) Google Scholar, Kang et al., 2016Kang X. He W. Huang Y. Yu Q. Chen Y. Gao X. Sun X. Fan Y. Introducing precise genetic modifications into human 3PN embryos by CRISPR/Cas-mediated genome editing.J. Assist. Reprod. Genet. 2016; 33: 581-588Crossref PubMed Scopus (196) Google Scholar) and diploid zygotes carrying mutations of the HBB and G6PD genes (Tang et al., 2017Tang L. Zeng Y. Du H. Gong M. Peng J. Zhang B. Lei M. Zhao F. Wang W. Li X. Liu J. CRISPR/Cas9-mediated gene editing in human zygotes using Cas9 protein.Mol. Genet. Genomics. 2017; 292: 525-533Crossref PubMed Scopus (143) Google Scholar). In these studies, HDR efficiency was low and several edited embryos were mosaic at the blastocyst stage. Recently, Ma et al. injected CRISPR/Cas9 proteins at the time of intra-cytoplasmic sperm injection, and reported more efficient HDR and a major reduction in mosaicism (Ma et al., 2017Ma H. Marti-Gutierrez N. Park S.W. Wu J. Lee Y. Suzuki K. Koski A. Ji D. Hayama T. Ahmed R. et al.Correction of a pathogenic gene mutation in human embryos.Nature. 2017; 548: 413-419Crossref PubMed Scopus (598) Google Scholar), but their results seem controversial (Egli et al., 2017Egli, D., Zuccaro, M., Kosicki, M., Church, G., Bradley, A., and Jasin, M. (2017). Inter-homologue repair in fertilized human eggs? bioRxiv 181255; doi: https://doi.org/10.1101/181255.Google Scholar). Concerns about potential undesired effects still remain to be addressed since off-target effects have been reported (Liang et al., 2015Liang P. Xu Y. Zhang X. Ding C. Huang R. Zhang Z. Lv J. Xie X. Chen Y. Li Y. et al.CRISPR/Cas9-mediated gene editing in human tripronuclear zygotes.Protein Cell. 2015; 6: 363-372Crossref PubMed Scopus (752) Google Scholar). Further research on human models is therefore absolutely needed to assess genome editing efficacy and safety and to improve the performance of the technique. Such research is not being undertaken in France, mainly because of the confusing nature of the French law. In France, the creation of “transgenic” embryos (article L2151-2 of the Public Health Code) is forbidden. While transgenic embryos are not “edited” embryos, the terms can be confounded by jurisdiction, and legal clarification is needed. Left-over affected embryos should be the first candidates for such research. Indeed, once the safety and efficacy questions have been resolved, we do not see any ethical barriers preventing the application of genome editing on PGD embryos affected by serious genetic conditions with no available treatment. Research on genome editing in human embryos should enable us to offer couples, one day, the possibility of rescuing their affected embryos. International discussions need now to focus on delineating consensus guidelines to prevent the slippery slope of “human enhancement.” In conclusion, the argument to ban research on embryo genome editing because alternatives exist needs nuance and is questionable, because PGD is far from ideal and many embryos are lost that could be used in the future thanks to gene editing. The authors thank Dr. Maya Chopra for her critical reading of the manuscript. J.S. and N.F. are supported by AFM grants (20304). Download .pdf (.74 MB) Help with pdf files Document S1. Table S1 and Figure S1