Detection of PAX5 Overexpression By RNA-Seq for B-Cell Lineage Detection in Acute Lymphocytic Leukemia

Background: Paired-box Pax gene family protein 5 (PAX5)/B-cell specific activator protein (BSAP) is a transcription factor encoded by the PAX5 gene and has an essential role in B-cell differentiation and maturation. High PAX5 expression is detected ensures commitment to B-cell lineage. PAX5 is normally downregulated at the plasma cell stage of B-cell development. Complete or partial deletion of the PAX5 gene has been found as secondary event associated with BCR-ABL1 or TCF3-PBX1 fusions in Acute Lymphocytic Leukemia (ALL) cases. PAX5 expression is a diagnostic marker for B-cell lineage and may help quantify minimal residual disease in B-ALL. Lineage determination of leukemic blasts is most often performed by flow cytometry, but also by immunohistochemistry (IHC). Evaluation of PAX5 is most commonly available by IHC and is not widely performed by flow cytometry. In cases with limited specimen for evaluation or aberrant loss of some B-cell markers, determining quantitative levels of RNA from lineage-specific genes, such as PAX5, could be a valuable clinical diagnostic tool for ALL patients.

Our existing single tube NGS based assay for simultaneous detection of DNA alterations and RNA fusions in heme malignancies from Total Nucleic Acid (TNA), can also be used to detect PAX5 gene expression through select exons enrichment along with a total of 213 genes. However, one of the current challenges for NGS-based gene expression profiling is to setup a threshold for overexpression. Here we developed a cutoff criterion for PAX5 overexpression and evaluated the performance of PAX5 gene expression analysis using the in-use heme assay and its potential use in clinical laboratory for cell lineage detection.

Methods: RNA sequencing was performed on TNA extracted from ALL samples and from 32 healthy donors using partial anchored amplicon based (Qiagen, inc) heme NGS assay. PAX5 RNA expression was calculated by TPM (transcript per million) counts normalized to TPM of the house-keeping gene GUSB. A commercially available qRT-PCR assay was used as orthogonal method to confirm the gene expression. The expression call by NGS based on the normalized value was confirmed by a commercial qRT-PCR assay in house validated through serial dilutions of template for six log scale. The analytical cutoff was determined from normalized TPM calculation from 32 healthy volunteers following CLSI guideline (CLSI_EP17-A2) and evaluated the outcome with IHC positive /negative clinical samples (a CLIA validated assay). Further, we used the established cutoff to evaluate the sensitivity or specificity in cohort of ALL samples.

Results: In this study we established the cutoff for PAX5 gene over-expression using the currently in-use heme NGS assay. First, a cutoff was established following the method in the CLSI guidelines and tested for sensitivity and specificity in the ALL sample cohort. PAX5 TPM normalization to GUSB or to the geometric mean of four house keeping genes (GUSB, PGD, RPL5 and RPL19) showed a strong correlation (R2>0.95), and GUSB was selected for further normalization since GUSB TPM values were most conserved across all the samples. Independent in-house evaluation for commercial qRT-PCR assay showed efficiency at 94.3 and 96% for GUSB and PAX5, respectively (with linearity R2>0.95), and been used to compare the NGS and IHC data as independent orthogonal assay. When a cohort of samples for Pax5 by IHC (positive and negative), a sensitivity at 67% and specificity at 100% were observed for the NGS based Pax5 detection. NGS results on the discordant samples were confirmed by qRT-PCR to have low RNA expression. Notably the discordant, IHC positive samples contained very low numbers of B cells. Alongside with other possible mechanisms of increased protein levels such as increased protein translation/increased protein stability could explain the discordance between RNA expression and the protein detection by IHC.

Conclusions: In this study we demonstrate that NeoGenomics's (heme) NGS assay can be used for PAX5 gene over-expression analysis on ALL. The heme NGS is inexpensive and is already integrated in the benchwork workflow without adding extra burden and can be used as an objective quantification of PAX5 levels overcoming the challenges associated with the relative signal intensity biases in IHC testing. This type of RNA testing can be useful especially with specimens having limited material.