Shear-induced Notch-Cx37-p27 axis arrests endothelial cell cycle to enable arterial specification

Abstract

Establishment of a functional vascular network is rate-limiting in embryonic development, tissue repair and engineering. During blood vessel formation, newly generated endothelial cells rapidly expand into primitive plexi that undergo vascular remodeling into circulatory networks, requiring coordinated growth inhibition and arterial-venous specification. Whether the mechanisms controlling endothelial cell cycle arrest and acquisition of specialized phenotypes are interdependent is unknown. Here we demonstrate that fluid shear stress, at arterial flow magnitudes, maximally activates NOTCH signaling, which upregulates GJA4 (commonly, Cx37) and downstream cell cycle inhibitor CDKN1B (p27). Blockade of any of these steps causes hyperproliferation and loss of arterial specification. Re-expression of GJA4 or CDKN1B, or chemical cell cycle inhibition, restores endothelial growth control and arterial gene expression. Thus, we elucidate a mechanochemical pathway in which arterial shear activates a NOTCH-GJA4-CDKN1B axis that promotes endothelial cell cycle arrest to enable arterial gene expression. These insights will guide vascular regeneration and engineering.

Fig. 1

NOTCH signaling regulates shear-induced endothelial cell quiescence. a Expression of several NOTCH signaling pathway effectors were significantly altered in whole-transcriptome analysis of HUVEC exposed to 6 h FSS (vs. 6 h Static), as were previously characterized flow-responsive genes KLF2, KLF4, and COX. b Cleavage of NOTCH receptor intracellular domain (NICD) was significantly increased with 1 h of 12 dynes/cm² FSS (mean densitometry ± SEM; n = 3 for all groups; Students’ t-test: p = 0.03), and blocked by 10 µM DAPT. Uncropped blots presented in Supplementary Fig. 1. c HEY1 and HEY2 transcript levels were elevated with 16 h FSS (mean relative mRNA expression ± SEM vs. Static; n = 3 for all groups; Students’ t-test: p = 0.003 (HEY1), p = 0.005 (HEY2)). d EdU incorporation, a measure of DNA synthesis and indicator of proliferation, was reduced over 18 h exposure to FSS in control cells, and this response was abrogated by inhibition of NOTCH signaling (via 10 µM DAPT) (Colors: Hoechst (cyan), EdU (green); scale bar = 100 µm; mean % EdU-positive nuclei ± SEM for 8 ROI; representative of n = 3 experiments; Students’ t-test: p = 0.09 (12 h), p = 0.0006 (18 h))

Fig. 2

Shear-sensitive NOTCH upregulates GJA4 expression. a Exposure of HUVEC to 6 h FSS significantly upregulated GJA4 protein, in parallel with sustained increase in NICD (mean densitometry ± SEM; n = 5 (Static, 3 h FSS), n = 4 (6 h FSS); one-way ANOVA: p = 0.04 (NICD), p = 0.004 (GJA4), asterisks indicate p < 0.05 in post hoc t-test vs. Static). Uncropped blots presented in Supplementary Fig. 3. b si-GJA4 and DAPT (alone or in combination) abrogated flow-induced suppression of EdU incorporation at 12–18 h shear (mean % EdU-positive nuclei ± SEM for 8 ROI; representative of n = 3 experiments; one-way ANOVA: p < 0.0001 (12 h FSS, 18 h FSS), asterisks indicate p < 0.05 in post hoc t-test). c GJA4 levels were reduced by treatment of confluent HUVEC with 8 h or 24 h DAPT (mean relative mRNA expression ± SEM vs. DMSO; n = 3 (8 h), n = 5 (24 h); individual values plotted where n < 5; Students’ t-test: p = 0.01 (8 h), p = 0.006 (24 h)). d 24 h DAPT also reduced GJA4 protein levels in confluent HUVEC. Uncropped blots presented in Supplementary Fig. 3. e Seeding of sub-confluent HUVEC onto recombinant DLL4 increased GJA4 mRNA levels (mean relative mRNA expression ± SEM vs. PBS; n = 5 for all groups; Students’ t-test: p = 0.03 (3 h), p = 0.002 (6 h), p < 0.0001 (9 h)). f GJA4 protein and NICD were increased following seeding of sub-confluent HUVEC onto Dll4 for 8 h. Uncropped blots presented in Supplementary Fig. 3. g Significant GJA4 promoter fold enrichment was detected in samples assayed by ChIP using anti-RBP-Jκ, compared to IgG controls. This effect was abolished with DAPT treatment (mean fold enrichment ± SEM vs IgG; n = 3 for all groups; Students t-test: p = 0.007 (anti-RBP-Jκ)). h si-NOTCH1, but not si-NOTCH4, abolished DLL4-induced upregulation of GJA4 mRNA at 6 h following cell seeding (mean relative mRNA expression ± SEM vs. PBS; n = 7 (si-Ctrl, si-NOTCH1), n = 8 (si-Ctrl, si-NOTCH4); Students’ t-test: p < 0.0001 (si-Ctrl), p = 0.02 (si-NOTCH4))

Fig. 3

GJA4 regulates vascular remodeling. a GJA4 expression was predominantly detected in the remodeling plexus at the R-M boundary, in addition to expression in SMA-invested arteries. (Colors: PECAM1 (red), SMA (cyan), GJA4 (green); scale bar: 250 µm (lower magnification), 100 µm (higher magnification); symbols: a = artery, v = vein). b At higher magnification, GJA4 expression was observed in vessels beyond the R-M boundary, including in large vessels not invested with SMA, as well as in adjacent smaller vessels. (Colors: PECAM1 (red), SMA (cyan), GJA4 (green); Scale bar = 50 µm; Symbols: a = artery). c GJA4 was absent from tip cells at the vascular edge. (Colors: PECAM1 (red), GJA4 (green); Scale bar = 50 µm). d In P6 retinas lacking GJA4 (Gja4 ^−/−) or NOTCH (WT + DAPT, or Notch1 ^iECKO−Cre+), the developing vasculature was hyperdense in comparison to wild-type (Gja4 ^+/+) tissues. (Colors: IB4 (red); scale bar = 100 µm; symbols: a = artery, v = vein). e Expression of NOTCH target genes HES1, HEY1, and HEY2 were significantly reduced in FACS-sorted endothelial cells (PECAM1+/PTPRC−, commonly CD31+/CD45–) with DAPT treatment (mean relative mRNA expression ± SEM vs. vehicle-treated controls; n = 4 for all groups; paired Students’ t-test: p = 0.0002 (Hes1), p = 0.04 (Hey1), p = 0.047 (Hey2)). f Vessel (PECAM1+ or IB4+) area was significantly increased in NOTCH-inhibited (WT + DAPT or Notch1 ^iECKO−Cre+) and GJA4-deleted (Gja4 ^−/−) mice compared to associated controls (mean % vessel area per retina ± SEM vs. control; n = 6 (WT + Veh, WT + DAPT), n = 3 (Notch1 ^{iECKO−Cre−}), n = 11 (Notch1 ^iECKO−Cre+), n = 4 (Gja4 ^+/+), n = 10 (Gja4 ^−/−); Students’ t-test: p = 0.001 (WT + Veh vs. WT + DAPT), p = 0.004 (Notch1 ^{iECKO−Cre−} vs. Notch1 ^iECKO−Cre+), p = 0.0006 (Gja4 ^+/+ vs. Gja4 ^−/−)). g Branchpoint number was also significantly increased in WT + DAPT, Notch1 ^iECKO−Cre+, and Gja4 ^−/− animals (mean % branchpoint number per retina ± SEM vs. control; n = 7 (WT + Veh), n = 9 (WT + DAPT), n = 3 (Notch1 ^{iECKO−Cre−}), n = 11 (Notch1 ^iECKO−Cre+), n = 4 (Gja4 ^+/+), n = 9 (Gja4 ^−/−) Students’ t-test: p = 0.03 (WT + Veh vs. WT + DAPT), p = 0.049 (Notch1 ^{iECKO−Cre−} vs. Notch1 ^iECKO−Cre+), p = 0.002 (Gja4 ^+/+ vs. Gja4 ^−/−)). h GJA4 expression was significantly reduced in remodeling vessels of DAPT-treated WT animals, as well as in Notch1 ^iECKO−Cre+ animals. (Colors: PECAM1 (red), GJA4 (white); scale bar = 50 µm; symbols: a = artery)

Fig. 4

NOTCH regulates endothelial cell cycle progression via GJA4. a In endothelial (CD31+/CD45−) cells isolated from WT + DAPT and Gja4 ^−/− P6 retinas, there was a decreased proportion in G0 or G1, and increased proportion in S/G2/M, compared to WT controls (mean difference in cell cycle phase % ± SEM vs. WT (Gja4 ^+/+); n = 6 (WT), n = 12 (WT + DAPT), n = 3 (Gja4 ^−/−); individual values plotted where n < 5; one-way ANOVA: p = 0.005 (S/G2/M), asterisks indicate p < 0.05 in post hoc t-tests). b pH3+ endothelial cells were more numerous in remodeling vessels of WT + DAPT and Gja4 ^−/− P6 retinas, compared to controls, and mitotic endothelial cells were mostly found at the R-M boundary (Colors: IB4 (red), ERG (blue), pH3 (magenta); scale bar: 100 µm (low magnification), 50 µm (high magnification); symbols: a = artery, r = remodeling, m = mature; mean number of pH3+/ERG+ nuclei ± SEM vs control; n = 8 (Gja4 ^+/+, Gja4 ^−/−), n = 6 (WT), n = 4 (WT + DAPT); Students’ t-test: p = 0.006 (WT + Veh vs. WT + DAPT, 200–300 µm), p = 0.03 (WT + Veh vs. WT + DAPT, 400–500 µm), p = 0.04 (WT + Veh vs. WT + DAPT. 500–600 µm), p = 0.04 (WT + Veh vs. WT + DAPT, 700–800 µm), p = 0.04 (Gja4 ^+/+ vs. Gja4 ^−/−, 200–300 µm, p = 0.02 (Gja4 ^+/+ vs. Gja4 ^−/−, 400–500 µm), p = 0.005 (Gja4 ^+/+ vs. Gja4 ^−/−, 500–600 µm), p = 0.005 (Gja4 ^+/+ vs. Gja4 ^−/−, 700–800 µm), p = 0.03 (Gja4 ^+/+ vs. Gja4 ^−/−, 900–1000 µm)). c The nuclei of endothelial cells in G1 were marked by CDT1-mOrange in remodeling vessels of Cdt1-mOrange+;Gja4 ^+/+ reporter mice. CDT1-mOrange signal in endothelial cells of remodeling vessels in Gja4 ^−/− reporter mice was greatly reduced, compared to WT controls. (Colors: IB4 (red), ERG (blue), pH3 (magenta), CDT1 (green); scale bar: 100 µm (low magnification), 50 µm (high magnification); symbols: a = artery). d In HUVEC, DAPT decreased the proportion of cells in G1 and increased the proportion in S/G2/M; constitutive GJA4 expression (via lenti-Gja4) abolished these cell cycle effects of DAPT (mean difference in cell cycle % ± SEM vs. DMSO; n = 9 (DMSO), n = 7 (DAPT), n = 8 (DAPT+lenti-Gja4); one-way ANOVA: p = 0.04 (G1), p = 0.03 (S/G2/M), asterisks indicate p < 0.05 in post hoc t-tests)

Fig. 5

GJA4 regulates endothelial cell cycle arrest via CDKN1B. a CDKN1B mRNA expression was significantly reduced in Gja4 ^−/− P6 retinal endothelial cells (CD31+/CD45−) (mean relative mRNA expression ± SEM vs. Gja4 ^+/− littermate controls; n = 3 (Cdk4, Gja4 ^−/−; Cdk6, Gja4 ^−/− Ccne1, Gja4 ^+/−; Ccne2, Gja4 ^+/−), n = 4 (Cdk4, Gja4 ^+/−; Cdk6, Gja4 ^+/−; Ccnb1; Cdkn1a, Gja4 ^−/−), n = 5 (Ccne1, Gja4 ^−/−), n = 6 (Ccne2, Gja4 ^−/−; Cdkn1a, Gja4 ^+/−; Ccnd1, Gja4 ^−/−), n = 8 (Cdk2; Tp53, Gja4 ^−/−), n = 9 (Ccnd1, Gja4 ^+/−; Cdkn1b; Tp53, Gja4 ^−/−); individual values plotted where n < 5). b CDKN1B expression was detected in Gja4 ^+/+, but not Gja4 ^−/−, endothelial cells of remodeling retinal vessels. (Colors: PECAM1 (red), CDKN1B (green); scale bar: 50 µm). c CDKN1B protein was elevated in HUVEC by 16 h FSS, and this effect was abolished by si-GJA4 (mean densitometry ± SEM vs. Static; n = 3 for all groups; one-way ANOVA: p = 0.002, asterisks indicate p < 0.05 in post hoc t-test). Uncropped western blots presented in Supplementary Fig. 14A. d Constitutive GJA4 expression (lenti-Gja4) arrested endothelial cells in G1 and reduced actively cycling cells in S/G2/M, while si-CDKN1B had the opposite effect, regardless of GJA4 expression (mean difference in cell cycle % ± SEM vs. Control; n = 6; one-way ANOVA: p = 0.0002 (G1), p = 0.0003 (S/G2/M), asterisks indicate p < 0.05 in post hoc t-test vs. si-Ctrl + lenti-Ctrl). e GJA4 knockdown (si-GJA4) reduced the proportion of endothelial cells in G1 and increased the proportion in S/G2/M, and constitutive expression of CDKN1B (lenti-CDKN1B) had the opposite effect, regardless of GJA4 expression (mean difference in cell cycle % ± SEM vs. Control; n = 3 for all groups; one-way ANOVA: p = 0.002 (G0), p < 0.0001 (G1), p < 0.0001 (S/G2/M), asterisks indicate p < 0.05 in post hoc t-tests vs. si-Ctrl+ lenti-Ctrl). f CDKN1B phosphorylation at serine 10 (pCDKN1B (S10)) and total protein levels were reduced by si-GJA4 and rescued with lenti-Gja4. Uncropped blots presented in Supplementary Fig. 14B. g MAPK/ERK signaling inhibition by 1 h treatment with 20 µM U0126, an inhibitor of MEK1/2, blocked GJA4-dependent CDKN1B phosphorylation at serine 10 and reduced total CDKN1B protein levels. Uncropped blots presented in Supplementary Fig. 14C

Fig. 6

FSS-NOTCH-GJA4-CDKN1B axis regulates arterial identity genes. a Disorganized or absent SMA investment was observed in Gja4 ^−/− and WT + DAPT mice, compared to WT controls. (Colors: PECAM1 (red), SMA (white); scale bar: 100 µm; symbols: a = artery, v = vein, R = remodeling, M = mature). b Radial SMA investment of arteries, as well as c radial distance of GJA5 (Cx40) expression, were reduced in Gja4 ^−/− and DAPT-treated animals (mean radial distance as % of total vascular outgrowth ± SEM vs WT; n = 3 (WT + DAPT; WT, GJA5), n = 4 (WT, SMA), n = 6 (Gja4 ^−/−); one-way ANOVA: p = 0.03 (SMA), p = 0.04 (GJA5); asterisks indicate p < 0.05 in post hoc t-tests vs. WT). d DLL4 activation of NOTCH signaling upregulated EFNB2 (EphrinB2) and GJA5, and this effect was abolished with si-GJA4 or si-CDKN1B (mean relative mRNA expression ± SEM vs. PBS; n = 3 (si-GJA4; si-CDKN1B, EFNB2), n = 4 (si-Ctrl, GJA5; si-CDKN1B, GJA5), n = 4 (si-Ctrl, EFNB2); Students’ t-test: p = 0.02 (si-Ctrl, EFNB2), p = 0.04 (si-Ctrl, GJA5), p = 0.0005 (si-GJA4, GJA5), p < 0.0001 (si-CDKN1B, GJA5)). e NOTCH activation was maximal with 1 h exposure to 18 dynes/cm² shear (mean NICD R.F.U. ± SEM; n = 7000 cells, representative of N = 3). f Expression of GJA4 (Cx37), GJA5 (Cx40) and EFNB2 (EphrinB2) were also maximal at 18 dynes/cm² (mean relative mRNA expression ± SE; n = 3, representative of 4 experiments). g Knockdown of NOTCH1, GJA4 or CDKN1B reduced basal and 10 h FSS-induced upregulation of EFNB2 and GJA5 (mean relative mRNA expression ± SEM vs. Static; n = 3 technical replicates; representative of two biological replicates)

Fig. 7

Endothelial cell cycle arrest, per se, enables arterial gene expression. a Treatment of HUVEC with 10 µM clotrimazole or 2 µM palbociclib reduced RB1, phosphorylated RB1 (pRb1), and E2F1, suggestive of G1 arrest. Clotrimazole tended to upregulate CDKN1B expression and preserve GJA4 expression, and CDK4 expression was lost only with CDK4/6i treatment. Uncropped blots presented in Supplementary Fig. 16. b Using FACS to assess cell cycle distribution, clotrimazole was found to arrest HUVEC in G1, even when CDKN1B was knocked down (via si-CDKN1B) (mean difference in cell cycle % ± SEM vs. DMSO; n = 3 (Clotrimazole), n = 5 (DMSO), n = 6 (Clotrimazole+ si-CDKN1B); individual values plotted where n < 5; one-way ANOVA: p = 0.007 (G1), p = 0.03 (S/G2/M); asterisks indicate p < 0.05 in post hoc t-tests). c Clotrimazole upregulated EFNB2 and GJA5 regardless of CDKN1B expression (mean relative mRNA expression ± SEM vs. DMSO; n = 4 (Clotrimazole, p27), n = 5 (Clotrimazole, EFNB2; Clotrimazole, GJA5), n = 7 (Clotrimazole+ si-CDKN1B), n = 8 (DMSO); individual values plotted where n < 5; one-way ANOVA: p = 0.03 (EFNB2, GJA5), p = 0.0007 (CDKN1B)). d Palbociclib treatment also arrested endothelial cells in G1 (mean difference in cell cycle % ± SEM vs. DMSO; n = 3 for all groups; Students’ t-test: p = 0.002 (G1), p < 0.0001 (S/G2/M)), e upregulated EFNB2 and GJA5 (Cx40) mRNA levels, abolished CDK4 expression, and had no effect on GJA4 (Cx37) and CDKN1B (p27), (mean relative mRNA expression ± SEM vs DMSO; n = 9 (GJA4), n = 12 (EFNB2), n = 14 (GJA5), n = 16 (CDKN1B; CDK4); Students’ t-test: p = 0.02 (EFNB2, GJA5), p < 0.0001 (CDK4)). f Endothelial cells within arteries and surrounding plexi of P6 retinas from Cdt1-mOrange+ reporter mice were predominantly in G1 phase, whereas endothelial cells in adjacent veins were not. (Colors: IB4 (red), ERG (blue), pH3 (magenta), CDT1 (green); scale bar: 100 µm). g We hypothesize that in remodeling vessels, arterial shear activates a novel Notch-Cx37-p27 signaling pathway that promotes endothelial cell cycle arrest to enable arterial gene expression