Hyaluronan-positive cell protrusions and microvesicles : studies on their structure, function and existence in vivo

(1)

VILLE KOISTINEN

HYALURONAN-POSITIVE CELL PROTRUSIONS AND MICROVESICLES

Dissertations in Health Sciences

(2)

(3)

Hyaluronan-‐‑positive cell protrusions and microvesicles

(4)

(5)

VILLE KOISTINEN

Hyaluronan-‐‑positive cell protrusions and microvesicles

Studies on their structure, function and existence in vivo

To be presented by permission of the Faculty of Health Sciences, University of Eastern Finland for public examination in lecture hall SN200, Kuopio, on Wednesday, June 21^st 2017, at 12 noon

Publications of the University of Eastern Finland Dissertations in Health Sciences

Number 425

Department of Biomedicine, School of Medicine, Faculty of Health Sciences, University of Eastern Finland

Kuopio 2017

(6)

Kiriprintti OY Helsinki, 2017

Series Editors:

Professor Tomi Laitinen, M.D., Ph.D.

Institute of Clinical Medicine, Clinical Physiology and Nuclear Medicine Faculty of Health Sciences

Professor Hannele Turunen, Ph.D.

Department of Nursing Science Faculty of Health Sciences

Professor Kai Kaarniranta, M.D., Ph.D.

Institute of Clinical Medicine, Ophthalmology Faculty of Health Sciences

Associate Professor (Tenure Track) Tarja Malm, Ph.D.

A.I. Virtanen Institute for Molecular Sciences Faculty of Health Sciences

Lecturer Veli-‐‑Pekka Ranta, Ph.D. (pharmacy) School of Pharmacy

Faculty of Health Sciences

Distributor:

University of Eastern Finland Kuopio Campus Library

P.O.Box 1627 FI-‐‑70211 Kuopio, Finland http://www.uef.fi/kirjasto

ISBN (print): 978-‐‑952-‐‑61-‐‑2527-‐‑5 ISBN (pdf): 978-‐‑952-‐‑61-‐‑2528-‐‑2

ISSN (print): 1798-‐‑5706, ISSN (pdf): 1798-‐‑5714

ISSN-‐‑L: 1798-‐‑5706

(7)

Author’s address: Department of Biomedicine/School of medicine/Anatomy University of Eastern Finland

KUOPIO FINLAND

Supervisors: Docent Kirsi Rilla, Ph.D.

Department of Biomedicine/School of medicine/Anatomy University of Eastern Finland

KUOPIO FINLAND

Professor Markku Tammi, M.D. Ph.D.

Department of Biomedicine/School of medicine/Anatomy University of Eastern Finland

KUOPIO FINLAND

Reviewers: Affiliate Professor Thomas N. Wight, Ph.D.

Institute for Stem Cell & Regenerative Medicine University of Washington

SEATTLE USA

Docent Varpu Marjomäki, Ph.D.

Department of Biological and Environmental Science University of Jyväskylä

JYVÄSKYLÄ FINLAND

Opponent: Professor Pekka Lappalainen

Institute of biotechnology

University of Helsinki

HELSINKI FINLAND

(8)

(9)

Koistinen, Ville

Hyaluronan positive cell protrusions and microvesicles, Studies on their structure, function and existence in vivo

University of Eastern Finland, Faculty of Health Sciences

Publications of the University of Eastern Finland. Dissertations in Health Sciences 425. 2017. 106 p.

ISBN (print): 978-‐‑952-‐‑61-‐‑2527-‐‑5 ISBN (pdf): 978-‐‑952-‐‑61-‐‑2528-‐‑2 ISSN (print): 1798-‐‑5706, ISSN (pdf): 1798-‐‑5714 ISSN-‐‑L: 1798-‐‑5706 ABSTRACT

Hyaluronan is a large unbranched glycosaminoglycan. The molecular mass of a single molecule can reach 10 million daltons and a stretched length of up to 25 µm. Hyaluronan is very hydrophilic due to a negative charge and it forms a highly viscous solution in water.

While hyaluronan has an important role as a space filler in tissues, it also contributes to essential cellular processes, including embryonic development, cell migration, proliferation, wound healing and cancer progression.

Hyaluronan is synthesized by three plasma membrane proteins, called hyaluronan synthases (HAS1-‐‑3). Previous studies on cultured cells have shown that overexpression of HAS2 and HAS3 induce the formation of extensive microvilli. The microvilli are covered with a hyaluronan coat and are dependent on filamentous actin. Nevertheless, several questions about the microvilli remain; do these structures exist in vivo? What is the exact role of hyaluronan in their maintenance? What might be their function? Therefore, the aim of this dissertation was to study the ultrastructure and function of the hyaluronan coated microvilli in more detail, and to investigate if these special structures exist in vivo.

The results show that HAS overexpression drives the assembly of actin filaments in the cell cortex and into the protrusions. The core of the HAS-‐‑induced microvilli consist of only

~8 actin microfilaments which implies that in order to maintain such long and slender protrusions, additional support is needed. Indeed, the hyaluronan coat around the microvilli acts as an extracellular cytoskeleton. Therefore, active hyaluronan synthesis is required for the growth and maintenance of the microvilli. The activity of hyaluronan synthesis was also found to correlate with the shedding of microvesicles, covered by hyaluronan, and budded off from the tips of the microvilli.

The HAS-‐‑induced microvilli in cultured cells, and the microvilli on mesothelial surfaces were ultrastructurally similar. Mesothelium was also positive for hyaluronan, but negative for the main hyaluronan receptor CD44. However, epithelial to mesenchymal transition induced by mesothelial wounding or epidermal growth factor increased hyaluronan synthesis, the expression of CD44 and HASs, the formation of microvilli and the shedding of microvesicles. These findings suggest that the HAS-‐‑induced microvilli serve as a source of microvesicles, a recently discovered vehicle for the transport of signals in the regulation of wound healing and progression of cancer. CD44 positive microvesicles could serve as a source for new biomarkers to detect EMT-‐‑related processes in tissue injuries and cancer.

In the future, the HAS-‐‑induced microvilli and microvesicles may prove to be a useful tool in clinical applications such as cancer diagnosis and therapy.

National Library of Medicine Classification: QU 83, QU 107, QU55.3, QU135, QU143, QU350

Medical Subject Headings: CD44; Cell Culture; Cell movement; Enzyme activation, Epidermal growth factor, Extracellular Matrix; Glycosaminoglycans; Hyaluronan; Microvilli; Rat

(10)

(11)

Koistinen, Ville

Hyaluronaanipositiiviset solu-‐‑ulokkeet ja mikrovesikkelit. Tutkimuksia niiden rakenteesta, toiminnasta ja esiintymisestä elävissä organismeissa

Itä-‐‑Suomen yliopisto, terveystieteiden tiedekunta

Publications of the University of Eastern Finland. Dissertations in Health Sciences 425. 2017. 106 s.

ISBN (print): 978-‐‑952-‐‑61-‐‑2527-‐‑5 ISBN (pdf): 978-‐‑952-‐‑61-‐‑2528-‐‑2 ISSN (print): 1798-‐‑5706, ISSN (pdf): 1798-‐‑5714 ISSN-‐‑L: 1798-‐‑5706 TIIVISTELMÄ

Hyaluronaani on vuorottelevista N-‐‑asetyyli-‐‑D-‐‑glukosamiini-‐‑ ja D-‐‑glukuronihappo-‐‑

molekyyleistä koostuva suuri sokeriketju. Hyaluronaani voi olla jopa 10 miljoonan daltonin kokoinen ja venytettynä molekyyli voi olla jopa 25 µm pituinen. Negatiivisen kokonaisvarauksensa takia se sitoo runsaasti vettä, ja muodostaakin hyvin viskoosin vesiliuoksen. Hyaluronaania tuottaa kolme solukalvolla toimivaa entsyymiä, hyaluronaani-‐‑

syntaasia (HAS1-‐‑3). Hyaluronaani on keskeinen molekyyli monissa elimistön tärkeissä toiminnoissa, kuten yksilön kehityksessä, solujen liikkumisessa, haavan paranemisessa ja syövän etenemisessä. Sillä on myös tärkeä rooli soluvälitilan täyttäjänä.

Aikaisemmissa tutkimuksissa havaittiin, että hyaluronaanisyntaasien HAS2 tai HAS3 ylituotanto aiheuttaa pitkien ja ohuiden hyaluronaanivaipallisten solu-‐‑ulokkeiden muodos-‐‑

tumisen. Tällaisia rakenteita ei ole aikaisemmin kuvattu esiintyvän elimistössä. Tämän väitöstutkimuksen tarkoitus oli tutkia näiden solu-‐‑ulokkeiden rakennetta, toimintaa ja hyaluronaanin merkitystä niiden muodostumisessa. Lisäksi tavoitteena oli selvittää esiintyykö samanlaisia solu-‐‑ulokerakenteita in vivo.

HAS3:n ylituotannon havaittiin aiheuttavan soluissa aktiinitukirangan uudelleen järjes-‐‑

täytymisen solun reunaosiin sekä solu-‐‑ulokkeisiin. Solu-‐‑ulokkeet sisältävät keskimäärin vain 8 aktiinisäiettä, joten ne tarvitsevat lisätukea pysyäkseen pystyasennossa.

Hyaluronaanivaipan havaittiin toimivan solu-‐‑ulokkeiden ulkoisena tukirankana. Käyttäen tutkimusmateriaalina runsaasti hyaluronaania tuottavia soluviljelmiä havaittiin, että aktiivinen hyaluronaanin tuotanto on sekä solu-‐‑ulokkeiden muodostumisen edellytys että seuraus. Hyaluronaanin tuotannon havaittiin myös korreloivan solunulkoisten kalvo-‐‑

rakkuloiden muodostumiseen. Tarkemmat selvitykset paljastivat, että solukalvorakkulat irtautuvat hyaluronaanivaipallisten solu-‐‑ulokkeiden kärjistä.

Hyaluronaanivaipallisten solu-‐‑ulokkeiden ja mesoteelisolujen mikrovillusten havaittiin olevan rakenteellisesti samankaltaisia. Mesoteelin hyaluronaanivärjäys osoitti, että rotan vatsakalvon mesoteelisolujen apikaalisella, solu-‐‑ulokkeita sisältävällä pinnalla on hyaluronaania, mutta ei CD44-‐‑molekyyliä, joka on hyaluronaanin tärkein reseptori ja ilmenee yleensä yhdessä hyaluronaanin kanssa.

Epiteeli-‐‑mesenkyymivaihdos (EMT) on tärkeä prosessi, joka liittyy muun muassa haavan paranemiseen ja syövän etenemiseen. Tässä tutkimuksessa havaittiin, että EMT:ssä solut alkavat tuottaa CD44:ä, muodostaa runsaasti solu-‐‑ulokkeita, hyaluronaania ja solukalvorakkuloita. Tulokset osoittavat, että hyaluronaanivaipallisilla solu-‐‑ulokkeilla voi olla tärkeä merkitys haavan paranemisessa ja syövässä. CD44 sisältävät solukalvorakkulat voivat toimia EMT:hen liittyvän kudovaurion ja syövän uutena biomarkkerina

Luokitus: QU 83, QU 107, QU55.3, QU135, QU143, QU350

Yleinen Suomalainen asiasanasto: hyaluronaani; solut; soluväliaine, rotta

(12)

(13)

Elämäni valolle

(14)

(15)

Acknowledgements

I finally did it!

This thesis work was carried out in the Institute of Biomedicine/Anatomy, School of Medicine at the University of Eastern Finland during the years 2009-‐‑2017. I want to thank all the people who helped me with my thesis work.

First of all, I want to express my deep gratitude to my principal supervisor, docent Kirsi Rilla, Ph.D., with whom we share a similar sense of humor and uttermost interest to cell protrusions. I admire your patience and help during these years and your skills on microscopy.

My warmest thanks go to my second supervisor, professor Markku Tammi, Ph.D., M.D., as humble, I keep marveling your expertise on hyaluronan science. I also thank you for your encouraging attitude towards my work.

I also want to thank professor Raija Tammi Ph.D., M.D. for her vast knowledge on immunohistochemistry, Sanna Oikari, Ph.D. and Tiina Jokela Ph.D. for their generous help concerning molecular biological and statistical methods.

Very warm thanks go also to my research colleagues Leena Rauhala, Ph.D., and Lasse Hämäläinen M.Sc. for helping diverse methodological issues. I want to express my gratitude the rest of the members of extracellular vesicle research group, especially to Kai Härkönen M.Sc. and Uma Thanigai Arasu M.Sc.

I am deeply grateful to my official reviewers docent Varpu Marjomäki, Ph.D. and professor Thomas Wight, Ph.D. for their careful review and essential suggestion and encouraging comments, which made this thesis considerable more readable and better. My sincere thanks go also to Gina Galli, Ph.D. for her detailed revision of the English language.

Special thanks go to the technical staff members. Virpi Miettinen, I keep wondering your exquisite skills on electron microscopy sample preparation, I asked you impossible, but still you did it! Riikka Kärnä, M.Sc., I admire your technical skills and precise pipette handling and Eija Rahunen, I look up to your firm expertise on immunohistochemical methods.

I also thank rest of the staff, Kari Kotikumpu, Eija Vartiainen, Arja Venäläinen, Eija Kettunen, Arja Winberg.

Warm thanks goes also to the rest of my co-‐‑authors Arto Koistinen Ph.D., Antti Arjonen Ph.D., Docent Sanna Pasonen-‐‑Seppänen Ph.D., Ashik Jawahar Deen Ph.D., Sara Wojciechowski M.Sc., Genevieve Bart Ph.D. and Kari Törrönen Ph.D.

I also want to thank all other members in hyaluronan research group and the whole personnel in the Institute of Biomedicine.

(16)

My superiors in the clinical work are thanked for their flexibility concerning my scientific work, Riitta Johannala-‐‑Kemppainen, Arja Jukkola-‐‑Vuorinen, M.D., Ph.D, Jussi Männistö, M.D., Ph.D. and Mika Huuskonen, M.D. Special and warm thanks go to my clinical mentors Sanna Kosonen, M.D., Laura Pusa, M.D. and docent Timo Muhonen, Ph.D, M.D., for the outstanding guidance for the fascinating world of oncology.

Lisäksi haluan kiittää vanhempiani Antti ja Seija Koistista heidän antamastaan tuesta ja rakkaudesta, siskoani Reeta Koistista sekä perheemme prinsessaa Tyttiä hauskoista hetkistä yhdessä. Haluan kiittää rakasta ystävääni ja kämppäkaveria LL Maija Peippoa tieteellisistä keskusteluista ja käsikirjoitusten tarkastamisesta sekä mukavista ja ei-‐‑niin-‐‑mukavista hetkistä kliinisen työmme alkutaipaleella. Kiitokset kuuluvat myös ystävilleni Lauri Jokiaiselle, TM Jonna Ojalammille, LL Hanna Jokelalle ja LL Hannes Holmalle. Sekä serkulleni Sari Hassiselle, joka tutustutti minut valokuvaukseen ja sai välillä unohtamaan tämän projektin.

Lopuksi kiitää elämäni valoa, LL Jukka Viikaria. Tämä työ vaati paljon, jaksoit kuitenkin aina tukea ja pitää minut maan pinnalla, kiitos!

Academy of Finland, Cancer center of Eastern finland, Sigrid Juselius foundations, North Savo cancer foundation are thanked for the financial support of this thesis

Kotka, May 2017

Ville Koistinen

(17)

List of the original publications

This dissertation is based on the following original publications:

I Koistinen V, Kärnä R, Koistinen A, Arjonen A, Tammi R, Rilla K. Cell protrusions induced by hyaluronan synthase 3 (HAS3) resemble mesothelial microvilli and share cytoskeletal features of filopodia. Exp Cell Res 337: 179-‐‑191, 2015.

II Koistinen V, Jokela T, Oikari S, Kärnä R. Tammi M, Rilla K. Hyaluronan positive plasma membrane protrusions exist on mesothelial cells in vivo. Histochem Cell Biol 145(5): 531-‐‑544, 2016.

III Koistinen V, Härkönen K, Kärnä R, Arasu Uma Thanigai, Oikari S Rilla K. EMT induced by EGF and wounding activates hyaluronan synthesis machinery and EV shedding in rat primary mesothelial cells. Matrix Biol., DOI:

10.1016/j.matbio.2016.12.00, 2016.

IV Rilla K, Pasonen-‐‑Seppänen S, Deen AJ, Koistinen VV, Wojciechowski S, Oikari S, Kärnä R, Bart G, Törrönen K, Tammi RH, Tammi MI. Hyaluronan production enchances shedding of plasma membrane-‐‑derived microvesicles. Exp Cell Res, 319: 2006-‐‑2018, 2013.

The publications were adapted with the permission of the copyright owners.

(18)

(19)

Abbreviations

4-‐‑MU 4-‐‑methylumbelliferone AC adenylyl cyclase

AMPK AMP-‐‑activated protein kinase atRA all-‐‑trans-‐‑retinoic acid

BSA Bovine serum albumin BVHyal Bee venom hyaluronidase cAMP cyclic adenosine mono

phosphate

CD44 Cluster of differentiation 44 CREB1 Cyclic AMP-‐‑responsive

element binding protein 1 CTGF connective tissue growth

factor

ECM Extracellular matrix EGF Epidermal growth factor EGFR Epidermal growth factor

receptor

EM Electron microscope

EMT Epithelial-‐‑to-‐‑mesenchymal transition

ER Endoplastic reticulum

ErbB2 Receptor tyrosine-‐‑protein kinase 2

EV Extracellular vesicle GFP Green fluorescent protein GT2 Glycosyltransferase module 2 HA Hyaluronan

HABP hyaluronan binding protein HARE Hyaluronan receptor for

Endocytosis

HAS-‐‑rs hyaluronan synthase related sequence

HB-‐‑EGF Heparin-‐‑binding epidermal growth factor like growth factor

HBSS Hank’s balanced salt solution HGF Hepatocyte growth factor HRG Heregulin

HYAL Hyaluronidase

IαI Inter-‐‑alpha-‐‑trypsin inhibitor IBD inflammatory bowel disease IL-‐‑1b Interleucin-‐‑1-‐‑beta

(22)

LYVE-‐‑1 Lymphatic vessel endothelial hyaluronan receptor

MTE motif ten elements

NF-‐‑kb Nuclear factor kappa-‐‑beta pathway

OGT O-‐‑GlcNAc transferase p65 Protein 65

PDGF-‐‑D Platelet derived growth factor-‐‑D

PGE2 prostaglandin E2

PH-‐‑20 Posterior head antibody PMA phorbol 12-‐‑myristate 13-‐‑

acetate

PTC renal proximal tubular epithelial cells

RAR nuclear retinoic acid receptor

RHAMM Receptor for hyaluronan-‐‑

mediated motility

SAP shrimp alkaline phosphatase Sp3 Secificity protein 3

STAT3 Signal transducer and activator of transcription 3 TGF-‐‑beta Transforming growth factor

beta

TIS transcription initiation site TNF-‐‑α/β Tumor necrosis factor

alfa/beta

TSG-‐‑6 Tumor necrosis factor stimulated gene 6 UDP Uridine diphosphate

UGDH UDP-‐‑glucose dehydrogenase YY1 Yin-‐‑Yang1 (transcription

factor)

(23)

1 Introduction

Hyaluronan is a linear macromolecule made of alternating N-‐‑acetyl-‐‑D-‐‑glucosamine (GlcNAc) and D-‐‑glucuronic acid (GlcUA), linked together via β-‐‑1,4 and β1,3 linkages. In mammals, hyaluronan is produced on the inner surface of the plasma membrane by three hyaluronan synthases (HAS1-‐‑3). Hyaluronan is bound to the plasma membrane by HASs and hyaluronan receptors including CD44 and RHAMM, and degraded by three hyaluronidases (HYAL1-‐‑2 and PH20) (Itano et al. 1999; Tien and Spicer 2005).

Hyaluronan has several functions. In many cases it forms a pericellular matrix around the cells (Evanko et al. 2007). This matrix is highly organized, can be cross-‐‑linked, and further extended by HA-‐‑binding proteins including versican and aggrecan (Day and de la Motte 2005; Evanko et al. 1999; Morgelin et al. 1995). Hyaluronan coat is often found in the pericellular matrices of dividing cells, and has been hypothesized to take part in the separation of dividing cells (Evanko et al. 2007). Additionally, hyaluronan synthesis correlates with the migration rate of cells (Lee and Spicer 2000; Spicer and Tien 2004).

Elevated hyaluronan production has been observed also in several inflammational processes (de La Motte et al. 1999; Majors et al. 2003; Teder et al. 2002).

Overexpression of GFP-‐‑linked HAS2-‐‑3 was observed to induce actin-‐‑dependent formation of numerous cell protrusions, that are coated with hyaluronan (Kultti et al. 2006).

Spontaneous formation of hyaluronan coat containing microvilli have been observed in cultured human embryonic lung and murine fibrosarcoma cells (Bard et al. 1983), chondrosarcoma cells, mesothelial cells (Rilla et al. 2008), and mesenchymal stem cells (Qu et al. 2014). These findings suggest that hyaluronan coated cell protrusions also exist in cells with naturally high levels of hyaluronan secretion.

Mesothelial cells form a monolayer in pleural, peritoneal and pericardial cavities and on the parietal surfaces of the internal organs. They lie on a thin basement membrane supported by sub-‐‑serosal connective tissue (Albertine et al. 1982; Ishihara et al. 1980; Wang 1974). Even though, mesothelial cells originate from mesodermal cells, they have several epithelial-‐‑like features, including a cobblestone polygonal morphology and thick microvillar border on the luminal surface (Czernobilsky et al. 1985; Mutsaers 2002). The mesothelial ultrastructure varies according to its location (Michailova et al. 1999). The microvillar border is remarkably dynamic with variable numbers and lengths of the microvilli under different physiological states (Madison et al. 1979; Mutsaers et al. 1996).

Mesothelial cells are capable of producing a hyaluronan-‐‑rich glycocalyx in vitro (Breborowicz et al. 1996; Breborowicz et al. 1998; Heldin and Pertoft 1993; Yung et al. 1994;

Yung et al. 2000). Some reports also indicate that mesothelium is capable of synthesizing hyaluronan in vivo (Breborowicz et al. 1998; Lai et al. 1999; Wang and Lai-‐‑Fook 1998; Yung et al. 1994). However, ultra-‐‑structural localization of mesothelial tissue hyaluronan and hyaluronan synthases have not yet been shown.

This thesis concentrated on the structure and function of the HAS-‐‑induced cell protrusions. The results show that the hyaluronan coat acts as an extracellular cytoskeleton to support the long and slender cell protrusions. This work also shows that hyaluronan coated cell protrusions exist in vivo on mesothelial cells, and may act as microvesicle-‐‑

forming organelles.

(24)

(25)

2 Review of the literature

2.1 HYALURONAN – HISTORY, STRUCTURE AND PROPERTIES

The story of glycosaminoglycans (GAGs), a family of animal polysaccharides which contain hyaluronan (HA), began in 1841, when the German anatomist Henle described the ground substance; an amorphous material between cells (Henle 1841). One of the main constituents of the ground substance was first identified in 1934 as a carbohydrate containing uronic acid and amino sugar, named hyaluronic acid (Meyer and Palmer 1934). It took a further 20 years before the chemical structure of hyaluronic acid was determined (Weissmann and Meyer 1954). The term “hyaluronan” was introduced in 1986 to cover all different forms of the molecule, hyaluronic acid and its salts such as sodium hyaluronate (Balazs et al. 1986).

The chemical structure of hyaluronan is simple, consisting of repeating disaccharide units of N-‐‑acetyl glucosamine and D-‐‑glucuronic acid linked with alternating β-‐‑1,4 and β-‐‑1,3 glycosidic bonds (Figure 1). The molecular weight can be as high as 6000-‐‑8000 kDa, and a stretched 6000 kDa hyaluronan chain would have an approximate length of 15 µm and a diameter of approximately 0.5 nm (Cowman and Matsuoka 2005). In a normal physiological state, it consists of 2,000-‐‑25,000 disaccharides with a polymer length of 2-‐‑25 µm (Toole 2004). Both sugar molecule units in the hyaluronan chain are derivatives of glucose. Because they are in the β-‐‑configuration, the hydroxyl and carboxyl groups in addition to the anomeric carbon are in a sterically favorable equatorial position, while the hydrogen atoms occupy the less sterically favorable axial positions. This in turn makes the hyaluronan molecule very stable (Necas et al. 2008).

Figure 1. The chemical structure of hyaluronan consisting of N-acetylglucosamine and D-glucuronic acid linked with alternating β-1,4 and β-1,3 glycosidic bonds.

Hyaluronan is a very hydrophilic substance; it can trap approximately 1000 times its weight of water. Hyaluronan chains interact with each other at very low concentrations, which may contribute to its unusual rheological properties. For instance, 1 % solution of a high molecular weight hyaluronan is jelly-‐‑like, but when subjected to pressure, it moves easily, which makes it an ideal lubricant (Necas et al. 2008). Hyaluronan has a large repertoire of biological functions, which suggests the existence of a large number of different conformations and binding interactions (Cowman and Matsuoka 2005). Traditionally,

(26)

hyaluronan has been thought to take an expanded, random coil form stabilized by hydrogen bonds in water solutions (Scott 1989). However, later studies have shown that hyaluronan takes a more ordered conformation under specific conditions (de la Motte et al.

2003). In addition, atomic force microscopy of hyaluronan on a hydrated mica surface has suggested relaxed coil and partially condensed conformations in liquid connective tissues, fully condensed rods tethered to a cell surface or in the intracellular space, and fibrous forms associated with proteins (Cowman et al. 2005).

Hyaluronan differs from other glycosaminoglycans in many ways (Table 1). For example, it is synthesized on the plasma membrane instead of the Golgi apparatus, it does not contain sulfate groups, nor is it covalently bound to a core protein during synthesis, like other GAGs (Dicker et al. 2014; Toole 2000).

Table 1. Properties of glycosaminoglycans (Fraser et al. 1997)

Glycosaminoglycan Sugar units Sulphate Size (Mr) Proteoglycan

Hyaluronan Glucuronic acid + glucosamine - 10⁵-10⁷ -

Chondroitin 4-(6-) sulphates Glucuronic acid + galactosamine + 10-50x10³ + Dermatan sulphate Iduronic acid + galactosamine + 10-50x10³ +

Keratan sulphate Galactose + glucosamine + X5-15x10³ +

Heparan sulphate Glucuronic and iduronic acid + glucosamine + 10-50x10³ + Heparin Glucuronic and iduronic acid + glucosamine + 5-20x10³ +

2.2 SYNTHESIS OF HYALURONAN

2.2.1 The vertebrate hyaluronan synthase gene family

Hyaluronan is synthesized by hyaluronan synthases (HAS). The first Has gene was cloned in group A Streptococcus pyogenes (DeAngelis et al. 1993). In mammalians, three hyaluronan synthase genes have been isolated and identified: HAS1 by functional expression cloning (Itano and Kimata 1996a; Itano and Kimata 1996b; Shyjan et al. 1996), HAS2 (Spicer et al.

1996; Watanabe and Yamaguchi 1996) and HAS3 (Spicer et al. 1997a) by using a degenerate RT-‐‑PCR protocol. Further, DG42 protein, expressed in Xenopus gastrulation, was found to be a hyaluronan synthase (Meyer and Kreil 1996) and was later identified as Has1 (Spicer and McDonald 1998). Interestingly, hyaluronan synthase is also found in the PBCV-‐‑1 virus that infects chlorella-‐‑like green algae (DeAngelis et al. 1997), but not in invertebrates although HAS-‐‑transfected Drosophila cells can produce hyaluronan (Takeo et al. 2004).

The Has gene family is highly conserved. Mammalian Has1, Has2 and Has3 genes share 55-‐‑71% sequence identity and approximately 25 % amino acid identity with Streptococcus pyogenes HasA (Spicer and McDonald 1998). The three Has genes are located in different chromosomes in both human and mouse genomes, which suggests that the Has gene family was formed quite early in vertebrate evolution. In humans, HAS1 is located in chromosome 19q13.3-‐‑q13.4, HAS2 in chromosome 8q24.12 and HAS3 in chromosome 16q22.1, while in the mouse they reside in chromosomes 17, 15, and 8, respectively (Spicer et al. 1997b).

The promoter regions for human HAS1 (Chen et al. 2012), HAS2 (Monslow et al. 2004) and HAS3 (Wang et al. 2015) have now been analyzed. All HAS genes show constitutive basal activity, of which HAS2 has the lowest basal level, suggesting that HAS2 is the main candidate modulating hyaluronan synthesis rate (Monslow et al. 2003). Analysis of the human HAS1 promoter region has revealed that the upregulation of HAS1 transcription by TGF-‐‑β1 is mediated via the mammalian homolog of Drosophila Mad 3 (Smad3), a key

(27)

mediator of TGF-‐‑β1 signaling, and that induction by IL-‐‑1β is dependent on the transcription activator specificity protein 3 (Sp3) (Chen et al. 2012).

Upstream of the HAS2 transcription initiation site (TIS) are located the putative Sp1, NF-‐‑

Y/CCAAT and NF-‐‑κB sites, common to all Has2 orthologs, suggesting evolutionally conserved transcriptional regulation (Monslow et al. 2004). Further analysis showed Sp1 and Sp3 binding 63 bp upstream of HAS2 TIS. siRNA knockdown of either Sp1 or Sp3 caused a significant decrease in HAS2 transcription (Monslow et al. 2006).

Human HAS3 promoter, which is highly conserved in the mouse and rat Has3 promoters, is mainly restricted to a 450 bp region upstream of the major TIS. It lacks canonical TATA box, but contains classical CG box and putative transcription factor binding sites for C/EBP and NFκB. Mutagenesis analysis showed that Sp1 and the core promoter motif ten elements (MTE) are required for HAS3 gene activity. Two TISs (Wang et al. 2015), distinct from earlier studies (Liu et al. 2001), were also observed, indicating complicated regulation of HAS3 expression.

All Has genes share at least one exon-‐‑intron boundary, suggesting that the genes have evolved from a common ancestral gene. In addition to Has1-‐‑3, non-‐‑functional Has-‐‑related sequences (Has-‐‑rs) have been identified in Xenopus. There are more similarity between Has2 and Has3 and on the other hand between mouse Has1 and Xenopus DG42, which seem to be orthologs (Spicer and McDonald 1998). Human and mouse Has1 gene comprises of 5 exons and Has2-‐‑3 of 4 exons (Monslow et al. 2003). HAS1 has been observed to have multiple aberrant splice variants in multiple myeloma (Adamia et al. 2005a; Kriangkum et al. 2013), Waldenström’s macroglobulinemia (Adamia et al. 2005b) and bladder cancer (Golshani et al. 2007). Inherited polymorphisms in HAS1 gene have been observed to influence aberrant splicing and predict the risk of B-‐‑cell malignancies (Kuppusamy et al.

2014). HAS3 has been observed to have two variants of mRNA for coding proteins (Sayo et al. 2002).

Exon-‐‑intron comparison suggests that the Has genes are divided into two classes;

Xenopus DG42/Has1, Has-‐‑rs and mouse Has1 comprises the first class and mouse Has2 and Has3 comprises the second class. Gene families arise usually by sequential gene duplication and divergence. Based on the genomic structure, comparisons between mammalian and xenopus Has genes indicate that there have been three sequential gene duplications: first to form Has1 and Has2 subfamilies, then to generate Has1 and Has-‐‑rs and Has2 and Has3, respectively (Spicer and McDonald 1998).

2.2.2 The structure and function of hyaluronan synthases

Mammalian and bacterial hyaluronan synthases are plasma membrane proteins (Markovitz and Dorfman 1962) with putative transmembrane and membrane associated domains (Weigel et al. 1997). HASs are divided into two or three categories based on their structure and function. Mammalian and Gram-‐‑positive HASs belong to Class I-‐‑R and Xenopus HAS to Class I-‐‑N and the HAS from the Gram-‐‑negative Pastorella multocida belongs to Class II.

Class I enzymes have one glycosyltransferase (GT2) module, and I-‐‑N enzymes add new sugar to the nonreducing end and I-‐‑R to the reducing end. Class II HAS has two GT2 modules, is membrane peripheral or soluble, and adds new sugars to the nonreducing end (Weigel and DeAngelis 2007). HASs have many unusual characteristics compared with other glycosyltransferases; they do not need a primer to start the synthesis (Markovitz et al.

1959; Stoolmiller and Dorfman 1969): 1) hyaluronan synthesis takes place in the plasma membrane, in contrast to the Golgi apparatus for other glycosyltransferases (Markovitz and Dorfman 1962; Prehm 1984), 2) the initiation of the hyaluronan polymer does not need a core protein attachment (Prehm 1983), and, in contrast to the “one enzyme -‐‑ one sugar linkage” dogma in glycobiology, 3) HASs catalyze two different linkages. Additionally, this relatively small enzyme (417-‐‑972 amino acid residues) has other functions as well (Table 2).

(28)

Table 2. Synthesis of hyaluronan requires multiple steps, that have been established within hyaluronan synthase Class I, modified from (Weigel et al. 2015).

Established activities within hyaluronan synthase Class I

1. GlcUA(α1→)UDP acceptor binding site 2. GlcNAc(α1→)UDP acceptor binding site

3. Hyaluronyl-GlcNAc(α1→)UDP donor binding site [for HA-GlcUA(β1,3)GlcNAc(α1→)UDP]

4. Hyaluronyl-GlcUA(α1→)UDP donor binding site [for HA-GlcNAc(β1,4)GlcUA(α1→)UDP]

5. Hyaluronyl-GlcNAc(α1→)UDP: GlcUA(α1→)UDP, hyaluronyl-GlcNAc(β1,4) transferase activity 6. Hyaluronyl-GlcUA(α1→)UDP: GlcNAc(α1→)UDP, hyaluronyl-GlcUA(β1,3) transferase activity

Because crystallographic data is yet not available, little is known about the exact structure of HASs, and thus, existing information is based on in silico analysis (Vigetti et al. 2014).

Mammalian HASs are predicted to contain 4-‐‑6 transmembrane domains and 1-‐‑2 membrane associated domains (Figure 2) (Heldermon et al. 2001a; Weigel and DeAngelis 2007). HASs are lipid-‐‑dependent (Tlapak-‐‑Simmons et al. 1998; Weigel and DeAngelis 2007) and in addition to the precursors UDP-‐‑glucuronic acid and UDP-‐‑N-‐‑acetylglucosamine, they need Mg⁺⁺, Mn⁺⁺ or Co⁺⁺ to be functional (Markovitz et al. 1959). Cysteine has catalytic, structural and functional roles in many enzymes (Carugo et al. 2003; Jose-‐‑Estanyol et al. 2004; Saito 1989). In fact, Streptococcus equisimilis HAS (seHAS), which is the smallest class I HAS, contains four cysteine residues (C²²⁶, C²⁶², C²⁸¹ and C³⁶⁷) that are conserved in the mammalian HAS family. Although Cysteine null mutants of seHAS are active (Heldermon et al. 2001b; Kumari et al. 2002), modifying individually or with specific combinations of Cys residues influence the rate and hyaluronan size independently (Weigel and Baggenstoss 2012). Although they are not directly involved in the catalysis (Kumari et al.

2002), Cys²⁶², Cys²⁸¹, and Cys³⁶⁷ cluster at the membrane-‐‑enzyme interface close to the substrate binding site (Kumari and Weigel 2005) and are functionally important.

Figure 2. Proposed structure and plasma membrane topology of Bacterial and eukaryotic HAS. Modified from (Kumari and Weigel 2005; Weigel 2002).

(29)

Transfection studies with N-‐‑terminally GFP-‐‑tagged mammalian HAS1-‐‑3 have shown that HASs travel through endoplasmic reticulum (ER) and via Golgi apparatus to their location of action, the plasma membrane. However, HASs do not have the typical N-‐‑terminal signal sequence, which would suggest that they are not processed via the usual secretory pathway (Rilla et al. 2005). Residence at the plasma membrane is imperative for all three HASs to activate hyaluronan synthesis. Although hyaluronan synthesis is thought to start at the plasma membrane (Rilla et al. 2005), in vitro incubations of membrane fractions show that HAS activity also exists in intracellular compartments (Vigetti et al. 2009a). O-‐‑GlcNAc modification of at least HAS2 and HAS3 enhance plasma membrane targeting and residence (Deen et al. 2016; Tammi et al. 2011; Vigetti et al. 2012).

Figure 3. Synthesis of the precursors of hyaluronan (modified from (Tammi et al. 2011)).

Hyaluronan is synthesized from UDP-‐‑N-‐‑acetylglucosamine and UDP-‐‑glucuronic acid. The precursors are formed from glucose via several steps (Figure 3). Although primers are not needed in hyaluronan synthesis (Prehm 1983; Stoolmiller and Dorfman 1969) there is a kinetic lag in the beginning of hyaluronan synthesis indicating that HAS assembly initiation is a rate limiting step (Baggenstoss and Weigel 2006). Class I HASs are able to produce chitin oligomers which have been suggested to act as self-‐‑primers initiating hyaluronan synthesis (Weigel, West et al. 2015). After initiation, class I-‐‑R enzymes add the alternating monosacharides to the reducing end of growing hyaluronan (see equation 1) (Asplund et al. 1998; Prehm 1983; Prehm 2006). Class I-‐‑N and Class II hyaluronan synthases use a non-‐‑reducing end of the growing hyaluronan (equation 2) (Bodevin-‐‑Authelet et al.

2005).

Equation 1. UDP-GLCNAc-HA + UDP-GlcUa+UDP-GlcNAc → 2xUDP + UDP-GlcNAc(ββ1,4)GlcUA(ββ1,3)GlcNAc-HA Equation 2. UDP-GlcUA + UDP-GlcNAc + HA → 2xUDP+ GlcUA(β1,3)GlcNAc(β1,4)-HA

Hyaluronan is mainly an extracellular molecule and its polymerization occurs on the inner surface of the plasma membrane (Prehm 1984). Therefore, hyaluronan must somehow be translocated outside of the cell. Two types of translocation mechanisms have been suggested; 1) ABC transporters and 2) HASs intrinsic translocase activity. The ABC transporter gene is located upstream of the Streptococcus pyogenes Has gene. Capsular

(30)

hyaluronan was reduced when the ABC gene was mutated (Ouskova et al. 2004). Further, inhibitors of multidrug resistance transporters decreased hyaluronan production (Prehm and Schumacher 2004). It is however possible that the reduced hyaluronan synthesis observed with ABC transporter inhibition is caused by an alteration of UDP sugar metabolism (Medina et al. 2012). In contrast, the ABC transporter was not shown to contribute to hyaluronan translocation in human breast cancer cells in vitro (Thomas and Brown 2010). Glycosyltransferase and translocation activities are coupled spatially (Hubbard et al. 2012). HAS2 activity is coupled to dimerization (Karousou et al. 2010) and all three HASs can form hetero-‐‑ or homomers (Bart et al. 2015). Therefore, it is possible that a transmembrane pore is formed in the interface of two or more HAS proteins (Bart et al.

2015) and the extrusion is energized by glycosyltransferase activity (Hubbard et al. 2012).

However, a conventional view suggests that a single HAS monomer forms an intraprotein pore (Weigel and Baggenstoss 2012). The intraprotein pore hypothesis is based on following aspects: 1) HAS protein contain several membrane domains (Heldermon et al.

2001a; Weigel and DeAngelis 2007) and 2) Streptococcal HAS is active as a monomere (Tlapak-‐‑Simmons et al. 1998)..

What controls the size of hyaluronan? Weigel and Baggenstoss showed that the size of hyaluronan is not coupled to, or dependent on, the specific polymerizing activity of hyaluronan synthase. They proposed a retain-‐‑release model, which explains why HAS makes a large polymer and does not release it into the extracellular space until a certain size is attained. Multiple hyaluronan-‐‑UDP and HAS interactions within the intraprotein pore could provide a relatively constant retention force. As the hyaluronan polymer increases in size, multiple interactions such as Brownian motion, fluid currents and matrix proteins in the extracellular environment can act to pull hyaluronan off the enzyme (Weigel and Baggenstoss 2012). Because the estimated lifetime of a single HAS protein is relatively short (4-‐‑6 hours), it has been suggested that one HAS protein produces only one hyaluronan molecule (Kitchen and Cysyk 1995). The estimated hyaluronan synthesis rate is 3 monosaccharides per second (Pummill and DeAngelis 2003) implying that the time for the synthesis of one 6x10⁶ Da hyaluronan chain is approximately 3 hours in rat epidermal keratinocytes (Karvinen et al. 2003).

The structural differences between the three mammalian Has genes are correlated with their properties. Has2 and Has3 genes share more similarities between each other than either one with Has1 (Spicer and McDonald 1998). Transfection studies with mouse Has1-‐‑3 showed that the pericellular coat produced by Has1 in COS-‐‑1 cells or rat fibroblasts was significantly smaller than those produced by Has2 and Has3 (Itano et al. 1999). Further, hyaluronan produced by HAS2 and HAS3 is concentrated on membrane protrusions (Kultti et al. 2006; Rilla et al. 2005), whereas the hyaluronan coat of HAS1 is more diffuse and not located on cell protrusions (Rilla et al. 2013). The K^m values for both precursors were significantly higher for Has1 than for Has2 and Has3 (Rilla et al. 2013). On the other hand, hyaluronan produced by Has1 and Has2 is larger than that produced by Has3 and the synthesis rate of Has3 is slower (Itano et al. 1999; Spicer and McDonald 1998; Wilkinson et al. 2006). In contrast, Brink and co-‐‑workers found that Has1 and Has3 transfected CHO cells produced smaller hyaluronan than Has2 transfected cells and that Has3 synthesis rate was the highest (Brinck and Heldin 1999). These inconsistencies suggest that the size and synthesis rate of hyaluronan are not solely dependent on a specific HAS isoenzyme.

Why are three distinct enzymes involved in the production of a fairly simple polymer?

Functional differences suggest that the HAS isoenzymes may be involved in different physiological states. Has2 knockout mice die as a result of cardiovascular defects (Camenisch et al. 2000) but Has1 (Kobayashi et al. 2010), Has3 (Bai et al. 2005) and Has1/Has3 double knockout mice (Mack et al. 2012) are viable and fertile. HAS isoenzymes are differentially expressed in developing mice embryos. HAS2 showed the most

Hyaluronan-positive cell protrusions and microvesicles : studies on their structure, function and existence in vivo

VILLE KOISTINEN

HYALURONAN-POSITIVE CELL PROTRUSIONS AND MICROVESICLES

Dissertations in Health Sciences

Hyaluronan-­‐‑positive cell protrusions and microvesicles

Hyaluronan-­‐‑positive cell protrusions and microvesicles

Studies on their structure, function and existence in vivo

Elämäni valolle

Acknowledgements

List of the original publications

Contents

Abbreviations

1 Introduction

2 Review of the literature

Hyaluronan-‐‑positive cell protrusions and microvesicles

Hyaluronan-‐‑positive cell protrusions and microvesicles