Single Cell Transcriptomics Software Targets Treatment-Resistant Cancer Cells

Designed to Cure

Human stem cells, which new research seeks to sequence individually in the context of cancer. Source: Wikipedia user Chinmaya Mahapatra.

Treatment resistance is unfortunately a common theme of oncotherapy.1 Tumors readily utilize natural selection to become immune to chemotherapies. But what if researchers could identify which cancer cells were the most likely to resist and nip them in the bud before starting treatment? A new paper published in Nature Medicine answers this question by bringing single cell transcriptomics and powerful analysis software to bear on cancer stem cells in chronic myeloid leukemia.2 Like many solutions to hard problems, the researchers behind the new paper bucked traditional methods and forged their own path forward.

Rooting Out Resistance

The authors of the paper proposed a single cell transcriptomic analysis protocol to examine the cancer stem cells most likely to survive chemotherapy and then subsequently bud off additional tumor cells to become a blastocyst. It’s not necessary to be an oncologist to understand that a blastocyst of cancer cells becomes a tumor as it grows.3

The researchers correctly assumed that cancer stem cells prone to survive chemotherapy would have a distinct molecular profile which could be detected before treatment and before pre-tumor formation blasting. It’s easy to see the appeal of a molecular profile of troublesome cells that could be turned into a therapy which would join other cancer treatments to reduce the length of treatment and reduce the chance of remission. Finding this molecular profile would prove difficult, however. The most common molecular biology tools that researchers know and love—RNA-seq and NGS—weren’t up to the task.

Preventing A Blast Crisis

Traditional transcriptomic methods like RNA-Seq generate a lot of data and are effective at detecting SNPs, but can’t always accurately get a handle on longer insertions or deletions without repeated runs.4  It’s likely that cancer stem cells have detectable SNPs, but there’s a good chance that there are other genetic anomalies going on, too.5 While some researchers have proposed getting around this obstacle by combining sequencing methods, doing so adds an ocean of data onto an already data-intensive process.6 For a problem as critical to research as robust sequencing, researchers are willing to suffer through the data and adjust their laboratory information systems accordingly, but the authors of the new article had a different idea.

To identify the troublemakers, the researchers invented a new method of single cell transcriptomic analysis which brings extreme sensitivity and also high accuracy to bear on minute samples of RNA isolated from cancer stem cells before treatment. The new method combines elements from smaller-scale implementations of the most popular techniques like RNA-seq and NGS with other, less used technologies which can keep track of long insertions and deletions.

In keeping with the scientific tradition of easily pronounceable nomenclature, the researchers dubbed their new technique BCR-ABL. To validate BCR-ABL’s efficacy, the researchers also ran a comparison between BCR-ABL and another single-cell sequencing technology, Smart-seq2.  7 Like other sequencing methods, both BCR-ABL and Smart-seq2 were ultimately detected by fluorescent probes in qPCR.

Though the researchers successfully identified a number of molecular markers which predict chronic myeloid leukemia treatment resistance at the cellular level, future researchers looking to apply similar techniques in transcriptomics need to stay abreast of the hefty data load that the relatively easy single-cell sequencing experiments will yield.

To successfully utilize single cell transcriptomics to pick out molecular profiles of chemotherapy-resistant cells, researchers will need to keep track of:

  • Patient exome data
  • Healthy cell genomic DNA
  • Healthy cell transcriptional RNA
  • Healthy cell transcriptional and genomic response to treatment
  • Normal tumor cell genomic DNA
  • Normal tumor cell transcriptional RNA
  • Normal tumor cell unique SNPs and indels
  • Normal tumor cell transcriptional and genomic response to treatment
  • Normal tumor cell molecular profile, incorporating the above information

That’s just the start, however. Once the healthy cell and treatment-vulnerable tumor cells are profiled, the real difficulties begin. To have a shot at building a profile for treatment resistant cells, even more data needs handling:

  • Treatment resistant tumor cell genomic DNA, harvested from patient after treatment failure
  • Treatment resistant tumor cell transcriptional RNA, SNPs and indels, measured immediately after treatment failure
  • Clonal data of retained resistant tumor cells for in vitro experimentation
  • Confirmation of treatment resistant tumor cell genetic information after clonal outgrowth in vitro; if it differs, entire process must restart from isolation
  • Cloned treatment resistant tumor cell transcriptional RNA before, during, and after treatment
  • Cloned treatment resistant tumor cell unique SNPs and indels before treatment
  • Cloned treatment resistant tumor cell molecular profile
  • Validation of cloned cell molecular profile by accurate prediction of cancerous pre-treatment cells which will later resist treatment

It’s all but certain that affiliated cellular and systems biologists will insist on gathering one last group of data to explain blasting behavior before closing out the data set: Treatment resistant tumor cell blasting factors—are oncogenic chemokines implicated, or do the cells blast without any signalling?

Future Transcriptomic Based Therapy?

There’s simply too much data to approach without one unified platform to guide research and subsequent therapeutic development. All of the normal molecular biology data caveats apply, except that when conducting this kind of research, new data sets– indels and genetic data before and after treatment– is mandatory rather than optional. Likewise, the cellular sample size required to form a believable molecular profile of such a sensitive state is bounded only by the grit of the laboratory staff. Given that running single cell transcriptomics doesn’t require much hands-on time, it’s clear that the future of oncology research is going to be rich in data. The clunky information systems that most labs use as a compromise aren’t going to cut the mustard.

Designed to Cure is the information technology platform for single cell transcriptomics research within oncology. Designed to Cure enables your team to handle and characterize thousands of molecular profiles, cellular states, SNPs, indels, and more. Contact us today to find out how you can use Designed to Cure to use big data built from the scale of individual cells to hunt down treatment resistant cancer cells before they can cause harm.

  1.  “Autophagy And Chemotherapy Resistance: A Promising Therapeutic Target For Cancer Treatment.” October 2013, https://www.nature.com/cddis/journal/v4/n10/abs/cddis2013350a.html.
  2.  “Single-cell Transcriptomics Uncovers Distinct Molecular Signatures of Stem Cells in Chronic Myeloid Leukemia.” https://www.nature.com/nm/journal/vaop/ncurrent/full/nm.4336.html.
  3. “The Cancer Sleeper Cell.” July 2014, https://www.nytimes.com/2010/10/31/magazine/31Cancer-t.html?pagewanted=4&_r=1&.
  4.  “Indel Detection From RNA-seq Data: Tool Evaluation and Strategies for Accurate Detection of Actionable Mutations.” July 2016, https://academic.oup.com/bib/article/2562816.
  5. “The Diverse Effects of Complex Chromosome Rearrangements and Chromothripsis in Cancer Development.” September 2015, https://link.springer.com/chapter/10.1007/978-3-319-20291-4_8.
  6.  “Combination of RNA- and Exome Sequencing: Increasing Specificity for Identification of Somatic Point Mutations and Indels in Acute Leukaemia.” December 2016, http://www.sciencedirect.com/science/article/pii/S014521261630220X.
  7.  “Full-Length RNA-seq From Single Cells Using Smart-seq2.” January 2014, https://www.nature.com/nprot/journal/v9/n1/full/nprot.2014.006.html.