The role of the cyclin D1-dependent kinases in ErbB2-mediated breast cancer

Abstract

Intact cyclin D1 functions are essential for transformation by erbB2 in tissue culture and murine models. Because cyclin D1 may alter cell proliferation through a variety of mechanisms, we used transgenic models and human tumor samples to particularly address the role of cyclin D1-cyclin-dependent kinases in transformation by erbB2. The p16 tumor suppressor specifically blocks cyclin-dependent kinase 4 and 6 activity. Here we show that an MMTV-p16 transgene blocked tumorigenesis by erbB2, demonstrating that deregulation of the cyclin-dependent kinase partner of cyclin D1 is an essential target of erbB2. ErbB2 overexpression was a determining factor in deregulation of cyclin D1-cdk4/6 interactions because neither transgenic cyclin D1 nor loss of p16 accelerated tumorigenesis in MMTV-erbB2-transgenic mice. ErbB2 was also a deciding factor in deregulation of cyclin D1-cdk4/6 in human tumors because no loss of pRb or p16 was found in tumors overexpressing erbB2, although erbB2-negative invasive breast adenocarcinomas frequently lacked expression of p16 or pRb. We conclude that deregulation of cyclin D1-Cdk4/6 interactions is a critical target of erbB2 function in human and mouse breast tumors, and erbB2's overexpression may be sufficient to deregulate cyclin D1-cdk4/6 activity in breast cancer.

Figures

Figure 1

Demonstration of the importance of erbB2 signaling through cyclin D1 using transgenic interactions between loss of p16/19, overexpression of cyclin D1, and inhibition by p16 in MMTV-erbB2-transgenic mice. A: Schematic diagram of transgenes used in this study. B: Shown are standard mortality curves demonstrating that expression of p16 using the MMTV-p16 transgene effectively blocks formation of tumors caused by MMTVerbB2. Fraction of mice remaining tumor-free is plotted on the y axis and compared to age in days on the x axis. Single erbB2-transgenic mice (▴) are compared to mice expressing both erbB2 and the human p16 transgene (▪). C: Shown are standard mortality curves demonstrating no acceleration of erbB2T-induced tumors (▴) by cyclin D1 (•) or heterozygous loss of p16/p19 (♦). The latter cross did result in the novel appearance of salivary tumors (not shown). D: Expression of individual transgenes used in this study was not altered by co-expression with additional transgenes. Protein lysates from mammary glands of the indicated genotypes (erbB2+ = MMTVerbB2 mice; CcnD1+ = MMTV-cyclin D1 mice; INK4a+ = MMTVp16 mice) were evaluated for expression of the erbB2 (erbB2), cyclin D1 (CcnD1), and human p16 (INK4a) proteins expressed by the various transgenes. E: Expression of the erbB2 (erbB2+) and human p16 (INK4a+) transgenes was evaluated in protein lysates from virgin (V) and tumor (T) tissues from mice of the indicated genotypes. F: Tumors that developed in the MMTVerbB2-INK4aARF+/− mice were evaluated for loss of the normal wild-type (W) compared with the knockout allele (K) using standard polymerase chain reactions.

Figure 2

p16-, pRB-, cyclin D1-, and erbB2-staining patterns in primary human breast tumors. A: Negative p16 control. Human tonsillar tissue with no primary antibody. B: Positive p16 control. Human tonsillar tissue with JC8 antibody. C: Negative pRb control. Human tonsillar tissue with no primary antibody. D: Positive pRb control. Human tonsillar tissue with G3-245 antibody. E and F: The same tumor stained for p16 (E) and pRb (F). Note that the endothelial cells stain p16+ in E, but both nuclear and cytoplasmic staining is seen in all tumor cells in F. G and H: A different tumor stained for p16 (G) and pRb (H). In G, note both nuclear and cytoplasmic staining. I and J: A third tumor stained for p16 (I) and pRb (J) as an example of low average staining for both genes in a single tumor. K and L: A fourth tumor stained for p16 (K) and pRb (L). Note the clonality of p16 loss in different portions of this tumor. M: Example of erbB2-negative staining. N: Example of erbB2-positive staining. Note the membrane localization of the erbB2 staining. O: Example of cyclin D1-negative staining. P: Example of cyclin D1-positive staining. Note the nuclear localization of the cyclin D1. Original magnifications: ×100 (A–D, K–N); ×250 (E, F); ×400 (G–J, O, P).

Figure 3

A: ErbB2 staining is found only in tumors when p16 and pRb are intact. Paraffin-embedded specimens were immunostained and scored for erbB2, p16, and pRb expression. Total tumors, n = 47. All tumors were invasive ductular carcinomas as described previously. Groups of tumors losing p16 (dark gray region of the bar graph color indicated in the legend to the figure), gaining cyclin D1 (black region of the bar graph), and losing pRb (white region of the bar graph) are indicated for erbB2-positive tumors (left column graph) and erbB2-negative (right). (Light gray regions of the charts indicate that none of these changes were identified.) The numbers in the bar graphs identify the number of tumors exhibiting the indicated change. Although cyclin D1 does not correlate with erbB2 staining, erbB2 is never found in tumors lacking pRb or p16. This result is statistically significant by Fisher’s exact test (P = 0.008). B: Cyclin D1 staining is found in ER-positive tumors, but rarely in ER-negative tumors. Loss of p16 and pRb are not associated with ER-positive or -negative tumors. “No change” identifies tumors with normal D1, p16, and pRb.

Figure 4

A and B: Log-fold changes in cyclin D1 mRNA measured by quantitative real-time (QRT) polymerase chain reaction (PCR) are plotted for each of 15 erbB2-negative tumors (A) and 8 erbB2-positive tumors (B). All tumors were invasive ductular carcinomas as described previously. The mean log change for cyclin D1 in the erbB2-positive cases was 0.59 (SD = 0.94) and 0.10 (SD = 0.71) for the erbB2-negative cases (P = 0.09 by t-test). C and D: Log-fold changes in p16 mRNA (black columns) and pRb mRNA (white columns) are plotted for the same individual tumors as in A and B. Criteria for the loss of expression of pRb or p16 is that the CT number is 42 for the sample based on 42 cycles in QRT-PCR. Therefore samples showing a downward bar did not necessarily lose the expression of the specific message, but instead the expression of the specific message was decreased compared to their respective controls. The mean log change for p16 in the erbB2-negative cases was 0.89 (SD = 1.25) and 0.85 (SD = 1.51) for the erbB2-positive cases (P = 0.47 by t-test). The mean log change for pRb in the erbB2-negative cases was −0.018 (SD = 0.71) and 0.25 (SD = 0.38) for the erbB2-positive cases (P = 0.04 by t-test).