If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Biogenesis of lipid droplets (LDs) in various cells plays an important role in various physiological and pathological processes. However, the function of LDs in endothelial physiology and pathology is not well understood. In the present work, we investigated the formation of LDs and prostacyclin (PGI2) generation in the vascular tissue of isolated murine aortas following activation by pro-inflammatory factors: tumor necrosis factor (TNF), lipopolysaccharides (LPS), angiotensin II (AngII), hypoxic conditions, or oleic acid (OA). The abundance, size, and biochemical composition of LDs was characterized based on Raman spectroscopy and fluorescence imaging. We found that blockade of lipolysis by the adipose triglyceride lipase (ATGL) delayed LDs degradation and simultaneously blunted PGI2 generation in aorta treated with all tested pro-inflammatory stimuli. Furthermore, the analysis of Raman spectra of LDs in the isolated vessels stimulated by TNF, LPS, AngII, or hypoxia uncovered that these LDs were all rich in highly unsaturated lipids and had a negligible content of phospholipids and cholesterols. Additionally, by comparing the Raman signature of endothelial LDs under hypoxic or OA-overload conditions in the presence or absence of ATGL inhibitor, atglistatin, we show that atglistatin does not affect the biochemical composition of LDs. Altogether, independent of whether LDs were induced by pro-inflammatory stimuli, hypoxia, or oleic acid, and of whether they were composed of highly unsaturated or less unsaturated lipids, we observed LDs formation invariably associated with ATGL-dependent PGI2 generation. In conclusion, vascular LDs formation and ATGL-dependent PGI2 generation represent a universal response to vascular pro-inflammatory insult.
Endothelium covers the innermost layer of blood and lymphatic vessels and fulfills many functions in maintaining cardiovascular homeostasis. The endothelial barrier between the vascular lumen and tissue, regulates the vessel diameter and blood fluidity, takes part in the immune system functioning and regulates the endothelial inflammation(
Aird, W.C., and M. Laubichler. 2007. Introductory essay: evolution, comparative biology, and development. In Endothelial Biomedicine. Cambridge University Press. 21-28.
). Over the course of evolution, endothelial cells (ECs) as the first cells in the vascular lumen subjected to alteration in the blood composition, have promoted the development of highly reproducible mechanisms of counteracting disturbances in homeostasis. Depending on the chain of events, the endothelial response to stimulant can be considered as activating or dysfunctional(
). The functional reaction of the endothelium in response to stress factors such as cytokines, toxins, hypoxia, and so forth should be viewed as a consequence of the physiological response of the endothelium, and then the endothelium is called activated. Nevertheless, variable stress factors might result in permanent alterations in endothelial phenotype viewed as dysfunctional (
). It is of importance in endothelial biology to delineate the mechanisms that offset endothelial activation and prevent endothelial dysfunction for example in response to endothelial pro-inflammatory insult.
Endothelial inflammation may be triggered by pro-inflammatory factors, for example, tumor necrosis factor (TNF), lipopolysaccharides (LPS), or angiotensin II (AngII). TNF-induced endothelial inflammation is mediated by the TNFR1 and death domain protein (TRADD), which, in turn, binds receptor-interacting protein 1 (RIP1) and TRAF2 forming the TRADD-RIP1-TRAF2 complex, activating pro-inflammatory endothelial response through the NF-κB pathway(
). Bacterial endotoxin lipopolysaccharide (LPS), a major component of the outer membranes of Gram-negative bacteria, acts on the endothelium through the soluble form of CD14 (sCD14)(
) or by TLR-4 receptor and the subsequent activation of nuclear protein complex NF-κB(5). The effects of AngII are mediated by its binding into the angiotensin type 1 (AT1R) or type 2 (AT2R) receptors with opposite actions(
): AT1R is primarily responsible for the pro-hypertensive and pro-inflammatory activities of AngII, whereas the AT2R is reported to induce vasoprotective effects. Activation of the AT1R increases the production of reactive oxygen species (ROS) and activates nuclear factor kappa B (NF-kB), and in consequence, inactivates nitric oxide (NO) and reduces its production by uncoupling endothelial NO synthase (eNOS)(
). Hypoxia is able to activate the endothelial cells via activation of hypoxia-inducible factor (HIF), and thereby initiates a cascade of reactions including diminished secretion of NO, and increased production of ROS(
). The common result of the action of aforementioned proinflammatory factors: TNF, LPS, AngII hypoxia or OA is the activation of vascular endothelium and the subsequent activation of nuclear protein complex NF-B responsible for the transcription of relevant genes including cyclooxygenase-2 resulting in increased synthesis of eicosanoids(
), the increased generation of prostacyclin (PGI2) in activated endothelium was associated with the formation of lipid droplets.
Lipid droplets (LDs) are spherical cellular organelles rich in triacylglycerols and cholesteryl esters intrinsically related to physiological cellular energy storage and metabolism(
). Although the general knowledge on endothelial LDs is growing, neither the precise function of LDs nor the pathway of their biogenesis have been fully revealed. The biogenesis of LDs was recently suggested to represent an integral part of vascular inflammation in TNF-activated isolated vascular wall ex vivo(
Rac1 regulates lipid droplets formation, nanomechanical, and nanostructural changes induced by TNF in vascular endothelium in the isolated murine aorta.
), or in LPS-(15), TNF-(16) activated ECs in vitro. It was shown that the formation and degradation of endothelial LDs were regulated by Rac1 and adipose triglyceride lipase (ATGL), respectively, in TNF-stimulated isolated murine aorta(
Rac1 regulates lipid droplets formation, nanomechanical, and nanostructural changes induced by TNF in vascular endothelium in the isolated murine aorta.
). These findings document the active role of LDs in endothelial homeostasis, however, further studies of the characteristics of vascular LDs are needed to reveal the complete pathophysiological role of LDs in the endothelium.
Here, we aimed to better characterize the formation of LDs in response to TNF, LPS, AngII, hypoxia, or OA in endothelial cells (ECs) within the isolated murine aorta. In particular, in response to each of the stimuli, the abundance, size and biochemical composition of LDs was characterized as well as ATGL-dependent PGI2 generation. The biochemical composition of LDs in this report was analyzed based on Raman imaging. Raman spectroscopy is a label-free and non-destructive technique that could provide valuable information about the biochemical compositions of the biological samples. In particular, this technique was used previously to study endothelium(
The distinct phenotype of primary adipocytes and adipocytes derived from stem cells of white adipose tissue as assessed by Raman and fluorescence imaging.
). Due to large Raman scattering cross-section for lipids, bands originating from lipids can be clearly detected in Raman spectra, what enables a reliable detection of lipid structures inside biological samples. Importantly, Raman imaging allows measurements of intracellular lipids without externally labelling them, and provides a possibility for the detection, indication of precise localization (endothelium vs. smooth muscle cells), quantitative assessment and, most importantly, the analysis of the biochemical composition of lipid droplets in biological samples including vascular wall(
Rac1 regulates lipid droplets formation, nanomechanical, and nanostructural changes induced by TNF in vascular endothelium in the isolated murine aorta.
). In this paper, the specificity of Raman imaging enabled recognition of newly formed LDs in the activated endothelium and their biochemical characterization in response to given pro-inflammatory stimulus, followed by discrimination of LDs using hierarchical cluster analysis.
Materials and methods
All experimental procedures involving animals were conducted according to the Guidelines for Animal Care and Treatment of the European Communities and the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). All procedures were approved by the 2nd Local Ethical Committee on Animal Experiments. Male C57BL/6J mice (aged 8–12 weeks) were purchased from Mossakowski Medical Research Institute, Polish Academy of Sciences, Warsaw, Poland, and were housed in a temperature-controlled environment (22–25°C), 12-hour light/day cycle and unlimited access to standard laboratory diet.
Mice were euthanized by an intraperitoneal injection (i.p.) of a mixture consisting of ketamine and xylazine (100 mg ketamine/10 mg xylazine/kg body weight). The chest (albo chest cavity) was exposed and the thoracic aorta was dissected. Then, the aorta was cleaned from the surrounding tissue, cut into rings and transferred into minimal essential medium (MEM, Sigma Aldrich) supplemented with 1% MEM vitamins (Sigma Aldrich), 1% antibiotics (penicillin 10,000 U/mL and streptomycin 10,000 μg/mL and Amphotericin B 25 μg/mL; Thermo Scientific), 1% non-essential amino acids (Sigma Aldrich), and 20% fetal bovine serum (Thermo Scientific. The aorta was incubated in the presence of tumour necrosis factor (TNF; 10 ng/ml, 24h, N=5), lipopolysaccharides from Escherichia coli O111:B4 (LPS; 10 μg/ml, 24h, N=5), angiotensin II (AngII; 1μM, 24h, N=6), hypoxic condition (5% O2, 24h, N=6), oleic acid (OA; 1 mM, 24h, N=6), atglistatin (Atgl; 10 μM, 24h, N=6), TNF with atglistatin (TNF+Atgl; 10 ng/ml and 10 μM, respectively, 24h, N=6), lipopolysaccharides with atglistatin (LPS+Atgl; 10 μg/ml and 10 μM, respectively, 24h, N=6), angiotensin II with atglistatin (AngII+Atgl; 1μM and 10 μM, respectively, 24h, N=6), hypoxic condition with atglistatin (hypoxia+Atgl; 5% O2 and 10 μM, respectively, 24h, N=5), oleic acid with atglistatin (OA+Atgl; 1 mM and 10 μM, respectively, 24h, N=5). The untreated aorta was maintained in the medium for 24h and was used as a control (N=6).
Immunostaining of aorta en face
For immunostaining imaging of aorta en face, the resected and split-open arteries were tightly glued to the Cell-Tak®-coated microscopic glasses and preserved by a 10-min soak in 4% paraformaldehyde. Tissue samples were blocked with TNB blocking buffer (0.1 M Tris-HCl pH 7.5, 0.15 M NaCl, and 0.5% (w/v) blocking reagent; PerkinElmer FP1020) for 3-4h, and then incubated with CD31 antibody (Abcam, 1:50) diluted in TNB blocking buffer overnight at 4°C. As secondary antibody, Alexa Fluor 647 nm goat-anti-rabbit (Jackson Immuno Research; 3:600) was used at room temperature for 3h. BODIPY 493/503 diluted in PBS at the final concentration of 0.1 mg/ml was applied for 1h to delineate LD, and Hoechst 33258 (Sigma; 1:1000) was used to highlight nuclei. Samples were measured by 40× magnification objective on CQ1 Confocal Quantitative Image Cytometer (Yokogawa).
Raman imaging of aorta en face
For Raman imaging of aorta en face, the resected and split-open arteries were tightly glued to the Cell-Tak®-coated calcium fluoride surface and preserved by a 10-min soak in 4% buffered formalin. Raman imaging was performed in fluid with a Confocal Raman Imaging system (WITec alpha 300, Ulm Germany) supplied with a UHTS 300 spectrograph (600 grooves·mm−1 grating, resolution of 3 cm−1) and a CCD detector (Andor, DU401A-BV-352) using a 63× water immersion objective (Nikon Fluor, NA=1, Netherlands). A laser excitation wavelength of 532 nm, a laser power of ca. 30 mW, and an integration time of 0.4 s per spectrum were used in all measurements. The nominal minimal lateral and depth resolution for our setup is 0.32 and 0.53 μm, respectively, and sampling density of 0.38-0.50 and 0.5-1.0 μm in x/y and z direction, respectively, were used. Depth profiling of the tissue was obtained by multiple imaging of the same line in several layers of the sample. Data matrices were analysed using WITec Project software (background subtraction using a polynomial of degree 3 and the automatic removal of cosmic rays). The analysis of the spectra was supported by Cluster Analysis (K-means, Manhattan distance, WITec Project Plus). For study of heterogeneity of endothelial LDs within aorta, the single Raman spectra were extracted from the centre of each LD, and then averaged. The OPUS 7.2 program was used for calculations of the integral intensity of the bands at ca. 1660 and 1445 cm-1 in the 1733-1767 and 1563-1712cm-1 spectral ranges, respectively. Integration was performed using method D: the integral was defined by the wavenumber limits and the horizontal baseline determined by a chosen baseline point (OPUS 7.2). The single spectra were used for HCA using the Ward’s algorithm in OPUS 7.2 applied to the following spectral regions: 600–3030 cm-1. These spectral regions were chosen because of the optimal separation of spectra.
Immunostaining of cross-section of aorta: expression of ICAM-1
Rings of the aorta were embedded in the OCT medium (Thermo) and frozen at −80°C using Leica CM1920 automatic cryostat (Leica, Wetzlar, Germany). The 5 μm thick cross-section slides were put on the microscopic glasses coated with poli-L-lysine, and fixed for 10 min in 4% buffered formalin (Merck) and used for immunostaining of ICAM-1. Aortic rings were preincubated with 5% normal goat serum (Jackson Immuno Research) and 2% dry milk in PBS, then immunostained using rat-anti-mouse ICAM-1 (eBioscience; 1:200) primary antibody overnight at 4°C. A secondary antibody Cy3-conjugated goat-anti-rat (Jackson Immuno Research; 1:300) was used, respectively, for 30 min at room temperature. Cell nuclei were visualized by Hoechst 33258 (Sigma; 1:1000) solution and unspecific fluorescence of aortic elastic fibres were used as a background counterstaining channel. Images were acquired using an AxioCam HRm digital monochromatic camera and an AxioObserver.D1 inverted fluorescent microscope (Carl Zeiss). ICAM-1 fluorescence was calculated as the ICAM-1 expression area divided by the tissue area expressed as a percentage (ImageJ program).
Measurements of PGI2 generation in isolated rings of mice aorta
PGI2 generation by aortic rings was quantified on the basis of the formation of 6-keto PGF1α, a stable metabolite of PGI2 in the supernatant of aortic rings, using 6-keto-PGF1α ELISA kit (Enzo Life Sciences).
Statistical analysis
Calculations were performed using Origin2019b software and all data were considered significant if p ≤ 0.05. After testing for normal distribution (Shapiro-Wilk test), the two-sample t-test or one-way ANOVA was performed. All values are given as mean ± SEM.
Results
Fluorescence-based detection of LDs in murine aorta in response to TNF, LPS, AngII, hypoxia or OA; effects of the blockade of adipose triglyceride lipase
An isolated aorta was stimulated by TNF (10 ng/ml, 24h), LPS (10 μg/ml, 24h), AngII (1 μM, 24h), hypoxic conditions (5% O2, 24h) or by OA (1 mM, 24h), and LDs were analyzed in en face aortic preparation. In aorta stimulated by TNF, LPS or AngII, LDs were sparsely observed while in aorta under hypoxic conditions or upon excess of OA, LDs were abundant (Fig. 1A, B). Blockade of the adipose triglyceride lipase (ATGL), a key lipase involved in triacylglycerols hydrolysis by atglistatin (Atgl, 10 μM) resulted in the increased number of LDs in the TNF-, LPS-, or AngII-stimulated aorta in comparison to the aorta treated by TNF, LPS, or AngII in the absence of atglistatin (Fig. 1C, D). Atgl alone in non-stimulated aorta did not result in LDs formation. Similarly, in hypoxic condition or in OA-stimulated aorta, the presence of Atgl significantly increased number of LDs. These results suggest that inhibition by atglistatin delayed LDs degradation in aorta irrespectively to the stimulus used and whether LDs was of low (TNF, LPS, AngII) or high abundance (hypoxia, OA) (Fig. 1D).
Fig. 1Effects of adipose triglyceride lipase (ATGL) inhibition by atglistatin on number of LDs formed in murine aorta in response to TNF, LPS, Ang II, hypoxia or OA and analyzed in en face aortic preparation. (A, B) Fragment of aorta incubated with TNF (10 ng/ml, 24h), LPS (10 μg/ml, 24h), or AngII (1 μM, 24h) showed negligible LDs formation, in contrast to the formation of abundant LDs in isolated blood vessel upon hypoxic condition (5% O2, 24h), or in the presence of OA (1 mM, 24h). (C, D) The ATGL inhibitor (Atgl, 10 μM), abolished LDs degradation in all mentioned-above condition, which results in the observation of LDs in the TNF-, LPS- or AngII-activated aorta, or an increase in the number of LDs in the hypoxic conditions, or in the presence of OA. LDs accumulation was quantified based on BODIPY intensity (green and blue fluorescence originating from LDs and cell nuclei, respectively).
In order to exclude the FBS in the medium as the source of lipids, we have tested the medium without FBS or with 1, 10 and 20% of FBS (Fig. S1). In the result, the high-abundance LDs were observed in conditions without FBS. Such data indicate that endogenous rather than exogenous lipids constitute the main source for LDs formation.
Label-free Raman-based detection of LDs in murine aorta in response to TNF, LPS, AngII, hypoxia or OA
In order to characterize biochemical composition of LDs formed in the isolated aorta stimulated by various stimuli Raman spectroscopy was used. Figure 2 presents two-dimensional Raman images obtained by integrating: the intensity of the C−H stretching vibrations of aliphatic molecules in the 2800-3100 cm−1 spectral range, nucleic acids in the 778-808 cm−1 range, and C−H stretching vibrations of CH2 groups originating mostly from lipids in the 2830-2900 cm−1. Representative images (Fig. 2) clearly showed that LDs were visible in the en face aorta after activation by TNF, LPS, AngII, hypoxic condition or excess of OA (Fig. 2C, white arrows). All experiments were performed in the presence of atglistatin (Atgl) to be able to detect easily LDs formed in response to stimuli resulting in low abundance of LDs (TNF, LPS, AngII). During stimulation of the aorta LDs were formed in both endothelium and smooth muscle cells (Fig. 2D). As the clear-cut border separating the endothelial cells from the smooth muscle cells is the basal membrane made mainly of elastin, we used the Raman signal of elastin and LDs to precisely determine the location of the LDs in endothelium or in smooth muscle cells in the aorta en face preparation. Thus, the Raman depth profiling separated LDs into those localized above or below the basal membrane, belonging to endothelial or smooth muscle cells layer, respectively (Fig. 2D-E).
Fig. 2Representative images of biochemical composition of endothelium and endothelial LDs induced by TNF, LPS, AngII, hypoxia or OA measured by label-free Raman-based confocal imaging in an face murine aorta. Raman images showing distribution of: (A) organic matter (integration in the 2800-3100 cm−1 spectral range), (B) nucleic acids (778-808 cm-1), and (C) lipids (2830-2900 cm-1) in en face control aorta or aorta activated by TNF (10 ng/ml, 24h), LPS (10 μg/ml, 24h), AngII (1 μM, 24h), hypoxia (5% O2, 24h) or OA (1 mM, 24h). Low-abundance LDs formed in response to TNF, LPS, AngII were analysed in the presence of atglistatin (10 μM, 24h), while high-abundance LDs induce by hypoxia or OA were analysed in the absence of atglistatin. (D) Due to the presence of Raman signal of elastin indicating the localization of the basal membrane, the LDs can be separated into those localized above or below the basal membrane, belonging to endothelial or smooth muscle cells layer, respectively. (E) The representative example of depth-profiling Raman measurement for en face aorta after stimulation with LPS was presented, however, analogous measurements were made for other groups of aorta samples activated by various stimuli.
Effects of atglistatin on biochemical composition and size of endothelial LDs in murine aorta in response to hypoxic condition or excess of OA
The effect of atglistatin on the Raman signature of endothelial LDs was compared for hypoxia- and OA-induced LDs, as these two stimuli induced abundant LDs. A direct comparison of the Raman spectra of endothelial LDs formed under the conditions of hypoxia (Fig. 3A) or in response to OA (Fig. 3B) in the absence and the presence of Atgl revealed that atglistatin neither affected the characteristics of the Raman spectrum nor the level of unsaturation of lipids building LDs. The size of LDs was also not affected by atglistatin (Fig 3A, B), suggesting that atglistatin affected only the number of LDs (Fig. 1) but no other characteristics of LDs.
Fig. 3Lack of effect of atglistatin on the biochemical composition of endothelial LDs formed in murine aorta in response to hypoxia or OA. (A) Direct comparison of average Raman spectra of endothelial LDs formed under hypoxic conditions (5% O2, 24h) in the presence or absence of Atgl (10 μM, 24h). (B) Similar comparison of average Raman spectra for endothelial LDs in aorta formed in response to the excess of OA (1 mM, 24h) in the presence or absence of Atgl (10 μM, 24h). The calculated level of lipids unsaturation (no. C=C), and size of LDs remain unchanged under hypoxic or OA-overload conditions irrespective of the use of atglistatin.
Comparison of biochemical composition of LDs formed in endothelium in response to various stimuli by label-free Raman-based imaging in murine aorta
To compare the chemical composition of LDs formed in ECs and analyzed in en face aorta preparation after stimulation with TNF, LPS, AngII, hypoxia or OA, the average spectra of LDs were presented and compared to spectra of arachidonic acid (AA), oleic acid (OA) and cholesterol (Chol; Fig. 4). Bands at 1269, 1662 and 3016 cm−1, assigned to the in-plane =C–H bending, C=C stretching, and =C–H stretching vibrations(
), respectively, indicated the unsaturated character of lipids inside LDs in ECs within en face aorta. LDs formed in response to all used stimuli were analysed in the presence of atglistatin (10 μM, 24h). Features observed at 1300–1310 cm−1 range (attributed to the CH2/CH3 twisting vibrations(
)). The latter assignment was confirmed by the presence of the band at ca. 1745 cm−1, attributed to the carbonyl stretching vibrations of cholesteryl esters, phospholipids, and triglycerides(
). The hallmark of LDs was a considerable increase of the components at ca. 2853 (due to the CH2 stretching vibrations in lipids and fatty acids) and 2889 cm−1 (due to the CH3 stretching vibrations in lipids and fatty acids) relative to band at 2940 cm−1 (resulting in an overall increase of the C-H band integral intensity)(
Fig. 4Comparison of chemical characterisation of LDs formed in endothelial cells in murine aorta in response to TNF, LPS, Ang II, hypoxia and OA. (A) Average Raman spectra of endothelial LDs formed in aorta treated with TNF (10 ng/ml, 24h), LPS (10 μg/ml, 24h), AngII (1 μM, 24h), hypoxia (5% O2, 24h) or OA (1 mM, 24h) with comparison to Raman spectrum of arachidonic acid (AA), oleic acid (OA), and cholesterol (Chol). LDs formed in response to all used stimuli were analysed in the presence of atglistatin (10 μM, 24h). (B) The comparison of the level of unsaturation of lipids in endothelial LDs and (C) size of endothelial LDs in aorta treated with various stimuli and analyzed in en face preparation.
The average spectra of LDs formed in ECs and analyzed in en face aorta following the response to TNF, LPS, AngII or upon hypoxic conditions (Fig. 4A) exhibited a number of similar features. All mentioned spectra did not express distinct band at ca. 703 cm-1 which suggested the negligible content of cholesterols within LDs, and all the spectra contained band at ca. 1745 cm-1, attributed to the carbonyl stretching vibrations indicating that the main component of LDs were triacylglycerols, not free fatty acids. Moreover, the level of unsaturation of lipids in endothelial LDs was determined as the n(C=C)/n(CH2) intensity ratio obtained by integration of the respective Raman bands (1662/1446 cm–1; Fig. 4B). Results showed slightly lower average level of unsaturation of LDs in ECs activated by LPS (1.97±0.05) in comparison to TNF (2.20±0.04) or hypoxia (2.3±0.05). Simultaneously, no statistically significant differences were observed between the level of unsaturation in LDs in AngII- (2.1±0.03), TNF- or hypoxia-treated aorta.
Endothelial LDs formed in the response to OA were definitely more saturated (average number of C=C bonds equaled 0.93±0.02) than LDs formed in the response to TNF, LPS, AngII or hypoxia, and theirs Raman signature resembled mainly glyceryl trioleate (Fig. 4A-B). Except for the statistically significantly larger LDs during incubation with OA (2.1±0.11), the remaining LDs observed in the response to TNF, LPS, AngII or hypoxia were of a similar size: 0.93±0.07, 2.3±0.05, 0.95±0.04, and 0.99±0.05, respectively (Fig. 4C).
Hierarchical cluster analysis (HCA) was applied to obtain classification of single spectra of LDs from endothelium stimulated by TNF, LPS, AngII, hypoxia or OA according to their heterogeneity. The best discrimination, with 100% accuracy, was obtained between Raman spectra of LDs in en face aorta stimulated by OA as compared with other factors (Fig. 5A; red and grey color-coding for LDs formed in the response to OA and other factors, respectively). HCA of Raman spectra did not allow to distinguish biochemical composition of LDs formed in endothelium in response to TNF, LPS, AngII or hypoxia (Fig. 5B). The Raman spectra of LDs from endothelium stimulated by TNF, LPS, AngII or hypoxia were mixed in HCA, which indicated that the existing differences in biochemical composition were too small to be separated by HCA classification.
Fig. 5Classification of single spectra of endothelial LDs formed in response to TNF, LPS, AngII hypoxia or OA assessed by HCA. (A) HCA of Raman spectra separated the spectra of LDs from endothelium stimulated by the excess of OA (red color), but (B) did not allow to distinguish LDs in endothelium stimulated by: TNF, LPS, AngII or hypoxia (blue, pink, green and yellow, respectively). LDs formed in response to all used stimuli were analysed in the presence of atglistatin (10 μM, 24h). Atgl alone did not altered biochemical features of LDs, only their size and number (see text for details).
Effects of TNF, LPS, AngII, hypoxia or OA on expression of ICAM-1 in endothelial cells in isolated murine aorta
To verify whether in our experimental conditions various stimuli used caused the endothelial inflammation, the immunohistochemical staining and fluorescence imaging of the intercellular cell adhesion molecule 1 (ICAM-1) was visualized and quantified (Fig. 6A-H). The images of ICAM-1-stained TNF-, LPS-, AngII- or hypoxia-activated ECs confirmed the overexpression of ICAM-1 in comparison to the control, which confirmed the development of endothelial inflammation by all stimuli used (Fig. 6H). Excess of OA induced increased expression of ICAM-1 that however did not reach the statistical significance. Atgl alone was without an effect on ICAM-1 expression (Fig. 6H).
Fig. 6Effects of TNF, LPS, AngII, hypoxia or OA on expression of ICAM-1 in ECs in isolated murine aorta. Microphotographs of ICAM-1 expression in cross-section of (A) control aorta, aorta treated with (B) TNF (10 ng/ml, 24h), (C) LPS (10 μg/ml, 24h), (D) AngII (1 μM, 24h), (E) hypoxic condition (5% O2, 24h), (F) OA (1 mM, 24h) or (G) Atgl (10 μM, 24h). Red, blue, and green fluorescence originate from ICAM-1, cell nuclei, and elastin fibres, respectively. (H) Quantification of ICAM-1 expression of studied groups of samples.
Effects of TNF, LPS, AngII, hypoxia or OA on prostacyclin generation in murine aorta; effects of the blockade of adipose triglyceride lipase
As, shown in Fig. 7 TNF, LPS, AngII, hypoxia or OA, all these stimuli resulted in an increased generation of PGI2 in aorta samples (measured based on its stable metabolite - 6-keto-PGF1α) after 24 or 48 hours of incubation, thought only for some among the stimuli the difference reached statistical significance. Importantly, in the presence of atglistatin generation of PGI2 induced by LPS, AngII, hypoxia or OA after 24 hours incubation (Fig. 7A) and induced by TNF, LPS, AngII, hypoxia or OA after 48 hours incubation (Fig. 7B) was significantly reduced as compared to aorta without atglistatin suggesting ATGL-dependent PGI2 generation.
Fig. 7Effects of atglistatin on vascular PGI2generation induced by TNF, LPS, AngII, hypoxia or OA. Prostacyclin generation measured based on its stable metabolite: 6-keto-PGF1α after (A) 24 or (B) 48h in unstimulated aorta or aorta activated by TNF (10 ng/ml), LPS (10 μg/ml), AngII (1 μM), hypoxia (5% O2) or OA (1 mM). In the presence of atglistatin 6-keto-PGF1α levels were lowered irrespectively of the stimulus used.
In the present work, we demonstrated that pro-inflammatory activation of endothelial cells in an isolated aorta using variety of stimuli such as TNF, LPS, AngII, hypoxia or OA, led to the formation of lipid droplets (LDs) and concomitant prostacyclin (PGI2) generation. The inhibition of adipose triglyceride lipase (ATGL) increased number of endothelial LDs and simultaneously abrogated PGI2 generation suggesting a key role of ATGL in dynamic formation of LDs in vascular inflammation and in the generation of PGI2. Importantly, biochemical composition of LDs as defined by Raman imaging was not uniform for all pro-inflammatory stimuli, but the formation of LDs was invariably associated with PGI2 generation. Altogether, our results suggest that LDs formation and ATGL-dependent PGI2 generation represent a universal responses to vascular pro-inflammatory insult.
LDs formation in endothelium was previously demonstrated in cultured endothelial cells in response to TNF(16), LPS(15) or OA(
). In contrast to the isolated aorta preparation, LDs formed in cultured cells under the influence of pro-inflammatory factors were stable and did not require the use of atglistatin(
), which indicates a difference in the formation and degradation of LDs in cell line as compared with the isolated blood vessel. These results seem to suggest that dynamics of LDs in vascular wall involve the interactions between endothelial cells and smooth muscle cells, and possibly other cells naturally occurring in the tissue. The endothelium in the isolated blood vessel retains its phenotypic characteristics of in vivo situation better than endothelial cells in culture. Therefore, research conducted on fragments of the vessel wall, may better mimic the in vivo condition, as regards LDs formation in the aorta in vascular inflammation as compared with endothelial cells in culture. However, obviously, isolated vessel is still far from the in vivo system, with blood flow, intercellular communication, and the presence of circulating cells (WBCs, RBCs, etc.) that could all affect composition and formation of LDs.
In the present work we thoroughly characterized LDs in endothelial cells in isolated aorta and extended the knowledge about vascular ATGL-dependent LDs formation and ATGL-dependent PGI2 generation in response to vascular pro-inflammatory insult. The development of endothelial inflammation was confirmed by the overexpression of intercellular adhesion molecule-1 (ICAM-1) in endothelium (Fig. 6). Dynamic formation and decomposition of LDs in isolated aorta incubated with: TNF, LPS or AngII (Fig. 1A-B) was demonstrated by showing that blockade of the adipose triglyceride lipase (ATGL) resulted on elevated number of LDs in stimulated aorta in comparison to non-treated tissues. However, the scale and the significance of the phenomenon were not uniform. After aorta stimulation by TNF, LPS or AngII, blockade of ATGL enables LDs observation, while under hypoxia or excess of OA, the action of atglistatin was not necessary for the observation of LDs, but increases their number (Fig. 1C-D). In the case of hypoxia, it should be considered that the inhibition of ATGL causes the reduction of fatty acids oxidation (FAO), and fatty acids accumulation in the form of LDs, even though, the literature previously showed that fatty acids are not the main substrate of mitochondrial respiration in the endothelium. Also, data from our laboratory showed that there is no accumulation of LDs during the blockage of mitochondrial respiration (antimycin and rotenone, 5%O2, 24h) compared to control (5%O2, 24h, data not shown). Thus, the increased number of LDs in hypoxia resulted from the blockage of their decomposition, and not from reduced oxidation of fatty acids.
These results stay in line with previous studies, where atglistatin enhanced LDs accumulation in OA-treated ECs in culture and in aorta(
) and extend them by showing that in the isolated aorta ATGL –dependent degradation of LDs is shared for all the stimuli inducing LD and used here.
Spectroscopic characterisation of inflammation-induced LDs in ECs within en face aorta releveled the heterogeneity in the biochemical composition of LDs. Raman signature of endothelial LDs under hypoxic or OA conditions in the presence or absence of atglistatin, revealed that atglistatin did not affect the biochemical composition of LDs (Fig. 3), which allows us to believe that atglistatin alone had no effect on the biochemical composition of LDs under stimulation conditions where atglistatin was necessary for the observation of LDs. The analysis of Raman spectra of LDs in the isolated vessels stimulated by TNF, LPS, AngII or hypoxia uncovered that they were all rich in highly unsaturated lipids and had a negligible content of phospholipids and cholesterols (Fig. 4A-B), which is consistent with previously analyzed endothelial LDs in TNF-treated aorta(
Rac1 regulates lipid droplets formation, nanomechanical, and nanostructural changes induced by TNF in vascular endothelium in the isolated murine aorta.
). A similar Raman signature of LDs from endothelium stimulated by LPS, AngII, TNF or hypoxia was confirmed by the lack of significant differences in the classification using hierarchical cluster analysis. The best discrimination, with 100% accuracy, was obtained for Raman spectra of LDs formed in response to OA as compared with other factors (Fig. 5B) suggesting a clearly distinct biochemical composition of LD induced by OA as compared with LDs induced by other factors.
Although the presented results clearly show the potential of Raman spectroscopy for the detection, indication of precise localization (endothelium vs. smooth muscle cells), quantitative assessment and the biochemical composition of LDs in biological samples, this technique is not without limitations. Imaging of LDs was carried out in an isolated blood vessel maintaining the original endothelial environment giving a possibility to distinguish vascular LDs formed in endothelium and LDs formed smooth muscle cells. In our study, the Raman signature of the elastin spectrum (which can be treated as the boundary between the endothelium and the SMC) was used to determine the precise location of the LDs, and on the other hand overlapped with the Raman signal of the LDs located in the SMC layer. As the consequence, the limitation was that collected Raman signal from LDs in SMC did not allow to analyze the degree of unsaturation of lipids building LDs in SMC(
Rac1 regulates lipid droplets formation, nanomechanical, and nanostructural changes induced by TNF in vascular endothelium in the isolated murine aorta.
). Of note, Raman-based analysis was determined on the basis of the intensity ratio of the bands representing an average value of the degree of lipid unsaturation. Thus, without additional mathematical modeling it was not possible to estimate the percentages of individual lipids class found in the LDs(
). Despite the limitations, the RI technique proved to be great for screening the biochemical composition of LDs under various stimulation conditions, and defining differences or similarities between them.
We have noticed the significantly lower degree of unsaturation of lipids building the endothelial LDs formed in aorta in response to OA in comparison to LDs in TNF-, LPS-, AngII-, or hypoxia-stimulated ECs. The lower degree of unsaturation of lipids building the endothelial LDs formed in aorta in response to OA was due to the uptake of fatty acid in the medium, where the endothelium acts as a buffer of excess acids present in the environment and protects the tissues from excess lipids(
). However, despite an excess of OA, endothelial LDs in the OA-stimulated aorta contained a mixture of OA and AA amounted to 91.3% and 8.70%, respectively(
). These findings supported the notion that the formation of LDs even in response to OA is related to the eicosanoids biosynthesis in vascular wall, similar to the well-known role of LDs in the generation of eicosanoids in leukocytes(
). Pro-inflammatory activation of endothelial cells causes the release of AA from the plasma membrane, the main source of eicosanoids generation via cyclooxygenase-2 (COX-2) or COX-1 pathway(
). The presence of AA in LDs could be indeed a reason for a high degree of lipid unsaturation observed in endothelial LDs in the aorta stimulated by TNF, LPS, AngII or hypoxic condition (Fig. 4).
The important finding of this work to show that despite the distinct biochemical composition of LDs induced by various stimuli formation of vascular LDs was invariably linked to the generation of prostacyclin (PGI2). Furthermore, we demonstrated that LDs formation (Fig. 1D) and PGI2 generation (Fig. 7) both responses were ATGL-dependent, which highlighted a key role of ATGL in lipids/eicosanoid-synthesizing machinery in vascular wall during inflammation. The function of LDs in the generation of eicosanoids was shown previously in leukocytes(
). Therefore, endothelial LDs also may represent a reservoir of reserve arachidonic acid, which is released according to the needs of the cell, and may be directed metabolized to prostaglandins or alternatively reincorporated to the phospholipids pool. The key enzyme regulating the decomposition of triacylglycerols from LDs, and thus indirectly releasing arachidonic acid, is ATGL(
). ATGL inhibition blocks the decomposition of lipid droplets, thus depriving the cell of the source of AA from LDs, and in the consequence reducing the availability of substrates for the production of eicosanoids. This explains the dependence of PGI2 generation on LDs formation, and the factor connecting both phenomena was ATGL.
The generation of vascular prostacyclin is tightly regulated to ensure a sufficient PGI2 level for vasoprotection (
). Here we linked the increased generation of PGI2 in vascular wall to LDs formation induced by pro-inflammatory factors, and to ATGL-dependent pathway. These results suggest the importance of ATGL activity not only for lipolysis, but also for vasoprotection. Interestingly, reduced expression of ATGL was shown to enhance TNF-induced ICAM-1 expression in human aortic endothelial cells via protein kinase C-dependent activation of NF-κB(
Reduced expression of adipose triglyceride lipase enhances tumor necrosis factor α-induced intercellular adhesion molecule-1 expression in human aortic endothelial cells via protein kinase C-dependent activation of nuclear factor-κB.
), while global ATGL deficiency leads to pronounced vascular endothelial dysfunction in mice, however in human subjects vascular symptoms were not reported for patients with vascular endothelial dysfunction(
1-Methylnicotinamide protects against liver injury induced by concanavalin A via a prostacyclin-dependent mechanism: A possible involvement of IL-4 and TNF-α.
In conclusion, we characterized the formation of endothelial LDs associated with ATGL-dependent prostacyclin (PGI2) generation in endothelium in isolated murine aorta in response to TNF, LPS, AngII, hypoxia or OA. The inhibition of adipose triglyceride lipase (ATGL) delayed LDs degradation in aorta irrespectively to the stimulus used and whether LDs was of low (TNF, LPS, AngII) or high abundance (hypoxia, OA). Biochemical composition of LDs as defined by Raman imaging was not uniform for all pro-inflammatory stimuli used, but the formation of endothelial LDs was invariably associated with ATGL –dependent PGI2 generation.
Data availability statement
All data are contained within the article. Any or additional data are available from the corresponding authors upon reasonable request.
Supplemental data
This article contains supplemental data.
Conflicts of interest/Competing interests
The authors declare that they have no conflicts of interest with the contents of this article..
Acknowledgements
Funding: This research was funded in whole the National Science Centre, Poland: SONATA-17 No.: DEC- 2021/43/D/ST4/02311. For the purpose of Open Access, the author has applied a CC-BY public copyright licence to any Author Accepted Manuscript (AAM) version arising from this submission.
Authors would like to address special thanks to Ms. Renata Budzynska (JCET UJ) for the technical assistance with ELISA measurements.
Aird, W.C., and M. Laubichler. 2007. Introductory essay: evolution, comparative biology, and development. In Endothelial Biomedicine. Cambridge University Press. 21-28.
Rac1 regulates lipid droplets formation, nanomechanical, and nanostructural changes induced by TNF in vascular endothelium in the isolated murine aorta.
The distinct phenotype of primary adipocytes and adipocytes derived from stem cells of white adipose tissue as assessed by Raman and fluorescence imaging.
Reduced expression of adipose triglyceride lipase enhances tumor necrosis factor α-induced intercellular adhesion molecule-1 expression in human aortic endothelial cells via protein kinase C-dependent activation of nuclear factor-κB.
1-Methylnicotinamide protects against liver injury induced by concanavalin A via a prostacyclin-dependent mechanism: A possible involvement of IL-4 and TNF-α.
All experimental procedures involving animals were conducted according to the Guidelines for Animal Care and Treatment of the European Communities and the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). All procedures were approved by the 2nd Local Ethical Committee on Animal Experiments.
00028-7&id=gr1.jpg){kind=link}
00028-7&id=gr2.jpg){kind=link}
00028-7&id=gr3.jpg){kind=link}
00028-7&id=gr4.jpg){kind=link}
00028-7&id=gr5.jpg){kind=link}
00028-7&id=gr6.jpg){kind=link}
00028-7&id=gr7.jpg){kind=link}