Posttranscriptional Regulation of the Human LDL Receptor by the U2-Spliceosome
Abstract
Background:
The LDLR (low-density lipoprotein receptor) in the liver is the major determinant of LDL-cholesterol levels in human plasma. The discovery of genes that regulate the activity of LDLR helps to identify pathomechanisms of hypercholesterolemia and novel therapeutic targets against atherosclerotic cardiovascular disease.
Methods:
We performed a genome-wide RNA interference screen for genes limiting the uptake of fluorescent LDL into Huh-7 hepatocarcinoma cells. Top hit genes were validated by in vitro experiments as well as analyses of data sets on gene expression and variants in human populations.
Results:
The knockdown of 54 genes significantly inhibited LDL uptake. Fifteen of them encode for components or interactors of the U2-spliceosome. Knocking down any one of 11 out of 15 genes resulted in the selective retention of intron 3 of LDLR. The translated LDLR fragment lacks 88% of the full length LDLR and is detectable neither in nontransfected cells nor in human plasma. The hepatic expression of the intron 3 retention transcript is increased in nonalcoholic fatty liver disease as well as after bariatric surgery. Its expression in blood cells correlates with LDL-cholesterol and age. Single nucleotide polymorphisms and 3 rare variants of one spliceosome gene, RBM25, are associated with LDL-cholesterol in the population and familial hypercholesterolemia, respectively. Compared with overexpression of wild-type RBM25, overexpression of the 3 rare RBM25 mutants in Huh-7 cells led to lower LDL uptake.
Conclusions:
We identified a novel mechanism of posttranscriptional regulation of LDLR activity in humans and associations of genetic variants of RBM25 with LDL-cholesterol levels.
Graphical Abstract
Novelty and Significance
What Is Known?
•
The LDLR (LDL [low-density lipoprotein] receptor) regulates LDL-cholesterol levels in blood by mediating the uptake of LDL into hepatocytes.
•
The transcriptional and posttranslational regulation of LDLR activity is targeted by cholesterol lowering drugs.
What New Information Does This Article Contribute?
•
Loss of subunits or interactors of the U2 spliceosome decreases the uptake of LDL into Huh7 hepatocarcinoma cells. Loss of subunits or interactors of the U2 spliceosome also causes intron 3 retention of the LDLR mRNA and, thereby, loss of LDR function.
•
Intron 3 retention of LDLR in human liver and peripheral blood cells is increased by nonalcoholic fatty liver disease and aging, respectively.
•
Single nucleotide polymorphisms of the spliceosome gene RBM25 are associated with higher RBM25 expression in tissues and lower LDL-cholesterol.
•
Expression of rare structural variants of RBM25 that are associated with familial hypercholesterolemia (FH) decrease LDL uptake into Huh7 cells.
LDL-cholesterol is a causal and treatable risk factor of atherosclerotic cardiovascular diseases whose plasma level is most strongly determined by hepatic removal through the LDLR. LDLR activity is known to be regulated both by transcription of the LDLR gene and degradation of the LDLR protein. By genome-wide RNA interference, we identified 15 genes encoding subunits and interactors of the U2 spliceosome to limit the uptake of LDL into Huh7 hepatocarcinoma cells. We identified intron 3 retention of the LDLR mRNA as the underlying mechanism. The mRNA expression analysis of human liver samples and peripheral blood cells showed the high interindividual variation of this newly identified posttranscriptional regulation of LDLR. Intron 3 retention increases in nonalcoholic fatty liver disease as well as with ageing. Moreover, genetic variation in the U2 spliceosome gene RBM25 is associated with differences in LDL-cholesterol. Overall, we identified a novel mechanism of LDLR regulation which might help to better understand the etiology and pathophysiology of LDL-hypercholesterolemia.
Meet the First Author, see p 4
Hypercholesterolemia is a causal and treatable risk factor of atherosclerotic cardiovascular diseases.1 The most important determinant of LDL-C (low-density lipoprotein cholesterol) levels in plasma is the hepatic removal of circulating LDL by binding to the LDLR (LDL receptor) for subsequent endocytosis and degradation.2 The expression of LDLR is tightly regulated by transcription factors, proteasomal and lysosomal degradation, endosomal recycling, and cleavage at the cell surface.1,2 The unravelling of this complex regulation led to the development of drugs that effectively lower plasma levels of cholesterol and, as the consequence, risk of atherosclerotic cardiovascular diseases.1
To identify novel regulators of LDL uptake into the liver, we performed an image-based genome-wide RNA interference (RNAi) screen in Huh-7 human hepatocarcinoma cells. Fifteen out of 54 genes significantly reducing LDL uptake upon knockdown encode for proteins involved in pre-mRNA splicing. The majority of them are either core components or interactors of the U2-spliceosome.3 By functionally validating this finding in vitro as well as in human tissues, we provide evidence that a functional U2 spliceosome is needed for the expression of full length LDLR and, hence, determining LDLR activity in humans.
Methods
Data Availability
The authors declare that all data and methods supporting the findings of this study are available in the Supplemental Material or from the corresponding authors on reasonable request.
A detailed description of materials and methods is provided in the text and Major Resources Table of the Supplemental Material.
Results
The U2-Spliceosome and Its Interactors Are Rate-Limiting for LDL Endocytosis
For the genome-wide RNAi screen of genes limiting uptake of LDL or HDL (high-density lipoprotein), Huh-7 human hepatocarcinoma cells were reverse-transfected using 3 different siRNA oligonucleotides against each of the 21 584 different human genes. To control efficacy and specificity of transfection, each plate contained wells with cells transfected with siRNAs against PLK1 whose knockdown results in cell death, and LDLR, respectively. Based on results of time and dose finding experiments, the cells were exposed 72 hours posttransfection to 33 µg/mL each of Atto594-labelled LDL and Atto655-HDL for 4 hours. As background controls, wells with cells transfected with a nontargeting siRNA were incubated in the absence of fluorescent lipoproteins. After washing, fixation, and staining of the nuclei with Hoechst 33258, the plates were imaged at 4× and 20× with 2 twin wide-field automated microscopes. Nuclei, the relative cytoplasm, and fluorescent LDL-containing vesicles were identified through automated image analysis (Figure 1A). Transfection efficiency was very high (Figure S1A). Analysis and validation of HDL image data will be subject of a separate report.
For the uptake of fluorescent LDL, the 5 best performing assay features (foci count per cell, foci mean intensity, cytoplasm granularity 1 and 2, cytoplasm median intensity) showed a high degree of correlation. Therefore and because of the widest dynamic range based on Z’-factor values from control wells, we identified gene hits by the redundant siRNA activity analysis of data from the median cytoplasm intensity feature. Z’-factor values for median cytoplasm intensity in each assay plate for both the background (median 0.00 [interquartile range, −023 to 0.20]) and positive control (median, −0.56 [interquartile range, −0.99 to −0.20]) clustered mostly around the 0-line, indicating a suboptimal but analytically exploitable signal-to-noise ratio (Figure S1B). Dimensionality reduction of main assay features did not significantly alter the outcome (Figure S1C and S1D). At a redundant siRNA activity P value cutoff of P<10−3, interference with 54 and 37 genes decreased and increased LDL uptake, respectively (Table, Table S1). By contrast to the findings of a previous genome-wide CRISPR-based screening in Huh7 cells,4 our list does not include LDLR or its modulators such as SCAP, MBTPS1, or IDOL/MYLIP except AP2M1, which is an essential contributor to clathrin-mediated endocytosis (Table). Gene Ontology enrichment analysis showed significant clustering for genes whose loss of function decreased LDL uptake (Table S2). Functional clustering of these genes with the STRING tool revealed 4 major groups: the ribosome (N=7), the proteasome (N=8), the spliceosome (N=15), and vesicular transport (N=5; Figure 1B). Out of the 15 spliceosome genes, 6 encode for core components of the U2 spliceosome, namely SF3A1, SF3A2, SF3B1, SF3B2, SF3B5, and SF3B6. Other proteins, interact with the U2-spliceosome either directly (AQR [aquarius intron-binding spliceosomal factor], ISY1 [ISY1 splicing factor homolog], and RBM25 [RNA binding motif protein 25]) or indirectly (RBM22).3
| Decreased LDL uptake | Increased LDL uptake | ||||||
|---|---|---|---|---|---|---|---|
| Gene | Assay score* avg† | Assay score* SEM‡ | RSA, P value§ | Gene | Assay score* avg† | Assay score* SEM‡ | RSA, P value§ |
| AP2M1 | −3.103179681 | 0.346648222 | 3.36×10−8 | PROX1 | 6.53057396 | 0.631260417 | 3.19×10−9 |
| CHMP2A | −3.130900347 | 0.359445533 | 2.51×10−7 | ITGAV | 7.431175355 | 1.519558432 | 2.96×10−8 |
| NFKB2 | −2.59157417 | 0.136886566 | 8.07×10−7 | TGFBR1 | 3.464028514 | 0.397588943 | 7.31×10−6 |
| AQR∥ | −2.484868551 | 0.199482589 | 4.57×10−6 | CDC37 | 3.747034032 | 1.072191825 | 2.35×10−5 |
| PSMD11 | −2.557101583 | 0.239773488 | 4.77×10−6 | DTNBP1 | 57.92944887 | 57.2617451 | 4.46×10−5 |
| SF3B2∥ | −2.107311389 | 0.015210399 | 4.81×10−6 | CYP27C1 | 32.06817221 | 31.61438081 | 8.92×10−5 |
| RPL35 | −2.346954606 | 0.150946677 | 5.45×10−6 | PNPLA2 | 2.279278207 | 0.266420784 | 1.26×10−4 |
| PSMD8 | −2.988677308 | 0.491086915 | 6.34×10−6 | C22orf39 | 7.995494448 | 8.342242785 | 1.78×10−4 |
| SON∥ | −2.164748153 | 0.201099955 | 1.46×10−5 | TMEM133 | 3.049034762 | 1.060165442 | 1.84×10−4 |
| COPA | −2.307675328 | 0.213018879 | 1.61×10−5 | TMEM130 | 6.466491664 | 5.317700355 | 2.23×10−4 |
| RBM25∥ | −1.993998657 | 0.055265194 | 1.92×10−5 | PM20D2 | 2.155202336 | 0.176491898 | 2.29×10−4 |
| RBM22∥ | −2.818121291 | 0.622885617 | 3.36×10−5 | PET117 | 3.001069341 | 1.652765441 | 2.68×10−4 |
| PSMD3 | −2.21302629 | 0.224903034 | 3.98×10−5 | CWF19L2 | 3.806757511 | 4.571696977 | 3.12×10−4 |
| SF3B5∥ | −2.285064158 | 0.25878823 | 4.32×10−5 | ENY2 | 2.420424347 | 0.514153659 | 3.28×10−4 |
| SF3B1∥ | −2.267932169 | 0.253122099 | 4.55×10−5 | NME4 | 2.711491413 | 0.954425612 | 3.39×10−4 |
| SALL4 | −1.937993523 | 1.13065979 | 6.02×10−5 | ZC3H4 | 4.545156994 | 3.551478266 | 3.57×10−4 |
| RPL5 | −2.106905493 | 0.373542859 | 7.40×10−5 | WASF2 | 2.310874515 | 0.449202822 | 3.61×10−4 |
| CCDC180 | −1.132459235 | 1.398333381 | 9.52×10−5 | HELZ2 | 2.546828237 | 0.984740435 | 3.87×10−4 |
| SF3B6∥ | −2.277000896 | 0.332074616 | 9.83×10−5 | RILP | 1.995550072 | 0.267567916 | 4.23×10−4 |
| HNRNPU | −1.724036435 | 0.093847304 | 1.23×10−4 | MAT2A | 3.705559066 | 3.611891772 | 4.91×10−4 |
| RPL17 | −2.226845162 | 0.329575956 | 1.46×10−4 | NRM | 1.710898743 | 0.050727817 | 5.02×10−4 |
| ISY1∥ | −2.74487386 | 0.698388989 | 1.55×10−4 | CEP295NL | 2.189792071 | 0.474598108 | 5.02×10−4 |
| ZNF641 | −1.034460324 | 1.453444444 | 2.58×10−4 | ACSM2A | 2.207444199 | 1.531809937 | 5.32×10−4 |
| COPB1 | −1.693933632 | 0.103029465 | 2.64×10−4 | RTL9 | 3.759306708 | 3.473297986 | 5.35×10−4 |
| SF3A1∥ | −2.225755015 | 0.412106586 | 2.72×10−4 | KIAA1522 | 3.362058267 | 3.27466253 | 6.25×10−4 |
| SNW1∥ | −1.76539067 | 0.142531611 | 2.76×10−4 | ZNF84 | 2.204388764 | 0.765657329 | 6.55×10−4 |
| EIF2S1 | −1.486721463 | 0.790741651 | 3.45×10−4 | TFAP4 | 3.032765033 | 3.340175625 | 6.69×10−4 |
| CCDC73 | −1.041204586 | 1.27682775 | 3.50×10−4 | TMEM182 | 3.227517874 | 1.666669492 | 7.29×10−4 |
| RPL9 | −1.715182797 | 0.249911985 | 3.55×10−4 | WDR55 | 1.967286849 | 1.365170916 | 7.32×10−4 |
| NXNL2 | −1.199311468 | 1.135835784 | 3.83×10−4 | DYNLL1 | 2.268266743 | 0.467997927 | 7.72×10−4 |
| WBP11∥ | −1.50591484 | 0.062444555 | 4.03×10−4 | ADPRHL2 | 2.078229013 | 0.322800093 | 8.51×10−4 |
| C2CD5 | −1.097951788 | 1.954971449 | 4.46×10−4 | ELAVL1 | 1.945364959 | 0.968117905 | 8.70×10−4 |
| RPL21 | −1.655773242 | 0.156797718 | 4.72×10−4 | CFAP298 | 1.883199038 | 0.378258022 | 8.87×10−4 |
| EPOP | −1.837314819 | 0.25795876 | 4.80×10−4 | PMM1 | 2.80926863 | 3.200260012 | 8.92×10−4 |
| RMND5B | −1.523957521 | 0.076773849 | 5.07×10−4 | CASKIN2 | 1.681223061 | 0.149986926 | 9.07×10−4 |
| TAPBPL | −1.52965773 | 0.154207886 | 5.27×10−4 | CIZ1 | 3.454694336 | 2.803876145 | 9.37×10−4 |
| STARD10 | −1.527795273 | 0.115135889 | 5.45×10−4 | BRICD5 | 1.962503862 | 0.408074057 | 9.41×10−4 |
| PSMD1 | −2.207116523 | 0.551747426 | 5.63×10−4 | ||||
| PFDN6 | −0.881689024 | 1.740601376 | 5.80×10−4 | ||||
| PSMA1 | −1.528976301 | 0.119805079 | 5.85×10−4 | ||||
| RTF2 | −1.573924771 | 0.169765686 | 6.14×10−4 | ||||
| LSM2∥ | −1.448888015 | 0.056454941 | 6.40×10−4 | ||||
| UBD | −1.171691024 | 1.530009178 | 6.69×10−4 | ||||
| LRRC14 | −1.258311764 | 1.067910962 | 6.84×10−4 | ||||
| SUPT6H∥ | −1.451332382 | 0.095214513 | 7.27×10−4 | ||||
| COPB2 | −2.037140764 | 0.468882876 | 7.34×10−4 | ||||
| SF3A2∥ | −1.347147433 | 0.758462926 | 7.89×10−4 | ||||
| ATP6V0C | −1.823918839 | 0.263639476 | 7.90×10−4 | ||||
| EMILIN3 | −1.598631472 | 2.238859705 | 8.03×10−4 | ||||
| DMTN | −1.559252376 | 0.142024687 | 8.20×10−4 | ||||
| MRPL19 | −0.755460842 | 1.688052373 | 8.92×10−4 | ||||
| MRO | −0.986783025 | 1.102624895 | 9.14×10−4 | ||||
| DDX59 | −1.380513222 | 1.040634076 | 9.25×10−4 | ||||
| PSMD12 | −1.761325035 | 0.367766123 | 9.45×10−4 | ||||
LDL indicates low-density lipoprotein.
*
Assay score: normalized score for the median cytoplasm intensity assay feature.
†
Average.
‡
SEM.
§
P values are not adjusted for multiple testing (P<3.6×10−6 after Bonferroni adjustment for 14 000 genes with expressed transcripts).
∥
The 15 hit genes involved in RNA splicing and validated.
To confirm the role of the U2 spliceosome in LDL endocytosis in vitro, we performed 125I-LDL cell association assays in Huh-7 and HepG2 cells. SF3B4 was also included in these experiments as it is part of the U2 spliceosome and barely missed the redundant siRNA activity P value cutoff (P=1.4×10−3). Knockdown was achieved using 4 pooled siRNA molecules against each hit gene acquired from vendors other than that of the siRNA screening library, namely Dharmacon or Sigma instead of Ambion (see Major Resource Table and Figure S2A). For RBM25, we replaced Dharmacon’s siRNAs with those from Sigma because of their presumable off-target effects on LDLR protein expression (Figure S3). Knockdown of each of these genes significantly decreased the specific cell association of 125I-LDL with both Huh-7 and HepG2 cells (Figure 1C and 1D and Figure S2B). The association of 125I-LDL was equally decreased by knockdown of SF3B1 (−45±5%), SF3A2 (−47±6%), AQR (−45±6%), and LDLR (−43±8%; Figure 1C). RNAi with RBM25 reduced the specific cellular association of 125I-LDL and fluorescent LDL by 27±8% and 52±5%, respectively (Figure 1D and Figure S3F). Of note, the specific cell association of 125I-HDL was unaltered or even increased upon knockdown of AQR and SF3A1 in either Huh-7 or HepG2 cells (Figure S2C and S2D).
Loss of U2-Spliceosome Genes and Their Interactors Causes Selective Retention of LDLR Intron 3 (IVS3)
To unravel the mechanism through which the U2-spliceosome and its interactors regulate LDL endocytosis, we applied RNA sequencing to Huh-7 cells, which were transfected with either siRNAs against eleven U2-spliceosome genes or a nontargeting control siRNA. Sequences can be accessed by codes PRJEB46899 and PRJEB46898 in the data bank of the European Nucleotide Archive (https://www.ebi.ac.uk/ena/browser/support). 72 hours after transfection, we measured both expression at the gene level and alternative exon usage in polyA-selected transcripts. Knockdown of all eleven genes except RBM25 induced a marked increase in the retention of intron 3 of LDLR in mature transcripts without altering the expression of the LDLR full length transcript (Figure 2A, Figure S4). This effect was confirmed in Huh-7 cells by quantitative real-time polymerase chain reaction (qRT-PCR) upon knockdown of AQR, SF3B1, or RBM25 by employing a primer set that was previously used to study the effects of the rare LDLR c.313+1, G>A intronic variant, which leads to LDLR loss of function by constitutively promoting intron 3 retention5 (Figure S5A). By contrast to the RNA sequencing (Figure S4), qRT-PCR unravelled increased expression of the LDLR IVS3 retention transcript upon knockdown of RBM25, albeit not as much as with knockdown of SF3B1 and AQR (Figures S5B and S5C).
Among all intronic or exonic sequences in the transcriptome, the expression of the intron 3 retaining LDLR transcript was altered most strongly. Upon knockdown of SF3B1, AQR, or SF3A2, the retained intronic sequence of LDLR ranked at the top of each respective data set when the exon-level expression data was plotted against each other (Figure 2B). The degree of intron 3 retention upon knocking down U2-spliceosome genes was significantly correlated with the decrease in 125I-LDL cell association, (r=−0.73, P=1.4×10−2, Figure 2C).
To investigate reasons for intron 3 retention in LDLR, we transfected HEK293T cells with 2 minigenes containing different portions of the LDLR genomic sequence flanked by 2 artificial exons (Figure 3A). The first minigene (MG1) encoding only for exon 3 of LDLR and the adjacent intronic regions cloned between 2 artificial exons (SD6 and SA2), displayed very low if any RNA sequencing reads mapping to the first ≈130 bp of intron 3. On the contrary, upon expression of the whole genomic sequence between the 3′-end of intron 2 and the 5′-end of intron 4 of LDLR (MG2) an increased number of reads mapped to the first section of intron 3. This indicates incomplete splicing of intron 3 when the physiological exon 4 acceptor site and the branch point site were present in the larger minigene MG2 (Figure 3B). The acceptor splice site of exon 4 of LDLR hence appears to be poorly defined. The bioinformatic analysis of the portion of intron 3 neighbouring exon 4 by the U2 branchpoint prediction algorithm SVM-BP-finder (http://regulatorygenomics.upf.edu/Software/SVM_BP/)6 identified one plausible U2-spliceosome dependent branch point site located 30 bp upstream of the acceptor site (Table S3). The gtgat pentamer in the center of the cggtgatgg branchpoint sequence was associated with very low U2 binding energy and occurs at low frequency in the branchpoint database.6 We discarded another predicted branchpoint 124 bp upstream of the acceptor site as the subsequent AG-exclusion zone does not reach up to the acceptor. Contrary to exon 4 of human LDLR, exon 4 of murine Ldlr contains a strong and frequently recurring branchpoint 33 bp upstream of the acceptor site (Figure 3C). This finding is in accordance with intron 3 of Ldlr being barely detectable at the RNA level by qRT-PCR in mouse liver (data not shown). Taken together, these data suggest that the branch point site of intron 3 in human LDLR is poorly defined and, therefore, very sensitive to alternative splicing.
Selective Intron 3 Retention Limits LDLR Cell Surface Abundance
The transcript with intron 3 retention encodes for a prematurely truncated proteoform of LDLR because the 5′-end of intron 3 encodes for 12 novel amino acids followed by a stop codon. Including the signal peptide, this theoretical 116 amino acid residues long and 12.7 kDa large LDLRret fragment encompasses the complete first and large part of the second class A domains (labelled as L1 and L2 in Figure 4A7) but lacks all other domains, including the transmembrane portion of LDLR. Western blots probed with an antibody against the C-terminus of LDLR revealed 60±30% and 61±13% lower LDLR protein levels upon knockdown of AQR and SF3B1, respectively (Figure 4B and 4C). A similar decrease in LDLR protein was seen upon knockdown of RBM25 with siRNAs from Sigma (−68±10%), whereas the knockdown of RBM25 with the siRNA of Dharmacon led to an increase in LDLR protein (122±109%), presumably due to off-target effects (Figure S3D and S3E). Flow cytometry experiments on alive Huh-7 cells after SF3B1 and AQR knockdown showed a −87±1% and –61±4%, respectively, lower cell surface abundance of LDLR (Figure 4D). The knockdown of RBM25 with siRNAs from Sigma and Dharmacon decreased the cell surface abundance of LDLR by 53±6% and 21±5%, respectively, as compared with nontargeting siRNAs from the respective manufacturers (Figure S3G).
To investigate whether cells produce and secrete the LDLRret fragment, we overexpressed a C-terminally HA-tagged version of the LDLRret fragment in HEK293T cells. Forty-eight hours after transfection, the HA-tagged LDLRret fragment was detectable in the cell lysates (Figure 4E) as well as in undiluted cell culture media (Figure 4F). The proteasomal inhibitor MG-132 decreased cellular LDLRret protein levels (Figure 4E) suggesting that the LDLRret fragment is not catabolized through the proteasome. We also overexpressed an untagged version of the LDLRret fragment in HEK293T cells. Targeted mass spectrometry recorded a peptide, which is present in both the full-length protein and in LDLRret, over its basal endogenous level in HEK293T cell lysates (Figure S6) but not in human plasma (data not shown).
A Large Proportion of LDLR Transcripts in Human Liver and Blood Cells Retains Intron 3
To investigate its physiological or pathological relevance, we quantified LDLR intron 3 retention in liver biopsies as well as in peripheral blood cells by three different methods, and explored associations with nonalcoholic fatty liver disease (NAFLD), demographic measures, lipid traits, and therapeutic interventions.
qRT-PCR of mRNAs of liver tissue from 17 patients with benign liver tumours and 9 patients with suspected NAFLD, found the LDLR intron 3 retention transcript expressed at considerable and interindividually variable amounts (Figure 5A). Taking the sum of the full length and intron 3 retention transcripts of LDLR as the reference, 43% (range, 23%–85%) of the transcripts retained intron 3 (Figure 5A).
The bioinformatics analysis of RNA sequencing data on liver samples of 13 healthy nonobese subjects, 12 obese subjects without NAFLD, 15 patients with NAFLD, and 15 patients with nonalcoholic steatophepatitis (NASH; Gene Expression Omnibus, accession number GSE126848)8 found 14 different LDLR transcripts (Figure S7). Four transcripts showed the largest interindividual variation, namely LDLR-201 and LDLR-208, encoding full length LDLR, as well as LDLR-206, which corresponds to the retained intron 3 transcript, and the likewise futile LDR-214 (LDLR transcripts are illustrated schematically in Figure S7A). Interestingly, the median concentration of LDLR-206 was substantially higher in patients with NAFLD or NASH than in normal weight or obese subjects without NAFLD. The median percentages of LDLR-206 reads relative to total reads from all transcripts of LDLR gene increased significantly from 1.8% (range, 0.7%–4.2%) and 1.7% (0.4%–3.7%) in normal weight and obese subjects without NAFLD, respectively, to 5.8% (1.1%–26.7%) and 5.0% (0.9%–29.0%) in patients with NAFLD and NASH, respectively (Figure 5B). Of the 2 most abundant full length encoding LDLR transcripts, LDLR-208 decreased significantly (Figure 5C) while the expression of LDLR-201 did not change (Figure S7).
We also investigated the expression of LDLR transcripts in liver biopsies of 155 obese nondiabetic subjects9 by using Affymetrix Human Gene 2.0 ST arrays (see Table S4 for clinical and biochemical characteristics). The signal intensities from a probe located in intron 3 of LDLR were significantly higher than the other intronic LDLR probes located in introns 2, 4, and 15 and comparable to probes located in coding exons (Figure 5D). The percent intensities of the IVS3 probe relative to the sum of all LDLR probes ranged from 7.5% to 82%. Intron 3 retention correlated significantly only with SF3B1 (r=0.26, P=1.4×10−2), while no U2-spliceosome gene showed any significant correlation with overall LDLR expression (Table S5). Relative intensities of neither the intron 3 probe nor any other of the 24 LDLR probes showed significant correlations with plasma levels of total, HDL- or LDL-cholesterol (Figure S8A–S8C and Table S6). Correlations with histological NAFLD stages were inverse by trend but not statistically significant (Figure S8D). Intron 3 relative probe intensity did not correlate with body mass index (Figure S8E). However, in a subgroup of 21 patients who underwent a second liver biopsy after bariatric surgery (median follow-up time, 13 months [interquartile range, 12–15]), the proportion of the intron 3 retention transcript relative to the full length LDLR transcript increased significantly after surgery (P=9.8×10−3; Figure S8F; Table S4). This increase was even more pronounced in eleven patients with NASH at baseline but no NASH at follow-up (P=3.6×10−2, Figure S8G).
Finally, we analyzed the RNA sequencing data in whole blood samples from 2462 subjects of the Dutch BIOS-consortium.10 The LDLR ENST00000557958 transcript, predicted to retain intron 3, was detectable in all subjects and represented 21±7% of the total LDLR transcripts. The ENST00000557958 transcript levels significantly correlated with age (r=0.25, P=9.2×10−36, Figure 6A) and less strongly with LDL-C (r=0.089, P=3.9×10−5, Figure 6B). The latter correlation lost its statistical significance after adjusting for age, suggesting age itself as the main driver of the association between ENST00000557958 levels and LDL-C. ENST00000252444, the only transcript encoding for full length LDLR and expressed in blood cells in all subjects in this data set, was also positively correlated with age (r=0.19, P=8.8×10−20, Figure 6C) but not with LDL-C (r=−0.033, P=4.0×10−1, Figure 6D). Correlation of neither transcript with body mass index was statistically significant.
Single Nucleotide Polymorphisms in RBM25 Are Associated With Lower LDL-Cholesterol
The analysis of whole exome sequencing data of 40 468 UK Biobank subjects11 did not unravel any significant association between our spliceosome hit genes and LDL-C or any other clinical lipid trait (Table S7). However, constraints data from the gnomAD database indicate a strong intolerance to functional genetic variation for our U2-spliceosome genes, with a probability of intolerance to loss of function12 of 0.91±0.17 (mean±SD; Table S8). The analysis of SNPs of 11 U2-spliceosome hit genes in 361 194 participants of UK Biobank found 24 SNPs of RBM25 significantly associated with lower levels of LDL-C (Figure 7A) and apoB (Figure S9A).
In Europeans, 4 SNPs in introns or downstream of the RBM25 coding sequence including the lead SNP rs17570658 and 2 upstream SNPs are in almost complete LD (Figure S9B). With R2 >0.8 no other SNP of RBM25 is in strong LD. A meta-analysis of 8 studies with 455 537 samples (https://cvd.hugeamp.org/variant.html?variant=rs17570658) and data of the Copenhagen City Heart and General Population Studies13 according to METAL14 showed the association of rs17570658 with LDL-C (Z Score=−4.181, P=2.9×10−5, Table S9).
RBM25 is widely expressed in many tissues, but expression is relatively low in liver (GTeX https://gtexportal.org/home/, data not shown). rs17570658 shows strong association with RBM25 expression in 15 different tissues including skeletal muscle and arteries (Figure 7B) as well as adipose and mammary tissue, lung, oesophagus, kidney, and skin. Carriers of the rare allele have higher mean RBM25 mRNA concentration, which is compatible with higher LDLR activity and lower LDL-C levels.
Impaired LDL Uptake by Cells Expressing Rare RBM25 Mutants Found in Patients With Familial Hypercholesterolemia
In the UK10K study, RBM25 was also among the genes identified to harbour an excess of rare novel variants in 71 patients with familial hypercholesterolemia who are negative for mutations in LDLR, APOB, and PCSK9, the known familial hypercholesterolemia (FH)-causing genes.15 We reanalyzed the burden of variants in the RBM25 gene, using previously published whole exome sequencing data from 71 FH patients negative for mutations in LDLR, APOB and PCSK9, and 56 352 European data provided by the gnomAD study.12 Missense, splice site, frameshift, and stop-gained variants identified by whole exome sequencing in both FH cases and gnomAD were filtered to select those with minor allele frequency (MAF) <1.0×10−4. After filtering, three RBM25 variants were found in the FH cohort and 163 in the gnomAD Europeans cohort. (Table S10). Two variants, p.I152F (c.454A >T) and p.A455D (c.1364C >A), were not found in any publicly available sequencing database and hence appear unique to the FH cohort. The third variant, p.L17P (c.50T >C; rs1167173761), was found in one European individual in the gnomAD cohort (MAF=9×10−6, allele count=1/251402). The comparison of variant numbers in FH cases versus gnomAD using a binomial test demonstrated the enrichment of rare variants in RBM25 in the FH cohort (P=1.0×10−3). Within the UK10K cohort, no other U2-spliceosome gene was found to carry a rare presumable LOF mutation.
We investigated the functional consequences of overexpressing the 3 FH-associated RBM25 mutants in Huh-7 cells. Overexpression of all RBM25 constructs was confirmed by qRT-PCR (Figures S10A and S11A) and—for wild-type RBM25—Western blotting (Figure S10B). The overexpression of neither wild-type RBM25 nor any RBM25 mutant in Huh-7 cells caused significant changes in the expression of full length or IVS3 retention transcripts of LDLR (Figures S10C, S10D, S11B, and S11C). Compared with empty vector, overexpression of wild-type RBM25 in Huh-7 cells changed neither the cell surface abundance of LDLR nor LDL uptake significantly (Figure S10E and S10F). Comparisons with cells overexpressing wild-type RBM25 revealed minor decreases of LDLR cell surface levels (Figure S11D) but more pronounced or even significant decreases of Atto655-LDL uptake of cells overexpressing the RBM25 mutants p.L17P (−15±16%), I152F (−23±12%), or p.A455D (−28±12%, P=2.6×10−2; Figure S11E).
Discussion
Through genome-wide siRNA screening, we discovered that the U2-spliceosome as well as some interacting proteins, control LDLR levels and LDL uptake in liver cells by modulating the selective retention of intron 3 of LDLR. The intron 3 retaining LDLR transcript encodes a truncated and most probably nonfunctional receptor. In several cohorts of healthy individuals and patients, we observed considerable interindividual variation of LDLR’s IVS3 retention in liver as well as in peripheral blood cells. Finally, we obtained initial evidence that rare genetic variants as well as SNPs associated with its expression levels in the U2-spliceosome-associated gene RBM25 are related to LDL-C levels in humans. Taken together, our findings suggest intron 3 retention of LDLR as a novel mechanism regulating LDLR activity and thereby plasma levels of LDL-C.
A previous siRNA screen also found U2-spliceosome genes to limit the uptake of LDL into EA.hy926 cells but the authors excluded them from further analysis and validation.16 Basic cellular functionality of spliceosome genes may be the reason why U2- spliceosome genes were not found by a previous CRISPR-based screen as limiting factors for LDL uptake into Huh-7 cells.4 As these authors discussed, CRISPR-based screens may overlook genes that are essential or confer a fitness advantage in culture, since guide RNAs targeting those genes will be progressively depleted from the pooled population.4
As a preliminary mechanistic explanation, our minigene data as well as our in silico predictions suggest that the branch point site in intron 3 of human LDLR is poorly defined and thereby highly sensitive to alterations in the activity of U2 splice factors. In this regard it is noteworthy that the rare c.313+1, G>A intronic variant leads to loss of LDLR function by constitutively promoting IVS3 retention.5
Medina and Krauss17 previously found alternative splicing of HMGCR, HMGCS1, MVK, PCSK9, and LDLR to be mediated by the splice protein PTBP1 and regulated by cellular cholesterol levels. Interestingly, PTBP1 works as an inhibitor of the U2AF splice component, and thus inhibits the recognition of 3′ splice sites by the U2-spliceosome.18 However, the knockdown of PTBP1 resulted in very limited changes in the expression levels of the different splice forms,17 especially when compared with the drastic changes observed in our study.
In our in vitro experiments, the knockdown of several U2-spliceosome genes and the resulting IVS3 retention compromised LDLR cell surface expression and LDL uptake as much as the knockdown of LDLR itself. The sensitivity of our mass spectrometric analysis only allowed detection of the tagged fragment after overexpression in the immortalized kidney cell line HEK293T. The artificial construct unlike an endogenously produced protein may have escaped nonsense-mediated decay. Nevertheless, we cannot rule out that the theoretical 116 amino acid long aminoterminal fragment of the differentially spliced LDLR is expressed in vivo and secreted. In fact, human plasma contains LDLR fragments, which are currently assumed to result from shedding of LDLR at the cell surface19 but may also correspond to secreted alternative splice variants.
The relative expression of LDLR’s IVS3 transcript in human liver varies strongly due to both analytical and biological reasons, namely between 0.4% and 29% upon RNA sequencing, between 7.5% and 81% upon chip array analysis, and between 23% and 85% upon qRT-PCR. Very likely, RNA sequencing yielded the most realistic data, because this method recorded the different LDLR transcripts most comprehensively. The large interindividual variation of IVS3 expression recorded by each method indicates relevant regulatory mechanisms and consequences. We made seemingly contradictory findings on the association of IVS3 retention with NAFLD. On the one hand, the percentage of IVS3 transcripts was significantly higher in 30 patients with NAFLD or NASH than in 25 normal weight and obese control subjects without NAFLD. On the other hand, the chip array analysis found significant increases of IVS3 transcripts after bariatric surgery, which rather causes regression of NAFLD. However, although causing regression of NASH, bariatric surgery may not necessarily undo all regulatory abnormalities associated with NAFLD. In this regard, it is noteworthy, that neither RNA sequencing nor chip array analysis found any significant effect of NASH on IVS3 retention (Figure 5B and Figure S8D). Larger studies are hence needed to answer the question how NAFLD influences the expression of functional and nonfunctional LDLR transcripts.
In peripheral blood cells but not in liver tissue, we found a significant correlation between plasma LDL-C levels and the IVS3 retention LDLR transcript, which was stronger than the correlation with the full-length LDLR transcript. Smaller sample size and narrower range of LDL-C levels but also differences between tissues may be the reasons, why no significant correlations of LDL-C with any hepatic LDLR transcript expression were found. However, the associations of RBM25 SNPs with differences in RBM25 expression and LDL-C levels and the higher than expected prevalence of rare RBM25 loss of function variants in FH patients with no mutation in canonical FH genes suggest that the regulation of LDLR splicing by the U2-spliceosome contributes to the determination of LDL-C levels in humans.
The lack of association of hypercholesterolemia with rare variants of any other U2- spliceosome gene may reflect their intolerance to gross variation as suggested by probability of intolerance values close to 1. Also of note, the analysis of whole exome sequencing data from the UK biobank only retrieved heterozygous mutations in U2-spliceosome genes whereas our knockdown experiments rather mimic homozygous conditions. Opposite effects on upstream regulators of LDLR may be another reason why the majority of SNPs and rare exome variants of the spliceosome genes do not show any association with LDL-C levels. The exclusive association of LDL-C with RBM25 variants may also indicate that RBM25 regulates LDL-C levels by mechanisms unrelated to the U2-spliceosome and the intron 3 retention. In fact, RBM25 also partakes in other spliceosomal subunits.20 Of note, RNAi with RBM25 had the weakest effects on LDLR splicing and overexpression of hypercholesterolemia associated RBM25 mutants in Huh-7 cells resulted in lower LDL uptake without affecting the expression of the LDLR IVS3 transcript.
The correlation between ENST00000557958 expression in blood cells with age makes us hypothesize that age-related changes in the activity of the U2-spliceosome contributes to the increase in LDL-C that parallels ageing21 but is not mechanistically understood. The functionality of the splicing process changes with ageing.22 Somatic mutations or decreased expression of splice factor genes, notably SF3B1 and RBM25 have been implicated in age-related processes, including cancer.22,23 The total number of alternatively spliced genes also increases with age.24 Until recently, SIRT1 is the only known gene involved in cholesterol metabolism and atherosclerosis25 whose alternative splicing may be disrupted with age.22 One may speculate that either the epigenetic dysregulation of the activity of splice factor genes or the accumulation of somatic loss of function variants in liver cells may promote increases in LDL-C with age.
Our study has several strengths and limitations. First, our screening unravelled several novel candidate genes that regulate hepatic LDL uptake but missed canonical LDL uptake regulating genes such as MYLIP, MBTPS1, PCSK9, or SREBP2. A general reason is the not optimal signal-to-noise ratio of our screening. A specific reason for missing MYLIP or PCSK9 is the optimization of our screening towards the discovery of loss of function effects. Second, our validation studies did not only confirm the limiting effect of U2-spliceosome genes on LDL uptake but unravelled a novel mechanism of LDL receptor regulation, namely IVS3 retention within an LDLR transcript which is translated into a truncated and nonfunctional receptor protein. In both human liver and peripheral blood cells, we demonstrate that this process happens at considerable quantity and interindividual variability, possibly influenced by aging and NAFLD. Third, RBM25 was the only spliceosome gene affected by mutations associated with differences in LDL-C, perhaps because RBM25 may tolerate loss of function better than other U2-spliceosome genes. However, we cannot rule out that RBM25 affects LDL metabolism beyond or even independently of LDLR splicing because both knockdown of RBM25 and overexpression of loss of function mutants associated with hypercholesterolemia exerted in Huh-7 cells stronger and more consistent effects on LDL uptake than on IVS3 retention in LDLR.
In conclusion, we identified IVS3 retention of LDLR upon loss of U2-spliceosome activity as a novel mechanism regulating LDLR activity in cells. The importance of this mechanism for the regulation of plasma LDL-C levels and thus determination of cardiovascular risk remains to be established by further studies.
Article Information
Supplemental Material
Materials & Methods
Figures S1–S11
Tables S1–S12
Major Resource Table
Original Western Blots
Acknowledgments
We acknowledge the use of data from BIOS-consortium (http://wiki.bbmri.nl/wiki/BIOS_bios) which is funded by BBMRI-NL (NWO project 184.021.007). Flow cytometry was performed with equipment of the flow cytometry facility, University of Zurich.
Footnote
Nonstandard Abbreviations and Acronyms
- AQR
- aquarius intron-binding spliceosomal factor
- FH
- familial hypercholesterolemia
- HDL
- high-density lipoprotein
- LDL-C
- low-density lipoprotein cholesterol
- LDLR
- LDL receptor
- NAFLD
- nonalcoholic fatty liver disease
- NASH
- nonalcoholic steatophepatitis
- qRT-PCR
- quantitative real-time polymerase chain reaction
- RBM25
- RNA binding motif protein 25
- RNAi
- RNA interference
Supplemental Material
File (318141_online.pdf)
- Download
- 2.94 MB
File (318141_online_data_set.csv)
- Download
- 20.49 MB
File (318141_uncut_gel_blots.pdf)
- Download
- 11.00 MB
File (circres_circres-2020-318141_supp1.pdf)
- Download
- 2.94 MB
File (circres_circres-2020-318141_supp2.csv)
- Download
- 20.49 MB
File (circres_circres-2020-318141_supp3.pdf)
- Download
- 11.00 MB
References
1.
Ference BA, Ginsberg HN, Graham I, Ray KK, Packard CJ, Bruckert E, Hegele RA, Krauss RM, Raal FJ, Schunkert H, et al. Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A consensus statement fromthe European Atherosclerosis Society Consensus Panel. Eur Heart J. 2017;38:2459–2472. doi: 10.1093/eurheartj/ehx144
2.
Zanoni P, Velagapudi S, Yalcinkaya M, Rohrer L, von Eckardstein A. Endocytosis of lipoproteins. Atherosclerosis. 2018;275:273–295. doi: 10.1016/j.atherosclerosis.2018.06.881
3.
Wilkinson ME, Charenton C, Nagai K. RNA splicing by the spliceosome. Annu Rev Biochem. 2020;89:359–388.
4.
Emmer BT, Sherman EJ, Lascuna PJ, Graham SE, Willer CJ, Ginsburg D. Genome-scale CRISPR screening for modifiers of cellular LDL uptake. PLoS Genet. 2021;17:e1009285. doi: 10.1371/journal.pgen.1009285
5.
Cameron J, Holla ØL, Kulseth MA, Leren TP, Berge KE. Splice-site mutation c.313+1, G>A in intron 3 of the LDL receptor gene results in transcripts with skipping of exon 3 and inclusion of intron 3. Clin Chim Acta. 2009;403:131–135. doi: 10.1016/j.cca.2009.02.001
6.
Corvelo A, Hallegger M, Smith CW, Eyras E. Genome-wide association between branch point properties and alternative splicing. PLoS Comput Biol. 2010;6:e1001016. doi: 10.1371/journal.pcbi.1001016
7.
Surdo PL, Bottomley MJ, Calzetta A, Settembre EC, Cirillo A, Pandit S, Ni YG, Hubbard B, Sitlani A, Carfí A. Mechanistic implications for LDL receptor degradation from the PCSK9/LDLR structure at neutral pH. EMBO Rep. 2011;12:1300–1305. doi: 10.1038/embor.2011.205
8.
Suppli MP, Rigbolt KTG, Veidal SS, Heebøll S, Eriksen PL, Demant M, Bagger JI, Nielsen JC, Oró D, Thrane SW, et al. Hepatic transcriptome signatures in patients with varying degrees of nonalcoholic fatty liver disease compared with healthy normal-weight individuals. Am J Physiol Gastrointest Liver Physiol. 2019;316:G462–G472. doi: 10.1152/ajpgi.00358.2018
9.
Lefebvre P, Lalloyer F, Baugé E, Pawlak M, Gheeraert C, Dehondt H, Vanhoutte J, Woitrain E, Hennuyer N, Mazuy C, et al. Interspecies NASH disease activity whole-genome profiling identifies a fibrogenic role of PPARα-regulated dermatopontin. JCI Insight. 2017;2:92264. doi: 10.1172/jci.insight.92264
10.
Zhernakova DV, Le TH, Kurilshikov A, Atanasovska B, Bonder MJ, Sanna S, Claringbould A, Võsa U, Deelen P, Franke L, et al; LifeLines cohort study; BIOS consortium. Individual variations in cardiovascular-disease-related protein levels are driven by genetics and gut microbiome. Nat Genet. 2018;50:1524–1532. doi: 10.1038/s41588-018-0224-7
11.
Cirulli ET, White S, Read RW, Elhanan G, Metcalf WJ, Tanudjaja F, Fath DM, Sandoval E, Isaksson M, Schlauch KA, et al. Genome-wide rare variant analysis for thousands of phenotypes in over 70,000 exomes from two cohorts. Nat Commun. 2020;11:542. doi: 10.1038/s41467-020-14288-y
12.
Lek M, Karczewski KJ, Minikel EV, Samocha KE, Banks E, Fennell T, O’Donnell-Luria AH, Ware JS, Hill AJ, Cummings BB, et al; Exome Aggregation Consortium. Analysis of protein-coding genetic variation in 60,706 humans. Nature. 2016;536:285–291. doi: 10.1038/nature19057
13.
Liu DJ, Peloso GM, Yu H, Butterworth AS, Wang X, Mahajan A, Saleheen D, Emdin C, Alam D, Alves AC, et al; Charge Diabetes Working Group; EPIC-InterAct Consortium; EPIC-CVD Consortium; GOLD Consortium; VA Million Veteran Program. Exome-wide association study of plasma lipids in >300,000 individuals. Nat Genet. 2017;49:1758–1766. doi: 10.1038/ng.3977
14.
Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics. 2010;26:2190–2191. doi: 10.1093/bioinformatics/btq340
15.
Futema M, Plagnol V, Li K, Whittall RA, Neil HAW, Seed M, Bertolini S, Calandra S, Descamps OS, Graham CA, et al. Whole exome sequencing of familial hypercholesterolaemia patients negative for LDLR / APOB / PCSK9 mutations. J Med Genet. 2014;51:537–544. doi: 10.1136/jmedgenet-2014-102405
16.
Kraehling JR, Chidlow JH, Rajagopal C, Sugiyama MG, Fowler JW, Lee MY, Zhang X, Ramírez CM, Park EJ, Tao B, et al. Genome-wide RNAi screen reveals ALK1 mediates LDL uptake and transcytosis in endothelial cells. Nat Commun. 2016;7:13516. doi: 10.1038/ncomms13516
17.
Medina MW, Krauss RM. Alternative splicing in the regulation of cholesterol homeostasis. Curr Opin Lipidol. 2013;24:147–152. doi: 10.1097/MOL.0b013e32835cf284
18.
Shao C, Yang B, Wu T, Huang J, Tang P, Zhou Y, Zhou J, Qiu J, Jiang L, Li H, et al. Mechanisms for U2AF to define 3′ splice sites and regulate alternative splicing in the human genome. Nat Struct Mol Biol. 2014;21:997–1005. doi: 10.1038/nsmb.2906
19.
Mayne J, Ooi TC, Tepliakova L, Seebun D, Walker K, Mohottalage D, Ning Z, Abujrad H, Mbikay M, Wassef H, et al. Associations between soluble LDLR and lipoproteins in a white cohort and the effect of PCSK9 loss-of-function. J Clin Endocrinol Metab. 2018;103:3486–3495. doi: 10.1210/jc.2018-00777
20.
Carlson SM, Soulette CM, Yang Z, Elias JE, Brooks AN, Gozani O. RBM25 is a global splicing factor promoting inclusion of alternatively spliced exons and is itself regulated by lysine mono-methylation. J Biol Chem. 2017;292:13381–13390. doi: 10.1074/jbc.M117.784371
21.
Balder JW, de Vries JK, Nolte IM, Lansberg PJ, Kuivenhoven JA, Kamphuisen PW. Lipid and lipoprotein reference values from 133,450 Dutch Lifelines participants: Age- and gender-specific baseline lipid values and percentiles. J Clin Lipidol. 2017;11:1055–1064.e6. doi: 10.1016/j.jacl.2017.05.007
22.
Deschênes M, Chabot B. The emerging role of alternative splicing in senescence and aging. Aging Cell. 2017;16:918–933. doi: 10.1111/acel.12646
23.
Seiler M, Peng S, Agrawal AA, Palacino J, Teng T, Zhu P, Smith PG, Buonamici S, Yu L; Cancer Genome Atlas Research Network. Somatic mutational landscape of splicing factor genes and their functional consequences across 33 cancer types. Cell Rep. 2018;23:282–296.e4. doi: 10.1016/j.celrep.2018.01.088
24.
Rodríguez SA, Grochová D, McKenna T, Borate B, Trivedi NS, Erdos MR, Eriksson M. Global genome splicing analysis reveals an increased number of alternatively spliced genes with aging. Aging Cell. 2016;15:267–278. doi: 10.1111/acel.12433
25.
Miranda MX, van Tits LJ, Lohmann C, Arsiwala T, Winnik S, Tailleux A, Stein S, Gomes AP, Suri V, Ellis JL, et al. The Sirt1 activator SRT3025 provides atheroprotection in Apoe-/- mice by reducing hepatic Pcsk9 secretion and enhancing Ldlr expression. Eur Heart J. 2015;36:51–59. doi: 10.1093/eurheartj/ehu095
26.
Aguet F, Brown AA, Castel SE, Davis JR, He Y, Jo B, Mohammadi P, Park YS, Parsana P, Segrè AV, et al. Genetic effects on gene expression across human tissues. Nature. 2017;550:204–213.
27.
Storey JD, Tibshirani R. Statistical significance for genomewide studies. Proc Natl Acad Sci USA. 2003;100:9440–9445. doi: 10.1073/pnas.1530509100
28.
Havel RJ, Eder HA, Bragdon JH. The distribution and chemical composition of ultracentrifugally separated lipoproteins in human serum. J Clin Invest. 1955;34:1345–1353. doi: 10.1172/JCI103182
29.
Velagapudi S, Yalcinkaya M, Piemontese A, Meier R, Nørrelykke SF, Perisa D, Rzepiela A, Stebler M, Stoma S, Zanoni P, et al. VEGF-A regulates cellular localization of SR-BI as well as transendothelial transport of HDL but not LDL. Arterioscler Thromb Vasc Biol. 2017;37:794–803. doi: 10.1161/ATVBAHA.117.309284
30.
McFARLANE AS. Efficient trace-labelling of proteins with iodine. Nature. 1958;182:53. doi: 10.1038/182053a0
31.
Kamentsky L, Jones TR, Fraser A, Bray MA, Logan DJ, Madden KL, Ljosa V, Rueden C, Eliceiri KW, Carpenter AE. Improved structure, function and compatibility for CellProfiler: modular high-throughput image analysis software. Bioinformatics. 2011;27:1179–1180. doi: 10.1093/bioinformatics/btr095
32.
Parsons BD, Schindler A, Evans DH, Foley E. A direct phenotypic comparison of siRNA pools and multiple individual duplexes in a functional assay. PLoS One. 2009;4:e8471. doi: 10.1371/journal.pone.0008471
33.
Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30:2114–2120. doi: 10.1093/bioinformatics/btu170
34.
Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29:15–21. doi: 10.1093/bioinformatics/bts635
35.
Liao Y, Smyth GK, Shi W. The Subread aligner: fast, accurate and scalable read mapping by seed-and-vote. Nucleic Acids Res. 2013;41:e108. doi: 10.1093/nar/gkt214
36.
Anders S, Reyes A, Huber W. Detecting differential usage of exons from RNA-seq data. Genome Res. 2012;22:2008–2017. doi: 10.1101/gr.133744.111
37.
Fedoseienko A, Wijers M, Wolters JC, Dekker D, Smit M, Huijkman N, Kloosterhuis N, Klug H, Schepers A, van Dijk KW, et al. The COMMD family regulates plasma LDL levels and attenuates atherosclerosis through stabilizing the CCC complex in endosomal LDLR traffcking. Circ Res. 2018;122:1648–1660. doi: 10.1161/CIRCRESAHA.117.312004
38.
MacLean B, Tomazela DM, Shulman N, Chambers M, Finney GL, Frewen B, Kern R, Tabb DL, Liebler DC, MacCoss MJ. Skyline: an open source document editor for creating and analyzing targeted proteomics experiments. Bioinformatics. 2010;26:966–968. doi: 10.1093/bioinformatics/btq054
39.
Nisson PE, Watkins PC, Krizman DB. Isolation of exons from cloned DNA by exon trapping. Curr Protoc Hum Genet. 2001;Chapter 6:Unit 6.1. doi: 10.1002/0471142905.hg0601s03
40.
Bray NL, Pimentel H, Melsted P, Pachter L. Near-optimal probabilistic RNA-seq quantification. Nat Biotechnol. 2016;34:525–527. doi: 10.1038/nbt.3519
41.
Cirulli, ET, Washington, NL. Helix Research. UK Biobank Exome rare variant analysis v1.3 [Internet]. Helix Blog.
42.
Whiffin N, Minikel E, Walsh R, O’Donnell-Luria AH, Karczewski K, Ing AY, Barton PJR, Funke B, Cook SA, MacArthur D, et al. Using high-resolution variant frequencies to empower clinical genome interpretation. Genet Med. 2017;19:1151–1158. doi: 10.1038/gim.2017.26
43.
Birmingham A, Selfors LM, Forster T, Wrobel D, Kennedy CJ, Shanks E, Santoyo-Lopez J, Dunican DJ, Long A, Kelleher D, et al. Statistical methods for analysis of high-throughput RNA interference screens. Nat Methods. 2009;6:569–575. doi: 10.1038/nmeth.1351
44.
König R, Chiang CY, Tu BP, Yan SF, DeJesus PD, Romero A, Bergauer T, Orth A, Krueger U, Zhou Y, et al. A probability-based approach for the analysis of large-scale RNAi screens. Nat Methods. 2007;4:847–849. doi: 10.1038/nmeth1089
45.
Roweis ST, Saul LK. Nonlinear dimensionality reduction by locally linear embedding. Science. 2000;290:2323–2326. doi: 10.1126/science.290.5500.2323
Information & Authors
Information
Published In
Copyright
© 2021 American Heart Association, Inc.
Versions
You are viewing the most recent version of this article.
History
Received: 3 September 2020
Accepted: 19 November 2021
Published online: 23 November 2021
Published in print: 7 January 2022
Keywords
Subjects
Authors
Disclosures
The RNAi screening was performed at the ETH ScopeM facility (R. Meier, M. Stebler, S.F. Norrelykke, A.J. Rzepiela, S. Stoma). As by contract, one-third of the service costs had to be paid by grants of A. von Eckardstein to cover part of the costs for personnel, infrastructure, and maintenance. S.E. Humphries received fees for Advisory Boards of Novartis and Amryt and is the Medical Director of a UCL spin-out company StoreGene that offers to clinicians genetic testing for patients with FH. The other authors report no conflicts.
Sources of Funding
This work was conducted as part of the TransCard project of the seventh Framework Program (FP7) granted by the European Commission, to J. Albert Kuivenhoven, A. Tybjaerg-Hansen, and A. von Eckardstein (number 603091) as well as partially the FP7 RESOLVE project (to J.T. Haas, B. Staels, A. Verhaegen, S. Francque, L. Van Gaal, and A. von Eckardstein) and the European Genomic Institute for Diabetes (EGID, ANR-10-LABX-46 to B. Staels). Additional work by A. von Eckardstein’s team was funded by the Swiss National Science Foundation (31003A-160126, 310030-185109) and the Swiss Systems X program (2014/267 [Medical Research and Development (MRD)] HDL-X). P. Zanoni received funding awards from the Swiss Atherosclerosis Society (Arbeitsgruppe Lipide und Atherosklerose [AGLA] and the DACH Society for Prevention of Cardiovascular Diseases). G. Panteloglou received funding from the University of Zurich (Forschungskredit, grant no. FK-20-037). J. Albert Kuivenhoven is an Established Investigator from the Dutch Heart Foundation (2015T068). J. Albert Kuivenhoven was also supported by GeniusII (CVON2017-2020). The Laboratory for Vascular Translational Science (L.V.T.S.) team is supported by Fondation Maladies Rares, Programme Hospitalier de Recherche Clinique (PHRC) (AOM06024), and the national project CHOPIN (CHolesterol Personalized Innovation), granted by the Agence Nationale de la Recherche (ANR-16-RHUS-0007). Y. Abou Khalil is supported by a grant from Ministère de l’Education Nationale et de la Technologie (France). J.T. Haas was supported by an EMBO Long Term Fellowship (ALTF277-2014). B. Staels is a recipient of an ERC Advanced Grant (no. 694717). Both are also supported by PreciNASH (ANR 16-RHUS-0006). Research at the Antwerp University Hospital was supported by the European Union: FP6 (HEPADIP Contract LSHM-CT-2005-018734). S. Francque has a senior clinical research fellowship from the Fund for Scientific Research (FWO) Flanders (1802154 N). S.E. Humphries received grants RG3008 and PG008/08 from the British Heart Foundation, and the support of the UCLH NIHR BRC. S.E. Humphries directs the UK Children’s FH Register which has been supported by a grant from Pfizer (24052829) given by the International Atherosclerosis Society.
Metrics & Citations
Metrics
Citations
Download Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Select your manager software from the list below and click Download.
- The low-density lipoprotein receptor: Emerging post-transcriptional regulatory mechanisms, Atherosclerosis, 401, (119082), (2025).https://doi.org/10.1016/j.atherosclerosis.2024.119082
- Refinement of a Published Gene‐Physical Activity Interaction Impacting HDL‐Cholesterol: Role of Sex and Lipoprotein Subfractions, Genetic Epidemiology, 49, 1, (2025).https://doi.org/10.1002/gepi.22607
- Association Between Gadoxetic Acid-Enhanced Magnetic Resonance Imaging, Organic Anion Transporters, and Farnesoid X Receptor in Benign Focal Liver Lesions, Drug Metabolism and Disposition, 52, 2, (118-125), (2024).https://doi.org/10.1124/dmd.123.001492
- Variants in LPA are associated with familial hypercholesterolaemia: whole genome sequencing analysis in the 100 000 Genomes Project , European Journal of Preventive Cardiology, (2024).https://doi.org/10.1093/eurjpc/zwae371
- Post-translational regulation of the low-density lipoprotein receptor provides new targets for cholesterol regulation, Biochemical Society Transactions, 52, 1, (431-440), (2024).https://doi.org/10.1042/BST20230918
- The mouse multi-organ proteome from infancy to adulthood, Nature Communications, 15, 1, (2024).https://doi.org/10.1038/s41467-024-50183-6
- RBM25 is required to restrain inflammation via ACLY RNA splicing-dependent metabolism rewiring, Cellular & Molecular Immunology, 21, 11, (1231-1250), (2024).https://doi.org/10.1038/s41423-024-01212-3
- RNF130 Adds Further Complexity to the Regulation of LDL Receptor Activity, Circulation Research, 132, 7, (864-866), (2023)./doi/10.1161/CIRCRESAHA.123.322555
- Metformin attenuates multiple myeloma cell proliferation and encourages apoptosis by suppressing METTL3-mediated m6A methylation of THRAP3, RBM25, and USP4, Cell Cycle, 22, 8, (986-1004), (2023).https://doi.org/10.1080/15384101.2023.2170521
- Altered splicing factor and alternative splicing events in a mouse model of diet- and polychlorinated biphenyl-induced liver disease, Environmental Toxicology and Pharmacology, 103, (104260), (2023).https://doi.org/10.1016/j.etap.2023.104260
- See more
Loading...
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginPurchase Options
Purchase this article to access the full text.
eLetters(0)
eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.
Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.