Cisd1 synergizes with Cisd2 to modulate protein processing by maintaining mitochondrial and ER homeostasis

Mice

Cisd1 KO mice were generated by Transgenic mouse models core facility (A4) of MOST and their genotype confirmed by Southern blotting and PCR. The Cisd1 KO mice was generated in R1 ES cells and then backcrossed to C57BL/6 background. Cisd1 was completely deleted in the Cisd1 KO mice. Cisd2 KO mice were generated as previously described [12]. Cisd1 +/− female mouse was mated with Cisd2 +/− male mouse to generate Cisd1 and Cisd2 double knockout (Cisd1&2 DKO) mice. Wild-type (WT) mice in this study were the littermates of the Cisd1&2 DKO mice. All mouse lines have congenic C57BL/6 background and were maintained in a specific pathogen–free facility with a 12/12 h light/dark cycle at 20°C to 22°C at the Laboratory Animal Center, National Yang Ming Chiao Tung University. Mice were weighed weekly starting at the age of 2 weeks. At the time of sampling, the mice were euthanized by carbon dioxide inhalation. The animal protocols followed local animal ethics regulations and were approved by the Institutional Animal Care and Use Committee of National Yang-Ming University (Approval No. 1021218).

Histological analysis

Skeletal muscle samples, soleus, gastrocnemius and femoris, were collected separately and fixed with 10% formalin buffered with phosphate. They then underwent tissue processing and were embedded in paraffin. Finally, Haemotoxylin and Eosin (H&E) staining of tissue sections (3–4 μm) were carried out using standard protocols.

Transmission electron microscopy (TEM)

The muscle tissue samples from the soleus and gastrocnemius were fixed separately in a mixture of 1.5% glutaraldehyde and 1.5% paraformaldehyde in 0.1 M cacodylate buffer at pH 7.3, and then post-fixed in 1% OsO₄ and 1.5% potassium hexacyanoferrate. Next the samples were rinsed in cacodylate, and 0.2 M sodium maleate buffers (pH 6.0) followed by block-stained with 1% uranyl acetate. After dehydration, the skeletal muscle samples were embedded in Epon (14120, EMS) and sectioned for TEM as described previously [25].

RNA sequencing and analysis

Total RNA was isolated from muscle tissue samples using TRIzol Reagent (Thermo Fisher Scientific, Waltham, MA, USA). The quality of total RNA was tested using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA); samples with a RNA Integrity Number (RIN) of greater than 8 were subjected to RNA sequencing. The RNA sequencing was performed on a NextSeq 550 platform by National Yang-Ming University VYM Genome Research Center. Counts were estimated by the rsem R package using raw fastq files; this filtered out genes that did not have an appropriate count, which was defined as having more than 150 counts detected in at least half of the samples. The PCA plot was generated by MetaboAnalyst (https://www.metaboanalyst.ca) using the filtered counts. Next, the differentially expressed genes (DEGs) were pinpointed using the DESeq2 R package; these had an FDR <0.1 (p < 0.005). The workstream is presented in Supplementary Figure 1. The differentially expressed genes (DEGs) calculated from the RPKM (reads per kilobase of exon model per million reads) in RNA-Seq data were annotated using the IPA approach (Ingenuity Systems^®, http://www.ingenuity.com/) and Go enrichment analysis (http://pantherdb.org/) [26].

Statistical analysis

The data of survival rate was presented as mean value and calculated using Log-rank Mantal-Cox tests (Prism 8.0). The data of other experiments was presented as mean ± SD. Comparison between two groups was carried out using the two-tailed Student t-test. P < 0.05 was considered significant when analyzing statistical differences among groups. For body weight, we evaluated the data using the Shapiro-Wilk normality test, which is more suitable for small sample sizes (N < 50). The results indicated that samples in some groups did not follow a normal distribution (p < 0.05 using the Shapiro-Wilk normality test). Therefore, we re-analyzed the body weight data using the Kruskal-Wallis test with Dunn’s multiple comparisons test, and the statistical results are presented in Supplementary Table 1. For muscle fiber size, the sample size was too small for the Shapiro-Wilk normality test. Consequently, we re-analyzed the data using the Kruskal-Wallis test with Dunn’s multiple comparisons test to identify any statistical differences.

Cisd1 and Cisd2 double knockout (DKO) mice had more severe phenotypes

Our previous report indicated that Cisd2 KO mice showed growth retardation, weakened skeletal muscles and a short lifespan [12, 18]. Cisd1 and Cisd2 belong to the same protein family and also have the identical domains, the transmembrane domain and the CDGSH iron sulfur domain [27]; therefore, whether the ablation of both Cisd1 and Cisd2 has a synergistic effect on skeletal muscle defects is an open question. We generated the Cisd1 KO mice as shown in Figure 1A. Genotype analysis was carried out by Southern blot and PCR (Figure 1B, 1C), and the Cisd1 protein was found to be completely deleted in the Cisd1 KO mice (Figure 1D). The survival rate of Cisd2 KO mice was lower than the WT and Cisd1 KO mice for both genders; however, the survival rate of DKO mice was lower than the three other groups (Figure 2A, 2B). Cisd1&2 DKO and Cisd2 KO mice showed growth retardation for both genders and showed early aging as shown in the photographs shown (Figure 2C–2F). This was obviously from 5 weeks old. Notably, Cisd1 deletion led to growth retardation in female but not in male mice, which is also shown in the photographs (Figure 2E, 2F). The probability value (p-value) across the four groups of mice at different ages are presented in Supplementary Table 1. Mild pathological defects were identified in the muscle sections from Cisd1 KO mice. Some muscle fibers exhibited slight hypertrophy, and a few showed losses of homogeneity of the sarcoplasm (Figure 3A, 3B). Partial muscle degeneration and pathological defects, such as fiber shrinking, vacuolated cytoplasm, and distortion of muscle fibers, were observed in Cisd2 KO mice (Figure 3A, 3B). More severe muscle defects were noted in muscle sections of DKO mice at 5 weeks old, including loss of homogeneity in the muscle cytoplasm, distortion of fiber boundaries, and altered and/or variable fiber sizes (Figure 3A, 3B). An increased number of larger muscle fibers were found in Cisd1 KO and DKO mice, particularly in male mice (Figure 3C). Additionally, in males, larger muscle fibers (>1000 μm²) averaged 36.7% in the Cisd1 KO and 28.4% in the DKO mice, compared to 7.5% in the Cisd2 KO and 2.6% in the WT mice. In females, larger muscle fibers (>1000 μm²) averaged 23.0% in the Cisd1 KO and 24.8% in the DKO mice, compared to 12% in the Cisd2 KO and 3.8% in the WT mice (Figure 3D). We also quantified the number of muscle fibers with central nuclei. The ratios of muscles with central nuclei in Cisd1 KO, Cisd2 KO, and DKO mice were comparable to those in WT mice, as shown in Supplementary Figure 2. Thus, it was clear that the genetic defects of Cisd1 and Cisd2 had synergistic effects and led to more severe phenotypes in skeletal muscle; however, muscle degeneration in our KO mice is unlikely to be caused by intense cycles of degeneration and regeneration. Remarkably, the body weight of the Cisd1 KO female mice was significantly lower than WT female mice, but this phenotype was not observed in the Cisd1 male, indicating that the mild muscle defects in the Cisd1 KO male mice might not seem to contribute to weight loss.

Mitochondrial defects, myofibril degeneration, ER stress and Golgi degeneration showed increased presence in Cisd1&2 DKO mice

To examine the cellular defects in skeletal muscles of these genetically defect mice, TEM was performed in order to observe the ultrastructure of soleus and gastrocnemius at 5 weeks old. The Cisd1 KO soleus showed mild mitochondrial defects (MD) and myofibril degeneration (MyoD); however, the Cisd2 KO soleus and Cisd1&2 DKO soleus had additional phenotypes, necrosis, and ER stress, besides MD and MyoD (Figure 4A–4D, 4A’–4D’ and Supplementary Figure 3). The Cisd1 KO and Cisd2 KO gastrocnemius showed MD, necrosis, and ER stress, while the Cisd1&2 DKO gastrocnemius showed more severe defects, including MD, ER stress, necrosis and MyoD (5A–5D, 5A’–5D’ and Supplementary Figure 4). In addition, nuclear envelope breakdown and Golgi degeneration were also found in the Cisd2 KO muscles, and these were more severe in Cisd1&2 DKO muscles (Figures 4C, 4D and 5C, 5D). Dysfunction of cellular organelles, such as the ER and Golgi apparatus, led to the errors in protein translation, modification, and transports, eventually resulting in cell death. As expected, the deletion of both Cisd1 and Cisd2 had the synergistic effects on various process and was able to induce premature aged phenotype as well as cell death.

The changes in RNA expression patterns in Cisd1&2 DKO mice

To understand the molecular mechanisms that are involved in the synergistic pathological process present in Cisd1&2 DKO mice, RNA sequencing analysis were performed on skeletal muscles with Cisd1 and/or Cisd2 gene deficiency. The PCA analysis showed four distinct clusters, WT, Cisd1 KO, Cisd2 KO and Cisd1&2 DKO. Cisd1 KO samples were clustered together and close to the WT samples. Cisd2 KO samples were also clustered together and further away from the WT samples. Finally, the Cisd1&2 DKO samples were also clustered and at the greatest distance from the WT samples (Figure 6A). The number of DEGs between Cisd2 KO and WT groups were 175 genes and between Cisd1&2 DKO and WT groups there were 657 genes; additionally, 155 genes were found in both sets of data (Figure 6B). The heatmaps of gene expression in both Cisd2 KO and Cisd1&2 DKO muscle (155 genes), in Cisd2 KO muscle alone (20 genes) and in the Cisd1&2 DKO muscle (502 genes) revealed there were significant changes of gene in expression across the different mouse groups (Supplementary Figure 5A–5C). Totals of 155 genes and 502 genes were separately mapped using IPA software (Figure 6C, 6D). In the top scoring canonical pathways produced by the IPA analysis, four pathways were identified, namely the unfolded protein response, tRNA charging, protein ubiquitination and EIF2 signaling; the genes involved in these pathways are presented in a heatmap (Figure 7A). The functional enrichment analysis results identified a number of significant Gene Ontology (GO) terms in Cisd1&2 DKO and Cisd2 KO muscles (Supplementary Figure 5D, 5E). Most of these GO terms were highly related to protein processing, including the initiation of translation, translational processes, modification, transport, and protein degradation. Except for protein process and EIF2 signaling, several critical pathways were significantly changed in Cisd1&2 DKO muscle sample and these functions are related to inflammation and immune response, redox reactions, and metabolism (Figure 7B). Finally, alterations in signaling and biological process that induce programmed cell death, such as ferroptosis (Figure 7C), as well as the non-regulated cell death, and necrosis were pinpointed (Figures 4 and 5). Solute Carrier Family 7 Member 11 (SLC7A11) is a critical regulator of ferroptosis and a key component of the cystine-glutamate antiporter. SLC7A11 was upregulated in Cisd2 KO and Cisd1&2 DKO mice. Additionally, hepcidin antimicrobial peptide (Hamp) and glutathione peroxidase 4 (Gpx4), which are closely related to ferroptosis, would seem to be upregulated in Cisd1&2 DKO mice.

To sum up, the ablation of Cisd1 and Cisd2 causes mitochondrial dysfunction, ROS generation and ER stress, which creates defects in protein translation and modifies the ER, as well as affecting protein transport from ER to Golgi apparatus (Figure 8). As mentioned earlier, the Cisd1 deletion mouse results in a milder phenotype in skeletal muscle; however, when a mouse lacks both Cisd1 and Cisd2 expression, the synergistic effects exacerbate the phenotypic consequence in muscle.

Defects affecting organelles

During the process of protein synthesis, including translation, modification, folding, transportation, and degradation, our previous reports have provided evidence that Cisd2 deletion contributes to ER stress and disrupts calcium homeostasis. Indeed, in this study, Cisd2 KO and Cisd1&2 DKO resulted in upregulated X-box binding protein 1 (Xbp1) and activating transcription factor 6 (ATF6) expression, both of which are the critical molecules in ER stress [28, 29]. In Cisd2 KO and Cisd1&2 DKO muscles, a group of genes involved in molecular chaperoning were upregulated, including DnaJ heat shock protein family (Hsp40) (Dnaj), Heat shock protein (Hsp), beta-subunit of prolyl 4-hydroxylase (p4hb), Calreticulin (Calr), and Calnexin (Canx), which are responsible for suppressing aggregation and shuttling of misfolded proteins for degradation [30–32]. Calr and Canx act as adaptors to recruit other ER chaperones to assist in different aspects [33]. If protein folding fails, the terminal misfolded proteins are transported for degradation via the ER associated protein degradation (ERAD) pathway. Several related genes to the above are also upregulated in Cisd2 KO and Cisd1&2 DKO muscles, that in the ubiquitin-associated mechanisms of protein degradation [34, 35]. In this situation, cells that are unable to produce functional proteins and initiate compensatory effects to start translation. As a result of this, most amino acid-tRNA synthetases were upregulated, and eIF2 signaling was also activated. Moreover, Cisd1&Cisd2 DKO mice had severe mitochondrial defects in their skeletal muscles. Several genes participating in muscle energy homeostasis and mitochondrial metabolism were found to be upregulated in the double knockout mice, such as Forkhead box class O 1 (Foxo1), Forkhead box class O 3 (Foxo3), Serine hydroxymethyltransferase 2 (Shmt2), Protein kinase AMP-activated non-catalytic subunit gamma 1 (Prkag1), and Protein tyrosine phosphatase mitochondrial 1 (Ptpmt1) [36–40]. Recent studies have indicated that Cisd expression levels are highly associated with mitophagy regulation [41, 42]. In Drosophila and in a human cell line, loss of Cisd (majorly Cisd1) interrupts the process of mitophagy and damaged mitochondria then accumulate in cells, finally triggering cell death. This may be one possibility for explaining why loss of Cisd1 exacerbates the phenotypes of Cisd2 KO mice. Both ER stress and mitochondrial damage directly or indirectly reduce MAMs stability, and all of the above will contribute the process of cell death.

Inflammation and oxidative stress

The alterations in several critical signaling pathways and mechanisms were only observed in the Cisd1&Cisd2 DKO muscles. Most of them are related to microenvironmental changes, such as inflammation, redox regulation (antioxidant capacity) and metabolic changes. Dipeptidyl peptidase 9 (Dpp9) was highly upregulated in Cisd1&Cisd2 DKO muscles. Dpp9/8 inhibition would activate the inflammasome sensor Nlrp1b resulting in procaspase-1-induced cell death [43]. Dpp9 acts as a multifunctional inflammasome regulator and is able to induce a lytic form of cell death, pyroptosis, in monocytes and macrophages [44]. Interestedly, MAMs are important to inflammasome activation, including the NOD-like receptor protein 3 (NLRP3)-mediated inflammasome [45]. Several genes involved in redox signaling show changes in Cisd1&Cisd2 DKO muscles, including Aldo-keto reductase family 1 member A1 (Akr1a1), Peroxiredoxin 6 (Prdx6), and different distinct classes of glutathione S-transferases. As previous reports indicated, Akr1a1 participates in redox-based cellular signaling and Prdx6 is involved in the nuclear factor erythroid 2-related factor 2 (Nrf2)-mediated antioxidant defense system [46, 47]. The dysregulation of inflammasomes and the presence of increased oxidative stress then destroy cellular homeostasis thus accelerating programmed cell death.

Limitation and prospectives

In this study, we demonstrated that Cisd2 is more essential than Cisd1 for maintaining skeletal muscle function and morphology in mice. Cisds gene deletion had huge influence on protein process and degradation; this results eventually in cell damage followed by cell death. Signaling regulation in our Cisds gene deleted models pinpoints critical mechanisms involved in modulating crosstalk and homeostasis among various organelles in cells. Additionally, immune response and metabolic regulation exacerbate muscle defects as the mice age. Intracellular defects typically induce the release of factors that circulate throughout the body, and this will lead to compensatory effects or disruption of the functioning of other cells or tissues. As a consequence of this, mice exhibit premature aging features, a retardation of growth, and a shortened lifespan. Every tissue and organ are able to release signals that modulate other tissues and organs, meaning critical aging-related factors may be secreted from both skeletal muscle and non-muscle tissues. Understanding the regulation of crosstalk between tissues will help to facilitate insights into the mechanisms involved in systemic aging processes.

Cisds seems to be a critical target when investigating muscle disorders and this may lead to novel therapeutic strategies. Our studies have shown that Cisd2 deletion leads to more severe physiological phenotypes compared to Cisd1 deletion, suggesting that Cisd2 may play a more important role than Cisd1. Previous research has demonstrated that pharmacological enhancement of Cisd2 expression can mitigate aging-related defects in the skin [48]. Therefore, there is a significant potential to rescue muscle defects by elevating Cisd2 expression.