Quantification of cell-associated HIV RNA (ca-RNA) is one of the most important and commonly used methods to evaluate the performance of latency-reversing agents (LRAs). Copies of HIV RNA measured by quantitative PCR (qPCR) are often normalized to the input RNA or cell number. However, these could be affected by biological variability and/or technical errors, which can be avoided by using an internal reference gene. To obtain reliable data, it is essential to select stable reference genes (RGs) of which the expression is not influenced by biological variability, the type of cells, or the LRAs used. However, to date, no study has carefully evaluated RG stability following LRA exposure. We analyzed the stability of six widely used RGs (GAPDH, TBP, YWHAZ, UBE2D2, HPRT1, and RPL27A) in human peripheral blood mononuclear cells (PBMCs) and CD41 T cells. LRA exposure significantly influenced the stability of these RGs. Overall, TBP, UBE2D2, and RPL27A were the most stable RGs under all tested conditions. TBP was generally the most stable RG, whereas GAPDH varied the most. Finally, we evaluated the impact of applying different RG normalizers to host genes and HIV ca-RNA data. Altered results were observed both in host and HIV gene expression when unstable RGs were used. Our data underline the importance of testing the stability of RGs utilized to evaluate LRA-induced HIV ca-RNA expression. To our knowledge, this is the first careful evaluation of the stability of RGs after LRA exposure and will significantly contribute to the quality of data analysis in regard to gene expression. [ABSTRACT FROM AUTHOR]