Main

Progressive calcification and fragmentation of elastic fibers are characteristic hallmarks of pseudoxanthoma elasticum (PXE), which is caused by mutations in the ABCC6 gene encoding the multidrug resistance-associated protein 6 (MRP6,1, 2, 3, 4, 5, 6). Recently, two different abcc6 knockout mouse models were generated, that exhibited a PXE phenotype, strengthening ABCC6/MRP6 as the candidate gene for PXE.7, 8 MRP6 is a member of the large ATP-binding cassette (ABC) transporter superfamily. To date, 49 ABC transporter genes have been identified as belonging to this family of membrane transporter proteins performing various functions in human cells. The most prominent and well-characterized ABC transporter proteins are P-gp (ABCB1/MDR1) and ABCC1/MRP1, mediating multidrug resistance in cancer cells (for review see Haimeur et al9 and Deeley et al10 and http://nutrigene.4t.com/humanabc.htm). MRP6 belongs to subfamily C currently consisting of 13 members, of which 9 belong to the MRP family known to be organic anion transporters. Until now, nothing is known about the physiological function of MRP6 and its role in PXE manifestation. The protein is predominantly expressed in the liver and kidney, whereas very low expression has been observed in tissues primarily affected by PXE, for instance the skin.2, 11 It was assumed that MRP6 serves as an efflux pump on the basolateral surface of hepatocytes transporting as yet unidentified substrates from the intracellular milieu to the blood.12, 13 Indeed, several studies reported alterations in systemic blood components in PXE patients and mouse models due to MRP6 deficiency.14, 15, 16, 17, 18, 19, 20 Accordingly, PXE was assumed to be primarily a metabolic disorder.12, 21, 22, 23 In contrast, results of several studies also point to a local role for ABCC6/MRP6 as expression of the protein was detected in dermal fibroblasts from healthy controls whereas it was found to be absent in fibroblasts from PXE patients.24 Moreover, dermal fibroblasts expanded from PXE patients' biopsies exhibited pathological alterations compared to dermal fibroblasts from healthy controls.25 These cells have been shown to synthesize abnormal proteoglycans and to exhibit a higher expression of elastin (ELN), which was further associated with pathological assembly of elastic fibers when cultured in the presence of PXE sera.16, 26 Quaglino et al27 reported that PXE fibroblasts have a raised matrix metallopeptidase 2 mRNA and protein expression. In addition, a mild oxidative stress was observed in PXE fibroblasts with higher malondialdehyde levels, increased superoxide dismutase 2 and reduced catalase and glutathione peroxidase activities.28 Therefore, it cannot be excluded that also low-level expression of ABCC6/MRP6 plays an important physiological role. MRPs are active in dermal fibroblasts as these cells exhibit a MRP efflux activity that can be blocked by specific inhibitors known to interfere with MRP function.29, 30, 31 Furthermore, it was reported that PXE fibroblasts have a reduced MRP transporter activity compared to normal dermal fibroblasts.31 Beside MRP6, other MRPs have also been reported to be associated with human diseases. Mutations in ABCC2/MRP2 lead to Dubin--Johnson syndrome, whereas mutations in ABCC7/CFTR are the cause for cystic fibrosis. A compensatory expression of ABCC3/MRP3 was previously observed in rats suffering from cholestasis due to MRP2 deficiency, as seen in patients with Dubin--Johnson syndrome and Eisai hyperbilirubinemic rats.32, 33 The question is whether a MRP6 deficiency has an effect on the expression of other ABC transporter proteins. A recent study by Li et al34 reported no alterations in the gene expression of other ABC transporter subfamily C members in abcc6 knockout mice. The aim of our present study was to investigate the expression profile of 47 ABC transporter genes, as well as the MRP transporter activity of dermal fibroblasts derived from skin biopsies of PXE patients and of healthy controls. Our results will provide new data implying a compensatory role for other ABC transporter proteins in PXE manifestation. Moreover, we show that the observed alterations in gene expression of other ABC transporters are associated with MRP6 deficiency due to inhibition of ABCC6/MRP6 synthesis. Finally, we will discuss a possible role for ABCC6/MRP6 in lipid metabolism.

MATERIALS AND METHODS

Skin Biopsies and Cell Culture Conditions

Dermal fibroblasts from seven PXE patients were expanded from skin biopsies. The diagnosis of PXE in all patients was consistent with the reported consensus criteria.35, 36 The status of the PXE patients was determined by the presence of dermal lesions and ocular findings. The dermal lesions were histologically confirmed by the observation of mineralized elastic fibers in biopsy samples following von Kossa staining. The study was approved by the institutional review board and all patients gave their informed consent. In detail: biopsy samples were digested with 0.5% pronase E (Sigma, Steinheim, Germany) for 30 min, and 1% collagenase/dispase (Roche, Penzberg, Germany) in Dulbecco's modified essential medium (DMEM) containing 10% fetal calf serum, 1% L-glutamine (200 mM) and 1% antibiotic/antimycotic solution (100 × ) at 37°C overnight. Afterward, the samples were washed twice with Dulbecco's phosphate-buffered saline (DPBS) and seeded in DMEM medium including all supplements. Media and supplements were obtained from PAA (Pasching, Austria) and Biowest (Nuaillé, France). Fibroblast cultures were grown out until they reached confluence and were then subcultured once a week with a ratio of 1:3. Dermal fibroblasts from six healthy controls (NHDF) were purchased from Promocell (Heidelberg, Germany), Genlantis (San Diego, USA) and Cambrex (Walkersville, USA). All cultures were checked for fibroblast-specific marker expression (Thy-1; CD90) by antibody staining following immunofluorescence, according to the manufacturer's protocol (Dianova, Hamburg, Germany). Main characteristics of PXE patients and healthy controls are summarized in Table 1. For all experiments subcultures of the 4th to 8th passage were used. Experiments were done on four different subcultures of each PXE patient and control subject to exclude variability from culture conditions. Measurements were done in duplicate or triplicate for each biological sample. HepG2 cells were maintained in DMEM supplemented with 10% fetal calf serum, 1% L-glutamine (200 mM) and 1% antibiotic/antimycotic solution (100 × ), subcultured with a ratio of 1:3 once a week.

Table 1 Main characteristics of dermal fibroblasts derived from PXE patients and healthy controls used in the present study
Full size table

ABCC6 Genotyping

Genotyping of the c.3421C>T. mutation and mutational analysis of ABCC6 was performed as previously described.6

Analysis of ABC Transporter Transcript Levels by TaqMan Low-Density Array

Transcript levels for 47 human ABC transporters and the reference gene 18S rRNA were analyzed by a well-established real-time PCR-based TaqMan low-density array (TLDA), as described.37 Total RNA was extracted from 1 × 106 cells using the RNeasy Midi Kit (Qiagen, Hilden, Germany). Genes that were regulated more than twofold (≥2.0 and ≤0.5) were considered significantly regulated.37 NHDF served as calibrators to which gene expression values of the PXE fibroblasts were compared.

Validation of ABC Transporter Transcript Levels and PXE Lineage Marker Gene Expression by Quantitative Real-Time PCR

Total RNA was extracted from 1 × 106 cells using the RNeasy Mini kit (Qiagen) according to the manufacturer's instructions and then stored at −80°C. The integrity of the RNA samples was checked by assessing ribosomal RNA bands on a denaturing agarose gel. Quantity and quality of the RNA samples were measured spectrophotometrically. Total RNA (1 μg) was transcribed to cDNA using oligo (dT) primers and Superscript II Reverse Transcriptase according to the manufacturer's instructions (Invitrogen, Karlsruhe, Germany). cDNA (equivalent to 12.5 or 25 ng of total RNA) served as a template for measurement of mRNA levels in quantitative real-time PCR (qPCR). Exon-intron-boundary-spanning PCR primers covering the major transcript forms were designed according to the published mRNA sequences (Table 2). QPCRs were performed using the Platinum SybrGreen PCR Supermix UDG kit from Invitrogen on Eppendorf Mastercycler Realplex2 System (Hamburg, Germany). PCR conditions were: 120 s incubation at 55°C, 120 s incubation at 95°C, followed by 45 cycles of degradation at 95°C for 10 s, primer-specific annealing for 15 s (Table 2), elongation and detection of the amplicon at 72°C for 20 s. Finally, a melting curve analysis of the amplicon was performed. Specificity of the amplicon was verified by standard methods. Data were analyzed using the method of Vandesompele et al to minimize interindividual sample variability.38 Initially, we defined the number of housekeeping genes necessary for accurate normalization by analyzing different housekeeping genes using Genorm software.38 Then, a normalization factor was determine by calculating a geometric mean of the expression values from three housekeeping genes (β-2-microglobulin=β-2m, glyceraldehyde-3-phosphate dehydrogenase=GAPDH, hypoxanthine phosphor-ribosyltransferase 1=HPRT1). Relative expression values were calculated by considering PCR efficiency and dividing by the normalization factor.

Table 2 Primer annotation and sequence, GenBank accession number, annealing temperature, amplicon size and melting temperature used in quantitative real-time PCR
Full size table

Measurement of MRP Efflux Activity by Flow Cytometry

Functional activity of MRPs was measured by quantifying calcein efflux as recently described.31 The principle of the assay in brief: cells were loaded with 0.1 μM of the nonfluorescent membrane-permeable calcein-acetoxymethyl (caAM) ester for 30 min at 37°C. The AM ester is cleaved by intracellular esterases to release the fluorochrom calcein (ca). Calcein is a specific substrate for MRPs, thus allowing MRP efflux activity to be determined. The cells were washed twice with DPBS afterward and then incubated at 37°C in DMEM depleted of FCS. Calcein efflux was quantified after 0, 60, 120, 240 and 360 min using Epics XL flow cytometer (Beckman Coulter, Krefeld, Germany). Data are presented as remaining intracellular calcein (cellular fluorescence) at time point 360 min (% intracellular ca=((ca)0 min−(ca)360 min)/(ca)0 min × 100).

Silencing of ABCC6/MRP6 Expression Using ABCC6-Specific Small-Interfering RNA

ABCC6-specific small-interfering RNA (siRNA) and FAM-labeled control siRNA oligonucleotides were purchased from Ambion (Cambridgeshire, UK; Table 3). NHDFs and PXE fibroblasts were reverse transfected using 4 μl/ml Lipofectamine 2000 (Invitrogen). Medium was replaced 12 h past transfection after washing cells twice with DPBS.

Table 3 Sequences of used ABCC6-specific siRNA-oligo and targeting sites
Full size table

Data and Statistical Analysis

All values are given as mean±s.e.m. Normality testing for Gaussian distribution of values was performed using Kolmogorov–Smirnov test. Statistical analysis was performed using Student's t-test and Mann–Whitney U-test where appropriate. P-values of less than 0.05 were considered significant. All tests were executed with GraphPad Prism 4.0 (GraphPad Prism Software, San Diego, CA, USA).

RESULTS

ABCC6 Genotyping

ABCC6 genotypes for the seven PXE patients investigated in the present study are shown in Table 1. Two patients were found to carry one ABCC6 mutation in homozygous state (P60F and P128M). Three patients were identified as being carriers of two different ABCC6 nonsense or splice-site mutations and one patient (P229F) was identified as carrying two missense mutations revealing a compound heterozygous status. Patient P308M revealed a novel 14-bp insertion in the promoter region in addition to a nonsense ABCC6 mutation. We failed to detect a second ABCC6 mutation for patient P341M.

Gene Expression Profile of ABC Transporter Genes in NHDFs

Characterization of the ABC transporter gene expression profile revealed 37 genes to be expressed in dermal fibroblasts from healthy controls (n=2) and PXE patients (n=4, Table 4). NHDFs revealed high expression levels (ΔCt 12–16) for ABC-A1, A2, A5, A6, B7, B8, B9, B10, C1, C4, C9, D1, D3, D4, E1, F1, F2, F3 and TAP1 and 2. Medium expression values (ΔCt 16.5–20) were observed for ABC- A3, A4, A7, A8, A9, A10, B4, B6, C3, C5, C10, D2 and G2. Low expression (ΔCt 20.5–25) was noticed for ABC- B5, C2, C6 and G4. TLDA analysis detected no gene expression for ABC- A12, A13, B1, B11, C7, C8, C11, G1, G5 and G8 in NHDFs.

Table 4 Gene expression profile of ABC transporters in dermal fibroblasts of healthy controls and PXE patientsa
Full size table

NHDF-Specific Gene Expression Profile of ABC Transporters Compared to HDFs of PXE Patients

No ABCC6 transcripts were quantifiable in fibroblast samples from PXE patients by TLDA analysis. Seven genes were found to be expressed more than twofold higher in PXE patients' fibroblasts (Table 4, depicted in dark gray). These include ABC- A6, A9, A10, B5, C2, C9 and D2. The relative mRNA expression of ABCA3 was found to be twofold reduced (≤0.5-fold) in the PXE patient's samples.

Validation of Altered ABC Transporter Transcript Levels Due to ABCC6/MRP6 Deficiency by Quantitative Real-time PCR

To verify the observed mRNA expression differences, we developed real-time PCR assays for selected members of the ABCC and ABCA subclasses, as well as for ABCD2, and reanalyzed gene expression in a larger group of PXE patients and healthy controls (each n=6). Samples of patient P60F were exchanged with samples of patient P308M in this analysis, due to a lack of material.

We developed a real-time PCR assay that permits relative quantification of ABCC6 mRNA expression. Our assay detects very low levels of ABCC6 mRNA without amplification of competing primer dimers (Figure 1a). To verify our assay, we also analyzed ABCC6 mRNA expression in HepG2 cells, a hepatocarcinoma cell line. ABCC6 mRNA expression levels were expressed 50-fold higher in HepG2 cells compared to NHDFs (Figure 1b). We found significantly lower ABCC6 mRNA expression in PXE patients compared to healthy controls (0.60±0.09 and 0.17±0.05, respectively, P<0.0001; Figure 2a). No ABCC6 transcripts were quantifiable in patients P265F, P3M, P128M and P341M (Figure 1c). Very low but clearly detectable ABCC6 mRNA levels were measured in PXE patients P229F and P308M.

Figure 1
figure 1

Quantification of ABCC6 mRNA expression. (a) Quantitative real-time PCR assay for detection of human ABCC6 transcripts. The figure shows the amplification plot, a corresponding melting curve of the amplicon and a standard curve generated with diluted NHDF cDNA (Slope −3.51; PCR efficiency 1.93). (b) Validation of the assay with cDNA synthesized from RNA samples of NHDF and HepG2 cells. (c) Quantification of ABCC6 mRNA expression in HDFs from healthy controls (CON, gray column; n=6) and PXE patients (white columns). Data are presented in arbitrary units as means with corresponding standard error; n.d.: not detectable.

Full size image
Figure 2
figure 2

Validation of ABC transporter transcript levels by quantitative real-time PCR. (a) Validation of ABCC subclass members mRNA expression by quantitative real-time PCRs in dermal fibroblasts derived from skin biopsies from PXE patients (white columns) and healthy controls (gray bars; n=6, each). (b) Validation of ABCA subclass members and ABCD2 mRNA expression by quantitative real-time PCRs in dermal fibroblasts derived from skin biopsies from PXE patients and healthy controls (n=6, each). Data are presented in arbitrary units as means with corresponding standard error. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

Full size image

The mRNA expression of the selected other ABC transporter genes were highly variable between the analyzed individuals but reflect the findings of the TLDA analysis. We found moderately but significantly higher ABCC2, ABCC4 and ABCC9 mRNA expression in PXE patients (Figure 2a). Significantly higher gene expression of ABCA6, ABCA9 and ABCA10 was detected in the patient's samples as already observed in the TLDA analysis (Figure 2b). ABCA3 transcript levels were moderately but significantly reduced in PXE patients (P<0.05). Analysis of ABCD2 mRNA expression revealed no significant difference comparing control and PXE fibroblasts.

Analysis of MRP Efflux Activity of Control and PXE Patient HDFs

Analysis of the MRP efflux activity revealed interesting results. We observed that dermal fibroblasts from male subjects exhibit significantly higher efflux rates. MRP efflux activities of dermal fibroblasts from male PXE patients were determined to be significantly reduced compared to male NHDFs (P<0.0001, Figure 3). In contrast, the dermal fibroblasts of female PXE patients showed no differences compared to female NHDFs (Figure 3).

Figure 3
figure 3

Analysis of MRP efflux activity of HDFs from healthy controls (gray bars) and PXE patients (white bars). Male HDFs show higher efflux rates than female HDFs. Male PXE patients exhibited a higher accumulation of calcein compared to male HDFs. Data are presented as remaining intracellular calcein (fluorescence) in % from default values at time point 0 min. CON=healthy controls. ****P<0.0001.

Full size image

Silencing of ABCC6/MRP6 Expression by ABCC6-Specific siRNA

We further examined the influence of MRP6 deficiency through usage of ABCC6-specific siRNA. HDFs from four different donors were transfected with ABCC6-specific siRNA and nonsense siRNA oligonucleotides. Three different ABCC6-specific siRNA oligonucleotides were tested in preceding experiments (Table 3). ABCC6-specific siRNA oligonucleotides 1 and 2 yielded a knockdown of ABCC6 mRNA expression of more than 90% (data not shown). ABCC6 mRNA expression was significantly downregulated with 40 nM of ABCC6-specific siRNA oligonucleotide 1 for 48, 96 and 168 h (Figure 4). Preliminary results demonstrated a reduced MRP efflux activity 168 and 240 h after transfection, although the reduction was only negligible (12% in the female and ∼5% in the male dermal fibroblast cultures compared to the scramble siRNA-transfected control culture, data not shown). Analysis of lineage markers of PXE reflected a PXE phenotype: ELN gene expression was significantly increased by up to 12-fold after 168 h incubation time in all four control fibroblast cultures (Figure 5a). Moreover, transcript levels of manganese superoxide dismutase (SOD2) were increased up to threefold (Figure 5b). Next, we analyzed the effect of the ABCC6-specific knockdown on the gene expression of other homologues ABCC transporter genes. We observed a slight effect of ABCC6-specific siRNA on ABCC1 and ABCC4 at early time points, returning to normal expression values after 168 h (data not shown). ABCC3 transcript levels remained unaffected throughout the experiment. Moreover, moderately increased transcript levels up to 1.5-fold were detected for ABCC5 at all time points analyzed (Figure 6b), whereas a significantly reduced gene expression was observed for ABCC2 and ABCC9 48 and 96 h after transfection (Figure 6a and b). We found a significantly elevated gene expression of ABCA3, ABCA9 and ABCA10 (Figure 7). Gene expression of ABCA6 was significantly reduced at early time points, returning to normal expression values after 168 h (data not shown). We performed the same experiments with ABCC6/MRP6-deficient cells from PXE patients P265F and P128M (Table 1). This experiments yielded interesting results: we observed similar significantly elevated ELN and SOD2 expression as already seen in the ABCC6/MRP6-deficient control fibroblasts (Figure 5). We also found elevated gene expression levels for ABCA3 and ABCA10, whereas transcript levels for ABCA9 were decreased in the silenced PXE fibroblasts at 48 and 96 h, with a moderate increase at time point 168 h (Figure 7). Moderately increased transcript levels up to 1.5-fold were detected for ABCC5 at all time points analyzed (Figure 6b). Futhermore, a significantly reduced gene expression was observed for ABCC2 and ABCC9 48 and 96 h after transfection (Figure 6a and c).

Figure 4
figure 4

Effect of siRNA-mediated knockdown on ABCC6 gene expression. Gray bars: cell cultures transfected with a scramble siRNA-negative control; white bars: ABCC6-specific siRNA-treated HDFs. Data are presented in arbitrary units as means with corresponding standard error at time points 48, 96 and 168 h. Experiments were performed with HDFs from healthy controls (n=4) and PXE fibroblasts (n=2). ****P<0.0001.

Full size image
Figure 5
figure 5

Effect of siRNA-mediated knockdown of ABCC6 gene expression on PXE lineage markers. (a) Relative ELN mRNA expression. (b) Relative SOD2 mRNA expression. Gray bars: cell cultures transfected with a scramble siRNA-negative control; white bars: ABCC6-specific siRNA-treated HDFs. Data are presented in arbitrary units as means with corresponding standard error at time points 48, 96 and 168 h. Experiments were performed with HDFs from healthy controls (n=4) and PXE fibroblasts (n=2). *P<0.05; **P<0.01; ****P<0.0001.

Full size image
Figure 6
figure 6

Effect of siRNA-mediated knockdown of ABCC6 gene expression on gene expression of ABCC transporter proteins. (a) ABCC2, (b) ABCC5 and (c) ABCC9. Experiments were performed with HDFs from healthy controls (n=4) and PXE fibroblasts (n=2). Gray bars: cell cultures transfected with a scramble siRNA-negative control; white bars: ABCC6-specific siRNA-transfected HDFs. Data are presented in arbitrary units as means with corresponding standard error at time points 48, 96 and 168 h. **P<0.01; ***P<0.001; ****P<0.0001.

Full size image
Figure 7
figure 7

Effect of siRNA-mediated knockdown of ABCC6 gene expression on gene expression of ABCA transporter proteins. (a) ABCA3, (b) ABCA9 and (c) ABCA10. Experiments were performed with HDFs from healthy controls (n=4) and PXE fibroblasts (n=2). Gray bars: cell cultures transfected with a scramble siRNA-negative control; white bars: ABCC6-specific siRNA-transfected HDFs. Data are presented in arbitrary units as means with corresponding standard error at time points 48, 96 and 168 h. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

Full size image

DISCUSSION

MRP6 deficiency is the cause for PXE manifestation due to mutations in the ABCC6 gene. To date, little is known about the implication of ABCC6/MRP6 in PXE pathogenesis. ABCC6/MRP6 is one of 49 more or less well-characterized ABC transporter genes. In the present study we investigated whether MRP6 deficiency leads to an altered gene expression profile of other ABC transporter genes. Several studies reported increased expression of related ABCC/MRP proteins as a mechanism to compensate a loss of function.32, 33

We detected altered gene expression of ABCA transporters and the homologous ABCC subfamily members 2, 4 and 9 in dermal fibroblasts derived from PXE patient's skin biopsies. The observed alterations in gene expression and MRP efflux activity reveal high interindividual variability due to different ABCC6 genotypes or genetic background. This may contribute to the great clinical variability observed in PXE patients. It is of important interest that ABCC6/MRP6-deficiency has a remarkable effect on the gene expression of ABCA subclass members. To date, ABCA1 is the best characterized member of this subclass. It is mainly involved in HDL biosynthesis and cholesterol efflux.39, 40 Polymorphisms in the ABCA1 were associated with cardiovascular disease and low plasma HDL levels.41, 42, 43 Wang et al previously reported that plasma HDL concentrations varied in a PXE patient and carriers of the ABCC6/MRP6 polymorphisms p.R1268Q.44 Furthermore, abcc6 knockout mice developed a 25% reduction in plasma HDL cholesterol.8 ABCA3 is a regulator of lamellar body metabolism and was reported to form lipid containing vesicles in human embryonic kidney cells.45 Little is known about the physiological function of the ABCA6-like transporters including ABCA6, ABCA9, ABCA10. The members of this subclass were reported to be regulated by cholesterol inversely to ABCA1 and therefore to act in opposed pathways.42 Gene expression of ABCA1 was previously shown to be positively and negatively regulated by products of the mevalonate pathway in different cell lines, as well in dermal fibroblasts.46 This pathway supplies polyisoprenoid-conjugates as percursors to further synthesize dolichol, menaquinone (vitamin Kn), tocopherol, ubichinone and finally cholesterol.47 It was previously hypothesized that vitamin K-conjugates might be possible ABCC6 substrates.48

We further observed enhanced ABCC2 gene expression in PXE patients samples; in female NHDFs and, most notably, in HDFs from female PXE patients (data not shown). These results are consistent with a previous study reporting gender differences in abcc2/mrp2 mRNA and protein expression in a mouse model for chronic kidney disease.49 Abcc2/mrp2, abcc3/mrp3 and abcc4/mrp4 expression in mice were also reported to be induced in the context of oxidative stress.50 Mild chronic oxidative stress might be involved in PXE manifestation.28 ABCC4/MRP4 and ABCC5/MRP5 were demonstrated to transport cyclic nucleotides and nucleotide analogs, even if controversial results were obtained regarding the strength of substrate affinity.51 Moreover, both transporters were thought to serve as overflow pumps under conditions where cGMP synthesis was impaired.52 Induction of ABCC4 and ABCC5 gene expression suggests an increase in cAMP and cGMP accumulation/efflux due to MRP6 deficiency in PXE. In contrast to the other members of subclass C, ABCC9/sulfonylurea receptor (SUR) 2 has no identified transport function. SURs are ATP-sensitive potassium ion channels which form dimers with other postassium-channels of the KIR-family. SURs are typical ABC transporter proteins due to their topology, and various compounds were demonstrated to modulate their activity as phosphoinositides and long-chain acyl coenzyme A derived from fatty acids.53 The fact that most of the genes with altered expression due to ABCC6/MRP6 deficiency belong to the ABCA subclass points to a role of ABCC6/MRP6 in lipid metabolism. This assumption is underlined by the identification of ABCC9/SUR2 as further candidate in PXE pathogenesis since ABCC9 is also affected by intermediates of lipid biosynthesis.

Our study design has several limitations due to the small sample size used and our major focus on investigating transcript levels. It is noteworthy that protein expression was not investigated due to study limitations, these may differ from mRNA expression. So, our new findings have to be confirmed on the level of protein expression which was behind the scope of our study. In order to investigate whether the impaired gene expression profile is really caused by MRP6 deficiency, and not by the high interindividual variability, we impaired ABCC6/MRP6 activity in vitro. Therefore, we knocked down ABCC6/MRP6 mRNA and protein expression by using target-specific siRNA to further analyze the influence of MRP deficiency on gene expression of other ABC transporter proteins and lineage markers of PXE manifestation. Indeed, specific downregulation of ABCC6 gene expression led to increased gene expression of the PXE marker genes ELN and SOD2. We further observed increased gene expression of ABCA3, ABCA9 and ABCA10 which corroborated the result from former experiments. In contrast, gene expression of ABCC2, ABCC9 and ABCA6 was controversially reduced. This needs to be further clarified and might be just due to off-target effects of siRNA silencing on highly homologous ABC transporter genes.

To our surprise, we observed similar siRNA-mediated effects on the gene expression profile in ABCC6/MRP6-deficient cells. To the best of our knowledge, siRNA silencing of target genes that cause genetic disorders has never been performed in cells, whose target expression is already disturbed. Therefore, it is quite difficult to evaluate the obtained results. The observed gene expression changes might be just due to off-target effects often mediated by siRNA silencing, particularly, in case of the PXE lineage markers ELN and SOD2 that are affected by several conditions, including stress. Besides we could detect small amounts of ABCC6 amplicons by performing standard PCRs pointing to a basal or residual ABCC6 gene expression in PXE fibroblasts (data not shown). Hence, another possibility is that also residual target gene expression might be affected by siRNA silencing leading to an intensification of an already existing phenotype. These concerns have to be clarified for instance by performing complementation studies and using further ABCC6-specific siRNA oligonucleotides targeting other sequences of the ABCC6 mRNA.

Our results show that the use of siRNA to downregulate ABCC6/MRP6 expression might be helpful for investigating PXE pathogenesis in cell culture models but that this strategy also reveals several limitations. We found no effect on the MRP efflux activity analyzing silenced control fibroblasts. The ABCC6/MRP6 knockdown mediated by siRNA might be accompanied by residual MRP6 protein. Moreover, we cannot exclude that ABCC6-specific siRNA has important off-target effects on other genes, particularly on highly homologues genes thus explaining the controversial results, especially obtained when analyzing samples from early time points after silencing. However, the usage of ABCC6-specific siRNA to knockdown ABCC6/MRP6 gene and protein expression is a valuable tool to determine what happens within the cell and to investigate the molecular mechanisms leading to PXE manifestation. Using different cell donors takes into account varying genetic backgrounds, which is important to detect common pathomechanisms induced by ABCC6/MRP6 deficiency. Therefore, our experimental design might be helpful in explaining the great clinical variability in PXE manifestation.

The observations of our study further support a role for ABCC6/MRP6 in transporting metabolites of sterol biosynthesis, as these function as important regulators of ABCA transporters. Many characteristic pathobiochemical hallmarks seen in PXE may arise from pathological changes in lipid metabolism, for instance, calcification, oxidative stress and enhanced matrix protein synthesis. It will be of great interest to uncover a possible role for ABCC6/MRP6 in this important metabolic pathway as this would be a major step toward disclosing the possible physiological substrate of ABCC6/MRP6.