| Dec
1, 2005 |
| By:
Michelle
Chen, Kaho
Minoura, Siqun
Wang, Tetsuo
Noda, Tetsuichiro
Muto, Yoshio
Miki |
| Pharmaceutical
Discovery |
|
Clinical biopsy samples of esophageal cancer and surrounding normal
cells were excised by laser capture microdissection and analyzed with DNA
microarrays for global gene expression studies. Genes involved in the
keratin biosynthetic pathway were found to be significantly affected in
these squamous cell carcinomas. Our findings provide a plausible mechanism
for the alteration of keratin synthesis observed in over 95% of hereditary
squamous cell carcinomas in esophageal cancer. This study also
demonstrates the feasibility of using a minimum amount of clinical sample
to decipher complex disease and disease processes at the molecular level.
Esophageal cancer is the sixth leading cause of cancer deaths
worldwide. Each year in the United States alone, 13,900 new cases are
diagnosed and 13,000 patients die from this lethal disease. However,
little is known about the pathogenesis or specific molecular pathways that
lead to the development of this type of cancer. To increase our
understanding of this disease and its etiology, we compared global gene
expression profiles of microdissected cells obtained from esophageal
cancer and surrounding normal tissues.
Figure 1. LCM of Esophageal Cancer
Tissue. Biopsy samples were cut into thin sections and stained.
The top panel shows a biopsy sample prior to microdissection and
the lower panel shows those regions of cells that were removed.
Dotted yellow lines indicate the regions where cancerous cells
were removed.
|
The microarrays (Agilent Technologies, Palo Alto, CA) used in these
experiments contain in situ-synthesized, oligonucleotide probes that
represent more than 17,000 well-annotated human genes. Although DNA
microarray technology is a powerful tool for molecular analysis, the
successful use of this technique can be challenging in clinical studies
where adequate amounts of biopsy tissue are difficult to obtain. Recent
developments in sample excision, such as laser capture microdissection
(LCM) employed in this study, can now overcome such limitations.
Experimental
Esophageal biopsy tissue samples were taken from four patients who were
diagnosed with different stages of esophageal cancer (stages I, IV, IV,
IVB). The samples were processed, embedded, and stained in thin sections
(Figure 1). LCM-extracted cells from biopsy samples were examined and
extracted cells from normal esophageal tissues were used as controls.
Figure 2. Bioanalyzer profiling of
total RNA input and cRNA targets. Electropherograms represent
total RNA from normal (2A) and tumor (2B) cells after dilution and
analyzed with the Agilent 2100 bioanalyzer using the RNA 6000 pico
kit. Cyanine-3 labeled cRNA targets from normal cells and
Cyanine-5 labeled cRNA targets from tumor cells were shown in (2C)
and (2D), respectively.
|
Total RNA from both normal and tumor cells was isolated from each sample
and evaluated to ensure RNA integrity prior to amplification (Fig. 2,
panels A and B). Isolated RNA concentrations ranged between 96-120 ng/mL
with OD 260/280 values between 1.87-1.96. RNA samples from the four
patients were pooled, amplified, and fluorescently labeled by
incorporating cyanine-3- or cyanine-5-tagged CTP in the amplification
reaction. As little as 50 ng of total RNA was used in these reactions.
Three independent amplification and labeling protocols were compared: The
Low RNA Input Fluorescent Linear Amplification kit, The Fluorescent Linear
Amplification kit (Agilent Technologies), and a two-round protocol (Tokyo
University Institute of Medical Sciences [IMS]).
Figure 3. The Detection of
Differentially Expressed Genes in Esophageal Cancer. Upregulated
genes are shown in red while downregulated genes are shown in
green.
|
Yields of cRNA were determined by UV
spectrophotometry and the amplified and labeled RNA was also assessed for
quality. The resulting profiles revealed a successful amplification for
the RNA samples (Fig. 2, panels C and D). The fluorescence of the
cyanine-3 sample (Fig. 2C) was lower than that of the cyanine-5 sample
because additional fluorescent emissions from the toluidine blue dye
originally employed to stain the tissue samples overlapped with the
wavelength of cyanine-5 fluorescence.
Table I: Experimental Details
|
Samples (500 ng) of quality-controlled
cRNAs were hybridized to human 1A microarrays consisting of in situ-synthesized,
60-mer oligonucleotide probes that represent more than 17,000
well-annotated human genes. Following hybridization, the arrays were
processed, dried under nitrogen, and scanned. The image pattern data was
extracted using feature extraction software and processed for further
analysis, such as for the elucidation and comparison of expression
profiles. A typical profile comparing the cRNA targets generated from
esophageal cancer cells with those from normal cells is shown in Figure 3.
Table 1 provides a detailed list of materials, equipment, and experimental
procedures.
Results and Discussion
Figure 4. Log Ratio Comparison of
Two Replicate Arrays Demonstrate Excellent Reproducibility. Data
analysis performed using Resolver® Gene Expression Data Analysis
System (Rosetta Biosoftware, Seattle, WA).
|
In this expression profile study, a
total of 3,463 signature genes were identified. Of these, 1,865 were
downregulated and 1,598 were upregulated (Figure 3). Signature genes
represent the sum of both up- and downregulated genes designated by the
software program as significantly above background. These results were
highly reproducible for all three amplification and labeling methods
tested. Figure 4 shows the data obtained from replicate microarray
analyses plotted along the x- and y-axes, respectively. The high
correlation coefficients (> 0.95) indicate excellent reproducibility
for these results. Comparable results were obtained with all three
labeling and amplification methods employed.
Table II. Examples of Highly
Downregulated Genes in Esophageal Cancer
|
We examined the top up- and
downregulated genes in these experiments and found that many of the up-
and downregulated genes in tumor cells are related to keratin synthesis or
keratinocyte differentiation. Some examples of the genes exhibiting the
greatest degree of downregulation and their associated functions are
listed in Table 2. The fold-change shown is an average from nine arrays
that used the three different labeling and amplification schemes
described. This finding suggests that there is a deregulation of keratin
synthesis and provides evidence for the reported link between esophageal
squamous cell carcinoma and the alteration of keratin synthesis (1).
Conclusions
This series of experiments identified
a unique pattern of both decreased and increased gene expression in
esophageal cancer compared with surrounding normal tissue. Many of the
affected genes are also involved in keratin synthesis or keratinocyte
differentiation. These findings provide a plausible molecular basis for
inferring that the alteration in keratin synthesis, observed in over 95%
of cases reporting this rare, autosomal dominant disorder, may also
predispose patients to the only known familial squamous cell carcinoma
(1).
The study also demonstrated the
feasibility of performing global gene expression profiling studies with as
little as 50 ng of RNA. Techniques such as LCM, used in conjunction with
appropriate amplification, labeling, hybridization, and data extraction
materials and methodologies, can now overcome the obstacles of working
with size-limited, clinical biopsy samples.
Acknowledgment
The authors would like to thank Steve
Kain for his comments.
Michelle Chen, Kaho Minoura, and
Siquin Wang represent Agilent Technologies, Inc. Headquarters: 5301
Stevens Creek Blvd., MS 53U-WG, Santa Clara, CA 95051; phone:
408-553-7006, fax: 408-553-7100; email: michelle_chen@agilent.com
Tetsuo Noda, Tetsuichiro Muto, and
Yoshio Miki represent the Cancer Insitute at the Japanese Foundation for
Cancer Research, Tokyo, 130-8455, Japan
Reference
1. J.M. Risk, H.S. Mill, J. Garde et
al., Dis Esophagus 12, 173-176 (1999).
All correspondence should be
addressed to Michelle Chen, Agilent Technologies, 5301 Stevens Creek
Blvd., MS 53U-WG, Santa Clara, CA 95051; phone: 408-553-7006, fax:
408-553-7100; email: michelle_chen@agilent.com
|
