Efficient isolation and culture of endothelial cells from venous malformation using the Rho-associated protein kinase inhibitor Y27632

Somatic GJA4 gain-of-function mutation in orbital cavernous venous malformations

Orbital cavernous venous malformation (OCVM) is a sporadic vascular anomaly of uncertain etiology characterized by abnormally dilated vascular channels. Here, we identify a somatic missense mutation, c.121G > T (p.Gly41Cys) in GJA4, which encodes a transmembrane protein that is a component of gap junctions and hemichannels in the vascular system, in OCVM tissues from 25/26 (96.2%) individuals with OCVM. GJA4 expression was detected in OCVM tissue including endothelial cells and the stroma, through immunohistochemistry. Within OCVM tissue, the mutation allele frequency was higher in endothelial cell-enriched fractions obtained using magnetic-activated cell sorting. Whole-cell voltage clamp analysis in Xenopus oocytes revealed that GJA4 c.121G > T (p.Gly41Cys) is a gain-of-function mutation that leads to the formation of a hyperactive hemichannel. Overexpression of the mutant protein in human umbilical vein endothelial cells led to a loss of cellular integrity, which was rescued by carbenoxolone, a non-specific gap junction/hemichannel inhibitor. Our data suggest that GJA4 c.121G > T (p.Gly41Cys) is a potential driver gene mutation for OCVM. We propose that hyperactive hemichannel plays a role in the development of this vascular phenotype.

A xenograft model for venous malformation

Vascular malformations are defects caused by the abnormal growth of the vasculature. Among them, venous malformation (VM) is an anomaly characterized by slow-flow vascular lesions with abnormally shaped veins, typically in sponge-like configuration. VMs can expand over years causing disfigurement, obstruction of vital structures, thrombosis, bleeding, and pain. Treatments have been very limited and primarily based on supportive care, compression garments, sclerotherapy, and/or surgical resection. Sirolimus treatment has recently shown efficacy in some patients with complicated vascular anomalies, including VMs. Activating somatic TIE2 gene mutations have been identified in up to 60% of VMs and PIK3CA mutations have been found in another 25%. Here, we report a xenograft model of VM that reflects the patients' mutation heterogeneity. First, we established a protocol to isolate and expand in culture endothelial cells (VM-EC) from VM tissue or VM blood of nine patients. In these cells, we identified somatic mutations of TIE2, PIK3CA, or a combination of both. Both TIE2 and PIK3CA mutations induced constitutive AKT activation, while TIE2 mutations also showed high MAPK-ERK signaling. Finally, VM-EC implanted into immune-deficient mice generated lesions with ectatic blood-filled channels with scarce smooth muscle cell coverage, similar to patients' VM. This VM xenograft model could be instrumental to test the therapeutic efficacy of Sirolimus in the presence of the different TIE2 or PIK3CA mutations or to test for efficacy of additional compounds in targeting the specific mutated protein(s), thus enabling development of personalized treatment options for VM patients.

Embryonic Stem Cell-Like Subpopulations in Venous Malformation

Background

Venous malformation (VM) consists of a network of ectatic anomalous thin-walled venous channels. A role for an activating TIE2 mutation in the development of the dilated luminal vessels in VM, and its proposed involvement of embryonic stem cells (ESCs), led us to investigate the expression of ESC markers in subcutaneous VM (SCVM) and intramuscular VM (IMVM).

Methods

Formalin-fixed paraffin-embedded sections of SCVM from seven patients and IMVM samples from seven patients were analyzed for the expression of Nanog, pSTAT3, OCT4, SOX2, SALL4, and CD44, using 3,3′-diaminobenzidine (DAB) immunohistochemical (IHC) staining. All these samples did not express lymphatic marker D2-40. NanoString mRNA analysis and RT-PCR were performed on snap-frozen samples of SCVM (n = 3) and IMVM (n = 3) from the respective original cohorts of patients included in DAB IHC staining. To confirm co-expression of two proteins, immunofluorescent (IF) IHC staining on two representative samples of IMVM and SCVM samples from the original cohorts of patients included for DAB IHC staining was performed.

Results

DAB IHC staining demonstrated expression of all of the above ESC markers in both SCVM and IMVM samples. IF IHC staining showed that these markers were localized to the endothelium within these lesions and that Nanog, pSTAT3, SOX2, and CD44 were also expressed by cells outside of the endothelium. NanoString mRNA analysis confirmed transcription activation of pSTAT3, OCT4, and CD44. RT-qPCR confirmed transcription activation of Nanog, SOX2, and SALL4.

Conclusion

Our findings support the presence of two ESC-like subpopulations, one within and one outside of the endothelium, of both SCVM and IMVM. Given that the endothelial ESC-like subpopulation expresses the more primitive marker, OCT4, it is exciting to speculate that they give rise to the non-endothelial subpopulation.