Single Cell RNA Sequencing of Blood Outgrowth Endothelial Cells from Patients with Low Von Willebrand Factor Reveal Multiple Novel Signaling Pathways

Background

The mechanisms that determine low VWF levels in patients with VWF levels between 30-50 IU/dL and no mutations in VWF are poorly understood.

Hypothesis/Objective

We hypothesize that the study of blood outgrowth endothelial cells (BOECs) from individuals with low VWF levels may reveal unique transcriptional profiles that contribute to the low VWF levels seen in these patients.

Methods

BOEC Derivation: Patients with low VWF levels and mucocutaneous bleeding (MCB) (30-50 IU/dL) were enrolled in an IRB-approved study. The mononuclear layer from whole blood was isolated and plated onto collagen coated plates. After extended incubation, BOECs were validated by visual inspection and flow cytometry.

Endothelial Transcriptional Characterization: A total of 9 cells lines including those from individuals with low VWF and HUVEC and BOECs from individuals with normal VWF levels as control were assayed via single cell RNA sequencing. Bioinformatic analysis included generalized transcriptional expression, Ingenuity Pathway Analysis (IPA), and expression of VWF. RNA-sequencing expression data was filtered according to a standardized algorithm. Cells that were defined as monocytes (TYROBP expression > 2 copies) were excluded. Following monocyte exclusions, cells were determined to be of endothelial origin if they demonstrated the presence of transcripts of PECAM1, CDH5, ROBO4, ESAM, TIE1, or NOTCH4 as previously described by Butler et al. (Cell Reports 2016).

Results

BOEC Derivation: A total of eight BOEC lines were generated, 6 from individuals with MCB and VWF levels between 30-50 IU/dL (5:1 female: male ratio, age range 11-54 years) and 2 from healthy controls (2 female, age range 22-39 years) with normal VWF levels and no symptoms of MCB.

Transcriptional Profiling of single endothelial cells from Low VWF Individuals: A 3D T-SNE plot that assesses unbiased differences in gene expression profiling was generated with each cell line represented by a different color (Figure 1A) demonstrating that individual cell lines have significant differences in their underlying transcriptional profiling.

VWF Expression in Low VWF Samples: Overall expression of VWF was significantly decreased in low VWF BOEC samples (5.341 transcripts/cell) vs. control (9.076 transcripts/cell) ECs (figure 1B), P<0.0001. Further, histogram and mixed model (multiple gaussian) analysis of VWF expression revealed changes in generalized expression of VWF in Low VWF BOECs compatible with multiple populations of VWF-expressing BOECs, demonstrating cell mosaicism within each sample.

IPA Analysis of Low VWF vs Control BOECs: IPA analysis demonstrated 64 pathways with a z-score difference >1 in Low VWF BOECs when compared to control BOECs (table 1), including multiple signaling pathways such as PI3kinase and AKT as well as several cytoskeleton pathways.

Conclusions

Single cell RNA sequencing of Low VWF BOECs reveal significant differences in transcriptional profiling when compared to control endothelial cell lines (control BOECs + HUVEC). BOECs from individuals with Low VWF levels demonstrate significantly lower VWF transcript expression than the control endothelial cells. Interestingly, BOECs from low VWF patients show significant differences in VWF transcript number within cells from the same individual demonstrating a degree of mosaicism previously described in murine endothelial cells. Finally, there are multiple cellular pathways that are differentially regulated in Low VWF BOECs as compared to control endothelial cells.

View largeDownload slide

View largeDownload slide

Close modal