Cyclin E and Survival in Patients with Breast Cancer

Abstract & Commentary

Synopsis: The hazard ratio for death from breast cancer for patients with high total cyclin E levels as compared with those with low total cyclin E levels was 13.3 to about 8 times as high as the hazard ratios associated with other independent clinical and pathological risk factors, including axillary node status.

Source: Keyomarsi K. N Engl J Med. 2002;347: 1566-1575.

Keyomarsi and colleagues introduce their article by noting that the prognosis for patients with newly diagnosed breast cancer is determined primarily by the presence or absence of metastases in draining axillary lymph nodes. However, about one third of women with breast cancer who are node negative suffer a recurrence, and about one third of those with positive nodes are free of disease at 10 years after local-regional therapy. The lack of predictability about survival derived from clinical disease stage and hormone receptor status drives the search for better prognostic markers of breast cancer survival. In this study, Keyomarsi et al examined the correlation between survival and tumor markers in 395 women with breast cancer followed for a median of 6.4 years. The molecular markers were cyclin E, cyclin D1, cyclin D3, and the HER-2/neu oncogene. Keyomarsi et al looked at levels of protein expression using Western blot analysis and traditional immunohistochemical techniques.

Cyclins are proteins that regulate cell growth. They are known as mitogens because high levels cause cell growth. Tumors often make aberrant versions of known proteins, and such is also true for breast cancer cells and cyclin E. In breast cancer cell lines, cyclin E is amplified, and the cyclin E protein is often constitutively overexpressed. Some of these cell lines overexpress not only the full-length version, but also up to 5 low-molecular-weight isoforms. In this study, Keyomarsi et al used a variety of assay methods to determine the levels of cyclin E and other molecular markers. Clinical data were obtained from tumor registries. Of the 114 patients with stage I disease, 12 had a recurrence of breast cancer and died of it with a median time to death of 4.1 years. All 12—and only those 12 of the 114 patients—had a high level of cyclin E. High levels of total cyclin E and high levels of low-molecular-weight cyclin E as assessed by Western blotting (but not by immunohistochemistry) were associated with hazard ratios for death from breast cancer of 33.0 and 20.8, respectively. In the article, Table 2 shows the factors that independently predict death from breast cancer. Those factors and their respective hazard ratios (with confidence intervals) are: positive nodes 1.8 (1.2-2.8); stage IIIB-IV 1.7 (1.1-2.5); negative estrogen-receptor status 1.8 (1.3-2.7); high level of low-molecular weight cyclin E 2.1 (1.1-4.0); and high total cyclin E 13.3 (5.8-30.2). Keyomarsi et al conclude that identifying those patients with poor prognosis breast cancer may spare some from the toxic systemic therapies now routinely administered.

Comment by Sarah L. Berga, MD

The reason to know about this study is that high levels of cyclin E, as assessed by Western blotting, were associated with a much larger hazard ratio for death from breast cancer than any other known predictor. In the accompanying editorial, Sutherland and Musgrove (N Engl J Med. 2002;347:1546-1547) point out that cyclin E may induce chromosomal instability, a hallmark of cells that metastasize and respond poorly to chemotherapies, be they endocrine or cytotoxic. As they note, it is metastases, not the rate of proliferation of the primary tumor, which causes death from breast cancer. Keyomarsi et al, in the source article, indicate the potential to identify patients who can be spared from aggressive intervention. It would also be wonderful to know if cyclin E correlates with past HRT use or whether "normal levels" can be used to identify patients likely to benefit from subsequent use of HRT. As the study of O’Meara so poignantly demonstrated, breast cancer survivors who used HRT lived longer and had fewer recurrences.1 Perhaps the patients who lived long enough to take HRT belonged to the group already destined to survive by virtue of the breast tumor’s biology, including, perhaps, low levels of cyclin E. In short, many of the questions with which we struggle, such as whether HRT causes breast cancer, make no sense unless breast cancer as an entity behaves homogenously. The present study makes evident that, indeed, not all breast cancers are the same. The corollary, then, is that we should not treat the primary breast cancers and the patients who have them as "the same." It is true that these findings are not ready for prime time because the assay methods are not widely available, but the lesson holds nonetheless. We must continue to find ways to appropriately individualize therapies, while eschewing simplistic assumptions about sameness and behavioral homogeneity. 

Dr. Berga is Professor and Director, Division of Reproductive Endocrinology and Infertility, University of Pittsburgh.


1. O’Meara ES, et al. J Natl Cancer Inst. 2001;93: 754-762.