Activation of Stat3 has been detected in many human neoplasias [37
], and it has been shown that IL-6-type cytokines induce Stat3 phosphorylation in various human and rodent cell lines [21
]. In addition, it has been demonstrated that these cytokines, including LIF, are expressed in breast cancer cells and in other tumor types [4
]. Interestingly, in certain myeloma and prostate cancer cell lines, IL-6 has been identified as the main cytokine responsible for Stat3 activation induction [40
]. In addition, a very recent report suggests a similar role for this cytokine in breast cancer cells [42
]. In mouse mammary glands during post-lactational involution, both induction of IL-6 and LIF expression and Stat3 activation have been demonstrated [17
]. Interestingly, in this context, Stat3 activation seems to be more dependent on LIF than on IL-6 status [43
]. In mammary tumors, to our knowledge, no report has yet been made linking LIF expression to Stat3 activation.
It has been reported that LIF and LIF-R expression in breast tumors is associated with favorable biological features such as diploidy and low S-phase fraction. In addition, in those tumors LIF-R expression was correlated with the presence of ER [3
]. On the basis of these data, those authors postulated that tumors expressing LIF/LIF-R would represent a phenotype that is closer to 'normal' and would therefore be less aggressive. Interestingly, although it has been shown that Stat3 expression and activation is commonly found in breast cancer cells [15
], and this transcription factor activation resulted in the malignant transformation of fibroblasts [43
] and the proliferation of mammary tumor cells [4
], there is also evidence that Stat3 activation in human breast cancer is associated with a better prognosis [46
Our results in mouse mammary tumors also show an association between LIF-R expression and Stat3 activation with a less aggressive phenotype. LIF-R+
MMTV(LA)-induced tumors appear in mid-pregnancy, when serum progesterone and estrogen levels are high, and continue to grow until delivery. After this, they soon regress and reappear in subsequent gestations, suggesting that progesterone and/or estrogen have a main role in their development. In addition, these tumors show high expression of estrogen and progesterone receptors (ER+
). Eventually, in the same mouse or after successive passages, these neoplasias progress to become autonomous: they lose hormone receptor expression and grow independently of the female's hormonal status [24
]. However, in spite of this drastic change in hormone dependence, MMTV(LA) HITs are not very aggressive and show a variety of histological patterns, from well-differentiated to very poorly differentiated architecture [24
]. They also remain LIF responsive and show Stat3 activation. In contrast, the LM3 cell line, which derives from a spontaneous BALB/c mammary adenocarcinoma, gives rise to ER-
poorly differentiated highly invasive (100% incidence of lung metastasis) tumors [47
]. Here we have shown that these tumors do not express LIF-R and show low levels of Stat3 activation. Our results from mouse mammary tumors are therefore in good agreement with data from human breast cancer samples. This suggests that, in certain cases, these experimental models can be better tools than breast cancer cell lines for reproducing particular aspects of human malignancies.
In transformed cells, there are no known naturally occurring mutations in Stat3 that lead to its constitutive activation. Alternatively, it has been proposed that Stat3 activation in tumors and in oncogene-transformed cells would be dependent on growth factor tyrosine kinase receptor activation or deregulation of JAK kinase's activity [48
]. The results shown here demonstrate that in well-differentiated mouse mammary tumors the constitutive activation of Stat3 would be mostly dependent on overexpression of LIF. The phosphorylation regulatory pathways of this transcription factor might therefore not be altered in these cancer cells. In addition, it has been shown that blocking ERK1/2 phosphorylation resulted in inhibition of Stat3 activation in Jak2-null cells, whereas no effect on pY-Stat3 has been observed in wild-type cells [43
]. Similarly, in HC11 cells we have not found a clear effect on Stat3 tyrosine phosphorylation when ERK1/2 activation was blocked. These results suggest that this MAPK could have a relevant role in mammary Stat3 activation only when the gp130/Jak2 pathway has been impaired in some way.
Our results from the crystal violet assays indicate that treatment with LIF can produce different biological responses in non-tumorigenic and tumorigenic cells, namely inhibiting and inducing cell survival, respectively. However, and in spite of this cytokine's being the principal one responsible for Stat3 phosphorylation, in both cases blocking Stat3 activation increased cell survival. It has been shown that expression of the phosphoinositide 3-kinase regulatory subunits p55α and p50α is directly induced by Stat3 during mammary gland involution [18
]. These proteins are involved in the downregulation of phosphoinositide 3-kinase signalling and Akt/protein kinase B activity, and abrogation of this survival pathway is essential for the induction of apoptosis in mammary epithelial cells [18
]. Our interpretation for the results shown here is therefore that in certain mammary tumor cells this apoptosis-inducing pathway is still functioning. However, these cancer cells are not fully responsive to the strategies for controlling cell survival because they are very sensitive to the activation of proliferative signaling pathways. For example, our results show clearly how much more susceptible tumor cells are to inhibition of MEK activity. We therefore believe that in both normal and neoplastic mammary cells LIF is able to induce both the survival and apoptotic pathways, the balance of which can lead to completely different outcomes in these cell types.
Stat3 biological activity depends on multiple factors, many still unknown. For example, in melanoma cells, IL-6/Stat3 function is modulated by the stage of tumor progression [49
]. The results shown here suggest that in well-to-moderately differentiated mammary tumor cells, LIF-induced Stat3 activation preserves the pro-apoptotic role of this factor in non-tumorigenic mammary cells. This activity might be altered in more aggressive or less differentiated tumors by different causes that need to be analyzed in future experiments. However, our results imply that in the development of therapeutic strategies for blocking Stat3 in breast cancer cells, the strong dependence on the cellular context that this factor activity displays should be taken into account.