Refining kidney organoid protocols

(1)

Refining kidney organoid protocols

THESIS

submitted in partial fulfillment of the requirements for the degree of

MASTER OFSCIENCE in

PHYSICS

Author : A.M. Vlaar

Student ID : s1535013

Supervisor : Dr. S. Semrau

Co-supervisors : Dr. M. Hochane

E.E. Adegeest MSc

2ndcorrector : Dr. M.J.A. de Dood

(2)

(3)

Refining kidney organoid protocols

A.M. Vlaar

Huygens-Kamerlingh Onnes Laboratory, Leiden University P.O. Box 9500, 2300 RA Leiden, The Netherlands

March 17, 2020

Abstract

Current kidney organoid protocols that differentiate human induced pluripotent stem cells into kidney cell types are still unable to grow an entire set of kidney cell types nor functional kidney structures. Protocols

of Morizane[1] and Taguchi[2][3] were adapted to improve their results on one human induced pluripotent stem cell line. When we adjusted the

Taguchi protocol we were unable to differentiate the cells beyond the metanephric mesenchyme stage. However, with the refined Morizane protocol, we observed podocyte-like cells which clustered together and formed small structures. Nevertheless, the cell line that we used did not

form tubular structures as was expected from the Morizane paper. Presumably, different cell lines respond differently to the same protocol.

Therefore, in addition to refining kidney organoid protocols, it is recommended to increase the docility of cell lines such that cells exhibit

(4)

(5)

Chapter

1

Introduction

Patients who suffer from kidney failure are often waiting for years before a matching kidney transplant is available, in the meantime, they are bound to dialysis which has a massive impact on their quality of life. And even after successful transplantation, the patient has to take medication for life, to prevent rejection of the transplant.

Recently, embryonic kidney development has been mimicked in the laboratory[1][3], resulting in the growth of kidney organoids (i.e. minia-ture organs). Although these kidney organoids are not complete func-tional organs, they have potentially a great impact on modern health-care. If kidney organoids could develop further into actual kidney organs, transplantation of these organs would, per definition, not result in trans-plant rejection because the kidney organoid can be made form the patients own cells. The kidney organoids can also find their application in phar-maceutical research for drug testing. And when stem cells from patients are used to create kidney organoids, the organoids can be used to model genetic diseases and to create personalized medicines[4].

Research into the use of stem cells to grow organs has developed since the middle of the last century. In 1960, Weiss[5] used differentiated em-bryonic cells to grow chick organs. Evans[6] succeeded in 1981 to culture mouse embryonic stem cells and it took until 1998 before Thomson[7] cul-tured successfully human embryonic stem cells. The controversial use of human embryonic stem cells restricts the research on human organ devel-opment. Until, in 2006 Takahashi and Yamanaka[8] generated a new type of stem cells, induced Pluripotent Stem cells (iPSCs). They reprogrammed mouse adult skin cells into pluripotent stem cells by introduction of only four factors: Oct3/4, Sox2, c-Myc and klf4. These mouse iPSCs have the same potential as embryonic stem cells, i.e. grow into a complete

(8)

organ-ism. One year after this major discovery, human skin cells were used to create human iPSCs (hiPSCs) which opened up the opportunity to grow human organs without major ethical concerns.

Nowadays different organoids are grown from induced pluripotent stem cells e.g. liver[9], heart[10], lung[11], brain[12] and kidney[1] [2] [3] [13] [14]. In the existing kidney organoid protocols, part of the kidney filter unit can be created which form lumen[1][2] and branching of the draining system is achieved[3]. But creating fully functional organoids re-mains a challenge. In the existing protocols, non-kidney cells have been produced e.g. neural and muscle cells[13]. The complete variety of cell types, present in the actual kidney, cannot be generated with the current protocols. Functional compartments of the kidney are not connected and cannot exhibit the kidney function i.e. filtering blood and collecting urine. Finally, cell types that are developed using these protocols do not reach a mature state.

By refining existing protocols, we wanted to improve the purity of kidney organoids that were grown from hiPSCs. First, the protocols of Morizane[1] and Taguchi[2][2] to generate kidney organoids were carried out as published. The obtained organoids lacked lumen formation. Next, the same protocols, but with several adaptations to improve structure for-mation, were carried out. These protocols resulted in a small curved align-ment of cells. Although the protocols need to be enhanced before actual kidney organs can be developed, results obtained during this project con-tribute to the elaboration of kidney organoids.

In the chapter hereafter, the anatomy and development of the human kidney will be explained, followed by a remark on the kidney organoid protocols that we used. Experimental results of the (refined) Morizane[1] and Taguchi[2][3] protocols will be reviewed and discussed, after which the methods will be elucidated.

(9)

Chapter

2

Theory

2.1 The kidney

2.1.1 Anatomy and function of the kidney

The kidneys are bean-shaped organs that have the size of a fist. The two kidneys of a human are located left and right of the spine. The main func-tion of the kidney is to excrete waste products of metabolism in urine [15]. The human kidney is composed of 4 to 8 lobes, where each lobe consists of three major structures: the cortex, which is at the periphery of the kidney, the medulla, which lies adjacent to the cortex, and the papilla, which is connected to the ureter (see Figure 2.1).

The functional unit of the kidney, called the nephron, is located in the lobes of the kidney. These nephrons filter the blood, recapture molecules and collect the urine. Each kidney contains approximately 1 million neph-rons[17] that are predominantly distributed in the cortex. The nephrons are also found in the medulla, which is mostly composed of collecting tubules that drain the urine into the papilla before being pumped into the ureters (see Figure 2.1).

When unfiltered blood enters the nephron via the afferent arteriole (see Figure 2.2), the first filtration steps take place in the glomerulus, a dense package of capillaries and epithelial and interstitial cells, which is sur-rounded by Bowman’s capsule.

The structure of the glomerulus is maintained by mesangial cells that reside in between the renal capillaries[18] (see Figure 2.2). The renal capil-laries consist of glomerular endothelial cells that are characterized by their multiple fenestrations (see Figure 2.2). When blood is pumped through the renal capillaries, the vast majority of molecules, except for large

(10)

pro-Figure 2.1: Anatomy of the human kidney. The human kidney consists of sev-eral lobes and contains around 1 million nephrons that are situated in the cortex and the medulla. Blood enters and leaves the kidney via the renal blood ves-sels. The nephrons filter the blood and produce urine. Blood is filtered in the glomerulus. The glomerulus is surrounded by Bowman’s capsule. This capsule collects the urine and is connected to the proximal tubule. The proximal tubule continues through the loop of Henle that is connected to the distal tubule. The distal tubule is linked to the collecting duct that comes together in the papilla from which the urine can flow to the ureter. Adopted from: Vecteezy.com and LifeMap Sciences[16]

teins, can enter these holes and pass through the glomerular basement membrane. On top of the glomerular basement membrane, podocytes are situated[15]. These podocytes have long arms extending from the cell body, out of which foot processes radiate. These foot processes, called pedicels, form a slit diaphragm with an aperture of around 40 nm[17]. Molecules that flow through the slit diaphragm will enter Bowman’s cap-sule. The inner lining of the Bowman’s capsule is constituted by Parietal epithelial cells[19].

The ultrafiltrate that is formed in the glomerulus will then pass through complex tubular structures (see Figure 2.1). These are respectively: the proximal tubule, the loop of Henle and the distal tubule. These tubules are responsible for the reabsorption of molecules and ions that are neces-sary for the homeostasis of body fluids, blood pressure and blood compo-sition. The proximal tubule reabsorbs glucose, amino acids, and essential minerals. Within the loop of Henle, the filtrate gets concentrated. In the distal tubule, water, sodium, and calcium are preserved and the pH is regulated[17]. The remaining filtrate, the urine, is gathered in the collect-ing duct which will lead it to the papilla and finally towards the ureter.

Besides filtering the blood and regulating the body fluid composition, the kidneys execute more tasks[17]. They maintain the blood pressure by managing the amount of blood that flows through the nephrons, due to

(11)

2.1 The kidney 11

Figure 2.2: Cell types in the glomerulus.The structure of the glomerulus is main-tained by mesangial cells that fill up the space between the glomerular capillaries. The glomerular capillaries are formed by endothelial cells that are fenestrated and form the first filter barrier. The glomerular basement membrane lies between the capillaries and podocytes. These podocytes are situated on the outer part of the glomerulus and extend into foot processes, forming 40 nm gaps (see inset). The glomerulus is surrounded by Bowman’s capsule, which is covered with parietal epithelial cells and connected to the proximal tubule. The proximal tubule is char-acterized by its brushed border and extends into the distal tubule, which loops back to pass the glomerulus from the side. There, Macula Densa cells are located in between the glomerulus and the distal tubule. The Macula Densa cells track the salt concentration of the urine and give signals to the juxtaglomerular cells. The Juxtaglomerular cells are located on the incoming blood vessel and control the blood pressure[17]. Adopted from: M. Hochane.

contraction and dilatation of the blood vessels, and by regulating the ion balance. The kidneys also activate vitamin D and produce erythropoietin, which stimulates the production of red blood cells.

The different stages in which the kidney develops is discussed here-after, starting with the general development of a human embryo.

2.1.2 Embryonic Development

Right after conception, the fertilised egg starts dividing. Around day 5, two cell types can be distinguished in what is now called the blastocyst[20] (see Figure 2.3). The cells at the edge of the blastocyst are called the outer cells, and the cells that clump together at one side form the inner cell

(12)

Figure 2.3: Formation of the three germlayersThe blastocyst forms in the first two weeks following fertilisation and consist of an outer layer of cells and an inner cell mass. The inner cell mass differentiates further into epiblast and hy-poblast cells. These two layers of cells form a bilaminar disk. At day 15 of de-velopment, the primitive streak appears on top of the epiblast-disk. Cells from the epiblast migrate towards the streak and then downwards to form a new layer below the epiblast layer. The newly formed layer is called mesoderm and consists of three segments: lateral plate mesoderm, intermediate mesoderm and paraxial mesoderm. The epiblast cells are now turned into ectoderm and the hypoblast cells have become endoderm.

mass[20]. The embryo is formed out of the inner cell mass[20]. The in-ner cell mass develops into the epiblast and hypoblast layers at about day 8 to 9[20]. These two layers form the bilaminar disk as depicted in figure 2.3 (upper right). On top of the epiblast layer, a structure arises around day 15. This structure is called the primitive streak[20]. The picture at the bottom right in figure 2.3 shows a cross-section of the bilaminar disk. Cells from the epiblast start migrating across the primitive streak to form the mesoderm layer in between the epiblast and hypoblast[21]. The epi-blast and hypoepi-blast differentiate into the ectoderm and endoderm layer, respectively. The ectoderm, mesoderm and endoderm together form the three germ layers. Each tissue in the human body originates from one of these three layers. The kidneys are formed in the mesoderm layer, more specifically in the intermediate mesoderm which is positioned between

(13)

2.1 The kidney 13

the paraxial mesoderm and lateral plate mesoderm[22] (see Figure 2.3, bottom left). How the kidney develops from the intermediate mesoderm is described below.

2.1.3 Development of the kidney

The kidney evolves from the intermediate mesoderm out of two compo-nents. In the anterior part of the intermediate mesoderm, the nephric duct is formed. When the nephric duct enters the caudal region of the embryo, interaction takes place between the nephric duct and metanephric mes-enchyme cells (see Figure 2.4). The metanephric mesmes-enchyme is formed in the posterior intermediate mesoderm [2]. The posterior intermediate mesoderm is generated by the cells that migrate from the primitive streak at the late stage[1]. The interaction between the nephric duct and metanephric mesenchyme induces the outgrow of the ureteric bud from the nephric duct[17], which is depicted in Figure 2.4. The outgrowth is promoted by the tyrosine kinase receptor (RET), its ligand the glial-derived neu-rotrophic factor (Gdnf) and the coreceptor glycophosphatidylinositol-linked GRFα1[22]. The metanephric mesenchyme cells condense around the tip of the ureteric bud and form the cap mesenchyme. The cap mesenchyme induces the ureteric bud to grow into the metanephric mesenchyme and the ureteric bud starts branching[22]. While branching of the ureteric bud continues, some metanephric mesenchyme cells form interstitial tis-sue (stroma) and other metanephric mesenchyme cells form clusters around the newly formed tips. These clusters, called pretubular aggregates, trans-form into a renal vesicle[22] (see Figure 2.4). As the renal vesicle grows, a notch occurs on one side of the vesicle, this changes the vesicle into a comma-shaped body. Opposite of the first notch, a second notch arises, transforming the comma-shaped body into an s-shaped body[22] (see Fig-ure 2.4). The cells at the distal end of the s-shaped body, the end closest to the ureteric bud, fuses with the ureteric bud to form a continuous col-lecting system[22]. On the proximal end of the s-shaped body, most dis-tant from the ureteric body, the cells of the s-shaped body differentiate into podocyte precursors that express podocyte-specific proteins: transcription factor WT1, Pod1; transmembrane proteins podocin, Glepp1, Nephrin; and Vascular Endothelial Growth Factor (VEGF)[22]. Endothelial cells are probably attracted by the VEGF and start to invade the proximal end of the s-shaped body[23]. The cells of the s-shaped body start to form Bowman’s capsule around the endothelial cells. These endothelial cells express the Flk1 receptor for VEGF and form the renal capillaries[22]. Growth and

(14)

de-velopment will shape the nephron and tubular structures until it becomes a fully functional kidney.

With the knowledge on kidney development, we have tried to regener-ate the kidney from stem cells by simulating kidney development. To do this, we used hiPSCs. What these stem cells are and how they are created is explained hereafter.

2.2 Stem cells

Stem cells are unspecialized cells that can either renew or, under particular conditions, specify into more specialized cells[24]. Cells that are extracted from the inner cell mass of an embryo (see section 2.1.2) are called em-bryonic stem cells. Since these emem-bryonic stem cells can generate every tissue present in the embryo, but not extra-embryonic tissue like the pla-centa, the cells are called pluripotent[24]. Research on human embryonic stem cells is considered controversial since these cells might be counted as individuals. Another type of stem cells is the multipotent cells that can be found in adult organs, called adult stem cells. Bone marrow is an ex-ample of adult stem cells and it can generate several cell types that can be found in the blood[24]. Bone marrow is multipotent since it only re-produces blood cells and not, for example, skin cells. In a recent study by Clevers et al[25], cells from the cortex of an adult kidney were used to grow tubular structures. The cortex cells can be considered as adult kidney stem cells, but are not capable of regenerating the whole kidney. Theoret-ically, hiPSCs are able to form a kidney. The hiPSCs are first described by Takahashi and Yamanaka[8], when they reprogrammed skin cells by the overexpression of four factors: Oct3/4, Sox2, c-Myc and Klf4. The cells that they created were able to differentiate into all three germ layers (see section 2.1.2) and had, therefore, the possibility to differentiate into every tissue of the human body. This makes hiPSCs useful to investigate the in vitro development of the human kidney.

By differentiating hiPSCs in a specific manner, 3D organized tissue can be obtained which expresses genes of specific cell types. These 3D tis-sues are referred to as organoids. Initiation of differentiation and mor-phogenesis is specified by the timing, concentration, and combinations of growth factors and signalling molecules that are added to the hiPSCs [26]. The protocols, in which these specifications are indicated, to grow kidney organoids will be discussed below.

(15)

2.2 Stem cells 15

Figure 2.4: Nephron development.The ureteric bud is formed out of the nephric duct when it interacts with the metanephric mesenchyme. The condensed mes-enchyme forms a cap around the ureteric bud and induces branching. Some mesenchyme forms vesicles below the ureteric bud tips. The renal vesicle will develop into a comma-shaped body and evolves further into an s-shaped body. One end of the s-shaped body will fuse with the ureteric bud to form the col-lecting system. The nephron is formed at the other end of the s-shaped body including Bowman’s capsule.

(16)

2.3 Kidney organoid protocols

There are several existing protocols on kidney organoids, such as Bon-ventre[27], Freedman[13], Morizane[1], Taguchi[2][3] and Takasato[14]. All differ in length, requirements and outcome. We have chosen to work on the protocol of Morizane since it is recommended for beginners by Nishinakamura[26]. We also worked with the protocol of Taguchi because it is relatively fast: it only takes 11 or 14 days whereas, for example, Freed-man continues for 21 days.

To differentiate the hiPSCs into kidney-like cell structures, several fac-tors are used. Some of these facfac-tors form the basis for all protocols i.e. CHIR99021, Activin and FGF9. The compound CHIR99021, further re-ferred to as CHIR, activates the Wnt signalling pathway[28]. Signalling pathways are triggered by extracellular signals and cause a cascade upon which the gene expression of a cell is altered[24]. During development Wnt signalling pathway plays an important role. Upon activation, Wnt signalling pathway affects the expression of genes that are needed for a cell to obtain a new identity or fate. The growth factor Activin supports the pluripotency of hiPSCs and during the formation of organs, it directs the proliferation and differentiation of progenitor cells[29]. The protein FGF9 (fibroblast growth factor 9) promotes maintenance of nephron pro-genitor cells[30].Without FGF9, branching of the ureteric bud stops and the development of the kidney fails[31]

The protocols of Morizane and Taguchi show an opposite approach concerning the administration of BMP4 at the start of differentiation. The protein BMP4 induces the Posterior Primitive Streak[32] that gives rise to lateral plate mesoderm. Therefore, Morizane uses noggin, that blocks the BMP4 activity[33]. Taguchi uses in his protocol of 2014 BMP4 to induce mesodermal lineage, based on research by Kattman[34] who showed that increased level of BMP4 induces mesoderm that is connected to the for-mation of Hematopoietic cells. These blood cell precursors originate from lateral plate mesoderm[35]. This might be a reason why, in the revised protocol of Taguchi in 2017, BMP4 is not administrated anymore in the first stage of differentiation. In-depth details of the Morizane and Taguchi protocols will be discussed below.

2.3.1 Morizane protocol 2015

The protocol designed by Morizane[1] starts with a 72h long preparation of the cells. The single hiPSCs are grown in a dish in ReproFF2, a medium that is designed to maintain the undifferentiated state and the

(17)

differentia-2.3 Kidney organoid protocols 17

tion capacity. FGF2 is added to the media to promote the pluripotent state of the cells[36]. Due to the dissociation into single cells, cells lose contact with their environment and other cells, which could lead to apoptosis[37], a form of cell death. To prevent apoptosis ROCK inhibitor Y27632 is added to the media. After three days of preparation, day 0 of differentiation starts. A schematic overview of the differentiation steps is displayed in Figure 2.5. During differentiation, the medium ReproFF2 is replaced for a basic differentiation medium that consists of Advanced RPMI 1640 sup-plemented with Glutamax. Glutamax is a stabilized form of L-glutamine which is an energy source for the cell.

At day 0 of differentiation, CHIR and noggin are added to the cells to drive them towards the late primitive Streak stage. After 4 days, the cells are guided towards the posterior intermediate mesoderm by Activin. At day 7, FGF9 is used to convert the cells into metanephric mesenchyme. Two days later, CHIR is added along with FGF9 to induce pretubular ag-gregates. From day 9, the protocol can be continued in 2D cell culture or transferred to 3D. Irrespectively of the culture environment, the proto-col steps remain the same. From day 11 to day 14 of differentiation, the treatment with FGF9 is continued to let the organoids grow renal vesicles. These vesicles can grow further into renal epithelial structures when the cells are kept in the basic differentiation medium.

Figure 2.5: Schematic overview of the Morizane protocol of 2015. The differen-tiation of cells starts in a 2D culture environment. In the first 4 days, CHIR and noggin converse the stem cells towards the late primitive streak stage. The three subsequent days, the cells are treated with Activin and become posterior inter-mediate mesoderm. Until day 9, only FGF9 is added. At this time point, the cells can be transferred to a 3D environment. The protocol remains the same for both 2D and 3D culture. From day 9, CHIR is added along with FGF9 for 2 days. At day 11, the CHIR treatment ends and FGF9 is continued until day 14. From this day on, the organoids are let to spontaneously differentiate for several weeks.

(18)

2.3.2 Taguchi Protocol

We have tested two protocols from Taguchi. The first one was published in 2014[2] and this protocol was modified by Taguchi in his paper of 2017[3]. Both protocols were performed by Taguchi on the same hiPSCs. A schematic overview of the protocols of 2014 and 2017 are displayed in Figure 2.6 and Figure 2.7, respectively.

In the protocol of 2014, the cells are differentiated towards the meta-nephric mesenchyme stage, which is comparable to the metameta-nephric neph-ron progenitor stage in the protocol of 2017. However, only the last two steps (day 9 and 11 in Figure 2.6) coincide with the last two steps in the 2017 protocol (day 7 and 9 in Figure 2.7). These two steps consist of a mix-ture of factors. First Activin, BMP4, CHIR, Retinoic Acid (RA) and ROCK inhibitor are added for two days. Where the retinoic acid signalling path-way is involved in apoptosis, differentiation and cell fate designation[38]. In the last step, CHIR, FGF9 and ROCK inhibitor are used to promote dif-ferentiation.

Where the protocol of 2014 starts with the administration of Rock in-hibitor and BMP4, in 2017 BMP4 is replaced by Activin and FGF2, the components that were used 24h later in the protocol of 2014. In both pro-tocols, the cells are treated with CHIR for 6 days. In 2014, CHIR is accom-panied by BMP4, which is replaced by ROCK inhibitor in the protocol of 2017. Rock inhibitor and BMP4 do not exhibit the same function and the motivation behind the modifications is not elucidated.

Figure 2.6: Schematic overview of the Taguchi protocol of 2014.After 24h, hiP-SCs form embryoid bodies in presence of BMP4 and ROCK inhibitor (ROCK I). Activin and FGF2 are added at day 1 to get the cells in the epiblast stage. From day 3 until day 9, BMP4 and CHIR are used to get the aggregates via the nascent mesoderm stage into the posterior nascent mesoderm stage. At day 9, Activin, BMP4, CHIR and Retinoic Acid (RA) bring the cells in the posterior intermedi-ate mesoderm stage. In the last step the organoids are differentiintermedi-ated towards the metanephric mesenchyme stage by adding CHIR and FGF9.

(19)

2.4 Cell type determination 19

Figure 2.7: Schematic overview of the Taguchi protocol of 2017. On day 0, ROCK inhibitor (ROCK I), FGF2 and Activin are added to the human induced Pluripotent Stem Cells (hiPSCs) to turn them into the epiblast stage. After 24h, CHIR and ROCK inhibitor are administered for 6 days to get the cells via the nascent mesoderm stage into the posterior nascent mesoderm stage. At day 7, Activin, Bmp4, CHIR and Retinoic Acid bring the cells in the posterior interme-diate mesoderm stage. In the last step, the organoids are differentiated towards the metanephric nephron progenitor stage by adding CHIR and FGF9.

2.4 Cell type determination

In the developmental process of the kidney and within the full-grown kid-ney itself, many different cell types can be distinguished, as is described in section 2.1.3. When executing a protocol to grow kidney organoids, as de-scribed in section 2.3, it is important to evaluate the process by determin-ing the cell types that are present. Cells can be classified in different ways, e.g. by morphology, density and gene expression. During this project, we have identified the cells based on their gene expression pattern i.e. specific genes are high, low, or not expressed. An overview of genes that are (+) or aren’t (-) expressed by different cell types, during kidney organogenesis, is given in table 2.1.

The expression of a gene can be validated by looking at the presence of a gene-specific protein. The protein can be detected using fluorescence microscopy. To do so, fluorescent molecules need to be attached to the protein. This can be achieved by using chimeric proteins[24], where the gene corresponding to the protein of interest is combined with the gene of a natural fluorescence protein. This technique creates the possibility to investigate the expression of a gene and dynamics of the protein within a living cell. We have used another method to attach fluorescent molecules to a protein of interest, called immunostaining. Immunostaining is based on the immune response of the body to foreign microorganisms. The pro-teins of the foreign microorganism consist of a specific part called the anti-gene. The antigene can be recognized by a protein of the body, this protein is called the antibody. The antibody binds to the antigen to start the elim-ination of the foreign microorganism[24]. The antibody itself also has an

(20)

antigene part, to which another antibody can bind. Immunostaining con-sists usually of one or two antibodies. The primary antibody binds to the protein of interest. If only one antibody is used, a fluorescent molecule is attached to the primary antibody, otherwise, the fluorescent molecule is attached to a secondary antibody which binds to the primary antibody. A fluorescence microscope can be used to excite the fluorescent molecule and detect the emitted fluorescent light. Details on the fluorescent light-sheet microscope, which is in general used to image large, dense biological specimen such as organoids, are discussed below.

2.5 Light sheet microscope

Figure 2.8: Regular and light-sheet microscopeThe regular fluorescence micro-scope uses the same objective to illuminate the whole sample and to detect the fluorescent light. The light-sheet fluorescence microscope uses a separate objec-tive to illuminate only part of the sample. Adopted from: E.E. Adegeest

To identify the different cell types which are discussed in section 2.4, in the big, 3D kidney organoids, a fluorescent light-sheet microscope is likely more suitable than a regular fluorescence microscope.

In conventional fluorescence microscopes, only one objective is used to shed the excitation light on the sample and to collect the emitted fluores-cence light (see Figure 2.8). As a consequence, all fluorescent molecules in the sample are excited at the same time. In live imaging, the amount of

(21)

2.5 Light sheet microscope 21

Cell type and expression of corresponding genes Late Primitive Streak

+ TBXT TBX6

- PAX2 LHX1 WT1 OSR1 HOXD11 FOXF1

Posterior Intermediate mesoderm

+ WT1 HOXD11 OSR1

- PAX2 LHX1 TBXT TBX6 SIX2

Metanephric Mesenchyme

+ SIX2 WT1 PAX2 SALL1 EYA1

Nephron progenitors

+ CITED1 SIX2 TMEM100

Pretubular Aggregates + PAX8 LHX1 - LAM Renal Vesicles + PAX8 LHX1 LAM S-shaped body + PTH1R CDH1 MAFB Proximal tubules + CDH2 AQP1 LTL Podocyte precursors + MAFB Podocytes

+ NPHS1 PODXL WT1 SYNPO MAFB

Renal Collecting Duct System

+ CLDN7 CDH1

Table 2.1: The different (progenitor) cell types of the kidney are stated with the corresponding proteins that they do (+) or do not (-) express. A cell type can be confirmed by demonstrating that certain proteins are (or aren’t) expressed. Composed with information from [1], [3], [39], [17] and [40].

light needed to illuminate the whole sample is fairly high, which can be phototoxic, i.e. can cause damage to the cell or even lead to cell death[41]. Exciting all fluorescent molecules at the same time can also be disadvanta-geous when imaging fixed samples. While focusing on one plane to get an image, the fluorescent molecules in the other planes are also excited and might degrade (i.e. photobleaching) before being imaged [41].

The light-sheet microscope uses a separate objective to illuminate the sample. The laser beam that exits this objective is shaped to a µm-thin

(22)

sheet that excites the fluorescent molecules in the focal plate of the de-tection objective (see Figure 2.8). Therefore, the sample is introduced to less energy than with a regular fluorescence microscope, making the light-sheet microscope a well-suited technique for imaging live biological samp-les[41]. The detection objective collects the fluorescence light from the fo-cal plane. Since the photobleaching effects are reduced to the fofo-cal plane, the detection efficiency and optical sectioning is greatly improved.

To image larger samples, the light dose needs to be increased. The stan-dard low energy that the light-sheet microscope uses, creates an opportu-nity to image larger samples with less photobleaching effects compared to a regular fluorescence microscope. This makes the light sheet micro-scope an appropriate instrument to image the large 3D kidney organoid structures[42].

(23)

Chapter

3

Results

3.1 Morizane protocol led to podocyte-like cells

The Morizane protocol was executed twice. In experiment I, the protocol of Morizane was repeated without major adaptations. Part of the cells was kept in 2D until day 28, others were transferred to 3D on day 9. In experi-ment II, the protocol was modified in several ways (see section 3.1.2).

3.1.1 Experiment I

2D culture

To initiate differentiation, the hiPSCs were grown along with CHIR and noggin for 4 days to drive the cells towards the late primitive streak stage, after which Activin was added to the cells to guide the cells towards the posterior intermediate mesoderm stage. In vivo, this stage leads to the metanephric mesenchyme from which the nephrons originate. To achieve the transition towards the metanephric mesenchyme stage in vitro, the cells were grown in FGF9 from day 7 to day 9.

At day 8 the cells were expected to be differentiated towards the meta-nephric mesenchyme stage. To see if this was the case, the cells were stained for LHX1 and PAX2. The nuclei were counterstained with DAPI. Figure 3.1 shows that LHX1 is not stained specifically, indicating that it is not present. Whereas the PAX2 marker is clearly visible, which suggests that the cells are differentiated into metanephric mesenchyme. However, no pretubular aggregates have formed yet.

The cells were further differentiated into pretubular aggregates using FGF9 with CHIR, from day 9 until day 11. Treatment with FGF9

(24)

admin-Figure 3.1: Cells grown in 2D at day 8 of differentiation in the Morizane pro-tocol have differentiated into metanephric mesenchyme. Cells immunostained with LHX1 (pretubular aggregates), PAX2 (Metanephric Mesenchyme) and DAPI. Scale bar: 100 µm, magnification: 20x.

istered on day 11 to day 14 was expected to differentiate the pretubular aggregates into renal vesicles.

At day 14 of the Morizane treatment cells were fixed and stained with PAX2 and TBX6. If the cells developed into metanephric mesenchyme or further, TBX6 is not supposed to be expressed since it is a marker of the posterior intermediate mesoderm. However, staining could not be clearly observed (see Supplementary A).

When the cells were cultured in basic differentiation medium for one week (day 21), to spontaneously differentiate, cells were stained with LTL which is specific to the brush borders of the epithelium of the proximal tubule (see figure 2.2), SYNPO which marks the podocytes and TMEM100 which is expressed by the nephron progenitor populations in the human developing kidney[39]. The nuclei were counterstained with DAPI. Re-gions of cells expressed SYNPO (see Figure 3.2, upper left). LTL signal was found around the whole cell (see Figure 3.2, bottom picture), but the LTL signal was expected to be found at the membranes of the cells since LTL stains the brushed borders of epithelial cells. The LTL staining was therefore not specific. TMEM100 was also not expressed, therefore, at day

(25)

21 of differentiation, there was no population of nephron progenitors nor proximal tubules were formed, only podocytes.

Figure 3.2: Cells grown in 2D at day 21 of differentiation in the Morizane proto-col have differentiated into podocytes.Cells stained with , SYNPO (podocytes), LTL (proximale tubules), TMEM100 (nephron progenitors) and DAPI. scale bar large images: 1000 µm, magnification: 10x, scale bar small images: 50 µm, mag-nification: 40x. Regions of cells expressed SYNPO (see image upper left). LTL staining was not specific, since it was not located at the membrane only (see bot-tom picture). Staining for TMEM100 was negative.

The cells were kept in basic differentiation medium and cultured until day 28 of differentiation. The cells were then fixed and stained for LTL, SYNPO and TMEM100. The nuclei were counterstained with DAPI. Fig-ure 3.3 displays two cell layers at the same position, 20 µm apart. Remark-able are the differences in the size of the nuclei and their density among the cell layers. Comparable to the cells at day 21, LTL and TMEM100 were not expressed, and the cells did express SYNPO (see Figure 3.3).

Thus the Morizane protocol in 2D at day 21 and day 28 resulted in podocyte cells and in contrast to results in the paper no tubular structures were formed.

(26)

Figure 3.3: Cells grown in 2D at day 28 of differentiation in the Morizane protocol have differentiated into podocytes. Cells were stained for SYNPO (podocytes) and DAPI. scale bar: 50 µm, magnification: 40x. Top (Z = 0) and bottom (Z = -20µm) image show different sections of the same position. The bot-tom image shows smaller nucluei. Cells in both layers expressed SYNPO.

(27)

From 2D to 3D

Until day 9, the cells were treated in the same way as the 2D culture de-scribed above. At day 9 the cells were expected to be in the metanephric mesenchyme stage. To transfer the cells from 2D to 3D, the cells were therefore dissociated and replated in a 96-well U-bottom plate with a den-sity of 100 000 cells per well. To promote aggregation of the cells, the plates were centrifuged at 200×g for 1 minute. During growth in the 3D environment the cells did not seem to form spheres but rather flat struc-tures with holes in the middle (see Figure 3.4 Day 14). As the treatment continued some of these structures did form dense organoids (see Figure 3.4 Day 27), noticeable are the smooth edges on the left side, which might indicate structure forming cells.

Figure 3.4: Cells grown in 3D at day 14 and 27 of differentiation in the Morizane protocol formed flat structures. Scale bar: 500 µm, magnification: 4x. Left: Organoids at day 14. The flat structures were ring or horseshoe-shaped. They showed widely-spread bright areas indicating differences in the densities of the forming structures. Right: Organoids at day 27. Some of the organoids formed closed, dense structures (top right). Sharp round boundaries are visible at the left side of the organoid. Other organoid formed a ring-shaped structure with less bright spots (bottom right).

(28)

induce the formation of pretubular aggregates. The FGF9 treatment was continued until day 14 when the organoids were expected to have formed renal vesicles.

Figure 3.5: Regions of cells grown in 3D at day 14 of differentiation in the Morizane protocol are metanephric mesenchyme. The organoids were stained with DAPI, TBX6 (Late primitive streak), LHX1 (Pretubular aggregates) and PAX2 (Metanephric Mesenchyme). scale bar: 100 µm, magnification: 40x. Upper Left: Staining for DAPI (top) and PAX2 (bottom) shows clear but locally-restricted staining for PAX2. Upper Right: Staining for PAX2 and TBX6. Cells that ex-pressed PAX2, did not express TBX6. Bottom: Staining for TBX6 (left) and LHX1 (right) overlapped, therefore, it was not possible to determine if TBX6 of LHX1 was stained specifically.

At day 14 of the Morizane treatment, organoids were fixed and stained with PAX2 (metanephric mesenchyme), LHX1 (pretubular aggregates), and TBX6 (late primitive streak). The nuclei were counterstained with DAPI. The staining for LHX1 and TBX6 was found to overlap but not with the PAX2 staining (see Figure 3.5). Therefore, it was not possible to deter-mine whether the staining of TBX6 or LHX1 was specific. Outside these TBX6/LHX1 regions, a cluster of cells was found to express PAX2. Which suggests that the cells are in the metanephric mesenchyme stage and did

(29)

not form pretubular aggregates. The organoids at day 14 of differentia-tion were also stained with CITED1 and SIX2, both indicators of nephron progenitor cells. SIX2 is also a marker for metanephric mesenchyme. The nuclei were counterstained with DAPI. The cells show specific expression of SIX2 and do not express CITED1 (see Figure 3.6). Since SIX2 is also a marker of metanephric mesenchyme, the cells are not in the nephron pro-genitor stage, but rather in the metanephric mesenchyme stage.

Figure 3.6: Cells grown in 3D at day 14 of differentiation in the Morizane pro-tocol have not differentiated into nephron progenitors.Organoids were stained with DAPI, SIX2 (nephron progenitors/ metanephric mesenchyme) and CITED1 (nephron progenitors). scale bar: 200 µm, magnification: 10x. The staining for CITED1 is not specific. Individual cells are recognizable in the SIX2 staining, indicating that these cells do express SIX2. The cells have differentiated into metanephric mesenchyme and not into nephron progenitors.

At day 21, the organoids were stained for LTL (proximal tubules), SYNPO (podocytes) and TMEM100 (nephron progenitors). The image could not be clearly interpreted and can be found in the supplementary, figure A.3.

The organoids at day 28 were cut in sections of 20 µm thick (see meth-ods 5.6) and stained for LTL (proximal tubules), SYNPO (podocytes) and CDH1 (i.e. E-cadherin). Where CDH1 is a marker of epithelialised tubular structures (e.g. the s-shaped body). The nuclei were counterstained with DAPI. The cells were not positive for LTL and CDH1 (data not shown), but most cells did express SYNPO (see Figure 3.7). The organoids have

(30)

there-fore differentiated into podocyte-like cells and did not form epithelialised structures. In some of the organoids, a remarkable pattern could be found where the cells formed a vortex (see figure 3.7).

Figure 3.7: Cells grown in 3D at day 28 of differentiation in the Morizane pro-tocol have differentiated into podocytes. Sections were stained with DAPI and SYNPO (podocytes). Big pictures, scale bar: 500 µm, magnification: 20x. In-serts, scale bar: 50 µm, magnification 100x. SYNPO is expressed throughout the organoid, indicating that the cells formed podocytes. In the middle of the section, the cells have a particular orientation and seem to swirl around a central point.

3.1.2 Experiment II

The Morizane protocol was executed again in the original form (Control) and in an adapted form (DMSO pre-treatment and 3D from the start). Dur-ing the control experiment, a small modification was made when the cells were transferred from a 2D cell culture to a 3D environment. Instead of ag-gregating the cells in a U-bottom plate and centrifuging the plate, the cells were replated in a 96-well V-bottom plate for 2 days without centrifuga-tion. At day 11 the cells were transferred to a U-bottom plate. This small modification of the protocol resulted in more spherical like structures (see figure 3.11, control) compared to the flat structures in experiment I.

(31)

Control experiment

The Morizane protocol to grow kidney organoids in 3D was executed sim-ilarly as described in section 3.1.1. At day 14, the organoids were expected to have formed renal vesicles and were stained for PAX2 (metanephric mesenchyme), LHX1 (renal vesicles) and TBX6 (late primitive streak). The nuclei were counterstained with DAPI. The staining did not give a clear indication if certain genes were expressed, figure A.4 can be found in the Supplementary.

Figure 3.8: Cells grown in 3D at day 28 of differentiation in the Morizane protocol control experiment formed a curved structure of podocytes. Sections were stained with DAPI, PTH1R (S-shaped body) and SYNPO (podocytes). Top pictures, scale bar: 200 µm, magnification: 20x. Other pictures, scale bar: 20

µm, magnification: 100x, maximum intensity z-projection. Several cells, with a

big nucleus (see white arrowhead), which is a characteristic of podocytes, clus-ter together and express SYNPO. The clusclus-tered cells seemed to have formed a glomerulus-like structure (see yellow arrowheads). The cells were negative for PTH1R.

Organoids harvested at day 28 were sectioned (see methods 5.6) and stained for SYNPO (podocytes), LTL (proximal tubules), PTH1R (s-shaped

(32)

body), CDH1 (s-shaped body), MAFB (podocytes precursors) and CLDN7 (renal collecting duct system). The nuclei were counterstained with DAPI. The stainings for LTL, CDH1 and CLDN7 were negative (data not shown), therefore, the organoids did not form epithelialised tubular structures. However, SYNPO was clearly expressed, see figure 3.8, indicating that the organoids have differentiated into podocyte-like cells.

Figure 3.9: Cells grown in 3D at day 28 of differentiation in the Morizane pro-tocol control experiment did not show clear staining for podocyte precursors: MAFB and PTH1RSections were stained with DAPI, PTH1R (podocyte precur-sors) and MAFB (podocytes precurprecur-sors). Top picture, scale bar: 200 µm, magni-fication: 20x. Other pictures, scale bar: 20 µm, magnimagni-fication: 100x, maximum intensity z-projection. Deconvoluted, Land Weber, 5 iterations. MAFB was ex-pressed at its expected localisation, that did not apply to PTH1R staining which seemed to overlap MAFB staining.

To indicate the developmental stage of the podocytes, sections were stained for MAFB. This is a marker for podocyte progenitor cells. MAFB is expected to be expressed in the nucleus and the cytosol[40]. According to the staining in figure 3.9, MAFB seems to be expressed by the whole cell. The staining might be specific, but the signal of MAFB is very low. The staining of PTH1R, another marker of podocytes, seems to be real staining

(33)

because the high signal is twice as high as the background signal outside the nucleus (5000:2000). The staining for PTH1R also seems to overlap with MAFB, but PTH1R should be expressed on the plasma membrane and in the nucleus. Based on their signal intensity and localisation, both PTH1R and MAFB are not stained specifically.

DMSO pre-treatment

To increase the efficiency of differentiation of the organoids into mature nephron cells, in the DMSO pre-treatment experiment, the cells were treated with 2% DMSO (Dimethylsulfoxide) for 24 h before differentiation, based on the protocol of Sambo[43]. DMSO is expected to extend the G1 phase of the cell cycle, the phase in which the cell decides to proliferate or differenti-ate[44]. The hiPSCs are characterised by a short stay in the G1 phase, due to which pluripotent stem cells are more willing to multiply[43]. By in-creasing the G1 phase, the DMSO prepares the cells for differentiation which increases the chance that cells follow the differentiation path to ma-ture nephron cells.

Figure 3.10: Cells grown in 2D from day 2 to 9 of differentiation in the Morizane protocol pre-treated with DMSO showed a different density pattern compared to cells of the control experiment. Bright-field images at different time points of one well that was pre-treated with DMSO (bottom) and one with-out DMSO pre-treatment (top). Scalebar day 2: 1 mm, scalebar day 4,7 and 9: 2 mm. From day 4 the control is more confluent than the DMSO well. Gaps of low density also appear in the control well, as can be clearly seen at day 7, but are more present in the DMSO well. On the contrary, the density of the concentrated cells in the DMSO well are much higher compared to the control well. Day 2, 1 tile. Day 4, tiled 3x3, Day 7, tiled 5x5, day 9, tiled 6x6.

(34)

Although the cells were only cultured in 2% DMSO for 24h, the DMSO had a great influence on their way of growing in a 2D culture dish, which became visible at day 4 of differentiation (see figure 3.10). The cells formed more dense areas, leaving a less confluent well. At day 9 the cells were dissociated into single cells and transferred to a 3D environment. At this time point, the viability, as well as the number of cells, was the same for both wells.

Figure 3.11: Organoids grown in 3D from day 14 to 18 of differentiation in the Morizane protocol pre-treated with DMSO decreased in size. Rows represent the same organoid at different time points. Scale bar: 500 µm. At day 10 the organoids are cultured in a V-bottom 96-well plate and transferred the next day to a U-bottom 96-well plate. From day 14 the organoids were cultured in basic dif-ferentiation medium without growth factors. After which all organoids decreased in size. However, the DMSO pre-treated organoids shrunk more compared to the organoids in the control experiment.

The cells pretreated with DMSO seemed to be prone to form high-density clusters, which was expected to be beneficial when the cells were transferred to a 3D cell culture. When the organoids were cultured in the V-bottom well at day 9 and 10, the DMSO organoids resembled those of the control experiment (see figure 3.11). From day 14, the organoids were cultured in basic differentiation medium without additional growth factors. At this point the DMSO pretreated organoids looked different from the control, there seemed to be more cell death around the DMSO organoids. Between day 14 and day 18 all organoids reduced in size, but

(35)

the DMSO organoids shrunk more. The control organoids were consistent in size from day 18 on and stayed condense until day 28. This condense structure seemed to disappear in the DMSO pretreated organoids from day 25 when an increase of cell death was observed.

Figure 3.12: Cells grown in 3D at day 14 of differentiation in the Morizane pro-tocol pre-treated with DMSO show PAX2 staining in small clusters.Organoids were stained for PAX2 (metanephric mesenchyme). Scale bar: 200 µm, magnifi-cation: 20x. The PAX2 staining seemed to be spread in small clusters around the organoid.

At day 14, the organoids were expected to have formed renal vesi-cles. The results of the organoids in the control experiment are discussed above in section 3.1.2. The organoids pretreated with DMSO were also stained for PAX2 (metanephric mesenchyme), LHX1 (renal vesicles) and TBX6 (late primitive streak). The nuclei were counterstained with DAPI. The DMSO organoids at day 14 were less dense than the control, making it possible to recognise individual cells (see figure 3.12). LHX1 and TBX6 were not expressed (data not shown). However, PAX2 had a sporadic ex-pression throughout the organoid. Compared to the organoids at day 14 in experiment I (see figure 3.5) where PAX2 seems to be expressed in larger regions, the DMSO treatment resulted in a different expression pattern.

3D from the start

The development of the kidney occurs in a 3D manner within the em-bryo, therefore, to mimic the development of the kidney we started the differentiation in a 3D culture instead of a 2D culture. But having a 3D culture is also a challenge, the aggregates are not attached and can easily

(36)

be destroyed when the media is changed. It is therefore not possible to completely refresh the media in contrast to a 2D culture. The Morizane protocol starts in 2D and moves to 3D after 9 days, the moment when the nephron progenitors start to form 3D tubular structures. But to go from 2D to 3D the cells need to be dissociated, a process that might negatively influence the cells. Single-cell dissociation of hiPSCs increase cell death by apoptosis. To circumvent the step from 2D to 3D, we cultured the cells in a 96-well U-bottom plate from the start at a density of 4000 cells per well. The plate was centrifuged until it reached 100×g, to promote aggregation of the single cells. The organoids were treated with growth-factors based on the Morizane protocol.

The organoids kept growing until day 14, from then the media did not contain any additional growth factors. Comparable to the control and DMSO organoids (see figure 3.11) these organoids become smaller after day 14. The amount of cell death increased and at day 28, most of the organoids do not appear vital anymore (see figure 3.13).

Figure 3.13: Cell death increased from day 18 to day 28 in organoids that were grown in 3D from the start in the Morizane protocol. Scalebar: 500 µm. Rows represent the same organoid at different time points. The organoids grow in size until day 14. At that time point, the organoids were left to differentiate without additional growth factors. During this process, the organoids reduced in size, which was stable from day 18 until day 21. The amount of cell death increased from day 18 and the compact structure seemed to fall apart at day 28.

At day 14 the organoids were expected to have formed renal vesicles. The organoids were stained for LHX1 (renal vesicles), TBX6 (late primitive streak) and PAX2 (metanephric mesenchyme). The nuclei were counter-stained with DAPI. LHX1 and TBX6 were not expressed (data not shown).

(37)

In figure 3.14 PAX2 expression seemed to be concentrated around a small area, indicating that a tiny region is metanephric mesenchyme. Probably the organoids are too condense to get a clear image on the individual cells, making it difficult to determine if the staining is specific.

Figure 3.14: Cells grown in 3D at day 14 of differentiation in the Morizane pro-tocol cultered in 3D from the start expressed PAX2 in a small region.Organoids were stained for PAX2. Pictures on the right, scale bar: 50 µm, 20x. Other pictures, scale bar: 200 µm, 20x. In a small region cells seem to express PAX2.

To conclude the results of the Morizane protocol, at day 14 the hiP-SCs were expected to have formed renal vesicles, but in all experiments, the marker LHX1 for epithelialised tubular structures was not expressed. The organoids only express the metanephric mesenchyme marker PAX2 in a small region. This expression pattern was different in the DMSO pre-treated organoids. At day 28, all organoids expressed the podocyte marker SYNPO. In the control experiment, when the organoids were more spheri-cal than in experiment I, podocyte-like cells clustered to form a glomerulus-like structure.

(38)

3.2 Taguchi protocols led to different expression

patterns

The Taguchi 2014 protocol[2] differentiates the hiPSCs into metanephric mesenchyme. In 2017, Taguchi published an adapted protocol[3] based on his 2014 protocol, to grow metanephric nephron progenitors. We have executed both protocols and tried to extend the protocol of 2017 to induce epithelialised tubular structures.

3.2.1 Protocol of 2014

In the Taguchi protocol of 2014, the cells are differentiated towards the metanephric mesenchyme stage. BMP4 was added for 24h to induce em-bryonic bodies (EBs). In the next two days, the cells were driven towards the Epiblast stage by adding Activin and FGF2. To differentiate the cells into the posterior nascent mesoderm stage, BMP4 and CHIR were added for 6 days. At day 9, Activin, BMP4, CHIR and Retinoic Acid were added to get the cells in the posterior intermediate mesoderm stage. Two days later, CHIR and FGF9 induced the differentiation into metanephric mes-enchyme.

At day 14, the organoids were expected to be metanephric mesenchyme and were stained for PAX2 (metanephric mesenchyme), TBX6 (late prim-itive streak) and LHX1 (pretubular aggregates). The nuclei were counter-stained with DAPI. In most organoids (8) it was not possible to identify in-dividual cells and therefore it was not possible to make a statement on the expression of genes. Two organoids were probably less dense and showed specific staining (see figure 3.15). The top part of the organoid stained pos-itive for PAX2 and LHX1, indicating the differentiation towards nephron progenitors. The bottom part expressed TBX6, a marker of cells in the late primitive streak stage.

(39)

Figure 3.15: Organoids grown in 3D at day 14 of differentiation in the Taguchi 2014 protocol show expression of both nephron progenitors and late primi-tive streak.Organoids were stained for PAX2 (metanephric Mesenchyme), LHX1 (pretubular aggregates), TBX6 (Late primitive streak). Scale bar: 200 µm, magni-fication: 20x. Z projection, maximum intensity. Most cells in the top part of the organoid seem to express PAX2. Small regions in the top part express LHX1. The bottom part expresses TBX6.

3.2.2 Protocol of 2017

In the Taguchi protocol of 2017, the cells were found to have differentiated into metanephric nephron progenitors within 11 days. In the first step, Ac-tivin was added for 24h to differentiate the cells towards the epiblast stage. CHIR, without additional BMP4, was added for 6 days to differentiate the cells towards posterior nascent mesoderm. Half of the medium was re-freshed every other day. To induce posterior intermediate mesoderm, the media was changed into CHIR, Activin, BMP4, Retinoic Acid. Two days later, at day 9, FGF9 and CHIR differentiated the cells in two days to the metanephric nephron progenitors.

At day 11, the organoids were expected to be metanephric nephron progenitors and were stained for PAX2 (Metanephric Mesenchyme), LHX1 (Pretubular aggregates) and TBX6 (late primitive streak). The organoids did not express TBX6, indicating that the organoids differentiated further than the late primitive streak stage. LHX1 was also not expressed, which

(40)

suggests that the cells did not start to form tubular structures. PAX2 was expressed in almost every cell (see figure 3.16) indicating that the cells were at the metanephric mesenchyme stage.

Figure 3.16: Organoids grown in 3D at day 11 of differentiation in the Taguchi 2017 protocol show expression metanephric mesenchyme. Organoids were stained for PAX2 (Metanephric mesenchyme) and TBX6 (late primitive streak). Scale bar: 50 µm, 40x. Shown is probably a segment of a bigger organoid and not a representative image. Other figures of bigger organoids were blurry and indi-vidual cells were not distinguishable. The TBX6 dot at the top right of the picture is not specific. In this fragment, all cells express PAX2.

3.2.3 Maturation of metanephric nephron progenitors

To further differentiate the metanephric nephron progenitors into epithe-lialized nephrons, Taguchi co-cultured the organoids with embryonic spinal cord or induced ureteric bud. Since other protocols like Morizane[1] can grow renal vesicles without embryonic spinal cord and ureteric bud, five experiments with growth factors were executed to mature the metanephric nephron progenitors. An overview of the experiments can be found in ta-ble 3.1. In the first experiment, spontaneous differentiation was promoted by refreshing basic differentiation medium on a regular basis. The sec-ond experiment was inspired by the protocol of Bonventre[27], CHIR was

(41)

added for 24h, which was then changed for FGF2 and Retinoic acid. Two days later the media was changed to FGF9 and Activin. After 3 days basic differentiation medium was added without additional growth factors. Ex-periments 3 and 4 are based on the idea that treatment with an activator of the Wnt signalling pathway for a short period would induce matura-tion of metanephric mesenchyme. In experiment 3 the non-canonical Wnt signalling pathway was activated with R-spondin 1 for 24 h. Experiment 4 activated the canonical Wnt pathway by adding CHIR for 24 h. The last experiment was inspired by the protocol of Morizane[1], where a pulse of CHIR was added next to FGF9 and ROCK inhibitor Y-27632 for 2 days. The media was then changed to only FGF9 and ROCK inhibitor Y-27632 for 3 days.

Day Exp 1 Exp 2 Exp 3 Exp 4 Exp 5

1 BDM CHIR Rspondin1 CHIR

CHIR FGF9 ROCK I 2 FGF2 RA BDM BDM 3 FGF9 ROCK I 4 BDM FGF9 Activin A BDM BDM 5 6 BDM BDM BDM BDM 7 BDM 8 BDM BDM BDM BDM

Table 3.1:Overview of the five experiments to differentiate metanephric nephron progenitors further into more mature nephron structures. The empty boxes in-dicate that the medium was not refreshed at that point. ROCK I: Rock inhibitor Y27632. BDM: basic differentiation medium.

After 6 days of treatment the organoids that could differentiate spon-taneously seem to grow bigger than the rest and showed dark and light regions. Which could indicate internal structures. The organoids of the second experiment showed small density differences within the organoid, their size was in between experiment 1 and the other experiments. The organoids of experiment 3, 4 and 5 were small and had no visible differ-ences in density. For the following days, all organoids became denser (see picture 3.17) and shrank every day. This indication made it very likely that

(42)

the organoids were not growing nor differentiating, on which the decision was based to abort the experiments.

Figure 3.17: Taguchi, 2017. Day 17. Representative pictures of organoids at day 17 after aggregation. The numbers correspond to the five experimental conditions (see table: 3.1). The organoids, kept shrinking and appeared to be in a non-vital state.

When executing the Morizane protocol, we expected to grow kidney organoids that would form epithelialised tubular structures. However, our organoids consisted only of podocytes. With the 2014 protocol of Taguchi, we aimed to grow metanephric mesenchyme. However, our organoids expressed, besides the metanephric mesenchyme marker PAX2, also TBX6 and LHX1 which are markers of the late primitive streak and pretubular aggregates respectively. The organoids that were grown ac-cording to the Taguchi protocol of 2017, did express the metanephric mes-enchyme marker PAX2. However, we were unable to let these organoids differentiate further into nephron-like structures.

(43)

Chapter

4

Discussion

In this project, we have refined protocols of Morizane and Taguchi to grow kidney organoids from hiPSCs. The Morizane protocol led mainly to podocyte-like cells. When we slightly adapted the method to transfer the cells to a 3D cell culture, several cells formed a glomerulus-like structure. The organoids that were grown using the 2014 protocol of Taguchi, ex-pressed other markers that were not expected in the intended metanephric mesenchyme. The 2017 protocol of Taguchi resulted in purer organoids that only expressed the metanephric mesenchyme marker. We extended the 2017 protocol to further mature the nephron progenitors, which was not successful.

We have carried out the Morizane protocol as published once in the 2D version, and twice in the 3D version. In all 3 experiments, we expected to have metanephric mesenchyme at day 8 and renal vesicles at day 14. The cells in the 2D experiment at day 8 expressed PAX2, indicating that the cells were metanephric mesenchyme. At day 14 the organoids in all 3 experiments did not express LHX1, the marker for pretubular aggregates. However, small regions within the organoids did still express PAX2.

The protocol of Morizane was altered by pretreating the cells with DMSO to enhance their differentiation towards mature nephrons. In the first 9 days, the cells that were pretreated with DMSO already showed a different behaviour than the control experiment. The cells formed more condensed regions but the number of living cells was the same as in the control experiment. It is interesting that, although DMSO was only ad-ministered for 24h, it influenced the behaviour of the cells even many days later. At day 14, the expression pattern of PAX2 deviated from the other organoids. The DMSO pretreated organoid showed patchy expression of PAX2. Due to lack of time, it was not possible to stain and section the

(44)

DMSO pretreated cells and see what effect the DMSO had on the cells at later stages.

Another adaptation was made to the Morizane protocol by culturing the cells in 3D from the start of the protocol. These organoids showed sub-stantial cell death around day 28. The dead cells by which the organoids were surrounded, might cause the cells to get stressed resulting in more cell death. Washing the organoids before adding new media could reduce the influence of dead cells on the organoids.

When the Morizane protocol was executed for the second time in ex-periment II, all organoids shrank in size between day 14 and day 18. The sizes of organoids in experiment I were not tracked and therefore cannot be compared. At day 14 the medium was switched to basic differentia-tion medium only and cells should spontaneously differentiate. It could be, that a passive selection process caused a large amount of cell death, selecting only those cells that can survive in basic differentiation medium without growth factors.

The main result of the Morizane experiments is that the majority of cells have differentiated into podocytes-like cells at day 28 of differentia-tion. There is a remarkable difference in appearance between the organoids at day 28 of experiment I and II, which might be caused by the difference in shape. The organoids in experiment I were rather flat compared to the more spherical structure of the ones from experiment II. Thus the way the cells are collected in 3D, a round bottom well with centrifuging or in a V-bottom well for 2 days, might influence the shape of the organoids.

The organoids at day 28 in our control experiment might have ex-pressed MAFB and PTH1R, both marker for podocyte precursors. To clar-ify their staining, it would be useful if, besides the nuclear staining with DAPI, another staining could be used for the plasma membrane. This could make it easier to see which part of the cell is stained by the marker and to state whether it is specific. According to the Morizane protocol, we also expected to have tubular structures beside podocytes on day 21 and 28 of differentiation. However, our organoids were only expressing podocyte markers. The different cell line that we used compared to Morizane might cause this contrasting result.

The Taguchi protocol of 2014 states that their organoids expressed mark-ers of the metanephric mesenchyme at day 14. Organoids that were cated during our project, using this protocol expressed PAX2 in a large re-gion were some cells expressed LHX1, another rere-gion expressed TBX6. This patterning indicates that not all cells are at the metanephric mes-enchyme stage. Some regions are still at the stage of the late primitive streak (TBX6). LHX1 is a marker of the intermediate mesoderm and

(45)

pre-45

tubular aggregates/renal vesicles. Since TBX6 is also expressed, which is a marker of an earlier developmental stage, it is also possible that the ex-pression of LHX1 is an indication of cells at the intermediate mesoderm stage. According to the protocol of Sambo[43], a pre-treatment of DMSO could increase the number of cells that differentiate into the same stage.

In the Taguchi protocol of 2017, it is not explicitly stated which genes were expressed at day 11, only that the cells were at the stage of metan-ephric nephron progenitors. The organoids obtained with this protocol did express PAX2 but not TBX6 nor LHX1. These results seem to match a stage of metanephric nephron progenitors.

Regardless of the protocol, the whole mount immunochemistry stain-ings were not always unambiguous. In this project a good, but time-consuming solution was to make sections. In the Morizane paper, whole mount stainings were much clearer than images presented in this thesis. Possible reasons for this difference in clarity and solutions to improve the images are describe hereinafter.

It is possible that the organoids were much larger than the ones grown by Morizane and Taguchi. At day 14 an organoid from the Morizane paper was approximately 400 µm, in experiment II a control organoid was about 1500 µm. This may be due to a faster proliferation of the cell line we used (sigma 0028) compared to the cells used by Morizane et al.

The fixing time that we carried out, was significantly longer than in Morizane’s and Taguchi’s protocols. Taguchi (2017) fixes for 60 min, Mori-zane fixes the organoids for 20 min, compared to several days of fixing as can be seen in table 5.1. The fixation time should increase with sam-ple thickness, to ensure that PFA can penetrate the samsam-ple fully. But if the fixation time is too long, it might form to many methylene bridges and pre-vents antibodies to reach their target side. Maybe the over-fixation could also cause more scattering during imaging, producing a blurry image.

The clearing protocol was used to improve the quality of the images. In the clearing protocol of Dekker[45], the samples are left at RT for 20 min after the fructose-glucose solution is added. Where the gastruloids, used in the clearing experiment, were almost invisible by eye, the organoids of Morizane exp. I. day 14 and Taguchi 2017, did not appear more transpar-ent. The time was increased from 20 min to 3,5h when the organoids of Morizane exp. I day 21 and 28 were cleared, but it did not improve the image quality (see figure A.3). Organoids at day 14 of exp. II and Taguchi 2014 were put at 28°C for 2,5h. Most organoids looked much more trans-parent except the organoids of the control experiment. The image of the latter, see figure A.4 was not clear enough to interpret. This adaptation of the clearing protocol seems to have slightly improved the image quality as

Refining kidney organoid protocols

Refining kidney organoid protocols

Refining kidney organoid protocols

A.M. Vlaar

Abstract

Contents

Chapter

1

Introduction

Chapter

2

Theory

2.1

The kidney

2.1.1

Anatomy and function of the kidney

2.1.2

Embryonic Development

2.1.3

Development of the kidney

2.2

Stem cells

2.3

Kidney organoid protocols

2.3.1

Morizane protocol 2015

2.3.2

Taguchi Protocol

2.4

Cell type determination

2.5

Light sheet microscope

Chapter

3

Results

3.1

Morizane protocol led to podocyte-like cells

3.1.1

Experiment I

3.1.2

Experiment II

3.2

Taguchi protocols led to different expression

patterns

3.2.1

Protocol of 2014

3.2.2

Protocol of 2017

3.2.3

Maturation of metanephric nephron progenitors

Chapter

4

Discussion