Small molecule inhibitors of the Wnt pathway potently promote cardiomyocytes from human embryonic stem cell derived mesoderm

Abstract

Rationale

Human embryonic stem cells (hESCs) can form cardiomyocytes when cultured under differentiation conditions. Although the initiating step of mesoderm formation is well characterized, the subsequent steps that enrich for cardiac lineages are poorly understood and limit the yield of cardiomyocytes.

Objective

Our aim was to develop a hESC-based high content screening (HCS) assay to discover small molecules that drive cardiogenic differentiation after mesoderm is established to improve our understanding of the biology. Screening of libraries of small molecule pathway modulators was predicted to provide insight into the cellular proteins and signaling pathways that control stem cell cardiogenesis.

Methods and results

About 550 known pathway modulators were screened in a HCS assay with hits being called out by the appearance of a red fluorescent protein driven by the promoter of the cardiac specific MYH6 gene. One potent small molecule was identified that inhibits transduction of the canonical Wnt response within the cell, demonstrating that Wnt inhibition alone is sufficient for deriving cardiomyocytes from hESC originating mesoderm cells. Transcriptional profiling of inhibitor-treated compared to vehicle-treated samples further indicated that inhibition of Wnt does not induce other mesoderm lineages. Notably, several other Wnt inhibitors are very efficient in inducing cardiogenesis, including a molecule that prevents Wnts from being secreted by the cell, confirming Wnt inhibition as the relevant biological activity.

Conclusions

Pharmacological inhibition of Wnt signaling is sufficient to drive human mesoderm cells to form cardiomyocytes, yielding novel tools for the benefit of pharmaceutical and clinical applications.

Introduction

Heart disease often leads to cardiomyocyte death and pathological remodeling, and the only replacement option is heart transplantation, but clinical complexity and limited donors have prompted research into stem cell-based alternatives.

Stem cell-based approaches include both cell transplantation and mobilization of an endogenous stem and progenitor pool, and show promise for therapeutic regeneration¹. Development of stem cell-based replacement therapies, however, is limited by incomplete understanding of the factors that drive differentiation of cardiomyocytes from either human embryonic stem cells (hESCs) or endogenous cardiac stem and progenitor cells. Therefore, we sought to develop high throughput screens for simultaneous testing of small molecules for cardiogenic potential in order to identify cellular signals that are required or need to be inhibited at different stages of cardiac development². Such a chemical biology approach would moreover generate small molecule tools to improve the yield in hESC differentiation protocols and may even result in drug leads for cardiac regeneration should they target hESC-derived progenitors resembling adult cardiac progenitor cells.

Here, we describe the development and implementation of a hESC-based assay that probes the signals that drive the conversion of PDGFRα⁺MESP1⁺ cardiogenic cells to cardiomyocytes as well as one of the hits that came forth from screening pathway modulator libraries.

Methods

MYH6-mCherry hESC³ were differentiated in StemPro 34 with growth factors as outlined in Figure 1A. For high content screening (HCS) day 4 embryoid bodies (EBs) were dissociated and plated in presence of small molecules.

Figure 1. A high content screen for cardiogenesis in hESC identifies a Wnt inhibitor.

(A) In EB, after treatment with the indicated growth factors, mesoderm is induced efficiently from days 1-4 (assessed by mRNA, upper half, blue full line) and cardiomyocytes occur spontaneously from day 6 onward (assessed by mRNA, upper half, red dashed line). For the HCS assay, day 4 EBs were dissociated, plated, and treated with molecules and/or growth factors as depicted in the lower half. (B) Day 4 cells have reached maximum mesoderm levels and were further characterized for early cardiac markers by RT-qPCR for MESP1 and for PDGFRα by flow cytometry (average ± standard deviation, n = 3). (C,D) The HCS assay was screened against a collection of 244 kinase inhibitors (InhibitorSelect) (C) and a collection of 305 pathway modulators (StemSelect) (D) in three concentrations (0.3, 1 and 3μM). All compounds with a Z score larger than 2 are shown for the StemSelect library (D). (E) Fluorescence image of cells expressing the MYH6-mCherry reporter under vehicle or IWR treated cultures.

For analysis, red fluorescence from the cardiac specific MYH6 reporter was imaged using a high throughput microscope and was quantified using an image analysis software package.

Results

In order to develop a HCS assay for small molecules that drive mesoderm to cardiomyocytes in hESC, we first profiled cardiac marker expression during differentiation in EBs (Figure 1A, upper half), in which cardiac-inducing signals are provided to mesodermal cells by closely juxtaposed endodermal and potentially other cell types present in early embryos⁴ (Online Figure I). Mesoderm was induced in EBs by addition of Activin A and BMP4 and results in mesoderm induction between days 1 and 4 (Figure 1A, blue full line), and cardiomyocytes appeared spontaneously from day 6 onwards (Figure 1A, red dashed line). As mesoderm induction appeared to be maximal at day 4, we further characterized the day 4 population for the early cardiac markers MESP1 and PDGFRα (Figure 1B)⁵. As this population appeared to be highly enriched for both markers, we developed a HCS assay that would probe these cells and ask what signals would drive differentiation towards cardiomyocytes (Figure 1A, lower half). HCS assays are image-based and have better dynamic range and sensitivity than traditional plate-reader-based assays. Day 4 EBs were collected and dispersed into 384-well plates, and subsequently exposed to three concentrations (0.3μM, 1μM and 3μM) of two small collections of well-characterized pathway modulators, namely about 244 protein kinase inhibitors (InhibitorSelect) as well 305 pathway agonists and antagonists (StemSelect). In the InhibitorSelect collection, no compound induced cardiogenesis (Figure 1C) whereas in the StemSelect collection one extremely potent hit was identified (Figure 1D). The hit IWR-1, a recently published inhibitor of the canonical Wnt signaling response (IWR)⁶, induced beating foci from monolayer cultures, whereas none were observed in DMSO conditions (Figure 1E, Supplementary Movies 1 and 2). Inhibitors of Egf, Vegf, Tgfβ, Insulin and Shh signaling did not result in increased cardiogenesis, suggesting Wnt inhibition alone is sufficient for cardiac induction (Online Table I). Moreover, agonists of the Wnt, Vegf and Shh pathways failed to promote cardiac fate (Online Table I).

Further characterization revealed that IWR’s maximal cardiac induction occurs when added from day 4 to day 5, with cardiac induction levels decreasing as the treatment window is shifted up towards day 10 of differentiation (Figure 2A). The EC₅₀ to induce cardiogenesis at the day 4 time window was 2241 nM (Figure 2B). Flow cytometry analysis demonstrated that IWR yields up to 30% cardiomyocytes, a 200-fold increase over vehicle treated cultures (Figure 2C). RT-qPCR analysis was performed to study the effect of IWR on mesoderm, cardiac progenitors and mesoderm derivatives such as cardiomyocytes, endothelial cells and smooth muscle (Figure 2D). The mesoderm specific gene T/BRACHYURY had a downward tendency as soon as 24 hours after IWR exposure, although not statistically significant at most days (Figure 2D). The earliest markers of cardiac fate, MESP1 and KDR, were not affected in the first few days after IWR treatment, suggesting IWR drives MESP1⁺, KDR⁺ cardiac mesoderm cells towards cardiomyocytes (Figure 2D). MEF2C and NKX2.5, two cardiac transcription factors, as well as the cardiomyocyte structural genes MYH6, ACTN2 and TNNT2 were increased dramatically by IWR (Figure 2D). Further characterization of other mesodermal lineages revealed smooth muscle lineages were not increased (ACTA2 and TGLN) and endothelial cell markers (CD31 and KDR) even decreased. Flow cytometry confirmed a >50% reduction in CD31⁺ cells (Online Figure II).

Figure 2. IWR specifically induces cardiomyocytes.

(A) Dose response curves for cardiogenesis by IWR with exposure at different time points show IWR is most active when added at day 4. (B) In the optimal window, IWR has an EC₅₀ for cardiogenesis of about 2241 nM. (C) Flow cytometry analysis for the MYH6-mCherry reporter of vehicle or IWR treated cells, gated against a negative control. Numbers reflect the percentage of MYH6 positive cells. (D) Time course RT-qPCR analysis of IWR treated monolayer samples (Red, squares) versus DMSO treated samples (blue, diamonds) normalized to the DMSO day 5 sample. Early markers for cardiogenic mesoderm include T (BRACHYURY), MESP1 and KDR, cardiac progenitor markers tested were MEF2C and NKX2.5, and cardiomyocyte markers studied were MYH6, TNNT2 and ACTN2. Other mesoderm lineages tested were endothelium (CD31) and smooth muscle (ACTA2 and TAGLN). Error bars represent SEM of 3 biological replicates. Asterisk (*) indicates p<0.05 compared to the corresponding DMSO control. (E) Schematic overview of the lineage diversification of a cardiovascular mesoderm progenitor. Wnt inhibition specifically drives mesoderm progenitors to cardiomyocytes.

We next questioned whether other inhibitors of the Wnt pathway were also able to induce cardiogenesis in this assay (Figure 3A). We tested 3 more small molecule inhibitors of Wnt signaling: (1) the PORCN inhibitor IWP-3 (IWP)⁶; (2) the more potent IWR-1 analog 53AH, and (3) the Tankyrase inhibitor XAV939 (XAV) which is cardiogenic in mouse ESC⁷. All three compounds promoted cardiogenesis, with 53AH having an EC₅₀ below the micromolar range (Figure 3B). Since EC₅₀ measurements are not indicative of the efficacy of each compound, we ran maximal effect doses of each compound, as well as the biological Wnt inhibitor DKK1 per comparison to evaluate their maximum efficacy in terms of cardiogenesis. IWR and its analog 53AH are most efficacious, followed by IWP and XAV, and any of the small molecule inhibitors were at least 40 times more effective in inducing cardiogenesis compared to DKK1, a purified recombinant Wnt inhibitor protein (Figure 3C).

Figure 3. Inhibition of Wnt at different levels of the pathway promotes cardiogenesis.

(A) Chemical structures of the Wnt inhibitors tested in the cardiogenesis assay. (B) Dose response curves and EC₅₀ values for cardiogenesis. (C) Maximum cardiac induction concentrations of Wnt inhibitors tested to indicate efficacy of each molecule, the natural inhibitor DKK1 is included to demonstrate superior efficacy by small molecules. Asterisk (*) indicates p<0.05 relative to the untreated condition. (D) Schematic overview of the canonical Wnt pathway, and active Wnt inhibitors (green). Green lines indicate general function of the inhibitors on their targets.

Discussion

Defining the pathways that control stem cell cardiogenesis is important to efficiently derive cardiomyocytes from human pluripotent stem cells and might be useful to mobilize endogenous cardiac stem cells. To discover new molecules and pathways that would direct cardiac differentiation and/or regeneration, we developed a hESC-based HCS assay that allows small molecule screens in serum-free conditions, in a scale that has not been reported for any hESC based assay. We prepared cardiogenic mesoderm (PDGFRα⁺MESP1⁺) cells in bulk EB culture (Figure 1B)⁵, and dispersed and plated the cells in a monolayer with the intent of identifying small molecules that are able to replace the natural signals provided in the 3D EB, thus yielding insight into signals that direct mesodermal progenitor to form cardiomyocytes.

By screening about 550 pathway modulators we have identified a small molecule inhibitor of the β-catenin dependent canonical Wnt pathway. While it was not entirely unexpected to identify a Wnt inhibitor as cardiac inducer since the natural Wnt inhibitor DKK1 is capable of directing for cardiogenesis in Xenopus⁸ and hESCs⁵, our study shows that Wnt inhibition is sufficient for driving hESC derived mesoderm to cardiac fate, in the absence of other signaling modulators, and no other inhibitors had comparable activity (Online Table I). Moreover, Wnt inhibition drives cardiomyocytes specifically and does not increase other mesoderm lineages, suggesting Wnt inhibition specifically drives a mesoderm progenitor towards cardiomyocytes (Figure 2D-E).

To further explore inhibition throughout the Wnt pathway, we evaluated structurally diverse Wnt inhibitors that target different cellular components of the pathway (Figure 3D). Importantly, all of the small molecules were much more cardiogenic than DKK1 (Figure 3C). The most interesting finding, however, was that IWP, an inhibitor that prevents cells from producing Wnts, is also potent, revealing that endogenous Wnt activity must be blocked. This is an important finding as it suggests that not only exogenously added factors are important for efficient differentiation, but that also endogenous cellular signals may have to be modulated to control differentiation. As these cultures are typically heterogeneous, further experimentation may unravel which cell type is, in this case, secreting inhibitory Wnts.

In summary, HCS of a hESC-based assay identified Wnt inhibition as critical for cardiogenesis. Active compounds function by blocking Wnt secretion and stabilize Axin to destabilize β-catenin. Our data further suggest that endogenous Wnt signals need to be inhibited to direct the formation of cardiomyocytes from a mesoderm progenitor.

Novelty and significance.

What is known?

• Human embryonic stem cells (hESC) show great promise as a source for generating myocardial cells for use in cell transplantation therapies.

• hESC form cardiac myocytes if they are treated appropriately. However, the yield is typically low, as the mechanisms that drive hESC towards a cardiac myocyte phenotype are poorly understood.

What new information does this article contribute?

• We developed a hESC-based high throughput small molecule screen assay, using a cardiac myocytes-specific fluorescent reporter, to identify molecules that drive hESC to cardiac myocytes.

• Small molecule inhibitors of the Wnt signaling pathways were identified as potent inducers of cardiac myocytes. They increased cardiac myocyte yield dramatically over recombinant protein inhibitors. Many other pathway modulators were inactive. Thus, Wnt inhibition is uniquely important for directed cardiogenesis

The development of a high throughput assay for hESC cardiogenesis was deemed essential as it allowed simultaneous probing of the many signals that may drive hESC to cardiac myocytes. Screening of 500+ pathway modulators demonstrated that inhibition of the Wnt pathway alone is sufficient to drive cardiac cell formation specifically and does not induce other mesodermal derivatives such as endothelial or smooth muscle cells. Small molecule inhibitors of the Wnt pathway are thus useful tools for increasing cardiac myocyte yield. These molecules have great potential benefits for clinical and pharmaceutical applications

Non-standard abbreviations and acronyms

hESCs

human embryonic stem cells

EBs

embryoid bodies

HCS

high content screen

RT-qPCR

reverse transcription quantitative polymerase chain reaction

IWR

inhibitor of Wnt response, IWR-1

IWP

inhibitor of Wnt production, IWP-3

XAV

inhibitor of tankyrase, XAV939

53AH

structural analog of IWR-1

InhibitorSelect

small molecule collection of 244 kinase inhibitors

StemSelect

small molecule collection of 305 pathway modulators