Page 18 - Read Online

suggested that these drug-resistant cells were TICs-like PDX models established from the very same patients
and could be responsible for tumor recurrence. on trial are being treated ahead of patient therapy or
concurrently, and results from the mouse trial is provided
Patient-derived Xenograft Models for in real-time to help guide clinical management of the
Pharmacology and Biomarker Studies patient’s tumor. Further powered by the molecular

Traditionally, pharmacology, biomarker and pharmacokinetics/ characterization of the tumors, this highly personalized
pharmacodynamics studies for oncology programs almost approach has the potential to revolutionize the drug
[77]
exclusively relied on tumor xenograft and to a much lesser development and patient care. For example, a recent
[78]
degree, syngeneic models. With the signifi cant increase in the study by Stebbing et al. reported 22 sarcoma PDX
availability and affordability of PDX models offered by both models were successfully established from 29 patients
academic institutions and contract research organizations, (76% take rate) and screened for drug sensitivity to a
PDX models have seen increasingly their utility in routine panel of therapeutic agents. The entire process typically
research activities. A quick survey of oncology discovery took 3-6 months depending on individual tumor growth
programs published in the past 3 years shows that increasing characteristics and treatment regimen. Of the 22 patients,
number of programs use PDX models at some point during the 6 died before data became available. Of the 16 remaining
preclinical discovery and translational research stages. [14,65-67] patients, 13 (81%) demonstrated a correlation between
In addition, there is an industry-wide trend to include the results from their PDX mouse trial and clinical
PDX model readout as a key component of the required outcome. Similar approach has also been reported in
[79]
[80]
data package for both internal use as well as regulatory advanced adenoid cystic carcinoma, ovarian, and
[81]
submission. The history of using incorporating PDX models other cancer types. The current data, although limited,
in drug discovery can be traced back to several decades ago. appears to support the use of PDX models to prioritize
For example, one of the earliest reports involving cancer therapeutic agents against individual tumors. However,
drugs and PDX models by Fiebig et al. studied a number some key challenges remain before this strategy can be
[68]
of chemotherapy drugs at their respective maximal tolerated broadly implemented in clinical practice. For example,
doses (MTDs) in PDX models derived from 34 patients, and establishment of PDX models is still a technically
demonstrated 92% accuracy in predicting effi cacy and 97% in challenging and time-consuming process, even after
predicting no-response. Similar predictive value was seen in a much progress has been made to improve the take rate
later study by the same group. However, additional studies and optimize the expansion scheme. In addition, the
[69]
suggest that the predictive value can fl uctuate due to factors algorithm for the selection of agents to be tested needs
such as tumor histology and location, stage of disease from to be further developed and refi ned. Lastly, to effectively
which the models are derived, the quality of PDX models, demonstrate the feasibility and clinical benefi t of the
sample size and dosing regimen. [64,70,71] In addition to selecting PDX-guided treatment prioritization in the patient care
models that are histologically, molecularly and genetically setting, properly controlled clinical trials are needed.
relevant to the patients in clinical, another important factor
for improving translatability of preclinical fi ndings is the Limitations of Patient-derived Xenograft Models
drug exposure. Not surprisingly, preclinical model species, in Although PDX models present an exciting opportunity
most cases immunocompromised mice, can exhibit different for improving predictive value of preclinical and
tolerability and adsorption, distribution, metabolism and translational studies, and offer a number of advantages
excretion property than those in human. It is commonly seen over conventional cell line xenograft models, just
that drug exposure levels at MTD dose in mice are higher like any other preclinical model platforms, there are
than clinically achievable levels in human. Therefore, a several limitations that one needs to be aware of. First,
[72]
compound given at mouse MTD to xenograft, allograft or the utilization of severely immune-compromised host
syngeneic models may generate exaggerated effi cacy that mouse strains, particularly the nonobese diabetic severe
over-predicts human response in the clinic. This phenomenon combined immune defi ciency gamma mice, while
has been seen for both chemotherapy agents [12,73,74] as well allowing higher take rate and more consistent growth of
as targeted agents such as vascular endothelial growth factor xenografted human tumors, is inherently inadequate in
receptor inhibitors and PI3K inhibitors. A key concept and modeling immune responses. Although human stroma
[75]
practice to avoid the pitfalls of using mouse MTD dose and components including immune cells originally present in
exposure as the sole basis for effi cacy prediction is to use the tumor biopsy can be grafted together with the tumor
CRD or clinically relevant exposure (CRE) whenever a CRD tissue, they normally cannot survive beyond the fi rst
[82]
or CRE can be determined. passage, and will be completely lost in the subsequent
expansion. The other stroma components including
[83]
Patient-derived Xenograft Models for Mouse fi broblasts and vasculature are quickly replaced by
Clinical Trial
murine counterparts. The lack of functional immune
[83]
An evolving concept and practice, PDX mouse clinical system limits the utility of these models in studies
trial, has started to yield positive results that had where immune responses are required. For example,
real-life impact on selected patients. In this setting, immunotherapy cannot be readily studied in the PDX
[76]
Journal of Cancer Metastasis and Treatment ¦ Volume 1 ¦ Issue 1 ¦ April 15, 2015 ¦ 11