Loss of ABCC6 in Human Mesenchymal Stem Cells Leads to Elevated Reactive Oxygen Species Formation and a Senescence-like Phenotype

TL;DR

This study shows that losing the ABCC6 gene in human stem cells increases harmful oxygen molecules and causes aging-like changes, suggesting antioxidants might help treat the disease.

Contribution

The study reveals a novel link between ABCC6 deficiency, oxidative stress, and a senescence-like phenotype in human mesenchymal stem cells.

Findings

ABCC6 knockdown increases reactive oxygen species and oxidative damage in hMSCs.

ABCC6 deficiency induces a p53-dependent senescence-like phenotype in human stem cells.

Antioxidants like Trolox can reverse oxidative stress and senescence caused by ABCC6 loss.

Abstract

Pseudoxanthoma elasticum (PXE) is an autosomal-recessive disorder caused by mutations in ATP-binding cassette subfamily C member 6 (ABCC6). In addition to the calcification and fragmentation of elastic fibers as the pathomechanistic cause of PXE, systemic and cellular oxidative stress have been reported. Human mesenchymal stem cells (hMSCs) with an ABCC6 knockdown were chosen to further investigate the oxidative stress associated with ABCC6 deficiency. The cells were treated with hydrogen peroxide to mimic external oxidative stress and the antioxidant Trolox to examine the cells’ reaction to decreased oxidative stress. The level of different types of reactive species (RS) like nitric oxide and reactive oxygen species, the senescent phenotype, oxidative damage and mRNA expression of oxidative stress-related genes were evaluated. Knockdown of ABCC6 was shown to increase RS levels in hMSCs, induce a p53-dependent senescence-like phenotype and increase oxidative damage, while the mRNA expression of oxidative defense genes was elevated. The ABCC6-deficient cells exhibited an altered reaction to additional oxidative stress and the incubation with Trolox reversed these changes induced by ABCC6 knockdown. Our findings provide further evidence linking ABCC6-deficiency to oxidative stress and a senescence-like phenotype, while pointing towards antioxidants as part of a potential treatment for PXE.

Figures

Full text

1. Introduction

The fragmentation and calcification of elastic fibers in connective tissues are hallmarks of the autosomal-recessive disorder pseudoxanthoma elasticum (PXE, OMIM 264800). The symptoms of PXE comprise extensive wrinkle formation of the skin, neovascularization and bleeding of the eye’s Bruch’s membrane and calcification of small and medium-sized arteries. The clinical consequences include impairment or loss of central vision, hypertension and intermittent claudication [1].

The genetic cause of PXE is mutations in the gene encoding the transmembrane protein ATP-binding cassette subfamily C member 6 (ABCC6, NG_007558.3) [2]. This ATP-dependent transporter is mainly localized in the liver and the kidneys and only to a small extent in the clinically involved tissues in PXE [3,4]. The substrate of ABCC6 remains unknown, leading to the emergence of two hypotheses on the pathobiology of PXE [5]. The metabolic hypothesis assumes that mutations in ABCC6 result in a lack of transport of the substrate from hepatocytes into the bloodstream leading to ectopic calcification. The cellular hypothesis claims that the loss of functional ABCC6 protein in peripheral cells alters their metabolism and, thus, promotes the calcification of elastic fibers in their proximity [6].

Excessive oxidative stress has been reported in the context of PXE, on both a cellular and systemic level. An increased mitochondrial membrane potential, elevated levels of mitochondrial superoxide and differential mitochondrial protein expression were detected in fibroblasts from PXE patients in comparison to those of healthy donors [7]. Pasquali-Ronchetti et al. were able to detect elevated levels of intracellular superoxide and an increase in hydrogen peroxide production. Furthermore, a decrease in antioxidative capacity and elevated levels of lipid peroxidation, a consequence of oxidative stress, were detected [8]. Garcia-Fernandez et al. detected a lower antioxidative capacity and higher levels of oxidative stress parameters in the circulation of PXE patients [9].

One mechanism of multicellular organisms to prevent the accumulation of DNA damage is called senescence, divided into replicative and acute senescence. Replicative senescence is typically induced by a shortening of the telomers of chromosomes during cell division, activating the DNA damage response and, thus, senescence [10]. Acute senescence is typically induced by stressors such as oxidative stress, oncogenes or ultraviolet radiation, ultimately leading to DNA damage as well [11]. The DNA damage in both types of senescence leads to the stabilization of one or more of the cyclin-dependent kinase inhibitors p53/p21 and p16 [12]. Consequently, cell cycle progression is terminated and further senescence characteristics are acquired. One of the latter is an elevated production of reactive species (RS), such as hydrogen peroxide (H_2_O_2_), superoxide (^·^O_2_^−^) and nitric oxide (NO), that are needed for the maintenance of senescence [13,14]. But RS can also be beneficial for the cell, since they exhibit signaling functions, [15,16,17]. As a result, the balance between the formation and degradation of RS is required for normal cellular function. Endogenous antioxidants, such as superoxide dismutase, catalase (CAT), and glutathione, synthesized by the glutathione synthetase (GSS) and glutathione peroxidases (GPX) or exogenous antioxidants, such as vitamin E and C or carotenoids, can be used to counteract excessive RS formation [18]. Sirtuins are a family of histone deacetylases that consist of seven members (SIRT1–7) with different functions in processes associated with oxidative stress defense [19,20]. The activation of SIRT1, for example, leads to reduced lipid peroxidation, measured by the peroxidation product 4-hydroxynonenal (4-HNE) [19]. Tumor necrosis factor receptor associated protein 1 (TRAP1) is a mitochondria-localized member of the heat shock protein family and known to decrease the production of RS [21].

Another important hallmark of senescence is the senescence-associated secretory phenotype, which is characterized by the increased expression and secretion of cytokines, such as interleukin (IL) 1β, IL-6 and IL-8 [22]. Furthermore, most senescent cells exhibit an increased activity of the senescence-associated β-galactosidase, which can act as a marker of the senescent state of the cells [23].

Recent studies in Abcc6 knockout mice have emphasized a possible role of the bone marrow in PXE pathogenesis [24]. Therefore, bone marrow-derived mesenchymal stem cells were chosen for this investigation. Furthermore, impaired lipid trafficking in adipocytes generated from ABCC6-deficient mesenchymal stem cells was detected [25]. The influence of ABCC6 knockdown on RS levels, senescence and oxidative damage and defense were investigated in this study. Incubation with H_2_O_2_ was chosen to induce acute senescence and mimic systemic oxidative stress, while the antioxidant Trolox was used to diminish RS levels and oxidative stress.

2. Materials and Methods

2.1. Cell Culture and Treatment

Human bone marrow-derived mesenchymal stem cells (hMSCs, see Table 1) were purchased from PromoCell (Heidelberg, Germany) and cultivated in a mesenchymal stem cell growth medium kit (MSCGM, Lonza, Basel, Switzerland) according to the manufacturer’s instructions. The cells were sub-cultured upon reaching approximately 90% confluence. Cells in passage 10 were used for experiments.

The hMSCs were seeded at a density of 4000 cells/cm^2^ in MSCGM for the induction of acute senescence and implementation of additional oxidative stress. The cells were treated 24 h later with 2 mM Trolox (1 mM Na_2_CO_3_, Merck, Darmstadt, Germany) with or without 1 mM H_2_O_2_ (Roth, Karlsruhe, Germany) or with 1 mM Na_2_CO_3_ (vehicle) in MSCGM with delipidated FCS (dFCS). After incubation with H_2_O_2_, cells were washed with 1× DPBS, given fresh dFCS MSCGM with or without Trolox and cultivated for an additional 72 h. Trolox concentration was selected by testing different concentrations using a WST1-cell proliferation assay. Control cells were cultivated in dFCS MSCGM without H_2_O_2_ following the same schedule.

2.2. Delipidation of FCS

The delipidation of FCS was performed by the incubation of 50 mL FCS (MCGS, Lonza, Basel, Switzerland) with 1 g of Cab-o-sil (Sigma-Aldrich, St. Louis, MO, USA) overnight at 4 °C. The mixture was then centrifugated at 4 °C and 10,000× g for 1 h and the delipidated FCS from the supernatant was sterile filtered using a 0.22 µm filter.

2.3. ABCC6 Knockdown via CRISPR/Cas9

The transfection of a ribonucleoprotein complex for CRISPR/Cas9-mediated ABCC6 knockdown was performed in accordance with Plümers et al. and the transfection protocol from IDT (Coralville, IA, USA) [25].

2.4. Nucleic Acid Isolation

The isolation of genomic DNA was performed using the NucleoSpin Blood Kit (Macherey-Nagel, Düren, Germany), following the manufacturer’s instructions. RNA was isolated using the NucleoSpin RNA Kit (Macherey-Nagel, Düren, Germany), according to the manufacturer’s instructions. The concentration of DNA and RNA was measured using the NanoDrop2000 spectrophotometer (Peqlab, Erlangen, Germany).

2.5. Assessment of Reactive Oxygen and Nitrogen Species

The levels of reactive oxygen and nitrogen species were measured using different fluorescent probes. The hMSCs were seeded and treated according to Section 2.1. After 72 h of incubation, the cells were stained with different concentrations of the fluorescent probes for different time periods (see Table 2). Cells were stained using the Viobility™ 640/770 Fixable Dye (Miltenyi Biotec, Bergisch Gladbach, Germany) 15 min before the RS staining was completed. The cells were washed with 1× DPBS and detached using 1× Trypsin (PAN Biotech, Aidenbach, Germany). After centrifugation for 5 min at 310× g, the cells were resuspended in 1× DPBS and the fluorescence was measured using the BD FACSCanto™ II Flow Cytometer (BD Biosciences, Franklin Lakes, NJ, USA). The fluorescence measurement was performed using biological triplicates and background fluorescence was subtracted using unstained hMSCs. Recorded events were gated for intact cells, singlets and viable cells using the Kaluza Software (Kaluza Analysis 2.3., Beckman Coulter, Brea, CA, USA) and the mean intensity of the events recorded was determined.

2.6. Immunofluorescence Staining and Fluorescence Microscopy

After the cell treatment, cells were seeded on 8-well µ-slides (ibidi, Gräfelfing, Germany) at a density of 4000 cells/cm^2^ and cultivated for an additional 24 h. The medium was removed, and the cells were washed with 1× DPBS and fixated using 4% paraformaldehyde (Roth, Karlsruhe, Germany) for 15 min. Permeabilization was performed using 0.1% Triton-X-100 (Roth, Karlsruhe, Germany) in 1× DPBS for 10 min and unspecific binding sites were blocked using 5% normal goat serum (Cell Signaling Technology, Cambridge, United Kingdom) in 1× DPBS for 1 h. Primary antibodies were diluted in 5% normal goat serum according to Table 3 and incubated on the cells for 2 h.

Species-specific fluorescence-conjugated secondary antibodies were diluted in 5% normal goat serum and incubated with the cells for 1 h before staining the nuclei using 5 µg/mL Hoechst (Abcam, Cambridge, MA, USA). Cells were mounted with ROTI^®^Mount FluorCare mounting media (Roth, Karlsruhe, Germany) and images were captured using the BZ-X810 microscope (Keyence, Osaka, Japan). The image processing and determination of fluorescence intensity was performed using ImageJ 1.52a (National Institutes of Health, Bethesda, MD, USA). For quantification of fluorescence intensity, images were converted to greyscale and identical acquisition settings were used for all compared conditions. Regions of interest were defined manually and mean fluorescence intensity was measured.

2.7. β-Galactosidase Assay

The β-galactosidase activity was measured using a fluorescence-based assay [26]. The hMSCs were treated as described in 4.1 and incubated for 72 h. Lysis of the cells was performed using 300 µL lysis buffer (pH 6.0, see Table 4) and the lysate was frozen at −80 °C until the assay was performed. The lysate was centrifuged at 12,000× g and 4 °C for 5 min. An amount of 100 µL of the supernatant was mixed with 100 µL of 2× reaction buffer (see Table 5) and incubated for 1 h at 37 °C.

The reaction was stopped using 600 µL of a 400 mM sodium carbonate solution and the fluorescence was measured on the Tecan Reader Infinite 200 Pro (Tecan, Männedorf, Switzerland). The fluorescence intensity detected was normalized to the protein concentration determined using a bicichonic acid assay (Merck, Darmstadt, Germany).

2.8. Gene Expression Analysis

An amount of 1 µg of isolated RNA was transcribed to complementary DNA using the GoScript™ Reverse Transcriptase Kit (Promega, Madison, WI, USA) following the manufacturer’s instructions. Quantitative real-time polymerase chain reaction (qRT-PCR) was performed using 5 µL LightCycler 480 SYBR Green I Master reaction mixture (Roche, Basel, Switzerland), 0.25 µL each of 25 µM forward and reverse primers (Biomers, Ulm, Germany), 2 µL water and 2.5 µL 1:10 diluted complementary DNA in a 384-well plate (Sarstedt, Nümbrecht, Germany). The PCR reaction consisted of an initial denaturation at 95 °C for 5 min, followed by 45 cycles of amplification including denaturation at 95 °C for 10 s, annealing at a primer-specific temperature (Table 6) for 15 s and elongation at 72 °C for 20 s.

The Ct values detected of genes of interest were normalized to the mean of expression of the three housekeeping genes ribosomal protein L13a (RPL13A), succinate dehydrogenase complex flavoprotein subunit A (SDHA) and beta-2-microglobulin (β2M) using the efficiency-corrected delta-delta Ct method.

2.9. Statistical Analysis

Statistical analysis was performed using the GraphPad Prism 10 software (GraphPad, San Diego, CA, USA). All experiments were performed in biological and technical triplicates. The data shown are presented as mean values ± standard error of mean and statistical significance was calculated using three-way ANOVA and Tukey’s multiple comparisons test with probability (p) values equal or below 0.05 being considered statistically significant.

3. Results

3.1. Elevated Levels of Reactive Species May Be the Cause of the Senescence-like Phenotype

Based on our previous work, a CRISPR/Cas9-based approach for ABCC6 knockdown was chosen [26]. This gene editing model was chosen to achieve a stable reduction in ABCC6 mRNA expression over several passages. An efficiency of the knockdown of above 70% was determined using the T7 endonuclease I assay, TA-cloning and qRT-PCR (Figures S1–S3).

Evaluation of the effect of ABCC6-knockdown on reactive species levels was performed by detecting different fluorescent probes using flow cytometry (Figure S4).

Total reactive oxygen species (ROS) was detected as 4.3-fold (±0.5) higher in H_2_O_2_-treated control cells compared to the untreated cells. The additional incubation with Trolox resulted in a 63% (±10%) decreased ROS level compared to H_2_O_2_-treated control cells. The knockdown of ABCC6 led to a 2.9-fold (±0.3) increase in the level of total ROS compared to untreated control cells and the incubation of knockdown cells with Trolox led to an 80% (±4%) reduction in the total ROS level compared to untreated knockdown cells. Hydrogen peroxide elevated the total ROS level 1.8-fold (±0.2) in relation to untreated knockdown cells, while the additional treatment with Trolox led to a 65% (±10%) decrease in the ROS level (Figure 1A).

The incubation of control cells with H_2_O_2_ elevated the nitric oxide (NO) level 5.7-fold (±0.6) compared to untreated cells, while the combined treatment of H_2_O_2_ and Trolox decreased the NO level to 54% (±2%) in relation to H_2_O_2_-treated cells. Due to the knockdown of ABCC6, a 2.8-fold (±0.4) higher DAF-FM fluorescence compared to control cells was detected. The incubation with H_2_O_2_ led to a 2.4-fold (±0.2) increase in the fluorescence intensity in knockdown cells compared to untreated knockdown cells and following treatment with H_2_O_2_ and Trolox, a 57% (±3%) decrease in the relative fluorescence intensity was detected compared to H_2_O_2_-treated knockdown cells. The incubation of control and knockdown cells with Trolox reduced the fluorescence intensity to the level of the internal unstained control, consequently, no measurable results were obtained (Figure 1B).

3.2. ABCC6 Knockdown hMSCs Reveal a Senescence-like Phenotype

A potential senescence-like phenotype in knockdown hMSCs was evaluated by measuring the senescence-associated β-galactosidase activity and the mRNA expression of the cell cycle inhibitors p21 and p53.

The incubation of control cells with H_2_O_2_ resulted in a 1.3-fold (±0.1) increase in the relative β-galactosidase activity compared to untreated control cells. The treatment of control cells with Trolox decreased the relative β-galactosidase activity by 30% (±2%) in relation to untreated control cells and the additional treatment of H_2_O_2_ incubated cells with Trolox reduced the β-galactosidase activity by 37% (±3%) compared to H_2_O_2_-treated cells. Due to the knockdown of ABCC6, a 1.3-fold (±0.1) increase in the β-galactosidase activity compared to control cells was detectable. In contrast to control cells, a treatment with H_2_O_2_ resulted in no significantly altered β-galactosidase activity in knockdown cells in comparison to untreated knockdown cells. Treating the knockdown cells with Trolox caused a 40% (±4%) reduction in the relative β-galactosidase activity compared to untreated cells. The combined treatment with H_2_O_2_ and Trolox resulted in a 26% (±6%) decreased activity relative to H_2_O_2_-treated knockdown cells (Figure 2A).

Due to the incubation of control cells with H_2_O_2_, a 104-fold (±25.2) increase in p21 mRNA expression was detected compared to untreated control cells, while the additional treatment with Trolox led to a 95% (±1%) decrease in mRNA expression in relation to H_2_O_2_-treated cells. The cultivation of control cells with Trolox alone did not alter the p21 mRNA expression compared to untreated cells. The knockdown of ABCC6 resulted in a 67.2-fold (±21.6) increase in p21 mRNA expression relative to untreated control cells, while none of the treatments tested had a significant effect on the p21 mRNA expression of the knockdown cells (Figure 2B).

After the incubation of control cells with H_2_O_2_, the p53 mRNA expression was detected as 30.4-fold (±6.1) increased compared to untreated control cells, whereas the additional treatment with Trolox resulted a 93% (±1%) decrease in the p53 mRNA expression in relation to H_2_O_2_-treated cells. The treatment of control cells with Trolox did not lead to any significant changes in the mRNA expression due to the knockdown of ABCC6; a 22.6-fold (±6.2) elevated p53 mRNA expression was measured compared to control cells. None of the treatments performed resulted in an altered p53 mRNA expression of the knockdown cells compared to the untreated knockdown cells (Figure 2C).

The senescent phenotype was further investigated by immunofluorescent staining of the cell cycle inhibitors p21 and p53, to validate the transcriptional data on the protein level. Using the nucleic counterstain, the area of the nucleus was marked and the fluorescence intensity of the stained cell cycle inhibitors was quantified exclusively in that area.

Following the treatment of control cells with H_2_O_2_, a 1.1-fold (±0.01) increase in relative fluorescence intensity compared to untreated cells was detected. Additional treatment with Trolox resulted in a 1.2-fold (±0.02) increase compared to the H_2_O_2_-treated control cells. Incubation with Trolox resulted in a 1.1-fold (±0.03) elevated fluorescence intensity compared to untreated cells. The knockdown of ABCC6 resulted in a 20% (±2%) reduction in the fluorescence intensity compared to control cells. The H_2_O_2_-treated cells exhibited a 1.5-fold (±0.03) increase in fluorescence intensity relative to untreated knockdown cells, while the additional treatment with Trolox resulted in a 1.1-fold (±0.02) elevated fluorescence intensity compared to H_2_O_2_-treated cells. Incubation of the knockdown cells with Trolox showed a 1.2-fold (±0.02) elevated relative fluorescence intensity (Figure 3A).

The cultivation of control cells with H_2_O_2_ increased the fluorescence intensity of p53 1.5-fold (±0.05) compared to untreated cells, while the additional treatment with Trolox resulted in a 34% (±2%) reduction in the fluorescence intensity compared to H_2_O_2_-treated control cells. The incubation of control cells with Trolox did not result in any significant changes in the fluorescence intensity in relation to untreated cells. Following the ABCC6 knockdown, an increase in the fluorescence intensity of 1.7-fold (±0.05) relative to untreated control cells was detected. The treatment of knockdown cells with H_2_O_2_ resulted in a 1.3-fold (±0.04) elevation of the p53 fluorescence intensity compared to untreated knockdown cells. The additional incubation with Trolox led to a 22% (±2%) decrease in the fluorescence intensity compared to H_2_O_2_-treated knockdown cells. The treatment of knockdown cells with Trolox did not induce any changes in the p53 fluorescence intensity in knockdown cells (Figure 3B).

3.3. Knockdown hMSCs Exhibit Differential Patterns of Oxidative Stress

An assessment of the oxidative damage was performed using immunofluorescence staining of the lipid peroxidation product 4-hydroxynonenal (4-HNE) (Figure 4A).

The incubation of control cells with H_2_O_2_ did not reveal any significant changes in the 4-HNE fluorescence intensity compared to the untreated control. The treatment of control cells with Trolox led to a 69% (±8%) decrease in the fluorescence intensity relative to untreated cells, while the incubation of H_2_O_2_-treated cells with Trolox resulted in a 66% (±8%) decrease in the fluorescence intensity compared to control cells. Following the knockdown of ABCC6, a 1.7-fold (±0.1) increase in the 4-HNE fluorescence intensity was detected. The treatment of knockdown cells with H_2_O_2_ did not alter the fluorescence intensity measured compared to untreated knockdown cells, while the additional incubation with Trolox did not result in any changes either compared to H_2_O_2_-treated cells. The treatment of knockdown cells with Trolox led to a 50% (±7%) decrease in the fluorescence intensity detected relative to untreated knockdown cells (Figure 4B).

The influence of the ABCC6 knockdown on the mRNA expression of different genes associated with oxidative stress defense, reactive species (RS) production and RS elimination was investigated using qRT-PCR.

Firstly, the mRNA expression of genes of the sirtuin (SIRT) family was evaluated. Following the H_2_O_2_ treatment, the SIRT2 mRNA expression of control cells was 17.9-fold (±2.9) increased compared to untreated control cells, while the additional treatment with Trolox resulted in a 93% (±1%) reduction in mRNA expression compared to H_2_O_2_-treated cells. The incubation of control cells with Trolox did not alter the mRNA expression in relation to untreated cells. The ABCC6 deficiency resulted in a 13.4-fold (±2.0) elevated SIRT2 mRNA expression relative to untreated control cells, while none of the conditions tested altered the mRNA expression in knockdown cells (Figure 5A).

Regarding the SIRT3 mRNA expression, the incubation of control cells with H_2_O_2_ resulted in a 3.4-fold (±0.6) increased expression. Additional treatment with Trolox resulted in a 36% (±14%) reduction in mRNA expression compared to H_2_O_2_-treated cells, while the knockdown of ABCC6 led to a 2.4-fold (±0.5) increase compared to untreated control cells. The other conditions tested did not show any changes in mRNA expression (Figure 5B).

The SIRT6 mRNA expression was not significantly altered in control cells after treatment with H_2_O_2_, Trolox or their combination. Due to the knockdown of ABCC6, a 22.3-fold (±7.8) elevation in SIRT6 mRNA expression compared to untreated control cells, was observed. The treatment of knockdown cells with H_2_O_2_ led to a 95% (±1%) reduction in SIRT6 mRNA expression compared to untreated knockdown cells, while the additional incubation with Trolox led to an 8.9-fold (±3.5) elevated SIRT6 mRNA expression compared to H_2_O_2_-treated knockdown cells (Figure 5C).

Following the H_2_O_2_ treatment, the mRNA expression of SIRT7 was detected as 11.5-fold (±2.5) increased relative to the untreated control. The additional incubation with Trolox resulted in an 81% (±5%) reduction compared to H_2_O_2_-treated control cells, while Trolox alone did not alter the mRNA expression in relation to untreated control cells. The knockdown of ABCC6 led to a 6.1-fold (±2.1) elevated SIRT7 mRNA expression compared to untreated control cells. None of the conditions tested resulted in changes in the SIRT7 mRNA expression in knockdown cells compared to untreated knockdown cells (Figure 5D).

The mRNA expression of glutathione synthetase (GSS) was detected as 6.9-fold (±2.3) increased due to the incubation of control cells with H_2_O_2_ compared to the untreated cells, while the additional incubation with Trolox led to a 71% (±4%) reduction in relation to H_2_O_2_-treated cells. The treatment of control cells with Trolox did not result in any changes in mRNA expression relative to the untreated control. Following the knockdown of ABCC6, the mRNA expression of GSS was detected as 4.3-fold (±1.4) elevated compared to control cells. None of the treatments performed led to an altered GSS mRNA expression in knockdown cells relative to untreated knockdown cells (Figure 6A).

After the incubation of control cells with H_2_O_2_, the mRNA expression of glutathione peroxidase 1 (GPX1) was detected as 6.4-fold (±1.4) increased compared to the untreated control. The additional treatment with Trolox led to an 84% (±4%) reduction in relation to H_2_O_2_-treated cells. The incubation of control cells with Trolox did not result in any changes in mRNA expression compared to control cells. The mRNA expression of GPX1 was measured as 6.0-fold (±1.3) elevated in knockdown cells compared to control cells, while none of the treatments performed was shown to have an effect on GPX1 mRNA expression in knockdown cells (Figure 6B).

A 2.6-fold (±0.3) increase in the mRNA expression of GPX4 was detected in control cells following the H_2_O_2_-treatment relative to the untreated control, while the additional treatment with Trolox reduced the mRNA expression to 60% (±6%) relative to H_2_O_2_-treated cells. The knockdown of ABCC6 led to a 2.0-fold (±0.2) elevated mRNA expression of GPX4 compared to control cells. No alterations in GPX4 mRNA expression were present compared to the respective control for the other conditions tested (Figure 6C).

The incubation of control cells with H_2_O_2_ induced a 2.1-fold (±0.3) increase in the mRNA expression of catalase (CAT) compared to untreated control cells, while the additional Trolox treatment reduced the expression to 73% (±4%) relative to H_2_O_2_-treated cells. The incubation of control cells with Trolox alone led to a 60% (±9%) reduction in mRNA expression. Neither the knockdown of ABCC6, nor the additional treatments resulted in any alterations in mRNA expression (Figure 7A).

The treatment with H_2_O_2_ led to a 34.1-fold (±6.5) increase in TNF receptor associated protein 1 (TRAP1) mRNA expression in control cells relative to untreated cells. Trolox alone did not induce any changes in TRAP1 mRNA expression, while the treatment with H_2_O_2_ and Trolox resulted in a 91% (±2%) reduction in mRNA expression compared to H_2_O_2_-treated cells. Due to the knockdown of ABCC6, a 29.4-fold (±5.9) elevation in mRNA expression compared to control cells was detected. The treatment of knockdown cells with H_2_O_2_ or Trolox resulted in 63% (±11%) and 62% (±10%) decreased TRAP1 mRNA expression, relatively, compared to untreated knockdown cells, while the combination of H_2_O_2_ and Trolox left the mRNA expression in knockdown cells unchanged in relation to H_2_O_2_-treated cells (Figure 7B).

4. Discussion

Pseudoxanthoma elasticum (PXE) is a complex disorder with an impact on both the systemic and cellular levels. Oxidative stress seems to play a vital role in the pathomechanism of PXE but still needs to be further investigated. There has been evidence of systemic oxidative stress, supporting the metabolic hypothesis of PXE pathogenesis. Furthermore, investigations on patient-derived fibroblasts indicated oxidative stress on the cellular level as well, which provides evidence for the cellular hypothesis. Analyses on ABCC6-deficient human mesenchymal stem cells (hMSCs) regarding their oxidative status and senescence were performed to gain further insight into the cellular component of PXE pathogenesis.

A recent publication from Brampton et al. points to the bone marrow as a potential contributing factor in the PXE pathogenesis [10]. They were able to show that restoring the bone marrow of Abbc6-deficient mice with that of wildtype mice led to a significant reduction in the calcification of the whiskers and kidney. Interestingly, the transplantation of bone marrow from Abcc6-deficient mice into wildtype mice did not increase calcification [10]. Furthermore, Plümers et al. detected impairments in lipid trafficking, intra- and extracellular lipolysis, release of lipids and fatty acid neogenesis in adipocytes generated from ABCC6-deficient human mesenchymal stem cells (hMSCs) [25]. Due to these findings linking MSCs to the pathogenesis of PXE, hMSCs were chosen for this investigation. In accordance with Plümers et al., a CRISPR/Cas9-based approach for an ABCC6 knockdown was chosen [25]. Due to the inability to generate single cell cultures, a cell pool containing homozygous, heterozygous and wildtype cells was used here and must be taken into account when interpreting the results of this study. Nevertheless, based on the results of this study regarding the CRISPR efficiency, the fraction of wildtype cells in the pool can be considered small. A knockdown efficiency of approximately 70% was confirmed using a T7 endonuclease I assay, TA-cloning and qRT-PCR. No mutations were detected in the three most common off-target gene regions, so off-target effects can be ruled out (Figure S2). However, it was not possible to confirm the knockdown at the protein level due to the lack of appropriate antibodies against human ABCC6 or ABCC6-specific activity assays. Given the sample size of n = 2, the results generated here were used solely to gain initial insights into the relationship between ABCC6 and oxidative stress in hMSCs. The consistency of the observed trends confirms the biological relevance of our results. Nevertheless, a larger sample size will be required for future studies in order to further confirm the observations made here.

The reactive oxygen species (ROS) and nitric oxide (NO) were investigated using fluorescent dyes to evaluate the oxidative status of knockdown cells in comparison to control cells. The incubation with H_2_O_2_ was chosen to mimic the systemic oxidative stress reported in PXE patients and the vitamin E derivate Trolox was used to counteract a potential increase in reactive species (RS) levels following ABCC6 knockdown or H_2_O_2_ induction. The incubation with H_2_O_2_ increased the levels of total ROS and nitric oxide in both control and knockdown cells. This can be explained by the RS-inducing properties of H_2_O_2_ reported in the literature [26,27,28]. The additional treatment with Trolox reduced the levels of ROS and NO, emphasizing its antioxidative properties. After the knockdown of ABCC6, total ROS and NO were determined to be increased, possibly indicating oxidative stress in knockdown cells by induced production or reduced degradation of nitric oxide. Nitric oxide in mesenchymal stem cells is reported to have opposite effects depending on the concentration. While small concentrations were shown to promote cell survival and proliferation, higher concentrations can induce cell cycle arrest and senescence [29,30,31]. This suggests the hypothesis that the increased NO levels following the knockdown of ABCC6 may induce a senescence-like phenotype. It should be noted that the fluorescent dyes used in this study may be influenced by experimental artefacts. The fluorescence of DCFDA may be influenced by pH or metal ions, while it has been shown that DAF-FM may form fluorescent products through auto-oxidation or superoxide [32,33]. Therefore, the results shown here must be interpreted with caution. A higher β-galactosidase activity was detected for knockdown cells, indicating the presence of a senescence-like phenotype, therefore, the expression of NO synthetase could be investigated in future studies. Furthermore, the specific inhibition of NO formation by asymmetric dimethylarginine could give further insights into the possible association of NO levels and the senescence-like phenotype following ABCC6 knockdown [34,35]. An indication of this connection was made by the incubation of knockdown cells with the antioxidant Trolox, resulting in decreased β-galactosidase activity and an attenuated senescence-like phenotype. This suggests that a decrease in oxidative stress in ABCC6 knockdown cells may result in the reversal of the senescence-like phenotype detected. In addition, the β-galactosidase assay did not reveal any elevated activity for knockdown cells incubated with H_2_O_2_, indicating a possible resistance to additional oxidative stress. The senescence-like phenotype associated with PXE was also detected in patient-derived dermal fibroblasts, with elevated β-galactosidase activity in addition to increased secretion of interleukin 6 (IL-6) and a higher mRNA expression of p21 but not p53 [36]. In this study, an elevated mRNA expression of p21 and p53 was detected. The differences between this study and that of Tiemann et al. could be explained by the use of different cell types, since the senescence-associated secretory phenotype is highly cell-type specific [37]. The decrease in β-galactosidase activity by incubation with Trolox was not accompanied by a decreased mRNA expression of p21 and p53 in knockdown cells, which could be explained by post-transcriptional regulations influencing protein synthesis and activity. Immunofluorescence staining of p21 and p53 was performed to further investigate the influence of the cell cycle inhibitors on the senescence-like phenotype, revealing a strong increase in protein abundance for p53. In contrast to the β-galactosidase assay, the treatment of knockdown cells with H_2_O_2_ further increased the fluorescence signal of p53, while the additional treatment with Trolox counteracted this increase. It might be useful here to investigate the protein abundance and localization of phosphorylated p53 using immunofluorescence or a Western blot, since phosphorylated p53 is the transcriptionally active form [38,39]. The discrepancy between mRNA and protein expression observed for p21 indicates pronounced post-transcriptional regulation, which has already been demonstrated for p21 [40].

The oxidative damage was evaluated using immunofluorescence staining with antibodies specific to 4-hydroxynonenal (4-HNE), a product of lipid peroxidation by oxidative stress [41]. It was shown that the knockdown of ABCC6 resulted in increased lipid peroxidation, possibly due to the elevated ROS levels. Surprisingly, the incubation of cells with H_2_O_2_ did not result in any increased lipid peroxidation. This could be explained by the degradation of 4-HNE with time, which can be mediated by proteins such as aldose reductase, glutathione-S-transferases and aldehyde dehydrogenases. The expression or activity of these proteins could be investigated in future studies to better understand the absence of lipid peroxidation products in H_2_O_2_-treated control cells. The same procedure could also be applicable in knockdown cells incubated with H_2_O_2_, since they show no increase in lipid peroxidation either. Regarding the knockdown cells, the unaltered level of lipid peroxidation after H_2_O_2_ treatment could also be in accordance with the results of the β-galactosidase assay, showing no increased activity after treatment. Trolox was shown to decrease the level of 4-HNE fluorescence in knockdown cells, proving to be a suitable inhibitor of oxidative stress damage induced by ABCC6 deficiency. In contrast to this, the additional treatment with Trolox reduced the 4-HNE fluorescence signal only in control cells.

Sirtuins (SIRT) are a family of histone deacetylases with different functions in processes associated with oxidative stress defense [19,20]. The mRNA expression of SIRT2, SIRT3, SIRT6 and SIRT7 was shown to be upregulated by incubation with H_2_O_2_ in control cells and following ABCC6 knockdown, while no change in mRNA expression was detectable in H_2_O_2_-treated knockdown cells. SIRT2 was shown to increase the expression of superoxide dismutase 1/2 and decrease oxidative stress and senescence through the inhibition of the p21/p53 pathway in nucleus pulposus cells [42]. In this study, mRNA expression of SIRT2 could be a compensatory mechanism for the senescence-like phenotype following the H_2_O_2_ treatment of control cells and ABCC6 knockdown. The sirtuins 3, 4 and 5 are localized in the mitochondria. Only the mRNA expression of SIRT3 is altered by incubation with H_2_O_2_ or ABCC6 knockdown; therefore, only a mildly, if any, altered mitochondrial oxidative stress can be assumed following H_2_O_2_-treatment and knockdown. Furthermore, SIRT3 was shown to deacetylate specific components of the electron transport chain to increase the oxidative phosphorylation and decrease its sensitivity to RS [43,44,45]. In addition, SIRT3 is an activator of fatty-acid oxidation in the mitochondria [46,47]. In accordance with the literature, these results suggest alterations in the mitochondrial energy metabolism and fatty-acid oxidation that require further investigation [7]. Knockout of SIRT6 promotes premature aging in mice [48]. It was shown that SIRT6 protects cells against oxidative stress-induced DNA damage and exerts highly cell type specific effects on proliferation [49,50,51]. Furthermore, a SIRT6 deficiency promotes the switch to anaerobic glycolysis, resulting in lactate generation and a decrease in mitochondrial oxidative phosphorylation [52]. In light of these reported functions of SIRT6, a compensatory upregulation due to the ABCC6 knockdown and the increase in RS levels would be possible. In addition, the increased mRNA expression of SIRT6 could be a sign of elevated aerobic glycolysis and ATP generation in mitochondria. This possibility has to be investigated in future studies. The function of SIRT7 is largely unknown, though it seems to regulate cell growth and metabolism, may be required for cell viability and modulates the stress response of cells to difficult circumstances [53,54,55,56]. A possible explanation for the upregulated mRNA expression of SIRT7 would be the adaptation to the increased oxidative stress following ABCC6 knockdown.

Regarding the mRNA expression of GSS, GPX1 and GPX4, the same pattern of induced expression following H_2_O_2_ treatment and ABCC6 knockdown with a lack of induction after the incubation of knockdown cells with H_2_O_2_ was detected. Glutathione is a very important cellular antioxidant protein that is synthesized by glutathione synthetase (GSS), whereby its antioxidant property is mostly mediated by glutathione peroxidases (GPX). Glutathione can be used to degrade H_2_O_2_ and lipid peroxides, generating glutathione disulfide that can be restored to maintain the cells’ antioxidant defense [57]. The upregulation of mRNA expression detected could be a compensatory mechanism due to the induction of oxidative stress by incubation with H_2_O_2_ or ABCC6 knockdown. Since no further elevation in mRNA expression was detectable after the H_2_O_2_ treatment of knockdown cells, the maximum induction may be reached due to the knockdown with no further elevation possible. The protective effects of increased GSS or GPX1 mRNA expression against oxidative stress were shown in genetically engineered mice and chicken embryo fibroblasts [58,59]. The redox status of the knockdown hMSCs could be evaluated in further studies to investigate changes by the application of additional oxidative stress.

Regarding the mRNA expression of catalase (CAT), no changes were detected following ABCC6 knockdown, indicating that the advanced oxidative defense of knockdown cells is based mainly on the glutathione system. The mRNA expression of TNF receptor associated protein 1 (TRAP1) was induced after the incubation of control cells with H_2_O_2_ and ABCC6 knockdown. Since TRAP1 is a mitochondria-localized protein that protects against oxidative stress, these results seem to contradict the lack of elevation of mitochondrial superoxide. However, there are different postulated functions of TRAP1 in the endoplasmic reticulum and the cytosol. TRAP1 seems to be a part of the tumor necrosis factor-α pathway and activates the transcription factor signal transducer and activator of transcription 3 (STAT3) [60]. In previous work, the JAK/STAT3 signaling pathway was shown to be activated in fibroblasts derived from PXE patients, resulting in increased inflammation [61]. The elevated expression of TRAP1 may be one contributing factor of STAT3 activation.

5. Conclusions

The data obtained in this study suggest that knockdown of ABCC6 leads to a p53-mediated senescence-like phenotype due to increased levels of reactive species. This is accompanied by increased lipid peroxidation and enhanced oxidative stress defenses, making knockdown cells insensitive to further oxidative stress. Some results point to defects in mitochondrial energy generation and fatty-acid oxidation that require further investigation. The antioxidant Trolox was shown to reduce reactive species levels in knockdown cells, leading to an attenuated senescence-like phenotype and lipid peroxidation but leaving the elevated oxidative stress defense unaltered. This study further emphasizes the impact of oxidative stress on the PXE phenotype and provides new indications for the use of antioxidants as a potential treatment for PXE.