PENK Knockdown Improves Mitochondrial Function in High-Glucose Cardiomyocytes

Researchers from Zhongshan Hospital Fudan University in Shanghai, China, have identified proenkephalin, or PENK, as a significant candidate gene involved in mitochondrial quality control dysregulation during diabetes-associated myocardial injury. The study, published in Frontiers in Cardiovascular Medicine, integrates transcriptomic screening of rat myocardial datasets with in vitro validation in human AC16 cardiomyocytes. The central finding is that PENK expression increases under hyperglycemic conditions and that suppressing this gene using two independent small interfering RNA sequences improves cell viability and mitochondrial functional phenotypes. This work provides experimental evidence linking PENK to the cellular mechanisms underlying diabetic heart injury, moving beyond previous associations with renal function and cardiovascular outcomes.

The investigation began with a discovery cohort from the Gene Expression Omnibus dataset GSE4745, which comprised 16 ventricular samples from male Wistar rats. These samples included eight controls and eight animals with streptozotocin-induced diabetes, with tissue collected 28 or 42 days after administration. The research team performed differential expression analysis, identifying 192 differentially expressed genes among 4,822 detected genes. By intersecting these genes with curated mitochondrial quality control gene sets, the team isolated 58 mitochondrial quality control-related differentially expressed genes. These genes were distributed across multiple mitochondrial processes, including fatty acid oxidation, oxidative phosphorylation, and mitophagy, indicating that diabetic myocardial injury involves broad transcriptional alterations in mitochondrial homeostasis.

To prioritize candidate genes, the researchers utilized weighted gene co-expression network analysis, an exploratory approach given the sample size of the discovery cohort. The brown module showed a strong positive correlation with diabetes status, with a Spearman correlation coefficient of 0.900. PENK emerged as a high-priority candidate, ranking fourth among the selected genes. It was significantly upregulated in diabetic myocardial samples, with a log2 fold change of 2.17 and a false discovery rate of 1.65 times 10 to the power of negative 5. The gene also exhibited high module membership and gene significance for diabetes status, supporting its relevance to the disease state. HMGCS2 ranked first in the discovery-stage analysis, serving as a comparative reference for the prioritization of PENK.

The study design included three supportive cohorts to assess the consistency of candidate-gene expression patterns: GSE5606, GSE6880, and GSE197999. These datasets contained myocardial tissue samples from diabetic rat models and corresponding controls. PENK exhibited a consistent upward expression pattern across all three datasets. In GSE5606 and GSE197999, PENK upregulation was statistically significant, while a similar upward trend was observed in GSE6880, although the difference did not reach statistical significance. Receiver operating characteristic analysis showed complete separation between diabetic and control samples within these supportive cohorts, with an area under the curve of 1.00 for PENK. These cohort-specific estimates characterize expression-based group separation rather than clinical diagnostic performance.

Functional validation was conducted in human AC16 cardiomyocytes, an established in vitro model for investigating hyperglycemia-associated injury. Cells were exposed to normal glucose, high glucose, or a mannitol osmotic control for 48 hours. High-glucose exposure significantly reduced cell viability compared with the normal glucose and mannitol groups, confirming that the injury was not solely due to increased osmolarity. RT-qPCR and Western blot analysis demonstrated that high-glucose exposure significantly increased PENK expression at both the mRNA and protein levels. This concordance between the transcriptomic findings in rat myocardium and the cellular expression in human cardiomyocytes provided a rationale for investigating the functional contribution of PENK to high-glucose-induced injury.

To assess the functional role of PENK, the researchers used two independent siRNAs targeting distinct PENK sequences to suppress gene expression in high-glucose-treated cells. In the initial experiment, PENK knockdown using si-PENK #1 significantly improved cell viability compared with the high-glucose and high-glucose with negative-control siRNA groups. Western blot analysis showed that PENK knockdown reduced the expression of cleaved caspase-3, an apoptosis-associated protein. These findings indicate that PENK suppression attenuates the cellular injury response to hyperglycemic stress, specifically by reducing markers associated with programmed cell death.

Mitochondrial function was evaluated using MitoSOX Red fluorescence to assess superoxide production and JC-1 staining to measure membrane potential. High-glucose exposure increased MitoSOX fluorescence intensity and reduced the JC-1 red-to-green fluorescence ratio, indicating enhanced mitochondrial oxidative stress and membrane depolarization. PENK knockdown significantly reduced MitoSOX fluorescence and increased the JC-1 ratio, suggesting an improvement in mitochondrial membrane potential. Western blot analysis further revealed that PENK knockdown decreased the protein abundance of PINK1, Parkin, and DRP1, while increasing MFN2 expression. These proteins are involved in mitochondrial dynamics and quality control, suggesting that PENK modulation affects these regulatory pathways.

An independent validation experiment using si-PENK #2 reproduced the principal cellular and mitochondrial phenotypes observed with the first siRNA. The second siRNA significantly improved cell viability, reduced cleaved caspase-3 abundance, decreased mitochondrial superoxide-related fluorescence, and improved mitochondrial membrane potential. The agreement between the two independent siRNA experiments strengthens the interpretation that the observed phenotypes are associated with PENK suppression rather than being artifacts of a single siRNA sequence. This reproducibility is crucial for establishing the functional relevance of PENK in the high-glucose injury response.

The study highlights the complexity of mitochondrial quality control in diabetes-associated myocardial injury. Transcriptomic analysis revealed increased fatty acid oxidation scores alongside reduced mitochondrial protein import and mitophagy scores in diabetic samples. The integrated mitochondrial quality control score was reduced in diabetic samples, although sensitivity analysis showed that this overall difference was strongly influenced by the fatty acid oxidation component. These findings suggest that diabetic myocardial injury involves coordinated changes in mitochondrial metabolic adaptation and quality control-associated transcriptional programs, with PENK positioned as a candidate linking these transcriptional alterations to measurable injury phenotypes.

The authors note that the current findings are derived from animal myocardial transcriptomic datasets and an AC16 cellular model, with weighted gene co-expression network analysis contributing exploratory network evidence. The study does not establish a direct causal mechanism by which PENK influences mitochondrial function, but rather demonstrates a functional association through loss-of-function experiments. Further investigation in complementary cardiac models and characterization of PENK-associated downstream signaling may help clarify its broader biological and therapeutic relevance. The identification of PENK as a high-ranking candidate provides a foundation for understanding mitochondrial dysfunction and cellular injury under hyperglycemic conditions.

To establish the functional framework for evaluating mitochondrial quality control, the investigators curated gene sets from Human MitoCarta3.0, MitoPathways3.0, and the Molecular Signatures Database collections, including Hallmark, curated pathways, and Gene Ontology categories. Human genes were mapped to rat orthologs using the babelgene package with a minimum support threshold of three, filtering out duplicates within each functional set. Preranked gene set enrichment analysis using fgsea and single-sample enrichment scoring through the GSVA package yielded standardized pathway activity metrics. Beyond examining individual pathways, the authors formulated an exploratory composite score balancing positive components, including mitophagy, protein import, and membrane potential, against negatively oriented components comprising fatty acid oxidation and oxidative phosphorylation.

This pathway-level scoring revealed distinct functional divergence in the discovery cohort. Diabetic myocardial samples demonstrated a marked rise in fatty acid oxidation scores, exhibiting a mean difference of 0.073 compared with control tissue, alongside significant reductions in mitochondrial protein import and mitophagy scores. While the integrated mitochondrial quality control score was significantly lower in diabetic tissue by a difference of 1.046, a leave-one-component-out sensitivity analysis demonstrated that this composite effect was heavily driven by the fatty acid oxidation component. When researchers sequentially excluded fatty acid oxidation, the between-group difference diminished to 0.036 and lost statistical significance, underscoring the critical influence of specific substrate oxidation shifts within the broader transcriptional remodeling.

The co-expression architecture was mapped using a signed network constructed from the 4,000 genes displaying the highest median absolute deviation, applying a soft-thresholding power of 16, a minimum module size of 300, and a merge cut height of 0.75. To confirm that the prioritization of PENK was not an artifact of specific algorithmic parameters, the authors tested 45 distinct parameter combinations and performed a 16-sample leave-one-out cross-validation. PENK maintained a median module membership of 0.965 across parameter adjustments, although network reconstruction revealed that excluding a single sample, GSM107428, caused the largest fluctuation in modular stability, illustrating the sensitivity inherent to small-cohort discovery networks.

The three supportive animal cohorts provided distinct experimental contexts that helped assess candidate robustness across diverse diabetic durations and rat strains. The GSE5606 cohort examined left-ventricular tissue 16 weeks following tail-vein streptozotocin administration in Wistar rats, while GSE6880 evaluated animals four weeks post-exposure. In GSE197999, which utilized Sprague-Dawley rats 96 days following intraperitoneal streptozotocin injection, count data were processed through the edgeR filterByExpr function and limma-voom transformation. In GSE6880, correlation testing revealed an inverse association between PENK expression and oxidative phosphorylation scores, demonstrating an exact Spearman rho of negative 1.000 across the cohort samples.

In the cellular loss-of-function experiments, human AC16 cardiomyocytes were transfected at 60 to 70 percent confluence with 50 nanomolar concentrations of small interfering RNA using Lipofectamine 3000, with culture medium replaced after six hours. High-glucose stress was initiated 24 hours post-transfection and maintained for 48 hours. Mitochondrial superoxide production was quantified following a ten-minute incubation with five micromolar MitoSOX Red at 37 degrees Celsius, while mitochondrial membrane potential was measured after 15 to 30 minutes of JC-1 staining. All image fields were averaged to yield one value per biological replicate across three independent runs, confirming reproducible phenotypic recovery following gene suppression.

Subject of Research: PENK involvement in mitochondrial quality control dysregulation during diabetes-associated myocardial injury

Article Title: Transcriptomic screening and in vitro validation suggest PENK involvement in mitochondrial quality-control dysregulation in diabetes-associated myocardial injury

Article References: Zuo, Z., He, B., Ming, Y., Liu, J., Zhu, K., Lai, H., Wang, C., Li, J., & Zhai, J. (2026). Transcriptomic screening and in vitro validation suggest PENK involvement in mitochondrial quality-control dysregulation in diabetes-associated myocardial injury. Frontiers in Cardiovascular Medicine, 13, Article 1970128. https://doi.org/10.3389/fcvm.2026.1970128

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1970128

Keywords: proenkephalin, mitochondrial quality control, diabetes-associated myocardial injury, AC16 cardiomyocytes, transcriptomic screening, high-glucose injury, mitochondrial membrane potential, mitochondrial superoxide, siRNA knockdown, WGCNA, cleaved caspase-3, PINK1/Parkin signaling, DRP1/MFN2 dynamics, fatty acid oxidation, mitophagy, cardiomyocyte viability, hyperglycemic stress, mitochondrial dysfunction, GEO datasets, cardiovascular metabolism