The first phase of sperm production is called spermatogenesis. In both rodents and humans, it starts with a mitotic division of diploid spermatogonial stem cells at the basal lamina of the seminiferous tubule in the testis [1]. The pair of new daughter cells either remain as part of the stem cell pool or migrate away from the basal compartment as chained pairs to follow the differentiation path [2]. The differentiating spermatogonia then enter the meiotic cycle and become the primary spermatocytes. After two rounds of meiotic division I/II without DNA replication, the germ cells are now haploid, denouncing the end of the meiotic cycle and becoming round spermatids, and then proceed to post-meiotic differentiation known as spermiogenesis [3].

While mitotic and meiotic germ cells undergo sophisticated networks of cellular processes that involve the reprogramming of DNA methylation, the histone marks and sncRNA, the epigenome and transcriptome of testicular spermatozoa are mainly established during the spermiogenesis that takes place at the adluminal compartment of the seminiferous tubules. During this process, a wave of transcriptional activity is observed in the round spermatids [4] during its extensive metamorphosis to transform into elongated spermatids [5]. It is believed that the transcriptional activity plays a pivotal role in establishing the repertoire of RNA species in the testicular spermatozoa expressed by the germ cell per se. The paternal genome is then reorganized through the process of histone-to-protamine transition, followed by nuclear condensation, which gives rise to the unique and compact nucleus seen in the spermatozoa. In parallel to the histone-to-protamine transition, the testicular spermatozoa further become transcriptionally inactive, and the bulk of their cytoplasmic content, containing typical cellular organelles including ribosomes, is shed as the residual body [6]. Within the condensing nucleus, the hyperacetylation and phosphorylation of histones result in the relaxation of DNA strands around the nucleosomes and thereby the chromatin structure as well. Chromatin remodeling occurs as nuclear histones are temporarily replaced with transition proteins, followed by subsequent protamine transition to further condense the paternal genetic material into a unique chromatin structure significantly smaller than the somatic cells at the interface [7]. The aberrant disruption of histone-patterning in sperm causes not only fertility problems but also defects in the development of the offspring [8,9]. Depending on the species, there is an estimate of 1% of histone in the sperm chromatin left in mice that is not replaced with protamine, and the estimate is roughly 15% in humans [10]. Evidence from chromatin immunoprecipitation (ChIP)-sequencing showed that the remaining histones are retained in the promoter region of genes linked to fertility and embryo development [11,12]. This is likely to facilitate embryo development in a timely manner, as studies found that, unlike paternal protamines, the paternal histones do not get replaced with oocyte-derived histones upon fertilization [13,14,15].

As protamine-bound genes remain inaccessible, the histone-code generated at the end of spermiogenesis in turn serves as an epigenetic instruction that can impact the offspring health (intergenerational) as well as subsequent generations (transgenerational) [16,17]. Up until recently, it was believed that the sperm chromatin compaction was fully established once it exited the testicular environment, as genome-wide methylation remained stable between the testicular spermatozoa and the cauda epididymal sperm [18]. New studies revealed continuous differential histone retention sites between the caput sperm and cauda sperm of the epididymis [19], suggesting that the alteration of histone marks acquired during the epididymis transit may play a role in the early embryo development and predisposition to offspring health.

Aside from histone marks, sperm also carries a distinct epigenetic profile made up of RNA species. The general consensus had viewed the sperm that has a condensed nucleus and sheds its cytoplasmic residue as transcriptionally inert. The insufficient ribosomal RNA species in sperm also suggested that they are devoid of translational activity [20]. Over the years, it was assumed that the sperm RNAs were mere remnants of spermatogenesis that were not fully discarded along with the cytoplasmic residue. Even so, after the first RNA was identified in the sperm of humans, mice and rats [21], numerous reports further identified RNA species in the male gamete, both coding [22,23,24] and noncoding [25,26]. Sperm-RNA transcripts that are normally not present in the oocyte can also be identified upon fertilization [27], suggesting contributions of the sperm RNA to the embryo, and subsequent studies have also reported the pivotal roles of sperm-RNA in embryonic development [23,28,29,30,31,32,33].

The small RNA (sRNA) profiling of mice revealed an extensive makeover of sRNA species between testicular sperm, caput and cauda sperm [34], where Piwi-interacting RNAs (piRNAs) account for about 85% of sRNA species in the testicular spermatozoa, followed by 10% of microRNA (miRNA). Upon entering the epididymis, transfer RNA-derived small RNAs (tsRNAs) become the dominant sRNA, with 75% of total sRNA species in the caput sperm and more than 85% in the cauda sperm. The level of miRNAs remains fairly constant during the epididymal transition, which represents 5% of the total sRNA, making it the second most abundant sRNA species in cauda sperm. The profiling of miRNAs from caput, corpus and caudal sperm reported a loss of 113 miRNAs, as well as a gain of 115 miRNAs, including a significant increase in miRNAs natively expressed in immature sperm [35].

Sperm are transcriptionally inactive after exiting the testis, despite RNA polymerase having been found in epididymal sperm in both bulls and mice [36,37]. Currently, there is no evidence of endogenous activity of sRNA transcription occurrence in the testicular spermatozoa to cauda sperm, which could explain the dynamic change in the sperm sRNA profile, nor is there evidence of the endogenous cleaving of intact transfer RNA (tRNA) in testicular spermatozoa, which could explain the abundance of tsRNA starting in the caput sperm. This raises suspicions that the abundance of tsRNA could be trafficked to the maturing sperm through extrinsic factors during their transit in the epididymis.

Early studies using the electron microscopy of sperm from Chinese hamsters showed small membranous vesicles over the acrosome of caudal sperm [38]. High-resolution microscopy, in turn, showed an interaction between extracellular vesicles (EVs) and the cytoplasmic droplets on the caudal sperm [39]. These EVs, later referred to as epididymosomes, are characterized, and their interaction with epididymal sperm had been established by Sullivan et al. through co-incubation and surface-protein labeling [40]. The binding of epididymosomes results in the incorporation of several proteins found in epididymosomes to the sperm membrane. This could either be because the epididymosomes do not fuse completely with the sperm to allow for the transfer to the whole protein repertoire or because only a subpopulation of EVs are delivered to the sperm, as the epididymosomes is a heterogeneous pool of vesicles.

In turn, using in vivo metabolic labeling of RNAs, Sharma et al. showed that RNAs that specifically synthesized the caput epididymis were taken up by the caudal sperm [34]. The co-incubation of testicular spermatozoa with caput epididymosomes also results in a significant increase in the abundance of tsRNAs and miRNAs. tRNA fragments (tRFs) corresponding to the 5′ end of tRNA Glu-CTC and Val-CTC, in particular, were increased 4–10 fold in testicular sperm co-incubated with caput epididymosomes when compared to the mock-treated control. This confirmed the ability of sperm to receive not only the protein but also the RNA cargo of the epididymosomes, thereby modifying both the sperm proteomic and RNA profiles.