Transcript Genomic signatures to guide the use of chemotherapeutics

Genomic signatures to guide the use of
chemotherapeutics
Authors: Anil Potti et. al
Presenter: Jong Cheol Jeong
Motivation
 What will be happened if ineffective
chemotherapy is used?
Increasing the probability of side effects
Decreasing the quality of life
Purpose
 Developing gene expression signatures which
predict responses to various cytotoxic
chemotherapeutic drugs.
 Giving us the direction for using cytotoxic agents
which best matches the characteristics of the
individual.
Outline
 Method
 Results
 Conclusion
Method
NCI-60: composed with 60 cell line and the sensitivity to 5084 compounds
Sensitivity: exposing each cell line to each compound for 48hours,
assessing the growth inhibition by sulforhodamine B
Method
 Using the cell line in the NCI-60 Panel
- 60 cancer cell line: sensitivity of 5084
compounds
1) Identifying cell line: most resistant or sensitive
to docetaxel
2) Identifying genes: their expression correlated
most highly with drug sensitivity
3) Bayesian binary regression analysis with
LOOCV
Results
Cell lines from NCI-60
Blue: lowest expression
Red : highest expression
GI 50 or IC50 The concentration of compound requiring 50% growth inhibition
LC50 The concentration of compound requiring 50% cytotoxic
Results
Validation of docetaxel response prediction model
30 lung and ovarian cancer cell lines
29 lung cancer cell lines
Significant correlation between predicted probability of docetaxel sensitive and IC50
Results
Applying
a Mann-Whitney
U-test
showing the
capacity of the
predictor
Results
Results
(developing series of expression profile from NCI-60)
Results
(developing series of expression profile from NCI-60)
Results
(developing series of expression profile from NCI-60)
Results
(developing series of expression profile from NCI-60)
Result
(predicting response of combinations of drugs)
 4 cytotoxic agents: paclitaxel, 5-FU, adriamycin,
and cyclophosphamide
 51 cell lines: 13 responders, 38 nonresponders
Individual chemosensitivity predictions
Result
(predicting response of combinations of drugs)
Statistically significant distinction between the responders and nonresponders
Result
(predicting response of combinations of drugs)
Breast cancer with 45 cell lines
38 responders
11 nonresponders
Result
(predicting response of combinations of drugs)
Kaplan-Meier survival analysis
PPV: Positive Predicted Value
NPV: Negative Predicted Value
FAC adjuvant chemotherapy
Blue: sensitive
Red: resistant
Result
(patterns of predicted chemotherapy response)
Step1. Chemotherapy response predictors calculates the likelihood of sensitivity to
the seven agents in a large collection of samples
Ex) breast, lung, and ovarian tumor
Step2. Clustering the samples according to patterns of predicted sensitivity to the
various chemotherapeutics and plotted a heatmap
Respond to 5-FU are resistant to Adriamycin and Docetaxel:
suggesting possibility of alternate treatments
Red: high probability of sensitivity of response
Blue: low probability of resistance
Result
(linking chemotherapy sensitivity to oncogenic pathway status)
Someone who initially responds to a given agent is likely to
eventually suffer a relapse; therefore the development of gene
expression signatures that reflect the activation of several
oncogenic pathways are needed
Step1: stratifying the NCI cell lines based on predicted docetaxel response
Step2: examining the patterns of pathway deregulation associated with docetaxel
sensitivity or resistance
Result
(linking chemotherapy sensitivity to oncogenic pathway status)
17 lung cancer cell lines
Red: high probability of sensitivity of response or activation
Blue: low probability of resistance or deregulation
Significant relationship between phosphatidylinositol 3-OH (PI3)-kinase
pathway deregulation and docetaxel resistance.
- Giving an opportunity to use a PI3-kinase inhibitor in this group
Conclusion
 The signature of chemosensitivity generated from
the NCI-60 panel have the capacity to predict
therapeutic response in individuals receiving
either single agent or combination chemotherapy
References
 Staunton, et. Al. “Chemosensitivity prediction by
transcriptional profiling”, PNAS, 98-19, 1078710792, 2001
 Potti, A. “Genomic signatures to guide the use of
chemotherapeutics”, Nature Medicine, 12-11,
2006