The use of cancer cell lines for in vitro studies of anti-cancer agents remains a primary component of preclinical drug development. These studies, however, struggle with significant pitfalls given the unique biology of cancer cell lines and the artificial assay format. This can lead to a failure of the research model to predict a drug’s potential clinical efficacy in a complex organism. To better capture organismal biology, there has been a shift towards utilizing xenograft models. Mouse xenograft studies, in which human tumors are grown in an immune-deficient mouse, have aided in bridging the gap between cell lines and the clinic in translational research.

Xenograft models come in three primary flavors:

  • Cancer cell line xenografts (CCLX): Cancer cell lines are implanted into immune deficient mice and allowed to form a tumor mass
  • Patient derived xenografts (PDX): Patient tumors are directly implanted into immune deficient mice and allowed to form a tumor mass
  • Humanized PDX: The human immune system is implanted into otherwise immune deficient mice along with the patient tumor, enabling research on the impact of the immune microenvironment on tumor growth

Early work by White et al (1989) utilized a xenograft model generated by the heterotransplantation of human retinoblastoma cells into the anterior chamber of the immunodeficient mouse eye. By evaluating the response of the tumor to six different chemotherapeutic agents this paper showed three of the agents were effective, while the other three were not. This and other similar studies laid the foundation for today’s PDX studies. Jumping forward a few decades, PDX options have greatly expanded, with consortiums, private collections, and CRO providers offering a wide variety of PDX models (listed below).

The power of PDX is demonstrated utilizing one of these large repositories in a study by Townsend et al (2016). The study is based on the hypothesis that large randomized drug trials in PDX models could help inform and improve the success of human clinical trials. Multiple PDX can be generated from a single tumor, allowing treatment to be randomized across the mice derived from that single patient tumor. Twenty-four different P53 WT, B Cell acute lymphoblastic leukemia PDX models were treated with the MDM2 inhibitor CGM097. Results show significant therapeutic efficacy and enabled the identification of candidate clinical biomarkers. Similar studies may reduce the time and expense of drug development, while mitigating the potential adverse effects of testing ineffective drugs in patients based only on cell line screening.

Beyond large drug studies, the concept of a mouse PDX avatar for cancer patients has also been proposed. In this model, multiple drug combinations can be tested on a PDX model derived from an individual’s own tumor for personalized efficacy. A recent article in The Scientist described a PDX avatar case study with a rare sarcoma patient for which there was no standard of care treatment. Despite the drawbacks to this method, including: the take rate of the tumor in the mice, the time it takes to expand the model to statistically power a study, and cost, an effective combination therapy was discovered in the mouse model. This avatar approach ultimately directly informed treatment decisions, resulting in the same positive result for the patient as seen in the mouse.

The field will continue to evolve, but PDX models are here to stay, as a powerful tool to study specific tumors in a highly controlled, organismal environment. However, these models remain precious, costing significant time and money to run even a single study. Therefore, it is imperative to get as much high quality data from these studies as possible. Adding NanoString’s 3D Biology™ Technology to a PDX study enables researchers to analyze DNA, RNA, protein, and phospho-protein expression, rapidly, producing reproducible, high-plex, multi-omic data from small sample inputs. 3D Biology Technology can be applied to PDX models to provide valuable insight into PDX model heterogeneity and the molecular determinants of therapeutic response, resulting in improved combination therapies and clinically relevant biomarkers. Click below to learn more about NanoString’s 3D Biology™ Technology and current PDX market leaders.

Learn more about NanoString 3D Biology Technology

Current large PDX suppliers: Large collections of PDX models have now been established by multi-institutional consortiums, (EuroPDX and PRoXe), large “private” collections (Novartis), institutions that provide access based on various collaborative and perhaps most commonly CRO-type pre-clinical service models including Taconic, Charles River Laboratories, The Jackson Laboratory and Champions Oncology among others.


FOR RESEARCH USE ONLY. Not for use in diagnostic procedures.


White, et al. (1989) Evaluation of response to chemotherapy in retinoblastoma heterotransplanted to the eyes of nude mice. Cancer Chemother Pharmacol. 23(2):63-67.

Townsend et al. (2016) The Public Repository of Xenografts Enables Discovery and Randomized Phase II-like Trials in Mice. Cancer Cell 29, 574–586 


Share this post

Posted by Doug Hinerfeld