Tag Archives: Sequencing

Diversity defines cancer.

Thanks to technological advances, we know that each person’s cancer is different. Just like a snowflake, no two are alike. This diversity between different individuals is termed intertumoral heterogeneity. Its corollary is that we need personalized therapy or individualized medicine, in order for cancer treatment to be effective.

Another layer of complexity is intratumoral heterogeneity, which means all cells in a single tumor are not necessarily the same (molecularly). Different groups of cells in a tumor (also called clones) likely have diverse molecular features. This is true in case of most cancers. Of these, glioblastoma is considered to be one of the most heterogeneous cancers.

An aggressive brain cancer, glioblastoma is very difficult to treat and recurs in most cases even after treatment. Out of every 100 patients with glioblastoma, 50 die in less than 15 months of diagnosis and very few live more than 5 years. An important reason for this dismal prognosis is the high degree of intratumoral heterogeneity. Individual cells within this tumor are different from each other, both genetically and functionally. Hence these cells respond to treatment differentially, making this tumor difficult to eradicate completely and more prone for recurrence.

Numerous previous studies have looked at genomic profiles of glioblastoma by analyzing chunks of tumors, each containing hundreds of thousands of tumor cells. One such landmark study, conducted by Verhaak and colleagues as part of The Cancer Genome Atlas (TCGA), used genomic analysis and found different tumors to have distinct genomic characteristics [1]. Based on these genomic profiles, they classified glioblastoma into 4 subtypes:

  1. Classical
  2. Mesenchymal
  3. Proneural
  4. Neural

These different subtypes of glioblastoma can each have variable degree of intratumoral heterogeneity. However, the diverse cellularity has never been systematically quantified. A recent study published in Science on June 12, 2014 does just that. Researchers from Broad Institute and Harvard use next-generation sequencing of individual cells in the tumor [2] and show that glioblastoma cells are far more heterogeneous than “previously thought”.

RNA-seq analysis of Glioblastoma

These researchers took 430 individual glioblastoma tumor cells isolated freshly from five different patients, and analyzed each cell by RNA sequencing (RNA-seq), an approach that involves profiling the transcriptome of the cell. The transcriptome includes all RNA in the cell – total RNA, messenger or mRNA, and other RNAs such as microRNA. Transcriptome sequencing or RNA-seq is a highly sensitive technique to detect genomic abnormalities commonly associated with cancer, such as gene fusion events or mutations. Change in expression of genes (either over-expression or decreased expressed) is an anomaly frequently seen in cancer; RNA-seq identifies gene expression levels in cancer cells as well.

RNA-seq glioblastoma
Single-cell RNA-seq reveals glioblastoma heterogeneity and may help design new, more effective therapies (Image Credit: cancer.gov)

In the Science study, RNA-seq analysis of glioblastoma cells revealed a high degree of cell-to-cell variability. Cells had different expression profiles of tyrosine kinase receptors, which are important targets for therapy. A direct clinical implication of this is that any single targeted therapeutic agent, no matter how effective will not kill all tumor cells. This provides a strong rationale for the use of combinations therapy for this and possibly other cancers.

This study also used RNA-seq to determine what state individual cells are in. Each tumor comprised cells in different states:

  • Some were differentiated, mature and hence sensitive to therapeutic agents,
  • Some were stem cell-like (glioma stem cells), had the potential for self-renewal and were resistant to most treatments, and
  • Some were in different intermediate states and showed variable responses to treatments

Considering this level of diversity, no single drug can completely kill all cells. Also, there are subtypes of cells that can reform the tumor after therapy. Hence, almost all glioblastoma tumors eventually recur even following the most aggressive therapies.

Researchers also determined which TCGA subtype (listed above) the individual tumor cells belong to. Surprisingly, every glioblastoma tumor was a heterogeneous mixture of cells from these different subtypes, pointing to the true diversity in tumor cells that we would miss when analyzing data from whole tumor chunks.

Clinical Implications

Important from a clinical standpoint, this study showed that increased heterogeneity in tumors is associated with poor prognosis (decreased patient survival). Heterogeneity thus has direct translational relevance and need to be considered for therapy. Bradley Bernstein from the Broad Institute, one of the senior authors on this study said in the press release, “Understanding the cellular landscape can provide a blueprint for identifying new therapies that target each of the various sub­populations of cancer cells, and ultimately for tailoring such therapies to individual patient tumors.”

This study is probably the first to quantify the extreme heterogeneity of glioblastoma. It reveals glioblastoma to be a formidable disease to manage clinically. While it underscores the challenge in successfully treating a cancer like glioblastoma, knowing this diversity helps us understand its basic biology. An ideal approach would be to leverage data on intratumoral heterogeneity to design new and effective therapeutic strategies against this deadly disease.


  1. Verhaak, R.G., et al., Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell, 2010. 17(1): p. 98-110. doi: 10.1016/j.ccr.2009.12.020
  2. Patel, A.P., et al., Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma. Science, 2014. DOI: 10.1126/science.1254257



Stop the presses! We have a new addition to the list of diseases that benefit from next-generation sequencing – infections.

In a case study published last week in The New England Journal of Medicine, routine medical and laboratory testing failed to identify the cause of encephalitis in a 14-year old patient [1], leaving him in a medically induced coma, with few treatment options and little hope. Encephalitis is brain inflammation, and can lead to severe neurologic abnormalities and death. Identifying the exact cause is critical for therapy but may be challenging. In this case, routine testing failed to provide a definitive diagnosis. Even a brain biopsy was inconclusive. As a last resort, doctors used a novel approach to figure out what was wrong with the patient. They analyzed his cerebrospinal fluid (CSF) using next-generation sequencing.

Next-generation sequencing infections
Sequencing for infectious diseases (Image Credit: Thomas Anthony Zampetti, Flickr)

Basically, researchers studied the CSF for evidence of microorganisms, in the form of DNA sequences. They used an unbiased approach to next-generation sequencing. I asked Charles Chiu, MD, PhD, Assistant Professor and Director of UCSF-Abbott Viral Diagnostics and Discovery Center, who is senior author on the study, about this approach.

“The term alludes to the fact that we are not targeting any specific pathogen or type of pathogen,” explained Chiu. It means that the researchers used sequencing and analysis to search for all known pathogens, including rare organisms.

Within 48 hours of receiving the CSF sample, next-generation sequencing and bioinformatics analysis revealed an obscure cause of encephalitis in this teenager – leptospirosis, an infection caused by the bacterium leptospira. Inability to accurately diagnose and treat this condition can be fatal. The good news: once diagnosed, leptospirosis is easily treatable with regular, old-fashioned penicillin. This antibiotic was administered in high doses, and the patient recovered completely.

Leptospirosis, Next-Generation Sequencing
Brain MRI of patient with encephalitis: Panels A, B, and C - images before treatment show signs of inflammation (arrows); Panel D - 7 days after penicillin treatment, shows inflammation resolved. From The New England Journal of Medicine, Wilson MR et al., Actionable Diagnosis of Neuroleptospirosis by Next-Generation Sequencing. Copyright © (2014) Massachusetts Medical Society. Reprinted with permission from Massachusetts Medical Society.

Next-generation sequencing identified the cause of infection in 2 days – something that months of traditional testing had not achieved. It saved the life of this teenager.

Clinical Translation of Next-Generation Sequencing

Next-generation sequencing is increasingly used in oncology for tumor profiling; in addition, it is a valuable tool for the diagnoses of various rare diseases and genetic disorders. The big question is: can successful diagnosis of an infection lead to the routine use of sequencing for other difficult-to-diagnose infections?

Charles Chiu believes so.

“I view this technology as being used as a broad-spectrum, second-line diagnostic assay after initial screening tests are negative and physicians have to resort to costly additional tests for rare and uncommon infections”, said Chiu.

However, this technology is far from clinic-ready. We need to come up with means to reliably identify disease-producing organisms, with high sensitivity and specificity, and differentiate them from the normal microbiome in relevant areas of the body. For example, the human gut contains 300-500 different species of bacteria [2]. So, one can appreciate the challenge of identifying an infection-causing microorganism from among the gut microbiome, by sequencing. Moreover, using next-generation sequencing routinely in the clinic for infectious diseases would require comprehensive testing and technology validation in order to obtain regulatory approval.

Clinically diagnosing infections (especially in cases of emergency) requires rapid sequencing and reliable analysis to deliver actionable results to the clinician. And all this needs to happen at affordable costs. Elaine Mardis, PhD, Professor of Medicine and Co-Director of The Genome Institute at Washington University, St. Louis, who was not part of The New England Journal of Medicine study agrees, “Probably the biggest hurdle is making it faster, cheaper and better than current assays.”

The Future

Once in routine clinical use, next-generation sequencing can prove critical for diagnosing cases of encephalitis and meningitis, like the one reported here. In addition, it will be beneficial for many zoonotic and infectious diseases that are difficult to diagnose using routine testing. Sequencing-driven diagnoses may be valuable, especially in critically ill patients with severe infections, including sepsis. In such cases, sequencing may not only identify the responsible microorganisms, but may also provide clues on drug resistance.

The successful use of next-generation sequencing by Chiu and colleagues provides us a rare window into a world where this technology can drive treatment decisions by diagnosing infections. As Elaine Mardis reflects, “(This study) beautifully illustrates how an unbiased look and smart bioinformatic analysis can provide answers that are life-saving.”

References Cited

  1. Wilson, M.R., et al., Actionable Diagnosis of Neuroleptospirosis by Next-Generation Sequencing. N Engl J Med, 2014. DOI: 10.1056/NEJMoa1401268
  2. Guarner, F. and J.R. Malagelada, Gut flora in health and disease. Lancet, 2003. 361(9356): p. 512-9. DOI: 10.1016/S0140-6736(03)12489-0