Dipeptidyl peptidase (DPP) 9 is a member of the DPP4 (DPP-IV) family, which are all members of clan SC of serine proteases, family S9, subfamily S9B. S9B proteases have a unique ability to remove Xaa-Pro dipeptides from the N-terminus of substrates. Fibroblast activation protein (FAP) and DPP8 are the other S9B peptidases.
DPP9 has been localized to human chromosome 19p13.3 (Olsen and Wagtmann 2002). The human DPP9 gene spans 48.6 kb and comprises 22 exons that are 53 bp to 1431 bp in length (Ajami et al. 2004). A predominant DPP9 mRNA transcript of 4.4 kb (AY374518; encoding 863 amino acids, the short form) is ubiquitous, with the highest levels in liver, heart, and skeletal muscle (Olsen and Wagtmann 2002; Ajami et al. 2004). A less abundant 5 kb transcript (AF542510; encoding 971 amino acids) that is abundant in skeletal muscle contains a second ATG translation start site that encodes a 892 amino acid protein (NM_139159.4), which is called the long form of DPP9 (DPP9-L) (Ajami et al. 2004).
Given that the crystal structures of DPP9 have not yet been solved, protein homology modeling has shown that DPP9 likely shares a similar tertiary structure with DPP4 and FAP. DPP9 is dimeric, with each monomer likely consisting of an α/β-hydrolase domain and an eight-blade β-propeller domain. The active site of the peptidase locates at the interface of these two domains, consisting of Ser730, Asp808, and His840 (Olsen and Wagtmann 2002; Ajami et al. 2004).
Until now, natural DPP9 has been purified only from bovine testes and was identified as the short form. Endogenous natural DPP9-L protein has not been isolated. Human recombinant DPP9, both the short and long forms, have been produced in various expression systems for research investigations and both forms are enzymatically active (Justa-Schuch et al. 2014; Zhang et al. 2015a, b).
Expression and Distribution
Proteins must be localized to their appropriate subcellular compartments to fully function. Different from DPP4 and FAP, which are cell surface expressed type II membrane glycoproteins, DPP9 lacks a transmembrane domain and is an intracellular protein. Inside the cell, the short and long forms of DPP9 localize differently. The intracellular distribution of the short form (DPP9-S) as a chimeric fluorescent protein in human hepatoma Huh7 cells and HeLa cells is diffuse, with some DPP9-S associated with mitochondria and with microtubules (Zhang et al. 2015a). Under steady-state conditions, DPP9-S was not seen at the plasma membrane, but upon stimulation with either phorbol 12-myristate 13-acetate or epidermal growth factor, some DPP9-S redistributes towards the inside of the ruffling membrane at the leading edge of moving cells (Zhang et al. 2015a). DPP9-L contains a nuclear localization sequence (NLS) near its N-terminus and is located in the cytoplasm and nucleus (Justa-Schuch et al. 2014). Overexpressed DPP9-L was shown not to be colocalized with a nuclear-rim protein RanBP2 and is wholly within the nucleus, not associated with the nuclear rim (Justa-Schuch et al. 2014).
DPP9 is ubiquitously expressed in normal tissues from mouse, human, baboon, cynomolgus monkey, and Sprague-Dawley rat (Zhang et al. 2013). Lymphocytes and epithelial cells from many organs, including lymph node, thymus, spleen, liver, lung, intestine, pancreas, muscle, and brain, express DPP9. The expression of DPP9 is altered in many disease conditions such as in liver disease, inflammatory bowel disease, and chronic lymphocytic leukemia (Zhang et al. 2013).
Regulation of DPP9 Activity
DPP9 is associated with H-Ras (Yao et al. 2011), and DPP9 activity reversibly increases in reducing conditions. DPP9 also interacts with small ubiquitin-related modifier 1 (SUMO1) and is allosterically regulated by SUMO1 binding (Pilla et al. 2013).
Proteases regulate many physiological processes by cleaving substrates. Synthetic substrates with a proline at P1 position have been widely used in the discovery of DPP9 inhibitors and in the investigations of DPP9 biochemical and biological functions. DPP9 is most active on Ala-Pro-, Arg-Pro-, and Gly-Pro- containing chromogenic (p-nitroaniline; pNA) or fluorogeneic substrates, such as 7-Amino-4-methylcoumarin (AMC) or 7-Amino-4-trifluoromethylcoumarin (AFC).
The cleavage of peptide hormones gastric inhibitory polypeptide (GIP) and glucagon-like peptide-1 (GLP-1) by DPP4 makes DPP4 inhibitors an effective treatment for type II diabetes. Therefore, the identification of DPP9 natural substrates can produce insights into biological function. The first natural substrate identified for DPP9 is the antigenic RU131–42 peptide which has a role in antigen presentation (Geiss-Friedlander et al. 2009). Mass spectrometry-based proteomic strategies have been successfully applied to identify substrates of DPP9 (Wilson et al. 2016). Using N-terminal amine isotopic labeling of substrates (TAILS), proteomic analysis of cytosolic proteins has revealed ten DPP9 substrates in cells stably expressing either enzymatically active or inactive DPP9 on a background of endogenous DPP9. Of these, nine were confirmed as having in vitro cleavable N-terminal peptides (Wilson et al. 2013), including several mitochondrial proteins. Thus, there are potential roles for DPP9 in cellular metabolism. Using 2D DIGE in a cell line expressing enzyme-inactive DPP9, nine proteins/peptides were confirmed as DPP9 substrates by MALDI-TOF or immunoblotting (Zhang et al. 2015b). Moreover, dipeptide Val-Ala was identified as a consensus site for DPP9 cleavage that was not recognized by DPP8, suggesting different in vivo roles for these closely related enzymes (Zhang et al. 2015b).
A DPP9 specific inhibitor would be a useful biochemical tool to identify DPP9 substrates and to study the functions of DPP9. However, there is none reported. Almost all the synthesized compounds targeting DPP9 also inhibit DPP8 to some extent, with 1G244 being the most potent one (IC50 values of 14 nM and 53 nM against DPP8 and DPP9, respectively) (Wu et al. 2009). A limitation for studies using 1G244 is that despite not inhibiting DPP4, DPP8 and DPP9 cannot be discriminated. Therefore, the development of a specific DPP9 inhibitor is of great urgency in the field of DPP9 research.
DPP9 Mouse Model and DPP9 in Metabolism
The only published DPP9 mouse line is a gene knock-in (gki) mouse line with a serine to alanine point mutation at the DPP9 active site (Ser729Ala). The nonsurvival of homozygous DPP9 gki mice shows an essential role of DPP9 enzyme activity in neonatal development. Pathology and histochemistry studies of embryos and neonates have not yet revealed a cause of death. The death of DPP9 gki neonates within hours of birth is intriguing. These DPP9 gki neonatal mice exhibit differential expression of genes involved in cell growth, innate immunity, and metabolic pathways in the DPP9 gki mice compared to wild-type littermates (Chen et al. 2016), suggesting that DPP9 enzyme activity is involved in gluconeogenesis and lipid metabolism in neonatal liver and gut.
A mitochondrial enzyme, adenylate kinase 2 (AK2), has been identified as a substrate for DPP9 and shown to colocalize with DPP9 (Wilson et al. 2013). DPP9 shows some colocalization with mitochondria (Zhang et al. 2015a) and potential substrates identified by 2D DIGE include several mitochondrial proteins (Zhang et al. 2015b), providing further evidence that DPP9 has access to mitochondrial proteins that may be substrates. Therefore, DPP9 probably has a role in energy metabolism.
Adipogenesis in preadipocytes can be blocked by a DPP8/DPP9 selective inhibitor 1G244 (Han et al. 2015). In addition, knockdown of DPP8 or DPP9 significantly impairs adipocyte differentiation from preadipocytes. Blocking the expression or activities of DPP8 and DPP9 attenuates PPARγ2 induction during preadipocyte differentiation. Addition of PPARγ agonist or ectopic expression of PPARγ2 is able to rescue the adipogenic defect caused by DPP8/DPP9 inhibition in preadipocytes (Han et al. 2015). Although the contribution of DPP9 compared to DPP8 cannot be elucidated due to the lack of a selective DPP9 inhibitor, this piece of data is indicative of the possible role of DPP9 in adipogenesis.
DPP9 in Cellular Functions and Implications for Cancer
Emerging evidence points to the role of DPP9 in a number of cellular processes, including extracellular matrix interactions, proliferation, and apoptosis (Yu et al. 2006, 2009; Yao et al. 2011; Zhang et al. 2015a). DPP9 is ubiquitously expressed in tumor cell lines and has so far been investigated in several types of tumors, including Ewing sarcoma, testicular, breast, and ovarian cancer, and leukemia and lymphoma (Ajami et al. 2004; Yu et al. 2009; Spagnuolo et al. 2013; Waumans et al. 2015). HEK293T cells overexpressing DPP9-S have impaired cell adhesion and monolayer wound healing, suggesting a role for DPP9 in cell adhesion and migration, but the underlying mechanisms remain elusive (Yao et al. 2011; Zhang et al. 2015a). Overexpression of either DPP9-S or DPP9-L causes increased apoptosis in human hepatoma Huh7 cells, suggesting overexpression provides limited and perhaps nonphysiological information on DPP9 function. Downregulating DPP9 by siRNA knockdown or enzyme inhibition of DPP9 by 1G244 in Huh7 cells has been found not to cause apoptosis, and these approaches target all endogenous DPP9, including both the short and long forms. DPP9-gene silencing or enzyme inhibition reduced cell adhesion and migration and expression of integrin-β1 and talin in Huh7 cells (Zhang et al. 2015a). There is a concomitant decrease in the phosphorylation of focal adhesion kinase and paxillin, indicating that reduced DPP9 suppressed the associated adhesion signaling pathway, thereby slowing cell movement (Zhang et al. 2015a). In Ewing Sarcoma, cell death can be enhanced by blocking DPP9 with either selective enzyme inhibition or siRNA knockdown, and DPP8/DPP9 inhibition enhances parthenolide’s tumor cell cytotoxicity (Spagnuolo et al. 2013). Similarly, inhibition of DPP8 and DPP9 probably contributes to the tumor regression induced by the compound Val-boro-Pro (Walsh et al. 2013). These findings may thus implicate DPP9 in tissue and tumor growth and metastasis. Possibly, the regulation of DPP9 in cell survival and proliferation is mediated by inhibiting Akt activation involving epidermal growth factor signaling pathway (Yao et al. 2011; Zhang et al. 2015a). However, selective inhibitors targeting DPP9 and conditional DPP9 knockout mice are needed to differentiate DPP9 from DPP8 for future investigations.
DPP9 in Immunology
The involvement of DPP9 in the immune system and the inflammatory response is an emerging interest. DPP9 has been implicated in lymphocyte proliferation and activation (Zhang et al. 2013; Waumans et al. 2015). DPP9 is expressed in lymphocytes and macrophages from immune organs and has been implicated in inflammatory diseases including atherosclerosis (Waumans et al. 2015). Particularly, DPP9 is expressed in macrophage-rich regions of atherosclerotic plaques, and DPP8/DPP9 inhibition attenuates macrophage activation, possibly due to the reduced IL-6 (interleukin-6) secretion (Waumans et al. 2016).
The identified substrates of DPP9 also suggest potential roles of DPP9. The antigenic RU131–42 peptide is the first identified natural substrate of DPP9, and it is possible that many more proline-containing antigens present in the cytoplasm are substrates for DPP9 (Geiss-Friedlander et al. 2009). DPP9 influences interferon gamma (IFNγ) secretion and antigen presentation on MHC class I molecules, suggesting a novel role for DPP9, but not DPP8, in antigen maturation and presentation (Geiss-Friedlander et al. 2009). Two important immune regulators, CXCL10 and IL1RA, have been shown to be DPP9 substrates (Zhang et al. 2015b). Cells and molecules of the immune system are a fundamental component of the tumor microenvironment. The chemokine CXCL10 is converted to its antagonist form by DPP4 in certain ovarian cancer tumors in which the active form of CXCL10 attracts lymphocytes that can attack tumor cells. The putative DPP9 substrates S100-A10, SET, and NUCB1 are mediators of immunity and/or inflammatory response (Zhang et al. 2015b), but a functional or physiological outcome of DPP9 cleavage has not been investigated. In the gki neonatal mouse, which lacks DPP9 enzyme activity, there is less TNF-α (tumor necrosis factor, alpha) and IL-1β (interleukin-1β) and more IκBα (nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha) in liver and gut (Chen et al. 2016), further pointing to a regulatory role for DPP9 in immune responses in vivo.
The understanding of DPP9 biology has made recent advances. The discovery of the DPP9 long form opens a new page of the DPP9 story revealing the potential regulation by and of DPP9 in nuclear events. Neonatal death of the DPP9 gene knockin mouse indicates a crucial role of DPP9 enzyme activity in survival. Recent examination of the underlying reasons for that neonatal death shows that DPP9 enzyme activity is involved in gluconeogenesis and lipid metabolism. DPP9 may also have important roles in cancers and immunology, and further investigations are needed on these topics.
MDG is supported by grants 1105238 and 1113842 from the Australian National Health and Medical Research Council.
- Han R, Wang X, Bachovchin W, Zukowska Z, Osborn JW. Inhibition of dipeptidyl peptidase 8/9 impairs preadipocyte differentiation. Sci Rep. 2015;5:12348. doi: 10.1038/srep12348. http://www.nature.com/articles/srep12348-supplementary-information
- Justa-Schuch D, Möller U, Geiss-Friedlander R. The amino terminus extension in the long dipeptidyl peptidase 9 isoform contains a nuclear localization signal targeting the active peptidase to the nucleus. Cell Mol Life Sci. 2014;71:3611–26. doi: 10.1007/s00018-014-1591-6.PubMedCrossRefGoogle Scholar
- Walsh MP, Duncan B, Larabee S, Krauss A, Davis JP, Cui Y, et al. Val-BoroPro accelerates T cell priming via modulation of dendritic cell trafficking resulting in complete regression of established murine tumors. PLoS One. 2013;8:e58860. doi: 10.1371/journal.pone.0058860.PubMedPubMedCentralCrossRefGoogle Scholar
- Waumans Y, Baerts L, Kehoe K, Lambeir A-M, De Meester I. The dipeptidyl peptidase family, prolyl oligopeptidase and prolyl carboxypeptidase in the immune system and inflammatory disease, including atherosclerosis. Front Immunol. 2015;6:387–405. doi: 10.3389/fimmu.2015.00387.PubMedPubMedCentralCrossRefGoogle Scholar
- Wilson CH, Indarto D, Doucet A, Pogson LD, Pitman MR, Menz RI, et al. Identifying natural substrates for dipeptidyl peptidase 8 (DP8) and DP9 using terminal amine isotopic labelling of substrates, TAILS, reveals in vivo roles in cellular homeostasis and energy metabolism. J Biol Chem. 2013;288:13936–49. doi: 10.1074/jbc.M112.445841.PubMedPubMedCentralCrossRefGoogle Scholar