Optimizing Nodal, Wnt and BMP signaling pathways for robust and efficient differentiation of human induced pluripotent stem cells to intermediate mesoderm cells

Several differentiation protocols have enabled the generation of intermediate mesoderm (IM)-derived cells from human pluripotent stem cells (hPSC). However, the substantial variability between existing protocols for generating IM cells compromises their efficiency, reproducibility, and overall success, potentially hindering the utility of urogenital system organoids. Here, we examined the role of high levels of Nodal signaling and BMP activity, as well as WNT signaling in the specification of IM cells derived from a UCSD167i-99-1 human induced pluripotent stem cells (hiPSC) line. We demonstrate that precise modulation of WNT and BMP signaling significantly enhances IM differentiation efficiency. Treatment of hPSC with 3 μM CHIR99021 induced TBXT+/MIXL1+ mesoderm progenitor (MP) cells after 48 h of differentiation. Further treatment with a combination of 3 μM CHIR99021 and 4 ng/mL BMP4 resulted in the generation of OSR1+/GATA3+/PAX2+ IM cells within a subsequent 48 h period. Molecular characterization of differentiated cells was confirmed through immunofluorescence staining and RT-qPCR. Hence, this study establishes a consistent and reproducible protocol for differentiating hiPSC into IM cells that faithfully recapitulates the molecular signatures of IM development. This protocol holds promise for improving the success of protocols designed to generate urogenital system organoids in vitro, with potential applications in regenerative medicine, drug discovery, and disease modeling.


Introduction
The intermediate mesoderm (IM) is a band section of mesodermal tissue lying between from mesoderm progenitors (MP) formed within the posterior primitive streak (PPS) (Saxón and Sariola, 1987;Nájera and Weijer, 2023).Gastrulation is a crucial stage in embryonic development, where the single-layered blastocyst transforms into a tri-layered structure, the ectoderm, mesoderm, and endoderm.The primitive streak, a transient structure formed during gastrulation, plays a key role in this process.Cells from the epiblast (outer layer) undergo epithelial-to-mesenchymal transition and migrate inwards through the primitive streak.The initial wave displaces the hypoblast cells to form the definitive endoderm (DE) (future gut lining).The second wave populates the space between the epiblast and endoderm, giving rise to the mesoderm layer.This mesoderm further differentiates into five subpopulations: paraxial, intermediate, lateral plate, cardiogenic, and notochordal mesoderm.Once the mesoderm is formed, the remaining epiblast cells stop migrating and become the ectoderm (outer layer) (Mouse development, 2007;Ghimire et al., 2021;Schoenwolf et al., 2021;Zhai et al., 2022).
Intricate signaling pathways, like Nodal, WNT, and BMP, activate specific sets of genes that dictate the fate of each cell, ultimately shaping the developing organism (Perrimon et al., 2012).The Nodal pathway plays a central role in specifying mesoderm fate (Schier, 2003;Ben-Haim et al., 2006).Traditionally, Nodal signaling has been viewed as a graded morphogen, where high levels trigger the formation of the endoderm layer and lower levels promote mesoderm development (Schier, 2003;Zorn and Wells, 2009).WNT signaling is crucial in mesoderm development.Studies have demonstrated its necessity for both initiating mesoderm formation and maintaining the pool of posterior mesodermal progenitors (Lindsley et al., 2006;Martin and Kimelman, 2008;Tanaka et al., 2011;Blauwkamp et al., 2012;Przybyla et al., 2016).BMP signaling also plays a critical role, with lower levels promoting the formation of IM, while higher levels favor the development of the LPM (Tonegawa and Takahashi, 1998;James and Schultheiss, 2005;Knarston et al., 2020).
Technical and ethical limitations pose a significant challenge to directly studying early human development.However, human pluripotent stem cells (hPSC) offer powerful tool to overcome these limitations.Recent studies have demonstrated the generation of various cell types from hPSC, including IM-derived cells from hPSC (Yucer et al., 2021;2017;Miyazaki et al., 2018;Knarston et al., 2020;Bejoy et al., 2022;Gong et al., 2022).This ability of hPSC to recapitulate different stages of human embryonic development, mimicking the natural developmental sequence, has revolutionized our understanding of human biology and disease.This has led to significant advancements in disease modeling, drug discovery, and regenerative medicine research (Thomson et al., 1998;Takahashi et al., 2007;Zhu and Huangfu, 2013;Kim et al., 2020).hPSC hold another valuable application in developing organoids for modeling genetic disorders.This ability to create patient-specific organoids reflects the growing recognition of genetic diversity and inter-individual variability in human biology and disease (Perez-Lanzon et al., 2018).The UCSD167i-99-1 hPSC line used in this study serves as a well-defined system for studying these factors.Its extensive genetic characterization makes it a reliable tool for investigating the impact of genetics on cellular differentiation (Panopoulos et al., 2017).
While these studies have demonstrated the successful generation of hPSC-derived IM cells, the differentiation strategies employed exhibit substantial variability between protocols, with variations in the choice of morphogens and the concentrations of the signaling molecules.Meanwhile, to accurately recapitulate IM development, hPSC differentiation should emulate the natural sequence of molecular signals that occur during embryogenesis as closely as possible.
In this sense, refining and optimizing the existing differentiation methodology to enhance the efficiency of IM cell generation would significantly increase the reproducibility and success of protocols to form urogenital system organoids in vitro.
For this purpose, we adopted the strategy outlined by Yucer et al., 2017 and investigated the roles of high levels of Nodal and BMP signaling activity, along with WNT signaling, in the determination of IM cells derived from a UCSD167i-99-1 hiPSC line, obtained from the WiCell Research Institute.
Here, we report a significant yet simple modification of the conventional methods, enabling IM development to faithfully recapitulate the molecular features described in embryological literature.We demonstrate that IM cells can be generated from a UCSD167i-99-1 hPSC line, addressing these features by suppressing the Nodal signaling during the mesoderm step and employing low concentrations of BMP4 for IM generation.Finally, our approach enhances the efficacy of IM differentiation and establishes a robust and reproducible protocol for differentiating hiPSC into IM cells.

Indirect immunofluorescence
Coating coverslips for hiPSC and embryonic tissue immunostaining Sterilized glass coverslips (cat.#11856933, Thermofisher Scientific) were placed into wells of 24-well plates and treated with fibronectin (0.33% vol/vol) (cat.#1918-FN,R&D Systems) for 2 h at RT. Subsequently, coverslips were washed twice in 1X PBS and allowed to dry in the laminar flow hood.Then, coverslips were coated with Matrigel ™ overnight.hiPSC were plated at a 1: 24 split ratio (wells from 6-well dish: wells in 24-well dish) onto Matrigel ™ -coated glass coverslips and maintained in mTeSR ™ 1 medium for 24 h before hiPSC immunostaining.For MP immunostaining, the medium was replaced with induction medium and cells were incubated for an additional 48 h before staining.For IM immunostaining, 70,000-140,000 cells per well were plated onto fibronectin-coated 24-well plates and maintained in induction medium for 48 h before staining.

Quantitative real-time RT-PCR (RT-qPCR) analysis
Total RNA was isolated using the Qiagen RNeasy Mini kit (cat.#74014;Qiagen) following the manufacturer's instructions.The concentration of total RNA was measured using a NanoDrop (Thermofisher Scientific).cDNA was synthesized using the High-Capacity cDNA kit (cat.#4368814;Thermofisher Scientific).Real-time qPCR was performed using the PowerUpSYBR Green mix (cat.#A25776,Thermofisher Scientific) on the Stratagene ™ MX3005P (Thermofisher Scientific).Absolute quantification was performed using a standard curve of serial diluted genomic DNA and normalized to β-Actin.Relative quantification was performed by comparing the samples to undifferentiated hiPSC as a reference control, and was determined using the ΔΔCt method.Primer sequences used in this study were as follows: β

Statistical analysis
The data were represented as the mean ± standard error of the mean (SEM) of n = 3 biological replicates from at least three independent experiments.For statistical comparisons, unpaired Student's t-tests were performed to calculate p-values between experimental conditions and controls and a p-value <0.05 was considered statistically significant.Significance of the analysis of the Student's t-test is indicated in the figures as *, p < 0.05; **, p < 0.01; ***, p < 0.001 and ****, p < 0.0001.Statistics were calculated with the Prism nine software (GraphPad Software).

Results
We aimed to establish an optimized protocol for generating UCSD167i-99-1 hiPSC line-derived induced IM cells.To achieve this, we modified the differentiation protocol described by Yucer et al., 2017 (Figures 4A, B).We first optimized the conditions by differentiating hiPSC into MP cells and then into IM cells, following a modified version of the Yucer et al., 2017 strategy (Figure 1A).
Gene expression analysis revealed a significant upregulation of TBXT within 24 h of treatment with 2 and 6 μM CHIR99021 compared to 0 μM CHIR99021, peaked at 48 h with 2, 4, and 6 μM CHIR99021 treatments compared to control and then decreased by 72 h with 4 and 6 μM CHIR99021 treatments compared to control.Treatments with 2 and 4 μM CHIR99021 induced significantly higher levels of TBXT at 48 h compared to 24 and 72 h (Figure 2A).
Increased expression levels of MIXL1 were observed significantly within 24 h of treatment with 2, 4 and 6 μM CHIR99021, compared to 0 μM CHIR99021, peaked at 48 h with all three concentrations of CHIR99021, and decreased by 72 h with 4 and 6 μM CHIR99021, compared to control.Treatment with 6 μM CHIR99021 induced significantly higher levels of MIXL1 at 48 and 72 h, compared to 24 h (Figure 2B).
However, qPCR analysis revealed a significant downregulation of pluripotency-associated genes POU5F1 and NANOG in cells treated with 4 and 6 μM CHIR99021 compared to 2 μM CHIR99021 and the untreated control group (0 μM) (Figures 2C, D).
Immunofluorescence analysis after 48 h of exposure showed that treatment >2 μM CHIR99021 induced the MP marker TBXT (Figure 2E).However, higher doses >4 μM CHIR99021 resulted in a concentration-dependent decrease in cell number (Figure 2E).Moreover, the percentage of TBXT was significantly higher with increasing concentrations of CHIR99021 (Figure 2F).To validate these findings and determine the optimal CHIR99021 concentration, we evaluated the expression levels of the MP markers TBXT and MIXL1 by RT-qPCR in hiPSC exposed to 0, 3, and 4 μM CHIR99021 for 48 h (Figures 2H, I).The results revealed a significant increase in both TBXT and MIXL1 expression in cells treated with three and 4 μM CHIR99021 compared to 0 μM CHIR99021.There were no statistically significant differences in the expression levels of either marker between the 3 μM and 4 μM CHIR99021 concentrations (Figures 2H, I).Similarly, while genes associated with pluripotency (POU5F1 and NANOG) remained downregulated compared to the 0 μM and 2 μM treatment groups, no significant differences were found between the 3 μM and 4 μM CHIR99021 treatments (Figures 2J, K).Thus, we chose an optimal CHIR99021 concentration of 3 μM and a treatment duration of 48 h for subsequent experiments.
We also detected co-expression of the pluripotency marker POU5F1 with MP marker TBXT.However, the increase in TBXT expression is consistent with a reduction in pluripotency, as evidenced by the decreased POU5F1 staining in confocal images (Figure 2E). Figure 2G further demonstrates a statistically significant decrease in the percentage of POU5F1+ cells within the MP cell population upon treatment with increasing concentrations of CHIR99021.
To assess the role of highly activated Nodal signaling in the generation of MP cells from hiPSC, we evaluated the effects of 100 ng/mL Activin A, a known agonist of Nodal signaling (Baharvand and Aghdami, 2012), alone or in combination with 3 μM CHIR99021 for 48 h.We found that the addition of Activin A to CHIR99021 significantly decreased the expression levels of TBXT (Figure 3A) and MIXL1 (Figure 3B), compared to CHIR99021 alone, although resulted in increased expression levels of SOX17 and FOXA1, DE markers (Séguin et al., 2008;Wang et al., 2011;Reizel et al., 2021), compared to CHIR99021 alone (Figures 3C, D).
Immunostaining analysis confirmed that undifferentiated hiPSC expressed POU5F1 but not TBXT.Treatments with CHIR99021 alone or in combination with Activin A clearly induced the expression of TBXT (Figure 3E).However, the percentage of TBXT + cells was significantly lower in Activin A in combination with CHIR99021 treatment than in CHIR99021 alone (Figure 3F).The pluripotency marker, POU5F1, was still expressed during the induction of MP cells from hiPSC (Figure 3E).
In order to evaluate the effect of BMP signaling activity on IM specification, we used different concentrations of BMP4, a key ligand in IM differentiation (James and Schultheiss, 2005), in combination with 3 μM CHIR99021, also involved in the differentiation of IM (Sepponen et al., 2017).After 48 h of exposure, CHIR99021-induced MP cells were exposed to 0, 4, 25 and 100 ng/mL BMP4 and 3 μM CHIR99021 for 48 h.We optimized the conditions for the generation of IM spheroids as described in (McCracken et al., 2011).Differentiation outcomes were assessed by immunofluorescence and qPCR analysis of known IM markers OSR1, GATA3 and PAX2 (Yucer et al., 2021;2017).
We first verified the absence or low gene expression levels of the IM markers OSR1, PAX2 and GATA3 in MP cells by RT-qPCR (Supplementary Figures 1A-C).We also confirmed low expression of OSR1 and PAX2 IM markers in MP cells by comparing their fluorescence intensity to hiPSC markers through immunofluorescence analysis (Supplementary Figure 1D).
Hence, we chose an optimal BMP4 treatment of 4 ng/mL and 3 μM CHIR99021 for 48 h in further experiments.
Additionally, we investigated the effect of low doses of BMP4 (4 ng/mL) on MP cells differentiated with 100 ng/mL Activin A, alone or with 3 μM CHIR99021 and studied OSR1, PAX2 and GATA3 IM markers by RT-qPCR (Figures 5A-C).MP cells treated with 100 ng/mL Activin A and 3 μM CHIR99021 showed a significantly decreased expression of the IM markers OSR1 and PAX2 compared to cells treated with CHIR99021 alone (Figures 5A,  B).However, we did not observe differences between treatments in the expression of the IM marker GATA3 (Figure 5C).

Discussion
Here, we report an efficient and reproducible protocol for generating IM cells from the UCSD167i-99-1 hiPSC line (Panopoulos et al., 2017).Our protocol faithfully recapitulates key molecular events of IM development described in the current embryological literature.We achieved this by optimizing the signaling factors Nodal, WNT, and BMP.Unlike existing methods that often rely on combined high Nodal and WNT activation typically associated with the anterior primitive streak (APS) origin, our protocol emphasizes WNT activation from a PPS origin, mimicking the natural developmental process of IM cells.Furthermore, we employ low BMP4 concentrations, compared to protocols using high BMP4 that can lead to undesired cell fate transitions.This aligns with the established role of BMP4 gradients in separating IM and LPM domains during development.Current IM differentiation protocols exhibit significant heterogeneity in culture conditions and utilized markers.Our protocol addresses this by focusing on optimizing the signaling pathways crucial for IM development, while acknowledging the limitations of current marker identification.Importantly, our method reduces the number and concentrations of morphogens compared to other protocols, while maintaining high yields of IM cells.Many established protocols require various morphogens or concentrations (Yucer et al., 2017;Miyazaki et al., 2018;Knarston et al., 2020;Bejoy et al., 2022;Gong et al., 2022), leading to inconsistent target cell generation and lacking a lineage-directed approach.By replicating features of embryogenesis not captured in previous protocols, our method offers a more robust, reproducible, and efficient approach for organoid development and disease modeling.
An essential parameter in optimizing the differentiation of MP cells from the UCSD167i-99-1 hiPSC line was the concentration of CHIR99021.Our results showed that high concentrations of CHIR99021 (>4 μM) achieve robust commitment of the UCSD167i-99-1 hiPSC line to TBXT+/MIXL1+ MP cells within 48 h, with a reduction in pluripotency.However, we observed a decrease in the number of cells at > 4 μM doses.This decrease aligns with previous findings on the detrimental effects of DMSO, a commonly used solvent in drug screening, on cell morphology and attachment, and viability in a dose-dependent manner (Pal et al., 2012).In addition, our results demonstrated a significant upregulation of both TBXT and MIXL1 expression in hiPSC treated with 3 μM and 4 μM CHIR99021 for 48 h, highlighting the potency of 3 μM CHIR99021 in inducing MP differentiation markers, as it achieved comparable levels of induction to a higher concentration >4 μM without compromising cell abundance.Conversely, the pluripotency markers POU5F1 and NANOG showed a significant downregulation in hiPSC treated with 3 μM and 4 μM CHIR99021 compared to the 2 μM treatment group.Therefore, we opted for 3 μM CHIR99021 as the optimal concentration, balancing efficient differentiation with maintaining a robust cell population.Notably, this selection aligns with established protocols for MP differentiation using CHIR99021 in UCSD167i-99-1 hiPSC, typically employing a 3 μM concentration (Musah et al., 2017;Bejoy et al., 2021).This further validates the use of CHIR99021 as a potent activator of WNT signaling in UCSD167i-99-1 hiPSC.
Another important finding of our research was that removing the Nodal pathway agonist yielded the most reproducible and efficient MP cell differentiation.In this regard, we observed that the presence of Nodal signaling, along with WNT signaling, increased the expression of the DE marker SOX17 and FOXA1.This finding is consistent with the observation that the Nodal signaling pathway has been described to differentiate hiPSC into DE cells in several induction protocols (Ardila Riveros et al., 2021;D'Amour et al., 2005;McLean et al., 2007;Pagliuca et al., 2014).However, while previous reports (Fleming et al., 2013) highlight the role of low Nodal signaling in modulating BMP signaling, which is essential for IM formation, current attempts at IM induction from hiPSC rely on high levels of Nodal signaling and WNT activation to specify mesoderm with an origin in the PPS, which is not consistent with the embryological literature (Yucer et al., 2017;Bejoy et al., 2022;2021).According to scientific literature, during gastrulation, a subset of epiblast cells invaginate through the primitive streak to form the endoderm and mesoderm (Tam and Behringer, 1997;Loebel et al., 2003).The APS generates DE (Levak-Švajger and Švajger, 1974;Takasato et al., 2014;Ardila Riveros et al., 2021), whereas the PPS forms mesoderm (Tam and Beddington, 1987;Lawson et al., 1991;Takasato et al., 2014;Takasato and Little, 2015;Ardila Riveros et al., 2021), and mesoderm will give rise to the axial, paraxial, intermediate, and lateral plate mesoderm (Prummel et al., 2020).
On the other hand, the fact that WNT pathway activation alone efficiently drives hPSC into mesodermal lineages but also specifies endoderm lineage is in line with the shared primitive streak origin of both germ layers during gastrulation, suggesting that these cells may represent mesendoderm progenitors (Levak-Švajger and Švajger, 1974;Tam and Beddington, 1987;Lawson et al., 1991).Indeed, DE and MP co-express TBXT, MIXL1, SOX17, FOXA2, among other markers (Kubo et al., 2004;Tada et al., 2005;Faial et al., 2015).
Regarding IM differentiation, our data demonstrated that BMP4 ligand could efficiently achieve IM differentiation at all concentrations tested using the UCSD167i-99-1 hiPSC line, aligning with the established role of BMP signaling in regulating mesodermal cell fate and specifically differentiating them into IM (Knarston et al., 2020;Magro-Lopez and Muñoz-Fernández, 2021).Notably, low doses of BMP4 efficiently promoted the differentiation of MP cells into OSR1+ IM cells, while high doses repressed IM differentiation.In line with this, prior studies have shown that mesodermal subtypes are specified along the mediolateral axis depending on BMPs activity, with low activity enhancing IM formation and higher activity in the LPM tissue (Tonegawa and Takahashi, 1998;James and Schultheiss, 2005;Knarston et al., 2020).However, several existing protocols for IM induction still employ high doses of BMP4 (Yucer et al., 2021;2017).
In contrast, we did not observe a similar pattern for PAX2 and GATA-3 marker expression.Interestingly, previous studies have identified OSR1 as an early marker of IM (James et al., 2006), while PAX2 and GATA3 have been widely used as late markers in the nephric duct, an epithelial duct differentiated from IM cells and essential for all further urogenital development (Grote et al., 2006;Boualia et al., 2013;Morizane and Lam, 2015;Sanchez-Ferras et al., 2021).Furthermore, we found that low-dose BMP4 treatment enhanced IM differentiation when MP cells were differentiated by WNT activation alone.This result suggests that IM differentiation may be optimized if it originates from PPS (WNT activation alone) rather than APS (combined Nodal and WNT activation) (Takasato et al., 2016;Knarston et al., 2020), supporting our previous results.
These findings highlight the heterogeneity not only in the culture protocols employed in current differentiation methods but also in the markers utilizied for identifying the mesoderm lineage.However, future efforts should focus on identifying more specific markers for early IM stages to enhance differentiation accuracy.In summary, optimizing the Nodal, WNT and BMP signaling pathways in a UCSD167i-99-1 hiPSC line, significantly enhanced the induction of IM cells from hPSC compared to existing protocols.We found the need to precisely define the specification of IM from hPSC and establish a standardized differentiation protocol that closely recapitulates in vivo IM development to efficiently generate human organoids that closely resemble the human embryo in both identity and function.Further advancements in IM protocols may involve exploring the impact of different Activin/Nodal signaling levels, alongside integrating other morphogenic signaling pathways and investigation of short-term culturing strategies.These approaches warrant further investigation to expand our understanding of human development and disease, enabling the development of innovative therapies for a wide range of conditions.

FIGURE 2 WNT
FIGURE 2 WNT-mediated MP cells differentiation.(A, B) Relative mRNA expression levels of MP markers TBXT and MIXL1 determined by RT-qPCR in differentiated cells at 24, 48 and 72 h of treatment with 0, 2, 4 and 6 µM CHIR99021.(C, D) Relative mRNA expression levels of pluripotent markers POU5F1 and NANOG determined by RT-qPCR in differentiated cells at 48 h of treatment with 0, 2, 4 and 6 µM CHIR99021.(E) Immunofluorescence images of differentiated cells stained for pluripotent marker POU5F1 (green), MP marker TBXT (red), merged with nucleus marker DAPI (cyan or magenta) at 48 h of treatment with 0, 2, 4 and 6 µM CHIR99021.Scale bars: 100 μm.Highresolution insets (×40 magnification) are presented directly below the main figure.These insets zoom in on representative cells, highlighting the details of stained merged nuclei (POU5F1 or TBXT with DAPI).This enhanced view complements the main figure, which displays images at a lower magnification (×20).(F, G) Quantification of TBXT and (Continued )

FIGURE 3
FIGURE 3 Nodal and WNT-mediated MP cells differentiation.(A-D) Relative mRNA expression levels of MP markers TBXT, MIXL1 and DE markers SOX17 and FOXA1 determined by RT-qPCR in differentiated cells at 48 h of treatment with 3 µM CHIR99021 alone and with 100 ng/mL Activin A. (E) Immunofluorescence images of hiPSC colonies and differentiated cells stained for POU5F1 (green), MP marker TBXT (red), merged with nucleus marker DAPI (cyan or magenta) at 48 h of treatment with 3 µM CHIR99021 alone and with 100 ng/mL Activin A. Scale bars: 100 μm.(F) Quantification of TBXT + cells in MP cells differentiated from hPSC after 48 h of treatment with 3 µM CHIR99021 alone and with 100 ng/mL Activin A. Error bars represent mean ± SEM, n = 3 independent experiments, n. s not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.Gene expression levels were determined by relative quantification compared to undifferentiated hiPSC.hiPSC human induced pluripotent stem cells, MP mesoderm progenitor, IM intermediate mesoderm, DE definitive endoderm.

FIGURE 4
FIGURE 4 BMP and WNT-mediated IM cells differentiation.(A) Schematic diagram of the modified differentiation strategy, without Activin A and low BMP4 concentrations, to induce hiPSC into MP and IM cells.(B) Representative bright-field images showing the stages of cell differentiation from hiPSC to IM cells.(C-E) Relative mRNA expression levels of IM markers OSR1, PAX2 and GATA3 determined by RT-qPCR in differentiated cells after 48 h of treatment with 0, 4, 25 and 100 ng/ mL BMP4 and 3 µM CHIR99021.(F) Immunofluorescence images of cells stained for IM markers OSR1, PAX2, merged with nucleus marker DAPI (cyan or magenta) at 48 h of treatment with 4 ng/mL BMP4 and 3 µM CHIR99021.Scale bars: 100 μm.Orange arrowheads indicate cells with a spheroid morphology, which is characteristic of IM cells.Error bars represent mean ± SEM, n = 3 independent experiments, n. s (Continued )