Our study demonstrated the feasibility of replicating the sperm genome via androgenesis and selecting the desired gamete before fertilization to preserve a specific paternal genotype confirmed by phenotypical observation and corroborated by genetic testing. We achieved a satisfactory pre-implantation developmental rate of the conceptuses generated using those replicated male gametes and were able to generate healthy offspring. Specifically, in our experiment using 8-cell stage androgenetic embryos, a single spermatozoon can yield up to 3 conceptuses that carries the identical paternal haplotype.
Interestingly, the embryo development of haploid pseudo-embryos reported in the first part of the study, whether androgenetic or parthenogenetic (Table 1), demonstrated a similar trend observed in previous studies, particularly the work done on digyneic and dispermic embryos, which underscores the importance of interdependence and complementation of both paternal and maternal genome in mammalian embryogenesis (35). Once the haploid androgenetic pseudo-blastomeres were used as gametes, the conceptuses generated demonstrated comparable embryo development comparing to control (Table 2) when 2- or 4-cell stage blastomeres were used. The utilization of 8-cell stage blastomere is less successful possibly caused by an intrinsic imprinting or epigenomic heterogeneity between sibling blastomeres since mouse embryo polarization occurs at 8-cell stage (36–39). However, more recent studies have suggested that the loss of totipotency can occur even in earlier stage at 2-cell in ~ 30% of the blastomeres (40) and further amplified in 4-cell stage embryos (41). Therefore, when an 8-cell stage pseudo-blastomere that is already differentiated was used as a male gamete, embryos generated by such cell may result in embryo development arrest, implantation failure and post-implantation development arrest.
Another explanation for why 8-cell androgenetic pseudo-blastomeres yielded less blastocyst per embryo reconstruction could be attributed to a higher level of asynchrony of epigenome between the native maternal genome and replicated paternal genome. During murine preimplantation development, zygotic genome activation occurs at 2-cell stage (42) and the embryo genome expression is heavily regulated by the paternal chromatin (43). Therefore, when a pseudo-gamete was used to generate a zygote, the intricate regulation of zygote genome activation and epigenome modification may be disrupted, therefore causing a lower pre-implantation development rate and low implantation rate. In contrary, when the nucleus of 2-cell or 4-cell were used as the source of paternal genome, the blastocysts rates were reported at 80.2% and 81.1% respectively, which is comparative to control (80.8%) and higher than experimental embryos generated from 8-cell stage pseudo-blastomeres (60.7%). This observation maybe supported by the fact that the haploid genome of earlier stage pseudo-blastomere did not underwent extensive demethylation comparing to later stage embryos (44), and therefore, the 2-cell and 4-cell stage pseudo-blastomere retains a more gamete-like epigenome. Nonetheless, despite a higher blastocyst rate per embryo reconstruction experiment, the eventual success rate per spermatozoon is still comparable. The use of 8-cell stage pseudo-blastomere is still preferred since more male gamete copies can be used to fertilize a larger cohort of oocytes.
When considering the use of this approach in higher mammal and even in human reproduction, this method of gamete replication is preferred over the utilization of haploid embryonic stem cells. Conventionally, haploid androgenetic embryonic stem cells (haESCs) are generated from haploid androgenetic embryos which has limitations including a low blastulation rate at 10–20%, supported by the data reported in our study at 11.2% (23, 45), a low cell line derivation rate about 20% (23), and prone to a rapid loss of haploidy unable to be blocked by FACS purification (46, 47). Comparing to haESC with low haploidy rate at 2–13% (17), as a precursor to haESCs, androgenetic blastomeres retain their haploidy at a higher rate (88.2%, Table 4).
Additionally, Y-chromosome bearing haploid androgenetic blastocyst are unable to yield haESC lines, which further inhibits the application of haESCs in reproductive semi-cloning (48, 49). However, our technique using haploid androgenetic blastomeres as gametes were able to generate male offspring. We do, however, observed a trend toward more viable female offspring (60.6%) comparing to male (39.4%), which could be explained by that an X-chromosome-bearing haploid blastomere has higher embryo competence comparing to the Y-chromosome-bearing counterpart. This trend was indeed true in early studies done on androgenetic and parthenogenetic embryos, where androgenones (two male pronuclei) arrest earlier while parthenogenetic embryos arrest post-implantation, signifying the role of X chromosome in these peculiar embryonic constructs (50).
During the generation of haploid androgenetic embryos, the sperm DNA decondenses in the ooplasm during pronuclear stage. The availability of an unraveled sperm genome creates a conducive environment for heritable genome editing experiments favoring the fore-coming progeny. This strategy can ensure the uniformity of the eventual gamete genotype, and the subsequent use of edited gamete can therefore prevent genetic mosaicism sometimes observed following CRISPR-Cas9-mediated genome editing on diploid embryo (51, 52).
One major concern of this technique is the species specificity. This tailored technique is currently optimized for murine gametes with limited generalizability if applied in human gametes due to the different physiology between rodent and primate reproduction. One major uncertainty is the role of sperm centrosome. In murine reproduction, the zygotic spindle is formed from the oocyte’s own microtubule organization center (MTOC), while the mitotic spindle of human zygote is contributed by the sperm proximal centriole (53, 54). Therefore, whether the human haploid blastomere will carry a functional MTOC still needs to be investigated. In term of epigenetic modification, we hypothesize that our technique may perform better on human gametes. Mouse zygotic genome activation occurs as early as 2-cell stage and this explains why the development of embryos generated from 8-cell stage blastomere is lower. Contrarily, higher mammal, including primates, has a later zygotic genome activation occur at 8-cell stage (55). Therefore, the maintenance of a more gamete-like epigenome may compensate for the development of conceptuses generated using our technique.
When considering clinical application of male gamete replication, our technique represents an alternative to traditional PGD of embryos. Screening gametes before fertilization may be more ethically palatable than embryo selection by PGD. The ability to identify healthy gametes prior to insemination can prevent the generation of excessive embryos with pathological genetic condition, therefore reducing embryo wastage. Moreover, this technique can also benefit men with severe oligozoospermia or even cryptozoospermia. By replicating scarce spermatozoa, the abundance of sperm copies permits the insemination of entire oocyte cohort using only few available male gametes.
In conclusion, haploid androgenetic blastomeres can function as replicated spermatozoa, demonstrated satisfactory blastocyst development, and yielded healthy offspring. This replication of sperm genome allowed us to select gametes of desired genotype prior to the generation of an embryo.