3.1.1 Quantitative defects

According to the diagnostic criteria of the World Health Organization, the total sperm count in the ejaculate can be categorized into normal count (> 39 million sperm cells), reduced sperm count (oligozoospermia < 39 million sperm cells), or no spermatozoa in the ejaculate. In case the lack of spermatozoa is due to a molecular cause originating from the testis, (i.e., primary spermatogenic failure), it is defined as non-obstructive azoospermia (NOA).⁵ Reduced sperm count affects around 5% of men worldwide, whereas 1% of men present NOA. “Oligozoospermia” is a broad term, ranging from a few spermatozoa up to 39 million, and therefore the following subcategories have recently been proposed based on rounded 25% and 5% percentile thresholds: i) extreme oligozoospermia (from > 0 to < 2 million; including the diagnostic term “cryptozoospermia”); ii) severe oligozoospermia (from > 2 to < 10 million); iii) moderate oligozoospermia (from > 10 to < 29 million); iv) mild oligozoospermia (from > 30 to < 39 million sperm)⁶ (Figure 1). Similarly, NOA can be heterogeneous based on the histopathological findings of testes. NOA patients can present an absence of germ cells in the testis (sertoli cell only [SCO]), maturation arrest (MA) of germ cells at different cell developmental stages (Spermatogonia, spermatocyte or spermatid), or hypospermatogenesis.

The genetic causes of quantitative defects are highly heterogeneous with hundreds of genes reported to be mutated.^{3, 7-9} With the exception of a small number of genes, the mutated genes typically follow the autosomal recessive or X/Y linked inheritance mode.^{3, 10, 11} It has been suggested that for NOA, as the most severe phenotype of spermatogenic defects, the genetic component may play a larger role in the manifestation of the disease as compared to oligozoospermia.¹² However, clear distinctions between the genetic causes of NOA and oligozoospermia, particularly extreme oligozoospermia, are challenging to make, as the same mutation or mutations in the same gene have been found in both phenotypes.^13-18

3.1.2 Qualitative defects

An excess of various qualitative defects of spermatozoa (≥4% of sperm cells in the ejaculate) due to morphological abnormalities that can affect the sperm head, neck, midpiece, or endpiece is defined as teratozoospermia,⁵ which includes acephalic spermatozoa syndrome (headless flagella; ASS), macrozoospermia (big-headed spermatozoa), globozoospermia (round-headed spermatozoa without acrosome) and MMAF among others. The latter is characterized by a mix of spermatozoa with short, absent, irregularly shaped, or coiled flagella that severely impair sperm motility. The teratozoospermia disorders are all considered fairly rare (< 0.01% of the male population) and follow an autosomal recessive inheritance pattern, although cases of X-linked MMAF have been identified via WES studies as well.

In cases where teratozoospermia occurs uniformly across all spermatozoa, affected by the same type of defect, that is, monomorphic or pure forms, genetic causes are more likely to be identified with respect to mild forms in which only a fraction of sperm cells are affected. This suggests a systemic genetic factor disrupting spermiogenesis as opposed to environmental stress and/or epigenetic alterations. The genetic landscape of teratozoospermia is heterogeneous and often specific to the phenotype. For instance, mutations in the AURKC gene are identified in most cases of pure sperm macrocephaly, while pure forms of globozoospermia are commonly associated with, but not limited to, mutations in the DPY19L2 gene.¹⁹ Similarly, ASS is a rare type of teratozoospermia with 30%–50% of the cases attributed to mutations in Sad1 and UNC84 domain containing 5 (SUN5) gene.^{20, 21} MMAF is a relatively new category in which the number of mutated genes, often belonging to the Dynein Axonemal Heavy Chain (DNAH) and Cilia and Flagella associated protein (CFAP) gene families, have significantly increased in recent years.³ This phenotype is mostly considered to be independent of the other types of asthenoteratozoospermia, such as PCD-related defects, although the phenotypic and genetic distinction between the two may remain complicated at times.^{22, 23}

3.3.1 Quantitative defects of primary spermatogenic disorders

According to the current literature on the genetics of spermatogenic impairment, the etiology of most of the studied infertility traits is characterized by a monogenic inheritance pattern, whereby typically a single recessive or X-linked gene is implicated in each affected case or a family (reviewed in Kasak et al.⁸ and Vockel et al.⁷). The advantage of applying WES to cohorts of infertile men is the ability to screen all the coding variation in the genome at once ideally pinpointing a disease-causing gene in each case and providing a collection of disease-linked genes per phenotype. Nevertheless, even for large patient cohorts the fraction of cases with a detectable genetic lesion may remain low in some traits.

This phenomenon applies to the phenotypes of reduced sperm count, whereby, in the majority of reports, a disease-linked variation is typically identified in less than 25% of studied cases independent of the cohort size (range n = 8–924 cases) (Table 1 and Figure 2A). Interestingly, comparable detection rates are reported for both recessive or X-linked (average across cohorts, 21.7%) and de novo variation (average across phenotypes 16.9%) (Table 1). Oud et al.⁴⁵ performed WES in a sample of 185 men with unexplained azoospermia or oligozoospermia and reported heterozygous or hemizygous de novo mutations in 29 possibly causative genes. Moreover, the authors identified significant enrichment of loss-of-function (LoF) and missense de novo mutations in genes that are intolerant to and functionally impacted by heterozygous deleterious variation in patients as compared to fertile controls. This is one of the rare examples suggesting the contribution of dominant variation in male infertility.⁴⁵

It has been described for various human disease phenotypes that addressing homogeneous subtypes of a disease may be of benefit and yield improved detection rates in genome studies.⁴⁷ In the context of NOA, patients can be subclassified according to their testicular histopathology as having SCO, MA, or hypospermatogenesis. Although the exact prevalence of each testicular phenotype in idiopathic NOA is not known, the majority of patients who undergo TESE present the testicular phenotype of SCO or hypospermatogenesis, whereas pure forms of idiopathic complete MA remain less common.⁴⁸ Interestingly, the highest detection rates observed across WES studies of quantitative spermatogenic failure have been reported for men who present MA (Table 1 and Figure 2A). According to Krausz et al., as much as 82% (n = 14/17) of MA cases were attributed to a recessive genetic variation.¹⁸ The successful assignment of a potential genetic cause was achieved by a stringent prioritization of MA cases that additionally had consanguineous parents and/or family history of male or female infertility. Comparable success was reported by Ghieh et al., whereby recessive deleterious variants were identified in n = 15/26 (58%) men with MA at the spermatocyte stage, with a 100% detection rate observed among consanguineous (n = 7/7) and 42.1% (n = 8/19) among non-consanguineous MA cases.⁴⁹ As the consequence of consanguinity, a recessive variation that is typically very rare or non-existent in a homozygous state in the general population is found in the background of large stretches of homozygosity in the genome and has commonly been reported in male infertility studies.^{8, 9} Similar tendencies were seen in the general NOA cohorts even without stratification by the testicular phenotype. In a WES study of 924 men with NOA, an improved detection rate of 76% was reported among 72 unrelated consanguineous cases as opposed to the 15% observed among the remainder of the men without any consanguineous background.³⁸ In the same line, Wu et al. demonstrated based on a collection of 12 consanguineous families affected by NOA and premature ovarian insufficiency (POI) that WES analysis is able to identify recessive deleterious variation in known and novel genes linked to the disease in 58.3% of the families (n = 5/9 unrelated NOA men) (Table 1 and Figure 2A).⁵⁰ An overlap in the genetic etiology between NOA and POI has previously been observed and suggests that a family history of POI may also be an indication of increased chances of heritable infertility among male members of the family, particularly in the presence of consanguinity.^{18, 51}

3.3.2 Qualitative defects of primary spermatogenic disorders

Unlike quantitative sperm defects, the majority of the studies on the morphological abnormalities of the sperm report a disease-linked variation in over 25% of the cases. The phenotype of MMAF has benefited the most from WES analysis in recent years with four comprehensive studies included in this review (Table 1). According to these studies, a potential disease-linked genetic variation is found in over half of the cases, with up to 76.7% (n = 23/30) in Tang et al. (range across four studies 47.8%–76.7%) (Table 1 and Figure 2A).⁵² Numerous single-gene reports have been published based on the WES data from the two largest collections of MMAF cases from North Africa/Middle East/Europe (total up to 167 cases) and China (n = 90), however a comparable overall detection rate of 51.5% and 60.0% has been mentioned for their cohorts, respectively.^{24, 25} These results are largely powered by the consanguineous background of the cases based on these single-gene reports (see references within Martinez et al.²⁴), as also observed in studies of men diagnosed with NOA (Figure 2A). A single comprehensive WES study by Oud et al. of patients with globozoospermia without mutations in DPY19L2 or spermatogenesis associated 16 (SPATA16), was similarly able to detect a candidate disease variant in 8/15 (53.3%) cases with 100% detection rate among seven men from consanguineous families and 12.5% (n = 1/8) among the remaining non-consanguineous cases.⁵³

As acephalic spermatozoa syndrome is a rare teratozoospermia phenotype (< 0.01%), the case cohorts have remained fairly small (n = 11, 12) and are able to ascertain a recessive disease-linked variation to 27.3%–58.3% of the cases (Table 1).^{54, 55} Li et al. have analyzed a fairly large group of patients (n = 82) categorized as “unexplained abnormal sperm head phenotypes”, which includes abnormal spermatozoa with globozoospermia, small or deformed acrosome, pyriform-headed, amorphous-headed, small-headed and tapered-headed (Figure 1). The authors report a low detection rate of 14.6% (n = 12/82) based on the affected genes directly related to the phenotype, likely due to a short list of known genes linked to the underlying etiology.⁵⁶ A varying success in the identification of a potential disease gene is evident in studies on asthenozoospermia, which features a severe reduction in sperm motility and may include a mosaic of cases that present MMAF and PCD-like symptoms, such as chronic sinus congestion or chronic respiratory disease (Table 1).^{57, 58}

3.4.1 Quantitative defects of primary spermatogenic disorders

The genetic causes of reduced sperm count form a complex list with each case typically presenting a unique disease-linked variation. For NOA and reduced sperm count combined, 443 out of 1995 cases carry variants in a total of 351 unique genes, of which the majority represent singleton observations (Figure 2B and Table S2). Among the genes affected in more than two carriers in the studies included in this review (AR, CCDC89, CEP250, DCAF12L1, DMRT1, FBXO15, KASH5, KLHL10, MAGEE2, MEI1, PNLDC1, SHOC1, SYCE1, TDRD9, TERB1, TEX11, TEX14, TTLL9, and ZNF711), most have been associated with NOA due to MA at the meiotic stage of spermatogenesis. The gene of most observations is X-linked testis expressed 11 (TEX11), a crucial factor in chromosomal synapsis, for which partial gene deletions were first linked to azoospermia due to meiotic arrest at pachytene stage by Yatsenko et al.⁵⁹ Subsequently, multiple patients carrying variants in this gene have been reported, with seven of them included in this review (Figure 2B).^{18, 37, 39, 49, 60} Interestingly, 46/1995 patients with reduced sperm counts (n = 10 studies) carry variants in more than one gene potentially pointing to a di/oligogenic disease model. However, further functional and facility studies are warranted to validate this scenario. It is noteworthy that 62 patients (16%) carry variants in X-linked variants highlighting the relevance of genes located in this chromosome in the etiology of quantitative disturbances of spermatogenesis. This is in line with a targeted WES study of X-linked genes in 2354 patients with non-obstructive azoospermia or extreme oligozoospermia, where authors reported deleterious variants in 11.9% of cases, with RB binding protein 7, chromatin remodeling factor (RBBP7), which is important in chromatin regulation and metabolism, being the most frequently affected gene observed in 10 NOA men.³⁴ This study underscored the significant role of the X chromosome in quantitative defects.³⁴ Of note, although the findings reported here largely mirror what is known based on published literature, the spectrum of genes and the number of patients carrying variants in the same gene are likely underestimated, as many single-gene centered and family studies that have reported recurrent observations of an affected gene have been excluded from this review.

3.4.2 Qualitative defects of primary spermatogenic disorders

Contrary to the quantitative defects of spermatogenesis, the morphological abnormalities of the sperm predominantly feature a small number of genes responsible for a large fraction of the studied cases (Figure 2B and Table S2). On average, the genes are typically observed among cases recurrently resulting in a higher detection rate indicating that disruptions of a limited number of pathways lead to the morphological abnormalities of the spermatozoa (Figure 2). As the most striking example, DNAH1 alone is responsible for 37.3% (n = 62/166) of MMAF cases across four WES studies, followed by 12.7% of cases with deleterious variants in three CFAP family genes CFAP43, 44 and 65 (n = 10, 7 and 4 carriers, respectively). Members of both DNAH and CFAP families are essential for the axonemal organization of cilia. These findings are in line with the known genetic etiology of MMAF, whereby 28%−44% of MMAF cases are attributed to DNAH1.⁵² As an added value of WES, 11 novel case-specific genes, of which six belong to the DNAH family, have been implicated in MMAF in these studies. As an extreme manifestation of morphological sperm abnormalities, acephalic spermatozoa syndrome represents a disorder in which the majority of sperm cells in the ejaculate are headless due to defects of the head-tail coupling apparatus. Although SUN5 has been proposed as the main gene responsible for the phenotype (about 40% of the cases), only 1/10 cases carry SUN5 mutations across two WES studies incorporated in this review, compared to 6/10 cases with disruptions in polyamine modulated factor 1 binding protein 1 (PMFBP1) (Figure 2B and Table S2).^{21, 54, 55} For globozoospermia, recessive deletions in DPY19L2 describe most of the cases with pure forms of the disease, as extensively demonstrated previously.⁶¹ Nevertheless, WES could potentially provide a diagnosis in half of the cases without the DPY19L2 deletions, as reported for a cohort of 15 men (n = 7/15 with consanguinity) by Oud et al. (Table 1 and Figure 2B).⁵³ Larger genetic complexity is observed for asthenozoospermia but also for the collection of cases with unexplained abnormal sperm head anomalies, partially due to the phenotypic heterogeneity of the cohorts.

Of note, 10 genes (CFAP44, DNAH1, DNAH5, DNAH6, FSIP2, MAGEA3, MDC1, PACRG, QRICH2, and TDRD9) reported in this review have been linked to both qualitative and quantitative spermatogenic defects. Interestingly, mRNA expression of all of these genes has been detected not only in late spermatids but also in early spermatogenic cells.⁶² In fact, a significant proportion of patients with qualitative defects also show reduced sperm counts, a semen phenotype called oligoasthenoteratozoospermia (OAT), indicating a detectable overlap between the genetic etiologies of the two phenotype categories of primary spermatogenic disorders.

3.4.3 Comparative view of biological pathways

The observed differences in the genetic architecture of the quantitative and qualitative abnormalities of sperm point to the varying levels of heterogeneity in the biological pathways that are essential for the accurate completion of spermatogenesis. This is clearly highlighted when comparing NOA and MMAF, as the most studied disease categories in quantitative and qualitative sperm defects, respectively (Figure 3). The genes that were identified among NOA cases by the WES platform (n = 324) affect a wide array of biological processes most of which are known to be relevant for germ cell survival. These include, for example, pathways regulating meiotic division (e.g., synaptonemal complex assembly or meiotic sister chromatid segregation) and metabolism of piRNAs, disruption of which leads to the MA of spermatogenesis.⁶³ On the other hand, only pathways related to cilium assembly dominate in MMAF, which is characterized by a high detection rate and a short list of recurrently affected genes (total, 16 genes; Figure 3). Although the progression of spermatogenesis partly differs in men and mice,⁶⁴ the comparative view of azoospermia and MMAF in mice largely mirrors the differences in genetic complexity observed in men (Figures 2B and 3). These findings indicate that the larger list of genes linked to NOA, as opposed to MMAF (n = 324 and n = 16, respectively), across the WES studies in men is not purely due to differences in studied sample sizes (n = 1759 and n = 166 cases total, respectively).