Dipeptidyl Peptidases: Substrates and Therapeutic Targeting in Human Health and Disease

1.1 Introduction

Dipeptidyl peptidase 4 (DP4), fibroblast activation protein (FAP), DP8, and DP9 are the enzymatic members of the serine protease S9b DP4-like gene family. One of the most important features of the DPs is their ability to preferentially cleave the N-terminal post-prolyl bond of regulatory peptides and small protein substrates. DP4 proteolysis results in the inactivation, activation, or alteration of its substrates function via changes in receptor selectivity; thus DP4 plays an important role in regulating biological function. Together, DP4 and FAP have been implicated in a number of diseases including liver disease, obesity, type II diabetes, arthritis, inflammatory bowel disease and cancer. Recently, evidence has emerged to implicate both DP8 and DP9 in innate immunity, and DP8/9 in vitro cleavage of well-known DP4 substrates, including neuropeptide Y (NPY), glucagon-like peptide (GLP)-1, and a number of chemokines, has been demonstrated. Despite this, the true pathophysiological roles of DP8/9 and their involvement within human biology and disease are still to be elucidated. Identification of the in vivo substrate repertoire of each DP will be an important step toward elucidating the biochemical pathways in which each protease is involved. This will allow us to unravel further the roles that the DPs play in human biology and disease, and evaluate further their suitability as therapeutic targets.

In this chapter, we will provide an introduction to the significance of post-proline cleavage, the DP4-like gene family enzymatic members, and the related enzymes prolyl endopeptidase (PEP) and DP2. This will be followed by an in-depth examination of the biochemical characteristics of DP4, FAP, DP8, and DP9, their natural substrates, and biological relevance following DP cleavage. Lastly, the suitability of targeting the DP family for the development of therapeutics for the treatment of human health and disease will be discussed.

1.2 Post-Proline-Cleaving Enzymes

A number of biologically active proteins and regulatory peptides such as cytokines, chemokines, growth hormones, and neuropeptides are protected from general proteolysis due to an evolutionary conserved N-terminal proline residue. Such protection results from the unique cyclic and imino structure of proline imposing conformational restrictions on the polypeptide backbone. Even proteases exhibiting a very broad substrate specificity are unable to attack peptide bonds where the prolyl residue is situated, and hence degradation of such peptides requires the use of proline specific peptidases.

Although a number of proline-cleaving enzymes have been identified, only a limited number of these proteases are capable of cleaving the N-terminal post-prolyl, X-Pro-, bond (Figure 1.1). The most notable of these enzymes are the N-terminal-specific, dipeptidyl peptidases of the serine protease SC clan, S9b sub-family such as DP4 (EC 3.4.15.5), the endopeptidase of the parent S9 family prolyl-endopeptidase (PEP; EC 3.4.21.26), and the S28 family member DP2 (EC 3.4.14.2). Lysosomal prolylcarboxypeptidase (PCP) (EC 3.4.16.2) also belongs to the S28 family, but in contrast to DP2, it functions as a C-terminal-specific protease cleaving the Pro-Xaa bond at the C-termini of proteins to release a single C-terminal amino acid. Carboxypeptidase P (EC 3.4.17.16) is a membrane-localized protease with similar cleavage specificity to lysosomal PCP; however, it is a metallo-, as opposed to serine, protease. Aminopeptidase P (EC 3.4.17.16), prolidase (EC 3.4.13.19) and prolinase are additional proline-cleaving metalloproteases. Aminopeptidase P is an N-terminal-specific protease, cleaving the pre-prolyl bond to release single amino acids at the N-termini of proteins. Importantly, aminopeptidase P is involved in cooperative activities with DP4 and related enzymes. Prolidase and prolinase cleave dipeptides with pre- and post-proline respectively to release two amino acids. The metalloprotease angiotensin-converting enzyme (ACE; EC 3.4.15.1) is also capable of cleaving prolyl bonds, although it is not renowned for its ability to do this. DP4, first discovered in 1966 by Hopsu-Havu and Glenner as the dipeptidyl cleaving glyclproline napthylamidase, was the first N-terminal post-proline-cleaving enzyme identified and hence the most readily investigated. Since its discovery, related enzymes have been identified including fibroblast activation FAP, then DP8 and DP9.

1.3 DP4-Like Gene Family and Related Enzymes

Members of the DP4-like gene family, S9b, make up a sub-family of the prolyl-oligopeptidase (POP) S9 family within the serine protease clan SC. In total the S9b family consists of six homologous members; the four pro- teases, DP4, FAP, DP8 and DP9 and two inactive protease homologs, DP6 and DP10 (Figure 1.2). Sharing similar features, PEP from the parent S9 family is also capable of cleaving the N-terminal post-prolyl bond but with endopeptidase specificity. In contrast to DP4, PEP is limited in its ability to cleave unblocked, free N-terminal dipeptides from substrates; however, inhibitors designed specifically for DPs can still bind to, and block, the enzymatic function of PEP. FAP, in addition to its DP activity, also functions as an endopeptidase. Structurally, the three crystallized S9b family members DP4, FAP and DP6 contain an α/β hydrolase domain and eight-bladed b-propeller domain (Figure 1.2) distinguishing it from that of parent PEP which contains a seven-bladed β-propeller domain. Protease members of this family and PEP all contain the non-classical arrangement of the serine protease catalytic triad (Ser, Asp, and His), located within the C-terminal portion of the α/β hydrolase domain (Figure 1.2). The inactive protease homologs DP6 and DP10 contain a mutation of the catalytic serine residue to an aspartic acid and glutamine residue respectively. All enzyme members contain the catalytic serine protease motif around the serine residue of GWSYG. Another distinguishing characteristic of the S9b subfamily that differentiates them from PEP is the pair of conserved glutamate residues situated within the β-propeller domain found to be essential for catalytic activity of DP4 and DP8 (Figure 1.2). DP2, belonging to the serine protease S28 family, is a non-structural homolog sharing a number of similarities with the DP4-like gene family including the conserved non-classical catalytic triad and similar substrate specificity with its ability to cleave N-terminal dipeptides from peptide substrates with proline in the penultimate position. As for PEP, a number of inhibitors designed for targeting DP4 are also known to interact and inhibit DP2. Hence, when investigating the expression and roles of the S9b family, due consideration must also be given to the potential involvement (or targeting in the case of inhibition) of this related enzyme. The subcellular localization of each enzyme is diverse and sometimes overlapping; they are found membrane-bound (DP4, FAP, PEP), in lysosomes (DP2), in the cytoplasm (DP8/9, PEP, DP2), or in secretory vesicles (DP2), so it is likely that their substrate degradomes will mostly differ.

Containing a trans-membrane domain within their N-termini and a number of glycosylation sites, DP4, FAP, DP6, and DP10 are all type II-membrane glycoproteins localized to the cell surface. Plasma-soluble forms of both DP4 and FAP have been identified with their presence in circulation believed to result from shedding events at the plasma membrane. DP8, DP9, and PEP lack a transmembrane domain and are cytosolic proteins. A number of membrane associated forms of PEP have been identified. Unique to DP9 is the presence of one of the best-known integrin-binding motifs the RGD (arginine-glutamine-aspartic acid; Arg-Gly-Asp) motif, within its N-termini (Figure 1.2). DP9 also contains two potential N-linked glycosylation sites; however, there is no evidence at present of DP9 glycosylation, so the significance of these features is yet to be revealed. DP4, FAP, DP8/9, and PEP all have neutral pH optimums of pH 7–8. In contrast, DP2 is a smaller, 492-amino-acid, lysosomal enzyme, active across a broad pH range with an acidic pH optimum of 5.5. DP2 also localizes within intracellular vesicles distinct from lysosomes in resting quiescent cells. It is likely that these vesicles contain a secretory component due to the release of fully functional DP2 in response to calcium stimulation. All S9b family members and the S28 member DP2 form homodimers with dimerization being essential for the catalytic activity of DP4, FAP and DP2. In contrast, PEP is a monomeric enzyme with SDSPAGE mobility of 70–80 kDa. S9b proteins have a monomer mobility on SDS-PAGE gel of 90–110 kDa and dimer mobility ranging between 150 and 200 kDa. Being much smaller, DP2 migrates in its glycosylated form as a ~60 kDa monomer on SDS-PAGE and forms a ~120 kDa homodimer.

A brief discussion of the homology and gene structure of the DP4-like gene family is provided below. Although both DP2 and PEP are of importance when discussing therapeutic targeting of DP4, FAP, and DP8/9, they will not be discussed here in detail; thus, the reader should refer to critical reviews on DP2 by Maes et al. and on PEP by Garcia-Horsman et al.. The inactive protease homologs DP6 and DP10 will be discussed in brief, as they are not a focus of this review (recently reviewed in McNicholas et al.).

1.3.1 Homology and Gene Structure

A high degree of homology exists within the DP4-like gene family with all members sharing 25–60% amino acid sequence identity and 47–77% amino acid sequence similarity. Notably, FAP and DP4, located adjacent to each other on the long arm of chromosome 2 (2q24.2 and 2q24.3 respectively), share 52% amino acid identity. The two non-enzyme members, DP6 and DP10, located on chromosomes 7q36.2 and 2q14.1 respectively, share 53% amino acid sequence identity, and the most recently identified enzyme members, DP8 and DP9, located respectively on chromosomes 15q22.32 and 19q13.3, share 61% amino acid sequence identity. The close proximity and high level of similarity between the DP4 and FAP genes suggest that they have arisen from a recent gene duplication event. Likewise, it is possible that DP10, also located on the long arm of chromosome 2, was derived from either DP4 or FAP, followed by the divergence of DP6 on chromosome 7. Similarly, the high sequence identity and shared cytosolic localization suggest that DP8 and DP9 have arisen from a gene duplication event. Although of varying gene length, the DP4 (82 kb), FAP (73 kb), DP6 (935 kb), and DP10 genes (1402 kb) all contain 26 exons, further supporting the likelihood of a common gene ancestor. In contrast, the DP8 (72 kb) gene contains 20 exons, and the DP9 gene contains 19 or 22 exons dependent on whether the gene is expressed in its short (863-amino-acid) or long (971 amino acid) forms. Like DP9, short and long transcript variants with variable length N-termini arising from alternate first exon use exist for DP6, DP10, and DP8. Differing patterns of expression and tissue specificity are often associated with these alternative transcripts, but a com- prehensive study investigating their significance to human biology is yet to be carried out.

1.4 Protease Members of the DP4-Like Gene Family

1.4.1 DP4 and FAP: Extracellular Proteins

DP4 is the best-characterized member of the DP enzyme family, followed to a lesser extent by FAP, which was identified almost 20 years after DP4. Both enzymes are implicated in a number of roles, including tissue remodeling/ wound healing, inflammatory bowel disease, arthritis, type II diabetes, obesity, and cancer (reviewed in several previous studies). In cancer, both DP4 and FAP play conflicting roles acting as either a tumour suppressor or tumour promoter depending on the cancer type. DP4 plays a key role in glucose homeostasis, regulating the activities of GLP-1 and glucose-dependent insulinotropic peptide (GIP) via its proteolytic activity, and has thus become a clinically validated target in the management of type II diabetes (reviewed in several previous studies). FAP is also of clinical relevance, particularly within a cancer and liver disease setting as discussed below. Here, the expression patterns and biological importance of DP4 and FAP will be introduced.

Human DP4 is ubiquitously expressed in a wide variety of tissues and cell types, exerting its functions on a broad range of physiological processes within the human body. DP4 is expressed at high levels on the surface of several endothelial, epithelial, and fibroblast cells of the lung, kidney, liver, intestine, bile duct, and other organs, thus affecting the gastrointestinal/ digestive, cardiovascular, endocrine, neurological, and immunological systems. In kidney, DP4 is one of the major microvillus membrane/brush- border membrane proteins. Relatively high levels of soluble DP4 activity are observed in human bodily fluids, and thus the activity of soluble, circulating DP4 is likely to also contribute to effects observed in many of these systems. Variations in the level of serum DP4 activity have been associated with numerous diseases, and the effectiveness of measuring soluble DP4 activity levels as a biomarker for diagnosis, monitoring, and prognosis of diseases such as cancer, arthritis, and psychiatric disorders has been investigated. In the immune system, expression of DP4, known as the surface antigen CD26, is detectable at low levels on the surface of some resting T cells, B cells, and natural killer (NK) cells. DP4 functions as a co-stimulatory molecule of T-cell activation, displaying increased expression on the cell surface of activated T-cells. On the surface of B cells, DP4 expression is upregulated following the immunogenic stimulation of B cells with pokeweed mitogen or the Staphlococcus aureus-derived immunogen streptokinase. Despite the widespread involvement of DP4 in animal biology, the absence of DP4 expression does not result in any known adverse defects, as demonstrated by the apparent healthy phenotype of knockout DP4 mice (DP4-/-) and their favorable protection against diet-induced obesity and associated insulin resistance. Recently, studies investigating the long-term chronic loss of DP4 expression in genetically induced DP4 deficient rats have identified some age-dependent alterations in immune system composition and stress-regulatory responses. Of interest to human pathological conditions, particularly cancer, is the apparent loss of DP4 expression during disease progression. Loss or alteration of DP4 surface expression has been observed in several malignant cells and carcinomas such as lung, breast and ovarian, prostate, and colon adenocarcinoma and during the progression of melanocytes to melanoma. It should be made clear, however, that loss of DP4 expression is not a hallmark of all cancers with many studies examining the expression and roles of DP4 in cancer producing conflicting results. In fact, DP4 upregulation has been observed in differing cancer studies to con- tribute to the metastatic cancer phenotype, thus acting as a tumor promoter contributing to disease progression.