Diana Törmä
FIBROSIS AND CARDIAC TISSUE SPECIFIC MARKERS OF TISSUE INJURY CAUSED BY HYPOXIA IN A HUMAN VASCULARIZED CARDIAC TISSUE MODEL
Fibroosi ja hypoksian aiheuttaman sydänvaurion sydänspesifiset markkerit tutkittuna ihmisperäisellä verisuonitetulla sydänmallilla
Faculty of Medicine and Health Technology Master’s Thesis January 2020
i
“ .. for it is good for the heart to be strengthened by grace, not by foods...”
Hebrews 13:9
ii
Master’s Thesis
Place: TAMPERE UNIVERSITY, Degree Programme in Biomedical Technology, specialization in Cell Technology
Author: Diana Törmä
Title: FIBROSIS AND CARDIAC TISSUE SPECIFIC MARKERS OF TISSUE INJURY CAUSED BY HYPOXIA IN A HUMAN VASCULARIZED CARDIAC TISSUE MODEL
Pages: 55 pages, 1 appendix page
Supervisors: Professor Tuula Heinonen and Tuomas Tolvanen Reviewers: Professor Heli Skottman, Professor Tuula Heinonen Date: January 2020
Abstract
Background and aims of the study: According to WHO, cardiovascular diseases, including ischemic heart diseases, are still the leading cause of death worldwide. Cardiac fibrosis occurs in various cardiovascular diseases and heart conditions as the response of the heart tissue to the injury, such as in ischemia. Injury causes extracellular matrix components to readjust in order to reconstruct the damaged area. The aim of this study was to determine if formation of fibrosis is triggered in the vascularized cardiac tissue model under hypoxic conditions of different durations.
Materials and methods: Human vascularized cardiac tissue model was constructed according to GLP quality principles and FICAM guidelines using human adipogenic stromal cells, human umbilical vein endothelial cells and human induced pluripotent stem cell derived cardiomyocytes. Oxygen deprivation was simulated using Minihypoxy-portable system. Oxygen levels were measured with fluorimetric oxygen sensor. Fibrotic features were studies with fluorescent and confocal microscopy.
Immunocytohemistry assaywas performed for troponin T, collagens I and III, and fibronectin. ELISA assay was performed to measure troponin T and lactate dehydrogenase.
Results: Experiment showed formation of two distinctive fibrotic features in the culture undergone 6- hour hypoxia. Changes in collagen I and III distribution were noticed in post-hypoxia cultures. Cell death, seen in increase in troponin T and lactate dehydrogenase values after reperfusion, suggests correlation of the event with ischemia-reperfusion injury occurring in a post infarction heart. This study presented preliminary results, bringing to light areas that require further development.
Conclusions: This experiment showed a novel in vitro approach for studying fibrosis of the cardiac tissue using human vascularized cardiac tissue model and portable Minihypoxy-system. Although this model showed similar fibrotic features with the ones occurring in the human heart upon ischemia- reperfusion injury, it still needs optimization in order to achieve repeatability and reproducibility for the conduction of the further studies. Thus this model is a potential new tool for development of therapies to battle cardiac fibrosis.
Keywords: cardiac tissue model, cardiac fibrosis, hypoxia, ischemia-reperfusion
The originality of this thesis has been checked using the Turnitin OriginalityCheck service.
iii
Pro-Gradu tutkielma
Paikka: TAMPEREEN YLIOPISTO, Biolääketieteen teknologian koulutusohjelma, soluteknologian suuntautumisvaihtoehto
Tekijä: Diana Törmä
Otsikko: Fibroosi ja hypoksian aiheuttaman sydänvaurion sydänspesifiset markkerit tutkittuna ihmisperäisellä verisuonitetulla sydänmallilla
Sivumäärä: 55 sivua, 1 liitesivu
Ohjaajat: Professori Tuula Heinonen ja Tuomas Tolvanen Tarkastajat: Professori Heli Skottman, professori Tuula Heinonen Aika: Tammikuu 2020
Tiivistelmä
Tutkimuksen tausta ja tavoitteet: Kansainvälisen terveysjärjestön WHO:n mukaan sydän- ja verisuonitaudit, mukaan lukien iskeemiset sydänsairaudet, ovat johtava kuolinsyy maailmanlaajuisesti.
Sydämen fibroosia esiintyy erilaisissa sydän- ja verisuonitaudeissa, kudoksen reagoidessa esimerkiksi iskemian aiheuttamaan vaurioon. Vaurio aiheuttaa muutoksia soluväliaineiden ainesosissa, jotka järjestäytyvät uudelleen vaurioituneen kohdan uudelleenmuotoilemiseksi. Tämän tutkimuksen tarkoituksena oli selvittää, muodostuuko verisuonitetulle sydänmallille fibroottisia piirteitä eripituisissa hapenpuutostiloissa.
Tutkimusmenetelmät: Ihmisperäinen verisuonitettu sydänmalli rakennettiin GLP-laatujärjestelmän ja FICAM:n ohjeiden mukaan ihmisperäisistä rasvan kantasoluista, ihmisperäisistä napanuoran laskimon endoteelisoluista ja ihmisperäisistä indusoiduista monikykyisistä kantasoluista erilaistuneista kardiomyosyyteistä. Hapenpuute aiheutettiin kannettavalla Minihypoxy-laitteella, jonka happitasoa mitattiin fluorimetrisellä happisensorilla. Fibroottisten piirteiden tutkimiseen käytettiin fluoresenssi- ja konfokaalimikroskopiaa. Kollageenien I ja III, troponiinin ja fibronektiinin esiintymistä tutkittiin immunosytokemiallisella analyysilla, lisäksi määritettiin troponiini T ja laktaattidehydrogenaasi ELISA:lla.
Tutkimustulokset: Kuusi tuntia vähähappisissa oloissa pidetyssä verisuonitetussa sydänmallissa havaittiin kaksi selkeää fibroosipiirrettä. Hapenpuutteessa olevissa viljelmissä esiintyi muutoksia kollageenien I ja III järjestäytymisessä. Sydänmallissa aikaansaatu solukuolema, joka ilmeni laktaattidehydrogenaasin ja troponiini T:n arvojen nousuna reperfuusion jälkeen, muistutti iskemia- reperfuusiovaurioita postinfarktisessa sydämessä. Tämä tutkimus esittää ainoastaan alustavat tulokset, paljastaen jatkokehitystä vaativat alueet.
Johtopäätökset: Tämä tutkimus esittää uudenlaisen in vitro lähestymistavan sydänperäisen fibroosin tutkimiselle, jossa käytetään ihmisperäistä verisuonitettua sydänmallia ja kannettavaa Minihypoxy- laitetta. Tutkimusmallilla aikaansaatiin samankaltaisia piirteitä, joita esiintyy iskemia- reperfuusiovaurion yhteydessä. Toistettavuuden ja luotettavuuden parantamiseksi malli tarvitsee kuitenkin vielä optimointia. Tämä malli tarjoaa uudenlaisen välineen sydänfibroosin hoitoon tarkoitettujen terapioiden kehittämiselle.
iv ACKNOWLEDGEMENTS
First, I would like to thank my supervisors, Tuula Heinonen for giving me this amazing opportunity to conduct this study and learn many new things, and Tuomas Tolvanen for guiding me through the whole process. I also want to thank all personnel of FICAM for many joyful moments and for providing help and assistance in technical issues. I thank Joose Kreutzer and professor Pasi Kallio’s group for providing the hypoxia equipment and technical support. I also want to say special thanks to my friend Sami Leppälä whose exemplary diligence and work ethics I admire. I’ve learned so much during these 6 months. Culturing cells, building this model and seeing “heart beat” on a dish was a wonderful experience.
I want to express enormous gratitude to my beloved family. I thank my dear husband, my mother, grandmother and sister for always supporting me in my studies and in achieving my goals. Their wisdom, encouragement and endless support made this incredible journey possible.
Tampere 11.1.2020 Diana Törmä
v TABLE OF CONTENTS
1 INTRODUCTION ... 1
2 LITERATURE REVIEW ... 3
2.1 Human cardiac muscle tissue ... 3
2.1.1 Cardiomyocytes ... 4
2.1.2 Fibroblasts ... 4
2.1.3 Endothelial cells and pericytes ... 5
2.1.4 Extracellular matrix of the cardiac connective tissue ... 5
2.1.5 Cardiac cell death ... 6
2.1.6 Cardiomyocyte metabolism ... 7
2.2 Hypoxia, ischemia and ischemia-reperfusion injury ... 7
2.2.1 Role of oxygen in cardiac function ... 7
2.2.2 Ischemia-reperfusion injury ... 8
2.2.3 Molecular pathways and effects of hypoxia ... 9
2.3 Cardiac fibrosis ... 9
2.3.1 Characteristic features of cardiac fibrosis ... 11
2.4 Cellular and molecular regulation of cardiac fibrosis ... 12
2.4.1 Cellular regulation of cardiac fibrosis ... 13
2.4.2 Molecular regulation of cardiac fibrosis... 14
2.5 Biomarkers of cardiac tissue injury ... 15
2.5.1 Troponin T ... 16
2.5.2 Lactate dehydrogenase ... 16
2.5.3 Collagens ... 16
2.5.4 α-smooth muscle actinin ... 17
2.5.5 Fibronectin and periostin ... 17
2.6 Studying cardiac tissue in vitro and in vivo ... 17
3. OBJECTIVES ... 21
4. MATERIALS AND METHODS ... 22
4.1 Setting up the vascularized cardiac tissue model ... 22
4.2 Inducing hypoxia ... 22
4.2.1 Post-hypoxia culture ... 23
4.3 Troponin T assay ... 23
4.4 Lactate dehydrogenase assay ... 24
4.5 Immunocytochemistry ... 24
4.6 Oxygen level measurement ... 25
4.7 Statistical analysis ... 25
vi
5. RESULTS ... 27
5.1 Experiment 1 ... 27
5.2 Experiment 2 ... 31
5.3 Oxygen level measurement ... 36
6. DISCUSSION ... 37
6.1 Study setting ... 37
6.2 Lactate dehydrogenase and troponin T ... 39
6.3 Immunocytochemistry and components of the extracellular matrix... 40
6.4 HVCTM in studying cardiac fibrosis ... 43
7. CONCLUSIONS ... 45
REFERENCES ... 47
APPENDICES ... 55
vii ABBREVIATIONS
AA ascorbic acid
Ang II angiotensin II
ANGm angiogenesis medium
ACE angiotensin converting enzyme
BSA bovine serum albumin
CM cardiomyocytes
CMmm cardiomyocyte maintenance medium
DPBS Dulbecco’s phosphate buffered saline
DMEM/F12 Dulbecco’s modified Eagle’s medium nutrient mixture F12 EBM-2 endothelial cell basal medium-2
EC endothelial cell
ECM extracellular matrix
EGM-2 endothelial cell growth medium-2
ELISA enzyme-linked immunosorbent assay
FB fibroblast
FGF fibroblast growth factor
GLP good laboratory practice
GLUT-1 glucose transporter -1 receptor
hADSC human adipose-derived stem cell
hASC human adipogenic stromal cell
hASCbm hASC basic medium
hASCgm hASC growth medium
HE heparin
hESC human embryonic stem cell
HIF-1 hypoxia inducible factor 1
hiPSC-CM human induced pluripotent stem cell derived cardiomyocytes HUVEC human umbilical vein endothelial cell
HUVECgm HUVEC growth medium
HVCTM human vascularized cardiac tissue model
HY hydrocortison
IL interleukin
LDH lactate dehydrogenase
MSC mesenchymal stem cells
MI myocardial infarction
MMP matrix metalloproteinase
MyoFB myofibroblast
PBS phosphate buffered saline
PDGF platelet-derived growth factor
PDMS polydimethylsiloxane
ROS reactive oxygen species
RT room temperature
SMA smooth muscle actinin
T3 3,3,,5-triiodo-L-thyronine sodium salt
TGF transforming growth factor
TIMP tissue inhibitors of metalloproteinase
TNF tumor necrosis factor
viii
TnC troponin C
TnI troponin I
TnT troponin T
VEGF vascular endothelial growth factor
1 1
1 INTRODUCTION
Heart, a vital organ, is a muscular pump that allows blood to flow through the circulatory system. Cardiac fibrosis develops in various cardiovascular diseases and pathophysiologic heart conditions, such as myocardial infarction (MI), ischemia and heart failure (Li et. al.
2018; Murtha et. al. 2017). According to World Health Organization, cardiovascular diseases are still the leading cause of death (WHO 2018), with ischemic heart diseases and stroke being the biggest killers worldwide, taking more than 9 million lives in 2016 (https://www.who.int/news-room/fact-sheets/detail/the-top-10-causes-of-death; 12.06.2019).
Cardiac fibrosis occurs as the response of the heart tissue to the injury, in which components of the extracellular matrix and their distribution become readjusted in order to reconstruct the damaged area, which is part of the process of wound healing and tissue remodeling.
Characteristic feature of fibrosis is pathological excess deposition of extracellular matrix (ECM) caused by imbalance between degradation and production of the ECM proteins, particularly the build-up of collagen (Li et al. 2018; Murtha et al. 2017). Because cells in cardiac tissue constantly interact with each other and surrounding milieu (Borg & Baudino 2011; Travers et al. 2016), knowledge on the roles and interactions of different cell types present in cardiac tissue, as well as their involvement in ECM remodeling, remains a target of many studies.
Constant oxygen supply is critical for normal function of the heart (Giordano 2005). Hypoxia, an insufficient oxygen supply to the cells, is closely associated with injury, tissue repair and fibrosis formation (Darby & Hewitson 2016). Decreased oxygen levels in cardiac tissue, seen in myocardial infarction (MI) and ischemic conditions, have been reported to alter gene expression profile of the cell, affecting cell signaling, molecular pathways and thus changing cell behavior (Darby & Hewitson 2016). Hypoxia of the cardiac tissue has been related to enhanced expression of pro-fibrotic genes, increased amount of collagen and increase in proliferation of cardiac fibroblasts (Watson et al. 2014).
Heart tissue’s limited capacity for repair and regeneration indicates that fibrosis, leaving a permanent scar, can cause severe cardiac dysfunction (Kim et al. 2018). This emphasizes importance of the studies on cardiac fibrosis. Thus, cardiac fibrosis presents a target for
2
development of new pharmaceuticals and therapies. Molecular mechanisms, signaling pathways and cell-cell interactions in cardiac fibrosis are nowadays widely studied. Numerous studies and reviews on fibrosis have been published in the last decade, reflecting the global concern and challenge which understanding of cardiac fibrosis presents to the scientific circles worldwide.
Studying molecular specifics of hypoxia and fibrosis in vivo in humans is challenging. There have been many recent studies conducted on genetically manipulated cells and fibrotic cardiac tissue, both human and animal derived. Animal models, mostly rats and mice, are widely used in studying various cardiovascular diseases, including cardiac fibrosis (Camacho et al. 2016;
Cowling et al. 2019). Cardiac (micro)tissue (Archer et al. 2018; Lee et al. 2019; Mannhardt et al. 2016), (vascularized) spheroids (Polonchuk et al. 2017; Yan et al. 2019), cardiac (vascularized) organoids (Richards et al. 2017; Voges et al. 2017), and other types of biomimetic cardiac (micro)tissue, with combinations of two to three cell types normally present in cardiac tissue, have been documented, offering the possibility to simulate in vivo conditions more closely. However, experiments with human embryonic stem cells (hESC) and animal models raise ethical concerns. Use of human hearts is invasive method that requires time and suitable donors, thus it can’t be conducted routinely. This emphasizes the need of more complex cardiac microtissue that can mimic desired in vivo conditions as closely as possible.
This thesis was conducted for Finnish Center of Alternative Methods FICAM. FICAM uses its expertise to establish new solutions for replacement of animal testing by carrying out tests on organ and tissue models that are derived from human cells (https://research.uta.fi/ficam;
17.05.2019). This study followed the change in cardiac fibrosis markers of fibronectin, collagen I and III distribution, amounts of troponin T (TnT) and lactate dehydrogenase (LDH) in the somatic human cell derived vascularized cardiac tissue model under hypoxic conditions, which were simulated with Minihypoxy system (Metsälä et al. 2018), caused by decrease in system oxygen level. It also discussed the correlation of the events with ones occurring in the human heart tissue, with the focus on the formation of fibrotic markers.
3 2 LITERATURE REVIEW
Expanding knowledge on human heart in health and disease is critical. Studies and researches provide valuable information crucial for understanding not only mechanisms and patterns of the disease progression and development, but also for proposing ideas on drug and therapy development to treat them.
There have been numerous excellent reviews and studies on such subjects as cardiac fibrosis, pattern of its development, its relation to various cells present in the cardiac tissue, known and suggested molecular pathways, cellular and molecular interactions, and their influence on disease features development, written recently (Kong et al. 2014; Li et. al. 2018; Ma et al.
2018; Murtha et al. 2017). This emphasizes that cardiac fibrosis is not only of a great importance, but also bears complexity, entirety of which is challenging to comprehend.
Amount of the related literature and publications depicts rapid development that has occurred in this field in the recent decades, in addition to continuing enthusiasm of professionals.
2.1 Human cardiac muscle tissue
Cardiac muscle, myocardium, is a highly organized specific type of muscle tissue unique to the heart that, due to its rhythmical contractile ability, allows blood to flow. Architecture of the myocardium is very complex: parallel cardiac muscle cells form sheets within the same layer, but their direction changes from layer to layer. Myocardium is the middle layer of the cardiac wall. On the outer side of the cardiac wall is epicardium, a layer of loose connective tissue and mesothelium. On the inner side is endocardium, a layer of loose connective tissue end endothelium (Brown et al. 2005b; LeGrice et al. 2005, 3-16; Parkkila 2016).
Cardiac tissue consists of various cell types, such as cardiomyocytes (CM), endothelial cells (EC), pericytes, smooth muscle cells, fibroblasts (FB) and cells of bone-marrow origin (Borg
& Baudino 2011; Jugdutt 2005, 23-24). Muscle tissue of the heart includes blood vessels, nerve endings and connective tissue that binds all together (LeGrice et al. 2005, 3-15).
4
Cellular and molecular elements of the cardiac tissue are in the constant interaction with each other. These interactions are dynamic and highly organized, with any molecular alteration affecting others, some able to lead to abnormal heart function (Borg & Baudino 2011).
Electrophysiological property of the heart is based on the electrical currents, caused by ion movement (K+, Na+, Ca2+ and others) through ion channels on CM membranes, and spontaneous depolarization of the CMs (Mäkynen & Mäkijärvi 2016). Electrophysiological properties of the heart and electrical disturbances caused by hypoxia and fibrosis will not be discussed in this study.
2.1.1 Cardiomyocytes
Cardiac muscle cells, or cardiomyocytes, are large cylindrical cells 50-150 µm in length and 10-25 µm in diameter that are branched and connected to each other end-to-end via special type of junction. CMs contain myofibrils – long rod-like structures comprised of actin, myosin, titin and other linker proteins that together form sarcomers, a contractile unit of the CMs that give the cell striated appearance (Parkkila 2016). CMs were found to comprise most of the myocardial mass and volume (Jugdutt 2005, 24), although, quantitatively they are not the most abundant cell type in the tissue (Pinto et al. 2016), with only 25-35% reported, as reviewed by Jugdutt (2005, 24).
2.1.2 Fibroblasts
Cardiac FBs have been earlier described to abundantly populate cardiac tissue, comprising most of (non-CM) cells (Baum & Duffy 2011; Brown et al. 2005b; Jugdutt 2005, 23-24;
Weber 1989). However, recent study on mouse heart by Pinto et al. (2016) suggested fibroblast population to be much smaller than described earlier, being below 20% of the non- CM cells. FBs are responsible for production of ECM proteins (Jugdutt 2005, 23-25;
Krenning et al. 2010), maintenance of the ECM in normal hearts, transmitting inflammatory response and remodeling of the injured tissue in various pathological conditions (Baum &
Duffy 2011; Brown et al. 2005b). Cardiac FBs also play important role in maintaining electrophysiological activity of the heart (Krenning et al. 2010).
5 2.1.3 Endothelial cells and pericytes
Blood vessels of all sizes are comprised of endothelial cells, with pericytes, a supportive cell type, lining smaller blood vessels – venules and capillaries (Sainio & Sariola 2015).
According to study conducted by Pinto and colleagues (2016), ECs appear to be most abundant non- myocyte cell type in the mouse cardiac tissue, contrary to earlier studies, with amount of over 60% reported.
2.1.4 Extracellular matrix of the cardiac connective tissue
ECM of the cardiac tissue provides a scaffold, balanced distribution of which is crucial for proper alignment of cells and cell layers that support structure of the tissue. ECM supports mechanical and electrophysiological activity of the heart (Weber 1989) and participates in maintenance of biochemical homeostasis (Krenning et al. 2010).
Cardiac connective tissue is organized in layers. Endomysium is a layer of cardiac connective tissue surrounding single cardiomyocytes and attaching adjacent muscle fibers to each other.
Groups of cardiomyocytes, myofibers, are covered by a sheet of connective tissue called perimysium. Cardiac connective tissue also connects cardiac capillaries to the cardiomyocytes (LeGrice et al. 2005, 3-7; Parkkila 2016; Weber 1989).
Cardiac ECM consists of a ground substance, glucosaminoglycans, proteoglycans, glycoproteins, a complex network of highly organized fibrous proteins (mainly collagens), and other structural proteins. It also includes various cells of non-cardiomyocyte origin, such as cardiac FBs, cardiac and vascular ECs, pericytes, macrophages, smooth muscle cells and others (Borg & Baudino 2011; Jugdutt 2005, 23-24). ECM also contains different growth factors, chemokines, cytokines, proteases and other molecules secreted by cells within the cardiac muscle tissue (Borg & Baudino 2011; Krenning et al. 2010).
Although cardiac FBs produce of most of the components of the ECM (Brown et al. 2005b;
Jugdutt 2005, 25; Weber 1989), it is suggested that other cells permanently or temporarily present in the cardiac tissue might also produce components of the matrix and thus affect ECM (Borg & Baudino 2011).
6
Five different types of collagen, types I, III, IV, V and VI, were described to be found in the ECM of the myocardium (Brown et. al. 2005b; Krenning et al. 2010; LeGrice et al. 2005, 7), with types I and III being the most abundant types (Brown et al. 2005a, 276). Collagens type IV and type V are found to be most abundant in basement membrane of the CMs (Brown et al. 2005b; LeGrice et al. 2005, 7-8).
ECM maintenance and turnover occurs by activity of matrix metalloproteinases (MMPs) that cleave matrix proteins, and tissue inhibitors of metalloproteinase (TIMPs) (Brown et al.
2005b; Jugdutt 2005, 27-29). There are several types of MMPs and TIMPs found in cardiac tissue, with prevalent MMPs types -2, -9 and -13, and TIMPs type -1 (at low levels), -2, -3, and -4 present in healthy cardiac tissue (Fan et al. 2012).
2.1.5 Cardiac cell death
Cardiac death is described to occur in vivo by three mechanisms, which are apoptosis, autophagy and oncosis (also necrosis) (Chiong et. al. 2011), each triggered by specific mechanisms and caused by various conditions and changes in tissue environment, all seen in cardiovascular diseases. Ability of the myocardium to regenerate itself is very limited, being a controversial topic and a subject of many studies (Buja & Vela 2008) for years. Most of the CM are permanently lost upon cardiac injury in humans, although recent studies suggest cardiogenesis occurring also in adult organisms (Cesselli et al. 2018; Uygur & Lee 2016).
Apoptotic (programmed) cell death, activated by external factors such as death ligands secreted by other cells, or intrinsically through mitochondrial injury, can be caused by hypoxia, oxidative stress, ischemia-reperfusion (Aalto-Setälä 2016, 27; Buja & Vela 2008;
Chiong et. al. 2011). Autophagy, which includes internal disintegration of the cell and its components, is another type of programmed cell death. Studies show connection between oxygen deprivation, ischemia, MI and autophagy in CMs, as reviewed by Chiong et. al.
(2011). Programmed cell death is also a part of organ development and aging processes, and it does not include inflammation (Aalto-Setälä 2016, 27; Buja & Vela 2008).
Oncotic cell death (necrosis) occurs in various pathological conditions, such as MI and heart failure (Aalto-Setälä 2016, 27; Chiong et. al. 2011), and is described as cell death caused by
7
injury, such as hypoxia and ischemia, toxic chemicals, or inflammation, in which cell contents are released into surrounding tissue. Such cell death is unprogrammed and is related to inflammation (Aalto-Setälä 2016, 27; Buja & Vela 2008).
Interestingly, cardiac muscle cells can undergo multiple types of cell death concurrently in the same heart as a response to injury in various pathological heart conditions and diseases, such as MI, ischemia and heart failure (Chiong et. al. 2011). Older studies have demonstrated simultaneous presence of different types of cell death, in the diseased cardiac tissue (Kostin et. al. 2003), and a coverslip in vitro model (Pitts & Toombs 2004). However, prevalence of each type varies and appears to be dependent on the specifics and duration of the disease (Buja & Vela 2008).
2.1.6 Cardiomyocyte metabolism
Cardiac muscle cells can use fatty acids, glucose, lactate, ketones and (branched) amino acids as their energy source to work efficiently throughout their lifespan, with fatty acids as their main energy source in aerobic conditions when oxygen supply is not restricted (Aalto-Setälä 2016, 25-26; Giordano 2005; Nabben & Glatz 2016, 7-14).
CMs can generate more energy through fatty acid metabolism than through glucose (Brown et al. 2016, 156). In hypoxic and anoxic conditions where oxygen level is decreased or oxygen lack is complete, such as in ischemic conditions, glycogen is used as primary energy source (Aalto-Setälä 2016, 25-26).
2.2 Hypoxia, ischemia and ischemia-reperfusion injury
2.2.1 Role of oxygen in cardiac function
Constant oxygen supply to the heart is critical for good performance of the organ.
Maintenance of cardiac activity is dependent on the oxygen consumption through oxygen- sensor proteins and hydroxylases. Oxygen participates in various molecular pathways, such as in cellular respiration (production of ATP), which ensures normal function of the cardiac tissue (Townley-Tilson et al. 2015). Oxygen participates in the electron transport chain in mitochondria, being the last electron acceptor in the chain. Myocardial gene expression is
8
dependent on oxygen supply and availability (Giordano 2005). Heart is thus unable to preserve its normal function in anaerobic conditions.
In ischemia, restricted blood flow causes decrease in oxygen supply and in nutrients to the cells, causing tissue hypoxia and cell death (Brown et al. 2016, 158; Townley-Tilson et al.
2015). Cardiac ischemia is related to many heart conditions and diseases, such as MI, hypertension and others. Insufficient oxygen and nutrient delivery cause cells to respond by breaking down glycogen storages. Anaerobic glycolysis takes place, featured with drop in ATP, short lactate and proton production, causing the decreases in pH (Brown et al. 2016, 155-156), and thus even more cell damage.
Tissue damage caused by lack of nutrients and oxygen activates healing process of tissue repair and regeneration, in which genes, responsible for cell migration, proliferation, angiogenesis induction, change in glucose metabolism, induction of pro-inflammatory response, and others, become activated in order to help tissue adjust to changes (Darby &
Hewitson 2016).
2.2.2 Ischemia-reperfusion injury
Studies and reviews describe damage caused to the heart in ischemic conditions after cardiac tissue experiences reperfusion, a recovery of blood flow, as ischemia-reperfusion injury.
Reperfusion injury, in which nutrient and oxygen deprived tissue gets rapid access to vital molecules (Turer & Hill 2010), is known to alter ion transport, crucial for cardiac muscle contractility, mitochondrial function, and change intracellular pathways (Brown et al. 2016, 156-161; Townley-Tilson et al. 2015), causing additional damage to the cardiac tissue (Turer
& Hill 2010). Timing seems to be important factor in ischemia-reperfusion injury, as well as interactions with other cells present or arriving to the site of damage. Short ischemia alters cell function, whereas prolonged ischemia causes permanent cell damage and cell death - apoptosis and mainly necrosis (Brown et al. 2016, 157).
9 2.2.3 Molecular pathways and effects of hypoxia
Hypoxia inducible factor 1 (HIF-1) is an important oxygen dependent transcription factor through which expression of genes that respond to hypoxia is increased (Giordano 2005;
Townley-Tilson et al. 2015). HIF-1 is involved in both normal tissue healing process with restoration of tissue function and structure, and pathological, with occurrence of fibrosis.
Expression of genes, coded for vascular endothelial growth factor (VEGF), transforming growth factor beta-1 (TGF-β1), platelet-derived growth factor (PDGF), among others, become upregulated by HIF-1 when tissue experiences hypoxia (Darby & Hewitson 2016). HIF-1 affects glucose transporters, along with some enzymes participating in glycolysis (Brown et al. 2016, 163). HIF-1 activates GLUT-1, a glucose transporter -1 receptor, expression of which is activated in hypoxia, thus affecting glucose metabolism during oxygen deprivation.
Angiogenesis is induced in hypoxia through activation of HIF-1 (Darby & Hewitson 2016).
2.3 Cardiac fibrosis
Fibrosis of the human cardiac tissue occurs as the response of the heart to the injury. In formation of fibrotic tissue (figure 1) components of the extracellular matrix and their distribution become readjusted in order to reconstruct the damaged area, as in normal tissue healing process. However, characteristic feature of pathological fibrosis is excess production of the ECM components caused by imbalance between degradation and synthesis of fibrillar ECM proteins, particularly the build-up of collagens I and III. This affects function of the heart and structure of the myocardium (Li et. al. 2018; Murtha et al. 2017; Weber 1989).
Fibrosis occurs in various cardiac diseases and pathological heart conditions, such as MI, ischemic conditions, pressure overload, volume overload, diabetic states, cardiomyopathies, heart failure, and others, and during aging (Kong et al. 2014; Li et. al. 2018; Murtha et al.
2017). There is a connection between inflammatory process, wound healing and fibrosis.
Injury to the cardiac tissue causes immediate cell response, mediated by various molecular pathways and different cell type interactions, through which tissue undergoes repair and remodeling, including inflammation with participation of immune cells, proliferation of cells and scar maturation, in which damaged tissue is processed, and injury site is filled with fibrotic tissue (Kong et al. 2014; Murtha et al. 2017; Suthahar et al. 2017).
10
Fibrosis has been described to occur in two ways: reactive fibrosis and replacement (also reparative) fibrosis (Krenning et al. 2010; Murtha et al. 2017; Talman and Ruskoaho 2016;
Weber 1989), with specifics depending on the cardiac disease or condition on the background.
Replacement fibrosis develops at the damaged area, where the site of injury with dead cells is replaced with fibrotic components (Kong et al. 2014; Li et al. 2018; Murtha et al. 2017;
Weber 1989) as a response to ischemia, ischemia-reperfusion injury, inflammation or toxicity (Li et al. 2018), or MI (Talman & Ruskoaho 2016). Areas with CM death, and cardiomyocytes not being able to regenerate themselves, are replaced with fibrotic tissue (Kong et al. 2014; Travers et al. 2016).
Reactive fibrosis can occur without loss of CMs (Travers et al. 2016; Weber 1989) in the areas that surround cardiac blood vessels (Krenning et al. 2010; Weber 1989), spreading into the cardiac muscle, also described as a chronic feature (Murtha et al. 2017). Interestingly, combination of both reactive and replacement fibrosis can occur (Travers et al. 2016) in several heart diseases, such as various types of cardiomyopathies (Li et al. 2018).
Figure 1. Difference between healthy and fibrotic cardiac tissue in replacement fibrosis.
Characteristic features of the replacement fibrosis occurring due to cardiac tissue injury are CM death, activation and transdifferentiation of cardiac fibroblasts into myofibroblasts, and increased production of ECM proteins. Such changes alter structure and function of the tissue.
(Talman & Ruskoaho 2016, picture modified).
11 2.3.1 Characteristic features of cardiac fibrosis
Normal function of the human heart depends on homeostasis of the ECM, in which production, degradation and inhibition of degradation of ECM proteins, regulated by MMPs, TIMPs, cytokines and growth factors, is balanced (Baum & Duffy 2011). Loss of homeostasis, abnormalities in ECM distribution and metabolism can severely alter heart function (Kong et al. 2014; Li et al. 2018). Thus, main characteristic feature of cardiac fibrosis is excessive accumulation of ECM (Kong et al. 2014; Li et al. 2018; Travers et al.
2016; Weber 1989). Production of collagens I and III increases (Kong et al. 2014), although exact rations seem to be disease specific. In MI production of collagen III is preliminary, followed by its decrease, with increased production of collagen type I (Nielsen et al. 2017;
Talman & Ruskoaho 2016).
Another characteristic feature of fibrosis is accretion of activated FBs, which in cardiac injury have been reported to originate from resident fibroblasts, epithelial or epicardial cells, or bone marrow hematopoietic cells, as mentioned in numerous reviews (Fan et al. 2012; Krenning et al. 2010; Murtha et al. 2017; Travers et al. 2016), although cell origin seems to vary depending on the disease (Krenning et al. 2010). FBs have an important role in repair and remodeling of the injured cardiac tissue (Fan et al. 2012; Li et al. 2018; Travers et al. 2016).
Important event occurring during the process of fibrosis formation is FBs transdifferentiation into myofibroblasts (MyoFB), presence of which has been detected in various cardiac diseases (Baum & Duffy 2011; Li et al. 2018; Tallquist & Molkentin 2017). MyoFBs are cells not present in a healthy cardiac tissue (Baum & Duffy 2011). Change in matrix composition and deposition, release of cytokines and growth factors, increase in mechanical stress and neurohormonal factors have been described to stimulate differentiation into MyoFBs (Kong et al. 2014). Apparently, only small number of FBs differentiates into MyoFBs. Exact proportion of main ECM components produced by FBs and MyoFBs remains unclear, since other cell types, CMs and ECs, among others, are also able to secrete specific types of collagens (Tallquist & Molkentin 2017).
12
Figure 2. Cell-cell interactions and key regulators in cardiac fibrosis formation Cell-cell crosstalk include fibroblasts, immune cells (leucocytes, macrophages and others), cardiomyocytes (CM) and endothelial cells. Fibroblast to myofibroblast transdifferentiation is an important step of fibrosis formation, as well as production of ECM components. Molecules expressed by cells and involved in regulation of fibrotic features formation include TGF-β, interleukins and other pro-inflammatory and pro-fibrotic factors. Complexity of intracellular communication involving autocrine and paracrine mediators is represented with bi-directional arrows. (Frangogiannis 2019, picture modified).
2.4 Cellular and molecular regulation of cardiac fibrosis
To better comprehend events occurring in the injured cardiac tissue upon disease or pathological condition, it is important to understand specifics of molecular regulation and molecular interactions. Because cardiac tissue is a complex, highly organized three- dimensional structure, both homeostasis and disease of cardiac tissue involve interactions on molecular, cellular and organ levels. Fibrotic response is mediated by various molecular pathways, which includes interactions of cell types present in the cardiac tissue and arriving cells. Molecular reactions in fibrosis formation involve cytokines, chemokines, transcription and growth factors and neurohormonal pathways (Kong et al. 2014; Murtha et al. 2017). Cell interactions and key molecular regulators can be seen in figure 2.
13 2.4.1 Cellular regulation of cardiac fibrosis
Cross-talk between cells present in the cardiac tissue upon injury is an important part of fibrosis and remodeling of the damaged area. It involves chemical interactions through paracrine manner, autocrine interactions, mechanical and electrical coupling, and interplay with the components of ECM (Banerjee et al. 2006; Borg & Baudino 2011; Murtha et al.
2017), all occurring in a three-dimensional environment.
FBs play important role in cardiac fibrosis. Maintenance and remodeling of the ECM occurs not only by production of proteins of the matrix, but also through cross-talk between other cells present in the tissue. Cardiac FBs also produce regulatory proteins, such as growth factors, cytokines (TGF-β, tumor necrosis factor -alpha (TNF-α), interleukins), MMPs and TIMPS, among others (Brown et al. 2005b; Fan et al. 2012). Secreted biomolecules regulate behavior of FBs and CMs though autocrine and paracrine signaling (Fan et al. 2012; Kakkar
& Lee 2010) involving TGF-β, FGF-2, IL-6 and -33 (Kakkar & Lee 2010). Study by Cartledge et al. (2015) showed that cross-talk between CMs and FBs is bidirectional, and involves TGF-β.
Activation of FBs is a step between their further differentiation into MyoFBs. Review by Tallquist and Molkenitn (2017) described FBs in the adult heart as mature FBs, further becoming activated FBs upon injury, with smaller number of them differentiating into MyoFB in a later response.
MyoFBs transdifferentiated from FBs have many important functions in enhancing cardiac fibrosis. One of the characteristic features of MyoFB is expression of α-smooth muscle actinin (α-SMA), usually not expressed in inactive fibroblasts (Baum & Duffy 2011; Fan et al. 2012), transcription of which is increased as a response to signaling pathways stimulated by cytokines secreted by cells arriving to the site of injury (Baum & Duffy 2011). MyoFBs also produce components of the ECM and cytokines, participating in cell-cell crosstalk and ECM remodelling. MyoFBs have been described to have better ability to produce collagen than FBs (Fan et al. 2012), shown in the earlier study on rat cardiac FBs by Petrov et al.
(2002).
14
Role of immune cells in the fibrosis formation has been studied, suggesting they also have an important role in generation and regulation of the fibrotic response (Brown et al. 2005b; Kong et al. 2014) by secreting pro-inflammatory and pro-fibrotic factors, importantly TGF-β, among others (Frieler & Mortensen 2015; Kong et al. 2014; Suthahar et al. 2017), although they also generate anti-inflammatory cytokines with possible anti-fibrotic effects (Kong et al.
2014; Murtha et al. 2017).
2.4.2 Molecular regulation of cardiac fibrosis
Molecular pathways in inflammatory processes and fibrosis formation involve numerous molecules, interactions of which are rather complex, being a subject of many studies. TGF-β, an important cytokine in maintenance of cell homeostasis, is known to regulate FB, CM, EC and other cells’ behavior in cardiac tissue injury, repair and remodeling, described in numerous reviews (Bujak & Frangogiannis 2007; Dobaczewski et al. 2011; Kong et al. 2014;
Murtha et al. 2017; Travers et al. 2016). Elevated levels of expression of TGF-β, seen in cardiomyopathy, heart failure, myocardial infarction, cardiac hypertrophy and other pathological heart conditions, is associated with pathogenesis of cardiac fibrosis, as has been seen in studies with animal models and patients (Bujak and Fragogiannis 2007; Dobaczewski et al. 2011; Frieler & Mortensen 2015). TGF-β is important in inflammatory response and healing of cardiac tissue, affecting inflammatory cell behavior, fibroblast activation and ECM deposition and remodeling (Bujak & Frangogiannis 2007; Dobaczewski et al. 2011; Kong et al. 2014; Murtha et al. 2017). Review by Dobaczewski et al. (2011) showed that effect of TGF-β can be protective to the heart, participating in maintenance of the cardiac tissue and matrix structure, or deleterious, advancing excessive ECM deposition and formation of scar tissue, with effect caused by either endogenous or exogenous TGF-β (Bujak & Frangogiannis 2007).
Differentiation of cardiac FBs into MyoFBs is guided by TGF-β in inflammatory response, which is an important part of the healing process (Baum & Duffy 2011; Fan et al. 2012; Kong et al. 2014). Correlation between increase in collagen production induced by TGF-β and expression of α-SMA has been earlier reported by Petrov et al. (2002).
Cardiac tissue experiences oxidative stress during fibrosis, as a result of damaged vessels, insufficient or lack of blood flow, and growing demand for energy for arriving cells.
15
Inflammatory cells that arrive at the injury site secrete inflammation inducing factors such as interleukin-6 (IL-6), TNF-α, TGF-β, PDGFs and fibroblast growth factors (FGF) (Kong et al.
2014), and reactive oxygen species (ROS) (Suthahar et al.2017). ROS participates in pathogenesis of cardiac fibrosis formation. It is involved in growth factor and cytokine signaling, expression of ECM proteins and regulation of ECM deposition and cardiac FB behavior (Kong et al. 2014). TNF-α and interleukins IL-6, IL-1β not only participate in inflammatory response (Suthahar et al. 2017) but have also been described to take part in fibrosis formation by affecting matrix remodeling and metabolism (Kong et al. 2014).
PDGF has also been shown to participate in regulation of fibrosis formation by affecting FB behavior, being an important factor, abnormal expression of which severely damages the heart tissue, as reviewed by Kong et al. (2014), Murtha et al. (2017) and Travers et al. (2016).
Interestingly, neurohormonal regulation has also been described as a part of pro-fibrotic process. Macrophages and FBs in the cardiac tissue produce angiotensin-converting enzyme, ACE, needed for angiotensin II (Ang II) production. Ang II, in turn, affects cardiac FB behavior and synthesis of matrix proteins (Kong et al. 2014; Murtha et al. 2017; Travers et al.
2016), and is related to regulation of TGF-β expression (Travers et al. 2016). Also effect of aldosterone has been associated with fibrosis formation through various proposed pathways (Kong et al. 2014).
Recent reviews describe participation of endothelin in remodeling (Brown et al. 2005b; Kong et al. 2014; Ma et al. 2018; Travers et al. 2016) and other signaling pathways to be involved in fibrotic response and regulation of fibroblast behavior (Travers et al. 2016). Also, matricellular components, such as osteopontin, periostin, thrombospondin, biglycan and decorin, have been documented to take part in pro-inflammatory, pro- or anti- fibrotic response and cardiac remodeling (Suthahar et al. 2017).
2.5 Biomarkers of cardiac tissue injury
As previously indicated, hypoxia and ischemia-reperfusion injury leading to formation of replacement fibrosis, has distinct features caused by changes in molecular regulation and cellular behavior with cell death and release of intracellular components into the extracellular
16
space. Certain changes occurring in the cells and injured tissue can be measured, which gives opportunity to estimate the level of cell damage and changes in molecular pathways. This is widely utilized in clinical analysis in order to evaluate disease development and progression, and as a diagnostic tool.
Cardiac cell death caused by damage in cardiac tissue can be studied with CM-specific markers, such as cardiac TnT. Overall cell death can be evaluated by LDH released form the cells. Fibrotic features can be studied by investigating changes in ECM rearrangement and changes in the amounts of secreted collagens.
2.5.1 Troponin T
TnT, together with troponin I (TnI) and troponin C (TnC), is a part of the cardiac troponin complex attached to the actin filament within the sarcomere (Mythili & Malathi 2015;
Parkkila 2016). Being present in the skeletal and cardiac muscle tissue only, TnT is used as biomarker due to its tissue specificity for striated muscle. Release of cardiac TnT occurs upon CM damage (Passino et al. 2015), also shown in the study by Streng et al. (2014).
Cardiac TnT thus serves as a biomarker of CM and cardiac tissue necrosis indicating injury of the myocardium (Mythili & Malathi 2015; Passino et al. 2015).
2.5.2 Lactate dehydrogenase
LDH is an enzyme present in all living cells. Release of LDH occurs upon cell death when cell membrane is damaged and its contents released, such as in necrosis. LDH can be used as a tool for estimating amount of cell death, as in the study by Streng et al. (2014) and it is one of the biomarkers for clinical diagnosis of MI (Mythili & Malathi 2015).
2.5.3 Collagens
Increase in deposition of fibrillar collagens, mainly types I and III, has been described as a fibrotic feature (Weber 1989; 2013), with their ratio increasing or decreasing, depending on the cardiac disease (Jugdutt 2005, 23-36). Early response to MI is degradation of the matrix proteins, increase in collagen III, following increase of collagen I later (Nielsen et al. 2017;
Talman & Ruskoaho 2016). Ischemia, autocrine and paracrine signaling, inflammation,
17
mechanical and other stimuli, affect the activity of MMPs, TIMPs, collagen secretion and matrix remodeling. Thus, following changes in expression of MMPs and TIMPs can also provide information on changes in matrix deposition upon injury (Passino et al. 2015).
2.5.4 α-smooth muscle actinin
To this day specific biomarker for fibroblast has not been found. α-SMA, a contractile protein, is expressed in activated FBs (Ma et al. 2018; Tallquist & Mollketin 2017) and MyoFBs, present during remodeling of the injured cardiac tissue, but not in mature FBs (Tallquist & Molkentin 2017). Thus, it has been used to identify FBs with activated phenotype. α- SMA is also expressed in pericytes and smooth muscle cells (Krenning 2010).
2.5.5 Fibronectin and periostin
Fibronectin, a glycoprotein, and periostin, a protein secreted by FBs and MyoFBs, have been reported as markers for fibroblast activation (Ma et al. 2018). Fibronectin is also present in the tissue when it undergoes changes, including wound healing and pathological conditions (Zollinger & Smith 2017). Fibronectin binds collagens I and III, other ECM components, and growth factors, thus affecting and guiding cell behavior, such as migration, proliferation, repair of the damaged area and other functions (Brown et al. 2005a, 276). Increase in fibronectin upon injury, such as MI, affects formation and assembly of collagen fibers (Murtha et al. 2017; Nielsen et al. 2017), during remodeling of the tissue (Zollinger & Smith 2017). Periostin is found in the fibrotic cardiac tissue (Kong et al. 2014) on site of cardiac injury, where it is secreted mostly by MyoFBs (Kanisicak et al. 2016).
2.6 Studying cardiac tissue in vitro and in vivo
Cardiac tissue is a complex environment which is comprised of different cell types and ECM components. To understand homeostasis of the cardiac tissue and changes occurring upon the disease or injury, it is important to view it as a combination of cells in a very specific microenvironment, which is affected by certain cell alignment and electrical features, with specific mechanical and biomechanical properties, all of which affect cell behavior. Many experiments have been conducted during the last decades to study fibrosis, hypoxia, cardiac
18
diseases and cardiac cell behavior in order to enlighten the mechanisms behind cardiac disease development, progression and potential treatment. Because of the complexity of the cardiac tissue and numerous players involved in cellular cross-talk and molecular interactions, mimicking in vivo environment with in vitro cultured models remains a challenge.
In vivo models continue to be an important tool in studying cardiac tissue. Mammalian models have been widely used, with smaller mammals, such as mice, rats and rabbits, having more advantages over bigger mammals (Camacho et al. 2016). Amount of studies conducted on mouse and rat cardiac tissues suggests rodents’ hearts (ex vivo) to be a good alternative.
Wessels and Sedmera (2003) earlier demonstrated similarities in anatomy of mouse and human hearts, with differences in size, heart rate and some structural features. Cardiac fibrosis and its mechanisms have been widely studied with animal, mostly murine, models, but also monkey, porcine and rabbit (Cowling et al. 2019).
Benefits of the in vivo models are that they represent native microenvironment of the tissue with all necessary components present. However, they can be expensive, hard to control and low-throughput (Kofron & Mende 2017). Despite genetic similarities between humans and animal models, significant differences in their gene and protein expression, cell signaling, and molecular pathways present a problem when comparing cells and organs of animal to the ones of the human being. These differences make experiments on toxic effects and disease modelling unpredictable (Heinonen 2015), forcing science to search for alternatives. Same disease might exhibit different reactions in different animals, presenting another disadvantage of the animal models (Heinonen 2015). Thus, animal models are not able to fully and reliably represent events in the human hearts (Capulli et al. 2016).
Using extracted diseased human hearts or histological human cardiac tissue samples provides a better look at molecular changes occurring upon specific heart conditions. However, their accessibility is challenging due to invasiveness, they are time consuming, and such studies can’t be conducted on regular bases, which makes them unsuitable for routine experiments.
Since such samples show end-point long-term changes, acute molecular responses and progression of the disease can’t be studied on them.
19
Growing understanding on the specifics of the cardiac tissue provides the opportunity for the development of more complex in vitro models. Unfortunately, in vitro models still have their limitations. Although they provide an environment that can be easily controlled, cells are cultured outside their natural microenvironment, with different topography, electrical, mechanical and biochemical properties than in the natural microenvironment inside the tissue.
(Kofron & Mende 2017). Cells in in vitro cultures may also vary in phenotype, depending on their origin, maturity state and activation state, which provides additional challenges in research and result interpretation. Often cultures consist of one or two cell types, which limits the effects of the direct cell-cell communication (Kofron & Mende 2017).
Traditional 2D-models with single layer of cells are very limited in mimicking complex cardiac microenvironment. However, depending on the application of the in vitro model, it is not essential to include all the components of the cardiac tissue environment (Kofron &
Mende 2017). Cardiac 3D models, with maximized cell-cell communication, independent assembly of the ECM, proper cell density and architectural features closer to the organ, are more relevant for studying cross-talk between cells and cell-ECM interactions (Kofron &
Mende 2017). There are also certain challenges with more complex models, such as microtissues and spheroids, related to diffusion of nutrients, oxygen and waste products.
Common challenge in in vitro cardiac models is morphological features of CMs, their deviation in size, shape and alignment (Kofron & Mende 2017), and proper ECM, although nowadays it can be artificially introduced, for example, with cardiac tissue specific scaffolds (Capulli et.al. 2016).
Numerous studies have described their achievements in constructing cardiac tissues in vitro with human derived cells, such as human pluripotent stem cells (hPSC), embryonic (hESC) or induced (hiPSC), differentiating them into cells of the cardiac tissue. Recent researches describe building cardiac (micro)tissues with hiPSC-CM (Mannhardt et al. 2016), hPSC-CMs and hPSC-ECs (Giacomelli et al. 2017), hiPSC-CMs, FBs and ECs (Archer et al. 2018), hESC-CMs and hESC- mesenchymal stem cells (MSC) (Lee et al. 2019), cardiac spheroids with CMs, FBs and ECs (Polonchuk et al. 2017), cardiovascular spheroids with hiPSC-CMs (Yan et al. 2019), cardiac organoids with hESCs (Voges et al. 2017), and hiPSC-CMs (Schulze et al. 2019), vascularized cardiac organoids with hiPSC-CMs, FBs, human adipose derived stem cells (hADSC) and human umbilical vein endothelial cells (HUVEC) (Richards
20
et al. 2017), cardiac microphysiological system with hiPSC-CMs (Mathur et al. 2015), to mention some.
Studying cardiac fibrosis (Galie & Stegemann 2014; Lee et al. 2019; van Spreeuwel 2017), ischemia-reperfusion injury (Pinho et al.2017), cardiac ischemia (Streng et al 2014; Hafez et al. 2018), ischemia and hypoxia (Pitts & Toombs 2004) and cardiac injury (Kanisicak et al.
2016; Voges et al. 2017) with in vitro cardiac tissue models have also been documented.
FICAM’s human vascularized cardiac tissue model (HVCTM) presents yet another new approach to mimic human cardiac tissue using human adipose stromal cells (hASC), HUVECs and hiPSC-CMs, providing new tool for studying molecular specifics of cardiac tissue, its development, disease modeling and drug testing.
21 3. OBJECTIVES
Topic for this thesis has been introduced by FICAM. FICAM has previously developed angiogenesis and cardiovascular in vitro models (Huttala 2019; Vuorenpää 2015) that mimic healthy human tissue. Purpose of this study was to test hypoxic conditions on HVCTM based on these models and to determine if the formation of the most prominent features of fibrosis, rearrangement of the ECM, particularly changes in collagens type I and III, occur in this model under hypoxia of different durations. Another goal of this study was to establish exact duration of hypoxia that causes formation of fibrotic features. The ultimate goal of this work was to discover the ability of this in vitro model to mimic the critical changes occurring in the cardiac tissue upon injury, thus introducing a new approach to study mechanisms and specifics of fibrosis formation de novo, which can be further used in development of drugs and therapies targeting cardiac fibrosis.
22 4. MATERIALS AND METHODS
4.1 Setting up the vascularized cardiac tissue model
HVCTM was constructed according to GLP quality system and FICAM guidelines. HVCTM was established using quality controlled hASCs and HUVECs isolated from human derived tissue (Vuorenpää 2015) and commercial CMs. Briefly, first, angiogenesis model was constructed from primary hASCs and HUVECs (Vuorenpää 2015; Huttala 2019). Cell banked hASCs were thawed and seeded into hASC growth medium (hASCgm) (see appendix 1).
Culture was grown in hASCgm for 1 week with medium change on the 5th day after seeding.
Cell banked HUVECs were thawed and seeded into HUVEC growth medium (HUVECgm) and cultured for 4 days. Both cell cultures were then detached and seeded together into HUVECgm into culture chambers (Metsälä et al. 2018) with following cell densities: 20 000 cells/cm2 for hASCs and 4000 cells/cm2 for HUVECs. Angiogenesis stimulation was performed with angiogenesis medium (ANGm) on 2nd day after seeding into culture chambers, with ANGm medium change every other day.
Construction of the HVCTM (Vuorenpää 2015) was continued with seeding of commercial hiPSC -derived CMs (iCell® Cardiomyocytes, Cellular Dynamics International, Fujifilm Company, USA) on top of the angiogenesis model. CMs were thawed in iCell Cardiomyocyte Plating Medium (iCell® Cardiomyocytes, Cellular Dynamics International, Fujifilm Company, USA) according to manufacturer’s instructions and seeded on top of the angiogenesis model after 1 week of its establishment, with cell density 156 000 cells/cm2. HVCTM was cultured for 1 week in iCell Cardiomyocyte Maintenance Medium (CMmm) (iCell® Cardiomyocytes, Cellular Dynamics International, Fujifilm Company, USA) and ANGm 1:1, with medium change every other day.
4.2 Inducing hypoxia
Hypoxia was achieved with Minihypoxy- portable system (Metsälä et al. 2018), allowing oxygen level decrease down to ~0% O2. To mimic oxygen and nutrient deprivation occurring in the human cardiac tissue during ischemia HVCTM was placed into Dulbecco’s phosphate buffered saline (DPBS) preconditioned in 1% O2 for 24h preceding the hypoxia treatment.
23
HVCTM was then placed into Minihypoxy-system, with air being replaced with 0% O2, 5%
CO2 and remaining N2, conducted in two separate experiments: Experiment 1 and Experiment 2.
The results of the Experiment 1 were used to established optimal duration of hypoxia for the Experiment 2. Hypoxia duration was 2, 4 and 6 hours for Experiment 1, and for 6 hours for Experiment 2. Control group was kept in preconditioned DPBS (1% O2 for 24h) and separately in CMmm and ANGm 1:1 (Experiment 1), and CMmm and ANGm 1:1 (Experiment 2). Ischemia-reperfusion was simulated by medium change CMmm and ANGm 1:1 immediately after hypoxia treatment in both Experiments 1 and 2. There were 16 HVCTM cultures (12 for hypoxia and 4 controls) for Experiment 1, and 12 HVCTM cultures (6 for hypoxia and 6 controls) for Experiment 2. Total of 4 HVCTMs were lost during both experiments due to contamination or technical issues.
4.2.1 Post-hypoxia culture
Changes occurring in post-hypoxia HVCTM were followed. HVCTM was cultured under normoxic conditions (19% O2) for 1 week after hypoxia treatment, with CMmm and ANGm 1:1 medium change every other day. HVCTM was fixated on the 6th day post-hypoxia for Experiment 1, and on days 1, 3 and 6 for Experiment 2 for immunocytochemistry assay (see chapter 4.5 and table 1).
4.3 Troponin T assay
For TnT assay medium samples (200 µL) were collected from each well of hypoxia treated HVCTMs and controls. Samples were taken immediately after hypoxia treatment, prior to every medium change and prior to cell fixation. Samples were stored in -70C° prior to analysis.
For the analysis samples were handled with Human Troponin T enzyme-linked immunosorbent assay (ELISA) Kit (EHTNNT1 Thermo Scientific™ Pierce™, USA) according to manufacturer’s instructions. Spectrophotometric analysis was performed with TEKAN Spark. Absorbance was measured at 450nm and 550nm. TnT results were obtained
24
using Microsoft Excel (version 2016, USA) and presented within linear measuring range of 0 mg/mL – 25 ng/mL. Amount of TnT in each sample was interpolated from the absorbance unit to the TnT concentration using the standard curve in Microsoft Excel (version 2016, USA). Values were calculated from all hypoxia samples (n= (2-12)) and controls (n= (2-6)), each run in duplicate once.
4.4 Lactate dehydrogenase assay
For LDH assay medium samples (200 µL) were collected from each well of hypoxia treated HVCTMs and controls. Samples were taken immediately after hypoxia treatment, prior to every medium change and prior to cell fixation. Samples were stored prior to analysis for 1-3 hours in room temperature (RT). Assay was performed with LDH Cytotoxicity Assay Kit (Thermo Scientific™ Pierce™, USA) according to manufacturer’s instructions.
Spectrophotometric analysis (ELISA) was performed and absorbance measured at 490nm and 680nm with TEKAN Spark. Results were obtained using Microsoft Excel (version 2016, USA). Values were calculated from hypoxic (n= (2-12)) and control (n= (2-6)) samples, each run in triplicate once.
4.5 Immunocytochemistry
Immunocytochemical staining was performed for collagens I and III, TnT and fibronectin in order to reveal changes in their amounts, deposition and structures. HVCTM were fist fixated with 70% ethanol for 20min at RT. After this, HVCTM were permeabilized with Triton X- 100 for 15min at RT. Blocking of non-specific binding sites was performed with 10% bovine serum albumin (BSA) for 30min at RT. HVCTM were stained with primary antibodies (see table 1) diluted with 1% BSA in DPBS overnight in +4C°. Secondary antibodies (see table 1), diluted with 1% BSA in DPBS, were applied on the next day, incubated for 45min at RT.
Nuclei were stained with Hoechst (33258 Solution, Sigma-Aldrich, USA) with dilution 1:1000. Washes in between were performed with phosphate buffered saline (PBS) or DPBS.
All samples were imaged with Nikon Eclipse Ti –fluorescent microscope (Nikon, Japan) and results obtained using NIS-Elements AR 5.11.00 image program (Nikon, Japan). Confocal imaging was performed with Zeiss LSM 780 confocal microscope (Zeiss, Germany).
Confocal images were obtained with Imaris x64 program (version 9.3.1, Bitplane, USA).
25
Antibody Dilution Manufacturer information Polyclonal rabbit anti-fibronectin 1:400 PA5-29578, Invitrogen, Thermo
Fisher Scientific, USA
Monoclonal mouse anti-collagen I 1:100 MA1-141, Invitrogen, Thermo Fisher Scientific, USA
Polyclonal rabbit anti-collagen III 1:100 PA5-34787, Invitrogen, Thermo Fisher Scientific, USA
Monoclonal mouse anti- TnT 1:100 MA5-12960, Invitrogen, Thermo Fisher Scientific, USA
Alexa fluor 594 donkey anti-rabbit IgG
1:400 A21207, Invitrogen, Thermo Fisher Scientific, USA
Alexa fluor 488 donkey anti-mouse IgG
1:400 A21202, Invitrogen, Thermo Fisher Scientific, USA
Table 1. Antibodies used in the immunocytochemical staining
4.6 Oxygen level measurement
Oxygen levels were measured from culture chamber (Metsälä et al. 2018) with preconditioned 1% O2 DPBS and CMmm and hASC basic medium (hASCbm) 1:1 separately from the Experiments 1 and 2. Measurements were performed in triplicate. During measurements, there were no cells inside the culture chamber. Medium was kept in the chamber until 19%
oxygen level was achieved. Medium was then changed into preconditioned DPBS (1% O2 for 24h) and its oxygen level decrease followed to obtain measurements for 2, 4 and 6 hours of hypoxia. Measurements were performed using fluorimetric oxygen sensor (Välimäki et al.
2017). Analysis of numerical data for oxygen measurement and graphical representation of oxygen level decrease was kindly provided by J. Kreutzer.
4.7 Statistical analysis
To determine statistical significance of the results, all numerical data from LDH-and TnT- assays was analyzed with Student’s t-test (Microsoft Office Excel, version 2016, USA).
26
Significance (*-***) was established as follows: p < 0,005 was considered significant (*), p <
0,001 (***) was considered highly significant.
27 5. RESULTS
Study process is outlined in figure 3. Timelines depict all the stages of Experiment 1 and Experiment 2. Also, differences in the processes can be seen.
Figure 3. Timeline for the Experiments 1 and 2. Three first weeks of both experiments were same and included seeding, separate culturing of hASCs and HUVECs and establishment of angiogenesis model, following the establishment of HVCTM. Experiment 1 was a single end-point experiment with duration of hypoxia treatment being 2, 4 and 6 hours, and culture fixation on the 6th day post-hypoxia. Experiment 2 was conducted to show gradual changes in the components of ECM caused by 6-hour hypoxia. Culture fixations occurred on days 1, 3 and 6 post-hypoxia in Experiment 2.
5.1 Experiment 1
Experiment 1 was a single end-point experiment. After successful establishment of angiogenesis model and HVCTM, culture was placed under hypoxic conditions with 2, 4 and 6 hours of hypoxia treatment in order to establish duration of hypoxia that causes fibrotic feature formation.
According to LDH and TnT results (figures 4 and 5), Experiment 1 cells experienced peak in total cell and CM necrosis on day 1, compared to controls in medium. 2-hour hypoxia
