ABSTRACT
TMEM16F (also known as ANO6), a Ca2+-activated lipid scramblase (CaPLSase) that dynamically disrupts lipid asymmetry, plays a crucial role in various physiological and pathological processes, such as blood coagulation, neurodegeneration, cell–cell fusion and viral infection. However, the mechanisms through which it regulates these processes remain largely elusive. Using endothelial cell-mediated angiogenesis as a model, here we report a previously unknown intracellular signaling function of TMEM16F. We demonstrate that TMEM16F deficiency impairs developmental retinal angiogenesis in mice and disrupts angiogenic processes in vitro. Biochemical analyses indicate that the absence of TMEM16F enhances the plasma membrane association of activated Src kinase. This in turn increases VE-cadherin phosphorylation and downregulation, accompanied by suppressed angiogenesis. Our findings not only highlight the role of intracellular signaling by TMEM16F in endothelial cells but also open new avenues for exploring the regulatory mechanisms for membrane lipid asymmetry and their implications in disease pathogenesis.
INTRODUCTION
TMEM16F (also known as ANO6) is a Ca2+-activated lipid scramblase (CaPLSase) that rapidly collapses membrane phospholipid asymmetry by bi-directionally and non-selectively transporting all major phospholipids down their concentration gradients, leading to surface exposure of phosphatidylserine (PS) (Bevers and Williamson, 2016; Le et al., 2019a; Suzuki et al., 2010). Despite the growing list of its functions in blood coagulation (Fujii et al., 2015; Schmaier et al., 2023; Suzuki et al., 2010; Yang et al., 2012; Yu et al., 2021), red blood cell function (Liang et al., 2024), cell fusion (Braga et al., 2021; Zhang et al., 2020a), plasma membrane repair (Wu et al., 2020), bone development (Ehlen et al., 2013; Ousingsawat et al., 2015), viral infection (Younan et al., 2018; Zaitseva et al., 2017) and neurodegeneration (Cui et al., 2023; Soulard et al., 2020; Zhang et al., 2020b), the role of TMEM16F in endothelial biology has just started emerging (Schmaier et al., 2023; Yu et al., 2021; Zhang et al., 2020a). We have recently shown that TMEM16F deficiency impairs blood vessel development in the murine placentas (Zhang et al., 2020a). An angiogenesis defect was also observed in the zebrafish embryos with TMEM16F morpholino knockdown (Delcourt et al., 2015). However, it is unclear whether the observed defects are directly derived from TMEM16F deficiency in endothelial cells (ECs) and how TMEM16F regulates vascular development.
Here, we report that TMEM16F CaPLSase is functionally expressed in human ECs and plays an important role in regulating neonatal retinal angiogenesis in vivo and human umbilical vein endothelial cell (HUVEC) tube formation in vitro. By investigating its role in angiogenesis, we uncovered a previously unknown intracellular signaling function of this CaPLSase. TMEM16F deficiency in HUVECs enhances Src phosphorylation and membrane association, which further phosphorylates and downregulates VE-cadherin, leading to subsequent defects in angiogenesis.
RESULTS AND DISCUSSION
TMEM16F is a major CaPLSase in human endothelial cell lines
Despite the implications of TMEM16F in EC functions (Delcourt et al., 2015; Schmaier et al., 2023; Yu et al., 2021; Zhang et al., 2020a), functional quantification of its CaPLSase activity in ECs has been missing. To fill this knowledge gap, we used siRNAs to silence TMEM16F in HUVECs and human aortic endothelial cells (HAECs), two widely used primary human EC models. Consistent with the nearly complete knockdown of TMEM16F transcripts and proteins (Fig. 1A–C; Fig. S1A,B), our fluorescence imaging-based CaPLSase assay, which reports surface exposure of PS through TMEM16F actions (Fig. 1D) (Le et al., 2019a,b), showed that TMEM16F siRNA largely abolished the ionomycin-induced time-dependent accumulation of fluorescently conjugated Annexin V (AnV), a PS-binding probe (Fig. 1E–G, Fig. S1C–E). Utilizing the moonlighting function of TMEM16F as a Ca2+-activated non-selective ion channel (Le et al., 2019a; Liang and Yang, 2021; Yang et al., 2012; Yu et al., 2015), our whole-cell patch clamp recording further demonstrated that TMEM16F siRNA eliminates the Ca2+- and voltage-activated outward rectifying TMEM16F current (Yang et al., 2012) (Fig. 1H,I). Taken together, our quantitative measurement establishes that TMEM16F is functionally expressed in ECs and is a major CaPLSase on the EC plasma membrane.
TMEM16F is functionally expressed in HUVECs. (A) qRT-PCR of TMEM16F in control and TMEM16F knockdown HUVECs (n=7, from five biological replicates). mRNA expression is normalized to GAPDH expression. (B) Representative western blots of TMEM16F in HUVECs with control or TMEM16F siRNA knockdown. (C) Densitometry quantifications of TMEM16F and loading control β-actin (n=4, from four biological replicates). (D) Schematic of the fluorescence-based scrambling assay. PS, phosphatidylserine; AnV, Annexin V. (E) Representative images of Ca2+ and AnV in control (left) and TMEM16F knockdown (right) HUVECs stimulated with 2.5 μM ionomycin. (F,G) Quantifications of the time course (F) or the maximum fluorescence intensity of AnV at 10 min post ionomycin stimulation (G) for control siRNA (n=37) and TMEM16F (n=46) (from four biological replicates). Each dot represents AnV signals from one cell. (H) Representative currents recorded in control or TMEM16F siRNA knockdown HUVECs. The currents were elicited by a voltage step protocol from −100 mV to +160 mV with a 20 mV increment. Holding potential was set at −60 mV. (I) Current-voltage (I-V) relationship of currents recorded in H (n=13 for control siRNA and n=14 for TMEM16F siRNA, from two biological replicates). Data are presented as mean±s.e.m. ****P<0.0001 (unpaired two-tailed t-test). a.u., arbitrary units.
TMEM16F is functionally expressed in HUVECs. (A) qRT-PCR of TMEM16F in control and TMEM16F knockdown HUVECs (n=7, from five biological replicates). mRNA expression is normalized to GAPDH expression. (B) Representative western blots of TMEM16F in HUVECs with control or TMEM16F siRNA knockdown. (C) Densitometry quantifications of TMEM16F and loading control β-actin (n=4, from four biological replicates). (D) Schematic of the fluorescence-based scrambling assay. PS, phosphatidylserine; AnV, Annexin V. (E) Representative images of Ca2+ and AnV in control (left) and TMEM16F knockdown (right) HUVECs stimulated with 2.5 μM ionomycin. (F,G) Quantifications of the time course (F) or the maximum fluorescence intensity of AnV at 10 min post ionomycin stimulation (G) for control siRNA (n=37) and TMEM16F (n=46) (from four biological replicates). Each dot represents AnV signals from one cell. (H) Representative currents recorded in control or TMEM16F siRNA knockdown HUVECs. The currents were elicited by a voltage step protocol from −100 mV to +160 mV with a 20 mV increment. Holding potential was set at −60 mV. (I) Current-voltage (I-V) relationship of currents recorded in H (n=13 for control siRNA and n=14 for TMEM16F siRNA, from two biological replicates). Data are presented as mean±s.e.m. ****P<0.0001 (unpaired two-tailed t-test). a.u., arbitrary units.
TMEM16F deficiency leads to defective retinal angiogenesis in mice
To evaluate the angiogenic role of TMEM16F in vivo, we quantified the vascular network in the retinas from postnatal day 5 (P5) wild-type (WT) and Tmem16F global knockout (KO) mice. Our IB4 staining of the whole mount retinas showed that the vasculature coverage of Tmem16F KO mice is significantly reduced compared to that of their WT littermates (Fig. 2A,B, top panels, 2C). Our analysis of retinal vasculature complexity (performed by a researcher unaware of the genotype) (Fig. 2A,B, middle and bottom panels; see Materials and Methods) further demonstrates that the average number of vascular loops per field of view is markedly lower (Fig. 2D) and the average hole areas of the vascular loop are significantly bigger (Fig. 2E) in Tmem16F KO retinas. Our analysis of the retinal developmental angiogenesis thus confirms that Tmem16F plays an important role in angiogenesis in vivo.
TMEM16F deficiency suppresses angiogenesis in vivo. (A,B) IB4 (green) staining of retinal whole-mounts from P5 wild-type (WT, A) and Tmem16F knockout (KO, B) littermates. White boxes indicate enlarged area shown in the middle panels. The bottom panels illustrate processed enlarged area used for the quantification of loop numbers and hole areas. White areas are defined as holes and vasculature (black lines) encircling the holes as loops. (C–E) Quantification of retinal vasculature coverage (C), average hole areas (D) and number of loops (E) calculated from four random fields of each retina. Data are presented as mean±s.e.m. Each dot represents data from one animal (n=7 pairs of WT and Tmem16F KO littermates). *P<0.05, **P<0.01, ***P<0.001 (unpaired two-tailed t-test).
TMEM16F deficiency suppresses angiogenesis in vivo. (A,B) IB4 (green) staining of retinal whole-mounts from P5 wild-type (WT, A) and Tmem16F knockout (KO, B) littermates. White boxes indicate enlarged area shown in the middle panels. The bottom panels illustrate processed enlarged area used for the quantification of loop numbers and hole areas. White areas are defined as holes and vasculature (black lines) encircling the holes as loops. (C–E) Quantification of retinal vasculature coverage (C), average hole areas (D) and number of loops (E) calculated from four random fields of each retina. Data are presented as mean±s.e.m. Each dot represents data from one animal (n=7 pairs of WT and Tmem16F KO littermates). *P<0.05, **P<0.01, ***P<0.001 (unpaired two-tailed t-test).
TMEM16F deficiency in ECs compromises angiogenesis in vitro
Given the Tmem16F-deficient mouse line is a constitutive KO, we used in vitro HUVEC tube formation to assess the endothelial-specific role of TMEM16F in angiogenesis. Our time-lapse imaging reveals that the HUVECs transfected with control siRNA successfully form a robust, three-dimensional network of tube-like structures on the Matrigel matrix within 4 h of seeding (Fig. 3A top; Movie 1). There were 4.56±0.44 loops/mm2 remaining at 24 h post-seeding (mean±s.e.m.; Fig. 3C). In stark contrast, the tube network formed by the TMEM16F siRNA transfected HUVECs was dramatically unstable, as evidenced by the large loops and a significantly decreased loop density (1.74±0.13 loops/mm2) 24 h post seeding (Fig. 3A bottom, 3C; Movie 1). Close examination of the tube network at different time points reveals that the tubes formed by the TMEM16F-deficient cells start to break apart at ∼8 h post seeding, and the separated cells tend to aggregate into large cell clusters at tube junctions (Fig. 3A, bottom), which results in a marked reduction of loop density (Fig. 3C).
TMEM16F deficiency in HUVEC impairs tube formation in vitro. (A,B) Representative images of tube formation for control and TMEM16F knockdown HUVECs (A), and TMEM16F knockdown HUVECs overexpressing wild-type (WT) and D703R loss-of-function mTmem16F (B). Yellow arrowheads indicate thin and stretched tubes. (C) Quantification of loop numbers at 24 h after seeding (control and TMEM16F knockdown, n=20 each; WT and D703R mTmem16F overexpressed in TMEM16F knockdown HUVECs, n=19 each; from three biological replicates). Each dot represents data from one well of tube formation μ-Slide. Data are presented as mean±s.e.m. n.s.; non-significant; **P<0.01, ***P<0.001, ****P<0.0001 (one-way ANOVA with Tukey's multiple comparisons test).
TMEM16F deficiency in HUVEC impairs tube formation in vitro. (A,B) Representative images of tube formation for control and TMEM16F knockdown HUVECs (A), and TMEM16F knockdown HUVECs overexpressing wild-type (WT) and D703R loss-of-function mTmem16F (B). Yellow arrowheads indicate thin and stretched tubes. (C) Quantification of loop numbers at 24 h after seeding (control and TMEM16F knockdown, n=20 each; WT and D703R mTmem16F overexpressed in TMEM16F knockdown HUVECs, n=19 each; from three biological replicates). Each dot represents data from one well of tube formation μ-Slide. Data are presented as mean±s.e.m. n.s.; non-significant; **P<0.01, ***P<0.001, ****P<0.0001 (one-way ANOVA with Tukey's multiple comparisons test).
To further establish the causal relationship between TMEM16F CaPLSase and EC-mediated angiogenesis, we reintroduced mouse Tmem16F (mTmem16F) wild-type (WT) and D703R, a loss-of-function mutation that abolishes Ca2+ sensing (Le et al., 2019a; Yang et al., 2012), into TMEM16F siRNA-treated HUVEC cells. Our fluorescence CaPLSase assay demonstrated that mTmem16F WT but not D703R mutant successfully rescued Ca2+-induced PS exposure in TMEM16F knockdown HUVECs (Fig. S2A–C). Consistent with the functional rescue, reintroducing mTmem16F WT also alleviated the tube formation defect in TMEM16F siRNA-treated HUVECs (Fig. 3B) with a significantly higher loop density (3.15±0.11 loops/mm2) 24 h post-seeding (Fig. 3C). By contrast, mTmem16F-D703R overexpression failed to show obvious rescuing effect (Fig. 3B,C). Our in vitro tube formation experiments thus further support that endothelial TMEM16F CaPLSase plays an important role in angiogenesis.
TMEM16F deficiency promotes Src and VE-cadherin phosphorylation
The instability of the tubes formed by the TMEM16F knockdown HUVECs (Fig. 3A,B; Movie 1) suggests that the EC cell–cell junctions might be impaired by TMEM16F deficiency. The weakened junctions could fail to hold the tubes together (Fig. 3, 8–24 h), leading to unstable tube networks. To assess whether TMEM16F regulates the adherens junctions of ECs, we measured the expression of VE-cadherin, an endothelial adhesion protein that is required at the cell–cell junctions to stabilize and maintain the newly formed vessels (Bentley et al., 2014). Our western blot results reveal that TMEM16F siRNA knockdown significantly downregulates VE-cadherin protein expression (Fig. 4B,C). Interestingly, our quantitative real-time RT-PCR (qRT-PCR) results indicate that VE-cadherin transcript levels were enhanced by TMEM16F knockdown (Fig. 4A), suggesting that TMEM16F CaPLSase likely regulates VE-cadherin expression through post-translational modification.
TMEM16F deficiency in HUVECs disrupts intracellular Src-VE-cadherin signaling. (A) qRT-PCR of VE-cadherin in control and TMEM16F knockdown HUVECs (n=7 from five biological replicates). The gene expression is normalized to GAPDH expression. **P<0.01 (unpaired two-tailed t-test). (B,C) Representative western blot of anti-VE-cadherin (B) and densitometry quantifications (n=9 from none biological replicates) (C). (D,E) Representative western blot of anti-Tyr685 phosphorylated and anti-total VE-cadherin (D) and densitometry quantifications (n=5 from five biological replicates) (E). (F,G) Representative western blot anti-Tyr416 phosphorylated and anti-total Src (F) and densitometry quantifications (n=4 from four biological replicates) (G). (H,I) TMEM16F knockdown increases Src phosphorylation without or with 100 ng/ml of VEGF stimulation. (H) Proximity ligation assay with antibodies against Src and p-SFK(Tyr416) (pSrc/Src, red) in HUVECs without (top) or with (bottom) 100 ng/ml of VEGF stimulation for 5 min. Cell junctions were stained for VE-cadherin (green) and nuclei with DAPI (blue). Enlarged views of the dotted boxes are shown on the right. (I) Mean fluorescence intensity quantifications of junctional PLA signals (n=6 from four biological replicates). (J) FITC–dextran (40 kDa, 1 mg/ml) Transwell assay (n=9 from three biological replicates). **P<0.01, ***P<0.001, ****P<0.0001 [unpaired two-tailed t-test (A,C,E,G); two-way ANOVA with Tukey's multiple comparisons test (I,J)]. All data are presented as mean±s.e.m. a.u., arbitrary units. (K) Cartoon illustration of the intracellular signaling role of TMEM16F in angiogenesis. TMEM16F-mediated PS exposure limits Src membrane association and phosphorylation, maintaining VE-cadherin membrane expression and normal angiogenesis (top). TMEM16F deficiency increases Src membrane association and phosphorylation and promotes its activation, which downregulates VE-cadherin and leads to defective angiogenesis (bottom).
TMEM16F deficiency in HUVECs disrupts intracellular Src-VE-cadherin signaling. (A) qRT-PCR of VE-cadherin in control and TMEM16F knockdown HUVECs (n=7 from five biological replicates). The gene expression is normalized to GAPDH expression. **P<0.01 (unpaired two-tailed t-test). (B,C) Representative western blot of anti-VE-cadherin (B) and densitometry quantifications (n=9 from none biological replicates) (C). (D,E) Representative western blot of anti-Tyr685 phosphorylated and anti-total VE-cadherin (D) and densitometry quantifications (n=5 from five biological replicates) (E). (F,G) Representative western blot anti-Tyr416 phosphorylated and anti-total Src (F) and densitometry quantifications (n=4 from four biological replicates) (G). (H,I) TMEM16F knockdown increases Src phosphorylation without or with 100 ng/ml of VEGF stimulation. (H) Proximity ligation assay with antibodies against Src and p-SFK(Tyr416) (pSrc/Src, red) in HUVECs without (top) or with (bottom) 100 ng/ml of VEGF stimulation for 5 min. Cell junctions were stained for VE-cadherin (green) and nuclei with DAPI (blue). Enlarged views of the dotted boxes are shown on the right. (I) Mean fluorescence intensity quantifications of junctional PLA signals (n=6 from four biological replicates). (J) FITC–dextran (40 kDa, 1 mg/ml) Transwell assay (n=9 from three biological replicates). **P<0.01, ***P<0.001, ****P<0.0001 [unpaired two-tailed t-test (A,C,E,G); two-way ANOVA with Tukey's multiple comparisons test (I,J)]. All data are presented as mean±s.e.m. a.u., arbitrary units. (K) Cartoon illustration of the intracellular signaling role of TMEM16F in angiogenesis. TMEM16F-mediated PS exposure limits Src membrane association and phosphorylation, maintaining VE-cadherin membrane expression and normal angiogenesis (top). TMEM16F deficiency increases Src membrane association and phosphorylation and promotes its activation, which downregulates VE-cadherin and leads to defective angiogenesis (bottom).
A prominent post-translational pathway that destabilizes VE-cadherin junctions is Src kinase phosphorylation at tyrosine (Tyr)685 of VE-cadherin, which induces VE-cadherin internalization (Vestweber, 2008; Wallez et al., 2007) and degradation (Vincent et al., 2004). To evaluate whether TMEM16F deficiency in HUVECs increases VE-cadherin phosphorylation by Src kinase, we examined VE-cadherin phosphorylation levels. Our western blot results indicate that TMEM16F knockdown in HUVECs robustly increased VE-cadherin phosphorylation at Tyr685 (Fig. 4D,E), suggesting that Src activity is upregulated. Indeed, the levels of the active form of Src kinase [phosphorylated at Tyr416; p-Src(Tyr416)], is significantly enhanced in TMEM16F knockdown HUVECs (Fig. 4F,G). As the anti-p-Src antibody could potentially recognize other Src family kinases, we utilized a proximity ligation assay (PLA) to detect Src kinase-specific phosphorylation. By using anti-Src and -p-Src(Tyr416) antibodies conjugated to the PLUS or MINUS PLA probes, only the phosphorylation in proximity to Src would be detected. Consistent with the western blot results (Fig. 4F,G), our PLA experiments revealed that the TMEM16F knockdown HUVECs have significantly enhanced p-Src(Tyr416) signal at the juxtamembrane region compared to the control knockdown cells at both basal conditions and with VEGF stimulation (Fig. 4H,I), suggesting that TMEM16F deficiency increases p-Src(Tyr416) association to the plasma membrane.
VE-cadherin downregulation and phosphorylation at Tyr685 are associated with a permeability increase in ECs (Orsenigo et al., 2012; Smith et al., 2020; Wessel et al., 2014). To examine whether TMEM16F knockdown leads to the functional consequence of permeability enhancement, we conducted a Transwell permeability assay. We found that TMEM16F siRNA significantly increased FITC–dextran (40 kDa) permeability under both basal and VEGF-stimulated conditions (Fig. 4J), further supporting that knockdown of TMEM16F in HUVEC impairs VE-cadherin-regulated EC functions.
Taken together, we established that there is functional expression of TMEM16F in ECs and uncovered a previously unknown intracellular signaling role for the TMEM16F CaPLSase in controlling angiogenesis (Fig. 4K, top). Our biochemical evidence indicates that TMEM16F deficiency in ECs promotes Src kinase phosphorylation and its plasma membrane association, which subsequently enhances VE-cadherin phosphorylation, leading to its internalization and degradation (Fig. 4K, bottom). Downregulation of VE-cadherin likely destabilizes the EC junctions, resulting in impaired stability of newly formed blood vessels and the observed angiogenesis defects.
Research on lipid scramblases has been primarily focused on their activities to promote PS exposure and subsequent PS-mediated extracellular events, such as the recruitment of PS-binding clotting factors for blood coagulation (Schmaier et al., 2023; Yang et al., 2012; Yu et al., 2021), PS receptors for phagocytosis (Zhang et al., 2020b), the ADAM proteases for proteolytic cleavage (Bleibaum et al., 2019; Sommer et al., 2016) and the viral envelop proteins for infection (Zaitseva et al., 2017). Distinct from these studies, here we reveal that TMEM16F CaPLSase controls intracellular Src–VE-cadherin signaling in ECs and subsequent angiogenesis, likely through regulation of Src membrane association and phosphorylation (Fig. 4K, top). Src kinase is anchored to the plasma membrane through myristylation and interaction with anionic phospholipids through a cluster of positively charged residues (Murray et al., 1998; Patwardhan and Resh, 2010; Sigal et al., 1994). It is likely that during angiogenesis, TMEM16F-mediated PS externalization reduces the negative charges at the cytoplasmic leaflet of the plasma membrane, preventing Src from excessive membrane association, phosphorylation and activation. This subsequently leads to reduced VE-cadherin phosphorylation and more VE-cadherin surface expression, which stabilizes cell–cell junction and facilitates new blood vessel formation (Fig. 4K, top). TMEM16F deficiency, on the other hand, prevents PS externalization and retains the highly negatively charged cytoplasmic leaflet. This can increase Src membrane association, phosphorylation and activation, which subsequently cause enhanced phosphorylation and downregulation of VE-cadherin, resulting in angiogenesis deficiency (Fig. 4K, bottom). Indeed, our PLA results show that more activated Src kinases are recruited to the plasma membrane in TMEM16F knockdown ECs, whereas much less phosphorylated Src is present at the juxtamembrane region of control knockdown ECs (Fig. 4H,I). Consistent with this, recent studies also show that in natural killer cells and macrophages, scramblase-mediated membrane lipid redistribution disrupts the membrane association of Src family protein tyrosine kinases, resulting in their translocation from the plasma membrane to the cytosol (Wu et al., 2021; Yeung et al., 2006).
A recent study on flippases, active lipid transporters that counteract scramblases by internalizing the exposed PS back to the cytoplasmic leaflet (Bevers and Williamson, 2016; Nagata et al., 2020), also hints at the importance of PS translocation on the plasma membrane in regulating Src phosphorylation (Zhang et al., 2019). The authors discovered that the loss of TMEM30A, a positive regulatory subunit of flippases (Bryde et al., 2010; Nagata et al., 2020), inhibits Src kinase phosphorylation in human retinal ECs (Zhang et al., 2019). In contrast to TMEM16F CaPLSase deficiency, which prevents PS exposure (Fig. 4K, bottom), TMEM30A deficiency promotes PS exposure on the outer leaflet (Nagata et al., 2020). The fact that TMEM30A- and TMEM16F-deficient ECs show opposite effects on Src phosphorylation at the activating Tyr416, supports that PS translocation plays an important role in regulating Src kinase activity and signaling. Intriguingly, both conditions, loss of TMEM30A and TMEM16F, result in impaired in vitro tube formation and in vivo retinal angiogenesis. This suggests that Src signaling must be correctly balanced to appropriately induce the correct angiogenic cues. Moreover, membrane lipid asymmetry might affect other signaling molecules required for the angiogenic process. Indeed, angiogenesis depends on highly coordinated spatiotemporal control of various signaling pathways that involve VEGF, VEGF receptors, VE-cadherin and Src kinase at different angiogenic stages (Bentley and Chakravartula, 2017; Eilken and Adams, 2010; Jeong et al., 2017). Future investigations are needed to further dissect the roles of scramblases and flippases in regulating these molecular pathways.
In addition to flippases, phospholipid scramblase 1 (PLSCR1) is implicated in facilitating Src activation (Nanjundan et al., 2003). However, it is important to note that mounting evidence indicates that, despite its name, PLSCR1 is not a bona fide scramblase (Fadeel et al., 1999; Zhou et al., 2000, 2002). How PLSCR1 regulates Src signaling warrants further investigation. Nevertheless, the study on flippases and our current study on TMEM16F scramblase collectively demonstrate the importance of lipid translocation on the EC plasma membrane in directly controlling intracellular signaling through PS translocation-mediated changes in Src–VE-cadherin signaling and angiogenesis.
Dysregulation of angiogenesis is a hallmark of highly prevalent diseases such as cancer, cardiovascular and inflammatory diseases, and ocular disorders (Carmeliet and Jain, 2000). However, the success of anti-angiogenesis therapies, which predominantly target pro-angiogenic factors, is curtailed by the activation of alternative angiogenic pathways and resistance mechanisms (Eelen et al., 2020). To overcome these limitations, it is crucial to further understand angiogenesis mechanisms and discover new pro-angiogenic pathways. Interestingly, PS exposure on the endothelium is frequently observed in conditions like cancer (Ran et al., 2002; Zhang et al., 2017), inflammatory bowel disease (Zhang et al., 2021) and choroidal neovascularization (Li et al., 2015). Future studies are needed to dissect whether TMEM16F CaPLSase participates in the excessive PS exposure under these disease conditions. Interestingly, masking exposed PS with anti-PS antibodies or PS-binding proteins is known to have anti-angiogenesis effects (Li et al., 2015; Ran et al., 2002; Zhang et al., 2017, 2021). Targeting PS exposure in ECs thus presents a promising strategy for designing more effective anti-angiogenesis therapies. Our identification of TMEM16F as an important EC CaPLSase and discovery of its key role in regulating blood vessel formation unlock a potential anti-angiogenesis target for combating diseases with abnormal angiogenesis.
MATERIALS AND METHODS
Mice
The Tmem16F knockout (KO) mouse line has been reported previously (Yang et al., 2012; Zhang et al., 2020a). Mouse handling and usage were carried out in strict compliance with protocols approved by the Institutional Animal Care and Use Committee at Duke University in accordance with National Institutes of Health guidelines.
Retinal flat-mount and staining
Enucleated eyes from WT or Tmem16f KO littermate pairs with both sexes at postnatal day 5 were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) overnight at 4°C with gentle shaking. The retinas were isolated from enucleated eyes and were blocked and permeabilized with 1% BSA and 0.5% Triton X-100 in PBS overnight at 4°C. The retinal vasculatures were then stained with fluorescein-conjugated isolectin B4 (IB4, 1:100 dilution, Vector laboratories, #FL-1201) overnight at 4°C with gentle shaking. After PBS washing, the retinas were flat-mounted and imaged with a Zeiss 780 inverted confocal microscope. We utilized a custom MATLAB script (available at https://github.com/superdongping/Retina_blood_vessel) to measure the vascular area, hole area and loop count in these subregions, leveraging the built-in function ‘imbinarize’ in MATLAB. The area of retinal vasculature holes and the number of loops per field of view were quantified. Processed samples of the vascular area can be viewed in Fig. 2A. For each retina, four square subregions were randomly chosen by an investigator who was not aware of the genotype of the retina.
Cell culture and siRNA transfection
HUVECs and HAECs (obtained from Lonza and authenticated by the Duke Cell Culture Facility) were cultured in EGM-2 BulletKit (Lonza, #CC-3162) in a humidified incubator at 37°C and 5% CO2-95% air. For imaging assays, HUVECs and HAECs cells were seeded on coverslips coated with poly-L-lysine (Sigma-Aldrich, #P2636) in 24-well plates. For western blot, HUVECs and HAECs cells were seeded in 10-cm cell culture dishes. For siRNA transfection, TMEM16F was knocked down by transfecting with non-targeting Silencer Negative Control No. 2 siRNA (Invitrogen, #AM4637) or TMEM16F SMARTpool siRNA (a pre-mixed pool of four siRNAs from Dharmacon Research, #M-003867-01-0020) using Lipofectamine RNAiMAX Transfection Reagent (Invitrogen, #13778075) following the manufacturer's instructions (volume ratio RNAiMAX: siRNA= 2:1). Fresh medium was changed the next day, and cells were cultured for another 24 h before conducting other assays.
RNA isolation and qRT-PCR
The RNeasy Mini Kit (Qiagen #74104) was utilized for total RNA extraction from control and TMEM16F knockdown HUVECs, followed by cDNA synthesis using SuperScript IV reverse transcriptase (Invitrogen) and random hexamers according to the manufacturer's protocol. The levels of TMEM16F and VE-cadherin expression were detected via quantitative PCR (qPCR) conducted on a StepOnePlus Real-time PCR system (Applied Biosystems), employing the synthesized cDNA, Power SYBR Green PCR Mix (Applied Biosystems), and primers. Gene expression levels were normalized to GAPDH levels.
Human GAPDH primers, forward, 5′-GGAGCGAGATCCCTCCAAAAT-3′; reverse, 5′-GGCTGTTGTCATACTTCTCATGG-3′; TMEM16F primers, forward, 5′-TGTCCCCGATTTGGGATCACT-3′; reverse, 5′-CGTATGCTTGTCTTTTCCTCCT-3′; VE-cadherin primers, forward, 5′-GCACCAGTTTGGCCAATATA-3′; reverse, 5′-GGGTTTTTGCATAATAAGCAGG-3′.
Fluorescence imaging of Ca2+ and PS exposure
The fluorescence-based scrambling assay was undertaken as reported previously (Fig. 1D) (Le et al., 2019a,b). Briefly, Calbryte 520 AM (AAT Bioquest, #20701) was used to monitor Ca2+ dynamics, and CF 594-tagged AnV (Biotium, #29011) was used to detect exposed PS. ECs were stained with 1 μM Calbryte 520 AM for 10 min before transferring to AnV solution (1:150 in Hank's balanced salt solution, Gibco, #14025-092). Time-lapse imaging was performed before and after 2.5 μM ionomycin stimulation with a Zeiss 780 inverted confocal microscope to monitor Ca2+ dynamics and AnV binding. A custom MATLAB code (available at https://github.com/YZ299/matlabcode/blob/main/matlabcode.m) was used to quantify the PS signal intensities as reported previously (Le et al., 2019a,b).
Mouse Tmem16F lentiviral overexpression
To overexpress WT and D703R mTmem16F in HUVECs, we used lentivirus carrying the pLVX mTmem16F plasmid (Addgene #62554). The D703R mutant was generated in this plasmid by site-directed mutagenesis using primers from IDT and subsequently confirmed by Sanger sequencing (Azenta). Lentivirus constructs of WT and D703R mTmem16F were packaged into lentivirus by transfecting HEK-293 T cells with lentiviral vectors, pMD2.G, and psPAX2 (Addgene #12259 and #12260) using lipofectamine 2000. Lentiviral particle-containing medium was collected and passed through a 0.45 µm syringe filter (VWR, #28145- 481) 48 and 72 h after transfection. HUVECs were then transduced with lentiviral supernatant and polybrene (10 µg/ml) and selected with puromycin.
In vitro tube formation assays
Prior to cell seeding, 10 μl of Matrigel (Corning, # 356231, lot #2213003) was added to each well of the µ-Slide 15 Well 3D (ibidi, #81506) without touching the wall of the wells to avoid unlevel matrix formation. Matrigel was then allowed to polymerize by incubating the µ-Slide at 37°C for 45 min to 1 h. After Matrigel solidified, 50 μl of 5.5×105 cells/ml HUVECs in EGM2 were added to each well. For time-lapse imaging, images were captured 30 min after seeding at a 10-min interval using the Zeiss Axio Observer Z1 microscope. For loop number quantification, µ-Slides were imaged 24 h after seeding using an Olympus IX73 or Olympus IX83 inverted epi-fluorescent microscope with a 4× objective. Loop numbers were quantified from the whole well with ImageJ and calculated as loop numbers/mm2 by dividing the µ-Slide inner well growth area.
Immunoblotting
After trypsinization and PBS washing, HUVECs with control or TMEM16F siRNA knockdown were harvested in lysis radioimmunoprecipitation assay (RIPA) lysate buffer (Thermo Fisher Scientific, #89900) supplemented with 1× Halt™ Protease and Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific, #78440). 20–30 μg of proteins were mixed with 1× Laemmli buffer (Bio-Rad, #161-0747) and 2-mercaptoethanol and loaded onto SDS-PAGE gels. After separation, the Trans-Blot Turbo Transfer system (Bio-Rad) was used to transfer the proteins from the SDS-PAGE gels to PVDF membranes. The membranes were blocked with 5% bovine serum albumin (BSA, Tocris, #5217) in TBS supplemented with 0.1% Tween 20 (TBST) at room temperature for at least an hour. Following blocking, membranes were incubated in anti-VE-cadherin (1:1000 dilution, Santa Cruz Biotechnology, #sc-9989), anti-phospho-VE-cadherin (Tyr685) (1:1000 dilution, ECM bioscience, #CP1981), anti-Src (1:1000 dilution, ABclonal, #A11707), anti-phospho-Src (Tyr-416) (1:1000 dilution, ABclonal, #AP0452) or anti-β-actin (1:10,000 dilution, Sigma, #A1978) overnight at 4°C. Membranes were washed with TBST the next day and incubated with HRP-conjugated goat anti-rabbit-IgG secondary antibody (Sigma, #A0545) or anti-mouse-IgG secondary antibody (Sigma, #A1917) for 1 h at room temperature before detecting with the Clarity Western ECL Substrate Kit (Bio-Rad, #170-5060). Raw images of immunoblots are included in Fig. S3.
In situ proximity ligation assays
Proximity ligation assays were undertaken using the Duolink PLA fluorescence kit (Sigma) following a protocol described in (Sjöberg et al., 2023). Briefly, HUVECs were seeded at 5×104 cells/well in 8-well glass slides followed by siRNA knockdown. Cells were starved for 2 h, stimulated with 100 ng/ml VEGF (Peprotech, #100-20-50UG) for 5 min, then fixed with 2% PFA. Cells were permeabilized for 0.1% Triton X-100 in TBS for 20 min, blocked with Duolink blocking solution for 1 h, and incubated overnight at 4°C in 1:100 rabbit anti-p-SFK(Tyr416) (Cell Signaling Technology, #2101 s) and mouse anti-Src (Abcam, #ab231081). PLA was performed with PLUS and MINUS PLA probes targeting mouse and rabbit primary antibodies. Cell junctions were visualized by counterstaining with goat anti-VE-cadherin antibody (1:100 dilution, R&D systems, #AF1002).
Transwell permeability assay
HUVECs with control or TMEM16F siRNA knockdown were seeded at 3×104 cells/Transwell insert (6.5 mm Transwell with 0.4 µm Pore Polyester Membrane, Corning, #3470) and were cultured until a confluent monolayer was formed. In VEGF stimulation groups, cells were pre-treated with 50 ng/ml of VEGF for 30 min. Permeability was measured by adding 1 mg/ml of 40 kDa FITC–dextran (Sigma, #FD40) to the top wells. 50 μl of FITC–dextran signals from bottom wells were collected after 2.5 h and measured for fluorescence at 520 nm excited at 492 nm with a spectrophotometer (SpectraMax M5, Molecular Devices).
Electrophysiology
TMEM16F currents were recorded in whole-cell configuration using an Axopatch 200B amplifier (Molecular Devices) and the pClamp software package (Molecular Devices). Glass pipettes were pulled from borosilicate capillaries (Sutter Instruments) and fire-polished using a microforge (Narishge) to reach a resistance of 2–3 MΩ. The pipette solution (internal) contained (in mM): 140 CsCl, 1 MgCl2, and 10 HEPES, plus 1 CaCl2 to avoid the long-delay activation of TMEM16F (Liang and Yang, 2021; Zhang et al., 2022). The bath solution contained 140 CsCl, 10 HEPES and 5 EGTA. Adjusted to pH 7.3 for both sides with CsOH. Currents were recorded for ∼2 min after whole-cell formation to ensure the activation of TMEM16F (Liang and Yang, 2021; Zhang et al., 2022).
Statistical analysis
All statistical analyses were performed using Prism software (GraphPad). Unpaired two-tailed Student's t-tests were used for single comparisons between two groups, and one-way or two-way ANOVA (with Tukey's multiple comparisons test) was used for multiple comparisons. Sample number (n) values are indicated in the results section and figure legends. All data are presented as the mean±standard error of the mean (s.e.m.).
Software and code
All relevant codes supporting the present study are available at GitHub: https://github.com/superdongping/Retina_blood_vessel and https://github.com/YZ299/matlabcode/blob/main/matlabcode.m. Figure cartoons were generated using BioRender. The raw data are shared upon reasonable request.
Acknowledgements
The authors thank Ssu-Yu Chen, Yasaman Setayeshpour, Jen-Tsan Chi, Pernilla Martinsson, Yangyang Su, Yindi Ding, Cole McCutcheon, Carolyn Coyne, Timothy McCord, Xi Chen, Christopher Kontos, and Vann Bennett for their help during the course of this work.
Footnotes
Author contributions
Conceptualization: K.Z.S., T.L., H.Y.; Methodology: K.Z.S., T.L., P.L., A.J.L., P.K., L.C., H.Y.; Software: P.D.; Validation: K.Z.S., T.L.; Formal analysis: K.Z.S., T.L., P.L.; Investigation: K.Z.S., T.L., P.L.; Resources: H.Y.; Writing - original draft: K.Z.S.; Writing - review & editing: K.Z.S., H.Y.; Visualization: K.Z.S., T.L.; Supervision: H.Y.; Project administration: H.Y.; Funding acquisition: H.Y.
Funding
This work was supported by the National Institutes of Health (DP2GM126898 and 1R35GM153196-01 awarded to H.Y.) and an American Heart Association Predoctoral fellowship (#20PRE35120162, awarded to T.L.). Deposited in PMC for release after 12 months.
Data availability
All relevant data can be found within the article and its supplementary information.
References
Competing interests
The authors declare no competing or financial interests.