RNA Sequence Analyses throughout the Course of Mouse Cardiac Laminopathy Identify Differentially Expressed Genes for Cell Cycle Control and Mitochondrial Function.

Lamin A/C (LMNA) gene mutations are a known cause of familial dilated cardiomyopathy, but the precise mechanisms triggering disease progression remain unknown. We hypothesize that analysis of differentially expressed genes (DEGs) throughout the course of Lmna knockout (Lmna-/-)-induced cardiomyopathy may reveal novel Lmna-mediated alterations of signaling pathways leading to dilated cardiomyopathy. Although Lmna was the only DEG down-regulated at 1 week of age, we identified 730 and 1004 DEGs in Lmna-/- mice at 2 weeks and 1 month of age, respectively. At 2 weeks, Lmna-/- mice demonstrated both down- and up-regulation of the key genes involving cell cycle control, mitochondrial dysfunction, and oxidative phosphorylation, as well as down-regulated genes governing DNA damage repair and up-regulated genes involved in oxidative stress response, cell survival, and cardiac hypertrophy. At 1 month, the down-regulated genes included those involved in oxidative phosphorylation, mitochondrial dysfunction, nutrient metabolism, cardiac β-adrenergic signaling, action potential generation, and cell survival. We also found 96 overlapping DEGs at both ages involved in oxidative phosphorylation, mitochondrial function, and calcium signaling. Impaired oxidative phosphorylation was observed at early disease stage, even before the appearance of disease phenotypes, and worsened with disease progression, suggesting its importance in the pathogenesis and progression of LMNA cardiomyopathy. Reduction of oxidative stress might therefore prevent or delay the development from Lmna mutation to LMNA cardiomyopathy.

Introduction

Lamin A/C (LMNA) gene mutations account for approximately 6–8% of known genetic dilated cardiomyopathies, and occur with frequent cardiac conduction system disease[1]. Lamins are type V intermediate filament proteins and major components of the nuclear lamina. They line the inner surface of the inner nuclear membrane, where they interact with chromatin associated proteins, nuclear envelope proteins (including nuclear pore complexes), and transcription factors, thereby regulating important cellular events such as chromatin organization, DNA repair/replication, transcription and cell division[2-5]. There are two types of lamins. Lamins A and C are A-type lamins coded by the Lmna gene, while B-type lamins (B1 and B2) are coded by two different genes (Lmnb1 and Lmnb2, respectively). Unlike B-type lamins, lamins A and C are also found in nucleoplasm and may play a role in regulating the expression of certain genes[5,6]. Much effort has been undertaken to identify signaling molecules implicated in Lmna mutations, and especially in LMNA cardiomyopathy, in order to gain insights on how lamins regulate a wide spectrum of cellular processing leading to cardiomyopathy and heart failure[7-16]. Previously, microarray analysis in a mouse model harboring the Lmna mutation (Lmna) for Emery-Dreifuss muscular dystrophy and cardiomyopathy showed that there was over-activation of mitogen-activated protein kinase 3/1 (ERK1/2) and mechanistic targeting of rapamycin kinase (mTOR) pathways[7-9,17,18]. Therapeutic inhibition of these pathways has shown delayed left ventricular dilation and improved cardiac function in the animal model of LMNA cardiomyopathy[8,10-14,18]. Recent RNA sequencing studies found the activation of FOXO transcription factors (in 2-week-old Lmna−/− mice) and E2F/DNA damage response/TP53 pathway (in 2-week Lmna mice) contribute to the pathogenesis of LMNA cardiomyopathy[15,16].

Compared to microarrays, RNA-Sequencing technology produces discrete, digital sequencing read counts and can quantify expression across a larger dynamic range (>105 versus 103 for arrays) and detect a higher percentage of differentially expressed genes. This is especially true for genes with low expression[19-21]. In our present study, we utilized RNA sequence analysis to investigate gene expression profiling throughout the lifespan of Lmna−/− mice with cardiomyopathy to investigate novel Lmna-mediated alterations of signaling pathways leading to dilated cardiomyopathy. While previous studies focused on specific pathways[8,10-16,18] in the pathogenesis of LMNA cardiomyopathy (caused by specific Lmna mutations or knockout), our study aimed to provide a systematic overview of the signaling pathways and pathophysiological changes that might synergistically contribute to the development of Lmna induced LMNA cardiomyopathy. To the best of our knowledge, this is the first systematic mechanism study covering the entire disease process of LMNA cardiomyopathy in the Lmna−/− mouse model. Rather than clarifying a specific pathogenic signaling pathway, this study reveals key signaling pathways involved in the pathogenesis of LMNA cardiomyopathy and provides guidance for further mechanism studies. In particular, our study is the first to identify the importance of impaired oxidative phosphorylation in the pathogenesis and progression of LMNA cardiomyopathy, suggesting the possibility that a reduction in oxidative stress might prevent or delay the development of LMNA cardiomyopathy in the presence of a Lmna gene mutation.

Results

Homozygous Lmna−/− mice exhibit severe growth retardation and cardiac dysfunction at 1 month of age

Homozygous mice harboring the targeted Lmna mutation, Lmnatm1Stw, were previously shown to have progressive muscular dystrophy and dilated cardiomyopathy[22-24]. We observed a similar disease progression from 1 week to 1 month. But unlike the former study, which showed a significant decrease in mouse weight in 2-week-old Lmna−/− mice[15], we did not identify a significant difference in growth between wild type (WT) and Lmna−/− mice at 1 week and 2 weeks of age (Fig. 1a). However, by 1 month of age Lmna−/− mice exhibited severe growth retardation and cardiac dysfunction (Fig. 1a,b). This cardiac dysfunction was accompanied by increased myocardial fibrosis in Lmna−/− mice at 1 month of age (Fig. 1c). No appreciable differences in myocardial fibrosis were observed between Lmna−/− and WT mice at either 1 week or 2 weeks of age.

Figure 1

Lmna−/− Mice Exhibited Severe Growth Retardation (A), Cardiac Dysfunction (B) and Increased Myocardial Fibrosis (C) at One Month of Age. (a) Significantly lower body weight was found in 1-month-old Lmna−/− mice compared with 1-month-old WT mice (p = 0.0003, n = 7 for each group). There was no statistical difference in body weight between Lmna−/− and WT mice at 1 week or 2 weeks old. (b) 1-month-old Lmna−/− mice demonstrated a significantly less fractional shortening percentage than 1-month-old WT mice (p = 0.0006, n = 7 for each group). (c) Picrosirus Red staining of left ventricular myocardium showed an increased ratio of fibrotic areas to the whole myocardium in 1-month-old Lmna−/− mice compared to 1-month old WT mice (p = 0.032, n = 7).

Early differentially expressed genes (DEGs) and related pathways in Lmna−/− mice

We first compared the gene expression profiles of WT and Lmna−/− mice at 1 week and 2 weeks of age. Although Lmna was the only gene with significantly reduced expression in 1-week-old Lmna−/− mice, we identified 730 DEGs (428 genes up-regulated, 293 genes down-regulated, 9 genes unmapped) in 2-week-old Lmna−/− mice. Based on these genes, 58 related canonical pathways (with p < 0.05) were identified through IPA pathway analysis. Figure 2a and Supplementary Table S1 list the six pathways with the highest p values. Table 1 lists representative key DEGs in the related pathways/functions among 2-week Lmna−/− mice. As shown in Table 1, Lmna−/− mice at 2 weeks of age demonstrated both down- and up-regulation of the key genes involving cell cycle regulation (e.g. chromosomal replication down-regulated, counteractive effects on G1/S and G2/M transition, M phase progression repressed), mitochondrial dysfunction, oxidative phosphorylation, and apoptosis. Meanwhile, Lmna−/− mice showed down-regulation in key genes in DNA damage response/repair pathways as well as up-regulation in genes involved in the oxidative stress response, acute phase response, cell survival/growth, cardiac hypertrophy, glycolysis, and triglyceride hydrolysis. We identified the up-regulation of FOXO3 (listed in Table 1 apoptosis pathway), which was also found in a previous study in 2-week Lmna−/− mice by Marian et al.[15]. However, instead of activation of the pathway in 2-week Lmna mice shown by Marian et al.[16], we identified overall down-regulation of DNA damage response pathway (z score = −1.265). But we did observe the overall up-regulation of TP53 pathway (z score = 0.302) due to up-regulation of TP53 effector (PERP) and other TP53 target genes, which was again similar to the findings from Marian et al. in 2-week Lmna mice[16].

Figure 2

Top Canonical Pathways Based on the Differentially Expressed Genes (DEGs) Between WT and Lmna−/− Mouse Hearts by Ingenuity Pathway Analysis (IPA). (a) Top 6 canonical pathways based on the DEGs between 2-week-old WT and Lmna−/− mouse hearts; (b) Top 6 canonical pathways based on the DEGs between 1-month-old WT and Lmna−/− mouse hearts; (c) Top 6 canonical pathways based on the overlapped DEGs between 2-week-old and 1-month-old Lmna−/− mice. The bar graph shows the percentages of differentially expressed (both down- and up-regulated) genes among the genes in the IPA random dataset involved in the listed pathways. The number listed at the end of each pathway bar represents the number of all the genes in the IPA random dataset for that pathway. The line graph shows the significant differences between the DEGs (in 2-week-old Lmna−/− vs. WT mouse hearts) and the genes in the IPA random dataset involved in the listed pathways. The point in each pathway line indicates the value of −log10 (p-value) for that specific pathway (between 5–9.25 in Fig. a, between 6–34 in Fig. b, between 2–5.5 in Fig. c). This transformation was performed because the p-values were very low (between 10−5–10–9.25 in Fig. a, between 10−6–10−34 in Fig. b, between 10−2–10−5.5 in Fig. c).

Table 1

Examples of key DEGs in related pathways/functions among 2-week Lmna−/− mice.

Pathways/Functions	Gene name	Gene symbol	Up(↑)- or down(↓)-regulated
Cell cycle regulation:
Chromosomal replication ↓	Cell division cycle 6/7	CDC 6/7	↓
	Cyclin dependent kinase 1	CDK1	↓
	Minichromosome maintenance complex component 2–7	MCM 2–7	↓
	DNA polymerase epsilon	POLE	↓
G1/S transition ↓	Cyclin E2	CCNE2	↓
G1/S transition ↑	RB transcriptional corepressor 1	RBL1	↓
	Checkpoint kinase 1	CHK1	↓
G2/M transition ↓	Cyclin dependent kinase inhibitor 1 A	CDKN1A	↑
	CDC28 protein kinase regulatory subunit 2	CKS2	↓
	Cyclin dependent kinase 1	CDK1	↓
	Cyclin B	CCNB	↓
G2/M transition ↑	Checkpoint kinase 1	CHK1	↓
M phase progression ↓	Polo Like Kinase 1	PLK1	↓
	Cyclin dependent kinase 1	CDK1	↓
	Cyclin B	CCNB	↓
	Protein regulator of cytokinesis 1	PRC1	↓
	Kinesin family member 11/23	KIF11/23	↓
Mitochondrial dysfunction/Oxidative phosphorylation	Genes for enzymes in Complex I		↑
	Genes for enzymes in Complex III, IV, V		↓
Apoptosis	Forkhead box O3	FOXO3	↑
	BCL2 like 11	BCL2L11	↑
	BCL2 interacting protein 3	BNIP3	↑
	NF-kappa-B inhibitor alpha	NFKBIA	↑
	TP53 apoptosis effector	PERP	↑
	Period circadian regulator 1	PER1	↑
	Protein phosphatase 1 regulatory subunit 15 A	PPP1R15A	↑
	Baculoviral IAP repeat containing 5	BIRC5	↓
DNA damage response/Repair ↓	Breast cancer type 1 susceptibility protein	BRCA1	↓
	BRCA1 associated RING domain 1	BARD1	↓
	Bloom syndrome RecQ like helicase	BLM	↓
	BRCA1 interacting protein C-terminal helicase 1	BRIP1	↓
	Breast cancer type 2 susceptibility protein	BRCA2	↓
	Fanconi anemia complementation group D2	FANCD2	↓
	Alpha thalassemia/mental retardation syndrome X-linked chromatin remodeler	ATRX	↓
	High mobility group box 2	HMGB2	↓
Oxidative stress response ↑	Nuclear factor erythroid 2 like 2	NRF2	↑
	Actin gamma 1	ACTG1	↑
	Jun pro-oncogene, AP-1 transcription factor Subunit	JUN	↑
	Sequestosome 1	SQSTM1	↑
	Ferritin light chain	FTL	↑
	Four and a half LIM domains 1	FHL1	↑
	Epoxide hydrolase 1	EPHX1	↑
	Glutathione S-transferase mu 5	GSTM5	↑
Acute phase response ↑	Angiotensinogen	AGT	↑
	Protein kinase B beta	AKT2	↑
	Complement component 1 subcomponent R	C1R	↑
	Complement Component 3 and 4	C3/4	↑
	Nuclear factor of interleukin 6	NFIL6	↑
	Ceruloplasmin	Cp	↑
	Serpin family A member 3	SERPINA3	↑
	Serpin family G member 1	SERPING1	↑
	Fos proto-oncogene, AP-1 transcription factor subunit	FOS	↑
	Jun pro-oncogene, AP-1 transcription factor Subunit	JUN	↑
	NF-kappa-B inhibitor alpha	NFKBIA	↑
Cell survival/growth	Protein kinase B beta	AKT2	↑
	Pyruvate dehydrogenase kinase 2	PDK2	↑
	BCL2 like 1	BCL2L1	↑
	Insulin receptor substrate 2	IRS2	↑
	Insulin like growth factor binding protein 3/5/6	IGFBP3/5/6	↑
	Fos proto-oncogene, AP-1 transcription factor subunit	FOS	↑
Cardiac hypertrophy	Adrenoceptor alpha 1 A	ADRA1A	↑
	G protein subunit α	GNA	↑
	Heart and neural crest derivatives expressed 2	HAND2	↑
	Jun pro-oncogene, AP-1 transcription factor Subunit	JUN	↑
	Mitogen-activated protein kinase kinase kinase 6	MAP3K6	↑
	Mitogen-activated protein kinase-activated protein kinase 2	MAPKAPK2	↑
	Ras homolog family member B/J	RHOB/J	↑
Glycolysis	Fructose-bisphosphate A aldolase	ALDOA	↑
	Glucose-6-phosphate isomerase	GPI	↑
	Phosphofructokinase	PFK	↑
	Triosephosphate isomerase 1	TPI1	↑
Triglyceride hydrolysis	Patatin like phospholipase domain-containing protein 2	PNPLA2	↑

Taken together, 2-week-old Lmna−/− mice demonstrated the activation of cell cycle arrest and apoptosis pathways and a failure of DNA repair in response to the targeted Lmna knockout mutation. Although the trend of cell cycle arrest and apoptosis seemed clear, some counteractive activation pathways were seen as well. For example, multiple cell survival factors were up-regulated to keep cells growing. In addition, while mitochondrial dysfunction (down-regulated genes for Complex III, IV, V) might compromise ATP generation and cause oxidative stress, other enzyme-coding genes for Complex I, III, IV, V and glycolysis were up-regulated to overcome ATP shortage, and the NRF2-mediated oxidative stress response pathway was dramatically activated to reduce oxidative damage. These counteractive effects might explain the lack of observed growth retardation at 2 weeks. In the meantime, up-regulation of the genes in cardiac hypertrophy, myofibroblast activation/fibrosis and acute phase response signaling might contribute to the early pathological changes before LMNA cardiomyopathy occurs.

Mitochondrial dysfunction and diminished nutrient metabolism associated with extensive late-stage DEGs in 1-month-old Lmna−/− mice

We identified 1004 DEGs between the gene expression profiles of WT and Lmna−/− mice at 1 month of age, with 699 genes had decreased expression levels while 290 genes had increased expression levels (15 genes unmapped) in Lmna−/− mice. Pathway analysis via IPA identified 66 canonical pathways (p < 0.05) based on these genes. Figure 2b and supplementary Table S2 list the top six pathways. Table 2 lists the examples of key DEGs in the related pathways/functions among 1-month-old Lmna−/− mice. Most impressively, these mice demonstrated extensive down-regulation of key genes associated with mitochondrial dysfunction, oxidative phosphorylation, a wide range of metabolic pathways (tricarboxylic acid cycle, glycolysis, glycogen synthesis, glycogenolysis, amino acid degradation, fatty acid β-oxidation, ketogenesis, and ketolysis), cardiac β-adrenergic signaling, G protein β γ signaling, caveolar-mediated endocytosis, and cardiac voltage-gated channels (Table 2). Such alterations in pathways of metabolism and energy status in Lmna−/− mice may explain their remarkable growth retardation. Down-regulation of multiple voltage-gated channels that have been reported to be related to the generation of arrhythmias[25] - KCNA5, KCND2, KCNE1, SCN4A, SCN4B (listed in Table 2) - might be the underlying basis for the increased risk of arrhythmogenic events in LMNA cardiomyopathy. Meanwhile, we observed down- and up-regulated genes among these mice involved in cell survival signaling, cardiac hypertrophy, calcium signaling and autophagy (Table 2).

Table 2

Examples of key DEGs in related pathways/functions among 1-month Lmna−/− mice.

Pathways/Functions	Gene name	Gene symbol	Up (↑)-/down (↓)-regulated
Mitochondrial dysfunction	Peroxiredoxin 3/5	PRDX 3/5	↑
	Superoxide dismutase 2	SOD2	↑
	Thioredoxin 2	TXN2	↑
	Apoptosis inducing factor mitochondria associated 1	AIFM1	↑
	Cytochrome C1	CYC1	↑
Oxidative phosphorylation	Genes for enzymes in all complex		↓
Cardiac β-adrenergic signaling	G protein subunit gamma	GNG	↑
	cAMP-dependent protein kinase inhibitor alpha	PKIA	↑
	Sarcoplasmic/endoplasmic reticulum Ca2+ ATPase 2	ATP2A2	↑
	Ryanodine receptor 2	RYR2	↑
G protein β γ signaling	G protein subunit gamma	GNG	↑
	G protein-coupled inwardly-rectifying potassium channel	GIRK	↑
Caveolar-mediated endocytosis	Caveolin 1	CAV1	↑
	Filamin B	FLNB	↑
	Integrin subunit alpha 1/6/8/9	ITGA1/6/8/9	↑
	Integrin subunit beta 6	ITGB6	↑
Voltage-gated channels (Action Potential generation)	Potassium voltage-gated channel subfamily A member 5	KCNA5	↑
	Potassium voltage-gated channel subfamily D member 2	KCND2	↑
	Potassium voltage-gated channel subfamily E regulatory subunit 1B	KCNE1B	↑
	Potassium voltage-gated channel modifier subfamily V member 2	KCNV2	↑
	Sodium voltage-gated channel alpha subunit 4	SCN4A	↑
	Sodium voltage-gated channel beta subunit 4	SCN4B	↑
Cell survival signaling	G protein subunit gamma	GNG	↓
	G protein-coupled inwardly-rectifying potassium channel	GIRK	↓
	Vascular endothelial growth factor A/B	VEGFA/B	↓
	Regulated protein tyrosine phosphatase receptor type F	PTPRF	↑
Cardiac hypertrophy	Calcium voltage-gated channel subunit alpha1G/H	CACNA1G/H	↑
	Transforming growth factor beta	TGFβ	↑
	Insulin like growth factor 1 receptor	IGF1R	↑
	Histone deacetylase 6	HDAC6	↑
	Protein phosphatase 3 catalytic subunit beta	PPP3CB	↓
Calcium signaling	Ryanodine receptor 2	RYR2	↓
	Troponin I3	TNNI3	↓
	Tropomyosin 4	TPM4	↓
	Troponin T2	TNNT2	↑
	Calcium/calmodulin dependent protein kinase 1G	CAMK1G	↑
	Calcium/calmodulin dependent protein kinase kinase 2	CAMKK2	↑
Autophagy	Autophagy related 4D cysteine peptidase	ATG4D	↓
	Autophagy related 3	ATG3	↑

Several up-regulated genes in 1-month Lmna−/− mice are associated with sarcomere structure and ERK1/2 pathway. Compared with previous microarray studies by Worman et al. on mouse LMNA cardiomyopathy from Lmna mutation[7,26], we identified up-regulation of a novel gene associated with sarcomere structure, myomesin 2 (MYOM2). MYOM2 is a protein coding gene for M-band in sarcomere that plays an important role in maintaining sarcomere structure[27,28]. In addition, we identified up-regulation of dual specificity phosphatase 4 (DUSP4) and DUSP5 in 1-month Lmna−/− mice, while only DUSP4 was up-regulated in mice with Lmna-induced cardiomyopathy based on the study by Worman et al.[17]. DUSP4 and DUSP5 can be transcriptionally induced by ERK1/2, which suggests the possibility of ERK1/2 pathway activation in LMNA cardiomyopathy progression. In addition, we found some compensatory up-regulated genes in Lmna−/− mice, e.g. AMP-Activated Protein Kinase Subunit Gamma-2 (PRKAG2) was up-regulated by higher cellular AMP/ATP ratio to inhibit key enzymes of ATP consuming pathways and induces ATP generation, and cAMP-dependent protein kinase type I alpha regulatory subunit (PRKAR1A) was up-regulated by cAMP accumulation for ATP generation and cell survival.

Comparison of 2-week and 1-month differentially expressed genes among Lmna−/− mice

While comparing the DEGs between 2-week-old and 1-month-old Lmna−/− mice, we found more extensive gene expression changes at 1 month than at 2 weeks (1004 vs. 730), and DEGs in 1-month-old Lmna−/− mice were mostly down-regulated while those in 2-week-old Lmna−/− mice were mostly up-regulated. In addition, we found 96 DEGs that overlapped at both ages. These DEGs could be important because their expression changes start early (between 1–2 weeks of age), before an identifiable phenotype, and remain into the later life stage of Lmna−/− mice, suggesting they may be related to the pathogenesis and progression of LMNA cardiomyopathy. Among the 96 overlapping DEGs, 64 genes are regulated in the same direction (either both up-regulated or both down-regulated) at the two time points while 32 genes are regulated in opposite directions (30 genes up-regulated at 2 weeks but down-regulated at 1 month and 2 genes down-regulated at 2 weeks but up-regulated at 1 month). Pathway analysis identified 23 canonical pathways (p < 0.05) based on the overlapping DEGs. Figure 2c and supplementary Table S3 listed the top six canonical pathways. Table 3 lists key overlapping DEGs, which were involved in oxidative phosphorylation, mitochondrial function, calcium signaling, G protein β γ signaling, and caveolar-mediated endocytosis. Although expression changes of some listed key genes could cause functional impairment at both time points, compensatory expression changes of other genes could prevent the occurrence of actual functional impairment as in 2-week-old Lmna−/− mice. For example, the coding gene for Complex V enzyme MT-ATP6 was down-regulated in these mice (Table 3), which could affect ATP generation in mitochondria. However, the coding genes for Complex I/IV enzymes NDUFA13, NDUFS7, NDUFV1 and COX6A2 are simultaneously up-regulated (Table 3), producing more substrates for MT-ATP6 and protecting the mice from ATP shortage.

Table 3

Examples of key overlapped DEGs in related pathways/functions.

Pathways/Functions	Gene name	Gene symbol	At 2 weeks	At 1 month
Oxidative phosphorylation/Mitochondrial function	Cytochrome c oxidase subunit 6A2	COX 6A2	↑ (up-regulated)	↓ (down-regulated)
	NADH: ubiquinone oxidoreductase subunit A13	NDUFA13	↑	↓
	NADH: ubiquinone oxidoreductase core subunit S7	NDUFS7	↑	↓
	NADH: ubiquinone oxidoreductase core subunit V1	NDUFV1	↑	↓
	Mitochondrial ATP synthase subunit 6	MT-ATP6	↓	↓
Calcium signaling	Calcium voltage-gated channel subunit alpha 1 H	CACNA1H	↑	↑
	Troponin T2	TNNT2	↑	↑
	Myosin heavy chain 7	MYH7	↑	↑
	A-kinase anchoring protein 5	AKAP5	↓	↓
	Tropomyosin 4	TPM4	↓	↓
	Troponin I3	TNNI3	↑	↓
G protein β γ signaling	G protein subunit gamma 11	GNG11	↓	↓
	Caveolin 1	CAV1	↓	↓
	G protein subunit alpha o1	GNAO1	↑	↑
Caveolar-mediated endocytosis	Caveolin 1	CAV1	↓	↓
	Integrin subunit alpha 1	ITGA1	↓	↓
	Integrin subunit beta 5	ITGB5	↑	↑

Other important overlapping DEGs include those involved in glycogen degradation (acid α-glucosidase, phosphofructokinase - PFKM), glycolysis (PFKM), and lactose degradation (prosaposin). All these were up-regulated in 2-week-old and down-regulated in 1-month-old mice, suggesting more glucose generation and usage in 2-week-old Lmna−/− mice and a dysfunction of glucose metabolism in 1-month-old mice. In addition, heat shock protein family B member 1 (HSPB1), nicotinamide riboside kinase (NMRK2) and dual specificity phosphatase 5 (DUSP5) were up-regulated at both time points. HSPB1 is important for stress resistance and actin organization while NMRK2 codes a key enzyme for NAD+ synthesis from nicotinamide riboside (NAD+ is one of the most important co-enzymes for redox reactions), and DUSP5, as a member of the dual specificity protein phosphatase subfamily, negatively regulates members of the mitogen-activated protein kinase (MAPK) superfamily (ERK1/2, stress-activated protein kinase/c-Jun N-terminal kinase - SAPK/JNK). DUSP1 and DUSP5 were up-regulated in 2-week-old Lmna−/− mice while DUSP4 and DUSP5 were up-regulated in 1-month-old Lmna−/− mice.

To validate our RNA-sequencing results, we performed RT-qPCR and western blot for representative genes associated with the top 6 pathways (for 2 weeks, listed in Fig. 2a) as well as oxidation phosphorylation, mitochondrial function, fatty acid metabolism, sarcomere structure, and cardiac development for 1 month. RT-qPCR analysis confirmed decreased expression of Aurora Kinase A (AURKA), breast cancer type 1 susceptibility protein (BRCA1), cyclin dependent kinase 1 (CDK1), checkpoint kinase 1 (CHEK1), cyclin B1, minichromosome maintenance complex component 5 (MCM5), polo like kinase 1 (PLK1) (for 2 weeks, Fig. 3a), peroxisome proliferator activated receptor alpha (PPARA), carnitine o-acetyltransferase (CRAT), ATP synthase membrane subunit c locus1 (ATP5G1), enoyl-CoA hydratase 1 (ECH1), acyl-CoA dehydrogenase very long chain (ACDVL), purinergic receptor P2Y1 (P2RY1), potassium voltage-gated channel subfamily D member 2 (KCDN2) (for 1 month, Fig. 4a) and increased expression of epoxide hydrolase 1 (EPHX1), NADH: ubiquinone oxidoreductase subunit A13 (NDUFA13), nuclear factor, erythroid 2 like 2 (NRF2), polo like kinase 1 (PLK3), sequestosome 1 (SQSTM1) (for 2 weeks, Fig. 3b), dual specificity phosphatase 4 (DUSP4), lysyl oxidase (LOX), ferritin heavy chain 1 (FHL1), myomesin 2 (MYOM2) and nicotinamide riboside kinase (NMRK2) (for 1 month, Fig. 4b) with statistically significant expression fold change (p < 0.05). Western blot analysis showed decreased protein levels of CDK1, cyclin B1, MCM5 (Fig. 3c), ECH1, PPARA (Fig. 4c) and increased protein levels of NRF2, PLK3 (Fig. 3c), DUSP4, FHL1 (Fig. 4c) in hearts from 2-week (Figs. 3c) and 1-month (Fig. 4c) Lmna−/− mice compared to WT mice. The full-length western blot images are provided in Supplementary Fig. S1 (for Fig. 3c) and S2 (for Fig. 4c).

Figure 3

Validation of Two-week-old Mouse Heart RNA-Sequencing Results by RT-qPCR and Western Blot. RNA and protein samples analyzed were extracted from WT and Lmna−/− mouse heart tissues at 2 weeks of age. (a) RT-qPCR validation of down-regulated genes (AURKA, BRCA1, CDK1, CHEK1, Cyclin B1, MCM5, PLK1) identified by RNA-Sequencing. The housekeeping gene S18 was used as a control gene for both WT and Lmna−/− samples. ∆∆ cycle threshold (Ct) method was used for RT-qPCR analysis. Average WT expression level was normalized to one and Lmna−/− expression levels were calculated over average WT level and showed as “mean ± SEM” on Y axis. The expression levels of all the listed genes were significantly lower in 2-week Lmna−/− mouse hearts than in 2-week WT mouse hearts (p < 0.01). (b) RT-qPCR validation of up-regulated genes (EPHX1, NDUFA13, NRF2, PLK3, SQSTM1) identified by RNA-Sequencing. The same ∆∆Ct method was used for analysis. The relative expression levels of all the listed genes are showed as “mean ± SEM” on Y axis. The expression levels in 2-week-old Lmna−/− mouse hearts were significantly higher than those in 2-week-old WT mouse hearts for all the genes (p < 0.05). (c) Western blot validation of protein expressions of representative down-regulated (CDK1, Cyclin B1 and MCM5) and up-regulated (NRF2 and PLK3) genes. GAPDH was used as a loading control. Three repeated experiments were conducted for each protein with similar results showed in the figure.

Figure 4

Validation of One-month-old Mouse Heart RNA-Sequencing Results by qRT-PCR and Western Blot. Samples (RNAs and proteins) being analyzed were extracted from WT and Lmna−/− mouse heart tissues at 1 month of age. (a) RT-qPCR validation of down-regulated genes (ACDVL, ATP5G1, CRAT, ECH1, KCDN2, P2RY1, PPARA) identified by RNA-Sequencing. The same ∆∆Ct method was used for analysis as described in Fig. 3 legends. The relative expression levels of all the listed genes are showed as “mean ± SEM” on Y axis. The expression levels of all the listed genes among 1-month Lmna−/− mouse hearts were significantly lower than the levels among 1-month WT mouse hearts (p < 0.05). (b) RT-qPCR validation of up-regulated genes (DUSP4, LOX, FHL1, MYOM2, NMRK2) identified by RNA-Sequencing. The same ∆∆Ct method was used for analysis. The relative expression levels of all the listed genes are showed as “mean ± SEM” on Y axis. The expression levels in 1-month Lmna−/− mouse hearts were significantly higher than those in 1-month WT mouse hearts for all the genes (p < 0.05). (c) Western blot validation of protein expressions of representative down-regulated (ECH1 and PPARA) and up-regulated (DUSP4 and FHL1) genes. β-actin (for ECH1 and PPARA) or GAPDH (for DUSP4 and FHL1) was used as a loading control. Three repeated experiments were conducted for each protein with similar results showed in the figure.

Discussion

Our study is the first to utilize RNA sequencing and pathway analysis to set up detailed gene expression and pathway profiles over the course of Lmna−/− induced LMNA cardiomyopathy, which may lead to further mechanistic studies of the pathogenesis and progression of LMNA cardiomyopathy. Based on an average lifespan of 5 weeks for Lmna−/− mice, we selected three different time points that correspond to the early (1 week), middle (2 weeks), and late (1 month) stages of the disease to investigate the detailed molecular mechanisms over time in hearts isolated from WT and Lmna−/− mice. The significantly increased number of genes with expression changes between 1 and 2 weeks (from 1 gene to 730 genes) and between 2 weeks to 1 month of age (from 730 genes to 1004 genes) was associated with the phenotypic observation with significant cardiac dysfunction and fibrosis.

The major findings of this study are the gene expression and related pathophysiological changes at different stages of LMNA cardiomyopathy. We have further identified how these changes evolved over time with disease progression. More importantly, we have identified 96 overlapping DEGs in 2-week-old and 1-month-old Lmna−/− mice, which are mainly involved in oxidative phosphorylation, mitochondrial dysfunction, calcium signaling, and G protein beta gamma signaling. These findings have provided insights on the pathogenesis of LMNA cardiomyopathy due to loss of LMNA function.

We found the DEGs in 2-week-old Lmna−/− mice were mainly involved in cell cycle regulation, DNA damage response, mitochondrial function, oxidative phosphorylation, oxidative stress response, and counteractive apoptosis/survival signaling. Although these mice seemed to fail in repairing their DNA damage, i.e. Lmna mutation (all related DEGs down-regulated) with down-regulated chromosomal replication and a trend of delayed cell cycle progression, we still observed some down-regulated genes meant to promote cell cycle progression. Similarly, up-regulated gene expression for mitochondrial complex I, III, IV and V enzymes might offset the consequences caused by down-regulated genes for complex III, IV, V enzymes for ATP synthesis from mitochondria. While oxidative stress may exist because of impaired oxidative phosphorylation, the oxidative stress response pathway was activated (all related DEGs up-regulated) to reduce oxidative injury. A similar pattern was observed with apoptosis and survival factors. There were up-regulated genes on both sides. These compensatory mechanisms might explain why the 2-week-old mice didn’t show any signs of LMNA cardiomyopathy or growth retardation yet.

For 1-month-old Lmna−/− mice, we observed significant growth and cardiac phenotypes associated with extensive down-regulation of genes involving oxidation phosphorylation, mitochondrial function, almost all nutrient metabolism (e.g. TCA cycle, amino acid degradation, and fatty acid β-oxidation), cardiac β-adrenergic signaling, G protein beta gamma signaling. Overwhelming oxidative stress and ATP deficiency caused by seriously impaired oxidation phosphorylation and mitochondrial dysfunction, together with extensively reduced nutrient metabolism might explain the remarkable growth retardation of Lmna−/− mice at this stage. Up-regulated cardiac β-adrenergic signaling (z score = 1) and dysregulated G protein beta gamma signaling (z score =0) in the heart could also contribute to cardiac dysfunction we observed. Our study provided a unique model that could identify comprehensive and accurate gene expression levels using RNA sequence analysis to further understand signaling pathways involved in the progressive dilated cardiomyopathy process from Lmna mutation.

Our study is the first to report that mitochondrial dysfunction and impaired oxidative phosphorylation can occur in Lmna−/− mice as early as 2 weeks of age, even before an apparent phenotype of LMNA cardiomyopathy. At late disease stage (1 month of age), when the phenotypes are dramatic, these impairments are more prominent. 47.5% genes involving oxidative phosphorylation and 35.2% genes involving mitochondrial dysfunction are down-regulated. Mitochondrial dysfunction has been described in dilated or hypertrophic cardiomyopathies and heart failure but its exact mechanism is unclear[29-32]. Mitochondrial dysfunction further promotes increased reactive oxygen species[29,33]. Mitochondrial dysfunction and increased oxidative stress have been noted in progeria laminopathies[34-37]. A specific gene targeted therapy or drug therapy to reduce oxidative stress has shown improvement of progeria phenotype[34,35]. It was previously shown that there was increased oxidative stress in heart failure and efforts had been made to target the oxidative stress pathway to prevent heart failure progression[38,39]. Our study suggested the possibility that reduction of oxidative stress might prevent or delay the development from Lmna mutation to LMNA cardiomyopathy.

Increased reactive oxidative stress and mitochondrial dysfunction can also affect ERK1/2 as well as mTOR pathways[40,41]. The over-activation of ERK1/2 and mTOR pathways has been well described in LMNA cardiomyopathy[8,9,11,13]. Our RNA sequence analysis shows increased gene expression level of dual specificity phosphatase 1 - DUSP1 (dephosphorylates and inactivates ERK2) and DUSP5 (inactivates ERK1) at 2 weeks of age as well as up-regulated DUSP4 (inactivates ERK1, ERK2 and JNK), DUSP5 and down-regulated DUSP7, DUSP8 (inactivates SAPK/JNK and p38) and DUSP18, suggesting over-activation of ERK1/2 signaling pathway. We also find that mitogen-activated protein kinase kinase kinase 6 - MAP3K6 (upstream kinase of MAPK pathways) and mitogen-activated protein kinase-activated protein kinase 2 - MAPKAPK2 (regulated through direct phosphorylation by p38 MAPK) are up-regulated in 2-week-old Lmna−/− mice. Meanwhile, we observe multiple up-regulated genes involving insulin like growth factor 1 (IGF1) signaling including Akt2 in 2-week-old Lmna−/− mice, which may promote mTOR activation. In addition, we identify a novel coding gene for nicotinamide riboside kinase (NMRK2) that is significantly up-regulated in Lmna−/− mice. NMRK2 has been shown to play a key role in myogenesis especially through its interactions with integrins thereby activating the integrin-mediated signaling pathways[42,43]. ERK1/2 pathway is a well-known downstream effector of integrin-mediated signaling pathway and our finding provides further insights of ERK1/2 signaling mechanisms in LMNA cardiomyopathy.

Besides identifying the importance of oxidative stress in the pathogenesis of LMNA cardiomyopathy, our study is the first to report the upregulation of counteractive factors and signaling pathways for both cell apoptosis and survival in 2-week Lmna−/− mice before the appearance of disease phenotypes. The upregulated pro-apoptotic factors include forkhead box O3 (FOXO3), BCL2 like 11 (BCL2L11, BIM), BCL2 interacting protein 3 (BNIP3), TP53 apoptosis effector (PERP), et al. (Table 1). Some of these factors are targets of FOXO3 and TP53, suggesting the upregulation of FOXO3 and TP53 pathways. Upregulated TP53 pathway (p = 0.00233, z score 0.302) is identified by IPA pathway analysis based on 2-week DEGs. The upregulated survival factors include protein kinase B beta (Akt2), pyruvate dehydrogenase kinase 2 (PDK2), BCL2 like 1 (BCL2L1, BCL-XL), insulin receptor substrate 2 (IRS2), insulin like growth factor binding protein 3/5/6 (IGFBP3/5/6), et al. (Table 1). We observed significantly upregulated IGF-1 pathway (p = 0.00545, z score 2.236), NRF2-mediated oxidative stress response (p = 1.59E-6, z score 3.207) and acute phase response signaling (p = 9.42E-4, z score 3.317), which may work together to promote cell survival and growth. It is reasonable to think these may be compensatory mechanisms under the background of DNA damage (deletion mutation), failure of DNA damage repair and counteractive factors on cell cycle progression in order to keep 2-week-old Lmna−/− mice in temporarily similar overall growth as that of their WT littermates. Interventions which inhibit cell apoptosis or enhance survival pathways may delay the appearance of disease phenotypes and improve the lifespan of Lmna−/− mice.

Activated FOXO1/3 and their target genes have been demonstrated in 2-week Lmna−/− mice by Auguste G et al.[15]. They also found suppression of FOXO1/3 activation by shRNA down-regulated the target gene transcription, improved apoptosis and prolonged survival of 2-week mice by 2-fold[15]. Conformed to this study, our study also finds the up-regulation of FOXO3 and its target genes (BCL2 L11/ BIM, BCL6, BNIP3 for apoptosis/autophagy, P21CIP1, PLK3, down-regulated PLK1/4 and cyclin B1/2 for cell cycle arrest). We don’t detect the up-regulation of the target genes for DNA repair and reactive oxygen species detoxification. Although FOXO1 gene expression is not found up-regulated in our analysis, we do identify up-regulated expression of its target genes as well. No up-regulation of FOXO4 and its target genes are found in 2-week Lmna−/− mice. But we observe mostly down-regulated target genes for FOXO 1/3/4 in 1-month Lmna−/− mice although no gene expression level changes for these three FOXO factors.

FOXO3/1 proteins can be activated by DNA damage or oxidative stress[44]. But when the intensity of these stimuli increases beyond a certain threshold, they activate pro-apoptotic genes and genes for cell cycle arrest. Meanwhile, Akt is a major upstream regulator of FOXO factors[44-46]. Our study shows up-regulation of Akt2 (phosphorylates and inactivates FOXO) through IGF1 pathway, suggesting over-activation of FOXO3/1. In addition, over-activation of ERK1/2 (phosphorylates and inactivates FOXO3) in these mice as discussed above supports the existence of over-activation of FOXO3[47]. Combined with downstream and upstream changes, our study provides some potential mechanisms on FOXO3/1 activation in 2-week Lmna−/− mice.

The activation of E2F transcription factor/DNA damage response/TP53 pathway has been found in a 2-week Lmna mouse model[16]. In our study, E2F is predicted to be inhibited (p = 1.30E-8, z score −2.798) in 2-week Lmna−/− mice because all of its 10 target genes (among DEGs) are inhibited. For DNA damage response, all the related DEGs in 2-week Lmna−/− mice are down-regulated. However, our study shows TP53 is predicted to be activated (p = 5.88E-24, z score 3.977) with 46 target genes up-regulated in 2-week Lmna−/− mice. The data differences between our study and the previous study may be due to the difference of mouse models and different lifespan: Lmna mice at 30 days vs. Lmna−/− mice at 42 days. At 2 weeks, these two kinds of mice may be at different disease stages, and their data may not be comparable.

In summary, our study identified failure of DNA repair, dysregulated cell cycle progression and mitochondrial dysfunction/impaired oxidative phosphorylation associated with loss of LMNA, thereby providing mechanistic insights to LMNA cardiomyopathy and heart failure. Compared with previous gene profiling using microarray in Lmna-induced cardiomyopathy (at 10 weeks of age)[7,26] and RNA sequencing study in 2-week Lmna−/− mice[15] and Lmna mice[16], our study not only confirmed the previously known genes and pathways related to LMNA cardiomyopathy but also identified novel genes and pathways that could contribute to the pathogenesis and progression of LMNA cardiomyopathy. It will be interesting to further investigate whether we can apply the same analysis to peripheral samples for its better practicality and feasibility. This application will provide us with detailed information at genomic level which can guide us to better understand this devastating disease in human subjects and develop advanced treatments including gene therapy.

Methods

Animals

Animal use and the study protocol were approved by the Institutional Animal Care and Use Committee at the Cleveland Clinic in accordance with the United States Public Health Service Policy on the Humane Care and Use of Animals, and the NIH Guide for the Care and Use of Laboratory Animals. Heterogeneous C57BL/6 mice harboring the targeted Lmna mutation, Lmnatm1Stw, were purchased from the Jackson Laboratory (stock# 009125). Heterozygotes and littermate wild type (WT) C57BL/6 mice were inbred and LMNA gene mutation was confirmed by genotyping[22,23,48]. 8 WT and 8 Lmna−/− mice were used for each time point at 1 week, 2 weeks and 1 month of age. Hearts were isolated from WT and homozygous Lmna−/− mice at each time point. Hearts were rinsed with cold saline to remove blood, snap frozen in liquid nitrogen, and stored at −80 °C for future use or fixed in 10% formalin for histology examination.

Transthoracic echocardiography

Mice at 1 month of age were anesthetized with isoflurane inhalation and placed on a table. Echocardiography was performed using a GE Vivid 7 ultrasound with a 14 MHz transducer (GE Healthcare, Chicago, IL). Left ventricular shortening fraction was measured in 2D mode.

Histological examination

Formalin-fixed, paraffin-embedded left ventricular (LV) sections were prepared and stained with Picrosirus Red. Five to six regions were randomly selected from each section and fibrosis areas were analyzed by an independent researcher, viewing in the whole heart longitudinal and cross sections with Image Pro Plus v7.0 software (Media Cybernetics Inc, Rockville, MD). Fibrosis area was presented as ratio of fibrotic areas to the whole myocardium areas[49].

Protein extraction and Western blot analysis

Each mouse heart was placed in an ice-cold lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Tergitol solution NP-40, 25 mM β-glycerolphosphate, 10% glycerol, proteinase and phosphatase inhibitors, and homogenized immediately using an Omni TH tissue homogenizer (Omni International, Kennesaw, Georgia). Samples were then centrifuged at 2,000 rpm for 10 min at 4 °C and supernatant was stored at −80 °C for Western blot analysis. 50–60 µg of total proteins were resolved in SDS-PAGE, and transferred to nitrocellulose membranes. The membranes were blocked with 5% skim dry milk in 1x TBST for 1 hour and blotted with primary antibodies including anti-CDK1 (1:500, #77055, Cell Signaling), anti-Cyclin B1 (1:250, #4138, Cell Signaling), anti-MCM5 (1:1000, ab75975, Abcam), anti-NRF2 (1:250, #12721, Cell Signaling), anti-PLK3 (1:250, #4896, Cell Signaling), anti-ECH1 (1:500, PA5-30012, Thermo Fisher Scientific), anti-PPARA (1:500, ab126285, Abcam), anti-DUSP4 (1:100, sc-17821, Santa Cruz), anti-FHL1 (1:500, ab133661, Abcam), anti-GAPDH (1:2000, PA1-987, Thermo Fisher Scientific) and anti-β-actin (1:100, sc-81178, Santa Cruz) in 5% skim dry milk in 1x TBST at 4 °C overnight except anti-GAPDH was incubated at room temperature for 1 hour. LI-COR IRDye 800CW goat anti-mouse IgG (925–32210), goat anti-rabbit IgG (925–32211) or donkey anti-goat IgG (925–32214) were used as secondary antibodies at 1:5000 dilution for 1 hour at room temperature. The membranes were scanned using an Odyssey infrared imager (LI-COR, Lincoln, Nebraska)[50].

RNA isolation and RNA sequence analysis

The hearts were harvested at three time points: 1 week, 2 weeks and 1 month of age, as postnatal growth of the homozygous mice is still not retarded at one week old of age but is severely retarded at 1 month of age. The entire heart tissue was used for RNA extraction to avoid variance caused by sampling different part of the heart tissue. Three hearts from WT and Lmna−/− at each time point (total 18 hearts) were used for this experiment. Total RNA was extracted and purified using RNeasy Fibrous Tissue Mini Kit (QIAGEN, Germantown, MD). RNA concentrations were measured using NanoDrop Spectrophotometers. Quality of the RNA samples was analyzed using the Agilent Bioanalyzer and all samples submitted for sequencing had a RNA Integrity number (RIN) > 8. Truseq Stranded Total RNA -RiboZero library preparation and RNA sequencing were performed by the Cleveland Clinic Lerner Research Institute Genomic Core. RNA sequencing was performed using the Illumina SBS v2 chemistry with 2 × 100 bp pair-end reads. The RNA-Sequencing data were aligned using STAR (v2.5.2b) program with default parameters[51]. Mapped read pairs were assigned to genes by collapsing all transcripts into a single gene model and then counting the number of reads that fully overlap the resulting exons. Reads that mapped to multiple locations were only counted once and those mapping to ambiguous regions were excluded. Reads uniquely mapped were considered for further analysis. DEseq. 2 (v3.8), a program using count-based matrices to identify differentially expressed genes (DEGs), was used to determine gene signatures[51]. To improve the reliability and accuracy of differential expression analysis, only genes with raw counts >5 in all individually sequenced samples were examined and compared between WT vs. Lmna−/− at 1 week of age, at 2 weeks of age and at 1 month of age. For differentially expressed genes, INGENUITY Pathway Analyses (IPA, Qiagen, Hilden Germany) were performed, including enrichment analysis for canonical pathway, disease and biological function, toxicology function.

Validation of the RNA-sequencing data

The most changed genes in the interesting signaling pathways were further confirmed using RT-qPCR and Western blot. RT-qPCR was performed using the TaqMan protocol in StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA). The TaqMan gene expression assays for the following genes were purchased from Applied Biosystems: AURKA (Mm01248177_m1), BRCA1 (Mm00515386_m1), CDK1 (Mm00772472_m1), CHEK1 (Mm01176757_m1), Cyclin B1 (Mm00838401_g1), EPHX1 (Mm00468752_m1), MCM5 (Mm01243769_m1), NDUFA13 (Mm00445751_m1), NRF2 (Mm00477784_m1), PLK1 (Mm00440924_g1), PLK3 (Mm00457348_m1), SQSTM1 (Mm00448091_m1), ECH1 (Mm00469322_m1), PPARA (Mm00440939_m1), ATP5G1 (Mm02601566_g1), ACADVL (Mm00444293_m1), CRAT (Mm00483985_m1), P2RY1 (Mm02619947_s1), KCND2 (Mm01161732_m1), LOX (Mm00495386_m1), FHL1 (Mm04204611_g1), MYOM2 (Mm00500665_m1), LMNA (Mm00497783_m1), DUSP4 (Mm00723761_m1), and NMRK2 (Mm01172899_g1). The housekeeping gene RPS18 (Mm02601777_g1) was used as control gene. Gene fold change calculation was determined by ∆∆Ct method over WT controls.

Statistical analysis

Data are expressed as the mean ± standard deviation (SD). Graphpad (Prism Software) was used for statistical analysis. Statistical significance was determined by 2-tailed Student’s t-test with a value of p < 0.05 considered significant.

Supplementary information.