2.2. Kallikrein gene families
The glandular or tissue kallikreins are a sub-family of serine proteases, with a high degree of substrate specificity and diverse expression in various tissues and biological fluids. The term "kallikrein" appeared in the literature for the first time in the 1930s, when large amounts of protease enzymes were found in pancreas isolates (pancreas is "Kallikreas" in Greek) (Kraut et al. 1930, Werle 1934). Nowadays kallikrein enzymes are divided into two groups, plasma (EC 3.4.21.34) and tissue (EC 3.4.21.35) kallikreins, which differ significantly in their molecular weight, substrate specificity, immunological characteristics, gene structure, and type of the kinin released. Plasma kallikrein is expressed solely in liver and it is involved in blood clotting, fibrinolysis, regulation of blood pressure and inflammatory reactions (Bhoola et al. 1992). Tissue kallikreins are a large group of enzymes which have substantial similarities at gene and protein level. Tissue kallikreins are involved in the post-translational processing of the polypeptides (like kininogen) and releasing potential biologically active peptides (like kinin) (Clements 1989, 1997). Tissue kallikreins are also called kininogenases. Kininogenases or kininases are enzymes that inactivate kinins (Yousef & Diamandis and the references therein, 2001). Pancreatic/renal or K1 (new nomenclature) kallikrein enzyme is the only one among human and animal tissue kallikreins possessing kininogenase activity, releasing lysyl-bradykinin from kininogen. hK1 has also an effect on blood pressure, electrolyte balance and inflammation, and it may also digest other substrates, like growth factors, hormones, and cytokines (Bhoola et al. 1992). The kallikrein-kinin system has been shown to activate angiogenesis in an in vivo mouse model (Emanueli et al. 2001). The concept "tissue kallikrein" does not only refer the enzymes with the above-mentioned functions, but also to group of enzymes with conserved gene and protein structures that are also located at same gene locus as the KLK1 gene.
2.2.1. Human kallikreins
The size of the tissue kallikrein gene family varies between species. The genes are located as clusters in the same chromosomal locus. In the house mouse, the kallikrein genes are clustered at chromosome 7 in a single locus (Mason et al. 1983, Evans et al. 1987). The mouse kallikrein gene family has at least 24 members, 14 of which seem to be produced as functional proteins, while the rest are pseudogenes (Mason et al. 1983, van Leeuven et al. 1986, Evans et al. 1987, Drinkwater et al. 1987). In the rat genome there is also a large kallikrein gene family, 15-20 genes, and at least 10-11 of the genes produce functional proteins, which are expressed with different patterns in various organs (Wines et al. 1989, MacDonald et al. 1996). The guinea pig tissue kallikrein family is smaller than in other rodents, it has only three members (Fiedler et al. 1999).
2.2.1.1. Characteristics of human kallikreins
The human tissue kallikrein gene family was thought to be formed of three genes (Riegman et al. 1992), but during the last few years studies on new kallikrein-like genes have shown that there are at least 15 members in this family. It is believed that members of this family are developed through gene or chromosomal duplication from a common ancestor gene, while they have similar activities or predicted activities (all of them have not yet been isolated in their native form or produced as recombinant protein), have conversed primary and tertiary structures (Gan et al. 2000). Tissue kallikrein genes are clustered on chromosome 19 (Riegman et al. 1992, Gan et al. 2000, Harvey et al. 2000, Yousef et al. 2000a), like other gene clusters, like the granzyme gene cluster on chromosome 14 (Barros et al. 1996) and the trypsinogen loci on chromosomes 7 and 9 (Rowen et al. 1996). All tissue kallikrein genes are located within the 320 kb region (Gan et al. 2000, Harvey et al. 2000, Yousef et al. 1999a, 2000a). Kallikreins are heat-stable glycoproteins with a single polypeptide chain, with molecular weight varying between 27 to 40 kDa (Clements 1989 and the references therein). Kallikreins are serine proteases like elastase, thrombin, plasmin, chymotrypsin and trypsin, with the characteristic serine residue and conserved Gly-Asp-Ser-Gly amino acid sequence in their catalytic site (Neurath et al. 1967). There are 29 conserved or invariant amino acids surrounding the catalytic site, which are also present in tissue kallikreins (Dayhoff 1976).
Three of the known human kallikreins are called classical kallikreins because of their earlier discovery, and the 12 kallikrein genes discovered during the last few years are termed new kallikreins. Yousef with co-workers (2000a) showed, with the linear genomic sequences around chromosome 19q13.3-q13.4 generated by the Lawrence Livermore National Laboratory, the precise location of three classical and 12 new kallikrein genes. The chromosomal organization of human kallikrein genes is presented in Figure 1. The nomenclature for kallikrein genes, especially for new genes has been miscellaneous. Some of the new kallikreins were nominated according to The Human Genome Organization"s (HUGO) proposed guidelines for the nomenclature of serine proteases with PRSS, and some with KLK prefixes. An international group of scientists suggested uniform nomenclature for all the human tissue kallikreins (Diamandis et al. 2000a), which generated vivid discussion on the HUGO Gene Nomenclature Committee home pages on the Internet (HGNC 2000), especially in the case of hKLK4 (Table 1).
All human tissue kallikreins share common features. All tissue kallikrein genes are located at 19q13.4 chromosomal region, they encode putative serine proteases, with a catalytic triad of amino acids in their active site. Histidine, aspartic acid, and serine are in conserved positions for trypsin-like activity, or asparagine instead of serine, predicting chymotrypsin-like activity. All genes have five coding exons, with identical or similar sizes (some have one or more untranslated exons). Intron phases are conserved for all tissue kallikrein genes. There are significant, 40-80%, sequence homologies at DNA and protein levels for all kallikreins, but within the new kallikrein genes the homologies areless than 25-44% (Clements et al. 2001). The major difference between classical and new kallikreins is that there are extra exons (either 5" or 3") present in many of the new kallikreins (Luo et al.1998, Hu et al. 2000, Mitsui et al. 2000, Yousef & Diamandis 2001, Yousef et al. 1999c, 2000c, d).
2.2.1.2. Classical kallikreins
There are three so-called classical kallikreins: tissue/renal/pancreatic kallikrein, prostate-specific glandular kallikrein and prostate-specific antigen. For prostate-specific antigen the most known symbol hPSA will be used throughout this review. hK1 is secreted from the kidney, pancreas and salivary glands (Schachter 1979). hK1 is the only member of kallikrein family which is involved in the regulation of blood flow, sodium equilibrium, inflammation, and cell proliferation (Bhoola et al. 1992). Nucleotide sequences for classical kallikreins were clarified in the 1980s, KLK1 (Baker & Shine 1985, Fukushima et al. 1985), KLK2 (Schedlich et al. 1987) and KLK3 (Wang et al. 1981). These three kallikreins have high homologies between each other in their DNA and protein levels; KLK1 and KLK3 have about 60%, KLK1 and KLK2 66%, while KLK2 and KLK3 have about 80% homology (Henttu & Vihko 1989). The genes cluster in a 60 kb region on human chromosome 19q13.3-q13.4 (Yousef et al. 2000a) as seen in Figure 1. The "Kallikrein loop", a 9-11 amino acid peptide sequence, is only found in classical kallikreins. This loop precedes the aspartatic acid in the active site/catalytic triad (Yousef & Diamandis 2001). Classical kallikreins have also splice variant forms. For KLK1 gene a new transcript has been isolated from colon (Chen et al. 1994). The KLK2 gene has also splice variants (Riegman et al. 1991, Liu et al. 1999). KLK3 has been described to have splice variants in addition to the main 1.6-kb transcript (Riegman et al. 1989, Henttu et al. 1990, Heuzé et al. 1999).
Figure 1. The localization of human tissue kallikrein gene family at chromosomal locus 19q13.3-q13.4.The kallikrein genes are represented as boxes bearing the names of the respective genes. The length of the gene (in base pairs, bp) is marked on the right-hand side of the respective box. The distances (in bp) between genes are shown between "boxes". Arrows on the left of each gene point to the direction of transcription. Modified from Yousef & Diamandis (2000) and Clements et al. (2001).
2.2.1.3. New kallikreins
The human kallikrein gene family was thought to be composed of 3-4 (Fukushima et al. 1985, Baker & Shine 1985) up to 19 genes (Murray et al. 1990). New technologies and information available (Human Genome Project) has made it possible to identify or predict the existence of new genes, and during the past few years new members of the human tissue kallikrein family have been discovered. The genes in the family, including the classical kallikreins, are named KLK1-KLK15 and the proteins they encode are termed hK1-hK15 (Table 1). KLK4-KLK15 genes are less related (25-44%) and do not have the usual kallikrein loop of classical kallikreins (Clements et al. 2001). The expression of new kallikreins in different tissues has been studied mainly using RT-PCR techniques (Gan et al. 2000 and the references therein, Clements et al. 2001 and the references there in) and a summary of this data is presented Figure 2.
Figure 2. Expression of kallikreins in different tissues. The expression detected is marked with a grey square, high expression detected is marked with black square and no expression detected is marked with a white square. For more information see text.
KLK4 with protein symbol hK4 was independently cloned by two groups (Nelson et al. 1999, Stephenson et al. 1999). It is a significantly divergent member of the human tissue kallikrein family. KLK4 is located down-stream from KLK2 gene transcribing
the same direction as KLK1 (Stephenson et al. 1999). The exon necessary for signal peptide formation is not present, and the green fluorescent tagged KLK4 is localized perinuclearally, suggesting that KLK4 has intracellular functions contrary to other members of the kallikrein gene family (Korkmaz et al. 2001).
KLK5 was identified from stratum corneum extracts (Brattsand & Egelrud 1999) and with positional candidate approach it was shown to be a member of the KLK family (Yousef & Diamandis 1999). The hK5 enzyme is proposed to be a part of turnover of stratum corneum and desquamation (Ekholm et al. 2000).
The KLK6 gene was discovered by several research groups (Anisowicz et al. 1996, Little et al. 1997, Yamashiro et al. 1997). KLK6 gene has high homology to trypsin (Anisowicz et al. 1996) and it might also possess amyloidogenic activity (Little et al. 1997). KLK6 has seven exons, two first ones are untranslated (Yousef et al. 1999c).
KLK7 was first linked to the desquamation of stratum corneum (Lundstrom & Egelrud, 1991). It was later cloned and characterized by two groups (Hansson et al. 1994, Yousef et al. 2000b) to be part of the tissue kallikrein gene family with chymotrypsin-like instead of trypsin-like activity, because aspartatic acid is replaced with asparagine in substrate binding pocket (Ekholm & Egelrud 1999).
KLK8 was first thought only to be a brain-related enzyme (Yoshida et al. 1998a), but it was found to be expressed in ovarian carcinoma (Underwood et al. 1999). KLK8 has two splice variants (Mitsui et al. 1999, Magklara et al. 2001), which are expressed abundantly in several normal human tissues, but mainly in human brain. KLK8 cDNA and the predicted amino acid sequence have 72% identity to corresponding one in mouse (Yoshida et al. 1998a).
When the human kallikrein gene locus was characterized, the KLK9 gene was discovered. The predicted enzyme activity for hK9 enzyme is chymotrypsin-like, while serine is replaced by aspartate in the catalytic triad (Yousef & Diamandis 2000).
KLK10 was isolated with subtractive hybridization between a nontumorigenic and tumorigenic breast cancer cell line (Liu et al. 1996). KLK10 acts as tumor suppressor in prostate, breast and possible other epithelial cells (Goyal et al. 1998). KLK10 has six exons, one untranslated (Luo et al. 1998). KLK10 gene expression is up-regulated by estrogens, androgens and progestins when studied with BT-474 breast cancer cell line (Luo et al. 2000). An assay measuring the free fraction of K10 protein in biological fluids and tissues has been developed using recombinant K10 (Luo et al. 2001a).
KLK11 was first cloned from human hippocampal cDNA with PCR (Yoshida et al. 1998b), and it is expressed in human skin, salivary gland, stomach, prostate, intestine and brain (Yousef et al. 2000c). KLK11 has six exons (first one untranslated) and five identical introns to other kallikreins (Yousef et al. 2000c). KLK11 has two splice forms, with a difference in their N-terminal part, which are expressed primarily in either the brain or prostate. Serine in catalytic triad predicts trypsin-like activity for hK11 (Mitsui et al. 2000).
The sequence alignment test indicated that KLK12 has seven exons, at least one of them untranslated. One extra 3" exon was also found, which has not been reported for any other members of kallikrein family. KLK12 has three splice variant forms. Using cDNA from 26 different tissues as template, three distinct bands were observed in many tissues.
The first is "classical form" with the typical five exons and four introns. The second is KLK12-related protein-1, in which the last exon is split into two separate exons with an additional intron. KLK12-related protein-1 is a protein 254 amino acids long, six amino acids longer than classical KLK12. The third is KLK12-related protein-2, which is otherwise similar to classical KLK12, but the fourth exon is missing (Yousef et al. 2000d).
The KLK13 gene was identified with the positional candidate gene approach. It has the typical gene structure for kallikreins with five coding exons. This gene is mainly expressed in prostate, salivary gland, breast, and testis. Aspartate in position 239 predicts that KLK13 protein is likely to have trypsin-like activity (Yousef et al. 2000e). KLK13 has also five splice variants, which are differentially expressed in testis and testicular cancer (Chan et al. 2001).
KLK14 has seven exons, five of which are coding exons (Hooper et al. 2001, Yousef et al. 2001b). KLK14 is expressed in various tissues, but mainly in central nervous system (CNS), which indicates that KLK14 might be involved in brain physiology. Aspartate at position 198 predicts trypsin-like activity (Yousef et al. 2001b).
KLK15 was found to be located between KLK1 and KLK3 at chromosome 19q13.4. There are five coding exons in KLK15. The possible presence of upstream-untranslated exon(s) could not be ruled out (Yousef et al. 2001c). Glutamic acid at the position 203 suggests that hK15 has unique substrate specificity. KLK15 has evenly distributed hydrophopic regions in the polypeptide chain like globular proteins, similarly to other serine proteases. At positions 148-155 hK15 has a unique 8-amino acid loop (HNEPGTAG). Out of the 29 conserved amino acids surrounding the active site, 28 were found in KLK15. One of the non-conserved amino acids (serine 173 instead of proline) is also found in hK5 and hK12 proteins, representing a conserved evolutionary change to a protein of the same group, according to protein evolution studies. The KLK15 gene is expressed in various tissues, with highest levels in the thyroid gland, and showing high expression in high-grade prostate tumors. The KLK15 gene might have a function in the physiology of normal thyroid gland since the expression of the gene is so high. KLK15 gene also has two splice variants which probably express truncated proteins. With phylogenetic studies KLK15 was grouped with KLK9 and KLK11 (Yousef et al. 2001c).
2.2.1.4. Kallikreins in diseases/malignancies
The exact role and involvement of tissue kallikreins in human diseases is not clear. Serine proteases are involved in tumor progression, such as invasion, proliferation and tumor metastasis (Tryggvason et al. 1987, DeClerck et al. 1997, Carroll & Binder 1999, Wolf et. al. 2001), but there are data supporting contrary functions of serine proteases in diseases. The serine proteases act as tumor suppressors (Goyal et al. 1998), as antiangiogenic factors (Fortier et al.1999), as apoptotic molecules (Balbay et al. 1999) or negatively regulate cell growth (Lai et al. 1996).
hK1 as a part of the kallikrein kinin system is involved in hypertension (Margolius et al. 1974), inflammation (Clements 1997), pancreatitis (Griesbacher & Lembeck 1997), renal diseases (Horwitz et al. 1978, Holland et al. 1980) and different types of cancers (Berg et al. 1985, Jones et al. 1989, 1990). High expression of the KLK4 gene has been reported in ovarian tumors, especially in aggressive forms of ovarian carcinomas (Dong et al. 2001, Obiezu et al. 2001). The expression of KLK4 gene is high in prostate tissue (Nelson et al. 1999, Yousef et al. 1999b, Harvey et al. 2000), but the connection between hKLK4 and prostate cancer has not been shown yet. KLK5 (Kim et al. 2001), KLK6 (Anisowicz et al. 1996, Diamandis et al. 2000b), KLK7 (Tanimoto et al. 1999), KLK8 (Underwood et al. 1999, Magklara et al. 2001) and KLK10 (Luo et al. 2001b,c) are also expressed in high levels in ovarian cancer and could be future ovarian cancer markers. KLK6 is expressed in primary mammary carcinoma cell lines (Anisowicz et al. 1996) and the connection of KLK6 to Alzheimer"s disease has been shown with its amyloidogenic activity (Little et al. 1997, Diamandis et al. 2000b). KLK7 was first linked to the desquamation of stratum corneum (Lundstrom & Egelrud, 1991), and it is suggested that it has a role in skin diseases, like pathological keratinization and psoriasis (Sondell et al. 1996, Ekholm & Egelrud 1999). KLK8 has also high expression in Alzheimer"s disease (Shimizu-Okabe et al. 2001) and its involvement in diseases of the CNS has been reported (Akita et al. 1997, Scarisbrick et al. 1997, Momota et al. 1998, Kishi et al. 1999, Mitsui et al. 1999, Yoshida & Shiosaka 1999). KLK9 is expressed at significantly higher levels in early stages of ovarian carcinoma (Yousef et al. 2001d). KLK10 expression is down-regulated in aggressive forms of prostate cancer (Luo & Diamandis 2000). It has been shown in animal models that KLK10 acts as a tumor suppressor in breast cancer (Liu et al. 1996, Goyal et al. 1998). hK10 has been suggested to be a marker for ovarian cancer (Luo et al. 2001b, c). Elevated serum levels of hK11 have been detected in ovarian and prostate cancer (Diamandis et al. 2002). KLK12-14 expression is low in breast cancer (Yousef et al. 2000b, Yousef et al. 2000d, e, 2001b). KLK15 shows higher expression in more aggressive, higher stage prostate cancers (Yousef et al. 2001c).
The role and suitability of new kallikreins as a marker for diseases, particularly in cancers, will be under investigations for next the few years. The role of hK2 and hPSA in diseases is discussed later in this review.
2.2.1.5. The steroid hormone regulation of kallikreins
Most of the studied kallikrein gene expression in different tissues is under regulation of androgens, estrogens and progestins. No clear up-regulation by steroid hormones for human KLK1 gene has been shown (Murray et al. 1990). High expression in the human endometrium in the middle of the menstrual cycle indicates up-regulation of KLK1 gene by estrogens (Clements et al. 1994). KLK2 gene expression is up-regulated by androgens and progestins (Riegman et al. 1991, Henttu & Vihko 1993, Hsieh et al. 1997, Shan et al. 1997, Magklara et al. 2000a). Androgens up-regulate also the expression of hPSA (Henttu & Vihko 1992, 1993, Henttu et al. 1992, Young et al. 1995, Shan et al. 1997, Magklara et al. 2000a). There are two androgen response elements in promoter region of KLK2 (Murtha et al. 1993), and three androgen response elements for KLK3 have been identified and their function has been studied (Luke & Coffey 1994, Schuur et al. 1996, Cleutjens et al. 1997).
The KLK4 gene is up-regulated by androgens in LNCaP prostate cancer cells and by androgens and progestins in BT-474 breast cancer cells (Nelson et al. 1999, Yousef et al. 1999b). Putative androgen response elements have been identified from the proximal promoter area of KLK4 (Stephenson et al. 1999). KLK5 expression is up-regulated with estrogens and progestins in breast cancer cell lines (Yousef & Diamandis 1999). Estrogens and progestins and to a lesser extent androgens up-regulate KLK6 expression in breast cancer cells (Yousef et al. 1999c). KLK7 expression is up-regulated in breast cancer cells by estrogens and glucocorticoids (Yousef et al. 2000b). KLK9 expression is up-regulated in breast cancer cell line by progestins, estrogens and androgens, the highest up-regulation having been detected with progestin induction (Yousef & Diamandis 2000).
KLK10 expression is up-regulated in breast cancer cell line BT-474 by steroid hormone stimulation with following potency: estrogens > androgens > progestins (Luo et al. 2000). The expression of KLK11 gene was shown to be regulated with steroid hormones in the same cells (Yousef et al. 2000c). KLK12 gene expression is up-regulated by androgens and progestins in LNCaP prostate cancer cells, whereas in two breast cancer cell lines KLK12 expression is up-regulated by steroid hormones with the following potencies: estrogens > androgens > progestins in BT-474 cells and androgens > progestins > estrogens in T-47D cells (Yousef et al. 2000d). KLK13 gene expression is up-regulated by androgens and progestins and to a lesser extent by estrogens in BT-474 cells (Yousef et al. 2000e). KLK15 gene expression is up-regulated in LNCaP cells; the highest induction was obtained with dihydrotestosterone and lowest with aldosterone (Yousef et al. 2001c).