Advertisement
MCP
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/mcp.M700115-MCP200 on October 13, 2007.
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
M700115-MCP200v1
7/1/163    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Glossary
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Song, X. C.
Right arrow Articles by Zhou, G. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Song, X. C.
Right arrow Articles by Zhou, G. W.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Molecular & Cellular Proteomics 7:163-169, 2008.
© 2008 by The American Society for Biochemistry and Molecular Biology, Inc.

Research

Protein Expression Profiling of Breast Cancer Cells by Dissociable Antibody Microarray (DAMA) Staining*,S

X. Cynthia Song{ddagger},§, Guanyuan Fu{ddagger},§, Xufen Yang{ddagger}, Zhong Jiang, Yingjian Wang|| and G. Wayne Zhou{ddagger},**

From the {ddagger} Department of Biological Sciences, Louisiana State University, Baton Rouge, Louisiana 70803, the Department of Pathology, University of Massachusetts Medical School, Worcester, Massachusetts 01605, and || Hypromatrix Inc., Worcester, Massachusetts 01606


   ABSTRACT
TOP
 ABSTRACT
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
Dissociable antibody microarray (DAMA) staining is a technology that combines protein microarrays with traditional immunostaining techniques. It can simultaneously determine the expression and subcellular location of hundreds of proteins in cultured cells and tissue samples. We developed this technology and demonstrated its application in identifying potential biomarkers for breast cancer. We compared the expression profiles of 312 proteins among three normal breast cell lines and seven breast cancer cell lines and identified 10 differentially expressed proteins by the data analysis program DAMAPEP (DAMA protein expression profiling). Among those proteins, RAIDD, Rb p107, Rb p130, SRF, and Tyk2 were confirmed by Western blot and statistical analysis to have higher expression levels in breast cancer cells than in normal breast cells. These proteins could be potential biomarkers for the diagnosis of breast cancer.

Protein microarrays have recently been attracting great attention for their potential use in high throughput studies of protein function (1–6). The ultimate goal of developing this technology is to construct ordered arrays of individual proteins for biochemical study at the molecular level. A number of different sources of peptides and proteins have been used for protein microarray manufacture, including synthetic peptides (7, 8), recombinant proteins (9–11), and monoclonal and polyclonal antibodies (12–14). Microarrays have been utilized to study protein expression profiles (15–17), protein-protein interactions (18, 19), drug analysis (11, 20), and the diagnosis of diseases such as cancer, food allergies, and infection by pathogenic viruses and bacteria (21–24). Most of the microarrays use the capture microarray platform (3–5). In a standard capture microarray procedure, an array of proteins is immobilized on a membrane or a glass slide to capture protein ligands from a protein mixture; the captured ligands are detected either with a different set of labeled antibodies (15–17) or a detectable tag attached to the ligands (25).

We have developed a different protein microarray platform, dissociable antibody microarray (DAMA)1 staining (26). This technology combines the power of immunohistochemical staining and the parallel analysis of antibody microarrays. DAMA staining provides a new approach in the global analysis of protein expression and subcellular localization. In DAMA staining, targeted cells are grown on a coverslip in a culture dish or mounted on a coverslip and fixed and permeabilized with the standard protocol for traditional immunostaining. In the next step, instead of adding a primary antibody, an array of antibodies immobilized on a membrane is placed on top of the specimen. Pressure is applied to maintain close contact between the antibody array and cells. During incubation, antibodies dissociate from the array support and bind to their respective antigens in the cells without significant lateral diffusion. In this way, hundreds of antibodies are delivered to the targeted cells or tissues in a position-dependent manner. The bound antibodies are then detected either with enzyme- or fluorophore-conjugated secondary antibodies. The expression profiles of hundreds of proteins can thus be determined simultaneously. Furthermore when stained with fluorophore-conjugated secondary antibodies, subcellular localization profiles of hundreds of proteins can be obtained from a single staining by a fluorescence microscope equipped with a computer-controlled motorized stage.

Here we report the development of DAMA staining technology and its application in identifying potential biomarkers for breast cancer. Breast cancer is the most common form of cancer in women with ~210,000 new cases diagnosed annually in the United States alone (27). This disease causes significant mortality, accounting for ~40,000 deaths in the United States per year (28) and many more fatalities worldwide. Earlier detection and better treatment will improve prognosis and survival of the disease (27). Detection of new molecular biomarkers will not only be very useful for breast cancer diagnosis but also may identify novel and key targets for therapy. In this report, we compared the expression profiles of 312 proteins among 10 different normal breast and breast cancer cell lines using DAMA staining. We also developed a data analysis program to identify differentially expressed proteins and validated the prediction by Western analysis in the same cell lines.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Preparation of Antibody Microarray Array-320—
Array-320 contains 312 antibodies in a 16 x 20 format. The antibodies were spotted on a membrane by a robotic arrayer from Gesim (Dresden, Germany) using piezoelectrically driven microdosage heads. This array has an overall size of 44 (width) x 42 mm (height) with 2 mm between different antibody spots. Each spot contains 50 ng of antibody and is ~500 µm in diameter. The antibody list for Array-320 is shown in the supplemental materials. Antibodies were selected from Hypromatrix's collection. All antibodies have been characterized and demonstrated to bind their targets in various assays.

Cell Culture—
Ten different breast cell lines, MCF10A, MCF12A, Hs578Bst, MCF7, T-47D, ZR-75-1, MDA-MB-231, BT549, Hs578T, and MDA-MB-435S, were purchased from American Type Culture Collection (ATCC). Cells were maintained and propagated as recommended by the ATCC and were grown on 10-cm cell culture dishes until 90–95% confluence for DAMA staining.

The DAMA Staining Procedure—
Cells were fixed with formaldehyde solution, permeabilized with Triton X-100, and blocked with goat serum. A 320-antibody microarray (Array-320) was placed over the cells and was incubated at room temperature under the optimal pressure of ~100g/cm2 as measured by a pressure detection sensor in an Economical Load and Force (ELF) single handle system (Tekscan, Inc.). Bound antibodies were detected by alkaline phosphatase-conjugated secondary antibodies (both goat anti-rabbit and goat anti-mouse) with 1-stepTM NBT/BCIP (nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate) substrate (Pierce). Images of protein expression profiles were scanned by an HP Scanjet 4890 and processed by Photoshop for intensity integration.

Intensity Integration and Initial Data Analysis—
The ScanAlyze program developed for DNA microarray analysis was used for intensity integration of the DAMA staining images. One image (e.g. the reference cell line, such as MCF10A) was used as the channel 1 data, and the other image (e.g. other sample cell lines, such as MCF7) was used as the channel 2 data. The scanned DAMA staining images were first compressed into eight-bit grayscale images with a resolution of 600 d.p.i. and color-inverted by Photoshop. The preprocessed images were then combined into an RGB overlay image with one in red channel 1 and another in green channel 2. A grid was created with a circular mask defining the boundary for each spot. For the purpose of intensity integration, the location of the mask was refined for every spot. MRAT(i, j) values, the median ratio of intensity between channel 1 and channel 2 within the mask, were determined and exported to an Excel file. The subsequent analysis was done with a customized program (DAMAPEP). Briefly MRAT(i, j) was converted to log2(MRAT(i, j)), normalized for all 312 proteins by subtracting the median log2(MRAT(i, j)) value, and scaled by dividing the root mean square of all normalized log2(MRAT(i, j)) values. The median intensity of every spot and their corresponding median and mean background values, CH1I, CH1B, and CH1AB for the reference cell line R and CH2I, CH2B, and CH2AB for the sample cell line S, were also exported for the analysis by the DAMAPEP program.

Western Blotting Analysis—
Ten breast cell lines were grown on 10-cm cell culture dishes in corresponding medium to around 90–95% confluence. The whole cell lysates were extracted, concentrated, then quantified with the Bradford assay, and adjusted to the same values. Equal amounts of total protein were loaded in each lane (29).

Gel Quantification and Statistical Analysis—
For every protein, relative band densities of 10 cell lines were scanned and digitized by UN-SCAN-IT software (Silk Scientific, Orem, UT). These values were normalized against the highest value and averaged over all experiments. Relative expression levels of five proteins between normal and cancer cells were analyzed by using the Student's t test. For every protein, expression levels in three normal cells and in seven cancer cells were analyzed by a two-sample t test assuming equal variances. The degree of freedom is 9 (n1 = 3, n2 = 7). The one-tailed p values were used for statistical inference.


   RESULTS
TOP
 ABSTRACT
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
Determining Protein Expression Profiles in Different Breast Cell Lines by DAMA Staining—
DAMA staining is a high throughput technology that can simultaneously determine the expression profiles and subcellular locations of hundreds of proteins (26). The goal of this work was to develop the DAMA staining technology for protein expression profiling and to demonstrate its application in identifying potential biomarkers for breast cancer. The technology includes the following steps: determination of protein expression profiles by DAMA staining, data extraction by the ScanAlyze program, data analysis and prediction by the DAMAPEP program, and evaluation by Western blot analysis.

For this purpose, protein expression profiles of 10 different cell lines from human mammary glands were obtained from DAMA staining (Fig. 1). The 10 breast cell lines include three normal cell lines (MCF12A, MCF10A, and Hs578Bst), three estrogen receptor-positive carcinomas (T-47D, MCF7, and ZR-75-1), and four estrogen receptor-negative carcinomas (MDA-MB-231, BT549, Hs578T, and MDA-MB-435S). The cells were grown on a 10-cm cell culture dish to 90–95% confluence and fixed and permeabilized by a standard protocol. Array-320 was then used to deliver an array of 312 primary antibodies to those fixed cells. Bound antibodies were detected by using alkaline phosphatase-conjugated secondary antibodies. The resulting array of gray dots was imaged and subjected to intensity integration and data analysis to generate expression profiles for the 312 proteins. The experiments were repeated at least twice for each cell line. The representative images and the Pearson's correlation coefficients (Rr) of the replicated images for all 10 breast cell lines are summarized in Fig. 1. Most cell lines had reproducible DAMA staining images with the Pearson's correlation coefficients higher than 0.5.


Figure 1
View larger version (53K):
[in this window]
[in a new window]

  		FIG. 1. Summary of DAMA staining images in 10 different breast cell lines. The images were obtained by using the 312-antibody microarray, Array-320. The distance between the spots is 2 mm, and the dimensions of the images within four corner spots is 38 mm in width and 30 mm in height. The cell line names are labeled either above or below the corresponding images. The spot positions are labeled from A to P for rows and from 1 to 20 for columns. Experiments were repeated at least twice, and one set of results is shown here. The Pearson's correlation coefficients (Rr) of two replicated images for all 10 cell lines are included.

 
Intensity Integration of the DAMA Staining Images by ScanAlyze—
The images obtained from DAMA staining were analyzed using standard methods for intensity integration, normalization, and scaling. Among the 10 breast cell lines, one normal cell line, MCF10A, was used as a common reference for intensity integration. The other nine cell lines were used as sample cell lines. The DAMA staining images of the nine cell lines were individually compared with the image of MCF10A by using ScanAlyze, an intensity integration program for DNA microarrays. Protein expression profiles between different samples were quantitatively compared by ScanAlyze with one image (e.g. the reference cell line R) as the channel 1 data and the other image (e.g. the sample cell line S) as the channel 2 data. MRAT(i, j), the median intensity ratio between channel 1 and channel 2 at the spot of the ith row and jth column (spot(i, j)), was exported. The logarithm of the MRATs of each pair was normalized and scaled by using a procedure similar to that used for data analysis of DNA microarrays.

Initial Data Analysis for the DAMA Staining Images with Program DAMAPEP—
A program, DAMAPEP (DAMA protein expression profiling), was developed to retrieve, normalize, and scale the data from the exported ScanAlyze files. The method for this data analysis is shown in Fig. 2. As the protein expression profiles for each cell line were determined at least twice, there are two independent DAMA staining images for each spot: two for the sample cell line (S1 and S2) and two for the reference cell line (R1 and R2). Therefore, four different sets of MRAT values, MRAT(1, 1), MRAT(1, 2), MRAT(2, 1) and MRAT(2, 2), corresponding to the intensity ratios of S1 to R1, S1 to R2, S2 to R1, and S2 to R2, respectively, were obtained for every spot(i, j) (Fig. 2a). The logarithms of those MRAT values were normalized and scaled. The average of those scaled log2(MRAT(i, j)) values, defined as SigS(i, j), represents the intensity change between the sample cell line (S) and the reference cell line (R) for the protein at spot(i, j) (Fig. 2c, Equation 1).


Figure 2
View larger version (36K):
[in this window]
[in a new window]

  		FIG. 2. Schematic of intensity integration and data analysis by DAMAPEP. a, intensity integration and initial data processing for the duplicated DAMA staining images. S1 and S2 represent two independent DAMA staining images for the sample cell line S, and R1 and R2 represent two images for the reference cell line R (such as MCF10A). For every sample cell line, each image was compared with one image of the reference cell line by the ScanAlyze program. Thus, four corresponding MRAT values were determined for each protein on the array. The average of the scaled log2(MRAT) values, SigS, represents the intensity change between the sample cell line S and the reference cell line R (Equation 1 in c). Similarly ScanAlyze was used to calculate MRATS and MRATR, the MRAT values between the two images of sample cell line S and between the two images of reference cell line R. The sum of the absolute value of their scaled values, BKS, represents the experimental error (Equation 2 in c). b, flow chart for data analysis. To compare the expression profile of 312 proteins in 10 breast cell lines, one of the normal cell lines, MCF10A, was chosen to be a common reference. Expression profiles of the other nine cell lines were compared with that of the common reference using the schemes in a. The resulting SigS values for normal and cancer cell lines were used to obtain the average intensity of the two normal cell lines and seven cancer cells over the reference cell line, namely SigNor and SigCan, respectively (Equations 3 and 4 in c). The ratio of the difference between SigNor and SigCan over SigNor was calculated to represent the average intensity change between cancer cells and normal cells relative to that of the normal cells. To qualify as differentially expressed proteins, the absolute ratio values of the proteins have to be larger than 2, and the channel intensities have to be greater than either the mean or median channel backgrounds. Additionally at least four of the cancer cell lines must have SigS greater than BKS. c, formulas used for data analysis.

 
To determine the experimental errors for the sample and reference spots, the MRAT value between the two sample cell line images, defined as MRATS, was calculated from ScanAlyze by including those two images as channel 1 and channel 2. Using a similar approach, MRATR was also calculated by using the two images of the reference cell line. As the theoretical MRATS(i, j) and MRATR(i, j) values for every spot(i, j) should be 1.0, the absolute average of the scaled log2(MRATS(i, j)) and scaled log2(MRATR(i, j)) values, defined as the background BKS(i, j), represents the experimental error for the protein at spot(i, j) (Fig. 2c, Equation 2).

Comparison of the Protein Expression Profiles with Program DAMAPEP—
After calculating signal (SigS(i, j)) and background (BKS(i, j)) values for all spots, the program DAMAPEP identifies those proteins differentially expressed between normal and cancer cell lines. For this purpose, three different criteria were utilized to identify those candidate proteins from the DAMA staining images (Fig. 2b).

First a decision was made according to the average intensity change of the protein at spot(i, j) between all cancer cells and all normal cells. SigCan(i, j), defined as the average SigS(i, j) for all seven cancer cells (T47D, MCF7, Zr-75-1, MDA-MB-231, BT549, HS578T, and MDA-MB-435s), was calculated for every spot(i, j) (Fig. 2, b and c, Equation 3). This value represents the average intensity change of the protein at spot(i, j) in all cancer cell lines relative to the reference cell line MCF10A. Similarly SigNor(i, j), the average SigS(i, j) for two normal cells (MCF12A and HS578Bst), was also calculated for every spot(i, j) (Fig. 2c, Equation 4). SigNor(i, j) corresponds to the average intensity change of the protein at spot(i, j) in other normal cell lines relative to the intensity in the reference cell MCF10A. The Ratio(i, j) calculated (Fig. 2c, Equation 5) thus represents the average intensity change of the protein at spot(i, j) between cancer cells and normal cells relative to that of the normal cells (Fig. 2b). The higher the absolute value of Ratio(i, j), the larger the intensity difference. Therefore, DAMAPEP can predict a list of proteins with different expression levels between normal and cancer cells based on the selected cutoff value of Ratio(i, j).

When 2.0 was used as a cutoff value of Ratio(i, j), 54 proteins were predicted to be differentially expressed proteins between normal and cancer cells. However, 22 of the 54 proteins had lower spot intensity in at least one of the 10 cell lines. The contribution of those weak intensity spots to the corresponding value of Ratio(i, j) could be enlarged. To test the possibility, the expression levels of eight proteins, randomly selected from the 22 proteins, were compared among 10 cell lines by Western blotting analysis. All eight proteins showed low or undetectable expression levels (data not shown). Therefore, to decrease false positives caused by lowered spot intensity, DAMAPEP automatically compares the intensity of each spot with their background from the ScanAlyze exported files (CH1I versus CH1B and CH1AB; CH2I versus CH2B and CH2AB). The spots with intensities lower than their corresponding backgrounds in any one of the cell lines were excluded from the final prediction list (Fig. 2b).

The third criteria, Sigcan(i, j) (the averaged SigS(i, j) values for all seven cancer cell lines), could easily be biased by large numbers in one or two cell lines. To eliminate this bias, the SigS(i, j) value was compared with the corresponding background BKS(i, j) for all seven cancer cell lines. Only those proteins at spot(i, j) having more than four cell lines with higher SigS(i, j) value than the corresponding background value BKS(i, j) were included in the final prediction list (Fig. 2b).

Identifying the Differentially Expressed Proteins between Normal and Cancer Cell Lines by DAMAPEP—
As discussed above, the expression profiles of 312 proteins obtained by the DAMA staining for 10 breast cell lines were repeated at least twice and were analyzed by the programs ScanAlyze and DAMAPEP. When using a minimum cutoff value of 2.0 for the absolute value of Ratio(i, j), 10 of 312 proteins (leukemia inhibitory factor receptor alpha, Tyk2, Rb p130, SRF, c-Kit, Rb p107, RAIDD, Mos proto-oncogene, I{kappa}B-β, and IL-2Rβ) were predicted by DAMAPEP to have different expression levels in cancer cells versus normal cells (Fig. 3a). The first seven proteins were predicted to have higher expression levels in cancer cells, and the latter three were predicted to have lower expression levels. Their corresponding dots in the DAMA images for all 10 cell lines demonstrated the consistency between the DAMAPEP prediction and the original DAMA staining data especially for those confirmed proteins (Fig. 3b and supplemental Fig. S1). When the cutoff value for Ratio(i, j) was decreased from 2.0 to 1.5 (representing a 50% intensity change between normal and cancer cells) six extra proteins were added to the prediction list. Among them, ErbB3, c-Raf, and Rad52 were predicted to have higher expression in cancer cells, and Fas (TNFRSF6)-associated death domain protein, FLICE inhibitory protein (short/long), and IL-1R were predicted to have lower expression. However, none of these six proteins was confirmed to have the predicted difference by Western blotting analysis (data not shown). Therefore, 2.0 was used as the cutoff value for Ratio(i, j) in the DAMAPEP prediction.


Figure 3
View larger version (18K):
[in this window]
[in a new window]

  		FIG. 3. Proteins predicted from DAMAPEP to have different expression levels between normal and cancer cells. a, the DAMAPEP-predicted proteins are shown with their name, position in the DAMA staining images, and their corresponding ratio values ranked from high at left to low at right. b, summary of the corresponding dots in the DAMA staining images shown in Fig. 1 for the seven proteins predicted by DAMAPEP with increased expression levels in carcinoma-originated cell lines. LIFR, leukemia inhibitory factor receptor alpha.

 
Expression Levels of the Predicted Proteins by Western Blotting—
To evaluate the prediction accuracy of DAMAPEP, expression levels of the above 10 proteins predicted with the 2.0 cutoff value were examined by Western blotting. Among the seven proteins predicted with increased expression in cancer cells, five proteins, RAIDD, Rb p107, Rb p130, SRF, and Tyk2, were confirmed to have higher expression (Fig. 4a). For the other five proteins predicted with DAMAPEP, increased expression level of leukemia inhibitory receptor factor alpha was only detected in some cancer cell lines. For c-Kit and IL-2Rβ, no bands were observed in Western blots in any cell lines probably due to extremely low expression levels. The expression levels of Mos proto-oncogene and I{kappa}B-β were similar in both normal and cancer cell lines (data not shown). The expression of RAIDD was almost undetectable in the three normal cell lines. The expression of Rb p107, SRF, and Tyk2 in normal cells was between 15 and 40% of their expression in cancer cells. The expression of Rb p130 in normal cells was about 20% of its expression in cancer cells (Fig. 4b). Mean values of the normalized expression of RAIDD, Rb p107, Rb p130, SRF, and Tyk2 in three normal cell lines are 5 versus 61% (p = 0.0032), 26 versus 81% (p = 0.0041), 18 versus 69% (p = 0.0055), 25 versus 73% (p = 0.0033), and 30% (p = 0.0012), respectively (Fig. 4c). Those data confirm that RAIDD, Rb p107, Rb p130, SRF, and Tyk2 are indeed overexpressed in the tested cancer cells.


Figure 4
View larger version (23K):
[in this window]
[in a new window]

  		FIG. 4. Validation of DAMAPEP-predicted results. a, Western blotting confirms that five proteins have higher expression levels in cancer cells. Total cell extracts of the 10 breast cell lines were prepared as described under "Experimental Procedures." Equal amounts of total lysates, as shown for actin, were loaded for Western blotting analysis. Proteins are labeled at the left of the gel, and tested cell lines are labeled at the top. N1, N2, and N3 represent three normal cell lines, ER+1, ER+2, and ER+3 represent three estrogen receptor-positive carcinoma cell lines, and ER–1, ER–2, ER–3, and ER–4 represent four estrogen receptor-negative carcinoma cell lines. b, relative expression of five proteins in 10 breast cell lines. Western blot analysis for every protein was repeated at least three times. The average values for every protein in different cell lines are shown here with error bars corresponding to their S.D. c, box plot of the relative expression of five proteins in three normal cell lines and seven cancer cell lines using the data shown in b. The ranges of expression levels in the group of normal cell lines versus the group of cancer cell lines are shown side by side for every protein. The interquartile ranges are shown by boxes with the median values in filled triangles. N, normal; C, cancer; HQ, higher quartile; LQ, lower quartile.

 

   DISCUSSION
TOP
 ABSTRACT
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
DAMA staining is a novel platform for protein microarray analysis that can determine both expression and subcellular localization profiles of hundreds of proteins in targeted cell and tissue samples.2 We developed this unique technology for protein expression profiling and developed a program, DAMAPEP, for intensity integration and data analysis. We demonstrated the application of this technology for potential biomarker identification by comparing the expression profiles of 312 proteins among 10 different breast cell lines. Seven proteins with higher expression levels and three proteins with lower expression in cancer cells were predicted from the DAMAPEP program. Among them, five proteins (RAIDD, Rb p107, Rb p130, SRF, and Tyk2) were confirmed to have higher expression in seven tested cancer cell lines relative to three normal cell lines. A similar approach could be widely applied to other biomedical research such as biomarker identification and signal transduction studies. Compared with other captured antibody microarrays, DAMA staining can determine the protein expression profiles for more than hundreds of proteins in a more convenient way.

The observed higher expression levels of two proteins in breast cancer cells are consistent with published reports. The increased expression of Rb2/p130 gene was observed in the majority of breast cancers (30). Up-regulated and highly active SRF was observed in squamous epithelial tumor cells (31). However, we are not aware of any studies addressing the expression changes of RAIDD, Rb p107, and Tyk2 in breast cancer.

Among the seven proteins predicted with higher expression level in breast cancer cells, five were confirmed by Western blot analysis. On the other hand, none of the three proteins predicted to have lower expression levels were confirmed by Western blotting. These errors could be caused by the quality of the DAMA staining images. For example, images of MCF7, ZR-75-1, and BT549 have a higher density in the lower part of the array (Fig. 1). This uneven intensity distribution could be caused by uneven pressure applied during primary antibody staining. In addition, the lower part of the array corresponds to the center of the culture dish, which could have higher cell density and therefore might result in the observed uneven distribution.

Two cell lines with lower Pearson's correlation coefficients, Hs578Bst and ZR-75-1, could also affect the prediction accuracy. The overlap coefficients for these two cell lines, a modified Pearson's correlation coefficient where the average gray value of each image was not subtracted from the original gray value of each pixel, were also lower. Those data suggested that superimposition of those images was not perfect possibly due to their lower spot intensities. The situation was improved by increasing the spot diameters for intensity integration by ScanAlyze. Therefore, the prediction accuracy could be improved by better DAMA staining images after further optimizing the experimental conditions.

Another possible reason for the inaccurate prediction of decreased expression levels is systematic bias in data analysis especially in the normalization procedure. As there are only 312 proteins in the dataset and most proteins are highly regulated, the assumption that those 312 proteins should have the same expression levels among different samples could generate systematic bias during data normalization. This systematic bias could be decreased by including additional antibodies on the array. In addition, deploying multiple replicate spots for each antibody on the array could also increase data accuracy, decreasing experimental error and increasing prediction accuracy.

The alkaline phosphatase-based detection with its narrow detection range of signal intensity could be another reason for the inaccuracy of the method. The intensity difference between the strongest spot and the weakest spot of the staining images as determined by the protein expression level and antibody affinity could be hundreds of thousands-fold. The weak spot may obscure the difference. For example, both ER-positive and ER-negative cell lines were utilized. However, the DAMA staining for antibody against ER at spot F5 in all cell lines showed weak signal. This could be improved by using a fluorescent conjugated secondary antibody and a fluorescence scanner. We tried several commercial microarray scanners but failed to obtain adequate signals. A more sensitive scanner may be needed for our purpose. In addition, the homogeneity of cell samples and the percentage of cell confluence could also affect the accuracy of protein expression profiling. For this reason, cells were grown to 90–95% confluence for consistency. Under current experimental conditions, 10 of 320 proteins were identified by the DAMA staining as differentially expressed proteins, and their expression patterns were consistent with those obtained by Western blot analysis, confirming the accuracy and broad application of the DAMA staining technology.


   ACKNOWLEDGMENTS
 
We thank Chandran Komma, Ling Ying, and Wei Wang for help on different stages of the project.


  FOOTNOTES
 
Received, March 15, 2007, and in revised form, September 17, 2007.

Published, MCP Papers in Press, October 13, 2007, DOI 10.1074/mcp.M700115-MCP200

1 The abbreviations used are: DAMA, dissociable antibody microarray; DAMAPEP, DAMA protein expression profiling; ER, estrogen receptor; RIP, receptor-interacting protein; RAIDD, RIP-associated ICH-1/CED-3-homologus protein with a death domain; SRF, serum response factor. Back

2 X. C. Song, G. Fu, X. Yang, Z. Jiang, Y. Wang, G. W. Zhou, unpublished results. Back

* This work was supported by National Institutes of Health Grant CA110047 and a grant from the Susan G. Komon Breast Cancer Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

S The on-line version of this article (available at http://www.mcponline.org) contains supplemental material. Back

§ Both authors contributed equally this work. Back

** To whom correspondence should be addressed. Tel.: 225-578-6733; Fax: 225-578-8011; E-mail: zhouw{at}lsu.edu


   REFERENCES
TOP
 ABSTRACT
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Emili, A. Q., and Cagney, G. (2000 ) Large scale functional analysis using peptide and protein arrays. Nat. Biotechnol. 18, 393 –397[CrossRef][Medline]

  2. de Wildt, R. M., Mundy, C. R, Gorick, B. D., and Tomlinson, I. M. (2000 ) Antibody arrays for high-throughput screening of antibody-antigen interactions. Nat. Biotechnol. 18, 989 –994[CrossRef][Medline]

  3. Kadadek, T. (2001 ) Protein microarrays: prospects and problems. Chem. Biol. 8, 105 –115[CrossRef][Medline]

  4. Macbeath, G. (2002 ) Protein microarrays and proteomics. Nat. Genet. 32, 526 –532[CrossRef][Medline]

  5. Bertone, P., and Snyder, M. (2005 ) Advances in functional protein microarray technology. FEBS J. 272, 5400 –5411[CrossRef][Medline]

  6. Hall, D. A., Ptacek, J., and Snyder, M. (2007 ) Protein microarray technology. Mech. Aging Dev. 128, 161 –167[CrossRef][Medline]

  7. Roep, B. O., Hiemstra, H. S., Schloot, N. C., De Vries, R. R., Chaudhuri, A., Behan, P. O., and Drijfhout, J. W. (2002 ) Molecular mimicry in type 1 diabetes: immune cross-reactivity between islet autoantigen and human cytomegalovirus but not Coxsackie virus. Ann. N. Y. Acad. Sci. 958, 163 –165[Medline]

  8. Collett, J. R., Cho, E. J., and Ellington, A. D. (2005 ) Production and processing of aptamer microarrays. Methods 37, 4 –15[CrossRef][Medline]

  9. MacBeath, G., and Schreiber, S. L. (2000 ) Printing proteins as microarrays for high-throughput function determination. Science 289, 1760 –1763[Abstract/Free Full Text]

  10. Ge, H. (2000 ) UPA, a universal protein array system for quantitative detection of protein-protein, protein-DNA, protein-RNA and protein-ligand interactions. Nucleic Acid Res. 28, e3[Abstract/Free Full Text]

  11. Zhu, H., Bilgin, M., Bangham, R., Hall, D., Casamayor, A., Bertone, P., Lan, N., Jansen, R., Bidlingmaier, S., Houfek, T., Mitchell, T., Miller, P., Dean, R. A., Gerstein, M., and Snyder, M. (2001 ) Global analysis of protein activities using proteome chips. Science 293, 2101 –2105[Abstract/Free Full Text]

  12. Lueking, A., Horn, M., Eickhoff, H., Bussow, K., Lehrach, H., and Walter, G. (1999 ) Protein microarrays for gene expression and antibody screening. Anal. Biochem. 270, 103 –111[CrossRef][Medline]

  13. Angenendt, P., Glokler, J., Murphy, D., Lehrach, H., and Cahill, D. J. (2002 ) Toward optimized antibody microarrays: a comparison of current microarray support materials. Anal. Biochem. 309, 253 –260[CrossRef][Medline]

  14. Delehanty, J. B., and Ligler, F. S. (2002 ) A microarray immunoassay for simultaneous detection of proteins and bacteria. Anal. Chem. 74, 5681 –5687[Medline]

  15. Knezevic, V., Leethanakul, C., Bichsel, V. E., Worth, J. M., Prabhu, V. V., Gutkind, J. S., Liotta, L. A., Munson, P. J., Petricoin, E. F., III, and Krizman, D. B. (2001 ) Proteomic profiling of the cancer microenvironment by antibody arrays. Proteomics 1, 1271 –1278[CrossRef][Medline]

  16. Sreekumar, A. (2001 ) Profiling of cancer cells using protein microarrays: discovery of novel radiation regulated proteins. Cancer Res. 61, 7585 –7593[Abstract/Free Full Text]

  17. Schweitzer, B., Roberts, S., Grimwade, B., Shao, W., Wang, M., Fu, Q., Shu, Q., Laroche, I., Zhou, Z., Tchernev, V. T., Christiansen, J., Velleca, M., and Kingsmore, S. F. (2002 ) Multiplexed protein profiling on microarrays by rolling circle amplification. Nat. Biotechnol. 20, 359 –365[CrossRef][Medline]

  18. Pawlak, M., Schick, E., Bopp, M. A., Schneider, M. J., Oroszlan, P., and Ehrat, M. (2002 ) Zeptosens' protein microarrays: a novel high performance microarray platform for low abundance protein analysis. Proteomics 2, 383 –393[CrossRef][Medline]

  19. Uetz, P., Giot, L., Cagney, G., Mansfield, T. A., Judson, R. S., Knight, J. R., Lockshon, D., Narayan, V., Srinivasan, M., Pochart, P., Qureshi-Emili, A., Li, Y., Godwin, B., Conover, D., Kalbfleisch, T., Vijayadamodar, G., Yang, M., Johnston, M., Fields, S., and Rothberg, J. M. (2000 ) A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae. Nature 403, 623 –627[CrossRef][Medline]

  20. Huang, R. P., Yang, W., Yang, D., Flowers, L., Horowitz, I. R., Cao, X., and Huang, R. (2005 ) The promise of cytokine antibody arrays in the drug discovery process. Expert Opin. Ther. Targets 9, 601 –615[CrossRef][Medline]

  21. Robinson, W. H., DiGennaro, C., Hueber, W., Haab, B. B., Kamachi, M., Dean, E. J., Fournel, S., Fong, D., Genovese, M. C., de Vegvar, H. E., Skriner, K., Hirschberg, D. L., Morris, R. I., Muller, S., Pruijn, G. J., van Venrooij, W. J., Smolen, J. S., Brown, P. O., Steinman, L., and Utz, P. J. (2002 ) Autoantigen microarrays for multiplex characterization of autoantibody responses. Nat. Med. 8, 295 –301[CrossRef][Medline]

  22. Bacarese-Hamilton, T., Bistoni, F., and Crisanti, A. (2002 ) Protein microarrays: from serodiagnosis to whole proteome scale analysis of the immune response against pathogenic microorganisms. BioTechniques 33, (suppl.) 24 –29[Medline]

  23. Mezzasoma, L., Bacarese-Hamilton, T., Di Cristina, M., Rossi, R., Bistoni, F., and Crisanti, A. (2002 ) Antigen microarrays for serodiagnosis of infectious diseases. Clin. Chem. 48, 121 –130[Abstract/Free Full Text]

  24. Woodbury, R. L., Varnum, S. M., and Zangar, R. C. (2002 ) Elevated HGF levels in sera from breast cancer patients detected using a protein microarray ELISA. J. Proteome Res. 1, 233 –237[CrossRef][Medline]

  25. Haab, B. B., Dunham, M. J., and Brown, P. O. (2001 ) Protein microarrays for highly parallel detection and quantitation of specific proteins and antibodies in complex solutions. Genome Biol. 2, research0004.1 –0004.13

  26. Wang, Y. (2004 ) Immunostaining with dissociable antibody microarrays. Proteomics 4, 20 –26[CrossRef][Medline]

  27. Cooper, C. S. (2001 ) Applications of microarray technology in breast cancer research. Breast Cancer Res. 3, 158 –175[CrossRef][Medline]

  28. Jemal, A., Murray, T., Ward, E., Samuels, A., Tiwari, R. C., Ghafoor, A., Feuer, E. J., and Thun, M. J. (2005 ) Cancer statistics, 2005. CA Cancer J. Clin. 55, 10 –30[Abstract/Free Full Text]

  29. Wu, C., Guan, Q., Wang, Y., Zhao, Z. J., and Zhou, G. W. (2003 ) SHP-1 suppresses cancer cell growth by promoting degradation of JAK kinases. J. Cell. Biochem. 90, 1026 –1037[CrossRef][Medline]

  30. Milde-Langosch, K., Goemann, C., Methner, C., Rieck, G., Bamberger, A. M., and Löning, T. (2001 ) Expression of Rb2/p130 in breast and endometrial cancer: correlations with hormone receptor status. Br. J. Cancer 85, 546 –551[CrossRef][Medline]

  31. Pschiari, E., Balmain, A., Plows, D., Zoumpourlis, V., and Pintzas, A. (2002 ) High activity of serum response factor in mesenchymal transition of epithelial tumor cells is regulated by RhoA signaling. J. Biol. Chem. 277, 29490 –29495[Abstract/Free Full Text]


Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?


This article has been cited by other articles:


Home page
 Cancer Res.Home page
O. Simma, E. Zebedin, N. Neugebauer, C. Schellack, A. Pilz, S. Chang-Rodriguez, K. Lingnau, E. Weisz, E. M. Putz, W. F. Pickl, et al.
Identification of an Indispensable Role for Tyrosine Kinase 2 in CTL-Mediated Tumor Surveillance
Cancer Res., January 1, 2009; 69(1): 203 - 211.
[Abstract] [Full Text] [PDF]

Home page
 J. Biol. Chem.Home page
M. Gakovic, J. Ragimbeau, V. Francois, S. N. Constantinescu, and S. Pellegrini
The Stat3-activating Tyk2 V678F Mutant Does Not Up-regulate Signaling through the Type I Interferon Receptor but Confers Ligand Hypersensitivity to a Homodimeric Receptor
J. Biol. Chem., July 4, 2008; 283(27): 18522 - 18529.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
M700115-MCP200v1
7/1/163    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Glossary
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Song, X. C.
Right arrow Articles by Zhou, G. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Song, X. C.
Right arrow Articles by Zhou, G. W.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 All ASBMB Journals   Journal of Biological Chemistry 
 Journal of Lipid Research   ASBMB Today 
Advertisement
spacer
Advertisement
Advertisement