Optimization of RNA extraction of fresh-frozen synovial sarcoma tumor samples

Elise Huang, Yvonne Vasquez, Anouk van den Bout, Olena Vaske

Introduction

Synovial Sarcoma (SS) is the second most common soft tissue sarcoma in young people, with 800 to 1000 diagnoses in the United States each year. Up to 64% of SS patients die within five years of diagnosis. Poor patient survival rates are partially due to a lack of effective, targeted treatments. There is a major push in the field toward personalized medicine, wherein the treatment of each patient is tailored to the genetic diversity of their tumors. For cancers like SS with few recurrent mutations, RNA sequencing (RNA-seq) technology is useful for exploring abnormal gene expression patterns. The identification of abnormal gene expression in patient tumor samples has implications for identifying molecular vulnerabilities that can be explored for therapy. However, a challenge in conducting such genomic analyses is small sample sizes. The Vaske lab has compiled one of the largest uniformly processed tumor RNA-seq data compendia with over 12,000 samples; however, this only includes 67 SS samples. To grow this compendium, the lab actively curates data and recently started generating our own RNA-seq data. The Vaske lab currently has an RNA-seq assay to identify gene fusions, expressed mutations, and gene expression abnormalities in pediatric tumors. This assay consists of RNA extraction from tumor samples, preparation of RNA-seq libraries, sequencing, and data analysis. The goal of this project was to optimize the Vaske Lab's RNA extraction protocol for processing fresh frozen tumor tissue samples. Further, I set out to apply the optimized RNA extraction protocol to fresh-frozen SS tumor samples and use the extracted RNA to generate SS RNA-seq data for the analysis of gene fusions and molecular vulnerabilities for treatment.

Methods

The Zymo RNA extraction protocol was optimized for processing frozen tissue samples.

Optimization was tested with the use of mouse samples and human lung tissue before moving on to the SS samples.

The following factors were tweaked:

homogenization time of the sample in the homogenizer
using the FastPrep-24 machine to homogenize the samples with DNA/RNA shield vs. using previous methods with the use of a vortexer
quantity of DNA/RNA shield added to the sample before and after homogenization
incubation time with lysis buffers and enzymes

To assess the quality of the RNA that was extracted, quality control (QC) checks were done. Our QC checks on the extracted RNA included determining the concentration via a Qubit, assessing the purity via Nanodrop, and determining the RNA integrity number (RIN) score via the Tapestation. All these QC checks helped us determine the quality of the sample as well as the next steps for the preparation of RNA sequencing libraries.

Data & Results

The following adjustments to the protocol when testing on mouse tissue improved the QC metrics and output of the RNA extraction protocol:

homogenization time of 6–12 minutes, depending on the tissue sample
incubation period upped to 1.5 hrs instead of 45 min.
transferring the supernatant of the homogenized sample and DNA/RNA shield to a new nuclease-free tube after incubation

The final optimized RNA extraction protocol is as follows:

Synovial sarcoma tumor RNA Data:

Tapestation Graph of 5 SS tumor RNA samples:

The tapestation was used to 1) calculate RIN scores and 2) assess transcript lengths.

1) RIN scores

The RIN scores are calculated by looking at the presence of the two most abundant RNAs, the small 18S subunit of ribosomal RNA and the large 28S subunit of ribosomal RNA. RIN scores range on a spectrum of 1–10, with 1 being the lowest quality and 10 being the highest quality. For tumor samples, we typically expect RIN scores greater than 6 to be good. All RIN scores for the SS RNA samples fell on the higher end of the RIN spectrum. This means that we were able to extract high-quality RNA for future sequencing.

2) Transcript lengths

To further assess the quality of an RNA sample, we look at transcript lengths to determine if we have full-length transcripts or degradation. Specifically, we look at the percent of transcripts that are greater than 200 nt; this is denoted as a DV200 score. In our lab, a DV200 score of >50% is sufficient to create good-quality RNA-seq libraries.

For all of our SS RNA samples, the DV200 scores ranged from 77-90%. This means that each sample contains mostly full-length transcripts and can be used to create RNA-seq libraries.

Discussion & Future Directions

The following adjustments to the protocol when testing on mouse tissue improved the QC metrics and output of the RNA extraction protocol:

homogenization time of 6–12 minutes, depending on the tissue sample
incubation period upped to 1.5 hrs instead of 45 min.
transferring the supernatant of the homogenized sample and DNA/RNA shield to a new nuclease-free tube after incubation

The lab's protocol for RNA extraction from fresh frozen SS tumor tissue samples was successfully optimized. Total cellular RNA from SS fresh-frozen tumor samples was successfully extracted and subsequently used to generate ribo-depleted RNA sequencing data. This, in turn, helped to grow the lab's riboD compendium for rare tumor types. The compendium, which previously had 19 SS samples, now has 24 riboD samples for transcriptomic analysis.

Acknowledgements

Thank you to the Koret Foundation and the generous donors who have offered their much-appreciated support to undergraduate research.

I would also like to acknowledge and thank the Vaske Lab for making this project possible and offering support. I would also like to specifically give a big thank you to my mentor, Yvonne Vasquez, who guided me throughout my project.

A big thank you as well to Koret organizer Don Bard!

Page updated

Report abuse