Problems of sperm fertility: A reproductive biologist's view

Abstract

For natural fertilization to occur successfully, millions of spermatozoa must be deposited in the lower end of the female genital tract during mating. Considerably fewer spermatozoa are required for fertilization when spermatozoa are deposited in the upper region of the female tract. The extreme case of assisted fertilization is the direct injection of a single spermatozoon into an oocyte in vitro, which is referred to as intracytoplasmic sperm injection or ICSI. All that is required to fertilize an oocyte is a single spermatozoon with a genetically and epigenetically normal nucleus. Even spermatozoa with grossly misshapen heads and no motility at all are able to produce normal offspring by ICSI as long as their nuclei are normal. There is a distinct difference between natural and ICSI fertilization. In ICSI the sperm plasma membrane and the acrosome, which never enter the oocyte's cytoplasm during natural fertilization, are injected into an oocyte. There are reports that the oocytes injected with acrosome-less spermatozoa develop better than those injected with acrosome-intact spermatozoa. At least in the mouse, pre-spermatozoal cells (e.g., round spermatids) are able to produce fertile offspring by injection into oocytes. Whether ‘spermatozoa’ produced from embryonic stem (ES) or induced pluri-potent stem (iPS) cells are functional remains to be determined. Will assisted reproduction increase overall male infertility by spreading genes involved in infertility? This is most unlikely in view of the widespread propagation of spontaneous mutation in the general population. When assisted reproduction technologies are perfected, pregnancy through assisted reproduction will become as safe as pregnancy through natural conception.

Keywords:

• fertility
• fertilization
• ICSI
• infertility
• oocyte
• sperm

Introduction

Under normal conditions, the life of all mammals begins with the fusion of an oocyte with a single spermatozoon in a process called fertilization. Thousands of genes and biochemical events are involved in (a) the formation of oocytes and spermatozoa within gonads, (b) the maturation of spermatozoa in the male epididymis, (c) sperm ascent through the female tract, and (d) interactions of spermatozoa with oocytes before and during their union. Failure in any one of these events results in fertilization failure and, consequently, infertility. Many clinical investigators and technical staff members engaged in assisted fertilization tend to ignore the processes and mechanisms of natural fertilization. This is partly understandable, but it is somewhat alarming to basic researchers like myself. To make assisted fertilization technologies safe and more efficient, we must keep learning about natural fertilization and embryo development, Mother Nature's masterpiece.

Spermatozoa participating in natural fertilization

Fertilization and embryo development in mammals take place deep inside the body of the female. Of the hundreds of millions of mature spermatozoa deposited in the lower end of the female tract during mating, only a very few reach the upper (ampullary) region of the oviduct where fertilization takes place. The vast majority of spermatozoa are either drained from the female's body or engulfed by uterine, oviductal and/or peritoneal epithelia, or phagocytized by leokocytes without participating in fertilization. In the golden hamster, 10–70 million spermatozoa are deposited in each uterus during mating [Yanagimachi and Chang 1963], yet less than two spermatozoa are seen swimming free in the lumen of the ampulla at any time during the progression of fertilization [Smith et al. 1987]. The number of spermatozoa in the ampulla exceeds the number of oocytes in this region of the oviduct only after more than 50% of oocytes are fertilized [Cummins and Yanagimachi 1982]. Even in the rabbit with much larger oviducts than the hamster, only 20–102 spermatozoa are present in the ampulla around the time of fertilization [Overstreet and Cooper 1979]. In humans, less than 500 [Williams et al. 1992] or even less than 100 spermatozoa [Morgenstern et al. 1966; Williams et al. 1993] are found in the ampulla during the estimated time of fertilization. In other words, less than 0.00005% of ejaculated spermatozoa reach the ampulla by the time of fertilization (Table 1). Although the female genital tract and its secretions (e.g., cervical mucus) prevent the ascent of poorly motile or highly deformed spermatozoa [Katz et al. 1990; Suarez and Pacey 2006], some deformed spermatozoa reach the oviduct [Mortimer et al. 1982]. It is obvious that not all fertilizing spermatozoa in vivo are genomically and epigenetically normal as evidenced by the birth of offspring with paternally inherited malformations and/or health problems [Miller and Therman 2001; Horsthemke 2006; Butler 2009].

Table 1. Number of Spermatozoa at the Site of Fertilization During Progression of Natural Fertilization.

Species	Number of sperm in oviduct ampulla
Hamster	2
Rabbit	20–100
Human	<100–500 (0.00005% of ejaculated sperm)

Number of spermatozoa necessary for natural and assisted fertilization

Since the vast majority of human spermatozoa do not reach the site of fertilization in vivo, men are most likely infertile when the number of spermatozoa in their semen is less than 39 million per ejaculate and the percentage of progressively motile spermatozoa is less than 30% [WHO 2010]. Since human in vitro fertilization (IVF) bypasses the entire female tract, far fewer (500–50,000) spermatozoa are required to fertilize oocytes [van der Ven et al. 1989]. When spermatozoa are injected directly in the perivitelline space of an oocyte a process known as subzona injection or SUZI [Krzyminska et al. 1992] only 3–5 spermatozoa are needed to fertilize each oocyte. Direct injection of a single spermatozoon into an oocyte, which is commonly referred to as intracytoplasmic sperm injection or ICSI [Palermo et al. 1992; Van Steirteghem et al. 1993], bypasses the need for sperm interaction with the oocyte's vestments (cumulus oophorus and zona pellucida) and even membrane fusion between the oocyte and the spermatozoon. A few, weakly twitching spermatozoa in semen are all that is required for successful ICSI [Palermo et al. 1992; Van Steirteghem et al. 1993]. Many healthy babies have been born after ICSI using extremely ‘poor’ sperm samples that failed completely to fertilize by conventional IVF [Nagy et al. 1995; Svalander et al. 1995]. Today, the number of infertility clinics using ICSI as the major means of assisted fertilization is increasing rapidly because of its efficiency.

Minimum requirements of spermatozoa for natural and assisted fertilization

The spermatozoa that are going to fertilize oocytes in vivo must have undergone normal processes of formation (spermatogenesis) in the testis, and physiological maturation in the epididymis, all without any errors. After entering the female tract, spermatozoa must interact properly with the female tract to undergo capacitation before undergoing the acrosome reaction on or near the oocytes, penetrating through the oocyte's zona pellucida and fusing with the oocyte's plasma membrane, all without any errors again. Even a single failure in any one of these long processes results in male infertility. Obviously, there are thousands of different causes of male infertility. This must also be true for female infertility. It was once thought, by medical professionals as well as by laypeople, that females were largely or even solely responsible for human infertility. It is now clear that both females and males are equally responsible for infertility. Figure 1 lists male and female factors involved in natural (in vivo) and assisted fertilization such as IVF and ICSI. Of course natural fertilization requires the largest number of male and female factors to succeed. IVF fertilization, which bypasses both the cervix and the uterus, requires fewer male and female factors than natural fertilization. ICSI fertilization requires the least male and female factors as it bypasses even sperm interaction with oocyte vestments (cumulus oophorus and zona pellucida) and even membrane fusion of oocytes and spermatozoa. All we need for successful human ICSI of an oocyte is a single spermatozoon with a genetically (and epigenetically) normal nucleus, sperm-borne oocyte-activating factor (SOAF), and the centriole. Whether sperm RNAs that enter an oocyte during fertilization [Ostermeier et al. 2004; Krawetz 2005] are essential for normal embryo development remains to be determined.

Figure 1.  Male and female factors necessary for successful natural and assisted fertilization (modified from [Yanagimachi 2009]).

Since ICSI bypasses sperm interaction with the cumulus oophorus and zona pellucida, as well as membrane fusion with the oocyte, even spermatozoa that are grossly misshapen and barely motile are able to ‘fertilize’ and produce normal offspring as long as their nuclei are genomically and epigenetically normal [Yanagimachi 2005]. If the oocyte-activating factor is absent in spermatozoa, ICSI oocytes remain inactivated. They can, however, be activated by injection of intrinsic SOAF (at present, the strongest candidate is phospholipase C zeta) [Saunders et al. 2002; Heytens et al. 2009], IP3 receptor antagonist such as adenophostin [Sato et al. 1998], Ca²⁺ ionophore [Eldar-Geva et al. 2003; Terada et al. 2009] or other agents, all of which initiate repetitive increases in the concentration of Ca²⁺ in the oocyte's cytoplasm. In most mammals, including humans, the centriole-containing centrosome in the neck of the spermatozoon is known to be essential for normal fertilization [Palermo et al. 1994; Schatten and Sun 2009]. Sperm centriole-centrosome introduced into the oocyte during fertilization (or ICSI) becomes the center of the sperm aster that brings both male and female pronuclei to the center of the zygote. After fertilization, the centriole-centrosome duplicates and serves as mitotic centers for the first and subsequent cleavages of the embryo. Injection of a spermatozoon with a defective centriole-centrosome may result in the failure of the union of female and male pronuclei. Since human spermatozoa retrieved from testes are able to produce normal offspring by ICSI [Schoysman et al. 1993; Devroey et al. 1996; De Croo et al. 2000], human sperm centriole-centrosomes must be functional before spermatozoa begin their ‘maturation’ in the epididymis. This may not be the case for some animals such as cats [Comizzoli et al. 2006].

Differences between natural fertilization and ICSI: how to make ICSI closer to natural fertilization

There are distinct differences between natural fertilization and ICSI. During natural fertilization and IVF, the outer acrosomal membrane of the principal segment of the acrosome and the overlying plasma membrane undergo multiple fusions, releasing the contents of the acrosome (acrosome reaction) before the spermatozoon passes the oocyte's zona pellucida and fuses with the oocyte (Fig. 2) [Yanagimachi 1994]. The sperm plasma membrane which did not participate in the acrosome reaction fuses with the oocyte plasma membrane to become a mosaic membrane covering the zygote surface [Gaunt 1983]. In other words, the sperm plasma membrane and contents of the acrosome never enter the oocyte's cytoplasm. In human ICSI, a virtually intact spermatozoon is injected into an oocyte. A single live spermatozoon is picked up in a pipette and ‘immobilized’ by scoring the sperm tail before injection into the oocyte [Palermo et al. 1992; 1996]. Since the spermatozoon is virtually free of cytoplasm, its plasma membrane is unable to repair even minor damage by itself. Sperm ‘immobilization’ means that the sperm plasma membrane is permanently damaged. The membrane will disintegrate progressively. When injected into an oocyte, the sperm plasma membrane sometimes disintegrates fast, sometimes slowly. SOAF residing between the sperm plasma membrane and the nucleus are exposed to the oocyte's cytoplasm only after the disintegration of sperm plasma membrane reaches an advanced stage. Variation in the time of onset of oocyte activation after ICSI can be explained by faster or slower disintegration of sperm plasma membranes within the oocyte [Yanagimachi 2005]. If a spermatozoon with an intact plasma membrane is injected, it may keep moving within the oocyte for 30 min or longer. Oocytes may never be activated or be activated only long after injection, depending on the speed of spontaneous plasma membrane disintegration. According to Takeuchi et al. [2004], ICSI fertilization rate was significantly higher when the sperm tail was aggressively scored than when it was gently scored. Electron microscopic examinations revealed an extensive disorganization of sperm head membranes and acrosome after aggressive scoring.

Figure 2.  Fertilization after IVF (diagrams) and ICSI (micrographs). The insert at the upper left of IVF shows the relative size of golden hamster and human gametes. Note that, in IVF the sperm acrosome and the plasma membrane do not enter the oocyte, whereas in ICSI both the sperm plasma membrane and the acrosome are injected into an oocyte. Abbreviations: ac: acrosomal cap region; cg: cortical granules; eq: equatorial segment of acrosome; iam: inner acrosomal membrane; pvs: perivitelline space; zp: zona pellucida. (A–G, from [Yanagimachi 1988]; A'–F', from [Kimura and Yanagimachi 1995]).

In the mouse, oocytes are activated much faster when sperm plasma membranes are chemically removed prior to ICSI (Fig. 3) [Morozumi and Yanagimachi 2005]. Development of ICSI embryos to live offspring is better when ICSI is performed using spermatozoa free of acrosome and plasma membrane than when using spermatozoa with intact plasma membranes and acrosomes. Injection of acrosomes into oocytes is potentially hazardous to oocytes and embryos [Morozumi and Yanagimachi 2005; Morozumi et al. 2006]. Spermatozoa of some animals (e.g., hamsters and guinea pigs) have very large acrosomes, and injection of a single acrosome-intact spermatozoon invariably causes cytolysis of an oocyte (e.g., [Yamauchi et al. 2002]). If human spermatozoa had large acrosomes, like hamster spermatozoa, injection of a single spermatozoon would inevitably result in the death of all oocytes unless sperm acrosomes were removed prior to injection [Yanagimachi 2005]. It is conceivable that spermatozoa of some men have a ‘tough’ plasma membrane or unusually high concentrations of acrosomal enzymes. Also, it is possible that oocytes of some women are extraordinarily sensitive to exotic enzymes like acrosomal enzymes. Removal of both plasma membrane and acrosomes before ICSI would make spermatozoa less stressful to the oocytes after ICSI. According to Lee et al. [1996b], human oocytes injected with acrosome-reacted spermatozoa fertilized and developed better than those injected with acrosome-intact spermatozoa. Gianaroli et al. [2010], who distinguished acrosome-reacted human spermatozoa from acrosome-intact spermatozoa by birefringence characteristics of sperm heads, reported that ICSI using acrosome-reacted spermatozoa resulted in higher implantation and delivery rates than that using acrosome-intact spermatozoa. To remove acrosomes from 100% of spermatozoa, low concentrations of membrane-disrupting agents like Triton X-100 and lysolecithin (=phosphatidylcholine) can be used [Morozumi et al. 2006], but intrinsic acrosome reaction-inducing factors such as the zona pellucida or progesterone can be used as well if we are able to distinguish (or separate) acrosome-reacted spermatozoa from acrosome-intact spermatozoa as reported by Gianaroli et al. [2010].

Figure 3.  Beneficial effects of the removal of sperm plasma (pl.) membrane and acrosome before ICSI.

Non-nuclear materials that spermatozoa bring into the oocyte during natural fertilization (e.g., axonema, dense fiber, mitochondria, and the fibrous sheath of the tail) disappear during preimplantation development of the embryo [Hiraoka and Hirao 1988; Sutovsky et al. 1996]. Lysosomes and/or ubiquitin-proteasome systems are believed to be involved in this process. It is known that sperm mitochondria are ubiquitinated in the ooplasm and subjected to proteolysis during embryo development [Sutovsky et al. 2000]. It is unknown at present whether sperm plasma membrane and acrosomal contents (including enzymes) that do not enter the oocyte during natural fertilization are subjected to the ubiquitin-proteasome system after ICSI.

According to former associates of the author who currently work at human fertility clinics, ICSI using spermatozoa freed from the plasma membrane and acrosome requires rigorous Institutional Review Board approval. If human ICSI were to start using spermatozoa free of the plasma membrane and acrosome, injection of the plasma membrane and acrosome-intact spermatozoa would certainly require more rigorous reviews by the same committees.

Spermatozoa and cells that can be used for ICSI

Spermatozoa with abnormal head morphology are not necessarily genetically abnormal. A good example of this is the spermatozoa of the BALB/c mouse. The majority (60–80%) of the spermatozoa of this strain of mouse are grossly misshapen. They are motile, but unable to fertilize under ordinary conditions, yet they are able to produce normal offspring by ICSI [Burruel et al. 1996]. Although the incidence of chromosomal abnormality is higher in spermatozoa with abnormal heads than those with normal heads (Table 2) [Lee et al. 1996a; Kishikawa et al. 1999], some spermatozoa with structurally abnormal heads are certainly normal in terms of their genome. This can explain why some, though not all, human spermatozoa with deformed heads are able to produce healthy offspring by ICSI [Battaglia et al. 1977; Lundin et al. 1995; Stalf et al. 1995; De Vos et al. 2003; Berkovitz et al. 2005]. Spermatozoa of some men are morphologically normal and highly motile, yet they are unable to fertilize any oocytes in vivo and even by IVF. This may be due to the inability of these spermatozoa to undergo the acrosome reaction on the oocyte zona pellucida [Liu et al. 1995] or to the presence of anti-sperm antibody on sperm surfaces [Lahteenmaki et al. 1995]. ICSI overcomes both these problems.

Table 2.  Chromosome Analysis of Human Spermatozoa With Normal and Abnormal Head Morphology.

Sperm head morphology	Chromosomally normal (%)	Chromosomally abnormal (%)
		Numerical	Structural
Normal	92	1	7
Abnormal	70	4	26

Spermatozoa of men with immotile cilia (Kartagener) syndrome are totally infertile under ordinary in vivo conditions, yet they are able to produce healthy offspring by ICSI [von Zumbusch et al. 1998; Kaushal and Baxi 2007]. Some men do not have live spermatozoa or have no spermatozoa at all in their semen. ICSI using spermatozoa from the epididymis or testes may produce healthy offspring [Schoysman et al. 1993; Tournaye et al. 1995; Weissman et al. 2008]. Mice lacking in transition protein genes (Tnp 1 and Tnp2) are infertile. They have extremely low numbers of epididymal spermatozoa with very poor motility [Shirley et al. 2004]. Interestingly, the testicular spermatozoa of these mutant mice are better than epididymal spermatozoa in producing offspring when injected into oocytes by ICSI [Suganuma et al. 2005]. Apparently, certain types of mutations ‘damage’ spermatozoa during their passage through the epididymis. Spermatogenic cells of the CREM-null mouse are not able to develop beyond the round spermatid stage. Round spermatids of such mutant mice, however, are able to produce fertile offspring when injected into oocytes [Yanagimachi et al. 2004; Yanagimachi 2005]. According to Tanaka et al. [2009], human round spermatids co-cultured with monkey Vero cells in vitro are able to develop into motile spermatozoa, but whether these spermatozoa are functional has not been determined. Birth of normal human babies after injection of round spermatid nuclear injection (ROSNI) has been reported by some researchers (e.g., [Tesarik and Sausa 1995; Saremi et al. 2002]), but many others did not succeed (e.g., [Yamanaka et al. 1997; Urman et al. 2002]). Human ROSNI is still considered experimental [ASRM 2003].

Animal and human spermatozoa can be kept frozen (−196°C or below) indefinitely without losing their fertility, but keeping frozen spermatozoa at this temperature for extended periods of time is cumbersome and expensive. Freeze-dried mouse spermatozoa can be kept at ambient temperature for weeks [Wakayama and Yanagimachi 1998; Kusakabe et al. 2001], but they must be kept at 80°C or below to keep them fertile for many years [Kawase et al. 2005]. Long-term chemical storage of spermatozoa (e.g., in 70% ethanol) at ambient temperature has not been successful [Tateno et al. 1998]. Perhaps, the simplest way to store (mouse) spermatozoa in a fertilization-competent state is freezing isolated testes or whole male bodies at −20°C. Spermatozoa recovered from defrosted testes and bodies are dead in the ordinary sense (plasma membranes were broken), but are able to produce normal offspring by ICSI even after being frozen for 15 years [Ogunuki et al. 2006].

Several investigators have claimed successful transformation of embryonic stem (ES) cells and induced pluri-potent stem (iPS) cells into haploid cells somewhat ‘similar’ in gene expression or appearance to spermatids and spermatozoa (e.g., [Toyooka et al. 2003; Nayernia et al. 2006; Kerkis et al. 2007; Aflatoonian et al. 2009; Kee et al. 2009]). Although Nayernia et al. [2006] reported the birth of 12 mice after ICSI using strange-looking ‘red (haploid) cells’, this has not been confirmed by other investigators. As far as the mouse is concerned, ‘haploid cells’ do not need to look like spermatozoa at all as long as they have the correct set of haploid chromosomes. One thing that is very clear is that the cells would be under enormous environmental stress during their development into ‘artificial spermatozoa.’ Unless these ‘spermatozoa’ are genomically and epigenetically identical with normal spermatozoa, ICSI using such artificial spermatozoa would result in the death of embryos or the birth of offspring with severe congenital malformation or serious health problems.

Safety of ICSI

The safety of assisted fertilization is our prime concern. Theoretical concern of assisted fertilization, in particular ICSI, stems from the use of the spermatozoa that are unable to participate in natural fertilization. Will babies born by assisted reproduction have serious medical and physical problems? Will in vitro manipulation of spermatozoa, oocytes, and embryos have lasting effects on the health of offspring? Some congenital defects and epigenetic disorders, which are not detected at birth, may become apparent as children grow.

Obviously, assisted reproduction is by no means risk-free, but reproduction through natural fertilization is not risk-free either. From time to time, laypeople encounter articles in popular magazines and newspapers that state that − babies conceived through IVF and ICSI are more than twice as likely as naturally conceived babies to suffer major birth defects and nearly three times as likely to be born small, with risk factors for later cardiovascular and skeletal muscular problems. Such statements may scare IVF and ICSI candidates, but they should be aware that their chance of having healthy babies is 90% or higher if they become pregnant [Wennerholm et al. 2000; Anthony et al. 2002; Bonduelle et al. 2002; 2003; 2004]. The decision of whether or not to initiate therapy must be made by couples who are fully informed of the nature and magnitude of potential therapy risks [Meschede et al. 1995]. For most couples who desperately want to have their own babies, benefits of assisted reproduction overweigh their concern [Tournaye and Van Steirteghem 1997].

The human is rather unique in that 21–37% of oocytes and 7–15% of spermatozoa are chromosomally abnormal even in normal, fertile couples (for references, see Table 2 of [Yanagimachi 2005]). Meiotic errors responsible for chromosomal abnormalities of oocytes and spermatozoa increase with advancement of maternal and paternal age. Since the oocyte and spermatozoa with abnormal chromosomes/genomic constitutions are not discriminated against during fertilization [Almeida and Bolton 1994; Meschede et al. 1995; Engel et al. 1966; Marchetti et al. 1999], one-third or even half of naturally fertilized human eggs and early embryos could be chromosomally/genomically abnormal. High pregnancy loss even in fertile women has been known for many years [Hertig et al. 1959; Carr 1971; Kerr 1971; Edwards 1986]. An interesting study using mutant mice was reported by Gropp et al. [1983]. When embryo in vivo development was examined, all except for trisomy 19 did not survive crucial phases of development; only embryos with trisomy 19 survived beyond birth (Fig. 4). Even though mother's stringent elimination of defective embryos must take place during pregnancies following natural conception (Fig. 5) and assisted fertilization, we cannot be too cautious about the handling of gametes and embryos in vitro. Experiments using mice warn us of the inadvertent introduction of bacterial or viral DNAs into oocytes during ICSI [Moreira et al. 2005]. ICSI using spermatozoa with DNA damage may produce potential health and behavioral problems in the later years of the offspring [Fernandez-Gonzales et al. 2008]. Even short term of embryo culture in vitro under suboptimal conditions may cause aberrant expression of imprinted genes during embryogenesis [Rivera et al. 2008].

Figure 4.  Development of mouse embryos with chromosomal monosomies and trisomies. Note that all monosomic embryos die before or soon after implantation. With the exception of trisomy 19, all trisomic embryos die during postimplantation development (from [Gropp et al. 1983]).

Figure 5.  A reason why many mutations related to male infertility factors are hidden in the general population. The germ cells developing from a single spermatogonium remain cytoplasmically connected to one another during spermatogenesis. Most of the genes responsible for male infertility are likely recessive. Therefore, there is no effect on germ cell differentiation if the spermatogonia contain a mutation in one allele of a gene important for spermatogenesis, the germ cells are capable of exchanging mRNAs and proteins effectively negating the mutation. All spermatozoa are morphologically and functionally normal and fertile (from [Engel et al. 1996]).

A major problem in IVF and ICSI has been the substantial increase in multiple pregnancies, which inevitably results in the ‘competition’ between growing fetuses and low birth weights of many newborns. It seems to be multiple pregnancy, not the IVF or the ICSI procedure itself, that causes a slight yet significant increase in birth defects and pediatric problems [Wennerholm et al. 2000; Anthony et al. 2002; Bonduelle et al. 2002]. In many infertility clinics today, a single embryo transfer is recommended to patients (except for older mothers) to eliminate the chance of preterm birth and multiple pregnancy. However, since the success rate of pregnancy is lower after a single embryo transfer than after a multiple transfer [Baruffi et al. 2009], the majority of couples of advanced age have preferred the latter if they thought there was even a slight chance of a higher pregnancy rate.

Will assisted fertilization technologies increase overall male infertility by the spreading of genes involved in infertility? Engel et al. [1996] and Faddy et al. [2001] do not think so in view of the widespread propagation of spontaneous genetic mutations in the general population. Engel et al. [1996] stated: “Let us assume that all fertility disorders are genetically determined and that all infertile men reproduce by ICSI, we could expect a rise in the frequency of infertile men over the following 20 generations (600 years) from 1.8% to 2.3%. Such a minimal rise in the frequency of infertile men can be explained on the basis of the biology of male germ cell development.” (Fig. 6.)

Figure 6.  Incidence of chromosomal aberrations in gametes, zygotes, embryos, and newborns before and during normal pregnancy in humans (from [Plachot et al. 1988]).

In many infertility clinics today, ICSI is used for nearly two thirds of couples who need assisted fertilization technologies. It might have been used for some couples who do not really need it partly because the female partners may feel that they are approaching the end of their fertile life and cannot afford any more time. Check et al. [2009] prefer IVF for men with mild male infertility factors because embryos obtained by IVF are more likely to implant than ICSI embryos. IVF, when it is properly performed, is likely less traumatic to oocytes than ICSI. It became increasingly clear that expression of some genes, in particular that of imprinted genes, is vulnerable to environmental stress [Doherty et al. 2000; Khosla et al. 2001]. Gametes, zygotes, and embryos exposed to less than optimal in vitro conditions may develop into offspring which suffer from epigenetic health problems throughout their lives (e.g., [Fernandez-Gonzales et al. 2004; Lucifero et al. 2004; Paoloni-Giacibino and Chaillet 2004; Maher et al. 2003; Shiota and Yamada 2005; Butler 2009]). Thus, we cannot be too cautious about in vitro handling of gametes, zygotes, and embryos. It is most unlikely that media and culture conditions we use today replace all of the benefits that the female reproductive tract provides [Bavister 2000].

Can we identify ‘good’ spermatozoa before ICSI?

We must first clarify the difference between sperm's fertilizing ability and sperm's fertility. While the former refers to the ability of spermatozoa to fertilize oocytes, the latter refers to the sperm's ability to produce viable offspring. Since the goal of infertility treatment is to bring children to infertile couples, the sperm's fertilizing ability itself is meaningless to these couples.

(1) When very few spermatozoa are available for ICSI: For ICSI we do not need many spermatozoa. If a patient has only a few spermatozoa in semen and testis/epididymis biopsy samples, we have no choice, we must use all that are available. If 100% of spermatozoa from a patient are completely immotile, we must evaluate if they are ‘alive’ or ‘dead’ by hypo-osmotic swelling test (e.g., [Jeyendran et al. 1984; van der Ven et al. 1986; Liu et al. 1997]) or dye-exclusion test (using 0.1 0.5% Eosin Y in 0.9% NaCl or fertilization medium) [WHO 2010]. If a particular spermatozoon is judged ‘dead,’ we should not use it as we never know if this spermatozoon died one minute ago or several days ago. We cannot expect genomically ‘intact’ sperm nuclei when spermatozoa died within the male's body several days ago. If more than ten spermatozoa are found ‘alive,’ which should be used for ICSI? At present, we are unable to distinguish chromosomally (genomically) normal spermatozoa and oocytes from those that are abnormal without killing them. Several investigators maintain that the presence of a large ‘vacuole(s)’ in the sperm nucleus is negatively associated with sperm's DNA integrity and fertility (e.g., [Berkovitz et al. 2006; Garolla et al. 2008; Franco et al. 2008]). Therefore, it is wise to avoid such spermatozoa for ICSI. Since the ultimate goal of assisted reproduction is to produce healthy offspring, the result of the union of male and female gametes, we must keep trying to distinguish genomically normal oocytes and spermatozoa from those that are abnormal before bringing them together.

(2) Selection of oocytes: Will it be possible to distinguish genomically normal oocytes and embryos from abnormal ones using non-invasive methods? It is not certain at present whether several methods proposed [Montag and van der Ven 2008; Scott et al. 2008; Sturmey 2008; Katz-Jaffe and Gardner 2008] can really evaluate the physiology of oocytes and predict the oocyte's ability to develop into normal offspring. Figure 7 illustrates a hypothetical protocol to select ‘good’ oocytes and spermatozoa for ICSI.

Figure 7.  Selection of ‘good’ gametes and embryos to guarantee the birth of healthy offspring by assisted fertilization. (A–E) Selection and preservation of oocytes. (F–J) Selection and preservation of spermatozoa. (K–M) Construction of zygotes and production of transferable embryos. Abbreviations: OGS: oocyte genetic screening; SGS: sperm genetic screening; ET: embryo transfer; PGS: preimplantation embryo genetic screening.

(A) Select oocytes of normal size, each having one polar body and good-looking cytoplasm [Serhal et al. 1997; Kahraman et al. 2000; Khalili et al. 2005]. Activate these oocytes for example by treatment with a Ca²⁺ ionophore. At present, we are not certain whether the proposed non-invasive methods such as optical imaging and measurement of oocyte respiration rate [Montag and van der Ven 2008; Nagy et al. 2008; Scott et al. 2008; Sturmey 2008] are able to determine normality and abnormality of individual oocytes. (B) Before activated oocytes are frozen (C), collect the second polar body and transfer its nucleus into a previously enucleated oocyte (D) which will then be activated. (E) Collect nuclei of a haploid parthenogenetic embryo to perform oocyte genetic screening including comprehensive screening for the entire chromosome complement [Wells and Delhanty 2000; Wells et al. 2008], microarray analysis [Patrizio et al. 2007], and determination of epigenetic status (e.g., DNA methylation) of imprinted genes. Ideally all genes should be examined, but it would be more efficient and practical to examine the statuses of genes which are known to be responsible for genetic and epigenetic diseases or health problems with high rates of prevalence in the general population. (C) The oocytes that pass the tests are kept frozen until use.

(3) Selection of spermatozoa: (F) Select spermatozoa with normal head morphology. If no structurally ‘normal’ and motile spermatozoa are available, any spermatozoa can be used as long as they are ‘alive.’ (G-I) Inject a single spermatozoon into a previously enucleated human oocyte. (J) When the haploid androgenic embryo reaches the 8-cell stage, collect 1–2 blastomeres randomly from each embryo for the genomic and epigenetic screening. This is referred to as sperm genetic screening. Embryos that passed the tests are kept frozen until use. Attempts to duplicate the sperm's haploid genome have begun [Kuznyetsov et al. 2007; Palermo et al. 2009].

(4) Construction of zygotes and production of transferable embryos: When oocytes and spermatozoa with perfectly normal genetic and epigenetic states are identified (E and J), defrost parthenogenetic and androgenic embryos one at a time. Inject the sperm-derived haploid nucleus into an oocyte to produce a zygote (C'-K), allowing it to develop into the 8-cell (L) or blastocyst stage (M). Genomic and epigenetic screening is performed once more at this stage. Only the embryos with correct genes, and in the correct epigenetic state, are transplanted.

Ideally, all oocytes used in the above screenings should be the patient's own. If donor oocytes are available, screening would be easier. When a patient's or a donor's oocytes are readily available for steps D-E, several blastomere nuclei (haploid) can each be used as the substitute for a single female pronucleus. If a parthenogenetic embryo at step E is used for genetic diagnosis only, the use of animal oocytes could be acceptable because the nuclei of these embryos are not going to be used to produce offspring. The use of animal oocytes for steps G-J (which involve exposure of the human sperm nucleus to the cytoplasm of an animal oocyte) is, of course, unacceptable. We can expect opposition to the use of human oocytes, even if they are enucleated for diagnostic purposes.

(5) Remarks on genetic and epigenetic screenings of gametes and embryos: In some countries where genetic screening of developing embryos is prohibited (e.g., Germany), polar body screening is perhaps the only way to allow us to perform genetic screening of pre-embryos (e.g., [Renbaum et al. 2007; Fiorentino et al. 2008; Griesinger et al. 2009; Montag et al. 2009]). Obviously, polar body screening is incomplete because we are unable to assess the paternal contribution to the embryos, which is as important as the maternal contribution.

Transmission of several hereditary diseases caused by chromosome aberrations and single-gene defects has been prevented successfully by screening and selective transfer of preimplantation embryos (e.g., [Kuliev et al. 1999; Verlinsky et al. 2004; Obradors et al. 2008; Chow et al. 2009; Gutierrez-Mateo et al. 2009; Munne et al. 2009]). In most preimplantation screenings, one or two blastomeres are removed from an 8-cell embryo or a blastocyst. Whether this cell removal is harmful to the embryos [Mastenbroek et al. 2007; Gossens et al. 2008 ; Staessen et al. 2008; Fauser 2008], and whether screening of only one or two cells can accurately assess the genetic and epigenetic status of embryos is controversial [Goossens et al. 2008; Basille et al. 2009; De Vos et al. 2009; Wilton et al. 2009]. Frequent genomic (chromosomal) mosaicism in blastomeres of early cleaving embryos [Coulam et al. 2007] makes accurate judgment of normality and abnormality of individual embryos difficult.

Conventional fluorescence in situ hybridization (FISH), which allows us to detect only a few chromosomal and genomic abnormalities in each embryo, has limited value in the preimplantation genetic screening of embryos. Other methods, such as comprehensive chromosomal screening using a few trophectoderm cells of a blastocyst [Wells and Delhanty 2000; Wells et al. 2008], which allow us to detect far larger numbers of gene defects, would be more appropriate for genomic screening of individual embryos [Schoolcraft et al. 2009]. Although trophectoderm cells are known to have developmental totipotency [Tsunoda and Kato 1998], it remains to be determined whether they are the best cells to assess genetic and epigenetic states of imprinted genes of the embryo itself. Genomic and epigenetic screening is labor-intensive and expensive. The screening of either the oocyte alone or the spermatozoon alone does not guarantee the production of normal offspring. Therefore, the optimal and most economical solution would be to run a one-time preimplantation diagnosis at either the cleavage or blastocyst stage. When the techniques of preimplantation genetic and epigenetic screenings are perfected, pregnancy through assisted fertilization will become as safe as, or even safer than, that achieved through natural fertilization.

Declaration of Interest: The author reports no conflict of interest. The author alone is responsible for the content and writing of the paper.

Abbreviations

ICSI:	=	intracytoplasmic sperm injection
IVF:	=	in vitro fertilization
ROSNI:	=	round sperm nucleus injection
SOAF:	=	sperm-borne oocyte-activating factor.

Acknowledgments

I am grateful to Drs. Thomas Huang, James Overstreet, Gianpiero Palermo, Heide Schatten, Peter Sutovsky, and Ted Tollner for providing me with valuable information on natural and assisted fertilization. I am indebted to Drs. Douglas Carrell and Steve Ward, and Mrs. Charlotte Oser for reading the original manuscript. I thank Dr. Hideaki Yamashiro for his assistance in the preparation of figures and diagrams.