Age-associated biomarker profiles of human breast cancer

https://doi.org/10.1016/S1357-2725(02)00052-3Get rights and content

Abstract

To explore the hypothesis that aging not only increases breast cancer incidence but also alters breast cancer biology, we correlated patient age at diagnosis with tumor histology, stage and biomarkers independently determined from two different tumor archives: an American collection of ∼800 paraffin-embedded and immunohistochemically analyzed primary breast cancers, and an European collection of ∼3000 cryobanked primary breast cancers analyzed by ligand-binding and enzyme immunoassay (EIA). The prognostic biomarkers chosen for comparison represented surrogate measures of tumor: (i) proliferation, growth and genetic instability (mitotic and apoptotic indices, Ki-67/MIB-1-positivity, nuclear grade, p53-positivity), (ii) endocrine-dependence (estrogen receptor (ER), progesterone receptors (PR), pS2, Bcl2), (iii) growth factor receptor-dependence (ErbB2, EGFR/ErbB1), and (iv) angiogenic, invasive and proteolytic potential (uPA, PAI-1, Cathepsin D, VEGF). No biomarker reflecting tumor angiogenic, invasive or proteolytic potential showed a significant correlation with patient age at diagnosis. In contrast, significant inverse correlations (|r|>0.1; P≤0.05) were observed for all measures of tumor growth and genetic instability as well as growth factor receptor overexpression (ErbB2 or EGFR positivity). Only one marker of endocrine-dependence, ER expression, showed a significant positive correlation with patient age at diagnosis. In summary, these findings support the hypothesis that breast cancer biology is significantly affected by patient age. In particular, breast tumors arising in older patients have slower growth rates, are more likely to be ER-positive, and are less likely to be p53-positive, EGFR-positive or ErbB2-positive.

Introduction

The association between human cancer and aging remains poorly understood [1]. The conventional wisdom among physicians is that “old people get old tumors” [2]; and the fact that most human cancers increase in frequency with aging has long dominated their descriptive epidemiology but received little basic research attention [3], [4]. The incidence of thyroid cancers, for example, increases with aging; but poorly recognized is the fact that its biological aggressiveness (controlled for disease stage and presence of activated oncogenes and loss of tumor suppressor genes) increases even more dramatically with patient age [5]. Yet this age-associated pattern of a worsening biological tumor profile is not universally observed with other age-dependent epithelial cancers [6]; prostate cancer, for example, shows a markedly increasing incidence with age but no age association with any biomarker of prognostic significance. Of all human epithelial cancers, breast cancer is the best studied to date with regard to clinical features distinguishing old versus young tumors; without controlling for tumor stage or biological markers, it appears that advancing age is associated with a more favorable breast tumor biology including a higher proportion of estrogen receptor (ER)-positive cases [7], [8], [9]. These breast cancer studies, however, have evaluated few of the many tumor biomarkers known to associate with patient prognosis and were seldom powered for decade-by-decade analyses across all patient age groups.

Scientists studying the biology of aging across various mammalian species have offered several theories for the observed increase in human cancer with aging [10], [11], [12], [13], [14], [15]: (i) longer exposure interval to carcinogen or tumor promoter (e.g. endogenous estrogen exposure and breast tumorigenesis) (ii) increased transformation susceptibility due to impaired intracellular repair (e.g. DNA) and detoxification (e.g. antioxidant defense) mechanisms and (iii) diminished immune surveillance and/or stromal tissue defense due to an age-associated increase in senescent cell populations. As has been long proposed [14] and recently elaborated upon in more mechanistic detail [15], the above theories all support a general paradigm that age-associated cancers possess an accumulated mutational load with loss of p53-dependent DNA damage checkpoints, acquisition of telomerase activity and a “mutator” phenotype resulting in progressive genetic instability. As well, the escape of transformed cells from an aging tissue population of senescent cells requires loss of cell cycle checkpoints and other biological changes typically found in tumors showing rapid growth and proliferation and a more invasive (i.e. metastatic) phenotype [13]. Unfortunately, clinical studies and direct evidence in support of predictions emerging from these theories are lacking for virtually all age-associated human malignancies including breast cancer [1], [6].

To explore the hypothesis that aging not only increases breast cancer incidence but also alters breast cancer biology, we correlated patient age at diagnosis with tumor histology, stage, and various biomarkers independently determined on two different tumor archives: an American collection of ∼800 paraffin-embedded and immunohistochemically analyzed primary breast cancers, and a European collection of ∼3000 cryobanked primary breast cancers analyzed by ligand-binding and EIA. Within the constraints of their assay methodologies, the multiple biomarkers assayed herein represent validated surrogate measures of tumor growth, invasive potential, genetic instability, endocrine and growth factor receptor dependence.

Section snippets

European tumor collective and biomarker analyses

The European (STB, Basel) collective of 3208 cryopreserved (−70 °C) primary breast tumors and its associated database of patient characteristics including age at diagnosis and survival with/without therapy, as well as tumor characteristics including histology, nuclear grade, stage (TNM with tumor size and number of positive nodes) and selected tumor biomarkers has been previously described [16], [17]. Tissue homogenates were prepared for biomarker measurement by validated radioligand-binding

Statistical methods

Associations between the continuous variable biomarker values from the European database (untransformed or transformed values, as indicated) for each patient age category, displayed as boxplots, were tested for significance using non-parametric Wilcoxon (compares median values of one age group to those of another age group) or Kruskal-Wallis rank-sum (compares median values of all different age groups) statistical tests. Tumor characteristics and IHC scores from the American database

Results

No significant association between breast tumor histology (proportion of invasive ductal versus lobular breast cancers) and patient age at diagnosis was observed in either the European or American tumor collectives. As well, neither collective observed any strong correlations (|r|>0.1; P≤0.05) between age and tumor stage at diagnosis, measured by tumor size (T, with median T values ranging from 2.0 to 2.5 cm across all age groups in both collectives) and nodal involvement (N, with all age groups

Discussion

In addition to the standard characterization of breast cancers by histology, grade and stage, this study evaluated two large European and American breast tumor collectives for prognostic markers that might correlate with increasing patient age at the time of diagnosis. Included in this retrospective analysis were previously validated biomarkers that represented surrogate measures of breast tumor: (i) proliferation, growth and genetic instability (mitotic and apoptotic indices,

Acknowledgements

We thank C. Wullschleger, K. Paris and A. Takahashi for technical assistance, data management and tumor banking in the analysis of the European collective. We thank S. Liu for technical assistance in the analysis of the American collective. This work was supported in part by NIH sponsored Grants R01-CA71468 and R01-CA36773 and DOD sponsored grant DAMD17-99-1-9111, the Hazel P. Munroe (Buck Institute) and Janet Landfear (Mt. Zion Health Systems) memorial funds, as well as funding from the

References (26)

  • K.M. van Tol et al.

    Differentiated thyroid carcinoma in the elderly

    Crit. Rev. Oncol. Hematol.

    (2001)
  • S. Costa et al.

    Predicitve value of EGF receptor in breast cancer (letter)

    Lancet

    (1988)
  • W.B. Ershler et al.

    Aging and cancer: issues of basic and clinical science

    J. Natl. Cancer Inst.

    (1997)
  • W.B. Ershler, Do old people get old tumors? in: M.C. Perry (Ed.), American Society of Clinical Oncology Educational...
  • D. Trichopoulos, E. Petridou, L. Lipworth, H.-O. Adami, Epidemiology of cancer, in: V.T. DeVita Jr., S. Hellman, S.A....
  • R. Yancik et al.

    Aging and cancer in America: demographic and epidemiologic perspectives

    Hematol. Oncol. Clin. North Am.

    (2000)
  • E.L. Trimble, Cancer biology and aging, in: M.C. Perry (Ed.), American Society of Clinical Oncology Educational Book...
  • G.C. Kimmick, H.B. Muss, Breast cancer in older women, in: J.R. Harris, M.E. Lippman, M. Morrow, C.K. Osborne (Eds.),...
  • A.J. Nixon et al.

    Relationship of patient age to pathologic features of the tumor and prognosis for patients with stage I or II breast cancer

    J. Clin. Oncol.

    (1994)
  • G.M. Clark et al.

    Correlations between estrogen receptor, progesterone receptor, and patient characteristics in human breast cancer

    J. Clin. Oncol.

    (1984)
  • S.N. Austad, Why We Age?, Wiley, New York, 1997, pp....
  • C.E. Finch, E.L. Schneider, Biology of aging, in: L. Goldman, J.C. Bennett (Eds.), Cecil’s Textbook of Medicine,...
  • E.A. Burns et al.

    Aging, immunity and cancer

    Cancer Causes Control

    (2000)
  • Cited by (91)

    • Treatment and survival outcomes in older women with primary breast cancer: A retrospective propensity score-matched analysis

      2022, Breast
      Citation Excerpt :

      Older patients were found to have a larger proportion of estrogen receptor (ER)- and human epidermal growth factor receptor 2 (HER2)-negative tumours [6–8]. Studies also demonstrated that older patients tended to develop tumours with less aggressive characteristics than their younger counterparts [9–12]. Moreover, studies indicated that comorbidities, toxicity tolerance, functional status, and life expectancy played increasingly essential roles in treatment decisions with ageing [13], which makes decision-making for older patients increasingly complicated [14].

    View all citing articles on Scopus
    View full text