Cellular Signaling by Fibroblast Growth Factors (FGFs) and Their Receptors (FGFRs) in Male Reproduction

Abstract

The major function of the reproductive system is to ensure the survival of the species by passing on hereditary traits from one generation to the next. This is accomplished through the production of gametes and the generation of hormones that function in the maturation and regulation of the reproductive system. It is well established that normal development and function of the male reproductive system is mediated by endocrine and paracrine signaling pathways. Fibroblast growth factors (FGFs), their receptors (FGFRs), and signaling cascades have been implicated in a diverse range of cellular processes including: proliferation, apoptosis, cell survival, chemotaxis, cell adhesion, motility, and differentiation. The maintenance and regulation of correct FGF signaling is evident from human and mouse genetic studies which demonstrate that mutations leading to disruption of FGF signaling cause a variety of developmental disorders including dominant skeletal diseases, infertility, and cancer. Over the course of this review, we will provide evidence for differential expression of FGFs/FGFRs in the testis, male germ cells, the epididymis, the seminal vesicle, and the prostate. We will show that this signaling cascade has an important role in sperm development and maturation. Furthermore, we will demonstrate that FGF/FGFR signaling is essential for normal epididymal function and prostate development. To this end, we will provide evidence for the involvement of the FGF signaling system in the regulation and maintenance of the male reproductive system.

• I. Introduction

• II. The FGF Signaling System

  • A. Discovery of FGFs/FGFRs
  • B. Conserved factors throughout the animal kingdom
  • C. Structural properties of FGFs and FGFRs
  • D. Diversity of FGF signaling
  • E. Heparin sulfate proteoglycans in FGF-FGFR ligand binding
  • F. Pathways of FGF signal transduction
  • G. Regulation of FGF signaling
• III. FGFs and FGFRs in Sexual Differentiation

  • A. The bipotential gonad
  • B. Sry/Sox9 initiation of testicular development
  • C. FGFs in Sertoli cell differentiation
  • D. SPROUTY2 inhibition of FGF signaling during sexual differentiation
• IV. FGFs and FGFRs in Male Germ Cells

  • A. FGFs in the testis
  • B. Differential expression of FGFR splice variants
  • C. FGF signal transduction in the acquisition of sperm motility
  • D. FGFs and testicular cancer
• V. FGFs and FGFRs in the Epididymis

  • A. Lumicrine regulation of epididymal function
  • B. FGFs as lumicrine regulators
  • C. Differential expression of Fgf/Fgfrs
• VI. FGFs/FGFRs in Seminal Vesicle Development

  • A. Branching morphogenesis
  • B. Seminal vesicle shape (svs) mutation
  • C. FGFs in the regulation of seminal vesicle development
• VII. FGFs and FGFRs in the Prostate

  • A. FGF signaling in the developing prostate
  • B. FGF signaling in the adult prostate
  • C. FGF signal transduction pathways in prostate cancer

• VIII. FGF/FGFRs and Disease in Humans

• IX. Mutations of FGFR1 Signaling as a Cause of Kallmann Syndrome

  • A. Kallmann syndrome
  • B. KAL1 encodes for anosmin-1
  • C. Association of FGFR1 (KAL2) with Kallmann syndrome

• X. Concluding Remarks

I. Introduction

THE MAJOR FUNCTION of the reproductive system is to ensure the survival of the species by passing on hereditary traits from one generation to the next. The primary reproductive organs are responsible for producing gametes and for generating hormones that function in the maturation and regulation of the reproductive system. It is now well established that normal development and function of the testis are mediated by endocrine and paracrine pathways including hormones, growth factors, and cytokines as well as by direct cell-to-cell contacts. Growth factors are hormone-related substances that promote cell proliferation, regulate tissue differentiation, and modulate organogenesis.

Fibroblast growth factors (FGFs) are a large family of multifunctional peptide growth factors of which there are at least 28 distinct members (1). The members of this peptide growth factor family have been identified in a variety of organisms (2) and have demonstrated pivotal roles in many cellular processes including mitogenesis, differentiation, migration, and cell survival (reviewed in Ref. 3). During embryonic development, FGFs play a critical role in morphogenesis by regulating cellular proliferation, differentiation, and migration. In the adult, FGFs are homeostatic factors functioning in tissue repair and wound healing, in the control of the nervous system, and in tumor angiogenesis (2,4). FGFs mediate their cellular function through binding to and activating a family of receptor tyrosine kinases (RTKs), which are designated the high-affinity FGF receptors (FGFRs). Since the isolation of the first complete FGFR cDNA, five distinct FGFRs, FGFR1 to FGFR5, have been identified; however, most FGF action is mediated through FGFRs 1–4 (5,6,7,8). FGFs also bind to heparin or heparin sulfate proteoglycans (HSPG), low-affinity receptors that do not transmit a biological signal; rather they function as accessory molecules that regulate FGF-binding and activation of the FGFRs (9,10,11,12). Like all RTKs, functional FGFRs are transmembrane proteins composed of an extracellular ligand-binding domain and a cytoplasmic domain containing the catalytic protein tyrosine kinase core (6,13,14). The extracellular ligand-binding domains of FGFRs are prototypically composed of three Ig-like domains (15). Alternative mRNA splicing of the Ig domains in FGFR1 through FGFR3 leads to distinct functional variants of these receptors (16). HSPG binding of FGF induces FGFR dimerization, which is followed by the transphosphorylation of receptor subunits and the initiation of intracellular signaling events (5,16,17).

FGFs /FGFRs have been localized in many cells of the male reproductive tract and have been shown to be important in regulating the growth and development of several reproductive organs (18,19,20,21,22,23,24,25,26). Components of the FGF/FGFR pathway and their associated signaling cascades have also been localized on spermatozoa and have demonstrated their ability to influence several aspects of sperm maturation (27,28,29,30), reducing the capability of the male to produce functional gametes and affecting male fertility. This review focuses on the way in which FGFs and their respective receptors play a role in the regulation and maintenance of the male reproductive system. Table 1 provides a summary of the localization/expression of FGF/FGFRs in the male reproductive system and was compiled from numerous studies that used techniques such as: in situ hybridization, immunolocalization, RT-PCR, and Western blot analysis. The role of these FGF/FGFRs in male reproduction will be discussed over the course of this review.

Table 1.

FGF/FGFR (ligand/receptor) localization/expression in the male reproductive system

FGF/FGFR	Reproductive organ	Cell type
Ligands
FGF1 (aFGF)	Testis	S
FGF2 (bFGF)	Testis	S
	Prostate	M/E
	Epididymis (IS)	?
FGF3	Prostate (cancer)	E
FGF4	Testis	S
	Epididymis (IS)	?
FGF5	Testis	S
	Prostate (cancer)	M
FGF6	Prostate	E
FGF7	Seminal vesicle	M/E
	Prostate	M/E
FGF8	Epididymis (IS)	?
	Prostate	M
FGF9	Bipotential gonad	S
	Prostate	M
FGF10	Seminal vesicle	M
	Prostate	M
FGF11, 12	?
FGF13	Bipotential gonad	Testis cord
FGF14–16	?
FGF17	Prostate	E
FGF18	Bipotential gonad	Testis cord
FGF19–25	?
Receptors
FGFR1 IIIb	Testis	Seminiferous E/I, peritubular cells, L, G-, elongating spermatids
	Epididymis (IS)	?
FGFR1 IIIc	Testis	L
	Epididymis (IS)	Principal cell E, cauda
	Prostate	E
FGFR2 IIIb	Epididymis (IS, cauda)	?
	Prostate	E
	Seminal vesicle	E
FGFR2 IIIc	Bipotential gonad	S
	Testis	G- Spermatocytes/spermatids, L, S
	Epididymis (IS, cauda)	?
	Prostate (cancer)	E
FGFR3 IIIb	Epididymis (IS)	?
	Prostate	E
FGFR3 IIIc	Testis	G- Spermatocytes/spermatids, L
	Epididymis (IS)	?
FGFR4	Testis	Seminiferous E, peritubular cells, G- Spermatogonia/spermatocytes/elongating spermatids, L, S
	Epididymis (IS)	?
	Prostate	E
FGFR5	Testis	G-Spermatogonia

Data presented within this table were compiled from numerous studies utilizing techniques such as in situ hybridization, immunolocalization, RT-PCR, and Western blot analysis. aFGF, Acidic FGF; bFGF, basic FGF; E, epithelium; G, germ cells; I, interstitium; IS, initial segment; L, Leydig cells; M, mesenchyme; S, Sertoli cells; ?, unknown. 

II. The FGF Signaling System

A. Discovery of FGFs/FGFRs

FGFs were first identified from bovine brain extracts based on their mitogenic and angiogenic activities (1). FGF1 and FGF2 were the first of the peptide family to be isolated and have both been characterized as potent mitogens of many cell types (13,31). A significant improvement in the purification of FGF1 and FGF2 was made when it was discovered that both factors bind to heparin, leading to the development of standardized purification protocols using heparin affinity chromatography (32,33). The highly purified preparations of FGFs were subjected to amino acid sequencing and thus made cloning of cDNAs possible for both FGF1 and FGF2 (13,34,35). Initial biological studies using purified or recombinant FGF1/FGF2 demonstrated that both factors acted as mitogens and chemoattractants for endothelial cells in vitro and exhibited potent angiogenic activity in vivo (36,37). The biological roles of more than half of the known mammalian FGFs have been investigated utilizing homologous recombination in mice to target individual FGFs. These studies, together with analysis of the role played by FGFs in specific developmental systems and expression patterns of FGFs in different tissues/cells, demonstrate that FGFs play critical roles during most stages of mouse development and organogenesis (2,4). Additionally, these studies show that certain members of the FGF family have specialized biological roles resulting in a highly specific phenotype (i.e., Angora mutant of FGF5 −/− mice), whereas the loss of other FGFs can be compensated for by the related members of the FGF family resulting in no obvious phenotype (i.e., no obvious defect was detected in FGF1 −/− mice) (4,38,39).

B. Conserved factors throughout the animal kingdom

FGF genes and those encoding FGFRs have been identified in multicellular organisms ranging from the nematode, Caenorhabditis elegans, to the mouse, Mus musculus, and the human, Homo sapiens. However, FGF-like sequences have not been identified in unicellular organisms such as bacteria, Escherichia coli and yeast, Saccharomyces cerevisiae (40). Although the Drosophila and C. elegans genomes have been sequenced, only one Fgf gene (branchless) was identified in Drosophila (41), whereas two (egl-17 and let-756) have been identified in C. elegans (16,42). This is in contrast to the large numbers of Fgf genes identified in vertebrates, indicating that both the Fgf and Fgfr gene families have greatly expanded during the evolution of primitive metazoans to vertebrates (2,40). Most human FGF genes are scattered throughout the genome, indicating that the FGF gene family was generated by both gene and chromosomal duplication and translocation during evolution (for review, see Refs. 2 and 43).

The Fgf gene family appears to have expanded in two phases during evolution. In the first phase, during early metazoan evolution, Fgf genes increased from two or three genes to six by gene duplication. In the second phase, during the evolution of early vertebrates, the Fgf gene family expanded through two large-scale gen(om)e duplications (reviewed in Refs. 40 and 44). Coevolution of Fgfs allowed for the development of increased ligand-receptor specificity, enabling the formation of preferred ligand-receptor interactions (40). The mechanisms that regulate alternative splicing of Fgfrs has been conserved since the divergence of vertebrates, enabling this signaling system to acquire diversity of function (45). The conservation of FGF/FGFR proteins throughout evolution and the extreme phenotypes of constitutive knockout models (2,5,46) are indicative that they have involvement in many developmental and physiological processes.

C. Structural properties of FGFs and FGFRs

FGFs have been cloned from many species, including mammals, fish, birds, amphibians, and worms. The most conserved sequence within the FGF proteins is located within a core of 120 amino acids, 28 of which are highly conserved and six that are identical amino acid residues (2,47). It has been shown that 10 of those highly conserved residues interact with the FGFRs to form functional receptor dimers (2,48). Within this region, FGF orthologs are 71–100% identical in amino acid composition, whereas the protein sequences of FGF paralogs are 22–66% identical (2). The prototypical FGF gene contains three coding exons, with exon 1 containing the initiation methionine. It is these three coding exons that encode the parallel β strands that fold into the distinct structural domain, the β trefoil (49). Regions involved in receptor binding are distinct from those regions that bind heparin (50,51,52,53). FGF gene organization is conserved in humans, mouse, and zebrafish, but its functionality is poorly understood. FGFs are known to exert their effects by binding to the four high-affinity FGFR proteins, although a putative fifth FGFR, FGFR5, has recently been cloned in mouse and human (7,54). This receptor lacks an intact kinase domain, and its function is undefined (7). Typically, the degree of identity between the receptors is 55–72% at the amino acid level (7).

D. Diversity of FGF signaling

Diversity in this signaling system is provided by multiple alternative splicing sites that generate receptor isoforms with altered ligand binding properties (13). For instance, FGF2 (basic FGF) interacts with the FGFR1 to -4 receptors, whereas FGF7 interacts only with a splice variant of FGFR2 known as KGFR or the FGFR2 IIIb isoform (2,55). Table 2 summarizes the specificity of the FGF ligands for specific FGFR isoforms. Binding of FGFs to their receptors induces FGFR dimerization, a step that is essential for subsequent receptor activation and triggering of downstream intracellular signaling events (13).

Table 2.

FGF/FGFR (ligand/receptor) interactions

FGF (ligand)	Interaction with receptors
	FGFR1	FGFR2	FGFR3	FGFR4
FGF1 (aFGF)	FGFR1: IIIb and IIIc	FGFR2: IIIb and IIIc	FGFR3: IIIb and IIIc	FGFR4
FGF2 (bFGF)	FGFR1: IIIb and IIIc	FGFR2: IIIc	FGFR3: IIIc	FGFR4
FGF3	FGFR1: IIIb	FGFR2: IIIb
FGF4	FGFR1: IIIc	FGFR2: IIIc	FGFR3: IIIc	FGFR4
FGF5	FGFR1: IIIc	FGFR2: IIIc	FGFR4
FGF6	FGFR1: IIIc	FGFR2: IIIc
FGF7		FGFR2: IIIb		FGFR4
FGF8	FGFR1a	FGFR2: IIIc	FGFR3: IIIc	FGFR4
FGF9		FGFR2: IIIc	FGFR3: IIIb and IIIc	FGFR4
FGF10	FGFR1: IIIb	FGFR2: IIIb
FGF11–14	?
FGF15	?
FGF16, 18	?
FGF17	FGFR1: IIIc	FGFR2: IIIc		FGFR4
FGF19, 21, 23	FGFR1: IIIc		FGFR3: IIIc	FGFR4
FGF20, 22, 24–25	?

Data were compiled from Refs. 2,4,5,13,41,64 and 66. aFGF, Acidic FGF; bFGF, basic FGF; ?, unknown. 

^aHave yet to identify FGFR isoform; i.e., IIIb or IIIc. 

The FGFR genes are characterized by up to three extracellular Ig-like domains, and alternative mRNA splicing of these Ig loops in FGFR1 through FGFR3 leads to distinct functional variants of the receptor (Fig. 1) (16,56,57). FGFR4 does undergo alternative splicing; however, its Ig domain is similar to a IIIc-like domain (58). The first and second loops of the Ig domains are separated by a region of acidic residues known as an acid box. This acid box is a unique feature of FGFRs and appears to be important in their function (59). The first FGFR1 cloned was found to contain three Ig loop domains in the extracellular domain, referred to as the long form (8). A second short form, in which the exon coding the first Ig loop domain was spliced out, was subsequently cloned (60,61,62). This same splicing phenomenon has also been demonstrated to occur within FGFR2 and -4 (58). Although the ligand binding region of FGFRs is typically defined by domains II and III, there is some biochemical evidence suggesting that domain I is also capable of affecting the affinity of ligand binding (63); however, a biological role for the different isoforms has yet to be shown.

Figure 1.

FGFR domain structure and isoforms generated by alternative splicing of FGFR transcripts. The main structural features of FGFRs are illustrated and include the signal peptide, the three Ig domains, acidic box, transmembrane domain, the JM domain/VT region and a split tyrosine kinase domain. The two forms of FGFR are generated by the alternative splicing of exons 8 and 9. The C-terminal portion of Ig domain III is encoded by exon 8 to generate the FGFR IIIb isoform, whereas the C-terminal portion of Ig domain III is encoded by exon 9 to generate the FGFR IIIc isoform (4,57,68,87,93).

The region often associated with determining the specificity of ligand binding is the carboxy terminal half of the Ig domain 3, or III loop. This region of the Ig domain III in FGFR1 to -3 can be derived from one of two exons, giving rise to the IIIb and IIIc forms, respectively (Fig. 1) (41,64,65,66). This splicing event does not occur within FGFR4 (58).These two alternative forms display different ligand binding characteristics. For example, it has been shown that the FGFR2 IIIb splice variant binds specifically to FGF7 but not FGF2, the ligand for FGFR2 IIIc (Table 2) (16). This alternative splicing event is regulated in a tissue-specific manner and dramatically affects ligand-receptor binding specificity (5,43). For example, it has been shown that the FGFR2 IIIb isoform is exclusively expressed in epithelial cells, whereas the FGFR2 IIIc isoform is expressed exclusively in mesenchymal cells. This lineage-specific expression of IIIb and IIIc FGFR isoforms enables interactions between the epithelial and mesenchymal layers during development in response to different FGFs (4). In addition to homodimerization, FGFRs are also able to heterodimerize, which adds another level of complexity for modulating responses to FGFs (63,67).

E. Heparin sulfate proteoglycans in FGF-FGFR ligand binding

FGFRs are also characterized by a heparin binding domain and a single transmembrane chain (transmembrane domain) that connects the extracellular part with the intracellular juxtamembrane (JM) domain (68). The JM domain is considerably longer in FGFRs than in other receptors (69). A split intracellular domain containing tyrosine kinase activity is also typical of these receptors (70). This kinase region is split in two by a short noncatalytic insert of approximately 15 amino acids (71) and contains two phosphorylatable tyrosine residues in FGFR1 and -2, one in FGFR3, and none in FGFR4 (72). The mitogenic potential of FGFR4 appears to be lower than that of the other receptors, which is due, in part, to the lack of the kinase insert tyrosine residues, with the receptor also containing what appears to be an intracellular cell adhesion domain (71,72).

An important feature of FGF biology involves the interaction between FGF and HSPG. The FGF-FGFR interactions and signaling are further regulated by the spatial and temporal expression of endogenous HSPG (43,73). Although the biologically active form of FGF is poorly understood, it has been established that heparin is required for FGF to effectively activate the FGFR in cells that are deficient in, or unable to synthesize, HSPG or in cells pretreated with heparin/heparin sulfate (HS)-degrading enzymes or inhibitors of sulfation (10). Biochemical and crystallographic studies have revealed the importance of HS sulfation for FGF signaling, which appears to be required for FGF1-FGFR2 and FGF2-FGFR1 interactions (15,43,74,75). In the absence of HS, FGFs can activate their receptors, but only at high concentrations, and in most cell systems that have been studied, FGFs do not induce cell growth in HS-deficient cells. The addition of HS to physiological concentrations of FGFs can increase their efficiency and also dictate specificity of FGF/FGFR binding depending on the fine structure of their sulfated domains (76). It has been shown that FGFR1 contains a 13-amino-acid-long heparin binding sequence that is required for the binding of FGF, suggesting that cell surface HSPGs directly interact with the tyrosine kinase FGFRs as well as FGF, forming a ternary complex before signal transduction commences (77).

Studies involving the syndecan family of cell surface proteoglycans, syndecan 1–4, have demonstrated that their expression pattern correlates with the distribution of several FGF family members during specific developmental stages (78,79). The members of this proteoglycan family have distinct expression patterns. Syndecan 1 is predominantly expressed in epithelial and mesenchymal tissues; syndecan 2 in cells of mesenchymal origin as well as neuronal and epithelial cells; syndecan 3 is almost exclusively expressed in neuronal and musculoskeletal tissues, whereas syndecan 4 is expressed in virtually every cell type (80,81). Given that every mammalian cell expresses at least one syndecan (82), it is probable that syndecans, FGFs, and FGFRs interact directly with one another to form a ternary structure that facilitates the signal transduction of the FGFR signaling pathway (83) (Fig. 2).

Figure 2.

Intracellular signaling activated through FGFRs. Formation of the FGF-Syndecan/heparin-FGFR complex leads to receptor autophosphorylation and activation of intracellular signaling cascades of the Ras/MAPK pathway, the PI3K/Akt pathway (left), and the PLCγ/Ca²⁺ /PKC pathway (right) (4,57,90,91). Syndecans facilitate the binding of FGFs to FGFRs, forming a ternary structure. EphA receptors induce tyrosine phosphorylation of FGFRs and FRS2 by forming a complex with the FGFR. The Ras/MAPK signaling cascade is activated by the binding of Grb2 to phosphorylated FRS2. The subsequent formation of the Grb2/SOS complex results in the activation of Ras. There are three routes by which FGFRs are able to activate the PI3K/Akt pathway: Gab1 can bind to FRS2 indirectly via Grb2, resulting in the tyrosine phosphorylation and activation of the PI3K/Akt pathway (2,96); PI3K can bind directly to a phosphorylated tyrosine residue of the FGFR (88); and lastly, activated Ras can induce membrane localization and activation of the PI3K catalytic subunit (57). The negative feedback signals that are mediated by or imposed on FRS2 are marked by the dotted line (94). The PLCγ/Ca²⁺ pathway is activated when autophosphorylation on a tyrosine residue in the carboxy terminal tail of the FGFR creates a specific binding site for the SH2 domain of PLCγ (91). The last activated components of the above-mentioned signal transduction cascades translocate into the nucleus and phosphorylate specific transcription factors of the Ets family, which in turn activate expression of FGF target genes. DAG, Diacylglycerol; IP3, inositol-1,4,5-triphosphate; SOS, son of sevenless.

FGFs 15 (the mouse ortholog of human FGF19), 19, 21, and 23 have recently emerged as a novel group of endocrine factors that have been shown to regulate diverse metabolic processes during adulthood (40). This subfamily of FGFs differ from other FGFs in that they require additional cofactors, besides HS to stabilize interactions with their corresponding FGFRs (84). The protein Klotho has been identified as a necessary cofactor for FGF19, -21, and -23/FGFR binding and activation (84,85,86).

F. Pathways of FGF signal transduction

FGFR signaling is mediated by direct recruitment of signaling proteins that bind to tyrosine autophosphorylation sites on the activated receptor and/or indirect recruitment of linked docking proteins that have become tyrosine phosphorylated in response to FGF activation (4,87). Like other RTKs, FGFR signal transduction results in the activation of several well-characterized intracellular signaling pathways, including: the direct activation of the MAPK and the phosphatidylinositol 3-kinase (PI3K) pathways, and the indirect activation of protein kinase C (PKC), via phospholipase C-γ (PLCγ) (4,88,89,90,91) (Fig. 2).

Studies have demonstrated that in the JM (68) region of alternatively spliced FGFRs, there may be deletions of Val423–Thr424, termed the VT region, and that this region may be phosphorylated by PKC (88). In functional studies, the activation of PKC significantly reduced the activity of the VT-containing isoform, VT+, while having little effect on the deletion isoform, VT− (88). The VT region also acts as a binding site for the phosphotyrosine binding domains of two members of the FRS2 family of docking proteins (87,92,93,94). FGFRs with mutations that result in the deletion of the VT region, VT−, fail to bind FRS2 (87,93). It has also been demonstrated that the ephrin A4 receptor (EphA4), a member of the largest family of RTKs, binds directly to the JM region of FGFRs, via the N-terminal portion of its tyrosine kinase core. In cells that express both EphA receptors and FGFRs, the interaction between these domains leads to receptor activation, resulting in the tyrosine phosphorylation of FRS2 and the subsequent activation of the MAPK transduction pathway (95). Recent studies have also established that the FRS2 family can modulate FGFR induction of the MAPK and PI3K pathways through the stimulation of a ternary Gab1/Grb2/FRS2 complex (90,95) (Fig. 2). It has therefore been hypothesized that through the VT region, FGFRs can modulate downstream cellular signaling, acting as a switch between the MAPK and PI3K signal transduction pathways (24,96,97,98).

G. Regulation of FGF signaling

The range of biological effects of FGFs and the variety of signaling pathways activated by this family imply that FGF signaling must be tightly regulated in regards to timing, duration, and signaling strength (43). FGF signaling can be regulated at many levels within the cell and in the extracellular space. There are a growing number of proteins that have been identified which specifically regulate the activity of FGF signals through the FGFRs. These proteins are part of what is known as the “FGF synexpression group.” Not only are they coexpressed with FGFs, but they themselves are also regulated by FGF signaling and inhibit FGF signals by establishing negative feedback loops (99,100,101,102) (Fig. 3).

Figure 3.

Regulation of FGF signaling. The stimulation of FGFR by FGF ligands results in the activation of specific downstream target genes. Among the targets of the FGF signaling pathway are several feedback inhibitors, which are involved in the attenuation of FGF signaling and are part of the FGF synexpression group. Spry acts at the level of Grb2 and/or the level of Raf. Sef and XFLRT3 are both located at the membrane and can interact directly with FGFR. Sef functions as a negative regulator, whereas XFLRT3 enhances FGF signaling resulting in the phosphorylation of the MAPK ERK. Sef has also been shown to affect the phosphorylation of the MAPK cascade, either at the level of ERK or by preventing the phosphorylation of ERK by MEK (99). MKP3 negatively regulates FGF signaling by dephosphorylation of activated ERK. P, Phosphorylation. [Adapted from Ref. 43 with permission from Elsevier.]

The first feedback regulator of the FGF pathway isolated was Sprouty (Spry). Gain- and/or loss-of-function experiments in the mouse, Xenopus, and zebrafish have demonstrated that Sprys antagonize FGF signaling (43,103,104,105,106,107) (reviewed in Ref. 43). In addition to the Sprys, several other genes, isolated through large-scale in situ hybridization screens, have been identified as regulators of FGF signaling (43,108,109). The XFLRT3 gene encodes a transmembrane protein that is characterized by a cluster of leucine-rich repeats and one fibronectin type III (FNIII) domain within the extracellular domain (110). Expression of this gene is induced after activation of FGF signaling, and its signaling results in phosphorylation of ERK and is blocked by the MAPK phosphatase 1 (110). The SEF protein is conserved across zebrafish, mouse, and human genomes. It was first identified as a gene coexpressed with FGF8 in embryogenesis (100). In zebrafish, loss- and/or gain-of-function studies revealed SEF functions as an antagonist of FGF signaling and interferes with FGF signal transduction by acting downstream of, or at the level of, mitogen effector kinase (MEK) and ERK (99,100). It has been proposed that human SEF acts as a spatial regulator for ERK signaling and blocks its translocation to the nucleus (111). MAPK phosphatase 3 (MKP3), a dual specificity phosphatase, negatively regulates the MAPK cascade through dephosphorylation of activated MAPK proteins (112,113,114). In chicken, it was shown that expression of MKP3 disrupted limb outgrowth, a phenotype that is characteristic of a disruption in FGF signaling (115,116). In zebrafish, it was demonstrated that expression of MKP3 limits the extent of FGF activity in the gastrula stage embryo (101). A more recent study also demonstrated the importance of MKP3 in the regulation of the FGF signal transduction cascade. This study showed that in mouse embryos, FGFR signaling is required for the transcription of Dusp6, which encodes MPK3 (117). Furthermore, it was demonstrated that targeted inactivation of MPK3 increases levels of phosphorylated ERK as well as the target transcription factor Erm (117). It was also observed that the mutant Mkp3 allele causes variably penetrant, dominant postnatal lethality, skeletal dwarfism, premature fusion of the cranial sutures (craniosynostosis), and hearing loss (117). All these phenotypes are characteristic of mutations that activate FGFR-meditated signaling pathways; however, they do not precisely mimic the results of any particular FGF gain-of-function mutation (117,118). Taken together, with the incomplete penetrance of this phenotype, the results suggest that MKP3 may function downstream of more than one FGFR and that there is probably some redundancy between MKP3 and other negative regulators of FGF signaling (118) (Fig. 3).

Regulation of FGF signaling has been demonstrated at different levels of the signal transduction pathways, from the membrane down to the levels of phosphorylation of the MAPKs. Several of these feedback loops are also known to negatively regulate other RTK signaling pathways (119,120).

III. FGFs and FGFRs in Sexual Differentiation

A. The bipotential gonad

The process of sexual differentiation, i.e., the development of the gonads, genital ducts, and external genitalia, produces fundamental differences between males and females. Unlike other organs in the developing embryo, the gonad forms with the potential to develop as one of two alternative organs, an ovary or a testis (121). Gonad development can be divided into two phases. The initial phase is characterized by the appearance of the “bipotential gonad,” or genital ridge, which is identical in both males and females (122,123). The second phase is characterized by the development of the gonad into a testis or an ovary (122).

B. Sry/Sox9 initiation of testicular development

In most mammals, sex is determined at the time of fertilization by the presence or absence of the Y chromosome: XY male or XX female (124). In all species, the developing gonad is comprised of a mixed population of both germ and somatic cells (125). Each cell has the potential to follow one of two fates, i.e., differentiate into either an ovarian or testicular cell. The somatic cells of the early gonad, the “supporting cells,” are the first to adopt a sex-specific fate (121). These cells differentiate either as Sertoli cells in XY gonads or as granulosa cells in the XX gonad (126,127). The expression of the Y-linked Sry gene initiates a cascade of events ultimately leading to testicular development (128,129,130). XY mice with no functional Sry gene develop ovaries, and the addition of Sry to XX mice triggers the testis development pathway (130,131,132). The direct targets of Sry are not known; however, Sry expression induces a subset of somatic cells to differentiate into Sertoli cells (122,133,134). Studies involving XY↔XX chimeras have shown that if a threshold number of Sertoli cells are present, they can recruit all other cells in the gonad (XX or XY) to the testis pathway (135,136). Sertoli cells then organize into cord-like structures that encase the germ cells. These cords later develop as the seminiferous tubules, which are the basic functional unit of the testis, responsible for the production, maturation, and export of sperm (137). Proper development of Sertoli cells is crucial because they are also thought to act as the control center of the male gonad, inducing the differentiation of all other cell types of the testis such as the Leydig cells that produce the masculinizing hormones (137).

Logical downstream targets of Sry would include intracellular factors that influence the development of Sertoli cells and secreted signaling factors that exert a paracrine effect on the surrounding cells (136). One such factor is Sox9. Transient expression of Sry usually results in the up-regulated expression of Sox9, which is continually expressed thereafter (121,138,139). Sox9, like Sry, is a member of the Sry-box (SOX) family of transcription factors. It is expressed at low levels in the “bipotential gonad” of both sexes and becomes dramatically up-regulated in the Sertoli cell immediately after the onset of Sry expression (140). It is not known whether Sox9 responds directly or indirectly to Sry; however, gain-of-function studies in mice and humans (reviewed in Ref. 141) have indicated that Sox9 is sufficient for male sex determination even in the absence of Sry (137). Heterozygous mutations in human SOX9 have been associated with the bone dysmorphology syndrome campomelic dysplasia, a disorder marked by cartilage and bone defects, XY sex reversal, and abnormalities of the heart, kidneys, brain, gut, and pancreas.

C. FGFs in Sertoli cell differentiation

Other extracellular factors are required for male-specific proliferation. The role of the signaling factor Fgf9 in sex determination is one of the most well studied. It is the only known case where the targeted deletion of a single gene resulted in a failure of Sertoli cell differentiation and complete male-to-female sex reversal (121,125,136,142). Analysis of the interactions between Fgf9, Sry, and Sox9 (as reviewed in Ref. 121) has shown that Fgf9 forms a regulatory network in conjunction with Sry and Sox9 during the early stages of Sertoli cell specification. In the absence of Fgf9, Sox9 expression is not maintained, despite normal expression of Sry and the initial up-regulation of Sox9, and progression along the male-specific pathway is aborted before Sertoli cell differentiation and testis cord formation can commence (121,125). These data strongly suggest that expression of Sry alone is not sufficient to establish the male-specific pathway. Remarkably, in this failing environment, the somatic cells do not die; instead, they undergo cell fate transitions (121). In Fgf9^−/− XY gonads, it was further found that Wnt4 expression, which is required for ovary development, was up-regulated and that other ovarian markers, which are normally expressed downstream of Wnt4 in XX gonads, were also up-regulated (121,142,143,144). This led to the idea that there is an antagonistic balance between WNT4 and FGF9 that holds the bipotential gonad between its two alternative fates (121). In XY gonads, Sry changes the antagonistic balance between WNT4 and FGF9 by inducing the up-regulation of Sox9. Sox9 then up-regulates the expression of Fgf9, which is required for the maintenance of Sox9 expression. This creates a positive feedback loop that leads to the predominance of Fgf9 signals in the gonad, the stabilization of Sox9 expression, repression of Wnt4, and commitment of the gonad to the testis pathway (121,123) (Fig. 4).

Figure 4.

Opposing signals regulate sex determination in the bipotential gonad. In both XX and XY gonads at 10.5–11.5 d postcoitum (dpc), Fgf9 transcripts (blue) are detected near the gonadal surface, whereas Wnt4 transcripts (pink) are detected near the gonad mesonephric boundary (123). At this time, in the bipotential gonad, there is a balance between these two competing signals. A genetic or environmental switch initiates the male pathway by creating an imbalance between these signals. Sry is activated in the XY gonad, and its expression diverts the XY gonad toward a testis-specific fate. Sry up-regulates Sox9, which has been implicated in the early steps of the male-specific pathway (121,123). Sox9 then up-regulates Fgf9, and Fgf9 (through FGFR2) maintains Sox9, forming a positive feedback loop in XY gonads (125). In this circumstance, the balance between FGF9 and Wnt4 signals is shifted in favor of FGF9, silencing Wnt4 signaling, and the dominance of the male pathway is established. In the absence of a feed-forward loop between SOX9 and FGF9 (e.g., in XX gonads), WNT4 suppresses Fgf9 transcription, initiating the female differentiation pathway.

The direct mechanism by which extracellular signals from Fgf9 influence the sex-specific fate of the gonad is poorly understood. Fgf9 is expressed in Sertoli cell precursors and can act as a paracrine factor through FGFRs to induce the expression of Sox9 (125,136,142). Interestingly, a recent study has identified Fgf13 and Fgf18 as new candidate genes for involvement in testes differentiation and pre-Sertoli cell function; however, the exact nature of this involvement requires further investigation (145). Immunostaining using an antibody raised against FGFR2 detected a sexually dimorphic expression pattern that colocalizes with Sox9 in the nucleus of Sertoli cells within the XY gonad (136), suggesting that this receptor plays an important role in the Fgf9 regulation of sex determination. In addition, it has been found that the cyclin inhibitor p21^Waf1/Cip1 (cyclin-dependent kinase inhibitor 1A) protein, in association with the transcription factor E2f1, binds to E2f1 response elements in the Wnt4 promoter, repressing Wnt4 expression in keratinocytes (146,147). E2f1 has been shown to be a strong activator of Fgfr2 gene expression (16,148). One might speculate that in the male gonad, Sry triggers Sox9 expression which, in turn, up-regulates Fgf9 expression. The concentrated action of these two proteins will directly or indirectly, through E2f1, activate FGFR2, which creates a feedback loop that leads to Wnt4 repression allowing Sertoli cell differentiation to proceed, committing the gonad to a testis fate (121,147) (Fig. 4). Although the importance of FGF9 signaling for sexual differentiation has been established, it has yet to be determined whether FGF9 signaling utilizes the MAPK pathway or other intracellular signaling pathways to exert their influence on the bipotential gonad. For more comprehensive reviews on the sexual differentiation of the bipotential gonad, see Refs. 121 and 123.

D. SPROUTY2 inhibition of FGF signaling during sexual differentiation

Recently, it has been demonstrated that a member of the FGF synexpression group influences FGF signaling during sexual differentiation. Spry2 expression was found to be confined to the developing testis and mesonephros tubules at the early stages of organogenesis. This study also determined that a gain of SPRY2 function in the embryonic testis and mesonephric tubules disrupted the formation of the seminiferous tubules and epididymis and reduced the amount of testicular interstitial tissue. It was determined that this phenotype was probably caused by an inhibitory effect on the migration of mesonephric cells into the testis that is promoted by FGF9 signaling (149). It is known that Sprouty-controlled signal transduction in developmental systems can inhibit FGF-induced ERK activation. SPRY2 expression was found to reduce the intensity of activated ERK in the embryonic testis and Wolffian duct (149). Thus, Spry2 signaling could contribute to normal testis development by controlling the male-specific action of FGF signaling (149).

IV. FGFs and FGFRs in Male Germ Cells

The testes are paired organs found within the scrotum, whose main function is the production of male germ cells, sperm, and hormones, all of which are essential for male reproduction. Fully functional sperm cannot develop independently of the testis and rely heavily upon the unique environment provided by the testis. The testis can be divided into two major compartments: the seminiferous tubule and the interstitial compartment (150). Spermatozoa are produced in the seminiferous tubules by a process termed spermatogenesis and are supported by the Sertoli cell, which extends from the basement membrane of the tubule to the lumen (151,152,153). Sertoli cells are also the principal source of estradiol production in the immature testis (154). The Leydig cells, which are contained within the interstitial compartment (151), also play an important role in sperm development by undertaking steroid biosynthesis, producing testosterone, a hormone that is crucial for spermatogenesis (155).

A. FGFs in the testis

FGFs have been localized to many cells throughout the testis, including Sertoli, Leydig, and germ cells (18,19,20). Germ cell localization included type A-spermatogonia, primary pachytene spermatocytes, and elongating spermatids (20,22,156). In fetal testes, it was demonstrated that FGF2 is a survival factor for Sertoli cells and a mitogenic factor for gonocytes (157), whereas FGF1 and FGF2 modulate Leydig cell steroidogenesis (55,158). FGF9 signaling has been shown to stimulate mesenchymal cell proliferation and migration of mesonephric cells into the testis during development, contributing to the formation of the interstitial compartment of the testis (142). The mRNAs for FGFs 4, 5, and 8 were also found to be expressed in the testis (25). Other studies have also suggested that FGFs are involved in the proliferation and differentiation of testicular cells and may also be involved in the regulation of spermatogenesis (19,159).

B. Differential expression of FGFR splice variants

The specificity of FGFs for each FGFR variant was determined using a mitogenic cell assay (16). Previous studies had already determined the expression pattern of FGFR1 to -4 during rat spermatogenesis (25). Briefly, it was observed that Sertoli cells were immunoreactive for FGFR4, whereas FGFR1 to -4 were present in germ cells, but each receptor was localized to specific stages of germ cell development. Although not commented on in the original paper, FGFR1 and -4 were present on the tails of elongating spermatids, suggesting a potential role for the FGF signaling cascade in sperm tail development or function (160,161).

A study by Cancilla et al. (25) identified the differential expression of FGFR variants and FGF ligands in fetal, immature, and adult rat testis. This study showed that five of the seven functional FGFR variants were expressed (25). FGFR1 had, in previous studies, been localized to the seminiferous epithelium and interstitium (21,156); however, it has been demonstrated that both variants of FGFR1, IIIb and IIIc, are present in the testis as well as in adult Leydig cells (25).

FGFR2 protein was immunolocalized to the spermatocytes and spermatids of immature and adult testis, as well as Leydig cells (21). The variant detected was FGFR2 IIIc (25). There are many FGF ligands that can signal through FGFR2 IIIc. The FGFR3 receptor variant present in the testis was FGFR3 IIIc (25), and it acts as a receptor to a similar range of ligands as FGFR2 IIIc (16). FGFR4 was localized to the Sertoli cells, Leydig cells, spermatogonia, spermatocytes, and elongating spermatids. FGFR4 is activated by FGFs 1, 2, 4, and 8, all of which are present in the testis (25).

Little is known about FGFR5 in terms of expression and regulation during testis development. Northern analysis of Fgfr5 expression, during mouse testis development, showed that it was strongly expressed during the early stages of development, in likely association with spermatogonia, and greatly diminished during the later stages of development. These data are consistent with the massive dilutions of spermatogonia and their mRNA as spermatogonia proceed to maturity. This suggests that FGFR5 is unlikely to be involved in any FGFR signaling that is occurring during the later stages of sperm development (L. M. Cotton and M. K. O’Bryan, unpublished observations).

C. FGF signal transduction in the acquisition of sperm motility

A large body of evidence has shown that human sperm capacitation is regulated by multiple signaling pathways that control sperm motility and tyrosine phosphorylation of sperm proteins (162,163). Components of the MAPK and PI3K pathways, signaling cascades downstream of FGFRs, are known to be present and active in sperm. It has been demonstrated that although transcription has ceased during spermatogenesis, the necessary components for the MAPK signaling pathway, b-raf, MEK1, and ERK, are incorporated into the sperm (27,160,164). It has been observed that the two forms of MAPK, ERK 1/2, are expressed in similar quantities in all spermatogenic cells at different stages; however, evidence of phosphorylation was only detected in early spermatogenic cells, from primitive spermatogonia to zygotene spermatocytes, suggesting that MAPK activation may contribute to the mitotic proliferation of spermatogonia and the early phases of meiosis (160). The first evidence for the potential role of MAPK in sperm function was provided when a MAPK substrate peptide was added to fowl sperm and a subsequent decrease in motility was observed (28). Furthermore, both ERK 1 and 2 are tyrosine phosphorylated in a time-dependent manner during in vitro capacitation of human sperm (27). The addition of a specific MAPK inhibitor, PD098509, to sperm was shown to inhibit ERK 1/2 activation; however, it did not affect the progesterone-induced acrosome reaction (29), suggesting that the MAPK pathway does not appear to be directly involved in acrosome reaction induction. A recent study by Luconi et al. (165) demonstrated that inhibition of PI3K with LY294002, a specific and cell-permeable inhibitor of PI3K, resulted in an increase of intracellular cAMP, tyrosine phosphorylation of A kinase anchoring protein 4, and recruitment of PKA to the sperm tail.

The presence of the adaptor protein FRS2, which binds to the VT region of the FGFR, and the identification of FGFR1 protein in association with sperm tails (25,30) suggested that the FGF/FGFR signal transduction pathway affects sperm function and/or development. A recent study has demonstrated that compromised FGFR1 signaling during spermiogenesis resulted in a significant decrease in daily sperm production and that sperm produced were functionally compromised in terms of their ability to undergo capacitation (30). Furthermore, this study showed the presence of functional FGFR1 signaling complexes on sperm and suggested that signaling occurred through the PI3K pathway. It was also suggested that FGFR1 signaling serves to suppress the MAPK pathway and prevent premature capacitation (30). The above data illustrate definitive roles for the FGFR1 signaling cascade in the establishment of normal spermiogenesis and male fertility.

D. FGFs and testicular cancer

The incidence of malignant tumors of the testis is on the rise in industrialized countries (166). Because FGFs are known to be involved in embryonic testis development and differentiation, several studies have attempted to elucidate the role of FGFs in testicular cancer. A recent immunohistochemical study on primary testicular germ cell tumors showed expression of FGF4, -8, and FGFR1 in nonseminomatous and highly proliferative components of the tumors (166,167). Analysis of human teratocarcinoma cells (Tera-2) in vitro showed that undifferentiated cells expressed mRNAs for FGFRs 1–4 (168). The induction of differentiation in Tera-2 cells by retinoic acid led to the loss of FGFR4 expression and to the down-regulation of FGF2 and -3 transcripts, although the FGFR1 mRNA levels remained unaltered (168). The down-regulation of FGFRs resulted in a significant reduction in the number of FGF4 binding sites (168). Further experiments with Tera-2 cells have shown that low concentrations of FGF2 stimulated cell proliferation, whereas high concentrations of FGF2 promoted cell migration (169). These results, together with recent clinical and in vitro observations, suggest that FGF4 and FGF8, and to a minor extent FGF2, are involved in testicular cancer (166).

V. FGFs and FGFRs in the Epididymis

The epididymis is a single convoluted tubule that is responsible for the protection of spermatozoa as they transit from the testis (via the rete testis and efferent ducts) to the cauda epididymis for storage. The main function of the epididymis is to provide a specialized microenvironment that allows for sperm storage, maturation (acquisition of motility and fertility), protection, and transport (170,171,172,173). One of the unique characteristics of the epididymis is its segmental nature, which develops during the postnatal period, before the appearance of spermatozoa in the epididymal lumen (174). The segmentation of the epididymis is illustrated by changes in morphology, function, and gene expression profiles of the cells along the length of the epididymal tubule. Many genes expressed in the epididymis show a defined pattern of expression during development and adulthood and are often confined to a particular region of the epididymis (175,176,177). The primary factor regulating epididymal function is androgens. However, there is now a large body of evidence suggesting that estrogens, retinoids, and other factors, such as growth factors, coming from the testis via the efferent ducts play important roles in the regulation and maintenance of the epididymis (178).

A. Lumicrine regulation of epididymal function

Growth factors that are produced in the testis reach the epididymis via rete testis fluid (RTF) where they influence receptors on the surface of epididymal epithelial cells. These growth factors are necessary for normal epididymal function, particularly within the initial segment, which has been shown to be more susceptible to luminal testicular factors. A fully developed and functioning initial segment has been shown to be important for male fertility (179). In 1983, Nicander et al. (180) proposed that a mitogenic factor of Sertoli cell origin, and not an androgenic factor, was responsible for preventing the cells of the initial segment from undergoing apoptosis. Since then, it has been established that efferent duct ligation (EDL), which causes the loss of testicular luminal fluid factors, results in the down-regulation of many genes, a decrease in protein synthesis and secretion, and the eventual apoptosis of a certain population of cells within the initial segment (180,181,182,183,184,185,186,187,188). No effect of EDL was observed in the remaining segments of the epididymis (180,185,188). Apoptosis is observed in the rat initial segment within 18–24 h after loss of testicular luminal fluid factors. The population of cells that undergo apoptosis is specifically located in the very proximal region of the initial segment (188). Thus, it appears that factors, which originate from Sertoli and/or germ cells, enter the lumen of the seminiferous tubule, pass out of the testis via the rete testis and efferent ducts, and enter the epididymal lumen where they influence receptors on the apical surface of initial segment epithelial cells. Here, second messenger pathways are activated, resulting in the activation of transcription factors and the transactivation of genes. This mode of regulation has been coined “lumicrine” regulation and suggests that growth factors, including FGFs and neurotrophins, are the putative luminal regulators of initial segment function (189).

It should be emphasized that regulation of the initial segment by this means is quite novel. Under normal physiological conditions, the cells of the initial segment are continuously stimulated by growth factors. This contrasts with other cell types that only activate growth-factor messenger pathways during periods of growth and differentiation. Epithelial cells of the initial segment only undergo proliferation and differentiation during embryonic and postnatal periods, not during adulthood, despite being continuously exposed to these growth factors.

B. FGFs as lumicrine regulators

The first piece of evidence that FGFs are putative luminal factors, regulating initial segment function, came from a study showing that, after EDL, the activity of the enzyme γ-glutamyl transpeptide (GGT) in the rat initial segment was decreased. However, if the tissue was treated with FGF2, the activity of the enzyme returned to control levels. Treatment of this tissue with epidermal growth factor gave no effect (23). Another study revealed that an ets transcriptional factor, polyomavirus enhancer activator 3 (PEA3), is involved in the transactivation of expression of GGT IV mRNA in the rat initial segment (190). The DNA-binding and transactivation activity of PEA3 has also been shown to be under the regulation of growth factors that are present in rat RTF (23), via the stimulation of the MAPK signal transduction pathway (190,191,192,193). The presence of FGFs 2, 4, and 8 has been confirmed in RTF, entering into the initial segment of the epididymis (23,24), and are potential ligands for the apically placed FGFRs (S. A. Crenshaw and B. T. Hinton, unpublished observations). Furthermore, it has been demonstrated that Fgfrs 1 to 4 are expressed in the initial segment of the rat epididymis, with both the IIIb and IIIc receptor isoforms of Fgfrs 1 to 3 being expressed (24). Fgfr1 IIIc has been shown to be present in the principal cells of the epididymal epithelium, cells that are responsible for GGT mRNA IV expression (23,24). The remaining Fgfr splice variants are thought to localize to the endothelial and interstitial cells of the epididymis (178).

C. Differential expression of Fgf/Fgfrs

Data collected from the mouse epididymal transcriptome (MET) database, which details gene expression through the segments of the mouse epididymis, demonstrated differential expression of Fgf and Fgfrs (175,194). For example, the database showed that Fgfr2 was more abundantly expressed in the cauda epididymis, whereas Fgfr1 was moderately expressed throughout the entire epididymis. Other potential components of the FGFR signaling pathway have been found to be expressed in the epididymis. Again, from data collected using the MET database, it was found that Syndecan 1 was highly expressed in the more proximal regions of the epididymis, with its transcripts being most abundant in the initial segment. Syndecan receptors 2 to 4 were also expressed throughout the epididymis, at significantly lower levels however (175). Furthermore, EphA1 and -4 receptors were also found to be expressed in the epididymis, with EphA4 demonstrating a higher expression level in the initial segment (175). These data suggest that Fgfs/Fgfrs, Syndecan 1, and EphA4 all interact directly with one another forming a complex that facilitates signal transduction of the FGFR pathway in the initial segment of the epididymis. The MET database is incorporated into the Mouse Reproductive Genetic Databases (MRG) and is available online at www.mrg.genetics.washington.edu. The MRG databases are an excellent resource for those wishing to investigate gene expression in the testis and epididymis.

Failure to maintain the constitutive stimulation of FGF signaling, via testicular luminal fluid growth factors, will result in apoptosis of the cells of the initial segment and ultimately lead to male infertility. Despite all the evidence demonstrating that the FGF signaling pathway is important throughout the epididymis, more functional information is required to understand the precise biological role of these growth factors in the development and regulation of the epididymal environment.

VI. FGFs/FGFRs in Seminal Vesicle Development

The prostate and seminal vesicles are the major sex-accessory glands that produce components of the seminal plasma in mammals (195). Seminal vesicles are present in many mammals, including mice and men; however, they are completely absent from some, such as monotremes, marsupials, carnivores, and cetaceans (195). In some mammals, such as the guinea pig, although the seminal vesicles are present, they appear as unbranched tubular glands rather than the highly branched structure found in men (196). Branching morphogenesis is a key feature of organogenesis of both the seminal vesicles and the prostate. It allows for an increased epithelial surface area for seminal plasma secretion, and it provides a greater storage capacity of seminal plasma than is found in vertebrates that lack these homologous glands (197).

A. Branching morphogenesis

Branching morphogenesis is a developmental process that is common to all organisms throughout the animal kingdom (198). In mammals, the kidney, lung, pancreas, and glands including mammary, prostrate, and seminal vesicles undergo branching morphogenesis during their development. Organogenesis of these branched organs can be divided into five steps: organ specification, epithelial bud initiation, epithelial duct elongation into the mesenchyme, bifurcation of the ducts leading to complex branching patterns, and cellular differentiation of the newly branched structure (199,200). Branching morphogenesis of the seminal vesicles is less complex than that of the prostate (201). The initial seminal vesicle buds from cane-shaped tubes off the Wolffian ducts before birth, after which the initial tubes develop lateral branches that elongate and often undergo secondary branching. Branching morphogenesis of the seminal vesicles is usually complete by 2 wk after birth in mice, whereas the prostate continues to generate new branch points until puberty is completed (202).

Insight into the mechanisms controlling seminal vesicle and prostatic branching morphogenesis has come from a range of experimental embryological work as well as from the study of mice and humans harboring mutations that alter branching morphogenesis. These studies demonstrated a requirement for androgens to initiate branching morphogenesis as well as a role for androgens in sustaining the normal rate and extent of branching (203,204). In addition, these studies have revealed a series of reciprocal paracrine signals between the developing prostatic epithelium and prostatic mesenchyme that are essential for regulating branching morphogenesis. Key growth factors that participate in these signaling events include members of the FGF, hedgehog, and TGF-β families (56,197,205,206,207,208,209,210,211,212,213).

B. Seminal vesicle shape (svs) mutation

The mouse svs mutation is a spontaneous recessive mutation that, during seminal vesicle and prostrate development, causes branching morphogenesis defects (214,215). Unlike other spontaneous and engineered mutations that reduce both organ growth and branching simultaneously, the svs mutation dramatically reduces branching without reducing organ growth (200,216,217,218). This unique feature of the svs mutation has provided an opportunity to investigate the molecular mechanisms that control branching morphogenesis in seminal vesicles and the prostate. The svs mutation has been mapped to a 2.7-cM segment of mouse chromosome 7, which is known to contain the Fgfr2 locus (214).

C. FGFs in the regulation of seminal vesicle development

Initially it was thought that Fgfr2 was the gene affected by the svs mutation. However, no sequence changes were detected in the open reading frame, and there were inconsistencies between the phenotypes of known Fgfr2 mutations and the svs mutant mice (219). Loss-of function Fgfr2 mutations caused agenesis of organs throughout the body and often resulted in embryonic lethality, whereas only branching defects were detected in the svs mice (114,200,219,220,221,222,223). Furthermore, gain-of-function mutations in Fgfr2 are known to cause several human diseases, including Crouzon, Jackson-Weiss, Apert, and Pfeiffer syndromes (222,224,225), the phenotypes of which are not observed in the svs mutant mice (200). Utilizing a positional cloning approach, it was determined that the svs mutant lesion was a 491-bp insertion into the 10th intron of Fgfr2 that resulted in changes in the pattern of Fgfr2 alternative splicing, from Fgfr2 IIIb to Fgfr2 IIIc forms (200).

It has been previously demonstrated that FGF7 and FGF10 were expressed in the mesenchyme of the developing seminal vesicles and act via FGFR2 IIIb, which is expressed in the developing epithelium. The addition of recombinant FGF7 or FGF10 stimulated both the growth and branching of developing seminal vesicles in vitro (207,208,209). The requirement of FGF10 for seminal vesicle development was illustrated by Fgf10-null embryos that showed degeneration of the caudal segments of the Wolffian ducts, which are the precursor structures of the seminal vesicles (218). In addition, grafting of the caudal Wolffian ducts, from rare Fgf10-null embryos in which the duct did not degenerate, revealed that Fgf10-null embryos had a limited ability to develop seminal vesicles, with only one in eight of the grafted Wolffian ducts resulting in a tissue that resembled an immature seminal vesicle (218). The fact that seminal vesicles exhibit no change in gland size in the svs mutant mice (214), whereas the Fgf10-null embryos exhibit a dramatic loss of growth, suggests that FGF10 can partially signal through the reduced levels of FGFR2 IIIb still present in the svs mutant glands and that this is sufficient to support gland growth (214). Branching morphogenesis fails completely in the svs seminal vesicles (214); this suggests that peak levels of FGF10 signaling through FGFR2 may be required to induce branching because partial loss of FGFR2 IIIb in the svs mutant mice blocked the branching of the vesicles (200). FGFR2 IIIb activates several downstream signaling pathways, one of which, the MEK1/2-ERK1/2 (or MAPK) pathway, has been shown to be important during seminal vesicle branching morphogenesis (214). The seminal vesicles of the svs mutants failed to maintain activation of the MAPK pathway during branching morphogenesis, despite similar levels of FGFR2 protein expression in wild-type and mutant vesicles. This suggested that the loss of activation of this signaling pathway was due to the shift in Fgfr2 alternative splicing. Because FGF7 and FGF10 are thought to be the ligands that activate FGFR2 during normal seminal vesicle development (26,209), and it is known that these ligands cannot signal through the FGFR2 IIIc isoform (2,5), this partial loss of the FGFR2 IIIb isoform may explain the loss of MAPK activation in svs mutant seminal vesicles (214).

Branching morphogenesis of the seminal vesicles and prostate is unique among other branched organs because of the critical role played by endocrine hormones, especially androgens. Currently, the list of important regulators of branching morphogenesis in the seminal vesicles (and prostate) include members of the Hedgehog, TGFβ, and FGF superfamilies of growth factors as well as several homeobox-containing transcription factors. Further studies are required to determine exactly how these families of genes interact in regulating the mechanisms of organogenesis and branching morphogenesis in seminal vesicles and the prostate.

VII. FGFs and FGFRs in the Prostate

The prostate is the largest accessory gland of the male reproductive system; the fluid and enzymes produced by the prostate contribute to semen and are required for male fertility. The prostate is not found in nonmammalian vertebrates, and among mammals there are dramatic variations in morphology and gene expression across species that arise from differences in bud induction, tubule elongation, and branching morphogenesis (195,197). There are at least three major variations in morphology. Multilobed, highly branched prostates like those found in rodents are also found in lagomorphs and insectivores. A disseminate prostate, in which glandular acini remain within the lamina propria around the urethra, is found in some marsupials and edentates, sheep, goats, hippopotami, and whales. Other species including men and dogs have solid, compact prostates. The adult human prostate lacks distinct lobes and is commonly described in terms of three zones that reflect the three separate sets of branched ducts present in the human prostate (226).

The prostate surrounds the urethra close to the base of the bladder and consists of a ductal gland that is highly secretory (26). It is a mixed epithelial and stromal organ that requires both androgenic stimulation and mesenchymal-epithelial interactions for its development, maintenance, and growth (26,227,228). The action of androgen in mesenchymal cells results in the proliferation of epithelial cells, and it is the accepted paradigm that this proliferation is mediated by paracrine factors made in the mesenchyme (26,229). The identity of these factors is unknown, although members of the FGF signaling family are considered excellent candidates (26).

A. FGF signaling in the developing prostate

In the prostate, depending on the species under investigation, the types of FGFs identified and evaluated vary. FGF signaling systems have been extensively studied in rodent prostate growth and branching using in vitro organ culture and knockout mice.

Male and female urogenital sinus (UGS) development diverges shortly after the fetal testes begin the production of testosterone (216). In males, testosterone specifies the fate of the UGS as prostate; this occurs in mice around embryonic day 15.5 (230). At this stage, the androgen receptor is localized in the mesenchyme of the UGS and suggests a paracrine mesenchyme-to-epithelial signal for prostatic induction (231). Androgens are necessary and sufficient to specify the UGS as prostate as shown by the absence of a prostate in mice lacking a functional androgen receptor and by the induction of prostates in female UGSs treated with androgens (232,233,234). Given that prostatic growth and development are regulated by the action of androgens in the mesenchyme, a significant effort has been put forward to identify paracrine-acting growth factors that mediate the effects of androgens. Although the process of identifying these factors has remained a challenge, Fgf7 and Fgf10 have been proposed to be candidate stromal-to-epithelial cell “andromedins” (235).

An andromedin, by definition, must be regulated directly or indirectly by androgens, its absence must result in a failure of the prostate to form, and it must be sufficient to induce prostatic development (218,234). Initially, Fgf7 and Fgf10 were proposed as andromedins because they are synthesized and secreted by mesenchymal cells, their cognate receptors are localized to epithelial cells, and in vitro studies showed that they were stimulated by androgens (209,210,218,234,236,237). This proposal was challenged by subsequent studies indicating that these Fgfs were not directly regulated by androgens in vivo (209). Analysis of Fgf knockout mice revealed that Fgf7 null mice showed no prostate abnormalities, whereas Fgf10 null mice, which die at birth, did not initiate prostate development. Treatment of cultured Fgf10 knockout UGSs with exogenous FGF10 alone did not induce prostatic bud induction; however, treatment of the cultured UGSs with exogenous FGF10 and testosterone partially restored prostatic bud induction (218). These experiments taken together provided more evidence for Fgf10 being an andromedin, mediating androgen action in the developing prostate (208,218,238). A recent study, utilizing tightly controlled rat organ culture conditions, demonstrated a consistent and significant up-regulation of Fgf10 expression by testosterone in both the ventral (VP) and lateral (LP) lobes of the prostate (235). However, it was also observed, as before (209), that testosterone is not required for Fgf10 expression because expression continues in the absence of androgens, although at a much lower level (235). This study suggests that Fgf10 is a critical, and possibly direct, target of testosterone action in the rat VP and LP (235). Furthermore, the investigators proposed that Fg10 up-regulates genes in the VP epithelium in response to testosterone stimulation, and as such, Fgf10 may be considered a mesenchymal-to-epithelial andromedin specific to that region of the prostate (235).

Observations of prostatic development, in both rodents and humans, have demonstrated that branching morphogenesis begins in a similar manner with several distinct sets of epithelial buds growing out of the UGS into a surrounding mass of mesenchyme (197,239). A number of different signaling families have been implicated in the regulation of branching morphogenesis in the prostate (212). In the prostate, as with seminal vesicles, Fgf10 expression is spatially restricted to the distal aspects of the gland, where it is thought to act as an inducer of ductal elongation and branching through stimulation of epithelial cell proliferation (209,210,218). Recombinant FGF7 or FGF10 has been shown to stimulate both growth and branching morphogenesis of the prostate in vitro, acting at least in part as proliferative signals for the epithelium (200,207,209). The requirement for FGF10 in prostate development was confirmed by experiments showing that Fgf10 null embryos develop minimal rudimentary prostatic organs (200,218). Analysis of the svs mutant mice, which have a mutation that resulted in changes in the pattern of Fgfr2 alternative splicing from Fgfr2 IIIb to Fgfr2 IIIc (see Section VI), demonstrated that branching morphogenesis was reduced by approximately 40% in svs prostates (214). This result suggests that maximum levels of FGF10 signaling through FGFR2 are required to induce branching in the prostate (200). FGFR2 IIIb activates several downstream signaling pathways, one of which, the MAPK pathway, has been implicated in prostatic budding and branching morphogenesis by several groups (210,240,241). A recent study demonstrated that ERK 1/2 was rapidly activated in the UGS by recombinant FGF10 in a dose-dependent manner, and that inhibiting FGFR activity ablated a majority of activated ERK 1/2 in the UGS, suggesting that FGF10 signals through the MAPK pathway in the UGS (234). The same study also showed that inhibition of FGFR or ERK 1/2 activation in the UGS blocked all androgen-induced changes, including epithelial budding, proliferation, and gene expression changes. From these findings it was concluded that FGFR-MAPK signaling is a critical part of prostate bud induction (234).

Androgens do not act exclusively through the FGF10-MAPK pathway because recombinant FGF10 does not induce UGS budding in the absence of androgens (218). However, it is possible that androgens activate the FGFR-MAPK pathway by induction of another FGF ligand, other than FGF10, or by modifying the extracellular environment between the mesenchyme and the epithelium to facilitate FGFR-MAPK signaling (234). Clearly, FGF and androgen signaling are important in prostatic development, although their exact mechanism of interaction remains unclear.

B. FGF signaling in the adult prostate

Adult prostates are androgen-dependent organs with respect to growth, tissue homeostasis, and function (242). As previously mentioned, FGF family members are partitioned in the epithelium and mesenchyme, mediating directional and reciprocal communications between these two compartments (242). Removal of this two-way communication in adult prostates disrupts tissue homeostasis and leads to prostatic intraepithelial neoplasia and progressively more severe lesions (243,244).

In normal human prostate FGF2, -7, and -9 are all produced by prostatic mesenchymal cells in biologically significant quantities (245). FGF2 does not contain a classical signal peptide; however, studies have revealed that this peptide is capable of being actively transported (inefficiently) across the plasma membrane (58). FGF7 is actively secreted, and its expression by mesenchymal cells can be regulated, at least in part, by androgens, although other factors are also likely to be involved (229,246). FGF9, like FGF2 and -7, is also expressed by prostatic mesenchymal cells and is actively secreted as a growth factor by both prostatic mesenchymal and epithelial cells in culture (229,247). In normal adult tissues, FGF8 is expressed in a very restricted manner or at very low levels (248). FGF10 is expressed by prostatic mesenchymal cells; however, it is present at low levels (209,249). In addition to these mesenchymal FGFs, there are epithelial-derived FGFs present in the normal prostate. FGF6 is present in the basal cells of the normal prostate, and FGF17 has also been shown to be expressed in small quantities by prostate epithelial cells (250,251). The expression of other FGFs in the human prostate has been evaluated, and it was found that they are expressed at much lower levels than FGF2, FGF7, and FGF9 (reviewed in Ref. 229). Changes in expression of FGF1, FGF2, FGF6, FGF8, FGF9, and FGF17 have been observed to be associated with prostatic lesions (refer to Section VII.C).

The epithelial cells of the prostate express multiple FGFRs. FGFR1 and FGFR2 are expressed in the basal epithelial cells of the prostate; however, they are not expressed in luminal cells (245). Studies of primary prostate epithelial cell cultures established that FGFR1 was present as the IIIc isoform, whereas FGFR2 was present as the IIIb isoform (252). FGFR3 isoform IIIb was also found to be predominantly present in the prostatic epithelium, whereas FGFR4 expression was found in the luminal prostatic epithelial cells (253,254). According to the known properties of FGF ligand and FGFR interactions (see Table 2), it can be concluded that prostatic epithelial cells express the required receptors to interact with the FGF ligands present in the normal human prostate (229).

C. FGF signal transduction pathways in prostate cancer

FGFs may promote tumor growth by different mechanisms, acting as angiogenic inducers, mitogens for tumor cells, or inhibitors of apoptosis (166). Studies investigating different forms of cancer have provided compelling evidence implicating the FGF signaling system in cancer genesis and progression. FGFs have been linked to mammary carcinogenesis, prostate carcinogenesis, skin tumorigenesis, urothelial cancer, testicular cancer, and hematological malignancies (reviewed in Ref. 255). Prostate cancer is the most common cancer in men in the United States and the second leading cause of cancer deaths in the Western world (256). Although there are radical treatments that may improve survival for localized disease, there are few therapeutic options available for metastatic prostate cancer (257). There is currently a large body of evidence linking alterations in the FGF/FGFR signaling pathway to the initiation and progression of many forms of cancer, including prostate (166,229,258). FGFs including FGF1, -2, -6, and -8 have all demonstrated increased expression levels and can act as paracrine and/or autocrine growth factors in these cancer cells (259,260,261,262).

The most studied FGF in prostate cancer is FGF2. Clinically FGF2 has been implicated in benign and malignant growth of the prostate. FGF2 levels are increased in serum and urine of patients suffering from proliferative disorders of the prostate (166,261,263,264). Apart from its angiogenic properties, the importance of FGF2 was illustrated in experimental prostatic tumors in nude mice. It was shown that when injected sc, the androgen-sensitive LNCaP cells rarely form tumors in nude mice. However, when nude mice were coinjected with LNCaP cells and a matrix absorbed with FGF2, tumors were consistently induced (265). Immunohistological analysis of tissue specimens from patients with prostate cancer showed that FGF2 is mainly expressed in the carcinomatous areas, indicating an enhanced production of FGF2 by the prostate tumor cells (166,261,266). FGF6, although undetectable in the normal prostate, has been shown to be elevated in prostate cancer. It was further determined to be mitogenic for primary epithelial and stromal cells (250). Like FGF6, FGF8 is highly elevated in prostate cancer, and its expression is linked with advancement of tumor progression and a poor prognosis (262). It has also been demonstrated that expression of this mitogen is regulated, at least in part, by the androgen receptor at the transcriptional level (267). FGF17 has been shown to promote epithelial proliferation in a paracrine manner, and its up-regulation during prostate cancer is thought to contribute to the increased epithelial proliferation associated with the disease (251). Taken together, these results suggest that the enhanced expression of FGFs contributes to a more aggressive phenotype in prostate cancer (166).

Prostate cancer epithelial cells express FGFRs 1 to 4 (71,268). FGFR1 inactivation results in extensive cell death in prostate cancer cell lines, suggesting a therapeutic basis for targeting of this receptor (269). Consistent with this, another study showed that FGFR1 overexpression was evident in both early and late stages of disease (270). It was further shown that forced overexpression of a FGFR1 kinase in premalignant epithelial cells resulted in accelerated progress to a malignant phenotype (71). These results suggest that FGFR1 has an important role in the initiation and progression of prostate cancer.

In the rat Dunning tumor system, tumor progression is associated with changes in FGFR2. During the early stages of prostate cancer, the predominant receptor form is the FGFR2 IIIb isoform, which preferentially binds to FGF7. The progression from a benign to a malignant cell is accompanied by a switch to the FGFR2 IIIc isoform, which no longer recognizes FGF7 but has a strong affinity for FGF2. This switch is followed by the activation of the FGF2, FGF3, and FGF5 genes (166,271). The loss of the FGFR2 IIIb isoform is also accompanied by a concurrent reduction of FGFR2 expression and a general increase and abnormal activation of the FGFR1 gene (245).

The role for FGFR3 in prostate cancer is unknown. Increased levels of FGFR4 have been seen in aggressive forms of prostate cancer (270). Suppression of FGFR4 signaling results in a reduction of cancer cell proliferation and invasion in response to exogenous stimulation (270). FGFR4 is a receptor for FGF8 and FGF17, both of which are overexpressed in clinical prostate cancers. Therefore, utilizing a targeted knockdown of FGFR4 may be an attractive option for disrupting FGF signaling with the hope of reducing cell proliferation and cancer progression (270).

Activation of FGFRs activates multiple signal transduction pathways including the PI3K and MAPK pathways, both of which have been shown to play roles in prostate cancer progression (229). Heparin and HSPGs are crucial for FGF/FGFR signaling; however, little is known about their protein interactions during prostate cancer. Syndecan 1, a heparin sulfate proteoglycan that modulates FGFR activity, has been shown to be significantly overexpressed in prostate cancer cells and is associated with decreased survival (272). In recent years it has become evident that FGF signaling is regulated by endogenous inhibitors that influence growth factor stimulation. Several of these inhibitors have been shown to affect FGF signal transduction; therefore, these inhibitors have the potential to decrease the biological activities of FGFs in prostate cancer cells, minimizing their ability to promote cancer progression. In a recent study, expression of Sef, a key inhibitory regulator of FGF signaling, was found to be down-regulated in advanced prostate cancer (273). Other studies have also demonstrated that Spry1 and -4 expression in prostate cancer is also down-regulated (274,275).

There is an increase in FGFR signaling in prostate cancer, and it has been shown to have wide-ranging effects, involving both cancer cells and their surrounding stromal cells. Overstimulation of FGFR signaling results in enhanced cancer proliferation, resistance to cell death, increased motility and invasiveness, increased angiogenesis, enhanced metastasis, androgen independence, and resistance to chemotherapy and radiation, all of which enhance tumor progression of prostate cancer (reviewed in Ref. 229). Knowledge pertaining to regulation of FGF signaling and identification of its inhibitory factors in prostate cancer may provide a new target approach for therapeutic intervention in this disease.

VIII. FGF/FGFRs and Disease in Humans

Almost all tissues express FGFs (46). Their expression patterns are often similar between the same subgroup, although their mode of action is highly variable and most FGFs are secreted even if they lack signal sequences (276). The expression of the FGF/FGFRs during development is important in regard to mediating FGF actions in a tissue-specific manner (277). To study functions of FGF signaling, rodent strains carrying different types of mutations, null, hypomorphic, isoform, or conditional knockouts, have been produced, and analysis of these mutant mice has provided valuable information about FGFRs within various aspects of mammalian development. The biological significance of this signaling system in human health and development is illustrated in recent observations that several FGF signaling disorders also result in human hereditary diseases. Some of these diseases (and their corresponding FGF/R mutation) include: Michel aplasia (FGF13); aplasia of lacrimal and salivary glands (FGF10); Börjeson-Forssman-Lehmann syndrome (FGF13); hereditary spinocerebellar ataxias (FGF14); Parkinson disease (FGF20); autosomal dominant hypophosphatemic rickets (FGF23); Pfieffers (FGFR1 mutation); Crouzon, Apert, Matthew-Wood, Pfieffers, Jackson-Weiss syndromes (FGFR2); and achondroplasia, thanatophoric dysplasia (FGFR3) (44,278,279,280). Patients presenting with these disorders often have multiple defects such as anosmia, mental retardation, and short stature (281). Due to these multiple defects, the treatment of infertility in patients is not a clinical priority; therefore the impact of FGF signaling on fertility in these disorders has yet to be determined.

Infertility can be caused by, among other things, minor deletions or translocations of genes on the X chromosome (282). The most profound defects of this type in testicular development are caused by the loss of a major portion of the X chromosome. Deletions of this nature cause problems for the development of a male fetus because the loss of one or more genes on the X chromosome cannot be compensated for by the presence of a second X chromosome as for females (283). The Xp22 contiguous gene syndrome is a rare disorder (281). De novo deletion of Xp22-pter affects several genes, including those responsible for glycerol kinase deficiency, adrenal hypoplasia congenital, Duchenne type muscular dystrophy (281), and KS (discussed in Section IX).

Despite all the evidence demonstrating that the FGF signaling system is crucial to many aspects of male reproduction, its affect on male fertility is poorly understood.

IX. Mutations of FGFR1 Signaling as a Cause of Kallmann Syndrome

A. Kallmann syndrome

In 1856, Aureliano Maestre de San Juan first reported the autopsy finding of a hypogonadal man with small testes and absent olfactory bulbs (284). In 1944, Franz Josef Kallmann subsequently demonstrated the heritable form of hypogonadotrophic hypogonadism (HH) with anosmia in three different families (284,285). KS is a developmental disorder that is characterized by HH and anosmia, and it occurs in one per 7000–8000 births (282). The anosmia is related to the absence or hypoplasia of the olfactory bulbs, and hypogonadism is due to GnRH deficiency and is likely to result from the failed embryonic migration of GnRH-synthesizing neurons (286). Classically, KS patients present with absent puberty; however, partial sexual development can rarely occur (287). KS can occasionally be diagnosed during the neonatal period on the basis of cryptorchism, microphallus, and the absence of the postnatal surge in LH, FSH, and testosterone (288). Neurological defects (synkinesia and hearing loss) and nonneurological defects (renal aplasia, high arched palate, labial or palatine cleft, and dental agenesis) occur with variable frequency in patients with KS (289,290,291). Usually, men and women diagnosed with HH cannot complete sexual maturation unless treated with GnRH or gonadotropins. Such patients require continual hormone replacement therapy to maintain sexual maturation (292). However, it has been observed that a rare subset of patients with KS can undergo subsequent spontaneous recovery of reproductive function (291,293,294). KS is genetically heterogeneous, with reports indicating autosomal dominant, recessive, and X-linked transmission (295,296). To date, two different genes have been implicated in the molecular basis of KS: specifically KAL1, which encodes the protein anosmin-1, and FGFR1 (KAL2) (296,297).

B. KAL1 encodes for anosmin-1

The gene responsible for the X-chromosome-linked form of the disease was identified as KAL1, and it encodes anosmin-1, an approximately 95-kDa extracellular matrix glycoprotein of poorly characterized function that shows significant homology with molecules known to have specific roles in neuronal development (286,298,299). Mutations in KAL1 account for approximately 10% of KS cases (300,301). In vitro experiments, carried out in COS cells, revealed a role for anosmin-1 in the control of different cell functions including cell adhesion, neurite/axonal elongation and fasciculation, and the migratory activity of GnRH-producing neurons (298,299,302). The predicted structure of anosmin-1 consists of different domains comprising an N-terminal peptide, followed by a cysteine-rich domain, a four-disulfide core domain typical of whey acidic proteins, four FNIII repeats, and a C-terminal histidine-rich region (303,304). Molecular modeling demonstrated that anosmin-1 domains are highly extended and have intradomain flexibility, which accounts for the known ability of anosmin-1 to interact with macromolecular ligands in the matrix (305,306). It has also been demonstrated that anosmin-1 binds strongly to heparan sulfate proteoglycans (304).

In many families with only males affected and presumably maternal transmission, no mutation has been found in the KAL1 coding region. Furthermore, molecular analysis of the gene in sporadic male cases rarely identifies a mutation in KAL1, suggesting that mutations in the X-linked gene do not account for the higher prevalence of this syndrome (307,308,309).

C. Association of FGFR1 (KAL2) with Kallmann syndrome

Recent studies have found that mutations in FGFR1 are involved in an autosomal dominant form of KS (KAL2) (286). Dode et al. (310) described heterozygous mutations in the 18 coding exons and splice sites of FGFR1, including one nonsense, two frame shift, two donor splice-site, and seven missense mutations in both familial and sporadic cases of KS. This study and others have demonstrated that several of these mutations affected residues involved in FGFR folding and/or in signal transduction (310,311), supporting the conclusion that mutations in FGFR1 underlie the autosomal dominant form of KS (310). Individuals affected by KS due to KAL2 mutations show not only typical HH and anosmia but also other defects, such as agenesis of the corpus callosum, unilateral hearing loss, and fusion of the fourth and fifth metacarpal bones (309,312).

The significant overlap in the clinical manifestation of KS argues in favor of anosmin-1 being involved in FGF signaling via FGFR1. A functional interaction between KAL1 and KAL2 gene products has been proposed (310,313); in particular, these two factors have been suggested to converge on the control of a common downstream intracellular signaling pathway (286). The exact mechanism of how these two proteins interact is still unknown. However, it is thought that such an interaction could explain the higher prevalence of the disease in males, assuming that the local concentration of ansomin-1 has a critical effect on FGF signaling (286). KAL1 has been shown to escape the X-inactivation process that takes place in female somatic cells (314,315). As a consequence, females are expected to synthesize larger amounts of anosmin-1 compared with males. Thus, it has been proposed that the physiologically higher concentrations of anosmin-1 protein in female embryonic tissue can compensate, in some cases, for a situation of FGFR1 haploinsufficiency (286).

As stated above and in the section describing FGFR signaling (Section II.E), both anosmin-1 and FGFs require interaction with HSPGs for their function (74,286,304). This interaction strengthens the notion that anosmin-1 and FGFs work together, in the presence of HSPG, to regulate a common signal transduction pathway. Anosmin-1 may also directly interact with FGFR1 because it has recently been shown that the two FNIII repeats of neural cell adhesion molecule bind to the extracellular domain of FGFR1 (286,316,317). Furthermore, anosmin-1 and FGFR1 are both found in the olfactory placode of human embryos as early as 4.5 wk and are known to be part of the “scaffold” for the migratory path of GnRH-expressing neurons toward the hypothalamus (284). The close proximity of anosmin-1 and FGFR1 at these stages supports the possibility of their interaction during olfactory GnRH system development (318). Recently, a functional in vitro interaction between anosmin-1 and the FGFR1-FGF2-HS complex has been demonstrated (318). The study, utilizing human embryonic GnRH olfactory neuroblasts, demonstrated that anosmin-1 induced neurite outgrowth and cytoskeletal changes through an FGFR1-dependent mechanism. Anosmin-1 enhanced FGF2 signaling in a HS-dependent manner, specifically through the FGFR1 IIIc isoform (284,303,318). Other studies have also suggested that mutations associated with KAL2 could hinder the formation of functional receptor dimers by a dominant-negative effect (310).

Spermatogenesis in patients with KS is either absent or significantly reduced, depending on the severity of the GnRH deficiency. The phenotypic variations seen among adult patients with KS range from normogonadotropic fertile patients to those patients who are reminiscent of prepubertal boys (282). Many patients respond to the induction of puberty with treatments of GnRH and testosterone, thus restoring spermatogenesis to near normal levels (282). However, there are some patients who fail to respond to this type of therapy. The discovery of a functional HSPG-dependent interaction between anosmin-1 and the FGFR1/FGF2 signaling pathway in KS illustrates the importance of the tight regulation of the FGFR1 signaling system. Furthermore, it has previously been shown that FGFR1 signaling is a requirement for normal spermiogenesis (30). When taken into consideration, the combined human and mouse research lends itself to the suggestion that KS, at least in a proportion of cases, is composed of a hypothalamic and a testicular component. Furthermore, disruption of FGFR1 signaling could be one of the underlying causes of male infertility in the GnRH nonresponding patients. Future work will involve determining the mechanism of anosmin-1-regulated FGFR1 signaling. An understanding of this mechanism of regulation will not only provide insight into the regulation of FGFR1 signaling, but may also provide potential therapeutic strategies to help combat the phenotypes/characteristics of this syndrome.

X. Concluding Remarks

The underlying mechanisms of FGF biology, including ligand binding, receptor activation, and subsequent activation of signal transduction, are all part of a complex network of signaling and transcriptional events that are regulated by multiple factors and therefore cannot simply be explained in a straightforward manner. There is substantial cross-talk between FGFs and other signaling pathways, and we are just beginning to understand the mechanisms of these cross-regulators. Over the last 30 yr, substantial progress has been made in understanding the role of FGF/FGFR signal transduction during male reproduction. In this review, we have discussed a variety of mechanisms by which FGF/FGFR signaling can produce diverse responses in different cell types of the male reproductive system (see Table 1). We have seen that FGF signaling is responsible for the proliferation and differentiation of testicular cells, the induction of quantitatively normal spermiogenesis, male fertility, the control of transcription in the initial segment of the epididymis, the development and regulation of the normal human prostate, and the enhanced proliferation of prostatic cancer. There is still, however, a lot of missing information. For example, what is the biological significance of the segmental expression of FGF/FGFRs throughout the epididymis? What roles do FGFRs play in spermatogenesis? Thus, additional investigations using biological and therapeutical approaches are required to further our understanding of FGF/FGFR signaling in regulation and maintenance of the male reproductive system. In their lifetime, one in six men will be diagnosed with prostate cancer and one in 25 men will be affected by infertility. Therefore, to enhance our ability to diagnose and treat these men, we require an understanding of the biochemistry and physiology of male reproductive function. This review brings together a body of evidence demonstrating the involvement of FGF/FGFR signal transduction in the regulation and maintenance of male reproduction. We aimed to draw attention to the importance of this signaling family, and we hope that by doing so we will inspire additional investigations, enabling advances in the development of novel male-based contraceptives, and ultimately aid in the diagnosis and treatment of testicular/prostate cancer and male infertility.