A computerized literature search was conducted for all publications in PubMed/Medline until December 2021 using the following MeSH or keyword terms: “Fertilization failure” or “Failed fertilization” or “Poor fertilization” or “oocyte activation deficiency” AND “In vitro fertilization” or “IVF” or “Intracytoplasmic sperm injection” or “ICSI” AND “Whole exome sequencing” or “WES” or “Next generation sequencing” or “NGS” or “Sanger sequencing” AND “genetic” or “gene” or “Mutation” or “Variant”. The search strategy is shown in Supplementary Table S1.

The title, abstract, and full texts of the retrieved articles were reviewed, and the relevant articles were included. Only English-language studies or articles in other languages, but with a detailed abstract in English were enrolled. In addition, in the present systematic scoping review, we have opted to focus on the studies that discussed specific gene mutations associated with poor or failed fertilization in IVF/ICSI treatments and provided strong and convincing evidence for the roles of the genes in the corresponding phenotypes. We did not include articles regarding the association between single nucleotide polymorphisms and fertilization failure after IVF/ICSI because we believe that these studies may have a low level of evidence. Searching and screening of the articles were separately performed by two authors.

The details of the enrolled articles, including author, year of publication, number of individuals studied, identified genes, the identified cDNA and amino acid alterations, and fertilization results of IVF/ICSI attempts are shown in Table 1 .

Recent descriptive data obtained from whole exome sequencing (WES) and Sanger sequencing in humans have revealed gene mutations causing poor or failed fertilization after IVF/ICSI attempts ( Table 2 ).

Intracellular Ca²⁺ oscillations are a remarkable signaling phenomenon observed during the process of mammalian fertilization (41–43). Sperm-specific phospholipase C (PLC) termed PLCzeta (PLCZ1) is widely considered to be the physiological stimulus responsible for generating Ca²⁺ oscillations (44, 45). As the smallest member of the PLC family (∼70 kDa in humans), PLCZ1 consists of a C-terminal C2 domain, four tandem Ca²⁺-binding EF-hand domains, and the X and Y catalytic domains (45, 46). Each domain plays a specific role in determining the function of sperm PLCZ1 as a trigger of oocyte activation and early embryonic development (46). PLCZ1 identified in sperm from fertile men usually expressed within the equatorial, acrosomal, and post-acrosomal regions of the head (17, 47, 48).

In 2009, a point mutation (H398P) in the PLCZ1 gene [MIM: 608075] was first discovered in a non-globozoospermic case linked to defects in the ability of the sperm to induce calcium oscillations in the oocyte following ICSI (49). To date, 24 PLCZ1 variants associated with poor or failed fertilization after IVF/ICSI attempts have been identified using Sanger sequencing and WES in infertile men. Segregation analysis and family pedigrees showed that it was an autosomal recessive mode of inheritance (13–16, 19, 20, 46, 50). It was indicated that these 24 reported mutationswithin the PLCZ1 gene include missense, frameshift, splicing, and nonsense that were localized in the C2 domain, EF-hand domains, and catalytic domains. The homozygous nonsense mutation c.588C>A (p.Cys196*), which maps to the catalytic domain, has been frequently reported (13, 14, 20, 50). Interestingly, Torra-Massana et al. reported a missense variant, c.1499C>T (p.S500L), located in the C2-domain of PLCZ1, which seems to be the most frequent mutation (19). This mutation was found in nine patients with poor or failed fertilization after ICSI (19).

The pathogenicity of these identified mutations and their possible effects on PLCZ1 protein were assessed through bioinformatics analysis and revealed that these important mutations probably weakened the stability of protein function (15, 19, 20, 46, 50). PLCZ1 protein structure consists of two main regions: the catalytic domain (X- and Y-domains) and the regulatory region (EF-hands and C2- domains), suggesting that the distinct locations of identified mutations may have different effects on the PLCZ1 protein function, which may lead to different phenotypes: total or partial fertilization failure (19, 20). Additionally, microinjection with the wild-type or mutant PLCZ1 cRNA into oocytes was performed to evaluate the effect of the identified mutation on protein function. Microinjection with the mutant cRNA into oocytes failed to activate oocytes to induce the formation of pronuclear, while the injection of wild-type cRNA could effectively do, indicating that these mutations have an adverse influence on protein function (13, 15, 16, 19, 20).

The histological examination performed in PLCZ1 knockout mice revealed that loss of PLCZ1 has no deleterious effects on spermatogenesis or quality parameters associated with the ability of the sperm to penetrate, bind and fuse with the egg (18). However, PLCZ1-null sperm failed to trigger Ca²⁺ oscillations in the egg in vitro. In addition, the incidence of polyspermy following IVF or in vivo fertilization increased significantly, indicating that PLCZ1-null sperm cannot induce Ca²⁺ oscillations, which is involved in the mechanism of preventing polyspermy (18).

WEE2 (WEE1 homolog 2, also known as WEE1B) [MIM: 614084] encodes a well-conserved oocyte-specific kinase which acts as an essential dual regulator of meiosis during prophase I and metaphase II by phosphorylating Tyr15 of the CDK1/cyclin B complex (M-phase promoting factor; MPF). In the GV stage of an oocyte, inhibition of WEE2 results in germinal vesicle breakdown (GVBD) and the resumption of meiosis. The WEE2 down-regulation leads to the failure of the MII stage exit and blockade of fertilization (51). WEE2 is predominantly expressed in the ovary of the rhesus macaque and weakly detectable in the testis, but not detected in any of the somatic tissues (52). Within the ovary, the expression of WEE2 persists in the germinal vesicle and cytoplasm of metaphase I and normal MII oocytes and reaches the highest level in preovulatory follicles (52).

In recent years, more and more evidence shows that WEE2 gene mutations may lead to fertilization failure and female infertility. Eight studies have so far reported a total of 27 mutations of WEE2 in patients with fertilization failure or poor fertilization, including 12 missense mutations, 10 frame-shift mutations, three splice-site mutations, and two nonsense mutations (21–24, 26–28, 53). The homozygous frameshift mutation c.1006_1007dup (p.His337Tyrfs*24) was repeatedly reported in three articles (27, 28, 53). Furthermore, in the study by Dai J et al., it was detected in four of nine patients, indicating that the incidence of this mutation is relatively high (53). These mutations follow an autosomal recessive pattern and proved to be truncated or loss of function, thereby reducing the protein level and disturbing the phosphorylation of WEE2 (22–24, 26–28, 53). In all reports describing female infertility linked to WEE2 mutations, nearly all the retrieved oocytes from affected individuals failed to form two pronuclei after IVF/ICSI attempts. Moreover, ICSI-AOA could not rescue fertilization failure (22, 26, 53). Interestingly, the researchers achieved phenotypic rescue by injection of WEE2 cRNA into affected oocytes, as indicated by the meiotic resumption, extrusion of the second polar body, and formation of pronuclei, embryonic development (28).

TUBB8 (Tubulin beta 8 class VIII) [MIM:616768] encodes a highly conserved β-tubulin isotype that only exists in primates. Human β-tubulin consisted of nine isotypes, including TUBB1, TUBB2A, TUBB2B, TUBB3, TUBB4A, TUBB4B, TUBB5, TUBB6, and TUBB8 (25). TUBB8 accounts for the majority of all the β-tubulin isotype expressed in human oocytes and early embryos and play a key role in oocyte meiotic spindle assembly (54, 55). Most mutations in TUBB8 reported during recent years were associated with the maturation arrest of human oocytes and pre-implantation embryonic development abnormalities, while A few mutations in TUBB8 are responsible for poor or failed fertilization (29, 30, 32, 56).

Since 2016, 12 TUBB8 mutations associated with poor or failed fertilization have been identified using WES and Sanger sequencing. These TUBB8 mutations are paternally hereditary as an autosomal dominant, recessive inheritance or arise from de novo. It was indicated that 12 reported mutations in the TUBB8 gene include missense, nonsense, and frameshift. The effect of the TUBB8 mutations on microtubule dynamics, α/β tubulin heterodimeric assembly, and spindle assembly in mouse and human oocytes was evaluated by microinjection of the corresponding cRNAs (54, 55). The expression of wild-type and mutant forms of TUBB8 was measured in cultured cells (31, 54, 55). These mutations were found to interfere with α/β-tubulin heterodimeric folding, microtubule assembly, and stability, or to affect the interaction of microtubules with kinesin and potentially other microtubule-associated proteins (MAPs). The association of TUBB8 mutations with poor or failed fertilization may be due to that it results in abnormalities in the oocytes morphologically identifiable as MII, causing failure of the second polar body extrusion or zygote cleavage. The pathogenicity of these TUBB8 mutations was usually evaluated by a variety of methods, including spindle morphological observation, proband gene detection, pedigree segregation analysis, and in vitro experiments. However, it should be noted that due to a high prevalence of mutations in TUBB8, some of the mutations were only observed in the proband, and further in vitro functional verification was not carried out. Therefore, the causal relationship between these TUBB8 mutations and observed phenotypes needs to be further studied.

NLRP5 (NLR family, Pyrin domain containing 5) is a maternal effect gene [MIM: 609658], originally identified in the mouse, which is exclusively expressed in the oocyte (57). The NLRP5-encoded protein belongs to the NALP protein family. Members of the NALP protein family typically consist of an amino N-terminal pyrin domain (PYD), a NACHT domain, and a carboxyl C-terminal leucine-rich repeat (LRR) region. PYD is a protein-protein interaction module, which has been identified in multiple human proteins involved in stress signaling pathways (58). The NACHT domain is presumed to bind ATP (59). LRR region is predicted to provide support for the formation of protein-protein interactions (60). As an oocyte-selective gene, NLRP5 plays a vital role in embryogenesis in mice (61), bovines (62), rhesus macaque monkeys (63), and humans (64). NLRP5 knockout female mice are infertile due to embryo development arrest at the 2-cell stage, although follicular development, oocyte maturation, and fertilization are fairly normal (61). In rhesus macaque, the NLRP5 knockout embryos had a block of embryogenesis between the 8-cell and the 16-cell stages (65). Peng H et al. reported that NLRP5 was required for preimplantation embryo development in cows, and NLRP5 knockout embryos were mainly arrested between the 2-cell and 8-cell stages (66). A recent report revealed that mutations in NLRP5 gene in human caused an early embryonic arrest (67).

NLRP5 mutations were shown to be associated with total fertilization failure after IVF/ICSI attempts. To date, two studies reported three mutations associated with fertilization failure (33, 68). In this regard, an autosomal recessive inheritance was suggested according to family pedigree. These identified mutations were missense and frameshift. The two missense variants (c.1598G>C and c.1919T>G; p.Arg533Pro and Leu640Arg), respectively located within the NACHT domain and LRR domain, are predicted to affect NLRP5 protein function, consequently, lead to total fertilization failure in IVF and ICSI cycles. Further studies are still required to reveal NLRP5 function in fertilization as well as its disruption effect on fertilization failure.

ACTL9 (Actin-like 9) [MIM: 619251] and ACTL7A (Actin-like 7A) [MIM: 604303] belong to the family of actin-related genes. They encode actin-like protein 9 and actin-like protein 7A, respectively. ACTL7A is an important paralog of ACTL9. ACTL9 and ACTL7A proteins are co-localized in the sub-acrosomal layer of perinuclear theca (PT) in sperm and interact with each other to maintain acrosomal anchoring to the nuclear membrane.

Two missense and one nonsense mutations in the ACTL9 gene were identified using WES and Sanger sequencing in three individuals with total fertilization failure or poor fertilization in IVF and ICSI attempts (34). ACTL9-mutated individuals exist a higher proportion of tapered-head sperm compared to those in fertile individuals. The sperm acrosome of affected individuals had an abnormal ultrastructure using transmission electron microscopy. These ACTL9 mutations were proved to weaken or lost the ability to interact with ACTL7A, resulting in the inner acrosomal membrane’s detaching from the nuclear and forming a loosened perinuclear theca structure. Mutant sperm show PLCZ1 absence or abnormal localization, which leads to failure to stimulate Ca²⁺ oscillations (34).

Two compound heterozygous variants in ACTL7A were identified in a family with total fertilization failure after ICSI. Pedigree analysis indicated a recessive pattern of inheritance (36). The nonsense variant located in exon 1 of ACTL7A causes a premature stop codon and is predicted to be disease-causing. The missense variant is predicted to be damaging and disease-causing by using a silicon tool. The ultrastructure of the mutation sperm shows defects in the acrosome and perinuclear theca (36). The ACTL7A knock-in mouse model showed reduced expression and abnormal localization of PLCZ1, which could be responsible for the fertilization failure (35).

Zona pellucida (ZP) is an extracellular glycoprotein matrix that surrounds oocytes and medicates several important roles in the acrosome reaction induction, blocking of polyspermy, and protection for the preimplantation embryo (69). ZP2 (zona pellucida glycoprotein 2) [MIM: 182888] encodes the ZP2 glycoprotein, which is a component of ZP composed of four glycoproteins (ZP1, ZP2, ZP3, and ZP4). ZP2 expression in humans was observed in oocytes and granulosa cells as early as the primordial follicle stage (70). Earlier studies on human ZP glycoproteins around the oocyte revealed that unlike ZP1, ZP3, and ZP4, ZP2 mainly binds to acrosome-reacted sperm and the N-terminus of ZP2 mediates the taxon-specific sperm-oocyte binding (71).

The human ZP2 gene, located on chromosome 16, has 19 exons and encodes a polypeptide of 745 amino acids (aa) (72). Two homozygous truncating pathogenic variants of the ZP2 gene were identified using Sanger sequencing in infertile women with IVF fertilization failure (73). The two identified variants were missense and frameshift respectively. Based on the family pedigree analysis, an autosomal recessive genetic mode of infertility was suggested. It was observed that oocytes from affected women had an abnormal ZP with a thinner matrix, and an enlarged perivitelline space compared with normal oocytes. By using transmission electron micrographs and polscope images, there was only one thin layer in the zona matrix of patient oocytes with an irregular network of filaments with large holes (73). The loss-of-function variants of the ZP2 gene led to defective ZP in gamete recognition. In conventional IVF, none of the oocytes were fertilized.

TLE6 (Transducin-like enhancer of split 6) [MIM: 612399] is an effective gene that encodes the TLE6 protein, an essential member of the subcortical maternal complex (SCMC). The TLE6 gene is expressed only in ovaries as well as the oocytes and preimplantation embryos (37). By combining with a variety of maternal effector proteins, TLE6 protein forms SCMC has been linked to key processes occurring during the oocyte-to-embryo transition: cytoskeleton reorganization, meiotic spindle formation, and positioning, organelle redistribution, and cell division (37). Functional knockout studies of the TLE6 gene in mice showed the termination of embryo cleavage as the most phenotype (74). Several mutations within the TLE6 gene were shown to be associated with low-quality embryos (39, 75, 76). In 2015, a homozygous substitution (c.1529C > A) in TLE6 was verified by whole-exome sequencing in three infertile women (77). In 2020, a homozygous missense mutation (c.1226G>A (p.Arg409Gln) in exon 13 of TLE6 was found to be responsible for recurrent total fertilization failure in ICSI cycles (75). According to the in silico computational algorithms, this mutation is predicted to be probably damaging (75). A recessive inheritance pattern was suggested based on family pedigree analysis. Correspondingly, further studies are still required to reveal TLE6 function for the fertilization process.

The axoneme is the core structure of sperm flagellum and comprises an intricate network of protein complexes, where a central pair of microtubules is surrounded by nine peripheral microtubule doublets (MTDs) in the fixed order (“9 + 2” pattern). Axonemal dynein consist of the outer and inner dynein arms (ODAs and IDAs, respectively). Each dynein arm is a multi-protein ATPase complex, which is composed of light, intermediate, and heavy chain proteins. DNAH17 (Dynein axonemal heavy chain 17) [MIM: 610063], belonging to a member of the dynein axonemal heavy chains (DNAHs) family, encodes a heavy chain protein of ODAs. DNAH17 is mainly expressed in the testis and sperm. Previous reports had revealed that the biallelic mutations of DNAH17 in human sperm were associated with impaired motility and multiple morphological abnormalities of the flagella (MMAF), a rare type of asthenozoospermia which is characterized by absent, short, bent, coiled, and irregular flagella (38, 78–80). Noticeably, the knockout of DNAH17 in mice resulted in morphologically abnormal spermatozoa, showing a phenotype similar to a typical human MMAF phenotype (79).

ICSI is an effective technique used to treat infertility related to MMAF. However, two novel compound heterozygous mutations within the DNAH17 gene were identified using whole exome and Sanger sequencing in an infertile man with markedly diminished sperm motility and caused TFF after two ICSI attempts (81). The nonsense mutation results in a premature stop codon, which leads to a truncated and nonfunctional protein lacking all ATPase domains as well as a microtubule-binding region. The missense mutation was absent in population databases (EXAC, GNOMAD, 1000 Genomes Project) (81). Based on the family pedigree analysis, an autosomal recessive inheritance pattern was suggested. Briefly, mutations in DNAH17 are one of the sperm-related gene mutations associated with TFF.

Human oocyte activation is characterized by a two-step pattern, called the trigger and oscillator (40). During natural fertilization, the trigger originates from the receptor-mediated interaction between the sperm and the oocyte plasma membrane. During ICSI, the trigger is replaced by a calcium influx called a ‘pseudotrigger’ generated by the injection procedure (40). The actions of the oscillator are described by a series of shorter calcium transients of high amplitude, resulting from the release of sperm-associated factors into the oocyte cytoplasm. The main cause of fertilization failure after IVF/ICSI is considered to be oocyte activation deficiency (OAD), which can be categorized into two kinds: oocyte-borne and sperm-borne activation deficiencies.

Several studies have tried to overcome fertilization failure through a variety of assisted oocyte activation (AOA) methods, including physical, mechanical, or chemical stimuli (82, 83). The application of ICSI combined with AOA improves fertilization rates in the majority of patients with ICSI failure (82, 83). However, not all patients benefit from AOA (84). Many reports have failed to identify which patients are candidates for AOA (85). Some cases of fertilization failure are not solely related to sperm-related OAD and may not require AOA treatment. Therefore, it is important to determine whether AOA is really necessary.

To differentiate between sperm-related OAD and oocyte-related OAD, the mouse oocyte activation test (MOAT), a heterologous ICSI model, can be used as a diagnostic tool (86). The sperm from patients was injected into mature mouse oocytes. Oocyte activation is assessed by examining the percentage of 2-cell formation. When not activated, sperm deficiency is assumed; otherwise, it is suspected that the oocyte is deficient. According to the results of the MOAT, AOA methods can be proposed to correct clear sperm-related OAD (84, 86). Due to inevitable limitations in the heterologous ICSI model, such as multiple steps, long time, and lack of reliability, the PLCZ1 screening assay was recommended to evaluate an oocyte- or sperm-related OAD by a recent study (87).

High throughput sequencing techniques have helped discover some causative gene mutations in infertile couples related to male or female factors associated with poor or failed fertilization after IVF/ICSI attempts. In recent years, a series of mutations in PLCZ1, ACTL7A, ACTL9, DNAH17, WEE2, TUBB8, NLRP5, ZP2, and TLE6 genes have been identified as potential markers for evaluating fertilization failure in humans ( Table 2 ). Table 3 presents that the use of AOA greatly improves fertilization rates in human cases with sperm-related mutations. However, the AOA method cannot effectively rescue the phenotypes of fertilization failure in cases with oocyte-related mutations (21, 22, 26, 32).

The reasons for fertilization failure caused by oocyte-related defects are often complicated and not easily accepted. Little is known about how to treat oocyte-borne oocyte activation failures. Tesarik J et al. reported that sperm-borne and oocyte-borne oocyte activation failures could be overcome by the use of a modified ICSI technique (40). Heindryckx B et al. recommend AOA as an efficient treatment option in cases of both oocyte- or sperm-related fertilization failure (84). However, these studies mainly focused on case reports. In a recent study, which included 114 patients with a history of extremely poor or complete fertilization failure, the modified stimulation protocol improved clinical outcomes in patients with an oocyte-related OAD (87). A large-sample investigation is needed to support the results.