# Normalization Challenges Across Adipocyte Differentiation and Lipid-Modulating Treatments: Identifying Reliable Housekeeping Genes

## TL;DR

This paper identifies the most reliable reference genes for normalizing RT-qPCR data in adipocyte studies involving differentiation and lipid treatments.

## Contribution

The study introduces a reliable normalization panel of housekeeping genes (Ppia, 36b4, B2M) for adipocyte research.

## Key findings

- Commonly used reference genes like Gapdh and Actb show high variability in adipocyte differentiation and lipid treatments.
- Ppia, 36b4, and B2M consistently display low variability across multiple analytical methods.
- A combination of algorithmic and statistical methods is recommended for reliable normalization in complex adipocyte models.

## Abstract

Accurate normalization of RT-qPCR data requires selecting stable internal control genes, particularly in models characterized by dynamic metabolic transitions, such as 3T3-L1 adipocytes. The current study compares the expression stability of nine widely used housekeeping genes (HKGs) (peptidylprolyl isomerase A (Ppia), glyceraldehyde-3-phosphate dehydrogenase (Gapdh), beta-2 microglobulin (B2M), ribosomal protein, large, P0 (36b4), hydroxymethylbilane synthase (Hmbs), hypoxanthine guanine phosphoribosyl transferase (Hprt), tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein, zeta polypeptide (Ywhaz), 18S ribosomal RNA (18S), and β-actin (Actb)) across key stages of differentiation (days 0, 9, and 18) and under treatments with palmitic acid and docosahexaenoic acid. Stability was assessed using four classical algorithms—geNorm, NormFinder, BestKeeper, and RefFinder—supplemented by the ΔCt method, conventional statistical testing, correlation, and regression analysis relative to two target genes, fatty acid-binding protein 4 (Fabp4) and sterol regulatory element binding transcription factor 1 (Srebf1). The obtained data indicate that no single HKG remains universally stable across these experimental conditions, and the expression of traditionally used reference genes (Gapdh, Actb, Hprt, 18S) is highly influenced by both the stage of adipogenesis and exposure to lipid-modulating factors. In contrast, Ppia, 36b4, and B2M—despite some of them being underestimated in use as references—consistently display the lowest variability across most analytical tools, forming a reliable and functionally diverse normalization panel. It should be noted that our initial stability assessment revealed apparent discrepancies among mathematical evaluation methods, emphasizing the need for a holistic, multiple-level approach strategy. The applied combination of algorithmic and statistical methods provides a more rigorous and objective framework for assessing the stability of reference genes, which is highly recommended in such a complex adipocyte-based model.