2 REVIEW OF THE LITERATURE
2.2 Hyaluronan synthesis
3 2.1.2 Structure of hyaluronan
Hyaluronan is composed of repeating disaccharides, D-‐‑glucuronic acid and N-‐‑acetyl-‐‑
glucosamine (Figure 1). There are alternating β(1-‐‑3) and β(1-‐‑4)glucuronidic bonds between the sugars (Weissmann et al. 1954). Hyaluronan chain may contain up to 25 000 disaccharides, and molecular mass of one hyaluronan molecule can be as high as 107 Da and length 25 µμm (Toole 2004). At physiological pH hyaluronan is highly hydrophilic due to negatively charged glucuronic acid groups. In aqueous solutions linear hyaluronan chain forms a random coiled structure. Unlike the other glycosaminoglycans, hyaluronan does not covalently attach to core protein to form proteoglycans.
Figure 1. The general chemical structure of the disaccharide unit of hyaluronan. Hyaluronan is composed of alternating residues of β-‐‑D-‐‑(1→3) glucuronic acid (GlcA) and β-‐‑D-‐‑(1→4)-‐‑N-‐‑
acetylglucosamine (GlcNAc).
2.2 HYALURONAN SYNTHESIS
2.2.1 Hyaluronan is synthesized on the plasma membrane
Hyaluronan is synthesized by hyaluronan synthases (HAS), integral transmembrane proteins that act on the inner face of the plasma membrane and extrude the growing hyaluronan chain through the plasma membrane into the extracellular space (Prehm 1984).
Mammals have three hyaluronan synthase isoenzymes, HAS1, HAS2 and HAS3 (Toole 2004). These enzymes utilize two precursors, UDP-‐‑N-‐‑acetylglucosamine and UDP-‐‑
glucuronic acid for hyaluronan synthesis. The new sugar units are added into the reducing end of the growing chain by the native vertebrate enzyme (Weigel et al. 1997). Studies with a recombinant enzyme and the enzyme from Pasteurella multocida show chain growth in the non-‐‑reducing end (Bodevin-‐‑Authelet et al. 2005, DeAngelis 1999).
The amino acid sequences of HAS isoenzymes are quite homologous between different species. The predicted structure of the enzyme consists of 4-‐‑6 transmembrane domains and 1-‐‑2 membrane associated domains (Figure 2). A large cytoplasmic domain is suggested to contain the enzymatically active area. The transmembrane domains that span the lipid
4
bilayer are suggested to create a pore in the plasma membrane for the protruding hyaluronan chain (Weigel et al. 1997). There has been speculation on the structure of the pore. A recent review by Weigel concludes that the HAS enzyme itself forms the pore (Weigel 2015), while previous studies using inhibition of multidrug resistance transporters had suggested that they are involved in the export of hyaluronan from the inner to the outer surface of plasmamembrane (Schulz et al. 2007).
Figure 2. A schematic structure of vertebrate hyaluronan synthase hyaluronan binding proteins and degradation enzymes. The synthase is composed of 7 transmembrane or membrane-‐‑associated domains and a large cytoplasmic domain, latter assumed to contain the enzymatic activity. Aggrecan is an example of extracellular hyaluronan binding protein.
CD44 is the main cell surface receptor for hyaluronan. Hyaluronidase 2 is a cell surface degradation enzyme of hyaluronan. Modified from Itano and Kimata (Itano and Kimata 2002), Anderegg et al. (Anderegg et al. 2014) and Chowdhury et al. (Chowdhury et al.
2016).
The human HAS1 gene is localized in chromosome 19 and the mouse gene in chromosome 17 (Spicer and McDonald 1998). HAS1 is the isoenzyme with lowest activity (Itano et al.
1999), (Rilla et al. 2013a), and HAS1 knockout mice have no apparent phenotype (Kobayashi et al. 2010). The hyaluronan chains synthesized by HAS1 are suggested to be smaller as compared to those of HAS2 (Itano et al. 1999). Hyaluronan production by HAS1 is highly dependent of the intracellular UDP-‐‑sugar concentration and high concentrations are necessary for full enzymatic activity (Rilla et al. 2013a). TGF-‐‑β stimulated synoviocytes
5
have elevated HAS1 expression and increased hyaluronan production (Stuhlmeier and Pollaschek 2004a). HAS1 promoter contains SP3 and SMAD3 elements and these elements regulate the level of its expression(Chen et al. 2012).
High expression of HAS1 is typical for some diseases like rheumatoid arthritis (Stuhlmeier and Pollaschek 2004b), osteoarthritis (Lambert et al. 2014) and infectious lung disease (Chang et al. 2014). Bone marrow mesenchymal progenitor cells isolated from myeloma patients have higher HAS1 mRNA expression as compared to cells collected from healthy people, with a corresponding elevation in HA production (Calabro et al. 2002).
The location of the human HAS2 gene is in chromosome 8 and that of mouse in chromosome 15 (Spicer and McDonald 1998). HAS2 deletion has lethal effects during embryonic development. HAS2 produces large hyaluronan polymers with an average molecular mass of >2 MDalton (Itano et al. 1999). The availability of UDP-‐‑HexNac in the cytosol limits the synthesis rate of hyaluronan and feedback regulates the expression of HAS2 (Jokela et al. 2011). On the other hand extracellularly applied UDP-‐‑Glucose activates HAS2 expression by binding to the P2Y14-‐‑ plasma membrane receptor, leading to the phosphorylation of STAT3 in tyrosine 705 and binding to the promoter of HAS2 (Jokela et al. 2014).
Interestingly, HAS2 is overexpressed in the hereditary cutaneous mucinosis of Shar Pei dogs (Zanna et al. 2009), and fibroblasts from Shar Pei dogs have higher numbers of plasma membrane protrusions (Docampo et al. 2011). The lifespan of naked mole rat is the longest among rodents, even exceeding 30 years (Buffenstein and Jarvis 2002). Skin fibroblasts of naked mole rat have high expression levels for HAS2 and produce extremely high molecular weight hyaluronan, which accumulates in the subcutaneous tissue (Tian et al.
2013).
HAS2 is overexpressed in fibroblasts isolated from patients suffering idiopathic pulmonary fibrosis. These fibroblasts are more invasive compared to fibroblasts from healthy people. This invasion capacity is regulated by CD44 (Li et al. 2011).
Human and mouse HAS3 genes are localized in chromosomes 16 and 8, respectively (Spicer and McDonald 1998). HAS3 knockout mice are viable and have no specific morphological phenotype (Bai et al. 2005), but they have epileptic phenotype (Arranz et al.
2014). Hyaluronan produced by HAS3 is usually shorter than hyaluronan made by HAS1 and by HAS2 (Brinck and Heldin 1999, Itano et al. 1999). The promoter area of HAS3 has been recently characterized. There are binding sites for C/EBP and NFκB and Sp1, which seem to be essential for promoter activity (Wang et al. 2015). HAS3 is abundant on the plasma membrane (Rilla et al. 2005), and a specific feature of HAS3 is its accumulation into plasma membrane protrusions that collapse after hyaluronidase digestion or inhibition of hyaluronan synthesis (Kultti et al. 2006). This suggests that the protrusions are dependent on HAS3 activity.
2.2.5 Regulation of hyaluronan synthesis
Hyaluronan synthesis is stimulated in many physiological and pathological states, like in inflammation, after tissue injury and during tumor progression (Cyphert et al. 2015). The HAS expression is regulated by numerous local and systemic stimuli, like growth factors, cytokines and hormones.
The effects of growth factors on HAS activity are mainly mediated at the transcriptional level, since they induce rapid changes in HAS mRNA levels, usually associated with a simultaneous increase in hyaluronan synthesis (Jacobson et al. 2000, Karvinen et al. 2003b, Pienimäki et al. 2001, Yamada et al. 2004). For example, keratinocyte growth factor (KGF) (Karvinen et al. 2003b) and epidermal growth factor (EGF) (Pienimäki et al. 2001) stimulate the expression of HAS2 and HAS3 in keratinocytes. In fact, HAS2 is one of the direct target genes for EGF signaling (Saavalainen et al. 2005). Examples of hormones that induce HAS
6
expression are estrogen (Tellbach et al. 2002) and progesterone (Uchiyama et al. 2005). Also several cytokines, like interleukin-‐‑ 1β (IL-‐‑1β) upregulate HAS expression and hyaluronan synthesis in many cell types, like fibroblasts (Yamada et al. 2004), endothelial cells (Vigetti et al. 2010) and lung adenocarcinoma cells (Chow et al. 2010). Hydrocortisone inhibits hyaluronan synthesis in human epidermis (Ågren et al. 1995). Glucocorticoids almost totally block HAS2 expression in dermal fibroblasts (Zhang et al. 2000)
4-‐‑methylumbelliferone (4-‐‑MU) has been reported to specifically inhibit hyaluronan synthesis in cultured mammalian cells (Kosaki et al. 1999a, Kultti et al. 2009, Nakamura et al. 1995, Nakamura et al. 1997, Sohara et al. 2001) and in Streptococcus equi FM100 cells (Kakizaki et al. 2002). It inhibits melanoma cell adhesion and locomotion (Kudo et al. 2004), and metastasis (Yoshihara et al. 2005). 4-‐‑MU also reverses the effect of HAS2 transfection on hyaluronan synthesis and colony formation of tumor cells (Kosaki et al. 1999a). Nowadays 4-‐‑MU is widely utilized as a research tool to inhibit hyaluronan synthesis.
At post-‐‑transcriptional level the availability of UDP-‐‑sugar precursors is a potential regulator of HAS activity (Itano et al. 1999). This hypothesis is supported with a finding that depletion of the UDP-‐‑glucuronic acid pool through glucuronidation of 4-‐‑MU (Kakizaki et al. 2004) reduces hyaluronan synthesis rate. Also reduction of UDP-‐‑GlcNA by mannose reduces hyaluronan synthesis rate (Jokela et al. 2011). In general, the availability of both UDP-‐‑sugars regulates the activity of hyaluronan synthesis (Deen et al. 2016, Jokela et al.
2011, Kakizaki et al. 2004, Kultti et al. 2009, Rilla et al. 2013a, Vigetti et al. 2009).
It has been suggested that hyaluronan synthases can form homo-‐‑ and heteromers in plasma membrane, which offers one more possible way of regulation for HAS activity (Bart et al. 2015, Karousou et al. 2010). Bart and coworkers showed reduced hyaluronan synthesis in HAS2 and HAS3 overexpressing cells cotransfected with Has1 (Bart et al. 2015). Other putative factors regulating HAS activity are post-‐‑transcriptional modifications in HAS2 enzyme, including phosphorylation (Goentzel et al. 2006, Vigetti et al. 2011), ubiquitination (Karousou et al. 2010) and O-‐‑GlcNAcylation (Deen et al. 2016, Vigetti et al. 2012).
Because HASs are known to be active only in the plasma membrane (Rilla et al. 2005), their traffic is potentially an important post-‐‑transcriptional factor regulating hyaluronan synthesis (Deen et al. 2014). All HAS isoenzymes follow the normal route for transmembrane proteins, travelling from ER to Golgi apparatus and further to the plasma membrane (Müllegger et al. 2003). Regulation of HAS trafficking is not fully understood, but posttranslational modifications of HAS and intracellular UDP-‐‑sugar levels affect it (Müllegger et al. 2003, Siiskonen et al. 2014). HAS1 seems to be less present on the plasma membrane (Siiskonen et al. 2014) than HAS2 and HAS3 (Rilla et al. 2005, Siiskonen et al.
2014). Accordingly, HAS2 and especially HAS3 are more active in inducing plasma membrane protrusions (Kultti et al. 2006, Rilla et al. 2005) and secretion of extracellular vesicles (Rilla et al. 2013b, Rilla et al. 2014).