Assessing Creatine-Related Gene Expression in Kidney Disease: Can Available Data Give Insights into an Old Discussion?

Abstract

The impact of creatine supplementation on individuals with kidney disease or pathological conditions with an increased risk of developing kidney dysfunction remains an active discussion. However, the literature on gene expression related to cellular creatine uptake and metabolism under altered renal function is scarce. Therefore, the present study utilized comprehensive bioinformatics analysis to evaluate the expression of creatine-related genes and to establish their relationships to normal and disturbed renal conditions. We identified 44 genes modulated explicitly in response to creatine exposure from a gene enrichment analysis, including IGF1, SLC2A4, and various creatine kinase genes. The analysis revealed associations with metabolic processes such as amino acid metabolism, indicating a connection between creatine and tissue physiology. Using the Genotype-Tissue Expression Portal, we evaluated their basal tissue-specific expression patterns in kidney and pancreas tissues. Then, we selected several pieces of Gene Expression Omnibus (GEO) transcriptomic data, estimated their expression values, and established relationships to the creatine metabolism pathways and regulation, shedding light on the potential regulatory roles of creatine in cellular processes during kidney diseases. These observations also highlight the connection between creatine and tissue physiology, emphasizing the importance of understanding the balance between endogenous creatine synthesis and creatine uptake, particularly the roles of genes such as GATM, GAMT, SLC6A8, and IGF1, under several kidney dysfunction conditions. Overall, the available data in the biological databases can provide new insights and directions into creatine's effects and role in renal function.

Keywords: creatine monohydrate; kidney; renal injury; review.

Figure 1

Chemical–protein interaction network of creatine obtained by STITCH (http://stitch.embl.de, accessed on 26 April 2024). Small nodes: protein of unknown 3D structure. Large nodes: protein 3D structure is known or predicted. Colored nodes: first shell of interactors. White nodes: second shell of interactors. Edges color represent interactions sources. Magenta: from curated databases; Pink: experimentally determined; Green, red and dark blue: predicted interactions from gene neighborhood, gene fusions and gene co-occurrence, respectively; Light green: predicted from text mining.

Figure 2

Number of gene–DiseaseID (MESH—Medical Subject Headings) associations from the creatine-related genes from Stitch and CTD databases. Names on bars indicate the Medical Subject Headings associated with each DiseaseID.

Figure 3

(A) Kidney and (B) pancreas tissue expression levels of the 17 creatine-related genes used in the present study (TPM—transcripts per million bases).

Figure 4

Gene Enrichment Analysis from EnrichR-Kg using the creatine-related gene list against the following databases: (A) ReactomeDB; (B) Gene Ontology; and (C) KEGG terms. Green circles represent the genes. The first ten results are represented, using p-values ≤ 0.05 as the threshold.

Figure 5

Expression of creatine-related genes and the kinases retrieved from gene enrichment analysis against kinase-specific databases (TPM—transcripts per million bases).

Figure 6

Heatmap with the expression of creatine-related genes, from the study of Neusser et al. [45] (GEO Accession number: GDS3712). The sample codes and gene symbols are represented horizontally and vertically, respectively. Sample states (TN—control; NSC—nephrosclerosis) are represented in the first colored line. The following cutoffs were used: p-values < 0.05 and Log₂Fold Change ± 1.

Figure 7

Expression analyses of creatine-related genes from transplanted patient kidney biopsies (dataset GDS724, [49]). (A) Acute rejection transplant (AR) versus control samples. (B) Well-functioning transplant (TX) versus control samples. (C) Well-functioning (TX) versus acute rejection transplant samples (AR). (D) Well-functioning (TX) versus dysfunctional non-rejected transplant samples (NR). The plot shows the log2 fold change for the genes that met the significance threshold in each comparison (p < 0.05). The error bars represent the standard deviation of a log2 fold change for genes with multiple measurements (distinct transcripts or probes for the same gene). This provides a measure of variability in the fold change estimate. Printed numbers in each bar describe the p-values.

Figure 8

Expression analyses of the creatine-related gene transcriptional profile of two stages of clear cell renal cell carcinomas (cRCCs), paired with controls [51,52] (GEO Accession Number: GDS2880). (A) Stage I cRCC versus Stage I control samples. (B) Stage II cRCC versus Stage II control samples. The comparisons of Stage I versus Stage II, within normal and tumor samples, did not show differential expression of creatine-related genes. The plot shows the log2 fold change for the genes that met the significance threshold in each comparison (p < 0.05). The error bars represent the standard deviation of log2 fold change for genes with multiple measurements (distinct transcripts or probes for the same gene). This provides a measure of variability in the fold change estimate. Printed numbers in each bar describe the p-values.

Figure 9

Expression analyses of creatine-related gene transcriptional profile from kidney fibrosis biopsies (GEO Accession Number: GSE137570). (A) Samples with tubulointerstitial fibrosis above 50% (fibrosis > 50%) versus non-chronic kidney disease (non-CKD) ones. (B) Tubulointerstitial fibrosis below 50% (fibrosis < 50%) versus non-CKD samples. (C) Samples with eGFR lower than 50 (eGFR < 50) versus non-CKD samples. (D) Comparisons between eGFR < 50 and eGFR > 50 samples. The plot shows the Log₂ fold change for the genes that met the significance threshold in each comparison (p < 0.05). Printed numbers in each bar describe the p-values.