Referee 1's review:

General comments:

This manuscript provides experimental data suggesting that, in the ERbeta1 negative breast cancers, the concomitant expression of alpha6beta4 and ErbB-3 receptors may be clinically relevant in predicting the response to Tamoxifen. While several methods were utilized to address the very important issue in breast cancer biology and treatment, there are some errors through out the manuscript, and the authors may also consider reperforming some of their experiments to obtain higher quality data in order to strengthen their points.

Specific comments:

1. It is not clear why western blot data were shown in fig. 1A, and RT-PCR data were used in fig. 1B.

2. The quality of fig. 2 is not impressive, and the levels of total Akt in fig. 2B should be added.

3. Statistical analysis of the bar graphs in fig. 3A, 3C may help the readers to better understand the authors' conclusion. A better PARP cleavage data should be provided in fig. 3B. In the figure legend, it is not proper to say "PARP degradation".

4. On page 9, it says "our data also suggest that a cooperative signaling between ErbB-3 and alpha6beta4 integrin could influence resistance to hormone therapy in vivo." However, figure 3 does not direct support this statement. SKBR3 cells are negative for both ERalpha and ERbeta, it is not clear through what mechanism this cell line was also response to tamoxifen treatment.

5. In the IHC data (fig. 4B), a negative control and an ErbB-2 positive case are needed.

6. On page 10, it says "the majority of high beta4-positive tumors were ERbeta1 negative (p<0.0001)", however this is really different from the data in table 2, which shows that 128 or 81% of the beta4-positive tumors are ERbeta1 positive.

7. There are descriptions of immunoprecipitation in the materials and methods, but no IP data was presented or mentioned in the manuscript.

8. In the materials and methods, a whole section was used to describe ERbeta immunocytochemistry by cytospin of breast cancer cell lines, however, there was no such data in the manuscript.

9. (a), (b), (c), (d) were used in fig. 5 and 6, while A, B, C were used in the other figures.

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.