Patient-Specific iPSC-Derived Endothelial Cells Uncover Pathways that Protect against Pulmonary Hypertension in BMPR2 Mutation Carriers

Abstract

In familial pulmonary arterial hypertension (FPAH), the autosomal dominant disease-causing BMPR2 mutation is only 20% penetrant, suggesting that genetic variation provides modifiers that alleviate the disease. Here, we used comparison of induced pluripotent stem cell-derived endothelial cells (iPSC-ECs) from three families with unaffected mutation carriers (UMCs), FPAH patients, and gender-matched controls to investigate this variation. Our analysis identified features of UMC iPSC-ECs related to modifiers of BMPR2 signaling or to differentially expressed genes. FPAH-iPSC-ECs showed reduced adhesion, survival, migration, and angiogenesis compared to UMC-iPSC-ECs and control cells. The "rescued" phenotype of UMC cells was related to an increase in specific BMPR2 activators and/or a reduction in inhibitors, and the improved cell adhesion could be attributed to preservation of related signaling. The improved survival was related to increased BIRC3 and was independent of BMPR2. Our findings therefore highlight protective modifiers for FPAH that could help inform development of future treatment strategies.

Keywords: bone morphogenetic protein receptor 2; cell adhesion; cell signaling; cell survival; endothelial dysfunction; induced pluripotent stem cell-derived endothelial cell; penetrance; pulmonary arterial hypertension; transcriptomic analysis; unaffected mutation carrier.

Figure 1. FPAH iPSC-ECs Show Similar Morphology and BMPR2 Expression but Impaired Adhesion and Survival, Compared with UMCs

(A) Endothelial cells (ECs) differentiated from skin fibroblast derived iPSCs exhibit typical cobblestone morphology, incorporate 3,3′-dioctadecylindocarbocyanine labeled low-density lipoprotein (Dil-LDL, red) and are positive for the EC surface marker VE-Cadherin (green). Blue: DAPI. Scale bar=100μm (top), 50μm (middle) and 10μm (bottom). (B) BMPR2 gene expression quantified by real-time PCR and (C) western immunoblot and densitometric quantification of BMPR2 protein in iPSC-ECs from controls, UMCs and FPAH patients in three families. (D–E) Allele specific BMPR2 expression measured by droplet digital PCR. Probe for wild-type allele was labeled by FAM (blue dots), and for mutant allele by HEX (green dots). Data represented missense mutations from family 1 and 3 and deletion mutation from family 2. (F) iPSC-ECs (10,000/well) were seeded onto 24-well plates either uncoated (plastic) or coated with the indicated matrices and allowed to adhere for one hour. Non-adherent cells in suspension were removed, adherent cells were washed with PBS then stained with Hoechst and imaged. The average number of nuclei was calculated by counting the total number in six random fields per well (10× magnification). (G) iPSC-ECs were exposed to either serum withdrawal overnight (0% FBS) or reoxygenation [Reoxy, 48h in hypoxia, (0.5%O2) followed by 48h at room air]. Apoptosis was measured by Caspase-Glo® 3/7 Assay. Bars represent mean±SD from n=3 families, 3 biological replicates per family. *p<0.05 vs. control, #p<0.05 vs. UMC, one-way ANOVA with Bonferroni post test. See also Figure S1 and Figure S2.

Figure 2. UMC iPSC-ECs Show Preserved pP38 Signaling and Incresed ITGB1, Leading to Improved Cell Adhesion

(A–C) Representative western immnoblots above and densitometric analyses for pP38 in iPSC-ECs from three families one hour after BMP4 (10ng/ml) stimulation. or PBS as vehicle. (D) Representative western immunoblots above and densitometric analyses of pP38 activation following iPSC-EC adhesion to a plastic surface for 15 and 60 min (n=3 families, 3 biological replicates per family). (E) iPSC-ECs from UMC (family 1) were pretreated with the pP38 inhibitor SB239063 for 30 min and adhesion of cells to the indicated substrates for one hour was measured as described in Figure 1F. Bars represent mean±SD from n=3 experiments. (F–G) iPSC-ECs from the FPAH patient were transfected with MKK6 construct (oeMKK6) for four days. oeCtrl: iPSC-ECs transfected with empty construct. (F) Representative western immunoblot and densitometry shows pP38 and β1-integrin (ITGB1) levels following cell adhesion assay (G) performed as in Figure 1F. Representative western immunoblot and densitometry shows integrin β1 (ITGB1) under baseline (H) or cell adhesion condition (I). Upper band is mature form of β1-integrin and lower band is the precursor (Salicioni et al., 2004). (J) UMC iPSC-ECs were treated with control siRNA (siCtrl) or ITGB1 siRNA (siITGB1). Cell adhesion was assessed four days later. (K) UMC iPSC-ECs were pretreated with ITGB1 blocking antibody (Anti-Human CD29, clone Mab13, 20μg/ml) for 60 min. at 37°C, and adhesion of cells to the substrates for one hour was measured as previously described. (L) ITGB1 was over-expressed in the FPAH iPSC-ECs and cell adhesion assessed four days later. Bars represent Mean±SD from n=3 experiments. (A–D, I) *p<0.05 vs. vehicle, #p<0.05 vs. same condition in UMC, two-way ANOVA with Bonferroni post-test. (E, G, J-L) *p<0.05, **p<0.01 vs. vehicle, oeCtrl or siCtrl, t test. (F) *p<0.05 vs. oeCtrl, unpaired t test. (H) #p<0.01 vs. UMC, one-way ANOVA with Bonferroni post-test. See also Figure S3.

Figure 3. Elevated Expression of BMPR2 Activators and Reduced Expression of Inhibitors in BMPR2 UMCs

(A–C) Representative western immunoblots on top and densitometric analyses below showing BMP regulators in iPSC-ECs from controls, UMCs and FPAH patients in the three different families. Bars represent mean±SD from n=3 biological replicates. *p<0.05 vs. control, #p<0.05 vs. UMC by one-way ANOVA followed by Bonferroni multiple comparisons test. (D) Diagram shows the balance of the BMP regulators related to pSMAD1/5-ID1 signaling and pP38 signaling in the iPSC-ECs from UMC. See also Figure S4.

Figure 4. Compensatory pP38 Signaling and Cell Adhesion in UMCs Are Related to BMP Regulators

(A–C) UMC iPSC-ECs were treated with control siRNA (siCtrl) or LRP1 siRNA (siLRP1) and measurements were made four days later. Representative western immunoblots on top and densitometry below for (A) pP38 (B) ITGB1 15min after cell adhesion to plastic; (C) Cell adhesion was measured as described in Figure 1. (D–F) LRP1 was over-expressed in the FPAH iPSC-ECs. Four days later we assessed (D) pP38 (E) ITGB1 15min after adhesion to plastic; (F) Cell adhesion. oeCtrl: iPSC-ECs transfected with empty construct. oeLRP1: iPSC-ECs transfected with construct over-expressing LRP1. (G-H) Co-transfection of LRP1 and ITGB1 or LRP1, ITGB1 and MKK6 was performed in FPAH iPSC-ECs and measurements were made four days later. (G) Representative western immunoblots and densitometry for LRP1, ITGB1 and Flag-MKK6; (H) Cell adhesion was measured as in Figure 1. (I, J) FPAH iPSC-ECs from family 3 were treated with Control or GREM1 siRNA and measurements were made four days later. (I) Representative western immunoblots (left) and densitometry (right) for ITGB1 and pP38 15min after cell adhesion to plastic; (J) Cell adhesion was measured as described in Figure 1. Bars represent mean±SD from n=3 experiments. (A–B, D–E, I) *p<0.05 vs. Vehicle, #p<0.05 vs. siCtrl or oeCtrl Adhesion 15′, two-way ANOVA with Bonferroni post-test. (H) *p<0.05 **p<0.01 vs. Control, #p<0.05 vs. oeCtrl, one-way ANOVA with Bonferroni post-test. (C, F, J) *p<0.05 vs. siCtrl or oeCtrl, t-test. See also Figure S5.

Figure 5. Gene Expression Analysis by RNA-Seq

(A) Heatmap displaying 71 DEGs between controls and UMCs vs. patients (n=11, p<0.01, fold change>2.0). Of these, 41 genes were up-regulated and 30 genes were down-regulated. (B) Quantitative real-time PCR of BIRC3 gene expression in FPAH patients. (C) BIRC3 gene expression reduced by siRNA. (D) Caspase activity of iPSC-ECs from UMCs treated by siBIRC3 in response to serum withdrawal (0%FBS) and reoxygenation after hypoxia (Reoxy). (E) Proposed Model: Compensatory Mechanisms in UMCs. Increased BMPR2 activators LRP1 and/or Caveolin1, and decreased BMPR2 repressors, Gremlin1 that blocks the BMP ligands and/or FKBP12 (FKBP1A) that blocks the BMPR2 type 1 co-receptor (BMPR1), enhance the downstream pP38 signaling pathway, improve β1- integrin recycling. Together with increased total amount of β1-integrin, these features lead to compensatory cell adhesion to a variety of extracellular matrices. Additionally, increased BIRC3 gene expression results in preserved cell survival in response to serum withdrawal and reoxygenation after hypoxia. These features protect the UMCs against the EC dysfunction associated with the development of PAH. (F) Functional enrichment analysis was performed with Reactome pathway database (FDR<0.1) using the 71 genes dysregulated in iPSC-ECs from the FPAH patients. (G–H) Gene-concept network displaying the gene names associated with the top signaling pathways identified in (F). (I, J) Transcription factor (TF) motif enrichment analysis of RNA-Seq: Distribution of the putative occupancy sites of TF was tested within 3kb ± transcription start site of the differentially expressed genes called by RNA-Seq. Analysis shows a unique overrepresentation of the SMAD2 and MAFF motifs in 75% and 55% of genes down-regulated in FPAH iPSC-ECs, and an overrepresentation of SOX8 and SPZ1 motifs in 44.1% of genes up-regulated in FPAH iPSC-ECs. Bars represent mean±SD from n=3 experiments. (B) *p<0.05 vs. UMC, one-way ANOVA with Bonferroni post-test. (C, D) *p<0.05 vs. siCtrl, unpaired t test. See also Figure S6 and Figure S7.

Figure 6. EC Adhesion, Survival, pP38 Signaling and Gene Expression Levels Are Normalized After Gene Editing in iPSC-ECs from Family 1 FPAH Patient

The BMPR2 mutation in iPSCs from the FPAH patient was corrected using CRISPR/Cas9 mediated homology directed repair. (A) Heterozygous point mutation 354T>G is shown as double peak on the diagram. (B) Following correction of the 354T>G mutation, only a single peak was detected. (C) Adhesion of iPSC-ECs in family 1 after gene editing. (D) Apoptosis of iPSC-ECs assessed by Caspase-Glo® 3/7 Assay. (E) Activation of pP38 signaling 1h after BMP4 treatment (10ng/ml) by western immunoblot with densitometry. (F–N) Expression of BMPR2 related genes as well as abnormally expressed genes identified by RNA-Seq were measured in iPSC-ECs. Bars represent mean±SD from n=3 experiments. *p<0.05, **p<0.01 vs. control or vehicle, #p<0.05 vs. UMC, δp<0.05 vs. FPAH, one-way ANOVA with Bonferroni multiple comparisons test (C–D, F–N), and two-way ANOVA with Bonferroni post-test (E).

Figure 7. Angiogenesis and Cell Migration Are Normalized with BMP4 Stimulation After Gene Editing in iPSC-ECs from Family 1 FPAH Patient

(A) Representative images of tube formation of iPSC-ECs from control, UMC, FPAH and FPAH after gene editing (FPAH-CRISPR) with quantitative analysis, indicating the number of tubes formed six hours after seeding cells on growth factor reduced matrigel. (B) Representative images and quantitative data related to wound closure. A scra12tch was applied to 100% confluent iPSC-ECs and migration of the cells toward the wound was imaged at 0 and 12h. (C, D) BMP4 (10ng/ml) was applied when seeding iPSC-ECs. Angiogenesis analyses (C) and migration assay (D) were performed as described in (A and B). (E) Proposed Model for disease development: (i) Control has normal BMPR2 and protective modifiers thus no FPAH or EC vulnerability; (ii) UMC has BMPR2 mutation, the protective modifiers protect against developing FPAH but may still show vulnerability when the level of BMP ligand is low; (iii) FPAH patient has BMPR2 mutation and lacks protective modifiers, conditions that lead to EC dysfunction associated with FPAH; (iv) After correcting the BMPR2 mutation, EC dysfunction related to the development of FPAH is reversed. However, lack of the protective modifiers still result in EC vulnerability when the level of the BMP ligand is low. Bars represent Mean±SD from n=3 experiments. *p<0.05, **p<0.01 vs. Control. #p<0.05 vs. UMC, δp<0.05 vs. FPAH, one-way ANOVA with Bonferroni multiple comparisons test.