REVIEW Endocrine - Related Cancer ( 2003 ) 10 359 – 373 Mammalian stanniocalcins and cancer

Stanniocalcin (STC) is a glycoprotein hormone that is secreted by the corpuscle of Stannius, an endocrine gland of bony fish, and is involved in calcium and phosphate homeostasis. The related mammalian proteins, STC1 and STC2, are expressed in a wide variety of tissues. The ovaries have the highest level of STC1, and this increases during pregnancy and lactation. STC1 is present in breast ductal epithelium, and its expression is induced by BRCA1, a tumor suppressor gene that has an important role in breast and ovarian cancer. The expression of STC2 is induced by estrogen, and there is a positive correlation between the level of expression of estrogen receptor and expression of both STC1 and STC2 in breast cancer. This article reviews the data currently available regarding the mammalian STCs, and discusses the roles they may play in normal physiology and in breast and other cancers. Endocrine-Related Cancer (2003) 10 359–373 Introduction Stanniocalcin (STC) derives its name from the corpuscle of Stannius (CS), an endocrine gland associated with the kidneys of fish (Stannius 1839). CS extract (Fontaine 1964) and purified STC (Lafeber et al. 1988a) were found to have an anti-hypercalcemic effect due to inhibition of whole-body Ca influx. The principal target organs for this effect are the gill (Wagner et al. 1988) and gut (Takagi et al. 1985, Sundell et al. 1992). STC was also found to stimulate resorption of inorganic phosphorus (Pi) by proximal tubule epithelium cells from fish kidney (Lu et al. 1994). A human ortholog of fish STC, STC1, was found by mRNA differential display of genes related to cellular immortalization, a key aspect of the cancer cell phenotype (Chang et al. 1995), and independently by random sequencing of a fetal lung cDNA library (Olsen et al. 1996). An STC1 paralog, STC2, was identified by searching for related sequences in expressed sequence tag (EST) databases (Chang & Reddel 1998, DiMattia et al. 1998, Ishibashi et al. 1998). Although mammalian STC1 and STC2 are not expressed ubiquitously, they are expressed in a wide variety of tissues, including endocrine glands and hormoneresponsive organs. The ovaries appear to have the highest STC1 levels out of all the tissues examined to date with increased expression observed during pregnancy and lactation (Deol et al. 2000). STC1 is also found in breast ductal epithelium (Welcsh et al. 2002). There is now mounting evidence that altered expression of the STCs may have a role in human cancer. cDNA Endocrine-Related Cancer (2003) 10 359–373 Online version via http://www.endocrinology.org 1351-0088/03/010–359  2003 Society for Endocrinology Printed in Great Britain microarray analyses showed that STC1 is upregulated in most primary hepatocellular carcinomas (Okabe et al. 2001). STC1 is downregulated 7-fold in ovarian cancer compared with normal ovarian epithelial cells (Ismail et al. 2000). An anonymous cDNA fragment shown to be downregulated in a breast cancer cell line (Liang et al. 1992) was subsequently identified as a portion of the 3′ untranslated region (UTR) of STC1 (Chang et al. 1995). However, STC1 and STC2 are both expressed in estrogen receptor (ER)-positive breast cancer (Bouras et al. 2002). Expression of STC2 is inducible by estrogen (Charpentier et al. 2000, Bouras et al. 2002) and repressed by anti-estrogen (Bouras et al. 2002). Expression of STC1 is upregulated by the tumor suppressor gene BRCA1, mutations of which predispose to breast and ovarian cancer, and loss of BRCA1 and STC1 expression are correlated in breast cancer (Welcsh et al. 2002). There is interest in the possible use of STC1 and/or STC2 expression for diagnosis and/or classification of breast cancer. This article reviews our current understanding of the mammalian STCs and their possible roles in breast and other cancers. Identification of mammalian STC genes As the CS gland has only been found in holostean and telostean fish, it had long been supposed that STC did not exist in other vertebrates (Wendelaar Bonga & Pang 1991). The first suggestion that there might be a mammalian STC was based on demonstrating STC immunoreactivity in the sera of Chang et al.: Mammalian stanniocalcins several species including human and in human kidney (Wagner et al. 1995), although a later study using more specific antibodies failed to detect STC in human sera (De Niu et al. 2000). Definitive evidence for the existence of human STC was provided by Chang et al. (1995), who identified a cDNA that was downregulated following immortalization of SV40-transformed human fibroblasts. When the cDNA was sequenced, the predicted amino acid sequence was found to share 60% identity and 73% similarity with various fish STCs. Due to its high degree of homology to fish STC, this novel mammalian protein was also named STC (Chang et al. 1995). When a second member of this gene family was identified, mammalian STC was renamed STC1. The same gene was independently isolated by Olsen et al. (1996) during random sequencing of an early-stage human fetal lung cDNA library. A second human and mouse STC gene, STC2, was identified by searching EST databases for sequences related to STC1 (Chang & Reddel 1998). Subsequently, the STC2 gene was independently identified using the same methodology by Ishibashi et al. (1998) as well as by DiMattia et al. (1998) who referred to the protein as ‘STC-related protein’. STC1 and STC2 proteins The human STC1 cDNA encodes a protein of 247 amino acids (Fig. 1). The level of sequence similarity to salmon STC is 92% over the first 204 amino acids, 118 residues of which are identical. However, the last 43 residues at the C-terminus are completely divergent. The human STC2 cDNA encodes a protein of 302 amino acids that has 34% identity to both STC1 and eel STC (Fig. 1). The relatedness of STC2 to STC1 and eel STC is greatest at the N-terminus, residues 41–160 being 40% identical. There is a greater similarity level, 53%, between STC1 and eel STC. Clearly, STC1 is more closely related to fish STC than to STC2. As with fish STC, mammalian STC1 and STC2 are predicted to be secreted glycosylated proteins with a signal peptide sequence of about 24 amino acids and a pro-sequence of about 15 amino acids that are subsequently processed to yield the mature proteins (Moore et al. 1999). A strikingly conserved feature of STCs is the presence and location of cysteine residues. There are 11 cysteine residues in STC1 with the same spacing as those in eel and coho salmon STC (Butkus et al. 1987, Wagner et al. 1992). This odd number of cysteine residues is consistent with STC1 being a homodimer in the native state, with the cysteines participating in interchain and intrachain disulfide bonding as in STC (Lafeber et al. 1988b). Disulfide linkages in chum salmon STC have been defined (Hulova & Kawauchi 1999). Chum salmon STC is a homodimer connected by a single intermonomeric disulfide bond at Cys169. The monomer contains 179 amino acids, with five intramonomeric disulfide bonds formed between Cys12-Cys26, Cys21-Cys41, Cys32-Cys81, 360 www.endocrinology.org Cys65-Cys95 and Cys102-Cys137. STC2 has 15 cysteines, of which ten have the same spacing as STC and STC1 (Fig. 1), and also exists as a disulfide-linked homodimer (Moore et al. 1999). The N-linked glycosylation site is also conserved (Fig. 1). The consensus sequence, Asn-X-Thr/Ser (around residues 62–72), is believed to be glycoslated in eel and coho salmon (Butkus et al. 1987, Wagner et al. 1992); however, direct sequence data are lacking. STC1, on the other hand, was analyzed to determine the saccharide content in both baculovirus and Chinese hamster ovary (CHO) cell expression systems (Zhang et al. 1998). At least one N-linked site was found to be glycosylated and no O-linked oligosaccharides were found. An unusual feature of STC2 is the presence of 15 histidine residues. This is five times the number seen in eel STC and more than twice the number in STC1. Four such residues are clustered towards the C-terminus of STC2 and possibly interact with transition metals. Evidence for this interaction has been obtained by Moore et al. (1999) who utilized a nickel chelating column to purify STC2. Chromosomal localization and genomic structure The human STC1 gene is on the short arm of chromosome 8 (8p11.2–p21). The STC1 gene contains four exons spanning 13 kb, and transcriptional start sites have been localized 284, 271 and 153 nucleotides 5′ of the initiator methionine codon (Chang et al. 1998, Varghese et al. 1998). The 5′ UTR is rich in the trinucleotide repeat, CAG. This area usually has 19 such repeats clustered within 102 nucleotides of the transcription start site, but is polymorphic (Chang et al. 1998). Trinucleotide repeats are associated with a number of genetic conditions including neurodegenerative diseases (Reddy & Housman 1997). In the case of STC1, there are four relatively small blocks, each consisting of three to six CAG repeats separated by 6–15 nucleotides. The intervening sequences, however, are 72% GC-rich and contain six CGG triplets. This region may represent a transcriptional control domain. The STC2 gene has been localized to chromosome 5q33 or 5q35 (White et al. 1998, Moore et al. 1999), and also contains four exons (Ishibashi et al. 1998). The exon–intron boundaries are fully conserved between STC2 and STC1, indicative of a common ancestral gene. In contrast to STC1, no CAG repeats are found in the 5′ or 3′ UTR of STC2. The fish STC gene, about 4 kb long, has been isolated from sockeye salmon (McCudden et al. 2001a), and contains five exons. From comparative alignment of the gene structures (Fig. 2), it can be seen that the size of exon 2 is highly conserved. Mammalian exon 3 corresponds to exons 3 and 4 in fish, so the mammalian exon 3 most likely represents a fusion of the fish exons 3 and 4. The existence of a fish STC2 Endocrine-Related Cancer (2003) 10 359–373