Ectopic Expression of Human Herpesvirus 1 Thymidine Kinase Induces Male Infertility

The herpesvirus family comprises several widespread infectious pathogens. They infect a variety of animal hosts, including humans and cause complex clinical outcomes. Recently, the possible correlation between genital infection by human herpesviruses (HHVs) and male infertility has attracted considerable attention. In this chaper, we investigated the mechanism of HHV‐1‐induced infertility in transgenic (Tg) rats and its possible correlation with infertility in human males. Ectopic expression of HHV‐1 thymidine kinase (TK) in the testis of Tg rats increased male infertility. In addition, truncated TK proteins were found in postmeiotic spermatids of Tg rat testis, leading to progressive degeneration of germ cells and vacuolization of the seminiferous epitheli‐ um. These findings suggest the possibility that a similar process occurs within HHV‐ infected human germ cells.


Introduction
Herpesviruses belong to a family of double-stranded DNA (dsDNA) viruses commonly causing herpes in animals. They present a unique four-layered structure. The outermost layer corresponds to the envelope, a lipid bilayer membrane interspersed with glycoproteins. This encases the tegument, a protein coat that surrounds the icosahedral nucleocapsid containing the viral linear DNA genome (Figure 1) [1]. At present, over 130 herpesvirus species have been identified, eight of which are known as human herpesviruses (HHVs): HHV-1 and HHV-2 (commonly called herpes simplex virus-1 and -2; HSV-1 and HSV-2), HHV-3 (varicella-zoster virus; VZV), HHV-4 (Epstein-Barr virus; EBV), HHV-5 (cytomegalovirus; CMV), HHV-6 (human herpesvirus 6, including HHV-6A and HHV-6B), HHV-7 (human herpesvirus 7), and HHV-8 (Kaposi's sarcoma-associated herpesvirus; KSHV). HHVs are divided into three subfamilies (α-, β-and γ-herpesviruses) based on their unique properties ( Table 1) [2]. HHVs are widespread among humans to the extent that more than 90% of adults are thought to be infected with at least one variety [2]. HHVs generally infect ectoderm-derived tissues, such as skin, mucoepithelial tissue and nervous tissue; however, they usually show preference for specific target cells/tissues and ensuing clinical outcomes ( Table 1) [2]. In addition, many people carrying the virus may be asymptomatic due to the latency of associated transcripts, which help HHVs evade the host's immune response [3]. Moreover, HHVs hide in specific cells during latent infection periods ( Table 1). Figure 1. HHVs structure. HHVs have a unique four-layered structure: a core of double-stranded DNA is surrounded by the nucleocapsid, which is enclosed by the envelope. The envelope is attached to the capsid through the tegument. Glycoproteins cover the surface of the envelope.
During the last decade, numerous studies have demonstrated that genital infections are often associated with human infertility, especially in males [4]. Most recently, genital infections caused by herpesviruses have attracted considerable attention [5][6][7]. Although herpes genitalis can be mainly caused by HHV-1 and HHV-2 [8], other HHVs have also been frequently detected in the genital organs of infertile male patients [9][10][11]. Although not fully demonstrated, there is a possible correlation between human male infertility and HHVs.
Our laboratory recently developed a line of transgenic (Tg) rats expressing a chimeric gene with the promoter of porcine follicle-stimulating hormone β subunit (FSHβ) and the open reading frame of HHV-1 thymidine kinase (TK). Unexpectedly, we observed that Tg rats showed infertility upon ectopic expression of HHV-1-TK in the testis, independent of the FSHβ promoter [12]. Indeed, we found that the accumulation of HHV-1-TK protein in postmeiotic spermatids was the cause of male infertility [13]. In this study, we used the Tg rat model to elucidate the HHV-mediated mechanism responsible for male infertility. 3. γ-herpesvirus: replication in lymphoblastoid cells, lytic cycle in some fibroblasts and epithelial cells.

HHVs infection and male human infertility
Human infertility is a widespread problem that has been increasing in recent decades. It affects 20-30% of couples in the world [14], and 40-50% is attributed to male infertility [15,16]. Many risk factors can disrupt male reproductive capacity and cause infertility. Among these are infections with pathogens [17], such as viruses [7], mycoplasma [18], chlamydia [19], and bacteria [20].
Although many studies have revealed a high prevalence of HHVs in male infertility patients, it is still difficult to conclude that HHVs infection is the causative agent. To address this issue, the capacity of HHVs to infect the testis and interfere with spermatogenesis should be proven regardless of the presence or absence of HHVs in the semen and spermatozoa. This would require direct evidence of a molecular mechanism for HHV-induced infertility. In this respect, studies in Tg animal models offer the opportunity to better understand human male infertility.

HHV-1-TK as a reporter system
TK (EC 2.7.1.21) is a key enzyme in the pyrimidine salvage pathway that catalyzes the transfer of the ATP γ-phosphate to thymidine to produce dTMP. HHV-1-TK is a phosphotransferase specifically required for viral DNA synthesis. HHV-1-TK shows broad substrate specificity, including pyrimidines (thymidine, deoxycytidine) and their analogs (azidothymidine), as well as purines (guanosine) and their analogs (acyclovir, ganciclovir [GCV], buciclovir and penciclovir). The ability to transform synthetic precursors, such as GCV, into toxic nucleotide analogs has resulted in effective cancer therapy agents [32,33]. GCV consists of a guanine linked to an acyclic sugar moiety at position 3'. HHV-1-TK converts GCV into GCV-monophosphate, which is further phosphorylated into GCV-triphosphate by the host's kinases. GCV-triphosphate has high affinity for DNA polymerase and functions as a competitive inhibitor of guanosine triphosphate (dGTP), becoming incorporated into the nascent DNA strand (Figure 2). GCV causes a distortion in the DNA sugar phosphate backbone, which blocks DNA replication [34][35][36] and induces cell apoptosis, thus selectively eliminating HHV-1-TK-positive cells [37,38]. Moreover, HHV-1-TK-expressing cells trigger apoptosis of neighboring cells due to the transfer of GCV-triphosphate via gap junctions in a phenomenon called "bystander killing" [39]. Therefore, HHV-1-TK can be used as a marker/reporter gene for removing specific target cells. Figure 2. Incorporation of GCV disrupts extension of replicating DNA strands. In HHV-1-TK-positive cells, GCV is phosphorylated into GCV-monophosphate and then converted into GCV-triphosphate. GCV-triphosphate has high affinity for DNA polymerase and is utilized as a dGTP analog. However, the acyclic sugar moiety distorts the sugar phosphate backbone, resulting in irregular extension of the replicating DNA strand.

Ectopic expression of HHV-1-TK
The use of HHV-1-TK offers a valuable tool for the ablation of specific cell types as well as in gene therapy. In the early 1990s, Tg rats were engineered to express HHV-1-TK under the control of the FSHβ promoter ( Figure 3A) in the anterior lobe of the pituitary gland [12]. However, the experiment was interrupted due to observed infertility of male rats. The same happened with Tg mice [40]. Moreover, Ellison and Bishop showed that ectopic expression of HHV-1-TK fused to the human immunodeficiency virus long terminal repeat (HIV-1-LTR) gene caused male infertility in Tg mice [41]. In the present work, we observed ectopic expression of HHV-1-TK gene in the testis of Tg rats ( Figure 3B). . Arrows indicate the region amplified using specific primer sets (FSH-tk-I and FSH-tk-II); product sizes are shown in parentheses. (B) Total RNAs of pituitary, gonads (testis and ovary), cerebellum, liver, kidney, adrenal glands, prostate, and uterus of Tg rats (male and female) were analyzed by real-time polymerase chain reaction (RT-PCR) using specific primers for rat FSHβ subunit, HHV-1-TK (I), HHV-1-TK (II) and cyclophilin A (control). Reproduced and modified from [12] with permission from the Society for Reproduction and Development.
Whereas the testes of 3-month-old normal rats exhibited morphologically normal germ cells through all stages of maturation, histological analysis of Tg rats revealed developmentally and Tg (C and D) 3-month-old rats. Arrowhead, elongated arrested round spermatid; arrow, multinucleated cell. (E and F) Spermatozoa obtained from the epididymis of normal (E) and Tg (F) rats stained with SYBR-14 and propidium iodide from a LIVE/DEAD sperm viability kit. Green and red indicate live and dead spermatozoa, respectively. Reproduced and modified from [12] with permission from the Society for Reproduction and Development. arrested spermatozoa and multinuclear cells, together with altered elongated spermatozoa (Figure 4) [12]. Loss of sperm motility and viability was also observed ( Figure 4F). Compared with normal rats, testis and epididymis weights were also decreased by 35 and 57%, respectively. In contrast, prostate and seminal vesicles weights were similar [12].

Disruption of spermatogenesis
Histological analysis revealed that in Tg rats abnormal spermatogenesis could be observed as early as 3 months in development. Although spermatocytogenesis and meiosis seemed to progress normally (Figure 5B), the number of spermatozoa in the epididymis decreased dramatically (Figure 5F). Furthermore, large numbers of degenerated germ cells  and immature/malformed spermatozoa were present (Figure 5F), indicating that maturation was likely to be disrupted. Degenerated germ cells with large vacuoles and cells lost throughout the tubules were observed in the testis of 6-month-old Tg rats ( Figure 5C). In 12-month-old Tg rats there was a complete loss of germ cells and only Sertoli cells remained in the tubules (Figure 5D). In the epididymis of 6-and 12-month-old Tg rats, spermatozoa could hardly be observed (Figure 5G-H). Thus, Tg rats showed a spermatidstage-specific defect in maturation and an age-dependent loss of germ cells.

HHV-1-TK accumulation increases the number of apoptotic germ cells
Whereas spermatocytogenesis and meiosis were confirmed in Tg rats, degeneration of germ cells by necrosis and apoptosis was frequently observed in the seminiferous tubules. In addition, terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay showed a stage-independent increase in TUNEL-positive tubules and cells (Figure 6). Reproduced and modified from [13] with permission from the European Teratology Society.

HHV-1-TK accumulation in postmeiotic spermatids
Tg rat testes (3-and 6-month-old) were simultaneously stained with anti-HHV-1-TK and nuclear DAPI stain (Figure 7). We observed that HHV-1-TK protein accumulated in the cytoplasm of postmeiotic spermatids while it was absent from spermatogonia, spermatocytes, Leydig cells and Sertoli cells. Furthermore, germ cells were obviously reduced in 6-month-old Tg rat testes (Figure 7B-C). Figure 7. Immunohistochemistry of Tg rat testis. Immunostaining of 3-month-old Tg rat testis with anti-HHV-1-TK antibody and nuclear DAPI stain. HHV-1-TK localizes to the cytoplasm of round spermatids, but not to other testis cells.

Mechanism of ectopic expression of HHV-1-TK
Ectopic expression of HHV-1-TK protein was confirmed in postmeiotic spermatids by immunohistochemistry. This suggested the existence of alternative promoters directing specific and ectopic expression in postmeiotic spermatids. Western blot analysis of HHV-1-TK in Tg rat testis revealed for the first time two bands at 37 and 39 kDa, corresponding to truncated products of the full-size 43 kDa HHV-1-TK protein ( Figure 8A) [13].
Analysis of the transcription start site by RNA ligase-mediated rapid amplification of cDNA ends (5'-RLM RACE) ( Figure 8B) showed that the start site corresponded to the first intron of the HHV-1-TK gene ( Figure 8C). This suggested that the second in frame ATG might have produced the 37 kDa band. Ellison and Bishop reported that HIV-1-LTR-driven HHV-1-TK ectopic expression was abolished by deleting the region between the multiple cloning site and the second ATG [41]. They suggested that removing this portion conferred DNA methylationindependent expression in the testis. Indeed, the corresponding region (144 bases) is GC-rich (67%) and contains 17 CpG sites. Currently, a HHV-1-TK vector free of the GC-rich region is available. Ectopic, testis-specific, expression of HHV-1-TK in Tg rats and mice raises the possibility that the same mechanism could also affect humans carrying HHV-1. Therefore, the transcriptional regulation exhibited by ectopic expression of HHV-1-TK in Tg rats and the effects of HHV-1-TK accumulation on spermatogenesis may be revelatory of the same mechanisms in humans.

Ultrastructural abnormalities of spermatozoa in Tg rat testis
Scanning electron microscopy revealed abnormal ultrastructure of epididymal spermatozoa in 10-week-old Tg rats [13]. We observed that the head and tail regions of spermatozoa from 10-week-old normal rats exhibited smooth surface and a regular morphological conformation ( Figure 9A). Indeed, the morphology of these spermatozoa was different from those of Tg rats (Figure 9B-F). Spermatozoa with malformed heads, the consequence of a defective acrosome or reduced genome, were observed at high frequency among Tg rats (Figure 9B-F). Transmission electron microscopy (TEM) revealed that the cell membrane was missing from segments of the head (Figure 9C), midpiece (Figure 9D), and flagellum ( Figure 9F). There were spermatozoa with looped tails (Figure 9D). TEM of 3-month-old rat testis showed that in contrast with normal rats (Figure 10A-C), Tg rats displayed massive vacuoles within the seminiferous epithelium ( Figure 10D-E, asterisks). These frequently presented degenerated spermatocytes in the seminiferous tubules (Figure 10E-F, arrowheads), in spite of confirmed spermatocytogenesis and meiosis. Moreover, TEM also showed elongated spermatids with disorganized heads partially devoid of cell membrane (Figure 10G, arrowhead), multiple flagella ( Figure 10H, arrow), and absence of the inner arms of flagellar axonemes (Figure 10I, arrowheads) in the lumen of seminiferous tubules. In addition, vacuolization was also observed in the cytoplasm of spermatocytes ( Figure 10F). TEM also revealed the low number of spermatozoa in the epididymis of Tg rats ( Figure 11B) as compared to normal rats ( Figure 11A). Spermatozoa from normal rats displayed intact cell membranes and normal-shaped heads with complete chromatin condensation ( Figure 11A1). Instead, those of Tg rats presented a number of defects: several immature spermatids detaching from Sertoli cells and sloughing into the epididymis (Figure 11B), degenerated round spermatids ( Figure 11B1) and many types of abnormal elongated spermatids (Figure 11B2-B4). In addition, whereas normal rats presented the typically assembled flagellar axonemes composed of nine outer doublet microtubules and a pair of central microtubules (Figure 11A2), most spermatozoa from Tg rats were dead and showed various ultrastructural defects, including breakage of the surface membrane ( Figure 11B5) and a decline in the number of outer dense fibers ( Figure 11B6). In addition, residual bodies of degenerated germ cells (B5) and missing outer doublet microtubules are frequently observed (B6, arrowhead). Reproduced and modified from [42] with permission from the Society for Reproduction and Development.

Disruption of Sertoli-germ junctions
Hematoxylin and eosin staining revealed various abnormal morphologies in Tg rat testis. Cytoplasmic vacuolation of Sertoli cells (Figure 12A, open arrowhead), disconnection between Sertoli and germ cells (Figure 12B, arrows), and disordered arrangement of elongated spermatids (Figure 12C, arrowhead) were present in many seminiferous tubules. In addition, immature spermatozoa falling off from the seminiferous epithelium were observed during all stages of spermatogenesis ( Figure 12D-H, arrowheads). Furthermore, sperm development was disrupted and germ cells with different levels of maturity were observed simultaneously (Figure 12D-H). These results indicate that cell junctions between germ and Sertoli cells may have been affected. This hypothesis may be confirmed by analyzing gene expression profiles of Tg rat testis.

Alterations in gene expression profiles of Tg rat testis
Changes in gene expression evoked by HHV-1-TK accumulation in Tg rat testis were examined by cDNA microarray analysis. We found that 200 genes, 0.67% of all transcripts on DNA chips, were differently expressed between Tg and normal rats. We sorted the genes by their functional categories, such as apoptosis, cell cycle, development, oxidative stress, proteolysis, signal transduction, transcription, translation, transport, metabolism, immune response, and cell adhesion. The highest number of affected genes was linked to metabolism, with 8 genes upregulated and 16 down-regulated by at least 1.5-fold in Tg rat testis ( Table 3) [42].
RT-PCR was performed for 10 genes involved in cell adhesion, signal transduction, and transport: Cnot7, Fgf7, Egfl6, Testin, Ostf1, and Atp2a2 (all up-regulated); and Sh3bp4, Lamc2, Versican, and Mamdc1 (all down-regulated). As listed in Table 4, Cnot7, Fgf7, Testin, Ostf1, Versican, and Mamdc1 were significantly up-or down-regulated, confirming microarray data. Testin is a signaling molecule responsible for monitoring Sertoli-germ cell adherens junctions. Increased secretion of testin by Sertoli cells causes disruption of these junctions [43,44]. Versican is expressed in adult mouse brain, heart, lung, spleen, skeletal muscle, skin, tail, kidney, and testis. It is suggested to play a role in cellular attachment, migration, and proliferation by interacting with cell surfaces and extracellular matrix molecules [45]. Mamdc1, a novel member of adhesion molecules of the immunoglobulin superfamily, is expressed in human Leydig cells of the testis. Mamdc1 mRNA was shown to be up-regulated by pro-inflammatory cytokines, such as tumor necrosis factor-α and interferon-γ, involved in cell adhesion, migration, and recruitment to inflammatory sites [46]. However, the function of versican and mamdc1 in spermatogenesis has not yet been established. Expression of Fgf7, Ostf1, and Cnot7 [47] also increased significantly, indicating a stress response in Tg rat testis. The roles of these genes are listed in Table 5. Note: The affected genes were classified into 12 functional categories according to "Gene ontology".
Gene ontology classification of genes whose expression changes by at least .5-fold in Tg with respect to normal rats, as revealed by cDNA microarray analysis. Reproduced and modified from [42] with permission from the Society for Reproduction and Development. Notably, the contraceptive adjudin has been reported to induce morphological alterations in the seminiferous tubules similar to the ones we observed here, and target directly testin and actin-myosin [48]. However, the exact mechanism by which HHV-1-TK disrupts spermatogenesis and adherens junctions will require further work. Reproduced and modified from [42] with permission from the Society for Reproduction and Development.

Number of virus infection in patients (153)
Reproduced and modified from [21] with permission from the Society for Reproduction and Development. The viral DNA-positive group showed a higher incidence of oligozoospermia compared with the viral DNA-negative group (38.5 vs. 16.7%, P < 0.05; Figure 14). Therefore, to further confirm the relationship between HHVs infection and changes to semen parameters, large-scale analysis according to WHO Guidelines (5 th edition) is strongly suggested. Figure 14. Comparison of oligozoospermia between viral DNA-negative and -positive groups. Analysis of oligozoospermia rates between viral DNA-negative (36 patients) and -positive (39 patients) according to WHO Guidelines (5th edition). Reproduced and modified from [21] with permission from the Society for Reproduction and Development.

Future perspectives
HHV-1 infection might strongly associate with human male infertility, possibly by testisspecific expression of the viral TK gene. However, conclusive evidence that HHV-1-TK causes male infertility is still missing.
As described above, HHV-1-TK is known as a suicide gene that kills target cells specifically in the presence of GCV. However, male infertility normally occurs in the presence of HHV-1 but absence of GCV. Degeneration of spermatogenesis was observed as early as 3 months in Tg rats ectopically expressing HHV-1-TK. With the exception of Sertoli cells, it later resulted in loss of germ cells, possibly by "bystander killing." In contrast, no abnormality in the pituitary gland of Tg rats was observed. This difference may indicate tissue-specific action of HHV-1-TK in the testis, which does not occur in the pituitary gland. At present, it is unclear why HHV-1-TK targets testis over other tissues and how it induces degeneration of spermatogenesis. Further work is required to elucidate this and other features of HHV-related infertility.
Expression of HHV-1-TK may be the main cause of infertility in HHV-1-infected human males.
To firmly establish a causative link, several players require identification: (1) target molecule(s) of HHV-1-TK; (2) the HHV-1-TK-dependent mechanism responsible for failed spermatogenesis; and (3) the mechanism dictating postmeiotic spermatid-specific expression of HHV-1-TK. The identification of target molecule(s) is particularly important, since there is no evidence that enzymatically active HHV-1-TK is required, raising speculation it may only serve as a binding protein. Even if HHV-1-TK functioned as an enzyme, the substrate(s) might be testisspecific. Currently, the only known activity of HHV-1-TK is to catalyze the transfer of the γphosphate from ATP to substrate nucleotides and analogs. Recently, a stable-isotope substrate ([γ- 18 O 4 ]ATP) for kinases has been developed [49]. Mass spectrometry could reveal the identity of [γ- 18 O 4 ]ATP-labeled products obtained by incubating recombinant HHV-1-TK with testis homogenates. The identified HHV-1-TK target(s) may help elucidate the mechanisms determining failed spermatogenesis and postmeiotic spermatid-specific expression of HHV-1-TK. Moreover, knowing the target(s) molecular properties may help design appropriate analogs for pharmacological use. The significance of our results is particularly clear in view of expected novel male contraceptives and drugs for HHV-related male infertility.