New opportunities for cell line generation at scale

The limited availability of appropriate disease models is a major bottleneck that limits research impact. If model systems representing the genetic diversity of diseases such as cancer were readily available, experiments to understand the biological consequence of particular gene mutations and to develop new therapeutic strategies could proceed more rapidly.

Until recently, despite enormous effort dedicated to the creation of cancer cell line models, success rates have been very low (often zero). However, exciting technology developments have recently changed the landscape of possibilities for systematic approaches to cell-line generation. These include protocols such as the Conditionally Reprogrammed Cells protocol developed by Richard Schlegel and colleagues at Georgetown University and the development of three dimensional organoid cultures by multiple groups including Hans Clevers at the Hubrecht Institute. Excitingly, many labs around the globe are starting to use these new methods.

There is great opportunity to come together to develop and deploy industrial strength SOPs to create models of every major cancer type at scale and make new models and methods broadly available. If hospitals and centers for cell line creation "factories" joined forces, the community could create large collections of genomically characterized disease models at scale and determine best practices to ensure open, unfettered access to the new models for the scientific community.



The mapping of cancer genomes is nearly complete. With this information in hand, the promise of developing and deploying customized cancer therapies that match the genetics of each particular patient's tumor has become the central goal for the translational cancer research community. While there is much that we can learn directly from cancer patients and their responses to therapeutics, a complementary approach of generating large-scale laboratory information is necessary to accelerate progress. We must actively manipulate cancer models (pieces of tumors that are propagated in the laboratory) to determine how the genes and mutations drive cancer vulnerabilities.

Over the last 50 years, the worldwide cancer research community has generated about 1000 cancer models. For decades, this seemed like a sufficiently large number to model many of the types of cancer that were known. But, now that we more fully understand the incredible genetic diversity of cancer it is clear that many thousands of new models are needed to ensure that faithful models of the mutations present in each cancer patient exist.

In addition, despite enormous effort dedicated to the creation of cancer cell line models, success rates for certain cancer types have been very low (often zero) and genetic factors promoting or restricting success are completely unclear. For instance, standard approaches nave netted almost zero successes in prostate cancer or neuroendocrine tumors. Even for cancer types in which many cell-line models exist (e.g. 150 lung cancer cell lines), many common mutations are poorly represented in existing cell lines and the complex genetic combinations found in individual patients are not present in any existing model.

An ecosystem of cancer cell line factories to support a cancer dependency map.

Boehm JS, Golub TR. Nat Rev Genet. 2015 Jun 16. doi: 10.1038/nrg3967

Read Manuscript


What's needed: a Cell Line Factory to serve the community

We have launched a pilot Cell Line Factory project at the Broad Institute, together with hospital partners, to explore how best to overcome current obstacles, produce faithful models at scale and explore the ethical issues associated with enabling rapid and unfettered access. Our pilot effort has 4 goals:

  1. Develop industrial strength protocols. For each cancer type and subtype, it will be necessary to establish robust SOPs and use these to create 100s of new cell lines to appropriately represent the genomic diversity of tumors. In addition, adhering to industrial strength processes (including standardizing and iteratively improving lab protocols, tube barcoding and molecular fingerprinting to avoid sample mix-ups, etc) will be necessary.
  2. Comprehensively characterize genomes. Protocols should be optimized based on tumor genomics. We will subject the primary tumor sample, matched normal tissue and associated cell lines to genome, exome or transcriptome sequencing. Recent large scale projects such as the Cancer Cell Line Encyclopedia and the Sanger-MGH Cell Line Profiling effort have clearly demonstrated the immense value of comprehensive characterization of cancer cell lines. Performing experiments in cell lines with such characterization will enable the rapid discovery of genetic biomarkers that predict tumor behavior (for instance drug sensitivities). We aspire to make de-identified data publicly available and easily accessible on rapid timeframes.
  3. Enable rapid and unfettered access. Pending appropriate patient consents and IRB approval, any lab should have immediate and unfettered access to new cancer cell lines that enable their research. In addition, individual sample contributors will likely want to draw upon models derived from other cancer types (either as a comparison population or since they harbor similar genetic events). Just as 3rd party non-profits such as Addgene and ATCC have provided highly enabling services for the rapid, accurate and cost-effective distribution of plasmids and existing cell lines, we envision that cell line distribution might be handled by an appropriate professional entity.
  4. Prioritize treatments for each cancer patient. Looking forward, if SOPs can be established that faithfully represent the biology of tumors (either on a patient-by-patient basis or in aggregate across 100s of models), it might be possible to use patient-derived models to test therapeutic hypotheses predicted by the tumor genome. Much work will be needed to overcome technical challenges and explore related ethical issues.

Current status of the Cell Line Factory:

Our pilot effort has made significant progress:

  • We are currently collaborating with over 16 clinical teams across multiple hospitals to execute pilots in 16 cancer types (Prostate, Glioblastoma Multiforme, Brain Metastases, Head and Neck, Thyroid, Neuroendocrine, Colon, Pancreas, Lung, Gastric, Esophageal, Pediatric, Breast, Sarcoma, Ovarian, Kidney, T-cell lymphoma and Rare tumors such as Chromophobe Renal Carcinoma, Chordoma and LAM-related malignancies).
  • We are scaling up efforts in Pediatric cancers, Brain cancers and Upper GI malignancies based on high initial success rates.
  • We've processed over 800 samples from our collaborators corresponding to over 400 cancer patients (as of 4/1/2016).
  • We continue to observe take rates nearing around 50% for culturing samples through five passages. We have completed sequencing of the first 200 samples and have validated the creation of over 70 new cancer models.
  • We are actively looking for new collaborators to submit appropriately consented and cryopreserved samples for cell line initiation (including from other diseases!). Please contact us for more information.
  • Register on our site to take a look at our preliminary list of verified tumor cell lines and to stay up to date with project news.
  • After a great response to our RFP for distribution, we are working on finalizing a partner and hope to make new verified lines available to the community later this year.

    Broad CCLF Phase 1 RFP

    Broad CCLF Phase 1 Principles



To collaborate with the Cell Line Factory team to generate new models, please contact (please note, due to the pilot nature of the project, we are not yet established a framework for distribution. We appreciate your patience - please stay tuned!)

Send Email