The Rho signaling network is perturbed during tumorigenesis [1
], and Rho GTPases are overexpressed and hyperactivated in breast tumors [4
]. In vitro
studies have shown that this signaling network has pleiotropic functions and can regulate tumor cell proliferation, survival, and migration [3
]. Despite the wealth of information available about how Rho GTPases affect cellular processes and signaling in vitro
, the cellular and molecular mechanisms by which Rho GTPases affect the stochastic process of tumor formation in the complex in vivo
environment are not well understood. We investigated this by examining the role of p190B RhoGAP, a major inhibitor of Rho GTPases, during MMTV-Neu induced mammary tumor formation and progression.
P190B is required for mammary gland development [12
], and we hypothesized that it would also be pivotal for mammary tumorigenesis. Previously, we examined the effects of loss of p190B function on MMTV-Neu tumorigenesis and showed that haploinsufficiency for p190B delayed tumor onset and decreased tumor penetrance [8
]. P190B was also deficient in the stroma, and transplantation assays indicated that its function in the stroma and/or vasculature was important for the angiogenic switch that is required to promote preneoplastic progression. Although these studies suggested that p190B also played an important role in the epithelium, it was difficult to assess this using the germline p190B knockout mice. We, therefore, turned to tet-regulatable p190B transgenic mice in which p190B is inducibly expressed in the mammary epithelium in response to Dox treatment.
Exogenous p190B expression enhanced MMTV-Neu induced tumorigenesis as indicated by an increase in the number of tumors that formed per mouse, which is consistent with an increase in tumor formation in mammary gland pairs one and four compared to control mice. We did not detect a difference in the number of preneoplastic lesions that formed in the p190B transgenic mice as compared to the control mice, nor did we detect a statistically significant difference in tumor angiogenesis. These data indicate that p190B expression in the epithelium does not facilitate tumor formation by affecting the angiogenic switch. Furthermore, these results are consistent with our data demonstrating that deficient expression of p190B in the stroma and/or vasculature affected progression of preneoplastic lesions in MMTV-Neu mice [8
Similar to the effects of p190B deficiency in MMTV-Neu tumors [8
], p190B transgene expression also had no impact on tumor growth, proliferation rates, or apoptosis as demonstrated by determination of tumor volumes, and quantification of BrdU incorporation and cleaved caspase-3 staining, respectively. Thus, p190B is not likely promoting tumor formation by enhancing proliferation or tumor cell survival.
In addition to promoting primary tumor formation, exogenous p190B expression also increased metastasis. Non-transformed p190B transgenic primary MECs exhibited increased adhesion to basement membrane proteins and formed invasive structures in a 3 D culture morphogenesis assay. These experiments demonstrated that even in the absence of an oncogenic stimulus, p190B promotes cellular invasion. We propose that these phenotypes contributed to the increase in metastasis in the p190B transgenic MMTV-Neu mice. However, we cannot rule out the possibility that the increase in tumor burden due to the presence of multiple tumors in the majority of p190B transgenic mice also impacted metastasis.
We attempted to elucidate the molecular mechanisms by which p190B facilitates invasion in primary non-transformed MECs by testing whether inhibition of Rac or ERK was sufficient to block the invasive phenotype promoted by exogenous p190B expression. However, inhibition of either Rac or ERK throughout MEC morphogenesis significantly perturbed polarity and lumen formation and enhanced invasion regardless of p190B expression. Dominant negative Rac1 expression has previously been shown to disrupt polarity and lumen formation in primary MECs [33
] so it was not surprising that Rac inhibition promoted these phenotypes. P190B transgene expression in the developing mammary gland [14
] and in MMTV-Neu tumors inhibited ERK activity, and it is possible that inhibition of ERK activity contributes to the invasive phenotype of p190B transgenic mammary acini.
Initially we were surprised that ERK inhibition dramatically altered lumen formation and promoted invasion in the majority of mammary acini in 3 D culture as several studies have demonstrated that sustained or elevated ERK activity promotes MEC transformation and invasion [34
]. In addition, elevated ERK activity has been linked to poor prognosis in breast cancer [37
]. Paradoxically, a number of other studies have demonstrated that ERK signaling can have opposite phenotypic effects in a cell type specific manner depending on the level, duration, and cellular localization of the signal [38
]. Furthermore, elevated ERK activity has also been linked to better prognosis in several studies of breast cancer patients [39
]. The seemingly paradoxical impact of ERK activation on mammary transformation and tumorigenesis may be dependent on the specific genetic alterations within individual breast tumors or cell lines, and the role of ERK activation should be considered within these specific contexts.
To further explore the molecular mechanisms underlying the pro-tumorigenic actions of p190B in MMTV-Neu tumors, we examined the activity levels of RhoA, Rac1, and Cdc42 in the tumors. We previously demonstrated that loss of p190B function in MMTV-Neu tumors leads to decreased Rac1 activity, and now conversely, we show that exogenous p190B expression results in elevated Rac1 activation. Similar to the studies in the p190B deficient mice, we did not detect statistically significant differences in RhoA or Cdc42 activities. However, we cannot rule out the possibilities that activity of other GTPases may be altered or that RhoA and Cdc42 are transiently regulated.
Rac1 has been implicated in transformation and metastasis [25
]. Deficiency of the Rac GEF Tiam1 in MMTV-Neu mice delayed tumor onset and decreased metastasis suggesting that Rac1 activity plays a causal role in MMTV-Neu induced tumor formation and progression [26
]. Taken together, these studies have lead us to hypothesize that Rac1 activity may be regulated by p190B during MMTV-Neu mammary tumorigenesis and that p190B may promote transformation and metastatic progression via a Rac1 dependent mechanism. Additional studies will be required to demonstrate whether p190B promotes tumorigenesis via a Rac-dependent mechanism.
A major question raised by our studies is why loss and gain of p190B function, which would be expected to increase and decrease GTPase activity, respectively, has the opposite effect on Rac1 activity? High GTPase activity and rapid turnover in the presence of elevated GAP expression seem to be contradictory. Computational modeling has suggested that although GAPs reduce GTPase activity, the GTPase turnover rate increases proportionally with increasing GAP concentration, and ultimately, the balance between GAP and GEF concentrations is critical for high GTPase activity [44
]. In addition, GTPase cycling rates in fast cycling mutants increase the transforming capacity of the Rho GTPases [45
]. It is possible that through direct interactions, p190B enhances the cycling rate of Rac1, thereby promoting elevated Rac1 activity. P190B may also act as a scaffold to bring together a large signaling complex at the membrane, and thus, exogenous p190B expression may increase Rac1 membrane localization and subsequent activity.
Another question raised by these studies is if Rac1 activity is elevated in p190B transgenic tumors, why is signaling to the downstream effectors AKT and ERK diminished? It is important to note that the Western analysis represents a fixed time point in late stage tumors that may not be indicative of the early changes in signaling that are responsible for tumor initiation and progression. However, p190B transgene expression in the developing mammary gland also inhibited signaling to AKT and ERK suggesting that similar signaling pathways are altered downstream of exogenous p190B expression during mammary gland development and tumorigenesis [14
]. In the p190B transgenic tumors the expression levels of PAK and ROCK proteins were only modestly decreased. Furthermore, phospho-ROCK was elevated in p190B transgenic tumors suggesting that despite effects on total protein levels, signaling through ROCK may still be important for MMTV-Neu induced tumorigenesis.
The Rho signaling network is highly complex, and it is likely that multiple signaling pathways converge to modulate PAK, ROCK, AKT, and ERK. For example, ErbB2 and ErbB3 cooperate to promote MMTV-Neu induced tumorigenesis, and activation of AKT and MAPK downstream of ErbB2/ErbB3 affects tumor cell proliferation [21
]. Interestingly, by Western analysis we detected a slight decrease in expression of the full-length form of ErbB3 concomitant with the appearance of a lower molecular weight form suggesting that ErbB3 may be proteolytically cleaved in the p190B transgenic tumors. Consistent with this possibility we detected an increase in MMP3 expression in the p190B transgenic tumors. Furthermore, MMP-dependent cleavage of the related ErbB4 produces a cytoplasmic fragment that regulates tumor cell proliferation [47
]. It is possible that decreased expression of full length ErbB3 at the cell surface is responsible for the diminished activation of the downstream effectors AKT and ERK in the tumors expressing exogenous p190B. Future studies will be aimed at elucidating how these complex signaling interactions regulate MEC morphogenesis and invasion utilizing primary MECs in the 3 D culture system.
To reconcile the apparent inconsistencies between the pro-tumorigenic effects of exogenous p190B expression and increased Rac1 activity with the diminished signaling to the downstream effectors we considered the possibility that ROS were elevated downstream of p190B/Rac1. ROS have been shown to act downstream of Rac1 to facilitate transformation and metastasis [15
], and in the presence of cellular stress, ROS promoted oxidative stress, diminished ERK phosphorylation, and cell death [49
]. Because of the lack of an in situ
assay to detect ROS in the tumor tissue, we examined protein carbonylation as an indicator of oxidative stress. Interestingly, protein carbonylation is elevated suggesting that oxidative stress may be increased in p190B transgenic tumors. Expression of matrix metalloproteases (MMPs) is also regulated by ROS [29
], and MMP-3 was upregulated in the p190B transgenic tumors as another potential indicator of an increase in ROS downstream of elevated Rac1 in the tumors expressing exogenous p190B. We hypothesize that p190B may promote tumor formation and progression in part by affecting Rac/ROS mediated signaling, however, additional studies will be required to prove this hypothesis.
Rho GTPases including RhoA, Rac1, and Cdc42 are elevated and hyperactivated in breast cancer [4
]. We, therefore, might have anticipated that exogenous expression of p190B, an inhibitor of Rho GTPases, would have tumor suppressor functions. However, our loss and gain of function studies indicate that p190B has pro-tumorigenic actions and that it may facilitate tumor formation and progression in part by modulating Rac1 activity. These results suggest that regulation of Rho GTPase activity in vivo
is likely more complex than what occurs in cultured cells.
Several lines of evidence suggest that the closely related p190A RhoGAP functions as a tumor suppressor. For example, p190A is located on chromosome 19q13.3, which is deleted in human gliomas [52
]. P190A also inhibits Ras-induced transformation in fibroblasts [53
] and suppresses platelet derived growth factor (PDGF) induced glioma formation in mice [54
]. Thus, these two related RhoGAPs may have distinct actions during tumorigenesis.