Infertility affects ~9% of couples worldwide [1] and about 12% in the Thai population [2]. Infertility results from various factors, including male ones, which are solely responsible for infertility in 20%-30% of cases and contribute to 50% of cases overall [3]. Oligozoospermia is associated with infertility. The World Health Organization (WHO) defined the lower reference limit for ‘normal’ sperm concentration and total sperm number to be 15 × 10⁶ spermatozoa per ml (5^th centile, 95% confident interval [CI] 12-16 × 10⁶) and 39 × 10⁶ spermatozoa per ejaculate (5^th centile, 95% CI 33-46 × 10⁶), respectively [4]. Among male factors, 10.1% of infertile men have hypogonadism with low serum testosterone (T) levels and 2.3% show primary hypogonadism [5].

Many hormones and other factors, especially FSH and T, are involved in male reproduction. The hypothalamus produces gonadotropin-releasing hormone, which stimulates the anterior pituitary gland to synthesize gonadotropins, including FSH and LH. FSH regulates the survival of spermatogonia [6] and stimulates B-spermatogonia to differentiate via receptors on Sertoli cells; whereas LH stimulates Leydig cells to synthesize T, which is indispensable for the normal function of Sertoli cells. Sertoli cells are the nursing cells of male gametes. Zhengwei et al. observed that 10 germ cells or 1.5 spermatozoa are embedded in each human Sertoli cell [7]. Hormones regulate spermatogenesis through their effects on Sertoli cells. In mice, many Sertoli cell markers and junctional proteins, such as N-cadherin, connexin-43, gelsolin, laminin-γ3, occludin, testin, nectin, zyxin and vinculin, are androgen-dependent [8]. Several of these molecules are associated with establishing the blood-testis barrier, the release of spermatozoa from the seminiferous epithelium and remodeling of germ cell-Sertoli cell junctions [9]. Testosterone and the androgen receptor are involved in two essential processes of spermatogenesis [10]. First, they act in maintaining the attachment of maturing spermatids to Sertoli cells. Second, they are required for the release of mature spermatids/spermatozoa from the Sertoli cells [10]. However, the collaboration between FSH and T is still inconclusive. Overall, FSH and T act in synergy to produce the factors required to maximize the production of spermatozoa. However, FSH inhibits T-mediated binding of germ cells to Sertoli cells [10, 11]. Consequently, altered FSH levels might be linked to changes in sperm attachment.

Hypogonadism is a rare cause of infertility. It is a clinical syndrome with unequivocally and consistently low serum total T and/or free T concentrations [12]. Serum T levels < 320 ng/dL are suggested to be symptomatic of hypogonadism [12]. Both serum and seminal T and dihydrotestosterone concentrations are correlated with overall sperm concentration and motile sperm concentration [13, 14]. Accordingly, androgen replacement therapy for men with hypogonadism aiming to raise the male sex hormone levels to an optimal level might help reverse abnormal semen production.

Topically applied T gel provides an effective method for T replacement in men with hypogonadism. It was reported to be beneficial in keeping serum T levels steadily within the adult male range [15-17], improving sexual function, mood, increasing lean body mass and decreasing fat mass [15-20]. Moreover, it has advantageous effects for bone mineral density with long-term safety [15-20]. However, data on the efficacy of T gel replacement therapy for spermatogenesis are still limited. A study of the influence of such treatment on spermatogenesis in hypogonadal men with normal semen parameters has shown that the application of 50 mg T gel daily could significantly increase serum T concentration and seminal volume after 3 months [21]. Testosterone is essential for spermatogenesis; nevertheless, supraphysiological T levels suppress LH excretion from the anterior pituitary gland, and consequently decrease intratesticular T synthesis and impair spermatogenesis [22-24]. Despite its suppressive effect, low dose T supplementation is supposed to improve sperm maturation [25] via a non-classical signaling pathway. Spermatogenesis occurs in the seminiferous tubules. It takes at least 64 days for the germ cell lineage to develop from spermatogonia to mature spermatozoa, or 74 days if including the time for spermatogonial renewal [26]. It takes another 12-21 days to transport spermatozoa from the testis to the ejaculatory duct via the epididymis and vas deferens [27]. Therefore, the whole process needs around 3 months to complete. As a result, any alteration in spermatogenesis induced by T therapy will be evident by about 3 months later.

The primary objective of this study was to investigate changes in sperm concentration in oligospermic infertile men with low serum T levels after 3 months of daily low-dose T gel applications. The secondary objective was to study changes in semen volume, motility, sperm morphology and viability, changes in both serum FSH, LH, T, estradiol (E) levels, seminal T and E levels and T/E ratios in serum and seminal plasma.

The baseline characteristics of the 16 infertile men are shown in Table 1. Two patients had underlying illnesses: one had essential hypertension and was taking losartan (50 mg) and atenolol (25 mg) once daily; the other had dyslipidemia that was controlled with lifestyle modification. All of them denied symptoms of hypogonadism except infertility. Two patients are current smokers (12.5%). Baseline semen parameters, seminal T and E levels and the T/E ratios are given in Table 2. Seminal T and E levels in all subjects were above the assay sensitivities.

Changes in semen characteristics are displayed in Table 3. The median sperm concentration increased significantly to 9 × 10⁶/mL (p = 0.04) and 8 × 10⁶/mL (p = 0.02) at the first and second month compare to baseline, respectively. However, at the third month the sperm concentration had rebounded to be similar to baseline. Other characteristics, including volume, pH, motility and viability, and percentage of normal sperm morphology at each month were not different from baseline. Serum FSH, LH, E, T levels and T/E ratios and seminal E, T levels and T/E ratios are shown in Table 4. Serum FSH significantly (p < 0.001) decreased after T gel therapy over all three months after starting T gel therapy compare to baseline. Serum T levels raised significantly (p < 0.001) in first and second months, while seminal T levels raised significantly (p = 0.008) in second and third months compared with the pretreatment data. Serum LH and serum and seminal estradiol (E) levels were similar to the baseline level. The seminal T/E ratios increased significantly (p = 0.004) in second and third months compare to baseline.

Some patients got other benefits of T gel application, such as improvements in libido and sleep quality. As for adverse side effects, two patients developed itching, but this resolved spontaneously. Unfortunately, an ischemic stroke developed in one patient with hypertension after 1 month of using T gel. Although he recovered fully in 3 hours, the study was discontinued.

The use of androgen treatment in cases of idiopathic male infertility can lead to a direct effect [30] and a rebound effect [31, 32]. The direct effect means that low-dose androgen application improves epididymal sperm maturation directly; while the rebound effect occurs after stopping androgen supplementation. When androgen-induced gonadotropin suppression is ceased, it leads to a transient rebound increase in the secretion of gonadotropins, which in turn stimulates spermatogenesis. Unfortunately, both rationales lack solid supporting evidence in humans. Failure to find a direct effect can be explained by the supraphysiological T level induced by improper supplementation, which in turn induces negative feedback on the gonadotropin releasing system, leading to a suppression of spermatogenesis.

In this study, the sperm concentrations increased significantly at the first and second month compared with baseline. This increase might have resulted from a synergistic action of two mechanisms. First, the increased intratesticular T levels, which were indirectly represented by increased seminal T level and seminal T/E ratios, could have augmented the release of mature spermatozoa from the Sertoli cells by the first month. Testosterone acts on Sertoli cells via three major pathways: the classical pathway and two non-classical pathways [11]. One non-classical pathway induces Ca²⁺ influx into Sertoli cells, the other is the activation of a series of kinases (within 1 min) which are critical to maintain spermatogenesis. The latter one acts via Src tyrosine kinase and extracellular signal-regulated (ERK) kinases to facilitate the adhesion of immature germ cells to Sertoli cells and also through Src to permit the release of mature spermatozoa [11]. Thus, increased intratesticular T levels probably facilitate this function. Second, the decreased serum FSH concentration during the first month of T supplementation might have reduced its inhibition of T-mediated binding of germ cells to Sertoli cells. FSH exerts its function on Sertoli cells through at least five signaling pathways, involving cAMP-dependent protein kinase A (PKA), mitogen-activated protein kinase (MAPK), calcium, phosphatidylinositol 3-kinase and phospholipase A₂ [33]. Shupe et al. found that FSH inhibited T-mediated activation of the ERK and MAPK pathways in Sertoli cells via the PKA-mediated inhibition of Raf kinase and reduced germ cell attachment to Sertoli cells in culture [10]. As a result, the reduced FSH concentration might theoretically have amplified the function of T. However, sperm concentration at the third month after starting T gel use was not different from baseline. This was similar to the study conducted by George et al. In that study, 50 mg daily T gel was used in eugonadotropic/hypogonadic men with normal semen parameters. There was normalizing of serum T levels without suppression of sperm concentration or motility [21]. This might have been associated with the significantly decreased FSH level. Reduced serum FSH might inhibit spermatogonial differentiation and maturation, reflected by the diminished sperm concentration 3 months later in our study. In other words, FSH has a dual effect on sperm production. It has an acute effect on the regulation of T-mediated binding of germ cells to Sertoli cells; and a delayed effect that promotes the transition from type A pale spermatogonia to type B spermatogonia during spermatogenesis [34]. The latter might have been reflected in the decreased sperm concentration around 3 months later in our study.

In the presented study, it was found significant increases in seminal T levels along with serum. The finding suggests that the gel application caused higher intratesticular T levels. Physiologically, these levels in men are about 50-100 fold higher than those in serum [34-37]. Intratesticular testosterone accumulates from at least two sources, one from local production by Leydig cells, the other by transportation from lymph spaces via a binding carrier molecule such as albumin [38]. Therefore, intratesticular androgen levels could be raised by lymphatic transportation, and excreted into the semen. Studies in animal models also found the restoration of testicular T via exogenous T administration [39, 40].

Testosterone is one of the most important sex hormones in men. It is responsible not only for sperm maturation, but also influences sexual function, muscle strength, and lipid and bone metabolism. Here we recruited asymptomatic men with low baseline T levels. After the daily use of 25 mg T gel, serum T levels were raised to the normal adult male ranges as seen in previous studies that used 50, 75 and 100 mg T gel applications [20, 21]. Thus, low-dose T gel applications were adequate for hormone replacement in these hypogonadal men. Even this small amount of gel could effectively suppress serum FSH levels. Besides the effects on serum FSH and T levels, we did not find any significant changes in those for serum LH and E, similar to previous study [41].

A balance between estrogens and androgens in seminal plasma is also important for normal male fertility. Luboshitzky et al. reported elevated seminal plasma E and decreased T levels in infertile men, which suggested a correlation between the seminal plasma T/E ratio with sperm concentration [42]. In the present study, seminal plasma T was significantly increased after treatment while the E level did not change, leading to increased T/E ratios. These increased ratios might explain the increased sperm concentrations for the first 2 months after T gel application because T/E ratios in the seminal plasma from oligospermic men were significantly lower than in a normospermic group [43]. Moreover, it has been suggested that selective estrogen receptor modulators and aromatase inhibitors can restore the T/E ratio and improve sperm concentrations [44]. Here, we found a slight positive correlation between serum and seminal T/E ratios when analyzed with Spearman’s log-rank test. The range of correlation coefficients was 0.15 to 0.40. Both hormones have their specific functions in the testes. Androgens can be converted into estrogens via the cytochrome P450 aromatase in spermatozoa [45], which is involved in the resorption of fluid in the efferent ducts secreted by Sertoli cells under androgen control [46, 47]. Therefore, a local balance between estrogens and androgens plays an important role in maintaining normal spermatogenesis.

Despite hormonal level alterations, semen volume was not changed after T gel application. Since semen is a mixture of testicular spermatozoa, secretion of seminal vesicle, prostate gland and mucous from bulbourethral gland; semen volume can be affected by other factors beyond testosterone, such as BMI and smoking status.

Transdermal T delivery in this study demonstrated long and steady improvements in serum T levels and was generally well tolerated. Only 10%-15% of the preparation is absorbed into the systemic circulation [48]. One serious adverse event that occurred in this study was an acute cerebrovascular incident. The patient was 50 years old and had essential hypertension. Although this seems not to have resulted from the medication, it was discontinued in any case.

Our study is the first clinical trial demonstrating the effects of T gel on hypogonadal men with oligozoospermia. We found positive effects over the first two months, but a negative effect by the third month. We investigated correlations between serum and seminal plasma T and E levels but the results were inconclusive. One limitation of the study was the inherent variations in semen parameters which might have affected the results. The small sample size was probably insufficient to evaluate the potential and long-term effects of the drug fully. Therefore, well-designed randomized controlled trials are needed to demonstrate the real effects of T gel applications on semen quality in hypogonadal men with oligozoospermia.

In conclusion, daily applications of 25 mg T gel improved sperm concentrations in these oligospermic infertile men with hypogonadism after 1 and 2 months of application. Nevertheless, the use of T gel beyond the third month reduced sperm concentrations. Moreover, daily application of 25 mg T gel in these hypogonadal men could raise serum T levels into the normal adult male range. Further studies are needed to clarify the mechanisms and optimal dosage of T required to improve sperm concentrations in eugonadotrophic hypogonadal men with oligozoospermia.

This clinical trial was conducted at the Infertility unit, Siriraj Hospital, Mahidol University, Thailand after obtaining ethical approval from the relevant Institutional Review Board (approval number: CoA Si014/2016). It was registered with the Thai clinical trials registry (TCTR20161030001) and conformed to the Declaration of Helsinki. All participants provided informed consent after counseling for infertility treatment and being informed about the study protocol.

The study was financially supported by a Siriraj Grant for Research Development, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand (Grant Number R015932020). The author would like to thank Miss Daunpen Orachon, Miss Chidchanok Kaewjunun and all staffs of the Infertility Unit, Siriraj Hospital for their assistance in recruiting the participants and laboratory procedures. We thank Assist. Prof. Dr. Chulaluk Komoltri from the Clinical Epidermiology Unit, Office for Research and Development, Faculty of Medicine Siriraj Hospital, Mahidol University for her assistance in statistical analysis.

The authors declare no conflict of interest.