Reference genes for mesangial cell and podocyte qPCR gene expression studies under high-glucose and renin-angiotensin-system blocker conditions

Background Real-time PCR remains currently the gold standard method for gene expression studies. Identification of the best reference gene is a key point in performing high-quality qPCR, providing strong support for results, and performing as a source of bias when inappropriately chosen. Mesangial cells and podocytes, as essential cell lines to study diabetic kidney disease (DKD) physiopathology, demand accurate analysis of the reference genes used thus far to enhance the validity of gene expression studies, especially regarding high glucose (HG) and DKD treatments, with angiotensin II receptor blockers (e.g., losartan) being the most commonly used. This study aimed to evaluate the suitability and define the most stable reference gene for mesangial cell and podocyte studies of an in vitro DKD model of disease and its treatment. Methods Five software packages (RefFinder, NormFinder, GeNorm, Bestkeeper, and DataAssist) and the comparative ΔCt method were selected to analyze six different candidate genes: HPRT, ACTB, PGAM-1, GAPDH, PPIA, and B2M. RNA was extracted, and cDNA was synthesized from immortalized mouse mesangial cells and podocytes cultured in 4 groups: control (n = 5; 5 mM glucose), mannitol (n = 5; 30 mM, as osmotic control), HG (n = 5; 30 mM glucose), and HG + losartan (n = 5; 30 mM glucose and 10−4 mM losartan). Real-time PCR was performed according to MIQE guidelines. Results We identified that the use of 2 genes was the best combination for qPCR normalization for both mesangial cells and podocytes. For mesangial cells, the combination of HPRT and ACTB presented higher stability values. For podocytes, HPRT and GAPDH showed the best results. Conclusion This analysis provides support for the use of HPRT and ACTB as reference genes in mouse mesangial cell studies of gene expression via real-time PCR, while for podocytes, HPRT and GAPDH should be chosen.


Introduction
Our goal was to evaluate the suitability and define the most stable reference gene specifically for mesangial cell and podocyte studies of an in vitro DKD model of disease and its treatment among six commonly used reference genes (HPRT, ACTB, PGAM-1, GAPDH, PPIA, and B2M).

Cell lines and cell culture
Immortalized mesangial cells (SV40 MES 13, ATCC) were cultured in DMEM (Invitrogen Corporation, Gaithersburg, MD, USA) containing 10% fetal bovine serum (FBS), penicillin (50 U/ml) and 2.6 g HEPES at 37˚C. Podocytes (Cell line E11, CLS) were cultured in RPMI 1640 medium (Invitrogen Corporation, Gaithersburg, MD, USA) supplemented with 10% fetal bovine serum and interferon-gamma (INF-gama) at 33˚C; after achieving the desired confluence, flasks were transferred to a 37˚C incubator for the differentiation process for 14 days without INF-gama. Both cell types were cultured until >90% confluence and remained in a 5% CO 2 environment. After 24 hours in 1% FBS, each group received the designated stimulus for 24 hours: pure medium (control group), medium containing 30 mM mannitol (as osmotic control, mannitol group), 30 mM D-glucose (high-glucose group) or 30 mM D-glucose combined with 100 μM losartan (losartan group). The study workflow is shown in Fig 1.

RNA extraction, quality parameters and reverse transcription
Total RNA was extracted using TRIzol reagent (Life Technologies, USA) according to the manufacturer's instructions. RNA concentration and quality (260/280 ratio >1. 8 and 260/230 ratio 2.0-2.2, indicating high purity) were assessed using a NanoVue spectrophotometer (GE Healthcare Life Sciences, USA). RNA integrity was also analyzed by gel electrophoresis. After RNA extraction, we performed DNAse treatment to avoid genomic DNA contamination. Two micrograms of total RNA was reverse-transcribed into cDNA (High Capacity cDNA Reverse Transcription Kit, Applied Biosystems, USA). The reaction mixture was incubated for 10 minutes at 25˚C, 120 minutes at 37˚C and 5 seconds at 85˚C. Study workflow. The figure shows the workflow for determination of the most stable reference gene for mesangial cells and podocytes exposed to mannitol, high glucose or high glucose and losartan. Five software packages (RefFinder, NormFinder, GeNorm, Bestkeeper, and DataAssist) and the comparative ΔCt method were applied to analyze six different candidate genes: HPRT, ACTB, PGAM-1, GAPDH, PPIA, and B2M. RNA was extracted, and cDNA was synthesized from immortalized mouse mesangial cells and podocytes cultured in 4 groups: control (n = 5; 5 mM glucose), mannitol (n = 5; 30 mM, as osmotic control), HG (n = 5; 30 mM glucose), and HG + losartan (n = 5; 30 mM glucose and 10 −4 mM losartan).

Software analysis for stability of candidate reference genes
To establish the best reference gene and best combination, we evaluated qPCR results in five different software applications: RefFinder, NormFinder, GeNorm, Bestkeeper, and DataAssist. We also evaluated the data with the comparative ΔCt method.
NormFinder is a freely available tool that provides the stability value for several candidate genes tested on a sample set. Any required number of samples is subject to the analysis, providing an estimation of expression variation [20]. GeNorm software works as an algorithm (M value) to determine the most stable reference genes among a collection of tested candidate

PLOS ONE
qPCR reference genes for mesangial cells and podocytes genes. The tool calculates a normalization factor for each sample, established according to the geometric mean of the reference genes number [10]. Bestkeeper is an Excel-based spreadsheet software that determines the best suited reference genes and combines them into an index, allowing a comparison with further target genes to decide which of them has the best suitability for normalization. The application acknowledges extremely deviating samples that can be removed from the calculation and improves the reliability of the results [21]. DataAssist is an Applied Biosystems software that quantifies relative gene expression across a given number of samples. It provides an "Endogenous Control Selection" tool that shows the Ct values of candidate genes for all samples as well as a score [22]. The ΔCt method compares the relative expression values between 'pairs of genes', implementing an elimination process according to a ranking of the variability among each pair. Subsequently, the most appropriate gene of reference can be selected [23]. The program GenEx was used to calculate the accumulated standard deviation across the samples, providing the necessary number of genes required for the minimum standard deviation [24]. Finally, we used the RefFinder software, an all-encompassing program developed with the aim of evaluating reference genes from experimental data. The tool includes available software algorithms and methods, all of which were previously mentioned: geNorm, Normfinder, BestKeeper, and the comparative ΔCt method. Supported by the ranking of each program, RefFinder calculates the geometric mean for an overall final ranking [25].

Statistical analysis
The entire dataset was analyzed regarding normality (Shapiro-Wilk test) and homogeneity (Levene's test). All comparisons were analyzed using ANOVA or Kruskal-Wallis tests, according to each test prerequisite. The level of significance considered was p<0.05. Analysis was performed using Jamovi software, version 1.0.1. The results are expressed as the mean ± standard deviation (SD).

Expression levels profile of candidate genes of reference
Raw Ct values were acquired in triplicate for both mesangial cell and podocyte samples and analyzed according to each stimulus received. Ct values are inversely proportional to gene expression. The Ct mean of the candidate genes ranged from 29.20 to 18.55 in mesangial cells. The highest Ct among the candidate genes in mesangial cells was achieved by ACTB (29.20 ± 1.09), and the lowest was achieved by PPIA (18.55 ± 0.79). HPRT showed a mean of 23

Stability of candidate genes
We applied the six algorithms described previously to determine the stability of each reference gene candidate according to the cell type. After analysis with different algorithms and a visual inspection of the ranked genes, we concluded that HPRT and ACTB for mesangial cells ( Table 2) and HPRT together with GAPDH for podocytes were the best reference genes for qPCR studies (Table 3).
NormFinder showed the lowest stability value for HPRT for both mesangial cells and podocytes, indicating that this gene is the best reference gene according to this algorithm. As recommended by software instructions, any gene with a stability value higher than 0.5 is considered unstable; all genes tested showed a stability value lower than the cutoff.
Individual results for BestKeeper showed, for mesangial cells, the lowest coefficient of variation (CV) for ACTB, indicating that this gene is the best for this cell type under the established conditions according to this software. For podocytes, the lowest CV was displayed for B2M, showing up as the most stable option. No candidate gene showed an SD higher than 1.0, the fixed software threshold for instability.
DataAssist software retrieved the PGAM-1 gene as the most stable gene for mesangial cells. Nevertheless, the lowest score for podocytes was achieved by HPRT.

Fig 3. Expression profile of the six candidate reference genes in mesangial cells (A) and podocytes (B).
A lower threshold value (Ct) represents a higher gene expression level. The data are presented as the mean +/-standard deviation. Each dot represents the average from triplicate ΔCt from each sample. All genes were tested for differences among the groups (Kruskal-Wallis test), and all comparisons showed a nonsignificant result (p>0.05).
https://doi.org/10.1371/journal.pone.0246227.g003 Regarding the ΔCt method, the lowest SD was obtained by HPRT for both cell lines. The highest SD was shown by B2M for mesangial cells and ACTB for podocytes, classifying them as the least stable genes according to the method.
For mesangial cells, GeNorm showed the best results for ACTB and GAPDH together according to M-values. B2M was considered the least stable gene. For podocytes, the best pair of M-values was given to HPRT and GAPDH. The least stable gene for podocytes was PPIA.
Based on these results and visual inspection of all data (Tables 2 and 3), HPRT was selected as the overall best reference gene for both mesangial cells and podocytes. For mesangial cells, HPRT and ACTB were considered the best combination of genes for qPCR normalization (Table 2). PPIA, otherwise, was classified as the least stable for mesangial cells. Along with HPRT, GAPDH was also ranked as the most stable candidate reference gene for podocytes, while ACTB, PGAM-1, and PPIA were found to be the least feasible genes ( Table 3).

Determination of the suitable number of reference genes
For each cell line, we determined the optimal number of genes to be used in a gene expression experiment via qPCR. This analysis was performed by Genex software, and the accumulated standard deviation (Acc.SD) parameter was considered for each cell line according to the number of genes used. For mesangial cells, we concluded that the Acc.SD decreased proportionally to the number of genes used. We also observed that the difference from one to two genes was higher than 0.1. However, the difference from two to three genes was smaller than 0.1 -a pattern that could be noticed in the following number of genes as well, achieving a plateau. Therefore, it would be reasonable to use two reference genes (HPRT and ACTB) and maintain a smaller source of error, since a higher number of genes increases the overall noise of the experiment as well as the cost (Fig 4A).
For podocytes, the lowest Acc.SD was acquired in the presence of 2 reference genes ( Fig  4B). In this case, the use of 2 genes-HPRT and GAPDH or B2M, the top genes according to the visual inspection-to analyze qPCR results would be the best option as well. Since GAPDH and B2M showed the same results on visual inspection ranking, we looked closely to the performance of each gene on all softwares: besides being the top gene for BestKeeper, B2M appeared in 5 th place for NormFinder, 3 rd for GeNorm, 5 th for DataAssist, 4 th for RefFinder and 4 th again for ΔCt method. GAPDH, however, in addition to being the top gene in GeNorm, appeared as 2 nd for NormFinder, 3 rd for BestKeeper, 2 nd for DataAssist, 2 nd for RefFinder and 2 nd for the ΔCt method. Considering the overall performance of both genes, GAPDH was selected as the best option to pair with HPRT as a reference gene for podocytes.

Correlation between the top candidates
After determining that the use of 2 reference genes would be the ideal option for mesangial cells and podocytes, we checked if the best 2 genes for each cell line were correlated and therefore could be used simultaneously. We found that the 2 recommended genes for mesangial cells, HPRT and ACTB, were strongly correlated (ρ = 0.80, p<0.0001, Fig 5A), providing support to the recommendation of using these genes at the same time to analyze qPCR data. The same occurred for podocytes: there was a strong correlation between HPRT and GAPDH expression data, again supporting the use of those genes together as reference genes for qPCR (ρ = 0.92, p<0.0001, Fig 5B).

Validation of the best reference genes
As the results showed that HPRT and ACTB were the best genes for normalization of mesangial cell qPCR data, we statistically confirmed that there was no difference among the four studied groups regarding the expression of these genes (p>0.05 by Kruskal-Wallis test) ( Fig  3A). We also confirmed that there was no difference among the groups of podocytes regarding HPRT and GAPDH expression, the best genes for this cell type (Fig 3B). In fact, there was no difference between the groups for all candidate genes in either cell line.

Discussion
The pipeline used in this work has been extensively used throughout many laboratories and is accepted by the literature as a reliable approach to determine the best reference gene to be used, specifically for qPCR in a predetermined biological sample and condition [24,26,27].
Here, we aimed to provide data to determine the most suitable reference gene to be used for

Fig 4. Evaluation of the optimal number of reference genes in (A) mesangial cells and (B) podocyte cells.
Accumulated standard deviation (Acc.SD) was accessed by GenEx software for the six candidate reference genes in all samples for each cell type. The Acc.SD value was used to determine the optimal number of genes to be used in a gene expression experiment by qPCR. Lowest value of Acc.SD indicate the best number of reference genes.
https://doi.org/10.1371/journal.pone.0246227.g004 mesangial cells and podocytes exposed to a high glucose environment and treated with losartan, a very known in vitro model for diabetic kidney disease [7,[28][29][30][31]. Many research groups have clearly shown a need for studies that approach reference genes for their specific study sample [32,33]. The importance of using the best-known reference gene and pragmatically looking at this question relies on the frequent inappropriate use of the least feasible reference genes, resulting in an inaccurate analysis of qPCR results and therefore in the loss of reagents, time, and samples. Sometimes the most known genes, such as ACTB and GAPDH, are used for samples and conditions that do not support their use. Even in the most recent years, researchers still normalize their qPCR data of in vitro studies based on the most frequently used genes, such as ACTB for podocytes and GAPDH for mesangial cells (opposite to the finding we had in our analysis), without literature support for this choice [34][35][36][37][38][39]. Unfortunately, some studies do not clearly provide which reference gene was used to normalize the data, or even if there was data normalization. This shows the need for systematic analysis to identify the best gene or genes to be used as references.
In fact, the literature frequently stands against the use of many popular reference genes. A systematic review performed on vertebrate studies found that 72% of the included studies used GAPDH, ACTB or 18S as normalizing genes. The same group shows that as the number of screened reference genes for a specific study design increases, the chance of one of these three genes being the most stable decreases [40].
In nephrology, few studies have addressed reference genes for qPCR normalization [12,13,16,17], exposing a lack of information regarding which gene must be used for gene expression studies for kidney samples and cell lines. The kidney itself is an organ specifically characterized by numerous cell types, justifying the need for reference genes regarding each different cell line [41,42]. The genes selected as best genes for the studied samples-HPRT and ACTB for mesangial cells and HPRT along with GAPDH for podocytes-are extensively described in the literature. Hypoxanthine phosphoribosyltransferase (HPRT) is mainly known for its role in the metabolism of purines, although impaired expression of this gene is also responsible for causing cell cycle dysregulation and multisystem regulatory dysfunction [43,44]. Actin beta (ACTB) is involved in cell structure, motility, and integrity, and as it is essential to multiple cell functions, the gene is highly abundant in many cell lines [45]. A previous study on reference genes for rat mesangial cells found ACTB to be one of the best reference genes for this cell line under high glucose conditions, in conformity with our findings [17]. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH), although it is reported to be involved in cellular survival, apoptosis and DNA repair, is mainly known to express a cellular energy enzyme determinant of the glycolytic process, functioning as a catalyzer of triose phosphate oxidation and, for this reason, ubiquitously distributed in all cell types [46,47].
The other three genes considered in this study (B2M, PPIA, and PGAM-1) are also known as common reference genes. Beta 2-microglobulin (B2M) expresses a low molecular weight protein related to immune processes linked to the major histocompatibility complex (MHC) [48,49]. B2M is considered a highly conserved molecule in many different vertebrate species and therefore may be considered a possible reference gene in our setting [49]. Peptidylprolyl isomerase A (PPIA) expresses proteins that catalyze proline imidic peptide bonds in oligopeptides and are also implicated in protein folding processes [50]. It is ubiquitously distributed in multiple cell types, including kidney cells [51]. Phosphoglycerate mutase 1 (PGAM-1) expresses an enzyme responsible for catalyzing 3-phosphoglycerate (3-PGA) to 2-phosphoglycerate (2-PGA), playing an essential role in glycolysis. It is also ubiquitously distributed in multiple cell types [52,53]. Shared characteristics related to vital processes in the organism, such as glycolysis, immune response, protein folding and cell structure, especially due to applicability to multiple species and most tissues, bring up those 6 genes as relevant targets to investigate as suitable reference genes in our study.
As long-established cell lines in the literature, mesangial cells and podocytes are important biological samples to determine the best reference gene-many researchers in the field are focused on these structures [54][55][56][57][58][59], and the data provided by our work could potentially influence many studies, providing support to avoid incorrect interpretation of results and their influence in downstream analysis and further conclusions.

Conclusion
We analyzed six different genes using five software applications and the ΔCt method to determine that the best genes to be used for mesangial cell studies with high glucose and angiotensin receptor II blockers are HPRT and ACTB, while under the same conditions, the best combination of genes for podocyte gene expression normalization is HPRT together with GAPDH. We believe our work may provide support to many research laboratories engaged in mesangial cell and podocyte cell culture studies, allowing them to improve the quality of gene expression studies via qPCR and, consequently, the overall quality of nephrology research.