Discovering long noncoding RNA predictors of anticancer drug sensitivity beyond protein-coding genes

Area of Science:

• Genomics
• Pharmacology
• Cancer Biology

Background:

• Large-scale screens identified protein-coding genes (PCGs) as anticancer drug response biomarkers.
• Systematic evaluation of long noncoding RNAs (lncRNAs) as pharmacogenomic biomarkers remains challenging.

Purpose of the Study:

• To evaluate lncRNAs as drug response predictors, distinguishing genuine associations from correlations with PCGs or other factors.
• To assess the overall predictive power of the lncRNA transcriptome versus the PCG transcriptome.
• To identify specific lncRNAs and lncRNA alterations associated with drug response.

Main Methods:

• Comparative analysis of lncRNA and PCG transcriptomes for predicting drug response.
• Statistical methods to adjust for proximal PCG effects and identify independent lncRNA associations.
• Analysis of lncRNA-specific somatic alterations and positive selection in cancer cells.
• Experimental validation of selected lncRNAs as drug response predictors.

Main Results:

• The lncRNA transcriptome is as potent as the PCG transcriptome in predicting response to hundreds of anticancer drugs.
• Nearly half of initially significant lncRNA associations were linked to proximal PCGs.
• Adjusting for PCGs revealed significant lncRNAs that enhance drug response predictions for most drugs.
• Identified lncRNA-specific somatic alterations associated with drug response.
• Experimentally validated lncRNAs EGFR-AS1 and MIR205HG as functional predictors of anti-EGFR drug response.

Conclusions:

• lncRNAs are a valuable, largely untapped resource for pharmacogenomic biomarker discovery.
• Specific lncRNAs can improve drug response prediction beyond established biomarkers.
• Functional studies confirm the role of lncRNAs in mediating drug response, offering new therapeutic targets.