| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Institute for Molecular Bioscience, University of Queensland, St. Lucia 4072, Australia
Address all correspondence and requests for reprints to: Michael J. Waters, Institute for Molecular Bioscience, University of Queensland, St. Lucia 4072, Australia. E-mail: m.waters{at}imb.uq.edu.au.
Hypophysectomy as a treatment for breast cancer was introduced by Luft and Olivecrona in 1953 (32), and together with oophorectomy and adrenalectomy, became a part of ablative endocrine therapy until the introduction of potent antiestrogens and aminoglutethimide in the 1980s. However, despite the obvious advantages of chemical suppression of estrogen, there were reports (1, 2) of regression of advanced breast cancers resistant to tamoxifen and aminoglutethimide as a result of hypophysectomy, leading to extension of survival (e.g. from 5.6–25.8 months) (2). A popular view has been that removal of prolactin was responsible for the improvement, rather than of GH/IGF-I, but recently, evidence has been accumulating to support a key role for the GH/IGF-I axis in mammary cancer, particularly in animal models. The publication by Shen et al. (3) in this month’s issue of Endocrinology provides striking evidence for the involvement of GH/IGF-I in promotion of breast cancer and evidence of its reversibility when the GH stimulus is removed. The important implication is that treatment of breast (and prostate) cancers by effective suppression of the GH/IGF-I axis could produce tumor regression.
Shen et al. (3) found that the carcinogen N-methyl-Nnitrosourea fails to induce breast tumors in dwarf (dw/dw) rats lacking circulating GH, that rat or bovine GH replacement increases tumor incidence and growth toward that seen in normal rats, and that cessation of the GH replacement results in the disappearance of the tumors within 4 wk. Although these tumors are estrogen dependent, in a second recent study, Swanson and colleagues (4) showed that estrogen-independent tumors induced in mice bearing the large T antigen (C3(1)/Tag mice) are strikingly reduced in size and number when these mice are crossed with GH receptor null mice. Moreover, the mammary tumors that were present were more morphologically differentiated in the GH receptor null mice.
These are not isolated findings. Ramsay et al. (5) found that Lewis dw/dw rats with less severe GH deficiency were resistant to dimethylbenzanthracene (DMBA)-induced mammary carcinogenesis, and administration of porcine GH to these increased tumor incidence. Mice overexpressing a GH antagonist likewise have a longer latency period and survival with regard to DMBA-induced mammary tumors (6), whereas mice overexpressing human GH (hGH) develop spontaneous mammary cancers (7). Furthermore, lit/lit mice lacking GH secretion and possessing low-serum IGF-I support the growth of MCF-7 human breast cancer cells only poorly compared with their wild-type controls (8). The loss of tumor support by circulating IGF-I is an important element in resistance to tumorigenesis by GH-deficient rodents, as exemplified by the decreased tumor incidence in liver-specific IGF-I gene-deleted mice, with hepatic IGF-I gene expression deleted, and only 25% of normal plasma IGF-I levels (9). However, it is notable that in both the DMBA and C3(1)Tag models of breast cancer, the extent of protection (latency, incidence, size) conferred by decreasing endocrine IGF-I to 25% is modest compared with that seen in GH receptor null or lit/lit mice, in which the plasma IGF-I is substantially lower, and GH action is absent. A similar situation applies to mice overexpressing the GH antagonist, which have serum IGF-I levels only about 50% normal (6)
Do these animal studies have any relevance to human breast cancer or other malignancies? In the only survey to date of cancer incidence in GH receptor mutant (Laron) dwarfs and those with isolated GH deficiency, 222 individuals were reported not to have a single malignancy, whereas 338 first and second-degree relatives were found to have a 10–24% incidence of various malignancies (10). This striking result needs to be tempered with the knowledge that the average age of the Laron dwarfs was considerably younger than the relatives, and even in the best comparison, the average age of the Laron mutants was 32 yr (n = 40) compared with 55 yr for the relatives (n = 99). Nevertheless, the oldest GH receptor mutant was 75 yr, so this appears to be a genuine finding.
What other epidemiological evidence is there for an association between the GH/IGF-I axis and breast and other cancers in man? First, in a comprehensive review of over 300 case-control and cohort studies, Gunnell et al. (11) compared individuals of less than 160-cm height with those over 175 cm, and found a 22% increase in breast cancer incidence, a 20% increase in prostatic cancer, and a 20–60% increase in colon cancer for both men and women with increased adult height. Second, in a large study (>100,000) of Danish children (12), the top height quintile at 14 yr of age had an adjusted relative risk of over 1.5 for breast cancer, substantially greater than for any other factor. Height at 14 yr may be taken as a surrogate for GH/IGF-I axis activity in the well-fed Danish population (13). Accordingly, a comprehensive meta-analysis of serum IGF-I across 21 publications (3609 cases of cancer, 7137 controls) found that high-normal concentrations of IGF-I are associated with a 2-fold increased risk of premenopausal breast cancer (P = 0.007), prostatic cancer (P = 0.009), and colorectal cancer (P = 0.09), but not for postmenopausal breast cancer or lung cancer (14). Interestingly, this study also found a positive association between serum IGF binding protein-3 (a target of GH action) and an increased risk of premenopausal breast cancer.
These epidemiological studies implying that IGF-I is an important factor in the development of breast and prostatic cancer fit with in vitro studies showing that IGF-I acts to promote tumor growth by increasing proliferation and decreasing apoptosis, and by promoting angiogenesis and metastasis (15). Indeed, breast cancer-associated gene-1 deficiency results in activation of the IGF-I axis in mammary cells (16). Is then IGF-I the culprit, with GH serving only to generate endocrine or paracrine IGF-I? This probably is not the case. The possibility of direct actions of GH on breast cancers is supported by the demonstration of GH receptor expression in breast cancer (17, 18), which is up-regulated compared with adjacent normal tissue, and inversely correlated with tumor grade. GH itself is expressed in breast cancers (19), together with pit-1 (20), which increases mammary hGH expression and cell proliferation, so GH could operate as an autocrine/paracrine tumor enhancer. Indeed, Raccurt et al. (19) found that metastatic breast carcinomas have the highest levels of hGH expression. More generally, Perry et al. (21) list a number of cases of ectopic hGH secretion associated with malignancy in lung, breast, stomach, and other tissues. The oncogenic role of such autocrine hGH has been extensively addressed by Lobie and colleagues (22), who have reported that sustained autocrine expression of hGH confers an invasive phenotype on MCF-7 mammary cancer cells and that this is a result of an epithelial-mesenchymal transition (22, 23). Moreover, forced autocrine hGH expression in the immortalized human mammary line MCF 10A transforms these cells into invasive carcinoma cells in vivo. That is, autocrine hGH is an oncogene (21, 24). It is notable that MCF-7 cells do not produce detectable levels of IGF-I (25), although IGF-I is present in the serum supplement. Using microarray technology, Lobie and colleagues (21) have characterized a number of key genes responsible for hGH-induced oncogenic transformation, including increased expression of the oncogene HOXA1 (homeobox A1), and of CHOP (C/EBP-homologous protein or GADD153), which confers resistance to apoptosis. Autocrine GH also increases the levels of the catalytic subunit of telomerase by stabilizing its transcript, which would increase cell life span. Importantly, IGF-I synergizes with the actions of autocrine GH in promoting cell proliferation and motility (26), and IGF-I would be available in vivo, both as endocrine IGF-I and as paracrine IGF-I generated from mammary stromal cells.
There is evidence that autocrine prolactin acts to promote breast and prostatic cancer (27, 28), so one would predict that autocrine hGH, able to activate both GH and prolactin receptors, would act as an oncogene in vivo. These two receptors share most signaling pathways, including those targeting extracellularly regulated kinase and signal transducer activator of transcription 3, which promote cell transformation. It is not difficult to envisage a combination of direct autocrine hGH/prolactin-mediated initiation of carcinogenesis, together with promotion by increased hGH-dependent IGF-I. Why then does clinical replacement of hGH in short children not increase cancer risk, at least in the short term? One can argue that autocrine GH is different in its actions from endocrine GH, and the work of Lobie and colleagues (22) would support this. Nevertheless, elevated IGF-I does increase risk, and acromegalics do have a significantly increased risk of some malignancies. We must conclude that the assessment of risk in GH-treated children is relative to the normal population, which itself has a quite significant cancer incidence. Comparison with non-GH treated individuals would presumably show a significantly elevated risk, but the imperative of increasing height and quality of life mitigates this concern.
What is most relevant here is the realization that effective treatment of breast, prostatic, and colon cancers (among others) will, based on the animal studies and Laron data, require suppression of serum IGF-I to less than 20% normal levels. It appears that there is a permissive level of IGF-I (
20% of normal) for cancer and that above this, there is only a slow increase in risk increasing to the 2- to 3-fold risk levels seen in acromegaly for colorectal cancer and lymphoma. None of the current therapies [pegvisomant/Trovert (Sensus Drug Development Corp., Austin, TX), somatostatin analogs, antisense oligonucleotides to GH receptor (29), or the small molecule receptor blocker, biovitrum-A] can reduce circulating IGF-I to less than 20% alone. Combinations may be more effective, and of course chemotherapy itself decreases serum IGF-I and growth velocity. Combination of some of these serum IGF-I lowering agents with blockers of IGF-I receptor activation such as the small molecule inhibitors now showing promise [picropodophyllin, PQ401, NVP-ADW742 (30)] is likely to be of greater benefit. Even GHRH antagonists, which block GH secretion as well as acting directly to inhibit tumor growth [JV-1-38 (31)] could profitably be trailed together with such IGF-I blockers. In any case, now that we have a clearer understanding of the requirement to drastically reduce GH action as evidenced by a very low serum IGF-I, we should now move to testing combinations that are likely to be clinically efficacious in combating breast and other cancers.
 |
Footnotes |
|---|
Abbreviations: DMBA, Dimethylbenzanthracene; dw/dw, dwarf; hGH, human GH.
Received June 27, 2007.
Accepted for publication July 6, 2007.
|
References |
|---|
 Top References |
|---|
This article has been cited by other articles:
 |
|
 |
|
  V. Pandey, J. K. Perry, K. M. Mohankumar, X.-J. Kong, S.-M. Liu, Z.-S. Wu, M. D. Mitchell, T. Zhu, and P. E. Lobie Autocrine Human Growth Hormone Stimulates Oncogenicity of Endometrial Carcinoma Cells Endocrinology, August 1, 2008; 149(8): 3909 - 3919. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
