Exploiting porphyrin metabolism to inhibit angiogenesis
Francesco De Giorgio, Giuseppe Mannino, Veronica Bonalume, Alessia Bibi, Cristina Martorana, Sabrina Digiovanni, Chiara Riganti, Tullio Genova, Luca Munaron, Christiana Ruhrberg, Alessandro Fantin, Sara Petrillo, Emanuela Tolosano
TL;DR
This paper shows that disrupting heme and porphyrin metabolism in endothelial cells can inhibit blood vessel growth, offering a potential treatment for diseases with abnormal blood vessel formation.
Contribution
The study reveals a novel pharmacological approach using ALA to inhibit angiogenesis by targeting porphyrin metabolism in endothelial cells.
Findings
ALA treatment caused porphyrin accumulation and reduced endothelial cell proliferation and migration.
Extracellular porphyrins contributed to anti-angiogenic effects in both ex vivo and in vivo models.
ALA inhibited pathological neovascularization in a mouse model of retinopathy.
Abstract
Heme is an essential iron-containing porphyrin that plays a critical role in endothelial cell (EC) function, regulating processes such as cell signalling and energetic metabolism. Nevertheless, the role of de novo heme synthesis and porphyrin metabolism during angiogenesis remains poorly understood. In this study, a pharmacological approach using 5-aminolevulinic acid (ALA) was employed to dysregulate heme/porphyrins homeostasis in EC. ALA treatment resulted in intracellular porphyrins accumulation and extensive release into the extracellular environment. ALA-treated EC exhibited diminished proliferation and migration, as well as reduced ability to form tubule-like structures, which led to impaired ex vivo angiogenic sprouting and in vivo angiogenesis in the developing retina. Moreover, ALA inhibited pathological neovascularization in the oxygen-induced retinopathy mouse model that…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 10
Figure 11
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9- —Università degli Studi di Torino
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsHeme Oxygenase-1 and Carbon Monoxide · Porphyrin Metabolism and Disorders · Photodynamic Therapy Research Studies
Introduction
Heme is an iron-containing porphyrin of vital importance for endothelial cells (EC). Indeed, besides its role as a cofactor in heme-containing proteins, it orchestrates a variety of cellular processes including cell signalling and energetic metabolism. To this end, intracellular heme levels are finely controlled at multiple levels, including biosynthesis, trafficking and catabolism [1, 2].
The first and rate-limiting enzyme of the heme biosynthetic pathway is aminolevulinic acid synthase 1 (ALAS1; Enzyme Commission number EC 2.3.1.37) which catalysed the condensation of succinyl-CoA and glycine to produce 5-aminolevulinic acid (ALA) [3]. Various studies have established that the physiological release of ALA from ALAS1 reflects the overall rate of heme synthesis [4]. Consistently, ALAS1 activity is strongly inhibited by heme, the final product of the pathway, ensuring that ALA production remains precisely regulated to meet heme synthesis demands [4].
From a metabolic point of view, two molecules of ALA are condensed by ALA dehydratase (ALAD) to form porphobilinogen (PBG) that is subsequently used to assemble the porphyrin ring. The first porphyrin intermediate is hydroxymethylbilane (HMB), whose production is catalysed by HMB synthase (HMBS; EC 2.5.1.61). HMB is progressively converted into uroporphyrinogen III (UPG), coproporphyrinogen III (CPG), and protoporphyrinogen III (PPG), ultimately leading to the formation of protoporphyrin IX (PPIX). The final step involves mitochondrial ferrochelatase (FECH; EC 4.99.1.1), which incorporates ferrous iron into PPIX, to synthetize de novo protoheme [4].
The endothelial targeting of enzymes involved in heme synthesis, including phosphoglycerate dehydrogenase (PHGDH; EC 1.1.1.95), the enzyme synthesizing glycine, and FECH, severely impairs angiogenesis due to intracellular heme deficiency which drives mitochondrial dysfunction [5–7]. In line with these findings, perturbed intracellular heme metabolism was previously showed to profoundly disrupts key endothelial functions. Indeed, models lacking the Feline Leukaemia Virus subgroup C Receptor 1, “a” isoform (FLVCR1a), a transmembrane choline/ethanolamine importer with a proven role in controlling ALAS1 activity, exhibit defective angiogenesis [8–12].
Nevertheless, the impact of overdriven heme and porphyrin metabolism on angiogenesis remains poorly explored. Here, a pharmacological approach was employed to address the effects of enhanced heme and porphyrin metabolism in EC. ALA was utilized to strain porphyrin production in EC while bypassing the negative feedback regulation of heme, the final product, on ALAS1 activity. ALA treatment led to intracellular porphyrins accumulation and extensive release into the extracellular environment. Notably, EC behaviour upon ALA treatment was significantly changed: their proliferative and migratory potential was reduced, as well as their ability to form complex tubule-like structures in vitro. Moreover, ALA was effective in reducing angiogenic sprouting from murine choroidal tissue ex vivo. Importantly, the release of excess porphyrins into the extracellular environment contributed, at least in part, to the observed anti-angiogenic effects. Finally, ALA injection in mice hampered in vivo angiogenesis in the developing retina and was effective in reducing aberrant angiogenesis in a mouse model of oxygen-induced retinopathy (OIR).
The results from this study highlight the biological role of endogenous porphyrins in affecting EC behaviour and give example of how to exploit this process in pathological contexts.
Results
De novo heme synthesis is modulated during angiogenesis
To elucidate the physiological relevance of endothelial heme metabolism during vascular development, mRNA levels of genes involved in heme biosynthesis were evaluated in murine angiogenic EC. Retinal and choroid EC were isolated by immunomagnetic separation (Fig. S1a) from wild-type mice at postnatal day (P) 3, P6 and P15. In mice retinal angiogenesis starts around P3, when most EC are actively proliferating and migrating, and is largely completed by P15, when EC have acquired a predominant quiescent state. Interestingly, most biosynthesis-related genes were significantly higher in EC (i.e. CD31+) as compared to the non-endothelial counterpart (i.e. CD31-) of the retinal tissue with the strongest difference observed at P6 (Fig. 1a–h). Furthermore, genes involved in heme biosynthesis were enriched in P6 retinal angiogenic EC compared to the quiescent choroidal vasculature (Fig. S1b, c). Analysis of a publicly available single cell RNA sequencing (scRNA-seq) dataset from murine P6 retinas, confirmed widespread detection in EC of transcripts for genes involved in heme biosynthesis, most prominently Alas1 and Alad transcripts (Figs. 1i–l; S1d).Fig. 1. Heme synthesis is up regulated during retinal angiogenesis a Schematic representation of heme synthesis and catabolism pathways. Up-regulated genes are displayed in red. b–h Quantitative real-time reverse transcription PCR (qRT-PCR) analysis of heme-related genes (i.e. Alas1, Alad, Hmbs, Uros, Cpox, Fech, Hmox-1) in CD31+ (EC, red line) and CD31- (retina parenchyma, grey line) cells isolated from developing retina at P3, P6 and P15. Data are representative of at least 3 independent experiments and are expressed as mean ± SEM. *p < 0.05; **p < 0.01; ****p < 0.0001. For statistical analyses ordinary two-way ANOVA test with Sidak’s multiple comparisons was used. i–l Single-cell RNA sequencing data from mouse retina on P6. A colour-coded UMAP of cell type identity is shown (i) alongside UMAPs illustrating the expression levels of the indicated genes in each cell (j and k). The corresponding violin plots illustrate the expression levels of the indicated genes for each cell cluster (l). Each data point represents the value for one cell. FC, fold change
To better investigate the role of heme in highly proliferative EC, human EC derived from breast tumor tissue (BTEC) and from healthy dermal tissue (HMEC) were used. In line with the higher proliferative rate of BTEC, both ALAS1 transcript and protein levels were markedly elevated in BTEC compared to HMEC (Fig. S1e–g) [9]. Additionally, BTEC displayed higher intracellular heme levels as compared to HMEC and same intracellular porphyrin levels (Fig. S1h, i). Taken together, these data indicate that EC rely on active heme synthesis to accomplish angiogenesis.
To further address the importance of de novo synthetized heme during angiogenesis, a strategy to mobilize the intracellular heme pool was used as an attempt to stimuli de novo heme synthesis. To this aim hemopexin (HX) was selected as a promising candidate due to its previously reported function as a heme export [13, 14]. Interestingly, endothelial Hpx was strongly upregulated during retinal vascular development, with a peak at P6 (Fig. 2a). This pattern suggests a potential role for HX in physiological angiogenesis and in maintaining vascular homeostasis. Consistent with reported function of HX, HX treatment in EC increased the extracellular heme content, rather than porphyrins which remained unchanged (Fig. 2b, c). Interestingly, enhanced heme export correlated with higher mitochondrial ALAS activity following HX administration (Fig. 2d), likely suggesting HX-mediated intracellular heme mobilization stimulates de novo heme synthesis. Nevertheless, intracellular heme and porphyrins levels remained substantially unchanged (Fig. 2e, f). In line with the crucial role of active heme synthesis in sustaining cell proliferation, HX supplementation enhanced the proliferation of human EC (Fig. 2g, h). To test the effects of HX on the angiogenic properties of EC, an ex vivo model of microvascular angiogenesis from choroidal tissue was employed. Choroids from adult mice were dissected, cut into small pieces, and embedded in a thick layer of Matrigel resembling the extracellular matrix [15]. HX treatment was started at in vitro day (DIV) 3, when the first sprouts appeared, and choroidal sprouting was then monitored until DIV7. HX treatment enhanced ex vivo choroidal neovascularization, as indicated by increased vascular area and radial outgrowth (Fig. 2i–k). Collectively, these data indicated that active heme synthesis sustains angiogenesis and that HX supports vascularization by favouring de novo heme synthesis.Fig. 2. Hemopexin (HX) promotes angiogenesis by stimulating endogenous heme synthesis in EC a qRT-PCR analysis of Hpx in CD31+ (EC, red line) cells isolated from developing retina at P3, P6 and P15. Plotted data are mean values of technical replicates obtained from a pooled sample of more than six eyes. b–c Extracellular heme (b) and porphyrin (c) levels in HMEC treated with 100 µM HX for 72 h (hrs). d ALAS mitochondrial activity in HMEC treated with 100 µM HX for 24 h. e–f Intracellular heme (e) and porphyrins (f) levels in HMEC treated with 100 µM HX for 72 h. g–h Proliferation of HMEC treated with 5 µM HX at the indicated time points. Scale bar: 500 µm. n = 6 i–k ex vivo choroid sprouting assay performed in presence of 5 µM HX. Representative images were taken at DIV 7. Scale bar: 1 mm. Vascular outcomes were evaluated by measuring the vascular area of the growing sprouts (j) and the maximal radial extension (k) at various time points. n > 6. Data are expressed as mean ± SEM. *p < 0.05; ***p < 0.001; ****p < 0.0001; ns, not significant. For statistical analyses, parametric unpaired t test (a–d) and ordinary two-way ANOVA test with Sidak’s multiple comparisons (f, h, i) were used. FC, fold change; hrs, hours; HX, hemopexin; NT, not treated controls
To further assess the importance of de novo heme synthesis in supporting angiogenesis, heme production was inhibited in BTEC using succinyl acetone (SA), an inhibitor of ALAD (EC 4.2.1.24) [16]. As expected, treatment with 0.5 mM SA caused a significant reduction in intracellular porphyrins within 24 h and a trend toward decreased heme content (Fig. S2a, b). Importantly, ALAS1 mRNA and protein levels were markedly elevated after 24 h of treatment (Fig. S2c, e), consistent with the induction of a heme-deficient state following SA administration. To assess the functional consequences of heme depletion on human EC, cell proliferation and migration were evaluated following SA treatment. In line with the essential role of heme in supporting EC functions, SA inhibited both proliferation and migration of BTEC in a dose-dependent manner (i.e. 0.5 mM, 1 mM) (Fig. S2f–i). To validate these findings in a more physiological system, ex vivo choroidal sprouting assay was exploited. Remarkably, SA reduced choroidal sprouting in a dose-dependent fashion, with a 5 mM concentration completely abolishing vascular sprouting, as evidenced by the progressive reduction in both vascular area and radial extension (Fig. S2j–l).
Taken together, these results indicate that de novo heme synthesis is critically required to support angiogenesis, and that pharmacological inhibition with SA impairs vascular growth.
ALA dysregulates de novo heme synthesis
De novo heme biosynthesis is essential for angiogenesis, but its fine-tuned regulation is critical to prevent heme or porphyrin intermediates accumulation, which may compromise EC survival and/or functionality. To specifically address the impact of dysregulated heme metabolism during angiogenesis, human-derived EC were treated with the heme precursor ALA. ALA supplementation in human EC promoted a prompt and sustained accumulation of intracellular porphyrins as indicated by fluorometric analysis (Fig. 3a, S3a). Fluorescence imaging on living EC confirmed the accumulation of porphyrins predominantly within the cytosol, as shown in Fig. 3b, c. Despite elevated porphyrins levels, intracellular heme did not significantly change after 24 h (hrs) of ALA treatment (Fig. 3d, S3b). Nevertheless, ALA-treated EC showed a diminished ALA synthase 1 (ALAS1) mitochondrial activity (Fig. 3e) together with a strong downmodulation of ALAS1 protein levels (Fig. 3f–g, S3c–d). Moreover, heme oxygenase-1 (HO-1; EC 1.14.99.3) was highly induced within the first 24 h of ALA treatment, possibly playing a role in maintaining intracellular labile heme balance (Fig. 3f, h). These results suggest that, following ALA administration, heme levels underwent fluctuations which are likely responsible for the cellular response mediated by ALAS1 and HO-1. Accordingly, qPCR analysis revealed that ALA treatment profoundly affected the expression profile of heme-related genes, including a significant down-modulation of enzymes involved in porphyrin synthesis (i.e. *UROS, *UROD, CPO, PPO) as well as a higher expression of ALAS1 and FECH (Fig. 3i–j). Notably, genes associated with porphyrin/heme transport (e.g., ABCB6 and ABCG2) were among the modulated genes (Fig. 3i–j). Consistent with changes in the expression of porphyrin transporters, culture media derived from ALA-treated EC were characterized by significantly higher levels of porphyrins and heme (Fig. 3k–l).Fig. 3ALA promotes porphyrin overdrive in human microvascular EC a Intracellular porphyrins levels in HMEC treated with 5 mM ALA for 4, 24, and 72 h. b–c Fluorescence imaging and relative quantification of intracellular porphyrins (magenta) in HMEC treated with 5 mM ALA and controls. Scale Bar: 50 µm. n = 12 d Intracellular heme levels in HMEC treated with 5 mM ALA for 4, 24, and 72 h. e Mitochondrial activity of ALAS in HMEC treated with 5 mM ALA for 24 h (hrs). f–h Representative Western Blot images (f) and their quantification g–h of ALAS1 and HO-1 protein levels in HMEC treated with 5 mM ALA for 24 and 72 h. i–j Differential expression of heme-related genes in HMEC upon 24 h of 5 mM ALA treatment. The heatmap shown in (i) displays differences in gene expression levels, represented as fold change (FC) respect to not-treated (NT) cells. Volcano plot shown in (j) represents the differences in gene expression levels versus -log10(q value). The dotted line indicates the significance threshold of FDR q < 1% (-log10(q value) = 2). The plotted results represent mean values obtained from at least 5 individual biological replicates. k–l Extracellular levels of porphyrins (k) and heme (l) in HMEC after 4, 24, and 72 h of 5 mM ALA supplementation. Data are expressed as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. For statistical analyses, ordinary one-way ANOVA test with Tukey’s multiple comparisons (a, d, g, h, k, l) and parametric unpaired t test (c, e) were used. FC, fold change; NT, not treated controls; hrs, hours; ALA, 5-amminolevulinic acid
Taken together, these data indicated that ALA supplementation in EC dysregulates heme metabolism by promoting intracellular porphyrin overload. Moreover, porphyrins are likely exported into the extracellular environment.
ALA inhibits angiogenesis
To investigate the functional effects mediated by dysregulated heme metabolism on human EC, cell proliferation, migration and in vitro angiogenic potential, were evaluated upon ALA treatment. Interestingly, ALA inhibited the proliferation of microvascular EC from human dermis (HMEC) in a dose dependent manner (Fig. 4a, b). Similarly, breast-tumor derived EC (BTEC) treated with ALA displayed a reduced proliferative rate as compared to the not-treated counterpart (Fig. S3e–f). Furthermore, ALA treatment impaired HMEC migration starting from a 100 μM dosage (Fig. 4C, D). Moreover, similar results were obtained using BTEC (Fig. S3g–h), further pointing out the detrimental effects of porphyrins on the migratory potential of EC. Finally, the ability of EC to form de novo vascular networks was evaluated by performing in vitro tubulogenesis assay. ALA-treated HMEC displayed a compromised ability to organize into a complex capillary network (Fig. 4e). Indeed, the total vascular length (Fig. 4f) as well as the number and complexity of interconnections (Fig. 4g, h) within the capillary network were significantly reduced following ALA administration.Fig. 4ALA inhibits in vitro angiogenesis on human microvascular EC (a and b) Proliferation of HMEC treated with increasing concentrations of ALA (500 nM, 0.1 mM, 5 mM) at various time points. Scale bar: 500 µm. n = 6 (c and d) Wound healing experiment of HMEC treated with increasing concentrations of ALA (500 nM, 0.1 mM, 5 mM) at various time points. Scale bar: 500 µm. n > 8 (e–h) in vitro tubulogenesis assay performed on 5 mM ALA-treated and control HMEC. Quantification of the total length (f) of the networks, number of nodes (g) and number of branches (h) are shown. Scale bar: 500 µm. Data are expressed as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. For statistical analyses ordinary two-way ANOVA test with Tukey’s multiple comparisons (b, d) and parametric unpaired t-test (f–h) were used. FC, fold change; NT, not treated controls; hrs, hours; ALA, 5-amminolevulinic acid
To further investigate the biological relevance of aberrant heme synthesis in EC, ex vivo neovascularization assay was performed. Consistent with data on human EC, ALA severely impaired choroidal sprouting in a dose dependent manner (Fig. 5a) as indicated by the reduced vascular area and extension (Fig. 5b, c). Moreover, ex vivo vascular defects due to ALA administration were associated with a strong increase in porphyrins and heme extracellular levels at DIV7 (Fig. 5g, h). To demonstrate the pivotal role of excess porphyrin metabolism in ALA-driven sprouting defects, the enzyme ALAD was inhibited by using SA to block downstream porphyrins production [16]. Notably, 0.5 mM SA was proficient in driving a complete recover of defective choroidal angiogenesis in ALA-treated samples (Fig. 5d–f). Remarkably, ALAD inhibition was also sufficient to abolish the increase in extracellular porphyrins and heme caused by ALA (Fig. 5g, h). These findings indicate that ALA impairs ex vivo sprouting by disrupting endothelial porphyrin homeostasis. Moreover, ALAD inhibition is sufficient to restore the dysregulated heme synthesis and defective angiogenesis caused by ALA, highlighting the critical role of ALAD in maintaining an adequate heme synthesis rate in EC. To further corroborate these data, aorta ring assay was performed upon ALA supplementation. Briefly, aorta from adult mice were dissected, EC inner layer exposed, cut into small rings and embedded in a thick layer of extracellular resembling matrix [17]. Treatment with ALA was performed starting from DIV3, when the first sprouts started to form, and aorta sprouting measured at DIV7. Consistent with previous data, ALA strongly impaired aorta angiogenesis as indicated by the reduced radial growth of the aorta sprouts (Fig. 5i, j).Fig. 5ALA inhibits ex vivo murine angiogenesis (a–c) ex vivo choroidal neovascularization assay with ALA at increasing doses (i.e. 2 µM, 0.5 mM, 5 mM). Representative images were taken at day in vitro (DIV) 7. Scale bar: 1 mm. Defective vascularization was evaluated by measuring the vascular area of the growing sprouts b and the maximal radial extension c at various time points. n > 3 d–f ex vivo choroidal neovascularization assay with 5 mM ALA or/and 0.5 mM succinyl acetone (SA). Representative images were taken at DIV 7. Scale bar: 1 mm. ex vivo sprouting was evaluated by measuring the vascular area of the growing sprouts e and the maximal radial extension f at various time points. n = 4 g–h Extracellular porphyrin g and heme h levels of DIV7 growing choroids treated with 5 mM ALA or/and 0.5 mM SA. Data are normalized over the area of the growing sprouts. Plotted data represent technical replicates obtained from pooled medium derived from four independent choroid cultures.** i**–j ex vivo aorta neovascularization assay in presence of 5 mM ALA. Representative images were taken at day in vitro DIV9. Scale bar: 1 mm. Aorta sprouting was evaluated by measuring the maximal radial growth (j) over the sprout. Data are expressed as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. For statistical analyses, ordinary two-way ANOVA test with Sidak’s multiple comparisons (b, c, e, f), ordinary one-way ANOVA test with Tukey’s multiple comparisons (g, h) and parametric unpaired t-test (j) were used. FC, fold change; NT, not treated controls; hrs, hours; ALA, 5-amminolevulinic acid; SA, succinyl acetone; DIV, day in vitro
Collectively, these data indicate that dysregulated heme metabolism is detrimental for EC and compromises their ability to accomplish angiogenesis.
Extracellular porphyrins are involved in ALA anti-angiogenic effects
To investigate whether changes in the extracellular porphyrin profile might directly contribute to the angiogenic defects, choroidal EC were treated with medium-derived from ALA-treated EC (Fig. 6a). Specifically, DIV5 choroids were treated with ALA for 24 h. Then, ALA containing medium was replaced with fresh ALA-free medium after an extensive washing phase. This medium was conditioned (CM) for other 24 h to enable the accumulation of porphyrins intermediates and heme. As expected, CM derived from ALA-treated choroids (CM-ALA) was enriched in porphyrins (Fig. 6b). Conversely, heme levels remained comparable to controls. This suggests that, within this experimental timeframe, ALA treatment was not sufficient to induce detectable extracellular heme accumulation, allowing to specifically assess the effect of porphyrins (Fig. 6c). Hence, CM-ALA was used to treat naïve choroids from DIV3 to DIV7 (Fig. 6a). Remarkably, choroids treated with CM-ALA displayed important sprouting defects (Fig. 6d–f) similarly to ALA treatment itself, highlighting the leading role of extracellular porphyrins in ALA-driven angiogenic defects. Similar results were also obtained in human EC by collecting medium from ALA-treated HMEC and using it to condition naïve HMEC. Indeed, CM-ALA diminished HMEC proliferation (Fig. 6g) and migration (Fig. 6h) as compared to NT-CM counterpart.Fig. 6. Extracellular porphyrins accumulation due to ALA treatment inhibits angiogenesis in a paracrine fashion a Schematic representation of conditioning experiments with medium derived from control or ALA-treated choroid. b and c Extracellular porphyrins (b) and heme (c) levels in choroid-derived CM-ALA or CM-CTRL at the end of medium conditioning. Data are normalized over the area of the growing sprouts. d–f ex vivo choroidal neovascularization assay in presence of CM-ALA or control CM. Representative images were taken at day in vitro (DIV) 7 are shown in (d). Scale bar: 1 mm. ex vivo sprouting was evaluated by measuring the vascular area of the growing sprouts (e) and the maximal radial extension (f) at various time points. g–h Proliferation (g) and wound healing (h) experiments on HMEC treated with CM-ALA or control CM at various time points. i Extracellular porphyrins levels in HMEC treated with 5 mM ALA and/or 100 μM HX for 72 h. Data are normalized over the confluence cell area. j Intracellular porphyrins levels in HMEC treated with 5 mM ALA and/or 100 μM HX for 72 h. k–l Extracellular and intracellular heme levels in HMEC treated with 5 mM ALA and/or 100 μM HX for 72 h. Data on extracellular heme are normalized over the confluence cell area. m Proliferation experiments on HMEC treated with 5 mM ALA and/or increasing doses of HX (i.e. 5 μM, 100 μM) at various time points. n = 6. n–o Extracellular porphyrins and heme levels in HMEC treated with 5 mM ALA and/or 5 uM HSA for 72 h. Data on extracellular heme are normalized over the confluence cell area. p Proliferation experiments on HMEC treated with 5 mM ALA and/or 5uM HSA at various time points. n = 6. Data are expressed as mean ± SEM. For statistical analyses, ordinary two-way ANOVA test with Sidak’s multiple comparisons (g,h,m,p), ordinary one-way ANOVA test with Tukey’s multiple comparisons (i–l, n–o) and unpaired Student’s t-test were used (b, c, e, f). FC, fold change; hrs, hours; ALA, 5-amminolevulinic acid; HX, hemopexin; HSA, human serum albumin; NT, not treated; CM, conditioned medium; hrs, hours; DIV, day in vitro
Collectively, these data indicate that ALA inhibits angiogenesis partly by causing a rewiring of the extracellular porphyrin profile.
To better investigate the role of porphyrins in these phenotypes, HX was used to enhance ALA-driven release of porphyrins and heme. Indeed, in addition to its heme-binding function, HX can also bind and facilitate the transport of metal-free porphyrins (e.g., PPIX), although with lower affinity [13]. Interestingly, HX treatment in combination with ALA further increased extracellular porphyrin accumulation compared with ALA alone (Fig. 6i). Additionally, intracellular porphyrin levels remained stable under the combined treatment, suggesting that de novo synthesized porphyrins may be rapidly exported by HX (Fig. 6j). Conversely, extracellular heme levels showed only a slight, non-significant increase upon ALA and HX combined treatment as compared to ALA treatment alone, while effectively restoring intracellular heme to control levels (Fig. 6k, l). Under these conditions, the pro-proliferative effect of HX treatment was abolished when combined with ALA, further reducing EC proliferation in a dose dependent manner (i.e. 5–100 uM) compared with ALA alone (Fig. 6m). Interestingly, similar results were also obtained in presence of human serum albumin (HSA), another known porphyrin and heme carrier protein [18]. Indeed, HSA in combination with ALA worsen EC proliferation as compared to ALA alone by promoting higher extracellular porphyrin and heme levels (Fig. 6n, o).
Collectively, these results suggest that ALA inhibits angiogenesis by modulating the extracellular porphyrin profile. Moreover, HX or HSA in combination with ALA further enhanced extracellular porphyrin accumulation, leading to an additional inhibition of EC proliferation.
Protoporphyrin IX impairs angiogenesis
To gain a deeper understanding of the significance of the changes induced by ALA in the extracellular porphyrin profile, heme metabolic intermediates were assessed. To this end, HPLC–DAD-MS/MS was employed to evaluate the content of heme intermediates in medium derived from both human EC (Fig. 7a–c) and cultured murine choroids (Fig. 7d–f) treated with ALA for 72 or 96 h, respectively. Interestingly, ALA-treatment led to a significant enrichment of PBG, PPIX and de novo synthesized heme b in both model systems (Fig. 6a–f). Given the high extracellular enrichment of PPIX caused by ALA administration, PPIX ability to influence angiogenesis was evaluated. Interestingly, PPIX inhibited human-derived EC proliferation (Fig. 8a, b) and migration (Fig. 8c, d) in a dose dependent manner (i.e. 250, 500 nM), pointing out the detrimental effect of PPIX on key EC functions. Moreover, PPIX compromised EC ability to arrange in a two-dimensional space (Fig. 8e). Specifically, the total length of the de novo formed capillary network, as well as the number of nodes and branches, were significantly reduced by PPIX (Fig. 8f–h). Similar results were also obtained in models of choroidal neovascularization (Fig. 8i). Indeed, PPIX inhibited ex vivo choroidal angiogenesis, as indicated by the reduced vascular area and radial growth (Fig. 8j–k). Collectively, these data provide evidence that ALA promotes PPIX export from EC resulting in extracellular PPIX accumulation that in turn exhibits a paracrine antiangiogenic effect.Fig. 7ALA promotes extracellular PPIX accumulation. a–f Quantification of PPGIX, HMB, CPH, HAEM B, PBG, and PPIX using HPLC–DAD-MS/MS in (a–c) ex vivo and in vitro (d–f) experiments. Panels (a, d) illustrate how individual variables discriminate between experimental conditions—control (gray) vs. ALA-treated (purple)—through orthogonal PCA analysis, with explained variance percentages indicated. Panels (b, e) present heatmap visualizations combined with hierarchical cluster analysis to depict variations of individual metabolites within groups -control class (white squares, each square corresponds to a sample) vs. ALA-treated class (purple squares, each square corresponds to a sample)—highlighting statistically significant differences (p < 0.05). Finally, panels (c, f) show loading plots of the respective principal components, indicating the contribution of each metabolite to the observed separation between groups. In b,c,e and f colour gradient from blue (low) to red (high) indicate the abundance of each compound under the specific experimental conditions; compounds with a VIP score > 0.5 are those contributing most to the class separation observed in the orthogonal PCA plots (a, d). Data are representative of at least 3 independent experiments and are expressed as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001Fig. 8Exogenous protoporphyrin IX (PPIX) impairs angiogenesis in human-derived EC (a, b) Proliferation of HMEC treated with increasing concentrations of PPIX at various time points. Scale bar: 500 µm. n > 5 (c, d) Wound healing experiment of HMEC treated with increasing concentrations of PPIX at various time points. Scale bar: 500 µm. n = 8 (e) in vitro tubulogenesis assay performed on HMEC treated with 500 nM PPIX and controls. f–h Quantification of the total network length (f), number of nodes (g) and branches (h) shown in (e). i–k ex vivo choroidal neovascularization assay with 500 nM PPIX. Representative images were taken at day in vitro (DIV) 7. Scale bar: 1 mm. Defective vascularization was evaluated by measuring the vascular area of the growing sprouts (j) and the maximal radial extension (k) at various time points. n = 6 Data are expressed as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. For statistical analyses, ordinary two-way ANOVA test with Sidak’s multiple comparisons (b, d, j, k) and unpaired t-test (f–h) were used. FC, fold change; NT, not treated controls; hrs, hours; PPIX, protoporphyrin IX; DIV, day in vitro
ALA administration in vivo reduces developmental angiogenesis
To investigate the anti-angiogenic effects of ALA in vivo, a physiological model of in vivo murine retinal angiogenesis was employed. First, the expression of ALA transporters (i.e., Solute Carrier Family 15 Member 1/2, Slc15a1/2) was assessed in postnatal murine retinas. Notably, the Slc15a2 isoform was expressed in retinal EC, with high levels detected during the angiogenic phase (P3 and P6), followed by a progressive decline as the vascular network matured (P15). These findings suggest that retinal angiogenic EC may be capable to uptake circulating ALA (Fig. 9a). To further assess ALA delivery in the retina and its conversion into downstream porphyrins, P5 mice were administered with a single intragastric injection of ALA 20 mg/Kg, corresponding to the recommended dose in humans. Porphyrin fluorescence was then evaluated 3 h later in freshly dissected retinas. Consistent with in vitro results, ALA-treated retinas displayed strong porphyrin fluorescence, which was instead absent in control retinas, indicating that retinal cells, including EC, can internalize and metabolize circulating ALA (Fig. 9b, c).Fig. 9ALA impairs physiological retinal angiogenesis a qRT-PCR analysis of the ALA importer Slc15a2 in CD31+ (EC, red line) and CD31- (retina parenchyma, grey line) cells isolated from developing retina at P3, P6 and P15. The plotted results represent technical replicates obtained from a pooled sample of more than six eyes. b–c Visualization and quantification of porphyrin fluorescence in freshly dissected retinas from ALA-treated and control mice. Representative images show (i) brightfield and (ii) porphyrin fluorescence. Porphyrins are visualized in magenta (Ex: 405 nm; Em: 633 nm). Scale bar: 500 µm. d Schematic representation of 20 mg/Kg ALA treatment in postnatal mice. e Whole mount CD31 staining (green) on P5 retinal primary plexus from ALA-treated or control mice. Scale bar: 1000 µm. f Quantification of radial extension of ALA or control P5 retinas. g Schematic representation of 5 mg/Kg ALA treatment in postnatal mice. h Whole mount CD31 staining (green) on P7 retinal primary plexus from ALA-treated or control mice. Scale bar: 500 µm. i–j Quantification of radial extension (i) and retinal radius (j) of ALA-treated or control P7 retinas. k–l CD31 (grey)/pHH3 (red) double staining of ALA-treated and control P7 retinas and quantification of pHH3+ EC. Scale bar: 50 µm. m Quantification of sprout numbers. n Lateral view of the vertical projection of P7 retinas stained with CD31 antibody (green). Scale bar: 500 µm. o–p Quantification of the vertical sprouting area (o) and vertical height (p). Data are representative of at least 3 independent experiments and are expressed as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. For statistical analyses ordinary two-way ANOVA test with Sidak’s multiple comparisons (a) or unpaired t-test (c, f, i–j, l–n, o–p) were used. CTRL, control; ALA, aminolevulinic acid; FC, fold change; ns, not significant; AF, angiogenic front; CZ, control zone; PP, primary plexus; IP; DP, deeper plexus
To evaluate the efficacy of ALA in modulating postnatal retinal angiogenesis, multiple doses and treatment regimens were tested. Firstly, a 20 mg/Kg dose of ALA was selected, and five consecutive intragastric injections were performed from P1 to P5. Retinas were then dissected at P5 (Fig. 9d). As shown in Fig. 9e and f, ALA 20 mg/Kg was effective in inhibiting EC sprouting on the retina primary plexus, demonstrating ALA efficacy in inhibiting angiogenesis in vivo. Importantly, no changes in pups body weight were observed following ALA treatment, likely confirming the safety profile of the drug (Fig. S4g). To limit systemic exposure while maintaining biological efficacy, a more conservative experimental approach was implemented, employing a lower ALA dose (5 mg/kg) and a shorter treatment regimen of three consecutive injections (Fig. 9g). Neonatal retinas were then dissected at P7 and retinal sprouting angiogenesis was assessed by IF. Consistent with the previous analysis, ALA-treated retinas displayed underdeveloped primary plexus as compared to control retinas (Fig. 9h). Indeed, the radial extension of the retinal vasculature was strongly delayed respect to control retinas (Fig. 9i). Notably, this was associated with no changes in the retina radius following ALA treatment (Fig. 9j), further suggesting a not-toxic and endothelial-specific effect of the drug. To elucidate if defective EC proliferation could be involved in the angiogenic defects upon ALA treatment, IF analysis for phosphorylated histone H3 (pHH3) was performed on P7 retinas (Fig. 9k). ALA-treated retinas displayed a diminished number of mitotic cells at the angiogenic front (Fig. 9l), suggesting that impaired angiogenesis could be partly due to defective EC proliferation. Beyond the impaired EC proliferation, the number of sprouts at the angiogenic front was decreased in ALA-treated retinas (Fig. 9m), overall suggesting a potential disruption of EC specification and migration across the retinal tissue. Next, ALA effects on retinal vertical sprouting were assessed. Starting from P6, EC from the primary plexus begin to migrate and proliferate toward the deeper retinal layers, giving rise to a secondary deep vascular plexus and an intermediate plexus. These additional vascular layers progressively develop and reach full maturation by P21. Remarkably, ALA delayed EC vertical sprouting toward the deeper layer of the retina at P7 (Fig. 9n). Indeed, the vertical sprouting area and the vertical vascular height were strongly reduced (Fig. 9o, p). To understand if pups treated with ALA could recover from the angiogenic defects, they were allowed to develop until P21(Fig. S4a). Interestingly, ALA-treated mice fully recovered from the angiogenic defects present at P7 (Fig. S4b). Indeed, the vascular areas of the three vascular layers (i.e. primary, intermediate, deep plexus) were similar (Fig. S4c), as well as the vertical sprouting vascular area and the vertical vascular height (Fig. S4d–f). Taken together these data strongly demonstrate the ability of ALA to inhibit in vivo retinal angiogenesis together with a non-toxic and reversible effect of the drug.
ALA administration hampers pathological angiogenesis in a mouse model of oxygen-induced retinopathy
To address the impact of dysregulated porphyrin metabolism on pathological vascularization, a murine model of oxygen-induced retinopathy (OIR) was used. In this model, sequential exposure of mouse pups to hyperoxia followed by normoxia promotes the formation of retinal neovascular lesions that resemble those ones observed in human pathological retinal neovascularization (RNV), including proliferative diabetic retinopathy (PDR) and retinopathy of prematurity (ROP) [19]. Specifically, mice were exposed to hyperoxia (i.e. 85% O_2_) from P8 to P11 to induce vaso-obliteration of the central retinal vasculature. Subsequent return to room air leads to retinal hypoxia and upregulation of proangiogenic factors, ultimately resulting in aberrant neovascularization and the formation of collagen IV-enriched tuft-like vascular structures by P16 (Fig. 10a), as previously showed [20, 21]. To assess the effect of ALA on modulating pathological RNV, mice received daily intraperitoneal injection of ALA or vehicle (i.e. PBS) from P12 to P15, and retinas were collected at P16 (Fig. 10a). First, the effects of 5 mg/kg ALA were evaluated, showing no effects on pathological neovascularization (data not shown). Therefore, a 20 mg/kg ALA dose was employed for subsequent experiments. Importantly, ALA did not impair murine development, as evidenced by the absence of significant differences in body weight gain compared to control animals, supporting its profile as a well-tolerated drug (Fig. S4h). Consistent with our data on physiological angiogenesis, 20 mg/Kg ALA was effective in reducing neo-vascularization and inhibited the formation of neo-vascular tufts (Fig. 10b, c), without significantly affecting the vaso-obliterated area (Fig. 10b, d). The formation of pathological arteriovenous shunts also appeared reduced compared to controls, although not significantly (Fig. 10b, e).Fig. 10ALA impairs pathological neovascularization in a OIR mouse model a Schematic representation of OIR model and ALA treatment in postnatal mice. b Whole mount αSMA (green) and collagen IV staining (black) on P16 OIR retinas from ALA-treated or control mice. Neovascular (NV, red) and vaso-obliterated (VO, yellow) areas superimposed on the collagen IV staining are shown. Scale bar: 1 mm. c–d Quantification of neovascular and vaso-obliterated area in ALA-treated or control P16 OIR retinas. e Quantification of arteriovenous shunts in ALA-treated or control P16 OIR retinas. Data are representative of at least 3 independent experiments and are expressed as mean ± SEM, > 7. **p < 0.01. For statistical analyses unpaired t-test (c–e) was used. ALA, aminolevulinic acid; PBS, Phosphate Buffer Saline; ns, not significant; NV, neovascularization; VO, vasoobliteration
Collectively, these data identified ALA as a potential therapeutic agent for conditions characterized by pathological angiogenesis such as RNV.
Discussion
This study elucidates the critical role of de novo synthetized heme in sustaining angiogenesis. Genes involved in heme biosynthesis (i.e. Alas1, Alad, Fech) were found upregulated in angiogenic retinal EC. Furthermore, highly angiogenic human EC isolated from breast tumor tissue exhibited increased ALAS1 protein levels and higher intracellular heme levels compared to less angiogenic EC from human dermis [9]. Consistent with the need of an active heme synthesis during the angiogenic process, previous studies showed that the targeting of genes involved in heme synthesis (i.e., Phgdh and Fech) impairs angiogenesis by limiting intracellular heme availability [5–7, 22]. In line with this, pharmacological inhibition of ALAD by succinyl acetone (SA), by causing heme deficiency, reduced tumor-derived EC proliferation, migration, and ex vivo angiogenesis. Consistently, recent data showed that SA impairs the formation of the deep capillary plexus in developing retina, further supporting an anti-angiogenic role for SA. However, heme depletion by SA was found to increase sprout length, only when induced after tip cell formation and initiation of angiogenesis. [https://iovs.arvojournals.org/article.aspx?articleid=2781139]. Overall, these data might reflect distinct requirements for endothelial heme in different phases of angiogenesis.
To further investigate the physiological relevance of heme synthesis during angiogenesis, hemopexin (HX) was also employed in this study. HX is a plasma glycoprotein that binds free heme with high affinity and delivers it to the liver or spleen for clearance [23, 24]. Despite being primarily produced in the liver, HX was also found expressed in angiogenic EC. Interestingly, HX has been also proposed as a heme/porphyrin export facilitator, even though its mechanism of action is still debated [13, 14, 18]. In line with this, HX was proved effective in mobilizing endothelial heme and, by doing this, it favoured de novo heme synthesis as shown by mitochondrial ALAS activity. Notably, HX promoted EC proliferation and vessel sprouting, consistent with its previously described pro-angiogenic function in the rat ischemic brain [25]. Taken together, these data support the pro-angiogenic role of de novo heme synthesis.
Nevertheless, our findings underscore the necessity of finely tuned heme metabolism to prevent the buildup of porphyrin intermediates, which are detrimental for angiogenesis. Indeed, the expression of genes involved in the steps of porphyrins production remained largely unchanged during retinal angiogenesis, likely reflecting the essential need to maintain intracellular porphyrins homeostasis despite the increased demand for heme. During heme synthesis, five intermediate porphyrins are produced (i.e. HMB, UPG, CPG, PPG, PPIX). These porphyrin intermediates are rapidly converted into their downstream metabolite to avoid harmful accumulation. Consistently, mutations interfering with the physiological flow of the pathway lead to a group of rare disorders named porphyria which are characterized by porphyrins buildup in the body leading to skin photosensitivity, biliary stones, hepatobiliary damage and liver failure [26, 27]. Importantly, EC dysfunction and long-term vascular complications are increasingly recognized as relevant features of the disease spectrum [28].
In this view, our findings demonstrate that endothelial porphyrin metabolism plays a crucial role in regulating angiogenesis (Fig. 11a). ALA, by promoting a “porphyrin overdrive” phenotype, impinged on key endothelial functions including proliferation, migration and the ability to form vascular network. Moreover, ALA reduced ex vivo choroid, and aorta sprouting and affected developmental angiogenesis in vivo.Fig. 11. Schematic representation ALA supplementation in EC increases porphyrins production which accumulates intracellularly and in the extracellular environment. The ALA-induced “porphyrin overdrive” phenotype impinges on physiological and pathological angiogenesis
A proper and balanced flow of heme biosynthesis is essential for maintaining adequate intracellular heme availability in EC and ensuring a “healthy” metabolic state [1, 29]. To this end, ALAS1 represents a crucial rate-limiting step, in line with its inherently low kinetic properties. ALAS1 activity is strictly dependent on the availability of its substrates, i.e. glycine and succinyl-CoA, which are condensed to form ALA. Moreover, ALAS1 activity is strongly inhibited by the final product of the pathway, i.e. heme, thus ensuring a proper amount of ALA is supplied not exceeding what is needed for heme synthesis [4, 30]. Consistently, ALA treatment in human EC led to a marked inhibition of ALAS1 protein expression and activity. Interestingly, decreased ALAS1 protein levels correlated with a strong induction of its transcript upon ALA treatment. This apparent discrepancy between ALAS1 mRNA and protein expression could be explained by the complex regulation of ALAS1 by heme. Indeed, ALAS1 protein may undergo rapid degradation, or its mitochondrial translocation—and consequently its activity—may be inhibited by fluctuations in intracellular heme levels. Conversely, the upregulation of ALAS1 mRNA might represent a compensatory response to an altered porphyrin-to-heme ratio. In line with this, ALAS1 mRNA levels were found elevated in patients with porphyria, in whom porphyrin accumulation is a hallmark feature, or upon administration of porphyrogenic compounds [31, 32].
Exogenous ALA administration bypasses the inhibition of ALAS1 and is rapidly converted into PBG molecules by ALAD, the fastest enzyme in the heme synthesis pathway. Under such conditions, HMBS becomes the secondarily rate-limiting enzyme due to its low kinetic, leading to PBG accumulation. Despite little is known about PBG role as a bioactive molecule in cells, previous works showed that PBG can diffuse across cell membranes [4, 33]. High levels of extracellular PBG were consistently measured in culture media derived from ALA-treated EC, likely reflecting an intracellular build-up of PBG. The fast kinetic of the enzymes downstream to PBG suggests that porphyrin intermediates are rapidly converted into the downstream byproduct to avoid accumulation. Consistently, inhibition of ALAD by SA was sufficient to prevent ALA-induced porphyrin accumulation and to rescue angiogenic defects, underscoring the pivotal role of endothelial ALAD in controlling ALA influx to ensure proper porphyrin and heme synthesis. Under conditions of dysregulated heme metabolism, cells have only two main strategies to counteract intracellular porphyrin accumulation: converting porphyrins into heme for subsequent degradation by HO-1 or exporting them into the extracellular environment. Although passive diffusion was expected due to the lipophilicity of porphyrins, a more rapid or reduced leakage has been observed in cell lines with modified expression of some members of the ATP binding cassette (ABC) family of transporters [34–37]. The ABC subfamily G member 2 (ABCG2) and the ABC subfamily B member 6 (ABCB6) are the most studied porphyrin/heme exporters [18, 38–40]. This evidence suggests the existence of active outward transport mechanisms for cellular porphyrins. Accordingly, media from ALA-treated EC were characterized by a marked accumulation of extracellular PPIX and heme b**,** possibly indicating EC ability to export porphyrins. Interestingly, the expression of ABCB6 was found to be significantly upregulated in ALA-treated EC, potentially indicating its role in facilitating porphyrin export.
Notably, anti-angiogenic effects like those obtained with ALA were also observed following PPIX exposure. Indeed, PPIX inhibited human-derived EC proliferation, migration and ability to organize in networks as well as ex vivo choroidal sprouting, suggesting its anti-angiogenic function. These results align with previous study demonstrating the antiangiogenic role of the porphyrin analogue ZnPPIX on postnatal retinal angiogenesis in mice [41]. Interestingly, exogenous PPIX was rapidly taken up by EC and converted to heme (data not shown), indicating that EC are sensitive to changes in the extracellular porphyrin profile. Nevertheless, these findings do not rule out the possibility of a direct intervention of PPIX to hamper angiogenesis from outside the cell. Consistently, previous works identified the porphyrin cationic analogue 5,10,15,20-tetrakis(methyl-4-pyridyl)-21H,23H-porphine-tetra p-tosylate salt (TMPP) to impair vascular growth likely due to inhibition of key angiogenic signalling pathways as Vascular Endothelial Growth Factor Receptor 2 (VEGFR2) and Fibroblast Growth Factor Receptor 1 (FGFR1) [42]. Moreover, porphyrin-enriched CM derived from ALA-treated EC was able to inhibit angiogenesis similarly to ALA or its downstream derivative PPIX.
While our data indicate a detrimental role of porphyrins in EC, they do not exclude a direct contribution of heme to the phenotypes observed following ALA treatment. Extracellular heme levels increased both after ALA and HX treatment; however, the angiogenic outcomes were opposite. Together, these observations suggest that endothelial-derived extracellular heme can exert a pro-angiogenic effect, provided that the process remains tightly controlled—namely in the absence of porphyrin accumulation—and occurs in the presence of an extracellular heme chelator such as HX. Conversely, HX supplementation combined with ALA induced a selective increase in extracellular porphyrins, while heme levels remained largely unchanged, further reducing EC proliferation compared with ALA alone. Similar results were obtained upon cotreatment with human serum albumin (HSA), consistent with its established role in facilitating porphyrin export through ABCG2 [18]. In this view, our data propose distinct biological effects exerted by extracellular porphyrins (e.g. PPIX) and heme. Although both molecules share the same tetrapyrrole ring structure, endogenous porphyrins are essentially planar molecules, whereas heme, due to the central coordination of iron, adopts a slightly non-planar, curved configuration. This structural difference may underlie their diverse biological effects. In addition, differences porphyrins side chains could further address their specificity and biological effects [43]. Further studies are needed to elucidate the distinct roles of porphyrin intermediates, as well as heme, in cellular processes. While our data suggest that antiangiogenic effect due to ALA is primarily mediated by porphyrins rather than heme, the differential impact of extracellular heme and porphyrins warrants further investigation in future studies. Notably, alterations in heme metabolism are expected to significantly influence EC metabolism, a major driver of angiogenesis. Indeed, heme is a crucial cofactor in the electron transport chain (ETC), and its synthesis represents a central hub in cell metabolism [2, 44, 45]. Consistently, inhibition of endothelial heme synthesis has been linked to defective mitochondrial respiration and impaired glycolysis, which may explain the anti-angiogenic effects observed upon SA treatment [5, 6]. Conversely, overdriven porphyrin and heme production following ALA treatment has been linked to increased sensitivity of cancer cells to metformin, suggesting a greater reliance on oxidative metabolism [46]. Similar metabolic alterations, including dysregulated TCA cycle flux and perturbed interconnected pathways, have also been reported in patients with porphyria [47]. Nevertheless, further studies are needed to determine how these metabolic phenotypes affect angiogenesis and to disentangle the specific contributions of porphyrins, heme, and heme synthesis flux to EC metabolism.
So far, most research on porphyrins in angiogenesis has focused on treating ocular neovascularization using anti-angiogenic drugs in combination with photodynamic therapy (PDT) [48]. In PDT, porphyrin analogues as verteporfin are exogenously delivered as photosensitizer agents to selectively kill angiogenic EC by subsequent laser ablation. Nevertheless, the pain and discomfort associated with PDT highlights the need for less invasive treatments to address retinopathies. Notably, a recent work pointed out the efficacy of using verteporfin to treat ocular neovascularization diseases in the absence of photoactivation [49]. In this perspective, ALA delineates as an intriguing strategy to inhibit pathological angiogenesis through stimulation of endogenous porphyrin production without resorting to PDT. Notably, ALA is a drug approved by The Food and Drug Administration (FDA) that is both safe and well tolerable by organism making it a valuable clinical opportunity [50–52]. Remarkably, ALA was effective in hampering aberrant neovascularization in a mouse model of OIR, making ALA a promising molecule for the treatment of ocular neo-vascularization diseases, including ROP and PDR.
Moreover, these data provide a foundation for future research to assess the effects of ALA in cancer biology, where aberrant vascularization is a critical hallmark. Indeed, ALA was also effective in tumor-associated EC (i.e. BTEC) and porphyrin accumulation has been previously described as a unique feature of the tumor microenvironment, including in tumor-derived EC [27, 50, 53–56]. Furthermore, high doses of ALA have been shown to inhibit proliferation of multiple cancer cell models [46]. Collectively, these data suggest that porphyrin overdrive might represent a metabolic vulnerability that can be exploited for cancer therapy.
In conclusion, this study provides first evidence of the biological relevance of endogenous porphyrin metabolism in the vascular system. Moreover, ALA, by promoting porphyrin overproduction in EC, emerges as a promising agent for inhibiting aberrant neovascularization. Further research is required to elucidate the bioactive properties of porphyrins, particularly in relation to the vascular system.
Methods
Cell culture
Human adult dermal microvascular endothelial cells (HMEC-1, RRID:CVCL_0307) were purchased by ATCC, propagated in MCDB131 (Thermo Fisher Scientific, Waltham, MA USA, catalog n°10,372,019) supplemented with 10% heat-inactivated low-endotoxin FBS (GIBCO by Thermofisher Scientific, Waltham, MA USA, catalog n10270106), 10 mM GlutaMAX™ Supplement (Thermo Fisher Scientific, Waltham, MA USA, catalog n°35,050,061), 10 ng/mL Epidermal Growth Factor (Thermo Fisher Scientific, Waltham, MA USA, catalog n° PHG0315), 1 µg/mL Hydrocortisone-Water Soluble (Sigma Aldrich, St. Louis, MO USA, catalog n° H0396) and used up to passage 12. Breast tumor-derived endothelial cells (BTEC) from human breast lobular-infiltrating carcinoma biopsy were isolated and characterized in the laboratory of Professor Benedetta Bussolati, Department of Molecular Biotechnology and Health Sciences, University of Torino, Italy [57, 58]. BTECs were maintained in EndoGRO-MV-VEGF Complete Culture Media Kit (SCME003 Merck Millipore). All cell media were ordinarily supplemented with antibiotics (100U/ml penicillin and 100 mg/ml streptomycin; GIBCO by Thermo Fisher Scientific, Waltham, MA USA, catalog n° 15,140,122). Cells were maintained in a 37 °C and 5% CO2 air incubator and routinely screened for the absence of mycoplasma contamination. 5-Aminolevulinic acid hydrochloride (ALA) and succinyl acetone (SA) were purchased by Sigma-Aldrich, St. Louis, MO, USA (Product Numbers are A3785 and D1415 respectively). Hemopexin (HX) and HSA (human serum albumin) were provided by CSL Behring SPA. Protoporphyrin IX (PPIX) was purchased by Frontier Specialty Chemicals (Catalog number P562-9).
Heme and porphyrin fluorimetric assay
Intracellular and extracellular concentrations of heme and porphyrins were measured by using spectrophotometry, as previously reported [59]. Firstly, EC were lysed in TBS 1% Triton and protein concentration was read using Bradford assay. Next, 10–50 ug were used to perform the intracellular quantifications. Culture media were instead centrifuged to remove cell debris and 100ul were read in the assay. Porphyrin concentration was determined at the Soret maximum (405 nm), and fluorescence was measured using an excitation filter AFC 405 nm and an emission filter in the following wavelengths’ range: 660–720 nm. To measure heme concentration, cellular proteins or culture media were firstly dissolved into 1 ml of 2 M oxalic acid. Each sample was then split into two tubes. The first tube was directly read at the spectrophotometer to measure non-heme fluorescence (background). The second tube was heated at 95 °C for 30 min to remove iron from heme and then heme-porphyrin fluorescence was read at the spectrophotometer (excitation filter: AFC 405 nm; emission filter: 660–720 nm). Heme content was obtained by subtracting the fluorescence of not-boiled samples (non-heme porphyrins) from that of boiled-samples. Fluorescence was read by using a Glomax Discover microplate reader (Promega Corporation). Data were normalized to total protein concentration (for intracellular heme/porphyrin quantifications), cell number (in the case of cell culture media) or choroid sprout’s area (in the case of culture media derived from choroids DIV7). All data are expressed as a fold change (FC) respect to the control condition or time zero (T0).
Porphyrin fluorescence imaging
HMEC were plated on IBIDI slides (IBIDI) in complete MCDB131 medium. Once attached, CMFDA 1:5000 (Thermo Fisher Scientific, Waltham, MA USA, catalog n°C2925) was added to stain cytoplasm. 5 mM ALA was supplemented to HMEC for 24 h to promote porphyrin synthesis. Endogenous porphyrin fluorescence was visualized with excitation filter 405 nm and emission filter 600–700 nm. Confocal images acquisition was performed with a Leica TCS SP8 confocal system (Leica Microsystems) with a PL APO 20 × /0.75 CS2.
To evaluate ALA delivery and conversion into porphyrins in the developing postnatal retina, P5 mice were administered 20 mg/kg ALA or vehicle (PBS) via intragastric injection. After 3 h, mice were sacrificed, eyes were collected, and retinas were dissected. Porphyrin fluorescence was analysed in freshly dissected retinas using Leica Thunder Imager (Leica Microsystems) equipped with a 4 × objective, with excitation at 405 nm and emission collected between 600 and 700 nm.
HPLC–DAD-MS/MS analysis
Samples of medium obtained from both in vitro and ex vivo experiments were purified and concentrated using SFE cartridges with the same stationary phase as the column employed for subsequent analyses. HPLC separation was performed on an Agilent Technologies 1200 system, equipped with a diode-array detector (DAD) and a 6330 Series Ion Trap LC–MS system (Agilent Technologies, Santa Clara, CA, USA) coupled to an electrospray ionization (ESI) source. Samples were injected via an automated and temperature-controlled (4 °C) injection module (Agilent Technologies, Santa Clara, CA, USA) fitted with a 200 µL loop. A porphyrin standard mixture was used for HPLC peak identification. Porphyrins were separated on a Hypersil BDS column (5 µm particle size, 250 × 4.6 mm i.d.; Thermo Hypersil-Keystone, Runcorn, UK) through gradient elution. The mobile phase consisted of: (A) 9% (v/v) acetonitrile in 1 M ammonium acetate/acetic acid buffer, pH 5.16, and (B) 9% (v/v) acetonitrile in methanol. The elution program ran from 0 to 60 min, transitioning from 100% A (0% B) to 90% B (10% A). The flow rate was set at 0.5 mL/min, with a source temperature of 110 °C and a cone voltage of 90 V. Data were acquired in positive ion mode and in Multiple Reaction Monitoring (MRM) mode, and data were collected using a collision energy of 35 eV with argon as the collision gas. The area under the curve (AUC) for each compound of interest was then normalized to the median value of biological and technical replicates, log-transformed (base 10), and scaled using Pareto scaling (mean-centered and divided by the square root of the standard deviation of each variable). The resulting data from MS/MS acquisition were normalized based on protein content, determined via the Bradford assay for in vitro experiments, and by retinal section area for ex vivo assays. The processed data were used to construct orthogonal PCA models and HeatMaps coupled with cluster analysis, employing MetaboAnalyst v.6.
Western blot analysis
To assess protein expression, HMEC and BTEC were lysed by rotation for 30 min at 4 °C in RIPA buffer (150 mM NaCl, 50 mM Tris–HCl pH 7.5, 1% Triton X-100, 0.5% Sodium deoxycholate, 0.1% SDS, 1 mM EDTA). The buffer was freshly supplemented with 1 mM phosphatase inhibitor cocktail (Sigma Aldrich, St. Louis, MO USA, catalog n° P0044), 1 mM PMSF (Sigma Aldrich, St. Louis, MO USA, catalog n° 93,482-50ML-F), and protease inhibitor cocktail (La Roche, Basel, CH, catalog n° 04693116001). The cell lysate was clarified by centrifugation for 10 min at 4 °C. Protein concentration in the supernatant was assessed by Bradford assay. For ALAS1 and HO-1 protein detection, 10 ug of protein extracts were incubated 5 min at 95 °C in 4X laemmli buffer freshly supplemented with 8% 2-mercaptoethanol. Before loading on 4–15% mini PROTEAN TGX precast gel (Bio-Rad, Hercules, CA USA, catalog n°4,568,084). The primary antibodies and dilutions are as follows: ALAS-H (Santa Cruz Biotechnology, Dallas, TX USA, catalog n° sc-137093; 1:1000), HO-1 (Enzo Life Sciences, Inc, Farmingdale, NY USA catalog n° ADI-SPA-896; 1:300), α-TUBULIN (Cell Signaling Technology, Inc., catalog n° 2144, 1:1000), vinculin (homemade, 1:10,000). The revelation was assessed using the ChemiDoc Imaging System (Bio-Rad, Hercules, CA USA).
Gene expression
Total RNA was extracted using PureLink RNA Mini Kit (Thermo Fisher Scientific, Waltham, MA USA) and 0.5–1 ug of total RNA was transcribed into complementary DNA (cDNA) by High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA USA). Quantitative real-time reverse transcription PCR (qRT-PCR) was performed using either predesigned TaqMan™ Gene Expression Assays (Thermo Fisher Scientific Waltham, MA USA) or SYBR Green dye-based PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific, Waltham, MA, USA). TaqMan assay IDs (Tables 1 and 2) and gene-specific primers used for SYBR Green–based detection (Tables 3 and 4) are listed below. To detect FLVCR1a expression, specific primers and probes were designed using Primer Express Software Version 3.0 (Thermo Fisher Scientific, Waltham, MA USA). qRT-PCR was performed on a QuantStudio™ 6 Flex Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA USA), and the analyses were done using QuantStudio Real-Time PCR software. Transcript abundance, normalized to 18 s messenger ribonucleic acid (mRNA) expression, is expressed as a fold increase over a calibrator sample.Table 1. Mouse Taqman assaysGene nameAssay IDAlas1Mm01235914_m1FechMm00500394_m1Hmox-1Mm00516005_m1HpxMm00457510_m1Cdh5Mm00486938_m1Table 2Human Taqman assaysGene nameAssay IDABCG2Hs01053790_m1ALAS1Hs00167441_m1FECHHs00164616_m1HPXHs00167197_m1FLVCR2Hs00215055_m1Table 3Mouse qPCR primers (SYBR green assay)Gene nameForward primerReverse primerAladCTGCGCTGTGTCCTGATCTCTGCAGAGCCCTGTTCATCHmbsAGAAAAGTGCCGTGGGAACTGTTGAGGTTTCCCCGAATUrosCAAGGGAGCCTGAAGAGCTAGGAGGGACTGAAAAACGTGACpoxCTTTGAAGTAGAGGAAGCTGACGTTCAAGTATGTCGGTGTGAGGTSlc15a1GACCACAATGGCAGTCCTGACAATCGCCACCAAACGCAGACACASlc15a2GTGAAGGCACTGACCAGGATACTGTTAGCTTGCAGAGTCCACCGTable 4Human qPCR primers (SYBR green assay)Gene nameForward primerReverse primerALADCGCTCTTACGCGGTCTGTGGCTGCAGGCTCTGTCTGHMBSTGTGGTGGGAACCAGCTCTGTTGAGGTTTCCCCGAATUROSTGATCCCTGTTTTATCGTTTGACCCGTAATCTTCAGGATGAGAAURODACAGTGGCCCCAAAGAAAGCACCAACTGCCCGATCTCCPOXCGCCAACTTTTCTGTGGACCGAAAACACACCCATCTTGAPPOXTCAGCCTCCAGGCAGAAGCTGAAGCTGGAATGGCACTAABCB6TGAAGAGGACCAAGATGTGGACCAAAATCTCGCCAGGTAGA
Mitochondrial extraction and ALAS activity
Mitochondria were extracted as reported in [60]. Cells were lysed in 0.5 ml mitochondria lysis buffer (50 mM Tris–HCl, 100 mM KCl, 5 mM Mg Cl2, 1.8 mM ATP, and 1 mM EDTA at pH7.2), supplemented with protease inhibitor cocktail III (Sigma Aldrich), 1 mM phenylmethylsulfonyl fluoride (PMSF), and 250 mM sodium fluoride (NaF). Samples were clarified by centrifugation at 650xg for 3 min at 4 °C. Supernatants were collected and centrifuged at 13,000xg for 5 min at 4 °C. Pellets, containing mitochondria, were washed once with lysis buffer and resuspended in 0.25 mL mitochondria resuspension buffer (250 mM sucrose, 15 mM K2HPO4, 2 mM MgCl2, and 0.5 mM EDTA). Part of the mitochondrial sample was sonicated and used to measure the mitochondrial proteins with the BCA Protein Assay kit (Sigma Aldrich) and for quality control: 10ug of each sonicated sample was analyzed by SDS-PAGE and immunoblotting with an anti-porin antibody (Abcam; clone 20B12AF2) to confirm the presence of mitochondrial proteins in the extracts. The remaining 200uL was used for ALAS activity measurement, according to [61]. A total of 10 uL of the reaction’s product was injected into the Waters Acquity ultra-performance liquid chromatography (UPLC) system, equipped with a binary solvent manager, sample manager, photodiode array detector (PDA), fluorescence detector, column heater, and an Acquity UPLC BEH C18, 1.7uM, 2.1 × 100 mm column. Detection of ALA-derivative was performed according to [61], setting the detector with λ excitation of 370 nm and λ emission of 460 nm, and the range of the PDA scanner was set between 210 and 500 nm. The results were converted into nmol/min based on a calibration curve and expressed as nmol/min/mg mitochondrial proteins.
Cell proliferation
To test the in vitro proliferative potential, 2500 cells were plated on a 96 multi-well plate. After 1 day, treatments were applied on cells and proliferation monitored with Incucyte SX5 Live-Cell Analysis System (Sartorius, Göttingen, Germany). Proliferation was evaluated as the fold increase of 96-well confluence respect to T0.
Wound-healing assay
To perform migration assay, 35 × 10^4^ HMEC or 30 × 10^4^ BTEC were pleated on a 96 multi-well. A scratch wound was made in cell monolayer after 1 day, using a sterile 200ul pipette tip. Cells were washed twice with PBS to remove detached cells and treatment supplemented for 24 h. Cell migration was monitored with Incucyte SX5 Live-Cell Analysis System (Sartorius, Göttingen, Germany) and evaluated as the fold change in cell-covered area over time, relative to the initial area occupied by cells (T0) or as percentage of wound closure respect to T0.
Tubulogenesis assay
In vitro formation of capillary-like structures of HMEC was studied on Geltrex™ LDEV-Free Reduced Growth Factor Basement (Thermo Fisher Scientific, Waltham, MA USA, catalog n°A1413202) in 24-well plates. Cells (3.5 × 10^4^ cells/well) were seeded onto Geltrex-coating in complete MCDB 131 Medium (Thermo Fisher Scientific, Waltham, MA USA, catalog n°A1413202). Cell organization onto Matrigel was monitored with Incucyte SX5 Live-Cell Analysis System (Sartorius, Göttingen, Germany). Images were acquired after 24 h. At least three independent experiments were done for each experimental condition. ImageJ’s Angiogenesis Analyzer was used to analyze the number of nodes and master junctions as well as total length.
Mice treatments
Neonatal C57BL/6 mice received daily intragastric injections of ALA dissolved in PBS. Two treatment regimens were applied: a higher dose of 20 mg/kg from P1 to P5, with mice euthanized at P5, and a lower dose of 5 mg/kg from P1 to P3, with mice euthanized at P7 or P21. Control mice received PBS alone. All the mice were provided with food and water ad libitum. All experiments with animals were approved by the Italian Ministry of Health. Mice were maintained according to international (EU Directive 2010/63/EU) and national guidelines (Italian decree No 26 of the 4th of March 2014, UK Home Office guidelines).
EC isolation from developing retina/choroid
Eyes were isolated from wild-type animals at multiple developing stages (i.e. P3, P6, P15) and the retina or the choroidal-retinal pigment epithelium-sclera complex, were dissected. After rinsing in Dulbecco’s Phosphate Buffered Saline (DPBS), tissues were incubated with the Neural Tissue Dissociation Kit-Postnatal Neurons (Miltenyi Biotec, Bergisch Gladbach, DE, catalog n 130-092-628) according to the manufacturer instruction. During incubation steps, cells were dissociated by pipetting multiple times. Cells suspension was then filtered through a 40 mm cell strainer (Corning Life Sciences, Corning, NY USA, catalog n 352,340) and rinsed in 0.5% BSA DPBS. Then, the cells were centrifuged (300 g, 10 min) and EC were isolated through MACS Technology by using nano-sized MicroBeads, following the manufacturer instructions. CD31 MicroBeads (Miltenyi Biotec, Bergisch Gladbach, DE, catalog n 130-097-418) was used to isolate EC (CD31+ fraction) from the retina and choroid; the CD31⁻ fraction was referred as the non-endothelial parenchymal compartment. Each sample is a pool of a minimum of 3 mice (i.e. 6 eyes).
Choroid sprouting assay
Choroid sprouting assay was performed as described in [15]. Briefly, animals were anesthetized and then euthanized by cervical dislocation. Eyes were immediately enucleated and kept in ice-cold medium before dissection. The choice of the medium was identical to the one used for later incubation. After removing the cornea and the lens from the anterior of the eye, the choroid-scleral complex was separated from the retina and cut into approximately 1 mm^2^. 30 µL of Geltrex™ LDEV-Free Reduced Growth Factor Basement (Thermo Fisher Scientific, Waltham, MA USA, catalog n°A1413202) were used to coat 24-well plates. After seeding the choroids into the Geltrex™ drop, plates were incubated at 37 °C in a cell culture incubator without medium for 30 min in order for the Geltrex™ to solidify. 500 µL of EndoGRO-MV-VEGF Complete Culture Media Kit (SCME003 Merck Millipore) were then added to each well and incubated at 37 °C with 5% CO2 for 72 h before any treatment. Medium was changed every 24 h for the first 3 days. Individual explants were monitored every 12 h using Incucyte SX5 Live-Cell Analysis System (Sartorius, Göttingen, Germany). The vascular growth of the sprouts was measured by using ImageJ.
Aorta ring assay
Aorta ring assay was performed as described in [62]. Briefly, thoracic aortas were removed from mice killed by cervical dislocation and immediately transferred into cold medium. The periaortic fibroadipose tissue was carefully removed with fine microdissecting forceps and scissors and 1 mm-long aortic rings were sectioned. Ring-shaped explants of mouse aorta were then embedded in 30 µL of Geltrex™ LDEV-Free Reduced Growth Factor Basement (Thermo Fisher Scientific, Waltham, MA USA, catalog n°A1413202) into 24-well plates. Plates were incubated at 37 °C in a cell culture incubator without medium for 30 min in order for the Geltrex™ to solidify. Additional 30 µL of Geltrex™ were put above the aorta fragments and left again to solidify. 500 µL of EndoGRO-MV-VEGF Complete Culture Media Kit (SCME003 Merck Millipore) were then added to each well and incubated at 37 °C with 5% CO2 for 72 h before any treatment. Medium was changed every 24 h for the first 3 days. Individual explants were acquired at day 9 using Incucyte SX5 Live-Cell Analysis System (Sartorius, Göttingen, Germany). The vascular growth of the sprouts was measured by using ImageJ.
Whole-mount retina immunofluorescence staining
Eyes were fixed for 2 h (hrs) in paraformaldehyde (PFA) 4% at 4 °C. Retina dissection was performed at postnatal day 7 or 21 as previously described [63]. After retina dissection, 4 to 5 radial incisions were done using spring scissors to create a ‘petal’ shape. The retinas were put in cold (− 20 °C) methanol for at least 20 min (until turning white) before proceeding with immunostaining. Next, the retinas were washed in phosphate-buffered saline solution (PBS) 1X and then covered with 100 μl of permeabilization/blocking solution (PBS + 0.3% Triton + 0.2% BSA) + 5% BSA on gentle shake for 1 h at room temperature (RT). Afterwards, the retinas were incubated with 100 μl of selected primary antibodies overnight at 4 °C with gentle shaking. Then, the retinas were wash retinas 4 times for 10 min in PBS + 0.3% Triton (PBSTX) and incubated with the appropriate fluorescent secondary antibody (diluted 1/200 in PBSTX) overnight at 4 °C. Finally, retinas were washed in PBS and mounted on a cover glass. The following primary antibodies were used: purified Rat Anti-Mouse CD31 (550,274 BD Pharmingen™); Anti-Histone H3 (phospho S10) antibody (ab47297 Abcam). Network morphology was analyzed using specific plugins of Fiji, a distribution of the open-source software ImageJ [64, 65]. Radial extension of the retinal vasculature was measured as the distance from the optic nerve to the vascular front. Retinal radius was determined by measuring the full length of each retinal leaflet (petal) from the optic nerve head to the peripheral edge. Sprout numbers were quantified by counting individual sprouts at the angiogenic front and normalizing the values to the angiogenic front length. Confocal image acquisition was performed with a Leica TCS SP8 confocal system (Leica Microsystems) with a PL APO 20x/0.75 CS2 or HC PL APO 40x/1.30 OIL CS2 objectives.
Oxygen-induced retinopathy (OIR) mouse model
OIR was induced as described [66]. Briefly, nursing mothers and their pups were maintained in 85% oxygen in a custom-made plexiglass chamber from P8 to P11 using a ProOx P110 oxygen controller (BioSpherix) and returned to room air for 5 days to P16. Mouse pups received daily intraperitoneal injection of 20 mg/kg ALA in PBS or vehicle only from P12 to P15. Retinas were isolated at P16 for immunolabeling with antibodies anti collagen IV (2150-1470 Bio-Rad) and alpha smooth muscle actin (αSMA; C6198 Sigma-Aldrich) and images were acquired with a stereo fluorescent microscope (SZX16; Olympus) equipped with a digital camera (Hamamatsu Photonics). Artero-venous shunts were manually quantified, whereas central avascular (AV) and neovascular (NV) areas were measured with the open-source software ImageJ as previously described [19].
Single cell RNAseq (scRNAseq) analysis
The mouse P6 retina dataset (GSE175895) was downloaded from GEO NCBI (https://www.ncbi.nlm.nih.gov/geo/) [67]. R Studio v.1.3.1056 and Seurat version 3.2.3 were used to explore scRNAseq data [68]. Cells containing less than 500 feature counts and genes expressed in less than 10 cells were removed. Downstream analysis included data normalisation (“LogNormalize” method and scale factor of 10,000) and variable gene detection (“vst” selection method, returning 2,000 features per dataset). The principal components analysis (PCA) was performed on variable genes, and the optimal number of principal components, PCs, was chosen using the elbow plot. The selected PCs were used for Louvain graph-based clustering. Uniform manifold approximation and projection (UMAP) was chosen as a non-linear dimensionality reduction method, and cluster cell identity was assigned by manual annotation based on known marker genes. Each relevant gene was then examined using the FeaturePlot and VlnPlot functions.
Statistics
Statistical comparisons were conducted in Prism (GraphPad Software, Inc., La Jolla, CA). Results are expressed as mean ± SEM (standard error mean). Whenever comparing the means of two groups, a Student’s t-test was used. Whenever comparing the means of three or more groups based on a single independent variable, a one-way ANOVA with Tukey’s test was used. Whenever comparing the means of three or more groups based on two independent variables, a two-way ANOVA with Sidak’s test was used. The level of significance was set at *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
Supplementary Information
Supplementary Material 1.
The reference list from the paper itself. Each links out to its DOI / PubMed record.