Upregulated MicroRNA-21 Drives the Proliferation of Lymphatic Malformation Endothelial Cells by Inhibiting PDCD4

Lymphatic malformations (LMs) are congenital anomalies of the lymphatic vasculature that affect approximately 1:4,000 live births and consist of dilated lymphatic channels surrounded by hypertrophic stroma, resulting in a compressible lesion with a mass effect. Somatic activating single nucleotide variants (SNVs) in PIK3CA, which encodes the p110α catalytic subunit of phosphoinositide 3-kinase, have been identified in approximately 80% of LMs and are localized to LM endothelial cells (LM-ECs) (Boscolo et al., 2015Boscolo E. Coma S. Luks V.L. Greene A.K. Klagsbrun M. Warman M.L. et al.AKT hyper-phosphorylation associated with PI3K mutations in lymphatic endothelial cells from a patient with lymphatic malformation.Angiogenesis. 2015; 18: 151-162Crossref PubMed Scopus (87) Google Scholar). Targeted therapeutics such as sirolimus have been used to suppress the hyperactivation of phosphoinositide 3-kinase/protein kinase B/mTOR signaling in patients with LMs and can reduce disease burden (Adams et al., 2016Adams D.M. Trenor 3rd, C.C. Hammill A.M. Vinks A.A. Patel M.N. Chaudry G. et al.Efficacy and safety of sirolimus in the treatment of complicated vascular anomalies.Pediatrics. 2016; 137e20153257Crossref Scopus (506) Google Scholar). Although promising, symptom relief is variable, and significant side effects have been reported (Rössler et al., 2021Rössler J. Baselga E. Davila V. Celis V. Diociaiuti A. El Hachem M. et al.Severe adverse events during sirolimus "off-label" therapy for vascular anomalies.Pediatr Blood Cancer. 2021; 68e28936Crossref PubMed Scopus (16) Google Scholar). The mechanisms underlying how PIK3CA SNVs drive the LM phenotype are unclear, and whether epigenetic dysregulation contributes to LM pathogenesis has not been investigated. MicroRNAs (miRNAs) regulate lymphangiogenesis and endothelial cell function (Chen et al., 2016Chen J. Zhu R.F. Li F.F. Liang Y.L. Wang C. Qin Y.W. et al.MicroRNA-126a directs lymphangiogenesis through interacting with chemokine and Flt4 signaling in zebrafish.Arterioscler Thromb Vasc Biol. 2016; 36: 2381-2393Crossref PubMed Scopus (36) Google Scholar) and drive cellular proliferation through the regulation of phosphoinositide 3-kinase (Catanzaro et al., 2018Catanzaro G. Besharat Z.M. Miele E. Chiacchiarini M. Po A. Carai A. et al.The miR-139-5p regulates proliferation of supratentorial paediatric low-grade gliomas by targeting the PI3K/AKT/mTORC1 signalling.Neuropathol Appl Neurobiol. 2018; 44: 687-706Crossref PubMed Scopus (32) Google Scholar). Although miRNA dysfunction has been identified in other vascular anomalies such as hemangiomas (Strub et al., 2016Strub G.M. Kirsh A.L. Whipple M.E. Kuo W.P. Keller R.B. Kapur R.P. et al.Endothelial and circulating C19MC microRNAs are biomarkers of infantile hemangioma.JCI Insight. 2016; 1e88856Crossref PubMed Scopus (36) Google Scholar), the contribution of miRNAs to LM pathogenesis is unknown. In this study, we identified the upregulation of the oncogenic miRNA-21 in LMs containing PIK3CA SNVs and conducted further studies to determine whether miR-21 inhibition may be a viable adjuvant molecular therapy for LMs. LM and normal skin/subcutaneous tissue samples were collected during resection from 18 pediatric patients after written informed consent and approval by the University of Arkansas for Medical Sciences Institutional Review Board (Institutional Review Board number 114012). Macrocystic components of the LMs, including the cyst wall, were dissected free of surrounding tissue and collected as LM samples. Once the entirety of the lesion was removed, skin including subcutaneous tissue and subdermal lymphatics was collected from an adjacent site as control tissue. Digital droplet RT-PCR detected PIK3CA SNVs (p.E542K, E545K, and H1047R) in 13 of the 18 LM samples. RT-PCR of miRNAs known to regulate phosphoinositide 3-kinase and/or normal lymphangiogenesis showed that miR-21 was significantly upregulated in the LM tissue samples containing PIK3CA mutations compared with that in normal skin/subcutaneous tissue from the same patients, despite similar lymphatic endothelial density as shown by equivocal podoplanin mRNA expression (Figure 1a). In situ hybridization probes targeting miR-21 showed the concentration of miR-21 in the lymphatic endothelium of two patient samples analyzed (Figure 1b). Owing to the limitations of this technique and its inability to quantitate endothelial miR-21 expression, lymphatic endothelial cells were isolated using podoplanin-conjugated magnetic beads from both LM tissue and normal control tissue from two patients, and their levels of miR-21 were compared using RT-PCR (Figure 1c). miR-21 was enriched in podoplanin-positive cells isolated from LM tissue compared with that in podoplanin-negative cells, and miR-21 expression was upregulated in LM-ECs compared with that in lymphatic endothelial cells isolated from control tissues. To confirm the successful isolation of PIK3CAmut LM-ECs, purified LM-ECs were cultured as previously described (Wu et al., 2015Wu J.K. Kitajewski C. Reiley M. Keung C.H. Monteagudo J. Andrews J.P. et al.Aberrant lymphatic endothelial progenitors in lymphatic malformation development.PLoS One. 2015; 10e0117352Google Scholar) and analyzed at successive passages by digital droplet PCR, which showed a PIK3CAmut variant allele frequency of 50% after 1−2 passages (Figure 1d). A proliferation assay comparing PIK3CAmut LM-ECs with normal human lymphatic endothelial cell (HDLECs) lines showed significantly higher proliferative rates in PIK3CAmut cells (Figure 1e). Taken together, these experiments show increased miR-21 expression and higher proliferative rates in PIK3CAmut lymphatic endothelial cells than in the wild type. miR-21 is an oncogenic miRNA that silences the expression of several tumor suppressors, and its upregulation has been shown to increase cellular proliferation in multiple cell types (reviewed by Arghiani and Matin, 2021Arghiani N. Matin M.M. miR-21: A key small molecule with great effects in combination cancer therapy.Nucleic Acid Ther. 2021; 31: 271-283Crossref PubMed Scopus (15) Google Scholar). To determine whether the upregulation of miR-21 in LM-ECs resulted in the inhibition of endothelial tumor suppressors, we compared the expression of five miR-21 targets (PDCD4 [Xu et al., 2017Xu X. Jiao X. Song N. Luo W. Liang M. Ding X. et al.Role of miR-21 on vascular endothelial cells in the protective effect of renal delayed ischemic preconditioning.Mol Med Rep. 2017; 16: 2627-2635Crossref PubMed Scopus (12) Google Scholar], SPRY-1 [Zhuang et al., 2020Zhuang H. Wang H. Yang H. Li H. Exosome-encapsulated MicroRNA-21 from esophageal squamous cell carcinoma cells enhances angiogenesis of human umbilical venous endothelial cells by targeting SPRY1 [retracted in Cancer Manag Res 2022;14:499−500].Cancer Manag Res. 2020; 12: 10651-10667Crossref PubMed Scopus (10) Google Scholar], TIMP-3 [Hu et al., 2016Hu J. Ni S. Cao Y. Zhang T. Wu T. Yin X. et al.The angiogenic effect of microRNA-21 targeting TIMP3 through the regulation of MMP2 and MMP9.PLoS One. 2016; 11e0149537Google Scholar], RhoB [Sabatel et al., 2011Sabatel C. Malvaux L. Bovy N. Deroanne C. Lambert V. Gonzalez M.L. et al.MicroRNA-21 exhibits antiangiogenic function by targeting RhoB expression in endothelial cells.PLoS One. 2011; 6e16979Crossref PubMed Scopus (214) Google Scholar], and PTEN [Luo et al., 2017Luo M. Tan X. Mu L. Luo Y. Li R. Deng X. et al.MiRNA-21 mediates the antiangiogenic activity of metformin through targeting PTEN and SMAD7 expression and PI3K/AKT pathway.Sci Rep. 2017; 7: 43427Crossref PubMed Scopus (59) Google Scholar]) by RT-PCR and western blot in HDLECs and LM-ECs (Figure 2a and b). PDCD4, SPRY-1, TIMP-3, and RhoB mRNA and protein levels were all significantly lower in LM-ECs than in HDLECs, whereas PTEN mRNA expression was equivocal. To determine whether the upregulation of miR-21 was responsible for the attenuation of PDCD4/SPRY-1/TIMP-3/RhoB protein expression observed in LM-ECs, HDLECs and LM-ECs were transfected with miR-21 antisense RNA (mir-21−), miR-21 mimic RNA (miR-21+), and scrambled miR-21 RNAs as controls (Figure 2c). In HDLECs, inhibition of miR-21 led to an approximate 10% increase in PDCD4 protein levels, whereas overexpression of miR-21 reduced PDCD4 protein levels by approximately 70% (Figure 2d). Conversely, in LM-ECs, miR-21 inhibition increased PDCD4 protein levels by approximately 60%, whereas miR-21 overexpression modestly reduced PDCD4 protein levels (∼20%) (Figure 2d). Transient transfection with miR-21 inhibitors or mimics did not affect SPRY-1/TIMP-3/RhoB protein levels in either cell type (data not shown). Because the miR-21−PDCD4 axis is known to regulate endothelial cell proliferation (Hua et al., 2020Hua R. Zhang X. Li W. Lian W. Liu Q. Gao D. et al.Ssc-miR-21-5p regulates endometrial epithelial cell proliferation, apoptosis and migration via the PDCD4/AKT pathway.J Cell Sci. 2020; 133cs248898PubMed Google Scholar), we compared the proliferative rates of HDLECs and LM-ECs in the presence of miR-21 inhibition or overexpression (Figure 2e). Overexpression of miR-21 increased the proliferative rates of both HDLECs and LM-ECs. In the presence of miR-21 inhibition, LM-EC proliferation was dramatically reduced compared with the effect of miR-21 inhibition in HDLECs. These results suggest that in PIK3CAmut LM-ECs that express abnormally high levels of miR-21, inhibition of miR-21 can reverse their hyperproliferative phenotype. In summary, this study identifies the contribution of an aberrantly expressed miRNA in LM-ECs that drives their hyperproliferative phenotype and is reversible with miRNA inhibition, which to our knowledge has not been reported. In addition, this study shows the downregulation of the tumor suppressors PDCD4, SPRY-1, TIMP-3, and RhoB in LM-ECs, potentially identifying additional molecular targets for treatment. miRNA-based therapeutics are emerging as methods of targeting aberrantly expressed genes in specific cell types. A recent study showed that transdermal nanocarrier delivery of a miR-21 mimic in a murine wound healing model could inhibit PDCD4 expression and increase cellular proliferation (Wang et al., 2020Wang S.Y. Kim H. Kwak G. Jo S.D. Cho D. Yang Y. et al.Development of microRNA-21 mimic nanocarriers for the treatment of cutaneous wounds.Theranostics. 2020; 10: 3240-3253Crossref PubMed Scopus (22) Google Scholar). A similar delivery of miR-21 inhibitors to LMs may have the opposite effect, representing a viable and minimally invasive method of reducing LM-EC proliferation in vivo. All human subjects signed written informed consent giving permission for tissue collection, use of their biologic specimens, and cellular and genetic analyses, according to Institutional Review Board protocols (University of Arkansas for Medical Sciences Institutional Review Board number 114012). All original data can be obtained by contacting the corresponding author at [email protected]. No datasets were generated or analyzed during this study. Ravi W. Sun: http://orcid.org/0000-0002-0494-7547 Haihong Zhang: http://orcid.org/0000-0002-4327-9027 Syed J. Mehdi: http://orcid.org/0000-0002-2388-0480 Gresham T. Richter: http://orcid.org/0000-0003-3423-1213 Hayden H. Bowman: http://orcid.org/0000-0002-0322-28355 Jessica Sifford: http://orcid.org/0000-0002-8858-5706 Chelsea Smith: http://orcid.org/ 0000-0002-4186-998X Alexander K. Burnett: http://orcid.org/0000-0002-4551-9194 Alexander Layman: http://orcid.org/0000-0002-2423-4106 Charity L. Washam: http://orcid.org/0000-0001-5761-9304 Stephanie D. Byrum: http://orcid.org/0000-0002-1783-3610 James T. Bennett: http://orcid.org/0000-0003-2843-5594 Dana M. Jensen: http://orcid.org/0000-0001-5981-5886 Victoria Dmyterko: http://orcid.org/0000-0001-5532-3868 Jonathan A. Perkins: http://orcid.org/0000-0003-0181-3997 Carrie J. Shawber: http://orcid.org/0000-0003-2654-3559 June K. Wu: http://orcid.org/0000-0003-2298-8040 Graham M. Strub: http://orcid.org/0000-0002-2585-4956 The authors state no conflict of interest. The authors thank Alan Tackett, Robert Eoff, and Jerry Ware for mentorship; Arkansas Children’s Vascular Biology Program for providing samples; the American Society of Pediatric Otolaryngology for funding; The Lyon Family for funding; COBRE Center for Translational Pediatric Research (P20GM121293) for funding; The Arkansas Biosciences Institute (AWD53593) and the University of Arkansas for Medical Sciences Systems Pharmacology and Toxicology Training Program (T32GM106999) for funding; James Suen for surgical sample contribution; Mary Dornhoffer for manuscript editing; and Jeffery Flowers for research administrative support. Conceptualization: GMS; Investigation: RWS, HZ, SJM, HHB, JS, CS, AKB, AL, JTB, DMJ, VD; Methodology: CS, JKW, GMS; Resources: GTR, CLW, SDB, JAP, CJS, JKW; Software: CLW, SDB; Supervision: GMS; Writing – Review and Editing: GMS, RWS; Writing - Original Draft Preparation: RWS All human subjects signed written informed consent giving permission for tissue collection, use of their biologic specimens, and cellular and genetic analyses, according to Institutional Review Board protocols (University of Arkansas for Medical Sciences Institutional Review Board number 114012). Lymphatic malformation (LM) tissue, cyst fluid, and normal tissue biopsies (harvested from adjacent, unaffected subcutaneous tissue) were collected from patients undergoing surgical resections. Whole-tissue specimens were either flash frozen in liquid nitrogen and stored at −80 °C or transported in normal saline at 4 °C for cell isolation. Hematoxylin staining was performed using the H&E stain kit (Abcam, Waltham, MA). Immunohistochemical staining was performed on paraffin sections with anti-podoplanin (ab77854, 1:40; Abcam). A Zeiss AXIO Imager.M2 microscope (Zeiss, Oberkochen, Germany) was used to obtain images with an Olympus DP73 digital camera (Olympus, Tokyo, Japan). Images were taken using CellSens entry 1.17 software (Olympus). Formalin-fixed paraffin-embedded LM sections were used for miR-21 in situ hybridization using miRCURY LNA miRNA ISH Optimization Kits (formalin-fixed paraffin-embedded), according to the manufacturer’s instructions (Qiagen, Germantown, MD) and as previously reported (Strub et al., 2016Strub G.M. Kirsh A.L. Whipple M.E. Kuo W.P. Keller R.B. Kapur R.P. et al.Endothelial and circulating C19MC microRNAs are biomarkers of infantile hemangioma.JCI Insight. 2016; 1e88856Crossref PubMed Scopus (36) Google Scholar). Images were collected using an Olympus BX43 microscope (Olympus) with an Infinity 3S digital camera (Teledyne Lumenera, Ottawa, Canada). Images were taken using Infinity analyze 7 software (Teledyne Lumenera). A multiplex digital droplet PCR assay to detect the four most common PIK3CA-activating variants was previously developed (Rowlands et al., 2019Rowlands V. Rutkowski A.J. Meuser E. Carr T.H. Harrington E.A. Barrett J.C. Optimisation of robust singleplex and multiplex droplet digital PCR assays for high confidence mutation detection in circulating tumour DNA.Sci Rep. 2019; 912620Crossref Scopus (56) Google Scholar; Wu et al., 2015Wu J.K. Kitajewski C. Reiley M. Keung C.H. Monteagudo J. Andrews J.P. et al.Aberrant lymphatic endothelial progenitors in lymphatic malformation development.PLoS One. 2015; 10e0117352Google Scholar). Samples were run in quadruplicate. For positive controls, gene blocks of 170 base pair were designed for each variant and associated wild type sequence (Integrative DNA Technologies, Coralville, IA). Data were analyzed with QuantaSoft software. For whole-tissue specimens, the RNeasy Plus Universal Mini Kit (Qiagen) was used to isolate total RNA. The RNeasy Plus Micro Kit (Qiagen) was used to isolate total RNA containing microRNA from cultured cells. RNA concentration was measured using the NanoDrop ND-1000 (Thermo Fisher Scientific, Waltham, MA). The real-time PCR amplification was performed on a QuantStudio 6 Flex Real-Time PCR System (Thermo Fisher Scientific). For endogenous controls, GAPDH (Hs03929097_g1) and RNU48 (assay identification 001006) were used for mRNA and microRNA assays, respectively. Human lymphatic endothelial cell lines (n = 6) were purchased (PromoCell, Heidelberg, Germany). These cells are characterized by the company using FACS for podoplanin and CD31 and were maintained in culture in our laboratory for less than 6 months. Lymphatic malformation endothelial cells were isolated as previously described (Zenner et al., 2021Zenner K. Jensen D.M. Cook T.T. Dmyterko V. Bly R.A. Ganti S. et al.Cell-free DNA as a diagnostic analyte for molecular diagnosis of vascular malformations.Genet Med. 2021; 23: 123-130Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar). For cells used for digital droplet PCR analysis, lymphatic malformation endothelial cells were isolated as previously described from whole LM tissue or LM cyst fluid, grown to ∼80% confluence, and then split 1:10 for each subsequent passage. At each passage, a cell pellet was collected, washed in PBS, and frozen at −80 °C for digital droplet PCR analysis. Total protein was extracted from tissues by T-PER tissue protein extraction reagent and from cell pellets by RIPA Lysis and Extraction Buffer (Thermo Fisher Scientific). Protein was loaded onto NuPAGE Novex 4-12% Bis-Tris Protein Gels (Thermo Fisher Scientific) and transferred to polyvinylidene fluoride membranes. The pictures were taken with Image Quant LAS 4000 (GE Healthcare, Little Chalfont, United Kingdom), and image results were analyzed with Image Quant TL 7.0 software (GE Healthcare). hsa-miR-21 mimic (assay identification MC10206) and its negative control (catalog number 4464058) and hsa-miR-21 inhibitor (assay identification MH10206) and its negative control (catalog number 4464076) were transfected into cells using Lipofectamine RNAiMAX Transfection Reagent (Thermo Fisher Scientific) for 24 hours following the manufacturer’s instructions. For proliferation assays, cells were seeded onto 96-well plates at 1,500 cells per well. Cell proliferation was measured at 24-hour intervals using the Cell Counting Kit-8 (WST-8) assay (Dojindo Molecular Technologies, Rockville, MD) using a standard curve for each time point. For qRT-PCR experiments, two-sample unpaired t-tests assuming unequal variances were calculated using GraphPad Prism 9.4.0 (GraphPad Software, San Diego, CA). For wound healing assays, the area under the curve of each migration experiment was calculated using Biotek Cytation5 software. For cell proliferation assays, unpaired t-tests assuming unequal variances were performed with Microsoft Excel 2019 with Data Analysis ToolPak 2.0. For all calculations, P-values < 0.05 were considered statistically significant and are either denoted directly on the figure or by asterisks (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001).