RNA splicing in the heart: Changing parts and performance - Chapter 7: Hypoxia induces alternative splicing changes in cardiomyocytes

(1)

UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

RNA splicing in the heart

Changing parts and performance

van den Hoogenhof, M.M.G.

Publication date

2018

Document Version

Other version

License

Other

Link to publication

Citation for published version (APA):

van den Hoogenhof, M. M. G. (2018). RNA splicing in the heart: Changing parts and

performance.

General rights

It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s)

and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open

content license (like Creative Commons).

Disclaimer/Complaints regulations

If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please

let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material

inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter

to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You

will be contacted as soon as possible.

(2)

516079-L-bw-hoogenhof 516079-L-bw-hoogenhof 516079-L-bw-hoogenhof 516079-L-bw-hoogenhof Processed on: 9-2-2018 Processed on: 9-2-2018 Processed on: 9-2-2018

Processed on: 9-2-2018 PDF page: 173PDF page: 173PDF page: 173PDF page: 173

changing parts and performance

RNA splicing in the heart

M.M.G. van den Hoogenhof

7

HYPOXIA INDUCES ALTERNATIVE

SPLICING CHANGES IN CARDIOMYOCYTES

Maarten M.G. van den Hoogenhof Simona Aufiero

Hanneke W.M. van Deutekom Yigal M. Pinto Esther E. Creemers

(3)

174

Abstract

Alternative splicing, a mechanism to generate multiple transcripts from a single gene, is increasingly being recognized as an important regulator of (heart) disease. The importance of specific induced alternative splicing events was recently illustrated by mechanistic insights into the hypoxia-dependent switch in ketohexokinase (Khk) isoforms. Khk is the central fructose-metabolizing enzyme critical for fructose metabolism in the heart, and a splicing-induced switch of Khk-A to Khk-C isoform is necessary for the transition from fatty acid metabolism to glycolysis in the failing heart. Preventing this Khk isoform switch protects against pathological cardiac growth. In the ischemic heart, decreased oxygen levels may lead to overall changes in alternative splicing, but to our knowledge, there are no studies available that interrogate hypoxia-induced alternative splicing on a genome-wide level. To investigate the extent at which hypoxia leads to alternative splicing changes in cardiomyocytes, we cultured neonatal rat cardiomyocytes (NRCM) in hypoxic chambers for 24 hours and analyzed gene expression levels and differential alternative splicing events using RNA-sequencing. We show that hypoxia induces alternative splicing changes in cardiomyocytes, but our analysis is hampered by limitations in RNA-sequencing analysis. Specifically, the incomplete annotation of alternative splice isoforms in the rat genome prevented the quantification of complete transcriptomic changes. Nevertheless, we did identify splicing changes in important cardiac genes, such as Titin (Ttn), calcium/calmodulin dependent kinase II delta (CamkIIδ), and LIM domain binding protein 3 (Ldb3). In conclusion, hypoxia may lead to alternative protein isoforms and may affect gene expression levels and thereby contribute to the development and progression of heart disease, especially in the setting of ischemic heart disease.

(4)

175

7 Introduction

Alternative splicing is a process that allows the production of multiple mature mRNA transcripts from a single gene, thereby increasing genomic complexity tremendously, since ~95% of all human genes can be alternatively spliced1_{. Alternative splicing is increasingly being recognized as an important layer of}

post-transcriptional gene expression, is highly regulated, can be altered in disease, and can be induced by various stimuli2_{. Hypoxia, which is defined as insufficient supply of oxygen to cells or tissue, is a state}

that is seen in multiple types of heart disease, such as hypertrophic cardiomyopathy or myocardial infarction, and could contribute substantially to the disease mechanism. Interestingly, it has recently become clear that hypoxia can also contribute to the development or progression of heart disease by inducing alternative splicing events in genes such as Ketohexokinase (Khk), X-box binding protein 1 (Xbp1), and Bcl-2/adenovirus E1B 19-kDa interacting protein 3 (Bnip3)3-5_{. Khk is the central}

fructose-metabolizing enzyme, and exists in two isoforms; Khk-A and Khk-C3_{. In the healthy situation, when}

the heart primarily relies on fatty acid metabolism, it uses mostly Khk-A for metabolizing fatty acids. A failing heart, however, relies more on glycolysis, and the switch to Khk-C is both necessary and sufficient to switch towards more glycolysis. The isoform switch in Khk is induced by Splicing Factor 3b subunit 1 (SF3B1), which is driven by the activity of Hypoxia inducible factor 1 alpha (HIF1a). Interestingly, cardiac specific knockout mice of SF3B1 or Khk both prevent the switch in Khk isoform and thereby the metabolic switch in the heart, which protects against pathological cardiac growth. It is intriguing that a single hypoxia-inducible alternative splicing event can have these profound effects on the heart. Another hypoxia-inducible alternative splicing event is the exclusion of exon 3 in the Bnip3 gene5_{. The}

expression of Bnip3 is induced by hypoxia, and provokes mitochondrial perturbations and cell death55_.

However, the exclusion of exon 3 in Bnip3 leads to a truncated protein, which acts as a dominant

negative protein by inhibition of the function of full length Bnip3, and thereby prolongs cellular survival

after hypoxic stress. In that sense, hypoxia can induce both protective and detrimental stress responses.

A third example of hypoxia-induced alternative splicing is that of Xbp1. Xbp1 is a transcription factor

that is normally rendered inactive. Interestingly, alternative splicing of the Xbp1 mRNA is responsible for

the activation of Xbp1. In normal conditions, a 26-basepair intron is retained in the mature Xbp1 mRNA

transcript, but upon hypoxic stress this intron is spliced out, resulting in a frameshift that alters the

open reading frame of Xbp1 from a 261 aa protein with unknown function to a 376 aa protein, which is known as the active Xbp1 transcription factor4_{. Xbp1 is involved in the unfolded protein response, and}

activates stress-responsive genes by binding to their regulatory elements6_{. Hypoxia-induced activation}

of Xbp1 seems cardioprotective, for example by increasing GRP94 and GRP78 expression, both of which are protective to cell death, at least in other cardiac cell types such as H9c2 cells7, 8_.

The importance of alternative splicing in the healthy and diseased heart has been studied extensively in the past decade2, 9_{, and even though examples of hypoxia-induced alternative splicing events exist,}

it is currently unknown to what extent hypoxia induces alternative splicing changes in the different cell types of the heart. We hypothesized that hypoxia is an important driver of alternative splicing in the cardiomyocyte, and used an in vitro hypoxia model to investigate to which extent hypoxia

(5)

176

leads to alternative splicing changes on a genome-wide scale. We found that hypoxia induces many alternative splicing events in neonatal rat cardiomyocytes, but the incomplete annotation of alternative splicing isoforms in the rat genome prevented the quantification of complete transcriptomic changes. Nevertheless, we demonstrated that several pivotal cardiac genes, such as Ttn, Ldb3, and CamkIId are alternatively spliced upon hypoxic stress. These results imply that hypoxia-induced alternative splicing changes may affect expression of protein isoform or gene expression levels and thereby contribute to the development and progression of heart disease, especially in the setting of ischemic heart disease.

(6)

177

7 Methods

NRCM isolation and hypoxia treatment

Neonatal rat cardiomyocytes (NRCM) were isolated as described previously10_{. In short, 1 x 10}6

NRCMs were plated on fibronectin-coated 6-well plates, serum-starved for 24 hours, and subjected to normoxia or hypoxia for an additional 24 hours. Hypoxia was established using hypoxic chambers (MIC-101, Billups-Rothenberg Inc.) which were flushed with 100% N₂ for 10 minutes to remove oxygen present in the chamber. NRCM were cultured in maintenance medium (DMEM 41965 : DMEM M199 (4:1), supplemented with antibiotics). Glucose-deprived NRCM were cultured in maintenance medium without glucose (DMEM 11966 : DMEM M199 (4:1), supplemented with antibiotics).

RNA isolation and (q)RT-PCR

RNA from cells was isolated using TRIreagent (Sigma-Aldrich) according to the manufacturers protocol. 1 µg of total RNA was subsequently treated with DNAse I (Invitrogen) and used for cDNA generation using SuperScript II (Invitrogen). End-point RT-PCR was performed with Hot Fire Taq Polymerase (Solis Biodyne). qRT-PCR was performed on a Lightcycler 480 system II (Roche) using SYBR Green (Roche). Primer sequences can be found in Supplemental Table 1.

RNA-sequencing and bioinformatical analysis

Control and hypoxic samples were selected from 3 different experiments. RNA quality was assessed using an Agilent Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and all samples had a RIN value of >9. RNA sequencing libraries were prepared with the TruSeq Stranded Total RNA Library Preparation

of >9. RNA sequencing libraries were prepared with the TruSeq Stranded Total RNA Library Preparation

Kit (Illumina, San Diego, USA) and sequenced on a HiSeq 2000 instrument (Illumina, San Diego, USA).

Sequencing depth was approximately 60 million reads per sample with read lengths of 100 bp.

Base-calling was performed using the bcl2fastq 2.0 Conversion Software from Illumina. Quality control of

calling was performed using the bcl2fastq 2.0 Conversion Software from Illumina. Quality control of

fastq files was performed using FASTQC (http://www.bioinformatics.bbsrc.ac.uk/projects/fastqc/).

Reads passing quality control were then aligned using TopHat 1.411_{to the ensembl release Rnor_5.0.}_{to the ensembl release Rnor_5.0.}

Differential gene expression was analyzed using DEseq12_{, alternative splicing analysis was done by}_{, alternative splicing analysis was done by}

calculating the Percentage Spliced In (PSI)-index using scripts from Schafer et al13_.

Gene ontology enrichment

Gene ontology enrichment analysis was done using the online tool DAVID version 6.714_{. Analyses were}

(7)

178

Results

Establishment of the in vitro hypoxia model

To investigate the extent at which hypoxia induces alternative splicing changes, we isolated NRCM and subjected them to normoxia or hypoxia for 24 hours, and performed RNA-sequencing on these cells (Figure 1A). We isolated RNA from normoxic and hypoxic cardiomycytes, and randomly selected 3 control samples, and 4 hypoxic samples from 3 different experiments. We examined the expression of known hypoxia-inducible genes GAPDH and VEGFa15, 16_{, and found that both genes were upregulated}

in the hypoxic NRCM (Figure 1B-C). We also validated two known hypoxia inducible splicing events, namely the exclusion of a 26 base-pair intron that is normally retained in Xbp1 and exon 3 in Bnip34, 5_{, and found that both splicing events were induced in our hypoxic model (Figure 1D). In addition}

to depriving cardiomyocytes of oxygen, we also deprived a subset of cardiomyocytes of glucose, by culturing in glucose-free medium. In the heart, hypoxia is caused by an insufficient supply of blood to cells or tissue, for example after myocardial infarction, or by insufficient perfusion of cells or tissue, for example in hypertrophic cardiomyopathy. The fact that in both conditions blood supply is inadequate, means that it is likely that hypoxic cardiomyocytes are also devoid of, at least to some extent, glucose and /or nutrients. Interestingly, differential splicing events that were present in hypoxic cardiomyocytes, were exacerbated in hypoxic cardiomyocytes that were also deprived of glucose (Figure 1D). However, to make sure that the differential splicing events were caused by hypoxia, and not by the glucose-deprivation, we chose to use hypoxic NRCM in the presence of glucose to study hypoxia-inducible alternative splicing changes by RNA sequencing.

Figure 1. Establishment of in vitro hypoxia model. A. Experimental set-up of the hypoxia experiment. NRCM,

neonatal rat cardiomyocytes. B-C. expression of hypoxic markers GAPDH and VEGFa. FPKM, Fragments Per Kilobase Million. D. RT-PCR of Xbp1 and Bnip3. Norm, normoxia. Hypo, hypoxic.

(8)

179

7 Hypoxia-induced and metabolic genes are upregulated and cell-cycle related genes are

downregulated in hypoxic NRCM

RNA-sequencing of 4 hypoxic and 3 normoxic NRCM samples revealed a total number of 2160 genes to be differentially expressed in the hypoxic NRCM, of which 1178 genes were downregulated, and 982 genes were upregulated (adjusted p-value < 0.05) (Table 1). The log2 fold changes ranged from -3.2 for the downregulated genes, to 4.9 for the upregulated genes in hypoxia. When using a more stringent cut-off of fold change > 2, 889 genes were downregulated, and 781 genes were upregulated (Table 1). The 10 most up- and downregulated genes are depicted in Table 2.

We next used the online tool DAVID to look for gene ontology enrichment in the differentially regulated genes in hypoxic NRCM, and found that the upregulated genes mainly belonged to two categories; hypoxia-inducible genes, or metabolic genes (Table 3). The downregulated genes mostly belonged to cell-cycle related processes (Table 4). The upregulation of hypoxia-inducible and metabolic genes, and the downregulation of cell-cycle related genes, are both known responses to hypoxia 16-19_{, again}

indicating that our model is suitable to test for hypoxia-inducible alternative splicing changes. Since we were interested in which genes might be involved in hypoxia-induced alternative splicing, we next looked for regulation of genes involved in, or related to RNA splicing in hypoxic cardiomyocytes, and found 21 splicing-related genes to be at least fold upregulated and 23 genes to be at least 1.4-fold downregulated (Table 5).

Table 1. Differential gene expression in hypoxic NRCM.

Differentially regulated genes in hypoxia Number of genes

Significantly downregulated 1178

Significantly downregulated with FC > 2 889

Significantly upregulated 982

Significantly upregulated with FC > 2 781

FC, Fold Change. An adjusted p-value of p < 0.05 was considered significant.

Table 2. 10 most upregulated and downregulated genes in hypoxic NRCM.

Upregulated genes Downregulated genes

Kcnh7, Car12, Sncb, Gipr, Pnliprp2, Mt1m, Zap70, Ndufa4l2, Mt2A, Srcin1

LOC100359539, Fbxl22, Pbk, Pi15, Diras2, Il12a, Olr59, Npas4, Cytl1, Car3

(9)

180

Table 3. Gene ontology enrichment analysis of upregulated genes in hypoxic NRCM.

GO term Process Fold enrichment p-value Bonferroni

GO:0001666 response to hypoxia 3.47 6.23E-08 1.61E-04

GO:0070482 response to oxygen levels 3.24 2.44E-07 6.31E-04

GO:0006006 glucose metabolic process 3.47 5.73E-06 0.015

GO:0006007 glucose catabolic process 5.34 7.56E-06 0.019

GO:0019318 hexose metabolic process 3.09 1.02E-05 0.026

GO:0046164 alcohol catabolic process 4.70 1.14E-05 0.029

GO:0019320 hexose catabolic process 5.08 1.28E-05 0.033

GO:0046365 monosaccharide catabolic process 5.08 1.28E-05 0.033 GO:0044275 cellular carbohydrate catabolic process 4.60 1.44E-05 0.037

GO:0006096 glycolysis 6.20 1.61E-05 0.041

Table 4. Gene ontology enrichment analysis of downregulated genes in hypoxic NRCM.

GO term Process Fold enrichment p-value Bonferroni

GO:0022402 cell cycle process 3.33 6.14E-19 1.48E-15

GO:0007049 cell cycle 2.89 1.57E-17 3.76E-14

GO:0022403 cell cycle phase 3.94 3.91E-17 9.41E-14

GO:0000279 M phase 4.40 9.45E-16 2.40E-12

GO:0000278 mitotic cell cycle 3.50 7.27E-15 1.73E-11

GO:0000087 M phase of mitotic cell cycle 4.67 5.21E-13 1.25E-09

GO:0048285 organelle fission 4.57 9.62E-13 2.31E-09

GO:0007067 mitosis 4.71 9.94E-13 2.39E-09

GO:0000280 nuclear division 4.71 9.94E-13 2.39E-09

GO:0007059 chromosome segregation 6.04 3.19E-10 7.67E-07

GO:0006260 DNA replication 3.85 1.15E-09 2.76E-06

GO:0000819 sister chromatid segregation 8.45 3.89E-09 9.36E-06 GO:0000070 mitotic sister chromatid segregation 8.45 3.89E-09 9.36E-06

GO:0051301 cell division 2.88 3.51E-06 0.008

GO:0006259 DNA metabolic process 2.13 5.26E-06 0.013

GO:0007076 mitotic chromosome condensation 10.01 1.84E-05 0.043

Table 5. Differentially regulated RNA-splicing related genes in hypoxic NRCM.

>40% Upregulated >40% Downregulated

Srrm4, Trpt1, Srsf12, Syncrip, Nol3, Pik3r1, Cir1, Wt1, Ahnak, Npm1, Clk1, Rbfox1, Crnkl1, Tsen34, Bud31, Prpf18, Snrnp35, Jmjd6, Pabpc1, Zfp259, Zrsr2

Slc39a5, Srsf10, Rbm15b, Gemin7, Prx, Gemin7, Sf3a1, Ncbp1, Hnrnph1, Pcbp4, Ppil3, Nova1, Tra2b, Rbm24, Lsm2, Lsm6, Mbnl3, LOC100360750, Srsf1, Rbm38, Lsm4, Snrpa, Magohb

(10)

181

7

Figure 2. Selected cardiac genes are alternatively spliced in hypoxic NRCM. A. RT-PCR of Ttn. B. RT-PCR of CamkIIδ. A. RT-PCR of Ttn. B. RT-PCR of CamkIIδ.

C. RT-PCR of Ldb3. HPRT was used as a loading control. Pre-mRNA schematics with exons numbers are presented on

the right. Exons that are subjected to alternative splicing are indicated in light blue.

Pivotal cardiac genes are alternatively spliced in hypoxic cardiomyocytes

We next analyzed alternative splicing changes in hypoxic cardiomyocytes by calculating the ‘Percentage Spliced In’ (PSI) of all exons, and comparing them to their PSI in normoxic cardiomyocytes. A difference in PSI of an exon between normoxic and hypoxic cardiomyocytes represents an alternative splicing event. Considering a difference in PSI of 10% (deltaPSI of 0.1), we found a total number of 2087 events in 1210 genes. When considering a deltaPSI of 0.2, we observed a total number of 417 events in 278 genes (Table 6). The 10 most dramatically changed alternative splicing events can be found in Table 7. It must be noted however, that due to the incomplete annotation of the rat transcriptome, we were unable to examine the full extent of hypoxia-induced alternative splicing events.

(11)

182

To validate some of the splicing events that we observed in the RNA-seq, we performed RT-PCR in pivotal cardiac genes such as Ttn, CamkIIδ, and Ldb3, of which it is known that changes in alternative splicing can affect the heart20_{. Ttn is a giant sarcomeric protein, that acts as a molecular spring in the}

sarcomere. The compliance of Ttn determines the passive stiffness of the cardiomyocyte, and this can primarily be modulated by alternative splicing and post-translational modifications of Ttn21, 22_{. In hypoxic}

NRCM, we observed a decrease in the inclusion of exons of the PEVK-region of Ttn, a region that is subjected to extensive alternative splicing. Exclusion of these exons leads to a less compliant and more stiff Ttn (Figure 2A).

CamkIIδ is a multimeric enzyme with a plethora of roles in cardiac physiology and pathophysiology23_.

It exists in multiple isoforms in heart, with distinct subcellular localizations and temporal expression patterns. The most known isoforms of CamkIIδ are CamkIIδ-A, CamkIIδ-B, CamkIIδ-C, and to a lesser extent CamkIIδ-9. CamkIIδ-A is localized at the intercalated disc and t-tubules, is expressed during the fetal period, and its expression is lost shortly after birth24_{. In the adult heart, the most expressed}

isoforms are CamkIIδ-B and CamkIIδ-C23_{. In hypoxic NRCM, we observed a loss of the CamkIIδ-A splice}

isoform (Figure 2B), suggesting there is less CamKIIδ availbable in the T-tubules for phosphorylation of key ion channels24_.

Ldb3, also known as Cypher/ZASP, is a sarcomeric protein that is also known to undergo alternative splicing25_{. Ldb3 exists as long and short isoforms, which are regulated by alternative splicing of the}

3’ end of the transcript. In addition, there are a cardiac and skeletal muscle splice isoform, which are determined by the inclusion of either exon 4, or exon 5-7. The most prominent isoform in NRCM is the cardiac isoform, which includes exon 4, but NRCM also express a small amount of the skeletal muscle isoform, which excludes exon 4, but includes exons 5-7. In hypoxic NRCM, we observed loss of the skeletal muscle isoform (Figure 2C). In addition, we observed a novel Ldb3 splice isoform, which excludes all exons in this variable domain, meaning that exons 4-7 are all spliced out (Figure 2C).

Table 6. Differential splicing events in hypoxic NRCM. PSI, Percentage Spliced In.

Alternative splicing events in hypoxia Number of events Number of genes

Difference in PSI >10% 2087 1210

Difference in PSI >20% 417 278

Table 7. 10 most differentially spliced genes in hypoxic NRCM. Genes

(12)

183

7 Discussion

Here, we established an in vitro model to investigate hypoxia-inducible alternative splicing changes in cardiomyocytes. Culturing NRCM in hypoxic conditions for 24 hours led to induction of known responses to hypoxia, as indicated by increased expression of the hypoxia-markers GAPDH and VEGFa, and by upregulation of hypoxia-response pathways and downregulation of cell cycle related genes in hypoxic NRCM. In addition, hypoxia induced splicing changes in a vast number of genes, including the pivotal cardiac genes Ttn, CamkIIδ, and Ldb3. Unfortunately, due to the incomplete annotation of the rat transcriptome, we were, for now, not able to ascertain the exact extent at which hypoxia induces alternative splicing changes in NRCM.

In a pathophysiological sense, hypoxia is the result of insufficient blood perfusion. Increased wall thickness, for example in hypertrophic cardiomyopathy, can lead to insufficient perfusion, and myocardial infarction leads to insufficient blood supply to cells or tissue. Therefore, cells that are hypoxic are likely also devoid, at least to some extent, of glucose and/or nutrients. In that sense, the phenomenon we observed that glucose deprivation exacerbates hypoxia-induced alternative splicing changes is interesting, since it could mean that hypoxia and glucose deprivation lead to a similar cellular stress response. If and how the response to hypoxia and glucose deprivation are similar and/or interconnected is yet unknown.

The differential splicing events in Ttn and CamkIIδ, which are likewise exacerbated after glucose deprivation, are also seen in the maturation of cardiomyocytes (i.e. the transition from fetal to adult cardiomyocytes)20, 23_{. The exclusion of exons in the PEVK-region in Ttn is part of the switch from the more}_{. The exclusion of exons in the PEVK-region in Ttn is part of the switch from the more}

compliant fetal N2BA-isoforms towards the stiffer N2B isoforms that are mostly expressed in the adult

heart. Similarly, the loss of CamkIIδ-A in cardiomyocytes is normally seen in the early postnatal stage, since

adult hearts mostly express CamkIIδ-B and CamkIIδ-C. In that regard, the hypoxia treatment seems to push

the NRCM towards a somewhat more mature phenotype, but whether this is a general effect of hypoxia

on alternative splicing, or that this is restricted to a selected set of genes, remains to be determined.

The exact effect of hypoxia on cellular proliferation is not yet entirely clear. While there are reports

showing that hypoxia inhibits cell cycling16, 18_{in murine embryonic fibroblasts and endothelial cells,}_{in murine embryonic fibroblasts and endothelial cells,}

and that HIF-1a is necessary for cell cycle arrest19_{, others have shown that hypoxia is in fact a driver}

of cellular proliferation, most notably in cardiomyocytes26-28_{. In this respect, it has been shown by}

Hif-1a fate mapping that proliferating cardiomyocytes are relatively hypoxic26_{. Moreover, the lack of}

proliferative capacity of postnatal cardiomyocytes has been attributed to a switch towards an oxygen-rich environment, which induces cell-cycle arrest through reactive oxygen species (ROS)-dependent DNA damage27_{. Postponing the switch towards an oxygen-rich environment by exposing neonates}

to a hypoxic environment delays cardiomyocyte cell-cycle arrest, and increases cardiomyocyte proliferation27_{. In that regard, it seems that hypoxia can indeed drive proliferation of cardiomyocytes,}

both at the fetal and early postnatal stage, and in the adult heart. In this study, however, we found a general downregulation of cell cycle genes in NRCM in response to hypoxia. It is conceivable that

(13)

184

the duration and severity of hypoxia determine whether hypoxic cardiomyocytes are able enter the cell cycle. Therefore, it would be interesting to test this hypothesis by culturing NRCM at different O₂ tensions, and examine their proliferative capacity.

Current RNA-sequencing analysis relies heavily on, and is therefore hampered by, transcriptome annotation of the species under investigation. Raw RNA-sequencing reads (represented in a fastq-file) are first mapped to a reference genome, which results in a ‘BAM (Binary Alignment Map)-file’ that provides sequence and positional information, but lacks any contextual information (i.e. whether the read aligns to an annotated gene/exon/intron). Next, a GTF (Gene transfer format)-file, which contains which contains gene structure (transcriptome) information, is used to extract RNA-seq reads from a BAM file that match to a genomic region defined in the GTF file. These RNA-seq reads are then used for further analysis such as gene expression and alternative splicing analysis. However, when reads align to a genomic region that is not annotated in the GTF-file (i.e. in an incompletely annotated transcriptome), these reads and their information are not used for further analysis, and are therefore lost. In this study, only ~60-70% of reads could be mapped to the reference genome (for reference, in a typical RNA-sequencing in humans >90% of reads are mapped) and an even smaller proportion of these reads mapped to genomic regions that where annotated in the GTF-file, meaning that over one-third of all reads were not used for gene expression and alternative splicing analysis.

In a general sense, even though RNA-sequencing is a genome-wide and unbiased technique to examine changes in RNA expression, the analysis is, unfortunately, not. The analysis is based on a priori knowledge of the transcriptome (the GTF-file), meaning that unknown genes or transcripts will not be analyzed.

Future perspectives

There are two ways to solve the problem with RNA-sequencing analysis on incomplete transcriptomes; either by completing the transcriptome annotation, or by using the unaligned (to the GTF-file) reads and use a de novo (or genome-guided) transcriptome assembly as reference in the GTF-file. The latter can be accomplished by using tools such as Cufflinks29_{or Trinity}30_{. Both will lead to a more complete}

understanding of the extent of hypoxia-induced alternative splicing changes, and will open the door to, more comprehensively, investigate how hypoxia induces these alternative splicing changes. When the extent of alternative splicing in response to hypoxia is fully mapped, it will be interesting to analyze putative RNA-binding protein binding sites within or flanking differentially spliced exons. This can be accomplished by using online motif-based sequence analysis tools (for instance Meme suite31_or

rMAPS32_{) or by performing a de novo motif enrichment analysis. This approach may point towards one}

or more hypoxia-regulated RNA-binding proteins or splicing factors. Finding the responsible splicing factor(s) can also be done by overlaying the expression of splicing factors with the differential splicing events, and look for correlations between the two. This approach has been used successfully for SF3B1 and its effect on Khk splicing, and could be used to identify pathological hypoxia responsive splicing factors3_{. Lastly, it will be interesting to investigate splicing changes in ischemic human hearts to examine}

which hypoxia-induced alternative splicing changes contribute to disease progression. Eventually, this will lead to a more comprehensive understanding of detrimental processes in the ischemic heart, and can open up new avenues to investigate novel therapeutic strategies.

(14)

185

7 References

1. Wang ET, Sandberg R, Luo S, Khrebtukova I, Zhang L, Mayr C, Kingsmore SF, Schroth GP and Burge CB. Alternative isoform regulation in human tissue transcriptomes. Nature. 2008;456:470-6.

2. van den Hoogenhof MM, Pinto YM and Creemers EE. RNA Splicing: Regulation and Dysregulation in the Heart. Circ Res. 2016;118:454-68.

3. Mirtschink P, Krishnan J, Grimm F, Sarre A, Horl M, Kayikci M, Fankhauser N, Christinat Y, Cortijo C, Feehan O, Vukolic A, Sossalla S, Stehr SN, Ule J, Zamboni N, Pedrazzini T and Krek W. HIF-driven SF3B1 induces KHK-C to enforce fructolysis and heart disease. Nature. 2015;522:444-449.

4. Thuerauf DJ, Marcinko M, Gude N, Rubio M, Sussman MA and Glembotski CC. Activation of the unfolded protein response in infarcted mouse heart and hypoxic cultured cardiac myocytes. Circ Res. 2006;99:275-82.

5. Gang H, Hai Y, Dhingra R, Gordon JW, Yurkova N, Aviv Y, Li H, Aguilar F, Marshall A, Leygue E and Kirshenbaum LA. A novel hypoxia-inducible spliced variant of mitochondrial death gene Bnip3 promotes survival of ventricular myocytes. Circ Res. 2011;108:1084-92.

6. Yoshida H, Matsui T, Yamamoto A, Okada T and Mori K. XBP1 mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor. Cell. 2001;107:881-91.

7. Vitadello M, Penzo D, Petronilli V, Michieli G, Gomirato S, Menabo R, Di Lisa F and Gorza L. Overexpression of the stress protein Grp94 reduces cardiomyocyte necrosis due to calcium overload and simulated ischemia. FASEB J. 2003;17:923-5.

8. Zhang PL, Lun M, Teng J, Huang J, Blasick TM, Yin L, Herrera GA and Cheung JY. Preinduced molecular chaperones in the endoplasmic reticulum protect cardiomyocytes from lethal injury. Ann Clin Lab Sci. 2004;34:449-57.

9. Lara-Pezzi E, Gomez-Salinero J, Gatto A and Garcia-Pavia P. The alternative heart: impact of alternative splicing in heart disease. J Cardiovasc Transl Res. 2013;6:945-55.

10. Medzikovic L, Schumacher CA, Verkerk AO, van Deel ED, Wolswinkel R, van der Made I, Bleeker N, Cakici D, van

den Hoogenhof MM, Meggouh F, Creemers EE, Remme CA, Baartscheer A, de Winter RJ, de Vries CJ, Arkenbout EK and de Waard V. Orphan nuclear receptor Nur77 affects cardiomyocyte calcium homeostasis and adverse

EK and de Waard V. Orphan nuclear receptor Nur77 affects cardiomyocyte calcium homeostasis and adverse

cardiac remodelling. Sci Rep. 2015;5:15404.

11. Trapnell C, Pachter L and Salzberg SL. TopHat: discovering splice junctions with RNA-Seq. BioinformaticsBioinformatics.

2009;25:1105-11.

12. Anders S and Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010;11:R106.. 2010;11:R106.

13. Schafer S, Miao K, Benson CC, Heinig M, Cook SA and Hubner N. Alternative Splicing Signatures in RNA-seq Schafer S, Miao K, Benson CC, Heinig M, Cook SA and Hubner N. Alternative Splicing Signatures in RNA-seq

Data: Percent Spliced in (PSI). Curr Protoc Hum Genet. 2015;87:11 16 1-14.

14. Huang da W, Sherman BT and Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID Huang da W, Sherman BT and Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID

bioinformatics resources. Nat Protoc. 2009;4:44-57.

15. Zhong H and Simons JW. Direct comparison of GAPDH, beta-actin, cyclophilin, and 28S rRNA as internal

standards for quantifying RNA levels under hypoxia. Biochem Biophys Res Commun. 1999;259:523-6.

16. Weigand JE, Boeckel JN, Gellert P and Dimmeler S. Hypoxia-induced alternative splicing in endothelial cells.

PLoS One. 2012;7:e42697.

17. Bristow RG and Hill RP. Hypoxia and metabolism. Hypoxia, DNA repair and genetic instability. Nat Rev Cancer.

2008;8:180-92.

18. Gardner LB, Li Q, Park MS, Flanagan WM, Semenza GL and Dang CV. Hypoxia inhibits G1/S transition through

regulation of p27 expression. J Biol Chem. 2001;276:7919-26.

19. Goda N, Ryan HE, Khadivi B, McNulty W, Rickert RC and Johnson RS. Hypoxia-inducible factor 1alpha is essential

for cell cycle arrest during hypoxia. Mol Cell Biol. 2003;23:359-69.

20. Weeland CJ, van den Hoogenhof MM, Beqqali A and Creemers EE. Insights into alternative splicing of sarcomeric

(15)

186

21. Guo W and Sun M. RBM20, a potential target for treatment of cardiomyopathy via titin isoform switching.

Biophys Rev. 2017.

22. Hamdani N, Herwig M and Linke WA. Tampering with springs: phosphorylation of titin affecting the mechanical

function of cardiomyocytes. Biophys Rev. 2017;9:225-237.

23. Gray CB and Heller Brown J. CaMKIIdelta subtypes: localization and function. Front Pharmacol. 2014;5:15. 24. Xu X, Yang D, Ding JH, Wang W, Chu PH, Dalton ND, Wang HY, Bermingham JR, Jr., Ye Z, Liu F, Rosenfeld MG,

Manley JL, Ross J, Jr., Chen J, Xiao RP, Cheng H and Fu XD. ASF/SF2-regulated CaMKIIdelta alternative splicing temporally reprograms excitation-contraction coupling in cardiac muscle. Cell. 2005;120:59-72.

25. Huang C, Zhou Q, Liang P, Hollander MS, Sheikh F, Li X, Greaser M, Shelton GD, Evans S and Chen J.

Characterization and in vivo functional analysis of splice variants of cypher. J Biol Chem. 2003;278:7360-5.

26. Kimura W, Xiao F, Canseco DC, Muralidhar S, Thet S, Zhang HM, Abderrahman Y, Chen R, Garcia JA, Shelton JM,

Richardson JA, Ashour AM, Asaithamby A, Liang H, Xing C, Lu Z, Zhang CC and Sadek HA. Hypoxia fate mapping identifies cycling cardiomyocytes in the adult heart. Nature. 2015;523:226-30.

27. Puente BN, Kimura W, Muralidhar SA, Moon J, Amatruda JF, Phelps KL, Grinsfelder D, Rothermel BA, Chen R,

Garcia JA, Santos CX, Thet S, Mori E, Kinter MT, Rindler PM, Zacchigna S, Mukherjee S, Chen DJ, Mahmoud AI, Giacca M, Rabinovitch PS, Aroumougame A, Shah AM, Szweda LI and Sadek HA. The oxygen-rich postnatal environment induces cardiomyocyte cell-cycle arrest through DNA damage response. Cell. 2014;157:565-79.

28. Guimaraes-Camboa N, Stowe J, Aneas I, Sakabe N, Cattaneo P, Henderson L, Kilberg MS, Johnson RS, Chen J,

McCulloch AD, Nobrega MA, Evans SM and Zambon AC. HIF1alpha Represses Cell Stress Pathways to Allow Proliferation of Hypoxic Fetal Cardiomyocytes. Dev Cell. 2015;33:507-21.

29. Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ and Pachter L.

Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol. 2010;28:511-5.

30. Grabherr MG, Haas BJ, Yassour M, Levin JZ, Thompson DA, Amit I, Adiconis X, Fan L, Raychowdhury R, Zeng

Q, Chen Z, Mauceli E, Hacohen N, Gnirke A, Rhind N, di Palma F, Birren BW, Nusbaum C, Lindblad-Toh K, Friedman N and Regev A. Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat Biotechnol. 2011;29:644-52.

31. Bailey TL, Boden M, Buske FA, Frith M, Grant CE, Clementi L, Ren J, Li WW and Noble WS. MEME SUITE: tools

for motif discovery and searching. Nucleic Acids Res. 2009;37:W202-8.

32. Park JW, Jung S, Rouchka EC, Tseng YT and Xing Y. rMAPS: RNA map analysis and plotting server for alternative

(16)

187

7

Supplemental Table 1. Primer sequences.

Gene Fwd primer Rev primer

Bnip3 GCTGGATGCGCAGCATGAAT CGACTTGACCAATCCCATATCC

Xbp1 ACGAGAGAAAACTCATGGGC ACAGGGTCCAACTTGTCCAG

Ttn GAGCCGTATGAGGAACCGTA CAGGAGCAGGTTTCTTTGGA

CamkIIδ AAGGGTGCCATCTTGACAAC TCGAAGTCCCCATTGTTGAT

Ldb3 TCCAAGCGTCCTATCCCCATC TGTATTCTGTCCCGGTCATCTG