Title: The building blocks for cardiac repair : isolation and differentiation of progenitor cells from the human heart

The handle http://hdl.handle.net/1887/62812 holds various files of this Leiden University dissertation.

Author: Moerkamp, A.T.

Title: The building blocks for cardiac repair : isolation and differentiation of progenitor cells from the human heart

Issue Date: 2018-06-12

I

I SOL ATION AND CHARACTERIZATION OF PROGENITOR CELLS FROM

HUMAN FETAL AND ADULT HEART TISSUE

41

3

M A B C19 TARGETS A NOVEL SURFACE MARKER FOR THE ISOL ATION OF CARDIAC PROGENITOR CELLS FROM HUMAN HEART TISSUE AND DIFFERENTIATED H ESC ^S

Moerkamp A.T. ^1,∗ , Leung H.W. ^2,∗ , Padmanabhan J. ³ , Ng S.W. ³ , Goumans M.J. ¹ , Choo A. ⁴

1. Department of Molecular Cell Biology, Leiden University Medical Centre, Leiden, The Netherlands.

2. Bioprocessing Technology Institute, Agency for Science, Technology and Re- search, Singapore; NUS Graduate School for Integrative Sciences and Engineering, National University of Singapore, Singapore.

3. Bioprocessing Technology Institute, Agency for Science, Technology and Re- search, Singapore.

4. Bioprocessing Technology Institute, Agency for Science, Technology and Re- search, Singapore; Department of Bioengineering, National University of Singa- pore, Singapore.

∗ Contributed equally to this work

Published in Journal of Molecular and Cellular Cardiology (2015)

43

3

Abstract

Cardiac progenitor cells (CPCs) have been isolated from adult and developing hearts

using an anti-mouse Sca-1 antibody. However, the absence of a human Sca-1 homo-

logue has hampered the clinical application of CPCs. Therefore, we generated novel

monoclonal antibodies (mAbs) specifically raised against surface markers expressed

by resident human CPCs. Here, we explored the suitability of one of these mAbs, mAb

C19, for the identification, isolation and characterization of CPCs from fetal heart

tissue and differentiating cultures of human embryonic stem cells (hESCs). Using

whole-cell immunization, mAbs were raised against Sca-1+ CPCs and screened for

reactivity to various CPC lines by flow cytometry. MAb C19 was found to be specific

for Sca-1+ CPCs and suitable for the isolation of multipotent CPCs from both human

heart tissues and differentiating hESCs. MAb C19 stained small stem-like cells in

cardiac tissue sections. Moreover, during differentiation of hESCs towards cardiomy-

ocytes, a transient population of cells with mAb C19 reactivity was identified and

isolated using magnetic-activated cell sorting. Their cell fate was tracked and found

to improve cardiomyocyte purity from hESC-derived cultures. MAb C19+ CPCs, from

both hESC differentiation and fetal heart tissues, were maintained and expanded in

culture, while retaining their CPC-like characteristics and their ability to further dif-

ferentiate into cardiomyocytes by stimulation with TGF β1. Finally, gene expression

profiling of these mAb C19+ CPCs suggested a highly angiogenic nature, which was

further validated by cell-based angiogenesis assays.

3

3.1. I NTRODUCTION

The heart has long been considered to be an organ with very limited regenerative capabilities. This has been challenged by several studies reporting that the heart harbors a resident population of cardiac progenitor cells (CPCs) that might be used to repair the injured heart [1]. A variety of isolation methods, such as cardiosphere formation, and the use of surface markers like C-Kit, kinase domain receptor (KDR) and stem cell antigen-1 (Sca-1), have been utilized to enrich for these cells from both adult and fetal hearts, but there is yet to be a consensus on what markers to use to define these cells [2]. In addition, different isolation methods yield distinct bind- ing profiles of stem cell markers, suggesting that they represent different progenitor populations or different stages of pluripotency. CPCs can potentially be used for cell-based cardiac repair, but in order to achieve feasibility and patient safety when applied for cell therapy purposes, there is a need to develop better tools to define a more homogenous population of resident CPCs. Therefore, the aim of this study was to generate a panel of novel monoclonal antibodies (mAbs) targeting surface epitopes on resident human heart-derived CPCs.

The generated antibodies were selected against heart-derived CPCs, which are isolated based on reactivity against mouse Sca-1 antibody, and can be found in both adult and fetal human hearts [3–5]. These Sca-1+ CPCs are multipotent, and able to generate all three cell types that make up the heart: cardiomyocytes, smooth muscle cells and endothelial cells. They are amenable to expansion in culture, making them an ideal stem cell source for cell-based cardiac repair. In addition, they have been shown to be able to engraft and differentiate in vivo when transplanted into the infarcted mouse heart and improve survival and long-term cardiac function [6]. The anti-mouse Sca-1 antibody has proven to be useful in identifying a homogeneous and robust population of heart-derived human CPCs. However, there is no Sca-1 homologue in the human genome [7]. To date it remains unclear which epitope on human CPCs is recognized by this antibody, thereby significantly hampering a clinical application. A new panel of mAbs raised against surface epitopes expressed by Sca-1+ CPCs will provide human specific alternatives to replace the anti-mouse Sca-1 antibody in the isolation of human CPCs.

Besides heart-derived CPCs, embryonic stem cell differentiation has been used

to model heart development, both in mice and human [8,9]. Key stages of human

heart development are known, and the transcription and signaling pathways in-

volved in its regulation are well-established from mouse models. However, conflict-

ing results often arise when looking at surface markers expressed on CPCs and com-

paring these markers between heart development and the differentiation process

of human embryonic stem cells (hESCs) towards cardiomyocytes. These discrepan-

cies are typically attributed to differences in maturity and developmental stages of

the progenitor cells [10]. Therefore, our panel of CPC-specific mAbs, can be a pow-

3

erful tool in addressing some of these fundamental questions regarding the similar- ities and differences between heart derived CPCs (heart-CPCs), and CPCs isolated from in vitro hESC differentiation (hESC-CPCs).

In this study, we describe the use of mAb C19, one of the antibodies from our panel of heart-CPC-specific mAbs, as a novel surface marker alternative for Sca-1 in the isolation of CPC populations. MAb C19 was found to recognize a glycosylated form of the glucose-regulated protein, 78 kDa (GRP78) found on the cell surface of CPCs. We characterized mAb C19 binding to Sca-1+ CPCs and performed heart tissue microarrays to show its ability to enrich for a multipotent cardiovascular CPC population from both differentiating hESC cultures and primary heart tissue.

3.2. M ETHODS

C ULTURE AND DIFFERENTIATION OF HEART - AND H ESC-CPC S

For human fetal tissue collection and atrial biopsies, individual permission was ob- tained using standard informed consent procedures and conforms to the Declara- tion of Helsinki. Prior approval of the ethics committee of the Leiden University Medical Centre was granted. Sca-1+ CPCs isolated from cardiac tissue were used for immunization and antibody screening (Supplementary methods). CPCs iso- lated from both heart tissue and hESC culture were maintained in culture as pre- viously described [3]. The differentiation of these CPCs was induced with 5 µM 5-azacytidine (Sigma) in the first 3 days of differentiation, followed by transforming growth factor beta-1 (TGF β1) (Peprotech) stimulation at 1 ng/ml.

H ESC CULTURE AND DIFFERENTIATION INTO CARDIOMYOCY TES

The HES-3 (ES Cell International) cell line was used in this study. hESCs were main- tained as co-cultures on inactivated mouse feeders as described previously [11].

The directed differentiation protocol follows that described by Lecina et al. [12].

Briefly, HES-3 co-cultures were mechanically dissociated with EZ-Passage to form

aggregates in ultra low attachment plates in serum-free medium supplemented

with 5 µM of the p38 MAPK inhibitor, SB203580 (Sigma), for the first 8 days. The

medium, including inhibitor, was refreshed every other day. For continued differ-

entiation after sorting, cells were harvested by incubation with TrypLE (Invitrogen)

at 37 ^◦ C for 5 min. The single cell suspension was re-aggregated by centrifugation at

100 ×g for 5 min in Aggrewell 800 plates (StemCell Technologies) to generate embry-

oid bodies (EBs) of 5000 cells in size. After 2 days, EBs were transferred to 24-well

plates at 80 EBs/well, and allowed to mature with medium change every two days.

3

E NRICHMENT OF M A B C19 AND ANTI -S CA -1 BINDING CELLS WITH MACS Magnetic bead separation based on the MACS technology (MiltenyiBiotec) was used to enrich for mAb C19+ cells from differentiating hESCs. Cells harvested on Day 7 of hESC differentiation were incubated at 4 ^◦ C with mAb C19 for 30 min, fol- lowed by anti-mouse FITC (1:500,DAKO) for 15 min, and finally with anti-FITC mi- crobeads (MiltenyiBiotec) for 15 min, before adding the cell suspension to the mag- netic column for cell separation. The flow-through (mAb C19-) and eluted (mAb C19+) fractions were collected for analysis. Heart-CPCs were isolated from heart tissue as previously described in [3].

G ENE EXPRESSION AND MICROARRAY STUDY

Total RNA was isolated using the RNeasy Mini Kit (Qiagen), and cDNA was gener- ated using the Maxima Reverse Transcription system (Thermo Scientific) accord- ing to the manufacturer’s protocols. Real-time PCR with SYBR green detection was performed using an ABI Prism 7500 Fast thermocycler (Applied Biosystems). Gene expression was normalized against GAPDH (unless otherwise stated) as the house- keeping gene and samples were run as triplicates. For microarray studies, total RNA was processed with the Affymetrix 3’ IVT Express Kit and the resultant labeled cRNA hybridized to GeneChip Human GenomeU133 Plus 2.0 Array. The GeneChip Com- mand Console Software was used for acquisition, normalization and analysis, and DAVID [13] was used for bioinformatics analysis and gene ontology classification.

A NGIOGENESIS ASSAY

In vitro tube formation assay for the sorted cells was conducted with the Angiogen- esis Kit (Millipore). Cells were cultured for 20h at 37 ^◦ C/5% CO2 in EGM-2 (Lonza) supplemented with 50 ng/ml vascular endothelial growth factor (VEGF). Phase con- trast images were taken and analyzed using the Angiogenesis Analyzer plugin for ImageJ (developed by Gilles Carpentier).

S TATISTICS

Unless otherwise stated, experiments were conducted minimally in triplicates. Un- paired Student’s t-test was used to calculate the p-value, of which p<0.05 is deemed to be biologically significant.

3.3. R ESULTS

3.3.1. C HARACTERIZATION OF M A B C19 AND ITS BINDING TO CPC S

A panel of monoclonal antibodies was raised against heart-derived Sca-1+CPCs

(Figure S1), of which mAb C19 was selected for further investigation. MAb C19 was

shown by flow cytometry to bind strongly (>90%) to the CPC line used for immu-

3 C

D

Figure 3.1: Characterization of mAb C19. (A) Flowcytometry analysis of mAb C19 binding against

heart-CPCs, hESCs and human lung fibroblast, IMR90, and human foreskin fibroblasts, hFF. (B) Im-

munostaining of Sca-1+ CPCs, under non-permeabilized (left), and permeabilized conditions (right),

with mAb C19 showing cell surface and a punctate intracellular binding pattern, respectively. (C)

Western blot of samples obtained by immunoprecipitation with mAb C19 on heart-CPC whole cell

lysates with biotin-tagged cell surface proteins. The antigen band was identified to be of ∼67 kDa

(boxed). Confirmatory test with a commercial antibody against GRP78 verified the identity of the

antigen for mAb C19. Data are representative from at least three independent experiments. (D) Co-

staining of mAb C19 with a commercial anti-GRP78 antibody demonstrated binding of both antibod-

ies to Sca-1+ CPCs. However, while mAb C19 showed a uniform staining on the cell surface, commer-

cial anti-GRP78 was intracellular localized, with stronger staining near the nucleus.

3

nization, and two other Sca-1+ CPC lines derived from separate patients. In com- parison, negligible binding (<5%) towards hESCs (HES-3) and 2 human fibroblast lines (IMR90 and hFF) was observed (Figure 1A and Figure S2A). Immunocytochem- istry staining on Sca-1+ CPC, hESC, and fibroblast lines verified the flow cytom- etry data. MAb C19 showed strong and even staining on the majority of Sca-1+

CPCs, with negligible binding to hESCs and hFFs (Figure S2B). In addition, mAb C19 bound uniformly to the surface of fixed Sca-1+ CPCs in culture, while perme- abilization of the cells resulted in a diffuse punctate staining pattern (Figure 1B).

Some cells showed stronger binding in regions closer to the nucleus, suggesting that the antigen of mAb C19 was present both on the cell surface and in the cytoplasm.

The antigen target for mAb C19 on Sca-1+ CPCs was further elucidated. West- ern blot and dot blot analysis of CPC lysates under denaturing conditions using SDS failed to identify any binding antigen targets, suggesting that antibody recog- nition is dependent on the native conformation of the antigen (Figure S3A). Perio- date treatment of the lysate abolished the binding between mAb C19 and its anti- gen, indicating that antigen-antibody recognition was dependent on glycosylation of the antigen (Figure S3B). In order to circumvent the antigen’s sensitivity to dena- turing SDS, proteins on the cell surface were conjugated with biotin and visualized indirectly with HRP-conjugated streptavidin. Immunoprecipitation with mAb C19 pulled-down a unique and strongly expressed band of ∼67 kDa. Mass spectrome- try analysis identified the band as 78 kDa glucose-regulated protein (GRP78), also known as bindingimmunoglobulin protein (BiP). A cross-probe with a commercial antibody recognizing GRP78 identified a band at a similar size after immunoprecip- itation with mAb C19, validating the antigen identity (Figure 1C).

While GRP78 and mAb C19 localized to the same cell, a difference in subcel- lular localization was observed, with mAb C19 and commercial anti-GRP78 being present on the cell surface and near the nucleus respectively (Figure 1D). As mAb C19 was raised against surface epitopes, we speculated that mAb C19 recognized only a subset of the GRP78 present in the cell, of which the intracellular form pre- dominated, accounting for the differences in staining pattern.

3.3.2. I SOL ATION AND CHARACTERIZATION OF M A B C19+ HEART -CPC S

To determine if mAb C19 could replace the anti-Sca-1 antibody for the isolation of

CPCs from human heart tissue, we used a similar method of sorting heart CPCs with

the newly identified mAb C19 and compared the mAb C19+ cells with the original

Sca-1+ CPC population. MACS with mAb C19 yielded more cells compared to iso-

lation using the anti-mouse Sca-1 antibody (Figure S4A top) which may reflect a

higher affinity of the human mAb C19 towards human cells. The mAb C19+ cells

had a morphology similar to Sca-1+ CPCs (Figure S4A bottom) and expressed early

cardiac markers (Figure 2A and Figure S4B). Markers for mature cardiomyocytes

3

mAb C19+ CPCs

m Ab C1 9 DA PI

Ep ica rd iu m Ca rd ia c V al ve Bl oo d Ve ss el s CP C- lik e ce lls

mAb C19 cTnI

mAb C19 mAb C19

PECAM-1 PECAM-1

mAb C19 mAb C19

aSMA mAb C19

PECAM-1

mAb C19 WT1 mAb C19

WT1 mAb C19

cTnI mAb C19

cTnI

MEF2C NKX2.5 0.0000

0.0002 0.0004 0.0006 0.0008

Relative expression to beta-actin

mAb C19+ CPCs Sca-1+ CPCs

ALCAM PDGFra GATA4 0.00

0.01 0.02 0.03

Relative expression to beta-actin mAb C19+ CPCs Sca-1+ CPCs

A B

C

Figure 3.2: Characterization of mAb C19+ heart-CPCs and localization of mAb C19 in human heart

tissue. (A) Relative gene expression of intracellular and surface cardiac progenitor markers in mAb

C19+ and Sca-1+ heart-CPCs isolated from fetal heart tissue (n=3). (B) Strong surface binding of mAb

C19 to mAb C19+CPCs isolated from a Week 15 heart. (C) Costaining of mAb C19 with markers for

myocardium (cTnI), epicardium (WT1), endothelium (PECAM-1) and smooth muscle cells (αSMA)

on sections of fetal heart tissue. MAb C19 was localized to the smooth muscle layer of large blood

vessels, cardiac valve and epicardium. In addition, mAb C19+ CPC-like cells were identified within

the myocardium.

3

like beta-myosin heavy chain ( βMHC) and cardiac troponinT (cTnT) were absent in both cell lines (data not shown). Furthermore, mAb C19+ heart-CPCs were ami- able to culture under conditions established for the Sca-1+ CPCs with comparable proliferation rate (not shown), and showed strong retention of mAb C19 binding for at least 16 passages (Figure 2B and Figure S4C).

Immunofluorescent staining on fetal heart tissue was performed using mAb C19 to determine the cell types being recognized (Figure 2C). MAb C19 co-localized with nuclear Wilm’s tumor 1 (WT1) in the epicardium which is the outer-most layer of the heart known to contain progenitor cells. Strong mAb C19 staining was also ob- served in blood vessels where it was present in alpha smooth muscle actin ( αSMA)+

cells, representing the smooth muscle layer, as well as in some platelet endothe- lialcell adhesion molecule 1 (PECAM-1)+ endothelial cells including the endothe- lial lining of the cardiac valve (Figure 2C). Interestingly, within the myocardial layer mAb C19 mainly localizes to αSMA+ cells but does recognize the PECAM-1+ lining of the cardiac valve suggesting a differences between valvular endothelial cells and endothelial cells in the remaining circulation (Figure S5). In addition, mAb C19 rec- ognizes only a part of the endothelial lining of the valve which may correspond to regional differences within the heart valve (Figure S5B). Finally, mAb 19 stains sin- gle cells within the myocardial layer of the heart which may represent a progenitor cell population (Figure 2C).

To determine whether the strong surface binding of mAb C19 is specific for CPCs, immunocytochemistry on various human (heart) cell types was performed (Figure S4D). The stainings revealed an intracellular localization of mAb C19 in cob- ble epicardial-derived cells, representing the inactive epicardium, endothelial cells, cardiac fibroblasts and smooth muscle cells. Altogether, this suggested that mem- brane localization of mAb C19 is specific for CPCs.

3.3.3. MA B C19 IN IDENTIFICATION AND ISOL ATION OF CARDIAC PROGENI -

TORS DURING H ESC DIFFERENTIATION

As established in the previous sections that mAb C19 recognized a surface anti- gen on human heart-CPCs, we speculated that it may also bind to a similar CPC- like population that arises transiently during the continuum of differentiation from hESCs to cardiomyocytes. HES-3 was differentiated into cardiomyocytes over a pe- riod of 20 days and the binding of mAb C19 was tracked by flow cytometry. A distinct mAb C19+ sub-population of cells was found to arise from Day 4 of differentiation onwards. This sub-population increased in proportion until it peaked on Day 7, be- fore gradually dropping in numbers as differentiation proceeded (Figure 3A). The peak in mAb C19 binding was found to occur just before the appearance of MF20+

and cTnT+ cells, and the onset of beating clusters, both indicators of functional car-

diomyocytes. In addition, mAb C19 staining showed a strong overlap with intracel-

3

A B

C

Figure 3.3: MAb C19 identifies cardiac progenitors during hESC differentiation. (A) Representative plot tracking the binding of mAb C19 based on flow cytometry, along with cardiac markers MF20 and cTnT during the differentiation of hESC to cardiomyocytes. MAb C19 was observed to have a temporal binding profile. The percentage of mAb C19+ cells increased during the initial phase of differentiation, and reached peak binding one day before the onset of beating clusters and expression of mature cardiac markers, before dropping in levels as differentiation progressed. This trend was observed in all three independent differentiation runs. (B) Flow cytometry with mAb C19 and cardiac progenitor marker Mef2c on Day 6 of differentiation, showing almost half of mAb C19+ cells being MEF2C+. (C) Co-stain of mAb C19 with cardiac transcription factors, MEF2C, GATA4 and NKX2.5, demonstrating co-expression of these markers in emerging cardiac clusters in Day 7 replated EBs.

Cells surrounding these cardiac clusters did not stain for mAb C19.

3

lular cardiac markers like MEF2C, GATA4 and NKX2.5 during the early stages (Days 5 to 7) of hESC differentiation when analyzed by flow cytometry and immunofluo- rescence (Figure 3C, Figure S6A and B). Whereas, on Day 9, when cells have matured past the cardiac progenitor stage, a sub-population of residual mAb C19+ cells was found in addition to the co-stained cells (Figure S6C). These results suggested that mAb C19 recognized a CPC population during hESC differentiation. To determine the lineage specificity of mAb C19 in identifying cardiac progenitors, its binding during differentiation of hESCs to neural progenitors was also determined, but no activity could be found (Figure S7A).

Co-stainings with other previously reported cardiac progenitor surface mark- ers, Sca-1 [3], KDR [10], and C-Kit [14], were also conducted 6 days after the onset of hESC differentiation (Figure S7B). As expected, <5% of the population stained for Sca-1 or C-Kit as these two markers are more commonly used in isolations from heart tissue. In contrast, KDR has been reported to identify cardiovascular progen- itor cells from hESCs undergoing cardiac differentiation. Of the 9% that was found to be KDR+ on Day 6 of hESC differentiation, half also bound mAb C19, suggest- ing that mAb C19 has utility in identifying a subset of the previously reported KDR+

hESC-CPCs [10].

Enrichment for hESC-CPCs on Day 7 (day of peak binding) using mAb C19, yielded a relatively high cell purity in both fractions and cells with a morphology similar to Sca-1+ CPCs, derived from the heart (Figure S8A and B). When compared with Sca-1+ heart-CPCs, expression of early cardiomyocyte markers (MESP1, GATA4 and NKX2.5) for these mAb C19+ hESC-CPCs was similar. However, expression of late cardiomyocyte markers (βMHC, C.ACTININ and cTnT) was significantly higher as the cells are undergoing differentiation down the cardiac lineage when sorting was conducted (Figure S8C).

Both positive and negative mAb C19 hESC fractions were successfully cultured under heart-CPC growth conditions. The mAb C19+ hESC CPCs had a high growth rate, with a doubling time of (31.8 ± 1.7) hours, allowing for 1:6 split ratio every 3 days, and could be maintained in culture for at least 15 passages while retaining high mAb C19 binding (Figure S9A and B) and cardiac gene expression (data not shown). In contrast, mAb C19- cells undergo senescence around passage 7, which may indicate that these cells have passed the progenitor cell stage at the time of isolation.

Altogether, mAb C19 was able to isolate a cell population during hESC differen-

tiation with CPC nature. These hESC-CPCs could be maintained in culture while

retaining their CPC characteristics such as high proliferative rates and expression of

cardiac markers.

3

3.3.4. A BILITY OF M A B C19+ HEART - AND H ESC-CPC S TO FORM CAR -

DIOMYOCY TES

Having isolated CPCs from both hESC differentiation and heart tissue using mAb C19, the fate of these cells was investigated. Heart-derived Sca-1+ CPCs are in- duced to differentiate into cardiomyocytes when stimulated with 5-azacytidine and TGF β1 [3]. Therefore, the isolated mAb C19+ CPCs were cultured and differentiated using this protocol to investigate if they responded to the same cues as the Sca-1+

CPCs.

MAb C19+ heart-CPCs were found to be able to differentiate into cardiomy- ocytes by stimulation with 5-azacytidine and TGF β1. After 3 weeks of differentia- tion, up-regulation of Troponin I (cTnI) as well as sarcomeric organizations were observed (Figure 4A). Similarly, mAb C19+ hESC-CPCs were able to differentiate into cardiomyocytes as demonstrated by the increased expression of the late car- diac markers βMHC and cTnT after 21 days of differentiation and the appearance of defined sarcomeres (Figure 4B).

In addition to stimulation with 5-azacytidine and TGFβ1, the continued dif- ferentiation of hESCs as EBs was simulated via forced reaggregation in Aggrewells.

While mAb C19 reactivity decreased for the unsorted control and C19- population, similar to the usual hESC differentiation protocol, mAb C19 binding remained high both 5 days and 10 days post-aggregation. In addition, when the extent of car- diomyocyte differentiation was quantified 5 days after MACS by flow cytometry, the proportion of MF20+ and cTnT+ cells was significantly higher for mAb C19+ hESC- CPCs, and this difference was further extended 10 days post-sort (Figure 4C). Beat- ing aggregates were only observed in the mAb C19+ population (data not shown).

Therefore, the enrichment for mAb C19+ hESC-CPCs was shown to improve car- diomyocyte differentiation (compared to a starting population of mixed cells, con- taining only about 20% mAb C19+ cells in the unsorted aggregates), and these mAb C19+ hESC-CPCs retained the ability to mature into cardiomyocytes after isolation in the absence of supporting mAb C19- cells.

Hence, mAb C19+ CPCs were responsive to both the TGF β1 stimulation, like Sca-1 + CPCs, and also to EB formation, like hESC cultures; and were capable to further mature and form cardiomyocytes after isolation and in vitro culture.

3.3.5. G ENE EXPRESSION PROFILING OF M A B C19+ H ESC- AND HEART - CPC S

In order to gain a better understanding of the mAb C19+ hESC and heart-CPC identity, microarray profiling was conducted. Biological triplicates of mAb C19+

and mAb C19- fractions sorted 7 days after differentiation were compared using

Affymetrix U133 human genome chips. Genes that showed more than a 4-fold dif-

ferential regulation were subjected to analysis by DAVID to categorize them based

3

A

B

C

Figure 3.4: Differentiation of mAb C19+ CPCs. (A) MAb C19+ heart-CPCs were differentiated into

mature cardiomyocytes using the same differentiation protocol used for Sca-1+ heart-CPCs, showing

cTnI (with faint sarcomeric patterning, white arrow and insert) after 21 days of differentiation. Like

the Sca-1+ CPCs, mAb C19+ heart-CPCs showed a similar increase in the expression of mature cardiac

markers 3 weeks after the initiation of differentiation (n=3). (B) MAb C19+ hESC-CPCs were cultured

for 7 passages before undergoing the same differentiation protocol used for heart-CPCs. Represen-

tative image of mAb C19+ hESC-CPCs on Day 21 of differentiation, showing cardiac markers cTnT

and NKX2.5 and the emergence of sarcomeric patterning of troponin (white arrows and insert). Gene

expression of cardiac markers was analyzed by qRT-PCR, with down-regulation of early progenitor

markers, and up-regulation of cardiomyocyte markers after 3 weeks (n=3). (C) Cell fractions collected

from MACS with mAb C19 on Day 7 of hESC differentiation were re-aggregated in Aggrewells to re-

form EBs for continuation of the differentiation protocol. Binding of mAb C19, and sarcomeric mark-

ers, MF20 and cTnT, by flow cytometry was analyzed 5 days and 10 days after MACS, and found to be

significantly up-regulated in mAb C19+ populations compared to mAb C19- cells (n=3 for unsorted

control and mAb C19- samples; n=2 for mAb C19+ samples) ( ^∗ : p<0.05; ^∗∗ : p<0.01; ^∗∗∗ : p<0.001).

3

on their gene-ontology biological function. Figure 5A summarizes the microarray results obtained for mAb C19+ hESC-CPCs, and the top ten categories based on the p-value are listed. Unexpectedly, from the list of genes with lower expression, a high frequency of genes involved in heart and muscle development was observed. This might be attributed to the possibility of early cardiomyocytes that have passed the progenitor stage and lost mAb C19 expression, and therefore found in the nega- tive fraction, skewing the expression of mature cardiac genes in that fraction. Ad- ditionally, highest amongst the up-regulated genes were genes involved in epithe- lium development, such as annexin A1, epithelial membrane protein 1, and uro- plakin 1B (Figure S10A), in addition to many other genes involved in transcription and metabolic regulation.

Similarly, gene expression using microarray was conducted on CPCs from Week 14 fetal human heart tissue that had been sorted using Sca-1 or mAb C19, and the same functional analysis was done for genes that showed a more than 5-fold differ- ence in expression (Figure 5B). Sca-1+ CPCs scored higher in genes related to heart and muscle development compared to the mAb C19+ cells. Conversely, mAb C19+

heart-CPCs are more related to endothelial and vascular cells, such as angiopoietin 2 and epiregulin (Figure S10B).

Figure 3.5: Microarray analysis of mAb C19 hESC- and heart-CPCs. (A) Large scale gene expression

profiling was done on mAb C19+ (n=3) and mAb C19- (n=3) cells sorted on Day 7 of hESC differenti-

ation using microarray analysis. The list of 5-fold down- and up-regulated genes were uploaded into

DAVID for Gene Ontology categorization based on their functional annotation. The 12 categories with

the highest significance based on the p-value are listed in the table. (B) A similar microarray analysis

was conducted on mAb C19+ heart-CPCs and normalized to Sca-1+ CPCs, both sorted from Week 14

fetal human heart tissue (Due to limitations in obtaining human heart tissue samples, this microarray

was based on one sample).

3

To validate the results obtained from our microarray gene expression analysis, qRT-PCR was done. Five genes from each of the top gene-ontology categories were selected for validation for both the up- and down-regulated genes in the hESC and heart group (Figure 6A). The qRT-PCR confirmed the findings from the microarray experiment, and the gene expression profiles of the selected genes follow similar trends, with the exception of phospholamban (PLN), which was found slightly up- regulated in the qRT-PCR.

At the same time, the gene expression profiles of the hESC-sorted fractions were compared to that of the undifferentiated HES-3 line [15] (Figure S11A). While both hESC-sorted populations were clustered separately from the HES-3 samples, mAb C19+ hESC-CPCs were clustered relatively closer to the parental HES-3 than the mAb C19- population were, suggesting that mAb C19+ hESC-CPC do have more

“stemlike” characteristics than their mAb C19- counterparts.

3.3.6. V ALIDATION OF ANGIOGENIC POTENTIAL

Concluded from the gene profile, mAb C19+ CPCs have an endothelial profile.

Therefore, a cell-based angiogenesis assay was done to supplement the results ob- tained from the qRT-PCR experiments for both hESC- and heart-CPCs. Cells were seeded onto Matrigel and stimulated with VEGF, which would result in the sponta- neous formation of small capillaries made of a single cell layer within 24 h (Figure 6B). Compared to the mAb C19- fraction, mAb C19+ hESC-CPCs formed a more extensive mesh-like network, while mAbC19- cells congregated to one spot and formed extensions of tubule structures outwards. The mAb C19+ hESC-CPCs were found to be more angiogenic than both the unsorted and mAb C19- fractions, based on the significantly higher mesh area and tubule lengths of the capillary network that was formed (Figure 6C). A similar observation was done for mAb C19+ heart- CPCs which formed a significantly bigger network of αSMA+/Vimentin+ tubes com- pared to Sca-1+ CPCs (Figure 7). These results showed that mAb C19+ hESC- and heart-CPCs were not only able to differentiate into cardiomyocytes but, under ap- propriate stimulation, also into the two other cell types in the cardiac lineage (en- dothelial and smooth muscle cells) that form the vascular system.

3.4. D ISCUSSION

In this study we characterized a novel mAb, mAb C19, that targeted a surface marker

useful for identifying and isolating CPCs from both human heart tissue and hESC

cardiac differentiation. We have shown that mAb C19, which was one of the novel

mAbs generated against Sca-1+ CPCs from heart tissue, was able to isolate a mul-

tipotent cell population with CPC characteristics. These CPCs were able to gen-

erate cardiomyocytes, either when allowed to continue differentiation in EBs, or

when stimulated with 5-azacytidine and TGF β1. The inherent predisposition of

3

A

B

C

Figure 3.6: Validation of microarray findings. (A) qRT-PCR was used to confirm the results obtained

from the microarrays. 5 genes were selected from each of the top GO categories gene lists for valida-

tion. For mAb C19+ hESC-CPCs and C19+ heart-CPCs, gene expression was normalized to mAb C19-

hESC-CPCs and the Sca-1+ CPCs respectively (n=3). (B) Representative phase contrast images from

the in vitro angiogenesis assay using mAb C19+ hESC-CPCs (after 4 passages in CPC growth condi-

tions). Segments, mesh areas and junctions were identified using the Angiogenesis Analyzer plug-in

in ImageJ, and overlaid onto the phase contrast image. (C) Statistics from the Angiogenesis Analyzer

are plotted. The mAb C19+ hESC-CPCs were found to have significantly higher angiogenic proper-

ties, based on the mean mesh area and total segment length of the capillary network formed (n=6; ^∗ :

p<0.05; ^∗∗ : p<0.01).

3

αSMA Vimentin

Sc a- 1+ C PC s m Ab C 19 + C PC s

Sca-1+

CPCs mAb C19+

CPCs 0

50 100 150 200

Number of tubes

* Number of tubes Sca-1+

CPCs mAb C19+

CPCs 0

5×10³ 1×10⁴ 1.5×10⁴ 2×10⁴ 2.5×10⁴ 3×10⁴ 3.5×10⁴

Length of tubes

Total length of tubes (µm) *

A B

Figure 3.7: Angiogenesis assay using mAb C19+ heart-CPCs. (A) Phase contrast images showing a tubular network for both Sca-1+ and mAb C19+ heart-CPCs which were positive for αSMA and Vi- mentin. (B) The mAb C19+ heart-CPCs were found to have significantly higher angiogenic properties, based on the total length and number of tubes ( ^∗ : p<0.05).

mAb C19+ cells, however, was towards the endothelial cell fate, as observed from the gene expression microarray and angiogenesis assay. This was also the first study in which CPCs, isolated during hESC differentiation and heart development using the same surface marker, were used to better understand the cell phenotype associated with each population.

For a long time, the heart was considered to be a terminally differentiated organ without a stem cell population for self-repair after injury. However, there is increas- ing evidence that cardiac stem or progenitor cells exist both in the adult heart, and during development. One of the earliest identified markers of cardiac progenitors is the Sca-1 antigen that was used by Oh et al. for isolating heart-CPCs from the adult mouse myocardium [16]. These cells were capable of homing to the myocardium and differentiated into cardiomyocytes after successful engraftment. More recently, although based on cross-reactivity with an unknown epitope, the anti-mouse Sca- 1 antibody has shown its utility in the isolation of CPCs from human hearts, both adult and fetal [4,6]. We have generated novel antibodies against this population of human Sca-1+ CPCs, of which mAb C19 was validated in this study.

Immunoprecipitation of heart-CPC lysates with mAb C19 yielded a band of around 67 kDa. Mass spectrometry analysis of this band identifiedit as GRP78.

GRP78 is one of the members of the heat-shock protein-70 family, and is involved in

correct protein folding and stress response when it is localized to the endoplasmic

3

reticulum. It also modulates important immunological activities when it is found on the cell surface or released into the extracellular space [17,18]. More impor- tantly, recent studies have shown that GRP78 features during heart development in mice, and its expression is up-regulated during early cardiogenesis [19,20]. They showed that its expression is regulated bythe cooperative binding of one of the car- diac transcription factors, GATA4 and Yin Yang 1 (YY1), to the promoter region of GRP78. In addition to its role in early heart organogenesis, GRP78 levels are up- regulated following hypoxia or hypoglycemic stress, and its expression may confer protective effects to surrounding cardiomyocytes [19,21]. This makes GRP78 an in- teresting candidate as a novel marker for heart-CPCs, both in fetal and adult hearts, and a potential role of GRP78 in CPC-biology requires further investigation.

Additionally, unlike commercial anti-GRP78 antibodies, mAb C19 was shown to recognize a unique glycosylated, surface-bound form of the protein. Glycans are gaining increasing prominence as biomarkers in the fields of cancer and stem cell biology [22,23]. Stage-specific embryonic antigens 3 and 4 (SSEA-3 and SSEA-4) are key glycolipid markers for embryonic stem cells, and changes in glycan pro- files have been used to track their differentiation [24]. In the cardiac field, B-type natriuretic peptide (BNP), a candidate biomarker for heart failure, was shown to undergo extensive O-glycosylation to stabilize the pro-BNP structure and affect the pro-peptide processing [25]. As such, the glycan motif on the antigen of mAb C19 may confer an additional level of specificity towards CPCs.

There are few intracellular markers, and no surface makers, used in the isolation

of CPCs from either the pre- or post-natal heart that have been concurrently used

in embryonic stem cell differentiation models. Furthermore, there are no studies

to date that, using the same method, compare the two types of CPCs isolated from

human cardiac tissue and hESC sources. Both the hESC- and heart-CPCs isolated

with mAb C19 were shown to be multipotent cardiac progenitors with the ability

to generate all three cell types in the heart. They also showed expression of early

cardiac markers like NKX2.5, GATA4, ISL-1 and MEF2C, indicating that they are in-

deed committed to the cardiac lineage. Leveraging on this, we used mAb C19 as a

common marker in both systems to compare and contrast the CPCs isolated. How-

ever, one of the limitations inherent in this study is the difference in the stage of de-

velopment and level of cardiac commitment of each CPC population. Co-staining

studies on differentiating hESC EBs showed that, while the correlation of expres-

sion of these cardiac transcription factors and mAb C19 was strong during the early

cardiac developmental phase, this association was weakened after the CPCs pro-

gressed past the progenitor state and attained functionality and sarcomeric features

of beating cardiomyocytes. The hESC-CPCs, being isolated on Day 7 of hESC differ-

entiation, will likely be much more “immature”, with some early cardiac transcrip-

tion factors like NKX2.5 and ISL-1 only just starting to get expressed. Conversely, the

3

heart-CPCs are already fully committed progenitors found in a Week 14 fetal heart.

We were fully aware of these differences when doing the comparison between the two populations of CPCs.

The microarray analysis showed that in terms of gene expression profiles, the hESC group (mAb C19+ and mAb C19-) clustered separately from the heart group (mAb C19+ and Sca-1+) (Figure S11B), implying that the differences resulting from the relative maturity and tissue source have a greater impact than the similari- ties conferred by the common marker. Despite such differences, one key similar- ity between the mAb C19+ hESC-CPC and heart-CPCs was their upregulation of endothelial/epithelial-related genes. The in vitro tube formation assay also indi- cated that mAb C19+ CPCs are more angiogenic than their mAb C19- counterparts, and we see this as an advantage. Many stem cell transplantation trials using either non-resident stem cells like skeletal myoblasts and mesenchymal stem cells, or resi- dent cardiac stem cells like cardiospheres, and C-Kit+ CPCs, reported that paracrine effects played a more important part in the improvements in heart function than any increase in cardiomyocyte numbers [26–29]. The parental Sca-1+ CPCs were also mediating repair largely by paracrine effects and improving vascular density in the initial 2 weeks of transplantation [30], and engraftment and differentiation into cardiomyocytes were only observed in a longer-term 12-week study [6]. In addi- tion, in a study conducting a direct comparison between various stem cell sources, often used in heart transplantation studies, a positive correlation was found be- tween the population’s angiogenic abilities and secretion of angiogenic growth fac- tors like VEGF and angiopoietin, and the eventual improvements in cardiac func- tion and suppression of ventricular remodeling [31]. As such, it can be implied that our highly angiogenic mAb C19+ hESC- and heart-CPCs may be a good candidate for stem cell transplantation purposes.

In summary, we have raised a panel of CPC-specific antibodies. The ability to use antibodies targeting surface antigens, such as mAb C19, facilitates the charac- terization and, more importantly, the isolation of cardiac progenitor populations from both hESC differentiation models and human heart tissues. This large panel of CPC specific antibodies will allow the use of multiple markers to obtain more homogenous sub-fractions of progenitor cell populations that as a result are better characterized. This is a vital step for the application of these cardiac stem cells in tissue engineering and heart developmental studies.

Sources of funding: This research is supported by funding from the Agency of Sci-

ence, Technology and Research (A ^∗ STAR) and the Netherlands Institute for Regen-

erative Medicine (grant no. FES0908).

3

Acknowledgments: We thank Dr. Dave Ow for the help with microarray analysis.

Special thanks to Dr. Marti Lecini, Dr. Allen Chen and Mr. Sherwin Ting for advice with the directed hESC differentiation protocol. We also thank Kirsten Lodder for technical assistance, and Dr. Filip Laco and Dr. Anke Smits for critically reviewing this manuscript and valuable inputs.

3.5. R EFERENCES

1. Reinecke H, Minami E, Zhu W-Z, Laflamme MA (2008) Cardiogenic differentiation and transdiffer- entiation of progenitor cells. Circ Res 103:1058–71.

2. Bollini S, Smart N, Riley PR (2011) Resident cardiac progenitor cells: at the heart of regeneration. J Mol Cell Cardiol 50:296–303.

3. Smits AM, van Vliet P, Metz CH, Korfage T, Sluijter JP, Doevendans PA, et al. (2009) Human car- diomyocyte progenitor cells differentiate into functional mature cardiomyocytes: an in vitro model for studying human cardiac physiology and pathophysiology. Nat Protoc 4:232–43.

4. Goumans MJ, de Boer TP, Smits AM, van Laake LW, van Vliet P, Metz CHG, et al. (2007) TGFbeta1 induces efficient differentiation of human cardiomyocyte progenitor cells into functional cardiomy- ocytes in vitro. Stem Cell Res 1:138–49.

5. De Boer TP, van Veen TAB, Jonsson MKB, Kok BGJM, Metz CHG, Sluijter JPG, et al. (2010) Hu- man cardiomyocyte progenitor cell-derived cardiomyocytes display a maturated electrical pheno- type. JMol Cell Cardiol 48:254–60.

6. Smits AM, van Laake LW, den Ouden K, Schreurs C, Szuhai K, van Echteld CJ, et al. (2009) Human cardiomyocyte progenitor cell transplantation preserves long-term function of the infarcted mouse myocardium. Cardiovasc Res 83:527–35.

7. Holmes C, Stanford WL (2007) Concise review: stem cell antigen-1: expression, function, and enigma. Stem Cells 25:1339–47.

8. Keller G (2005) Embryonic stem cell differentiation: emergence of a new era in biology and medicine. Genes Dev 19:1129–55.

9. Huber TL (2010) Dissecting hematopoietic differentiation using the embryonic stem cell differenti- ation model. Int J Dev Biol 54:991–1002.

10. Yang L, Soonpaa MH, Adler ED, Roepke TK, Kattman SJ, et al. (2008) Human cardiovascular pro- genitor cells develop from a KDR+ embryonic-stem-cell-derived population. Nature 453:524–8.

11. Choo A, Padmanabhan J, Chin A, Fong WJ, Oh SKW (2006) Immortalized feeders for the scale-up of human embryonic stem cells in feeder and feeder-free conditions. J Biotechnol 122:130–41.

12. Lecina M, Ting S, Choo A, et al. (2010) Scalable platform for human embryonic stem cell differen- tiation to cardiomyocytes in suspended microcarrier cultures. Tissue Eng Part C Methods 16:1609–19.

13. Huang DW, Sherman BT, Lempicki RA (2008) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4:44–57.

14. Bearzi C, Rota M, Hosoda T, Tillmanns J, Nascimbene A, De Angelis A, et al. (2007) Human cardiac stem cells. Proc Natl Acad Sci 104:14068–73.

15. Munoz J, Low TY, Kok YJ, Chin A, Frese CK, Ding V, et al. (2011) The quantitative proteomes of human-induced pluripotent stem cells and embryonic stem cells. Mol Syst Biol 7:550.

16. Oh H, Bradfute SB, Gallardo TD, Nakamura T, et al. (2003) Cardiac progenitor cells from adult my- ocardium: homing, differentiation, and fusion after infarction. Proc Natl Acad Sci U S A 100:12313–8.

17. Panayi GS, Corrigall VM, Henderson B (2004) Stress cytokines: pivotal proteins in immune regula- tory networks: opinion. Curr Opin Immunol 16:531–4.

18. Quinones QJ, de Ridder GG, Pizzo SV (2008) GRP78: a chaperone with diverse roles beyond the

endoplasmic reticulum. Histol Histopathol 23:1409–16.

3

19. Barnes JA, Smoak IW (2000) Glucose-regulated protein 78 (GRP78) is elevated in embryonic mouse heart and induced following hypoglycemic stress. Anat Embryol 202:67–74.

20. Mao C, Tai W-C, Bai Y, Poizat C, Lee AS (2006) In vivo regulation of Grp78/BiP transcription in the embryonic heart: role of the endoplasmic reticulum stress response element and GATA-4. J Biol Chem 281:8877–87.

21. Hardy B, Raiter A (2010) Peptide-binding heat shock protein GRP78 protects cardiomyocytes from hypoxia-induced apoptosis. J Mol Med 88:1157–67.

22. Adamczyk B, Tharmalingam T, Rudd PM (2012) Glycans as cancer biomarkers. Biochim Biophys Acta Gen Subj 1820:1347–53.

23. Lanctot PM, Gage FH, Varki AP (2007) The glycans of stem cells. Curr Opin Chem Biol 11:373–80.

24. Nairn AV, Aoki K, dela Rosa M, Porterfield M, Lim J-M, Kulik M, et al. (2012) Regulation of glycan structures in murine embryonic stem cells: combined transcript profiling of glycan-related genes and glycan structural analysis. J Biol Chem 287: 37835–56.

25. Peng J, Jiang J, WangW, Qi X, Sun X-L, Wu Q (2011) Glycosylation and processing of pro-B-type na- triuretic peptide in cardiomyocytes. Biochem Biophys Res Commun 411:593–8.

26. Menasché P, Alfieri O, Janssens S, McKenna W, Reichenspurner H, Trinquart L, et al. (2008) The myoblast autologous grafting in ischemic cardiomyopathy (MAGIC) trial first randomized placebo- controlled study of myoblast transplantation. Circulation 117:1189–200.

27. Psaltis PJ, Zannettino ACW, Worthley SG, Gronthos S (2008) Concise review: mesenchymal stro- mal cells: potential for cardiovascular repair. Stem Cells 26:2201–10.

28. Tang XL, Rokosh G, Sanganalmath SK, Yuan F, Sato H, Mu J, et al. (2010) Intracoronary admin- istration of cardiac progenitor cells alleviates left ventricular dysfunction in rats with a 30-day-old infarction. Circulation 121:293–305.

29. Yoon CH, Koyanagi M, Iekushi K, Seeger F, Urbich C, Zeiher AM, et al. (2010) Mechanism of im- proved cardiac function after bone marrow mononuclear cell therapy role of cardiovascular lineage commitment. Circulation 121:2001–11.

30. Den Haan MC, Grauss RW, Smits AM, Winter EM, van Tuyn J, Pijnappels DA, et al. (2012) Car- diomyogenic differentiation-independent improvement of cardiac function by human cardiomyocyte progenitor cell injection in ischaemic mouse hearts. J Cell Mol Med 16:1508–21.

31. Li TS, Cheng K, Malliaras K, Smith RR, Zhang Y, Sun B, et al. (2012) Direct comparison of different

stem cell types and subpopulations reveals superior paracrine potency and myocardial repair efficacy

with cardiosphere-derived cells. J Am Coll Cardiol 2012;59:942–53.

3

3.6. S UPPLEMENTARY FIGURES

Figure S1: CPC-specific mAb panel. Binding to the respective cell lines were scored by flow cytometry, and based on averaged values of triplicates. +++: 70–100%; ++: 40–70%; +: 15–40%; -/+: 7–15%; -: <7%;

N.D.: not done. Isotypes were identified using a mouse isotyping kit from Roche – IsoStrip.

3

A

B

Figure S2: Screening of mAb C19 binding. (A) Binding of mAb C19 to several Sca-1+ CPC lines by flow cytometry and immunocytochemisty. (B) Immunostainings conducted under the same staining and exposure conditions, showing relatively strong and homogeneous staining for Sca-1+ CPCs (L9), and negligible staining for HES-3 and hFF cell lines.

A B

Figure S3: Antigen characterization. (A) Western blot of Sca-1+ CPC whole cell lysate, showing non-

binding of mAb C19 under both reducing (R) and non-reducing (NR) conditions ran under denaturing

SDS-PAGE. (B) Dot blot of CPC whole cell lysate was sensitive to periodate treatment when probed

with mAb C19, but binding of anti-actin antibody to the same lysate was not affected.

3

A B

C

D

Figure S4: MAb C19+ progenitors from human heart tissue. (A) Cells were isolated from a fetal

human heart, using Sca-1 antibody and mAb C19, showing more cells isolated for mAb C19 and a

similar cell morphology between both cell populations at passage 7. (B) Nuclear localization of the

cardiac transcription factors NKX2.5 and GATA4 in both CPC populations. (C) Binding of mAb C19 to

mAb C19+ heart-CPCs, showing retention of the marker for at least 16 passages. (D) Representative

immunofluorescence stainings for mAb C19, and corresponding cell markers, on cardiac fibroblasts,

spindle and cobble epicardial-derived cells and endothelial cells (HUVECS), showing absent or intra-

cellular staining for mAb C19, as opposed to surface binding on CPC-like cells (Scale bar: 100 µm).

3

A

mAb C19αSMA mAb C19PECAM1

B

*

Figure S5: MAb C19 localization in human heart tissue. Immunofluorescence staining on week 15 human heart tissue showed that mAb C19 recognized (A) the smooth muscle layer of the vasculature and (B) part of the endothelial lining of the cardiac valve. Asteriks indicate the lining of the cardiac valve that is negative for mAb C19 (Scale bar: 100 µm).

A

B

C

Figure S6: MAb C19 co-staining with cardiac markers during hESC differentiation. (A–C) Co-

staining on Day 5, 7 and 9-replated EBs respectively, showing strong correlation of mAb C19 with

intracellular cardiac progenitor markers, MEF2C, GATA4 and NKX2.5 in emerging beating clusters.

3

A

B

Figure S7: MAb C19 binding during neural differentiation and co-staining with reported CPC sur-

face markers. (A) Absence of mAb C19 binding when tracked during the early stages of directed dif-

ferentiation of HES-3 towards the neural lineage, despite the high differentiation efficiencies achieved

as indicated by the strong staining of NCAM, a neural progenitor cell marker. (B) Co-staining of mAb

C19 with other reported multi-potent CPC surface markers, Sca-1, C-Kit and KDR, on Day 6 of differ-

entiation. A sub-population of cells stained double positive, especially for KDR and mAb C19.

3

A

B

C

Figure S8: Isolation of mAb C19+ hESC-CPCs. (A) Flow cytometry analysis of mAb C19 binding on day 7 of hESC-CM cultures, for the unsorted (left), mAb C19 negative (centre), and positive (right) populations after MACS, showing an enrichment (from 23.8% to 76.2%) after sorting. (B) Cell mor- phology of CPC lines, compared to hESC-CPCs sorted using mAb C19, cultured under CPC growth conditions for 2 passages. (C) Gene expression of various cardiac markers for Sca-1+ heart-CPCs and mAb C19+ hESC-CPCs, normalized against expression levels of adult (L7) Sca-1+ heart-CPCs (n=3,

∗∗∗ : p<0.001)

3 A B

Figure S9: Culture and differentiation of mAb C19+ hESC-CPCs. (A) Growth curve and doubling

times of mAb C19- and mAb C19+ hESC-CPCs taken at P4 (n=2). (B) Flow cytometry analysis of mAb

C19 binding to mAb C19+ hESC-CPCs at P1 and P4 after isolation, showing increase and stabilization

of mAb C19 binding with increasing passage. MAb C19 binding remained above 80% when tested at

P13. P: passage

3

Figure S10: Gene list of differentially regulated GO categories. (A and B) Gene list of differentially

regulated genes in the top GO categories, with the respective fold-expression of each gene.

3 ^A

B

Figure S11: Microarray Cluster Analysis. (A) Cluster analysis of microarray data from the undiffer-

entiated HES-3 line, mAb C19+ and mAb C19- hESC-CPC, showing clustering of the hESC-CPCs away

from the parental HES-3 gene expression profile. In addition, mAb C19+ hESC-CPCs share more sim-

ilarities to HES-3 than the mAb C19- population. (B) Cluster analysis of microarray data from heart-

CPCs and hESC-CPCs, with the heart- and hESC- group clustering separately, indicating that the cell

source, instead of mAb C19, played a more important role in determining the gene expression profile.

3

A. A PPENDIX : C OMMERCIAL GRP78 ANTIBODY CANNOT RE -

PL ACE M A B C19 IN THE ISOL ATION OF HUMAN CARDIAC PRO -

GENITOR CELLS

Moerkamp A.T. ¹ , Leung H.W. ² , Lodder K. ¹ , Smits A.M. ¹ , Choo A. ^2,3 , Goumans M.J. ¹

1. Department of Molecular Cell Biology, Leiden University Medical Centre, Leiden, The Netherlands.

2. Bioprocessing Technology Institute, Agency for Science, Technology and Re- search, Singapore.

3. Department of Bioengineering, National University of Singapore, Singapore.

GRP78 ^N20 ISOL ATES A CELL POPUL ATION FROM HUMAN HEART TISSUE WITH LIM -

ITED SELF - RENEWAL AND EXPANSION ABILITY

We questioned whether the glycosylated form of GRP78, recognized by mAb C19, is specific for human CPCs. Therefore, we made use of a commercial available anti-GRP78 antibody (N20 from Santa Cruz; GRP78 ^N20 ). GRP78 ^N20 binds to the N- terminus of human GRP78 reported to be exposed at the cell surface [15], and can interfere with binding of its ligand Cripto thereby preventing downstream signalling to occur [5]. GRP78 is expressed by both Sca-1+ and mAb C19+ CPCs isolated from human fetal heart tissue (Figure 1A). Furthermore, these human CPCs express the epitope for GRP78 ^N20 (Figure 1B).

We next isolated cells from human heart tissue using GRP78 ^N20 and compared this population with Sca-1+ and mAb C19+ CPCs. Although GRP78 ^N20 was able to isolate a cell population from human heart tissue with no obvious differences in initial cell numbers between the GRP78 ^N20 + and Sca-1+ fraction (data not shown), GRP78 ^N20 + cells (Figure 1C) had limited self-renewal and expansion ability and adopted a fibroblast-like morphology between passage 10 and 15 (Figure 1D). Inter- estingly, performing immunofluorescent stainings for mAb C19 and GRP78 ^N20 on Sca-1+ CPCs showed a difference in subcellular localization. MAb C19 was mainly present at the cell surface, while GRP78 ^N20 was predominantly intracellular (Fig- ure 1E) suggesting that both antibodies recognize a subfraction of total GRP78 in human Sca-1+ CPCs.

GRP78 ^N20 AND M A B C19 ONLY PARTLY COLOCALIZE IN HUMAN HEART TISSUE

In the human heart, mAb C19 localizes to the smooth muscle layer of blood vessels,

the epicardium and partly to the heart-endothelium, like the lining of the cardiac

3

GRP78

+ cells p13

A

GRP78

+ cells p3

C

mAb C19+ CPCs Sca-1+ CPCs

B

D

G R P7 8

G R P7 8

GRP78

+ cells

GRP78

mAb C19 mAb C19

GRP78

E

Sca-1+ CPCs

Sc a- 1+ C PC

mAb C19+ CPCs

Sca-1+

CPCs GRP78^N20+ cells 0.00

0.05 0.10 0.15

0.20 GRP78 expression

Relative expression to housekeeping gene mAb C19+

CPCs Sca-1+

CPCs 0.0

0.1 0.2 0.3 0.4

0.5 GRP78 expression

Relative expression to housekeeping gene

Figure 1: GRP78 ^N20 isolates a cell population from human heart tissue with limited self-renewal and expansion ability. Human CPCs express GRP78 at the (A) mRNA (N=3) and (B) protein level.

(C) Upon isolation GRP78 ^N20 + cells express GRP78 at the mRNA (N=3) and protein level. (D) Upon expansion in culture, GRP78 ^N20 + cells adopted a fibroblast like morphology. (E) Immunofluorescent staining for mAb C19 and GRP78 ^N20 on Sca-1+ CPCs showing only partial overlap. (Scale bars: 100 µm)

valve (as shown in this chapter) [6]. Most importantly, mAb C19 seems to recognize CPC-like cells dispersed throughout the myocardium. Colocalization between mAb C19 and GRP78 ^N20 was observed in the endothelium lining of the cardiac valve (Fig- ure 2A). However, and in contrast to mAb C19, GRP78 ^N20 was mainly localized to the endothelium like the endothelial layer of blood vessels (Figure 2B). Altogether, these data show that the staining pattern of both antibodies only partly overlap suggest- ing that the affinity of both antibodies is towards different forms of GRP78 and as a consequence towards different cell types.

D ISCUSSION

GRP78 ^N20 + cells, isolated from the human fetal heart, exhibited limited self-renewal

ability. Furthermore, the expression of mAb C19 and GRP78 ^N20 in vitro and in vivo

did only partially overlap suggesting that these antibodies have an affinity for dif-

3

A

B

GRP78

mAb C19 mAb C19 GRP78

mAb C19 GRP78

mAb C19 GRP78

C ar di ac v al ve s Bl oo d ve ss el s

SMC EC

Figure 2: GRP78 ^N20 and mAb C19 only partly colocalize in human heart tissue. Immunofluores- cent stainings on human fetal heart tissue show that (A) mAb C19 and GRP78 ^N20 colocalize to the endothelial lining of the cardiac valve (Scale bar: 100 µm). The high-magnification views in lower panel correspond to the outlined area. (B) Within the vasculature, GRP78 ^N20 has a higher affinity for the endothelium, while mAb C19 mainly localizes to the smooth muscle layer (Scale bar: 20 µm).

SMC: Smooth muscle cell layer and EC: Endothelial cell layer

ferent forms of the GRP78 protein as well as towards different cell types within the human heart.

Heterogeneity within the GRP78 protein repertoire was also reported by

Rauschert et al. (2008). They showed that an O-linked variant of GRP78 was spe-

3

cific for malignant tumor cells and not expressed by epithelial cells and fibroblasts [9, 10]. Furthermore, we observed overlap between mAb C19 and GRP78 ^N20 in the endothelial lining of the valve. Interestingly, disturbed flow may occur around the valves which affects the phenotype of the endothelial cells [1] and has been reported to cause ER stress which increased the expression of GRP78 both in vitro as in vivo [1, 2, 3].

In conclusion, although we cannot exclude that GRP78 ^N20 + cells require a dif- ferent culture condition, our data suggest that mAb C19 cannot be replaced by a commercial GRP78 antibody for the isolation of CPCs from the human heart. Fur- thermore, it signifies that the membrane localized and glycosylated form of GRP78, recognized by mAb C19, is specific for human CPCs and shows the importance of our mAb C19 antibody as tool for CPC isolation. Whether mAb C19 recognizes an O- linked or N-linked sugar will be discussed in chapter 4, however, to which epitope of GRP78 mAb C19 is able to bind remains subject to future studies.

M ETHOD : CELL ISOL ATION AND CULTURE

Cells were isolated and cultured as previously described for Sca-1+ CPCs [12].

Briefly, heart tissue was digested to a single cell suspension using collagenase A (Roche) overnight at 4 ^◦ C followed by magnetic-activated cell sorting (MACS) ac- cording to the manufacturer’s recommendations (Miltenyi Biotec). For isolation with commercial GRP78 (noted as GRP78 ^N20 ) the following antibodies were used:

GRP78 (goat; SC-1050/N20 from Santa Cruz), anti-goat coupled to FITC and anti- FITC coupled to magnetic beads (Miltenyi Biotec).

All cells were cultured on gelatin (0.1%; Sigma) coated dishes and maintained in growth medium which consists of a 1:4 mixture of EGM-2 complete (Lonza) and Medium 199 (M199; Invitrogen) supplemented with 10% fetal calf serum (FCS;

Gibco), 1x non-essential amino acids (Gibco) and 100 U/ml pen/strep (Gibco).

R EFERENCES

1. Chiu JJ, Chien S (2011) Effects of disturbed flow on vascular endothelium: pathophysiological basis and clinical perspectives. Physiol Rev 91:327–387.

2. Davies PF, Civelek M, Fang Y, Fleming I (2013) The atherosusceptible endothelium: Endothelial phenotypes in complex haemodynamic shear stress regions in vivo. Cardiovasc Res 99:315–327.

3. Feaver RE, Hastings NE, Pryor A, Blackman BR (2008) GRP78 upregulation by atheroprone shear stress via p38-, alpha2beta1-dependent mechanism in endothelial cells. Arterioscler Thromb Vasc Biol 28:1534–1541.

4. Goumans MJ, de Boer TP, Smits AM, van Laake LW, van Vliet P, Metz CHG, Korfage TH, Kats KP, Hochstenbach R, Pasterkamp G, Verhaar MC, van der Heyden MAG, de Kleijn D, Mummery CL, van Veen TAB, Sluijter JPG, Doevendans PA (2008) TGF-β1 induces efficient differentiation of human car- diomyocyte progenitor cells into functional cardiomyocytes in vitro. Stem Cell Res 1:138–149.

5. Kelber JA, Panopoulos AD, Shani G, Booker EC, Belmonte JC, Vale WW, Gray PC (2009) Blockade of

Cripto binding to cell surface GRP78 inhibits oncogenic Cripto signaling via MAPK/PI3K and Smad2/3

3

pathways. Oncogene 28:2324–36.

6. Leung HW, Moerkamp AT, Padmanabhan J, Ng SW, Goumans MJ, Choo A (2015) mAb C19 targets a novel surface marker for the isolation of human cardiac progenitor cells from human heart tissue and differentiated hESCs. J Mol Cell Cardiol 82:228–37.

7. Liu M, Chen Z, Chen L (2016) Endoplasmic reticulum stress: a novel mechanism and therapeutic target for cardiovascular diseases. Acta Pharmacol Sin 37:425–443.

8. Ni M, Zhang Y, Lee AS (2011) Beyond the endoplasmic reticulum: atypical GRP78 in cell viability, signaling and therapeutic targeting. Biochem J 434:181–8.

9. Pohle T, Brändlein S, Ruoff N, Müller-Hermelink HK, Vollmers HP (2004) Lipoptosis: Tumor-specific cell death by antibody-induced intracellular lipid accumulation. Cancer Res 64:3900–3906.

10. Rauschert N, Brändlein S, Holzinger E, Hensel F, Müller-Hermelink HK, Vollmers HP (2008) A new tumor-specific variant of GRP78 as target for antibody-based therapy. Lab Investig 88:375–386.

11. Shani G, Fischer W, Justice N, Kelber J, Vale W, Gray P (2008) GRP78 and Cripto form a complex at the cell surface and collaborate to inhibit transforming growth factor beta signaling and enhance cell growth. Mol Cell Biol 28:666–677.

12. Smits AM, van Vliet P, Metz CH, Korfage T, Sluijter JP, Doevendans P a, Goumans MJ (2009) Hu- man cardiomyocyte progenitor cells differentiate into functional mature cardiomyocytes: an in vitro model for studying human cardiac physiology and pathophysiology. Nat Protoc 4:232–243.

13. Spike BT, Kelber JA, Booker E, Kalathur M, Rodewald R, Lipianskaya J, La J, He M, Wright T, Klemke R, Wahl GM, Gray PC (2014) CRIPTO/GRP78 signaling maintains fetal and adult mammary stem cells ex vivo. Stem Cell Reports 2:427–439.

14. Toyoda K, Fukuda T, Sanui T, Tanaka U, Yamamichi K, Atomura R, Maeda H, Tomokiyo A, Take- tomi T, Uchiumi T, Nishimura F (2016) Grp78 Is Critical for Amelogenin-Induced Cell Migration in a Multipotent Clonal Human Periodontal Ligament Cell Line. J Cell Physiol 231:414–427.

15. Zhang Y, Liu R, Ni M, Gill P, Lee AS (2010) Cell Surface Relocalization of the Endoplasmic Reticulum

Chaperone and Unfolded Protein Response Regulator GRP78/BiP. J Biol Chem 15065–15075.