Transcript Rossitta Yung Presentation

Identification of Genes Expressed in
Malignant Cells That Promote Invasion
Authors: Jennifer Walter-Yohrling, Xiaohong
Cao, Michele Callahan, William Weber, Sharon
Morgenbesser, Stephen L. Madden, Clarence
Wang, and Beverley A. Teicher
Published in Cancer Research, December 15,
2003
Presented by Rossitta Yung
Date: March 25, 2004
Outline
Background
 Objective
 Methods
 Results
 Conclusion
 Critique
 Discussion

Cancer
Malignant tumor or forms of new
tissue cells that lack a controlled
growth pattern
 Invade and destroy normal tissue cells
 Tend to spread to other parts of the
body

Stromal Cells
Supporting cells for an organ
 made up of connective tissues (except
brain and spinal cord)
 E.g. endothelial cells – form cellular
sheets that cover line its cavities of
heart, blood vessels and serous of the
body, originating from mesoderm
 E.g. myofibroblasts – responsible for
wound closure after tissue injury

Stromal cells (cont’)
Actively contribute to tumor growth,
invasion and metastasis
 Often comprise a major portion of
solid tumors
 E.g. myofibroblasts:
- ECM proteins  host desmoplastic
response
- GFs  involved in tumor angiogenesis

Objective

investigate the genes expressed by
tumor cells that promoted interactions
between tumor cells and stromal cells
Materials
eight tumor cell lines:
1. MDA-MB-231, human breast carcinoma cell line
2. A-375, human melanoma cell line
3. LNCaP, human prostate carcinoma
4. A-549, human lung carcinoma
5. PC-3, human prostate carcinoma cell line
6. DU-145, human prostate carcinoma cell line
7. SKOV-3 cells, human ovarian carcinoma cell line
8. Mel624, melanoma cell line

-human adult dermal microvascular endothelial cells (HMVECs)
- human adult dermal fibroblasts (HDFs)  comprised of
myofibroblasts
Method – In vitro model

to determine whether tumor cell lines
could be grouped by their ability to
enable stromal invasion
Method – In vitro model






Added Matrigel to each well of a 24-well
plate
Polymerized
Removed a ~1mm plug
Filled with tumor cells
Added labeled endothelial cells (green) and
myofibroblasts (red)
Captured both fluorescent and bright field
images
Result
Result (cont’)
High [fluorescence]
Low [fluorescence]
Stromal Invasion Positive Stromal Invasion Negative
MDA-MB-231, human
breast carcinoma cell line
A-549, human lung
carcinoma
SKOV-3 cells, human
LNCaP, human prostate
ovarian carcinoma cell line carcinoma
A-375, human melanoma
cell line
PC-3, human prostate
carcinoma cell line
Mel624, melanoma cell
line
DU-145, human prostate
carcinoma cell line
Result (cont’)
MDA-MB-231 (human breast carcinoma cell line) and PC3 (human
prostate carcinoma cell line):
Method – SAGE Analysis

to obtain a comprehensive, unbiased
comparison of gene expression
between the human tumor cell lines
that underwent efficient invasion by
the myofibroblasts and endothelial
cells with those cell lines that did not
Method – SAGE Analysis
Method – Construction and
Analysis of SAGE library
Retrieved from either the Genzyme
proprietary database or the public
Gene Expression Omnibus database
 Normalized tag counts to 50,000 total
library counts for each library
 GeneSpring software: an initial
hierarchical clustering

Result (cont’)
Result (cont’)

1.
-
2.
-
2 statistical tests were used to extract
tags with significant differential
expression
T test
For group comparison
Unable to eliminate tags with very low
counts in one group and zeros in the other
group
Chi square test
For the comparison of the averages of
each group
Result (cont’)
T test: significant value P≤ 0.05
 Chi square test: significant value ≤ 0.1


99 SAGE tags emerged
Result (cont’)
Result (cont’)

a statistical comparison was carried
out on the eight SAGE libraries that
exceeded 20,000 tags

87 SAGE tags emerged
Result (cont’)
Method and Result – Quantitative
PCR and In Situ Hybridization
to validate the differential gene
expression profiles identified by
analysis of the SAGE libraries
 30 genes were selected for real-time
PCR examination in the eight tumor
cell lines
 9 of the genes analyzed by real-time
PCR showed good correlation with
normalized SAGE tag counts

Result
- Strongest Correlation:
 bone marrow stromal antigen 2
(BST2 or HM1.24)
 stathmin-like 3 (STMN3)
 tumor necrosis factor receptor
superfamily member 5 (TNFRSF5)
 hepatocyte growth factor tyrosine
kinase substrate (HGS or HRS)
Result (cont’)
Result (cont’)

in situ hybridization for BST2 and
STMN3 was performed on samples of
metastatic and nonmetastatic ovarian
cancer
Result (cont’)
Result (cont’)

high levels of both BST2 and STMN3
were observed in the metastatic
ovarian cancer tissue than the
nonmetastatic ovarian cancer tissue
Conclusion

the combination of the stromal
invasion assay amd SAGE analysis
appears a useful tool for identifying
genes involved in tumor-stromal
interactions

the genes identified through this
analysis may be useful as therapeutic
and/or diagnostic targets
Critique
No mention about the selection of tumor cell lines
for the experiment
 Figure 1: no mention about the significance of
bright field images
 How do they perform the two statistical tests???
 How do they select the 30 genes for RT-PCR???
 Why do they choose BST2 and STMN3 for in situ
hybridization???
 Personally, I am not familiar with the histochemical
pictures in Figure 6

Discussion

Can results imply to other
experimental samples or conditions???
(e.g. other tumor cell lines, other
stromal cells)

In-vivo model???