In a recent Research Article by Sutovsky et al., the authors interpret their results to mean that in the normal animal the epididymis ‘...provides a mechanism for sperm quality control...’ and that ‘ubiquitinated (i.e. defective) sperm are subsequently phagocytosed by the epididymal epithelial cells’ (Sutovsky et al., 2001). This was based on observations that abnormal sperm were ubiquitin-labelled and that the percentage of ubiquitin-labelled sperm decreased from 5% in the caput to 1% in the cauda epididymidis of two animals.
This hypothesis is an old one that was discredited over 30 years ago. For it to be valid, two important points need to be established: a loss of sperm in the epididymis and the removal of abnormal spermatozoa by its epithelium. First, the evidence from quantitative testicular histology and cannulation experiments in the normal animal (specifically in the species studied here, the bull) is that the number of sperm leaving the epididymis via the vas deferens is no lower than the number entering it from the testis (Lino et al., 1967; Amann et al., 1974; Amann and Lambiase, 1974), discounting loss during epididymal passage. In this paper, data from only two animals were provided and the potential counting errors of assessing the low percentage of anti-ubiquitin antibody-coated cells found (around 2 percentage points) (World Health Organization, 1999) and the small differences observed render the evidence for a loss of ubiquitinated sperm weak.
Second, phagocytosis of sperm by the epididymal epithelial cells does not feature in any of the classical histological and ultrastructural investigations (Nicander, 1957a; Nicander, 1957b; Nicander and Glover, 1973; Hamilton, 1975; Hamilton, 1972) and many others on normal epididymides of many species, including the bull. Given the large number of sperm present in the epididymis at any one time (several billion in the case of the bull, boar and ram), phagocytosis of even 0.1% would be readily detectable on tissue sections examined in the microscope, especially since sperm heads are difficult to digest intracellularly. Reports of sperm within epithelial cells in the normal male tract, mainly in the rete testis - efferent ductules and vas deferens-urethral junctions (Cooper and Hamilton, 1977a), amount to very small numbers indeed. The presence of 5-10% dead sperm in the epididymal lumen has been described on numerous occasions at both the light microscopic and ultastructural levels (Cooper and Hamilton, 1977b; NagDas et al., 2000) but no mention is made in these reports of phagocytosis of sperm by epididymal cells.
Thirdly, the apparent decrease of labelled cells observed by Sutovsky et al., does not necessarily imply epithelial spermiophagy. Although spermiophagy normally is not seen in the epididymis of all mammals thus far investigated, under certain conditions such as after obstruction or during advanced ageing, spermatozoa can be rarely seen in the cells of the epididymal epithelium. Equally plausible explanations for the observations reported are the well-documented phenomena of antigen epitope masking by secreted epididymal glycoproteins (Yeung et al., 2000) and the tendency for immature sperm to bind IgG non-specifically (Yeung et al., 1997).
We conclude that the paper by Sutovsky et al., presents unconvincing evidence for removal of defective spermatozoa by the normal epididymis and that the data certainly cannot be interpreted as a role for the epididymis in control of sperm quality.
We appreciate the opportunity to discuss our paper (Sutovsky et al., 2001a) and thank Cooper et al. for their respected opinion. Our work provides morphological and biochemical evidence that ubiquitin, a universal proteolytic marker that is also implicated in endocytosis, is found predominantly on the surface of defective spermatozoa during and after, but not prior to, their passage through the epididymis. Since the covalent ligation of ubiquitin serves as a selective, substrate-specific signal for proteolytic degradation (Laney and Hochstrasser, 1999), we suggested the term ‘epididymal sperm quality control’. The presence of ubiquitin in human sperm (Sutovsky et al., 2001b), seminal plasma (Lippert et al., 1993) and epididymal epithelium (Fraile et al., 1996; Santamaria et al., 1993) has been shown previously, and we performed an extensive set of control experiments to test the specificity of our anti-ubiquitin antibodies. Other proteins including clusterin (Ibrahim et al., 2000) and HEP64 (NagDas et al., 2000) are associated with the defective epididymal spermatozoa. It is somewhat difficult to explain the loss of ubiquitin-crossreactive spermatozoa by the epitope masking phenomenon (Yeung et al., 2000), which itself has not been explained satisfactorily. One can speculate that the transient loss of crossreactivity to certain antibodies could be caused by antigenic modification such as glycosylation or ubiquitination, the latter being a reversible ligation. Sperm phagocytosis was not the primary focus of our paper, but is a rather enticing explanation for the partial loss of ubiquitinated spermatozoa during epididymal passage, which we believe is beyond the 2% error margin. The title of our paper uses the word ‘putative’, and we hint to a ‘possible’ role of the epididymis in sperm quality control.
We would like to point out some pertinent, recent reports that address some of the issues in the continuing debate on epididymal sperm phagocytosis. Several studies document the loss of visibly defective spermatozoa during epididymal passage in normal, fertile animals (Arya and Vanha-Perttula, 1986; Axnér et al., 1999; Chenoweth et al., 2000; Goyal, 1982; Lopez Alvarez and Bustos Obregon, 1995; Ramamohana et al., 1980) which, in some cases, explain this sperm loss by epididymal phagocytosis (Ramamohana et al., 1980). Respected resources attribute the sperm-phagocytotic activity to epididymal epithelium (Berndtson, 1977; Barth and Oko, 1989). Ultrastructural and light microscopic evidence of sperm phagocytosis in the rete testis, efferent ducts and the epididymis of fertile animals was published in recent years (Arya and Vanha-Perttula, 1986; Goyal, 1982; Lopez Alvarez and Bustos Obregon, 1995). Solid evidence exists for protein turnover and endocytosis by the epididymal epithelium (Hermo et al., 1991; Hermo et al., 1992). Besides secretory proteins, epididymal epithelial cells (EEC) can internalize large particles such as sperm cytoplasmic droplets (Hermo et al., 1988; Temple-Smith, 1984). Temple-Smith asserts the existence of a specific ‘recognition factor on cytoplasmic droplet’ (Temple-Smith, 1984), and our studies show that ubiquitin is abundant therein (Sutovsky et al. 2001a) (see figure 4D). Although a valuable attempt, the cannulation studies should be interpreted with certain caution. Such surgical intervention could affect the delicate balance of sperm maturation and mobility throughout efferent ducts, and the catheters may remain patent for a very limited amount of time (Berndtson, 1977). Estimated daily sperm output ranges from 46% (rabbit) (Orgebin-Crist, 1968) to 92% (bull) (Amann et al., 1974) of the daily sperm production, calculated by histological or homogenization techniques. In 1977, Berndtson contended that the major disadvantage of such studies, in addition to altering reproductive organs, is that ‘spermatozoal resorption may occur in the epididymis’ (Berndtson, 1977).
Sperm phagocytosis is not readily detectable by light microscopy of bovine epididymal tissue sections processed with DNA stains and anti-ubiquitin antibodies, whereas spermatozoa intermingled with the apical microvilli/cilia of EEC are seen regularly and in large quantities (Sutovsky et al., 2001) (see figure 4G). It should also be acknowledged that such tissue sections, prepared by repeated dehydration/extraction with ethanol and xylene, may not reflect the natural conformation of the sperm mass within the epididymal lumen. Nevertheless, spermatozoa at various stages of disintegration were evident in the apical cortex of EEC, and in the epididymal lumen (Sutovsky et al., 2001) (see figure 5A-C) by electron microscopy in situ, and the immobilized spermatozoa were quickly ‘ingested and digested’ by cultured EEC in vitro (Sutovsky et al., 2001) (see figure 8B). Our findings are consistent with the internalization and turnover of epididymal proteins between the EEC cytoplasm and epididymal lumen (Hermo et al., 1992). In the case of ubiquitin, this turnover could be facilitated by the ubiquitin recycling enzyme, PGP 9.5, a ubiquitin-C-terminal hydrolase abundant in the epididymal epithelium (Fraile et al., 1996). The ubiquitinated spermatozoa may disintegrate while still in the epididymal lumen, and could be internalized by parts, as suggested by electron micrographs of sperm fragments in the epididymal epithelium in situ(Sutovsky et al., 2001) (see figure 5A-C). Perhaps the most accurate account of these events was provided by Barth and Oko, referring to the ‘resorption, phagocytosis and liquefaction’ of the epididymal spermatozoa (Barth and Oko, 1989). Sperm disposal by intraluminal dissolution (Barrat and Cohen, 1987; Flickinger, 1982) and intraluminal phagocytosis (Flickinger, 1982) following vasectomy has been proposed.
The occurrence of (sperm) cell surface ubiquitination is a surprising finding that, we believe, deserves detailed examination, and may have implications for infertility diagnostics and treatment. We acknowledge that the fate of the ubiquitinated epididymal spermatozoa can be interpreted in different ways and welcome further discussion.