Development of the Gubernaculum During Testicular Descent in the Rat
Article first published online: 25 MAY 2011
Copyright © 2011 Wiley-Liss, Inc.
The Anatomical Record
Volume 294, Issue 7, pages 1249–1260, July 2011
How to Cite
Nation, T.R., Buraundi, S., Farmer, P.J., Balic, A., Newgreen, D., Southwell, B.R. and Hutson, J.M. (2011), Development of the Gubernaculum During Testicular Descent in the Rat. Anat Rec, 294: 1249–1260. doi: 10.1002/ar.21393
- Issue published online: 15 JUN 2011
- Article first published online: 25 MAY 2011
- Manuscript Accepted: 22 FEB 2011
- Manuscript Revised: 16 FEB 2011
- Manuscript Received: 5 JUL 2010
- testicular descent;
- genitofemoral nerve;
- mammary bud
Gubernacular elongation during inguinoscrotal testicular descent and cremaster muscle development remains poorly described in mammals. The role of the genitofemoral nerve (GFN) remains elusive. We performed detailed histological analysis of testicular descent in normal rats to provide a comprehensive anatomical description for molecular studies. Fetuses and neonatal male offspring (5–10 per group) from time-mated Sprague-Dawley dams (embryonic days 15, 16, and 19; postnatal days 0, 2, and 8) were prepared for histology. Immunohistochemistry was performed for nerves (Class III tubulin, Tuj1) and muscle (desmin). At embryonic days 15 and 16, the gubernaculum and breast bud are adjacent and both supplied by the GFN. By embryonic day 19, the breast bud has regressed and the gubernacular swelling reaction is completed. Postnatally, the gubernacular core regresses, except for a cranial proliferative zone. The cremaster is continuous with internal oblique and transversus abdominis. By postnatal day 2 (P2), the gubernaculum has everted, locating the proliferative zone caudally and the residual mesenchymal core externally. Eversion creates the processus vaginalis, with the everted gubernaculum loose in subcutaneous tissue but still remote from the scrotum. By P8, the gubernaculum has nearly reached the scrotum with fibrous connections attaching the gubernaculum to the scrotal skin. A direct link between GFN, gubernaculum, and breast bud suggests that the latter may be involved in gubernacular development. Second, the cremaster muscle is continuous with abdominal wall muscles, but most of its growth occurs in the distal gubernacular tip. Finally, gubernacular eversion at birth brings the cranial proliferative zone to the external distal tip, enabling gubernacular elongation similar to a limb bud. Anat Rec, 2011. © 2011 Wiley-Liss, Inc.
It is generally agreed that testicular descent in mammals occurs in two distinct steps with different anatomy and hormonal regulation (Hutson,1985; Amann and Veeramachaneni,2007; Foresta et al.,2008; Hughes and Acerini,2008; Feng et al.,2009). In the first, or transabdominal phase, the genitoinguinal ligament, also known as the “gubernaculum,” undergoes enlargement or the “swelling reaction.” This is seen clearly in pigs and rodents as well as in humans (Wensing,1986; Heyns,1987). The enlarged gubernaculum remains short during growth of the fetal abdomen, holding the testis near the future inguinal region as the fetus enlarges (8–15 weeks in humans; embryonic day (E)16–E19 in rats) (Hutson et al.,1997). By contrast, in females, the gubernaculum just elongates in proportion to the abdomen, and no swelling reaction occurs. The developing ovary in rodents, therefore, remains in a high abdominal position as the cranial suspensory ligament develops further to hold the ovary just caudal to the kidney (van der Schoot,1993).
During the second, or inguinoscrotal phase, the gubernaculum actively elongates out of the abdominal wall to reach the scrotum. This is under the control of the genitofemoral nerve (GFN) and androgen. It has been shown that GFN transection and treatment with sensory neurotoxins disturb the inguinoscrotal phase of descent (Beasley and Hutson,1987; Al Shareef et al.,2007). Furthermore, humans and animal models with androgen insensitivity have undescended testes (Nation et al.,2009b).
Although respiratory and cardiovascular physiology are highly conserved across mammalian species, there are important differences in reproductive strategies, leading to some anatomical variations. The anatomy of the prostate and penis in rodents, for example, is quite different from humans. In addition, the timing of testicular descent varies, with both phases occurring prenatally in humans compared with inguinoscrotal migration occurring postnatally in rodents. The different anatomy of the gubernaculum and its contained cremaster muscle in rodents versus humans has been highlighted recently by Klonisch et al. (2004). However, most of the key features of the mechanism of descent are similar, secondary to evolution with minor differences from a common ancestor. This is supported by a study in dogs, which shows the immunohistochemical labeling of Insl3-RXFP2 signaling as well as the presence of calcitonin gene-related peptide in nerves supplying the developing cremaster muscle in the gubernaculum (Arrighi et al.,2010).
Despite our increasing understanding of this complex mechanism in sexual development, there are still many unresolved issues in both animal models and humans. How, for example, does the GFN regulate gubernacular migration, and where does the cremaster muscle develop from? Is the cremaster muscle derived from the abdominal wall muscles, transversus abdominis and/or internal oblique, or does it develop separately within the gubernaculum? Recent studies from our own laboratory suggest that the gubernaculum develops a growth center that is similar to a progress zone within an embryonic limb bud (Na et al.,2007), but the origin of this putative growth center is unknown.
In this study, we aimed to perform a detailed histological and anatomical analysis of the gubernaculum across the time period of testicular descent (embryonic day 15 to postnatal day 8) in normal male rats to address some of these unanswered questions.
Sprague-Dawley female rats were time mated and maintained in standard rat cages in a temperature-controlled environment with 12-hr light–dark cycle. The morning that a mucus plug was found in the vagina was defined as embryonic day 0. They were fed commercial rat chow ad libitum. Five to ten male offspring per group were collected for embryonic (E) days 15, 16, and 19 and postnatal (P) days 0, 2, and 8 for anatomical and histological analysis, as these time points span the two phases of descent in the rat.
Transverse and Sagittal Sections
The pelvic regions were dissected from the rest of the body after sacrifice and then fixed in 4% paraformaldehyde overnight. Tissues were processed through graded alcohol and xylene, embedded in paraffin blocks, sectioned in either transverse or sagittal planes at 5 μm, and then floated onto silane-coated slides.
Sections were stained with hematoxylin and eosin (H&E) or Masson's trichrome and processed for immunohistochemistry to identify nerves and muscle layers. Nerve labeling was done with mouse monoclonal antibodies against Class III β-tubulin (mouse anti-Tuj1, Covance, MMS-435P, Emeryville, CA), and muscle fiber location was visualized with mouse anti-desmin antibody (Dako, M0760, Carpintaria, USA), followed by 3,3′-diaminobenzidine (DAB) reaction.
Sections were dewaxed in xylene and rehydrated through graded alcohols. Antigen retrieval was achieved with 0.1 mol/L citrate buffer (pH 6.0) for 20 min in an 800-W microwave. Hydrogen peroxide (3%) was used for 10 min of incubation to quench endogenous peroxidase. Slides were rinsed with phosphate-buffered saline (PBS) and incubated with 1% sheep serum for 30 min to block nonspecific binding. Primary antibodies were then applied at the following concentrations: mouse anti-Tuj1 (1:800); mouse anti-desmin (1:200), and slides were incubated overnight at 4°C. Slides were rinsed with PBS, and anti-mouse biotinylated secondary antibody (1:200; Dako, E354; Carpintaria, USA) was added for 2 hr. Slides were rinsed again with PBS, and streptavidin (1:200; Millipore, SA202, Melbourne, Vic, Australia) was applied for 30 min. After a final PBS wash, peroxidase activity was detected with DAB (Dako, 1 mg/mL; Carpinteria, USA). Haematoxylin was used for counterstaining nuclei.
In each immunostaining procedure, a negative control, omitting the primary antibody, was run. As both primary antibodies are very well characterized, with Tuj1 labeling neuronal Class III β-tubulin in all neuronal cells except β-tubulin in glial cells, and desmin labeling Class III intermediate filaments, which are present in skeletal, cardiac, and smooth muscle, separate positive controls were not required.
To provide three-dimensional views of the anatomy, some embryos were prepared for whole-mount immunohistochemistry using the anti-human anti-neurofilament antibody to identify the pathway of the genitofemoral and other nerves (Neurofilament Protein; Clone 2F11; Dako; Golstrup, Denmark) or desmin antibody to identify muscle cells. Embryos were dissected in cold PBS and then fixed in 4% paraformaldehyde overnight at 4°C. Embryos were washed in PBS containing 0.5% Triton X-100 twice, 10 min each, and then dehydrated on ice through 25% methanol in 75% PBS, 50% methanol in 50% PBS, 75% methanol in 25% PBS, and 100% methanol, 15 min for each step. The specimens were then soaked in methanol with 3% hydrogen peroxide (1–2 hr at 4°C), with rocking, and then washed in methanol (10 min). Then, the specimens were rehydrated on ice through 75% methanol in 25% PBS, 50% methanol in 50% PBS, 25% methanol in 75% PBS, 100% PBS + 0.1% Triton X-100, 15 min each step, and washed two times in PBSST (5% donkey serum, 0.1% Triton X-100 in PBS) for 1 hr at 4°C with rocking.
The specimens were incubated with anti-neurofilament antibody (1:800 dilution in PBSST; Dako, Golstrup, Denmark) overnight at 4°C with rocking. The following day the antibody solution was washed out (5 × 15 min) at room temperature and then 5 × 1 hr in PBSST at 4°C with rocking. Then, the specimens were incubated with biotin donkey anti-mouse antibody (1/200 in PBSST) overnight. The specimens were washed 2 × 15 min at room temperature, then 5 × 1 hr at 4°C with PBSST with rocking, and then incubated overnight at 4°C with HRP-streptavidin (1/200 in PBSST) with rocking. Specimens were then stained with DAB (Dako, 1 mg/mL; Carpinteria, USA), cleared in 50% glycerol in 50% PBS, and then 70% glycerol in 30% PBS. Whole-mount embryos were viewed using a Leica DC300F microscope and camera (Wetzlar, Germany). Negative controls were run with omission of the primary anti-neurofilament antibody.
In the rat embryo at embryonic day 15, a transverse section of inguinal region shows a condensation of dense undifferentiated mesenchyme that is the future gubernaculum. This is located between the future rectus abdominis muscle medially and the developing oblique muscles laterally (Fig. 1A). This relationship is more obvious under high power (Fig. 1B) and clearly depicted with desmin labeling to identify the muscle fibers (Fig. 1C). In addition, at this age, the close anatomical relationship between the gubernaculum and the most distal breast (mammary) bud is revealed (Fig. 1B,C). The GFN can also be seen passing from the gubernaculum and supplying the adjacent mammary bud (Fig. 1D,E). The mammary bud is identified as a swelling of the epithelial layers lateral to the phallus, cranial to the scrotum, which is located caudal to the phallus (Fig. 2).
By embryonic day 16, the gubernaculum in the male has enlarged into a conical structure, with a central core of dense but undifferentiated mesenchyme, and peripheral band of muscle differentiation (Fig. 3A,B). The muscle layer is in continuity with the developing abdominal wall muscles, but separate from them (Fig. 3C,D). Tuj1 labeling of the nerves clearly shows the main branches of the GFN coming through the abdominal wall behind the gubernaculum, supplying it, with branches extending beyond it to the skin and mammary bud (Fig. 3E,F).
The gubernacular cord connecting proximally to the gonad is still relatively long at embryonic day 16, as the swelling reaction within the gubernaculum bulb is not yet complete. This can be seen in a dissection of the fetus, revealing the testis and developing epididymis attached to the gubernacular bulb by a long cord, which retains a connection with the posterior abdominal wall by a fold of peritoneum (Fig. 4).
By embryonic day 19, the swelling reaction in the gubernaculum is complete. The gubernacular bulb has enlarged to include a discrete central core of mesenchyme surrounded by two layers of future cremaster muscle (Fig. 5A–D). Nerve labeling with Tuj1 shows small branches of the GFN supplying the developing cremaster muscle. Both the mammary bud and its GFN branch have disappeared (Fig. 5E,F). Muscle labeling with desmin shows that the two layers of the developing cremaster are in continuity with the transversus abdominis and the internal oblique muscle (Fig. 5G,H).
Staining of the nerves with Tuj1 in whole-mount preparations shows the GFN branching before its entry into the future inguinal canal to supply the gubernaculum (Fig. 6A,B). Magnified views of the gubernaculum show a dense innervation by the GFN (Fig. 6C). Desmin staining of the whole-mount embryonic day 19 specimen clearly shows the developing cremaster myoblasts forming two discrete layers at 90-degree angle from each other around the gubernacular bulb (Fig. 6D,E).
By the day of birth, the mesenchyme within the core of the gubernacular bulb is mostly regressed, except for a proliferative zone at the cranial end of the cone (Fig. 7A). With loss of the mesenchymal core, the outer layers of the cremaster muscle are now approximating each other (Fig. 7A–D). The position of the GFN remains unchanged at the back of the future inguinal canal (Fig. 7E), whereas some cutaneous branches still extend beyond the gubernaculum in the transverse plane (Fig. 7F). Muscle staining shows continuity between the developing cremaster muscle in the gubernaculum and the internal oblique and transversus abdominis muscles as well as the persisting mesenchymal core cranially (Fig. 7G,H).
By day 2 postnatally, the gubernacular bulb, which was initially a solid mesenchymal cone bulging into the abdominal cavity, has now remodeled completely and everted into the subcutaneous space, thereby creating the inguinal canal. The dense, undifferentiated mesenchymal core of the gubernacular bulb, which on E19 and P0 was at the cranial end of the intra-abdominal cone, is now located caudally at the end of the everted gubernaculum. The eversion has changed the relationship between the mesenchymal core and its surrounding cremaster muscle layers, which have been completely reversed: at E19, the mesenchyme was covered by the muscle layers, but now it is external and caudal to the muscle layers (Fig. 8A–D). Nerve staining shows that the location of the GFN is unchanged (Fig. 8E,F), whereas muscle staining confirms relocation of the undifferentiated mesenchyme from inside the cranial end of the gubernacular cone prenatally to the outside of the caudal end of the gubernacular diverticulum postnatally (Fig. 8G,H). The processus vaginalis (PV) has been created by this eversion. The early postnatal gubernaculum lacks not only any distal fibrous attachments (Fig. 9A) but also much of the previously conspicuous subcutaneous connective tissue. In many of the sections, a loose plane of cleavage, which may be fixative artifact, is seen around the everted gubernaculum and between the gubernaculum and the scrotum (Fig. 9A).
By day 8 postnatally, the everted gubernaculum has elongated most of the way to the future scrotum, and there has now appeared some fibrous connections between the outside of the gubernaculum and the adjacent subcutaneous tissue of the upper scrotum (Fig. 9B,C). In the elongated PV within the gubernaculum, the caudal epididymis and testis remain connected to the apex of the everted gubernaculum by the cord in its peritoneal fold (mesorchium) (not shown in Fig. 9).
This detailed study of the rat gubernaculum confirms many previous reports about the anatomy and demonstrates several new findings (Hadziselimovic,1983; Wensing,1986). First, an intriguing anatomical relationship can be seen between the embryonic gubernaculum and the mammary bud at embryonic day 15. At this time, they are adjacent structures and both supplied by branches of the GFN. Second, the two layers of the developing cremaster muscle are connected to but separate from the two inner muscles of the abdominal wall. Third, the swelling reaction within the gubernacular bulb leads to substantial growth of both the mesenchymal core of the gubernaculum as well as the cremaster muscle layers between embryonic days 16 and 19. Between embryonic day 19 and birth, most of the central mesenchyme and its extracellular matrix within the gubernacular bulb disappears, except for what will become the putative growth center of the everted postnatal gubernaculum. This is located at the cranial end of the intra-abdominal fetal gubernaculum at birth, but by eversion becomes located on the outside, caudal tip of the postnatal gubernaculum at postnatal day 2. The cremaster muscle progressively elongates, and the gubernaculum migrates toward the scrotum. Postnatally, the subcutaneous tissues around the gubernaculum change from moderately dense connective tissue (that will form the mammary fat pad in females) to extremely loose and sparse mesenchyme that facilitates gubernacular eversion and then elongation. Finally, once the gubernaculum has reached the site of the future scrotum, dense connective tissue reappears, anchoring the outside of the PV to the scrotal skin (Fig. 10).
The results presented here provide a different view of the mechanism of descent from that described previously (Wensing,1986), but the broad outline is similar. The role of eversion of the solid gubernacular bulb to form a hollow cone is clearly not the only process required for inguinoscrotal descent, as shown by Lam et al. (1998), although some authors have previously suggested this (Husmann,2009). Lam et al (1998) showed that eversion of the gubernacular cone would only enable the gubernaculum in rodents to cover a ¼–⅓ of the distance to the scrotum, so clearly there is active growth required as well as a mechanism for orientating the direction of elongation (Fallat et al.,1992).
The idea is emerging that the mammary bud and gubernaculum may be linked.
The close association between the distal mammary bud, GFN, and gubernaculum has been described recently. Interestingly, androgen blockade with flutamide causes mammary tissue to form around the gubernaculum in male rats (Nation et al.,2009b; Balic et al.,2010). In marsupials, there is a clear link between the gubernaculum and the breast where the marsupial homolog of the cremaster muscle, called ilio-marsupialis, is attached to the mammary gland and nipple and acts as a suspensory muscle for the lactating breast (Griffiths and Slater,1988; Nation et al.,2009b). In addition, the mammary bud has androgen receptors and responds to androgen by regression (Hens and Wysolmerski,2005).
The ectoderm and adjacent specialized mesenchyme of the mammary bud is a potential source for regulatory signals controlling the gubernaculum, in a way similar to the apical ectodermal ridge of an embryonic limb bud (Balic et al.,2010). Indeed, the mammary line develops along the boundary of the early embryonic trunk between the sites of the limb buds, so its inductive properties are likely to be similar (Cho et al.,2006).
The GFN supplies both the gubernaculum and regressing male mammary bud, raising the possibility that the subsequent sexual dimorphism of the nerve may be mediated by trophic signaling from the mammary tissue. The GFN is sexually dimorphic, but the exact location of the androgen receptors that are likely to regulate this process remains unknown. It has been suggested that the androgen receptors are in the gubernaculum itself or may be in the dorsal root ganglion (Husmann and McPhaul,1991; Schwindt et al.,1999), but this remains unresolved. The location of the mammary bud and its innervation now suggests an additional site that needs to be studied, that is, the mesenchyme adjacent to the mammary bud.
The early muscle cell differentiation within the bulb of the gubernaculum raises the question about the origin of the cremaster muscle. Recent work from our laboratory suggests that the cremaster muscle may be primarily developing within the gubernaculum itself, and therefore, it is not secondarily derived by stretching of one or more of the abdominal wall muscles during descent. We have shown that the rat gubernaculum grows from a distal growth center during the inguinoscrotal phase (Hrabovszky et al.,2002; Shenker et al.,2006), and that this process is analogous to a progress zone in a limb bud (Huynh et al.,2007; Nightingale et al.,2008). In this study, during the transabdominal phase, there is continuity between what will later become the bilaminar cremaster sac and the two inner muscles of the abdominal wall. The interpretation of this remains, however, uncertain. The separate nerve supply of the cremaster muscle by the GFN supports independent development, as does the elongation of the cremaster sac by proliferation of new muscle cells in the distal growth center during inguinoscrotal descent.
Rapid remodeling of the mesenchymal cells and extracellular matrix within the gubernaculum bulb at the onset of the inguinoscrotal phase allows the gubernaculum to evert and explains how the distal growth center develops (Costa et al.,2002). Abnormalities of this process would be expected to occur in models of cryptorchidism and have been shown in the flutamide-treated rat, as has been reported recently (Nation et al.,2009a).
Finally, the loose mesenchymal tissue around the elongating gubernaculum at postnatal days 0 and 2 is similar to the free end of the gubernaculum in human dissections reported by Heyns (1987). What molecular steps could produce this loss of extracellular matrix? This process has not been investigated fully, but may be important to understand human cryptorchidism, as its failure may prevent or impede gubernacular migration. Certainly, there are reports of a “fibrous barrier” at the neck of the scrotum in some children with cryptorchidism, but it is not known whether this is the cause or the result of cryptorchidism (Backhouse,1982). The appearance of a dense fibrous attachment between gubernaculum and scrotum at the completion of migration, as seen here, suggests that the “fibrous barrier” in children may be just the normal process developing after migration has ended prematurely. Another key implication of the fibrous adherence occurring at the end of descent is that it provides some experimental support for the view that perinatal testicular torsion in babies is caused by the gubernaculum becoming twisted within the scrotum before adherence has occurred (Samnakay et al.,2006). The twisting itself may be innate, secondary to spiral migration, or may be extrinsic, and caused by bumping by the thigh of the fetus during the last few weeks of pregnancy.
In conclusion, this study demonstrates that the mechanism of testicular descent in the rat is a complicated, multistaged process. The “swelling reaction” initiating the transabdominal phase has been known for nearly 30 years, but new evidence suggests that the mammary line may play a role in initiating the migration of the gubernaculum to the scrotum in the inguinoscrotal phase. The cremaster muscle forms early in development and may have a function in gubernacular eversion, bringing a proliferative zone to the external tip, allowing elongation to the scrotum. Future studies should concentrate on the interaction between the gubernaculum and the inguinoscrotal mesenchyme and the role of remodeling of the extracellular matrix. The fate of the cells in the mesenchymal core of the swollen bulb needs to be determined to find out if they die or migrate away. Finally, further studies at the molecular level are required to define the relationship between the cremaster muscle and the oblique muscles of the lower abdominal wall.
- 2007. Exogenous calcitonin gene-related peptide perturbs the direction and length of gubernaculum in capsaicin-treated rats. Pediatr Surg Int 23: 305–308. , , .
- 2007. Cryptorchidism in common eutherian mammals. Reproduction 133: 541–561. , .
- 2010. An insight into testis and gubernaculum dynamics of INSL3-RXFP2 signalling during testicular descent in the dog. Reprod Fertil Dev 22: 751–760. , , , , .
- 1982. Development and descent of the testis. Eur J Pediatr 139: 249–252. .
- 2010. Hidden in plain sight: the mammary line in males may be the missing link regulating inguinoscrotal testicular descent. J Pediatr Surg 45: 414–418; discussion 418. , , , , , , .
- 1987. Effect of division of genitofemoral nerve on testicular descent in the rat. Aust N Z J Surg 57: 49–51. , .
- 2006. Molecular interactions between Tbx3 and Bmp4 and a model for dorsoventral positioning of mammary gland development. Proc Natl Acad Sci USA 103: 16788–16793. , , , , , , , .
- 2002. Testicular migration: remodeling of connective tissue and muscle cells in human gubernaculum testis. J Urol 167: 2171–2176. , , , .
- 1992. Histologic evaluation of inguinoscrotal migration of the gubernaculum in rodents during testicular descent and its relationship to the genitofemoral nerve. Pediatr Surg Int 7: 265–270. , , , .
- 2009. INSL3/RXFP2 signaling in testicular descent. Ann N Y Acad Sci 1160: 197–204. , , , , , , , , , , .
- 2008. Role of hormones, genes, and environment in human cryptorchidism. Endocr Rev 29: 560–580. , , , .
- 1988. The significance of striated muscle in the mammary glands of marsupials. J Anat 156: 141–156. , .
- 1983. Cryptorchidism: management and implications. Berlin: Springer-Verlag. p 1–135. .
- 2005. Key stages of mammary gland development: molecular mechanisms involved in the formation of the embryonic mammary gland. Breast Cancer Res 7: 220–224. , .
- 1987. The gubernaculum during testicular descent in the human fetus. J Anat 153: 93–112. .
- 2002. Role of the gubernacular bulb in cremaster muscle development of the rat. Anat Rec 267: 159–165. , , , , , .
- 2008. Factors controlling testis descent. Eur J Endocrinol 159 ( Suppl 1): S75–S82. , .
- 2009. Testicular descent: a hypothesis and review of current controversies. Pediatr Endocrinol Rev 6: 491–495. .
- 1991. Localization of the androgen receptor in the developing rat gubernaculum. Endocrinology 128: 383–387. , .
- 1985. A biphasic model for the hormonal control of testicular descent. Lancet 2: 419–421. .
- 1997. Anatomical and functional aspects of testicular descent and cryptorchidism. Endocr Rev 18: 259–280. , , .
- 2007. Signalling molecules: clues from development of the limb bud for cryptorchidism? Pediatr Surg Int 23: 617–624. , , , .
- 2004. Molecular and genetic regulation of testis descent and external genitalia development. Dev Biol 270: 1–18. , , .
- 1998. Does the gubernaculum migrate during inguinoscrotal testicular descent in the rat? Anat Rec 250: 159–163. , , .
- 2007. Cell membrane and mitotic markers show that the neonatal rat gubernaculum grows in a similar way to an embryonic limb bud. J Pediatr Surg 42: 1566–1573. , , , , , .
- 2009a. The antiandrogen flutamide perturbs inguinoscrotal testicular descent in the rat and suggests a link with mammary development. J Pediatr Surg 44: 2330–2334. , , , , , , .
- 2009b. The hormonal control of testicular descent. Pediatr Endocrinol Rev 7: 22–31. , , , , .
- 2008. The migrating gubernaculum grows like a “limb bud”. J Pediatr Surg 43: 387–390. , , .
- 2006. Spin on perinatal testicular torsion. J Paediatr Child Health 42: 734–736. , , .
- 1999. Localization of calcitonin gene-related peptide within the genitofemoral nerve in immature rats. J Pediatr Surg 34: 986–991. , , , , .
- 2006. A new role for androgen in testicular descent: permitting gubernacular cell proliferation in response to the neuropeptide, calcitonin gene-related peptide. J Pediatr Surg 41: 407–412. , , , .
- 1993. The name cranial ovarian suspensory ligaments in mammalian anatomy should only be used to indicate the structures derived from the foetal cranial mesonephric and gonadal ligaments. Anat Rec 237: 434–438. .
- 1986. Testicular descent in the rat and a comparison of this process in the rat with that in the pig. Anat Rec 214: 154–160. .