Loss of Akt1 or Akt2 delays mammary tumor onset and suppresses tumor growth rate in MTB-IGFIR transgenic mice

The insulin-like growth factor (IGF) family has been implicated in a number of human cancers including breast cancer [1‐5]. In particular, the type I IGF receptor or IGF-IR has been found to be expressed at high levels in 39-93% of human breast cancers [6‐9]. Originally, the IGF-IR was associated with luminal breast cancer however more recent studies have found the IGF-IR in all breast cancer subtypes [9‐14]. Two different transgenic mouse models have also shown the importance of IGF-IR in mammary tumorigenesis. In a study by Carboni et al. [15] expression of a constitutively active IGF-IR in mammary epithelial cells induced mammary tumor development. The other transgenic model was developed in our lab and this model overexpressed the wild type IGF-IR in mammary epithelial cells in a doxycycline inducible manner [16]. Overexpression of wild type IGF-IR also resulted in the formation of mammary tumors.

One important signalling molecule downstream of the IGF-IR is Akt. Akt is a serine-threonine kinase that lies downstream of PI3K signaling [17‐21]. There are 3 Akt isoforms in mammals and each is transcribed from separate genes [22‐25]. Based on genetic ablation of each isoform in mice it appears that Akt1 regulates cell survival and/or proliferation based on the observation that Akt1^−/− mice display increased perinatal lethality and are smaller in size [26]. Akt2 appears to regulate glucose homeostasis and Akt2^−/− mice develop insulin-resistant diabetes while Akt3 appears important in the brain as Akt3^−/− have impaired brain development [27, 28]. The function of Akt isoforms with respect to breast cancer has been investigated in cell lines and transgenic models of mammary tumorigenesis. In MDA-MB-231 human breast cancer cells, which express low levels of activated Akt, the overexpression of constitutively active forms of Akt1 or Akt2 inhibited cell proliferation and migration with little effect on apoptosis [29]. In a non-transformed human mammary epithelial cell line (MCF-10A) genetically altered to express high levels of IGF-IR, Akt1 downregulation reduced proliferation and enhanced migration while Akt2 downregulation reduced proliferation but had not effect on migration [30].

The function of Akt isoforms has been studied in two transgenic mammary tumor models, MMTV-neu and MMTV-PyMT. MMTV-neu transgenic mice constitutively overexpress erbB2 in mouse mammary epithelial cells using the mouse mammary tumor virus (MMTV) promoter. MMTV-PyMT transgenic mice express the polyoma virus middle T antigen in mammary epithelial cells using the MMTV promoter. In both models overexpression of constitutively active Akt1 was shown to delay mammary tumor formation but had no effect on metastasis while the overexpression of activated Akt2 did not affect tumor latency but did enhance tumor metastasis [31]. In addition, ablation of Akt1 in MMTV-neu or MMTV-PyMT transgenics inhibited tumor formation while the ablation of Akt2 accelerated tumor formation [32]. In addition, loss of Akt1 in MMTV-neu tumors enhanced their invasiveness.

In our MTB-IGFIR transgenic mice, we found that Akt1 or Akt2 ablation significantly increased in tumor latency and decreased tumor growth rate. Moreover, loss of Akt1 or Akt2 did not alter tumor histology or cytokeratin expression.

Previous characterization of the MTB-IGFIR transgenic mice showed that mammary tumors expressing high levels of IGF-IR also contained high levels of phosphorylated Akt [16]. However, it was not determined which of the Akt isoforms were important in the MTB-IGFIR mammary tumors. Therefore, as an initial experiment, the levels of the three Akt isoforms were evaluated in normal, wild type (WT) mammary tissue and in mammary tumors of MTB-IGFIR transgenic mice. Evaluation of Akt1, Akt2 and Akt3 revealed that Akt1 and Akt2 can be detected in both WT mammary tissue and mammary tumors (Figure 1) while Akt3 could not be detected in either WT mammary tissue or mammary tumors.

Since Ak1 and Akt2 were consistently expressed in the mammary tumors, MTB-IGFIR mice were crossed with Akt1^−/− and Akt2^−/− mice to determine the roles of Akt1 and Akt2 in mammary tumorigenesis. To minimize any effects that mouse strain may play on mammary tumorigenesis, the Akt1^−/− and Akt2^−/− mice were backcrossed 7 times into a FVB/N background so that all mice were in a FVB/N background (MTB-IGFIR transgenic mice were created and maintained in a FVB/N background). MTB-IGFIR/Akt1^−/−, MTB-IGFIR/Akt2^−/−, and MTB-IGFIR mice were fed food supplemented with 2 g of doxycycline per kilogram of rodent chow beginning when the mice were 21 days of age. Mice overexpressing IGF-IR and containing Akt1 and Akt2 developed tumors at approximately 56.5 days of age or 35.5 days after the initiation of doxycycline supplemented food. This onset is similar to what we previously published for these mice [16]. In the MTB-IGFIR/Akt1^−/− mice, tumor onset was 47.3 ± 3.1 days after supplying the mice with doxycycline supplemented food (Table 1). This tumor onset was approximately 12 days longer than the MTB-IGFIR mice and this difference was statistically significant. In the MTB-IGFIR/Akt2^−/− mice tumor onset was 43.6 ± 3.6 days after the addition of doxycycline to the food and this delay in tumor onset was also statistically different from the MTB-IGFIR mice (Table 1). Graphs showing the tumor onset are presented in Figure 2. There were no obvious differences in the number of tumors that developed in each animal or the number of animals with lung metastasis (Table 1).

Tumor growth rates were also evaluated using the formula for Specific Growth Rate (SGR) [33]. It was observed that mammary tumors in MTB-IGFIR/Akt1^−/− mice had SGRs that were approximately half that of the MTB-IGFIR mice. This difference in growth rate was significant. The mammary tumors in MTB-IGFIR/Akt2^−/− mice had growth rates that were approximately 36% slower than MTB-IGFIR mice and this difference was also statistically significant. Figure 3 shows a scatter plot with best fit exponential tumor growth curves to illustrate the differences in mammary tumor growth rates. As indicated in these plots there was some variability regarding growth rates of mammary tumors in MTB-IGFIR/Akt1^−/− and MTB-IGFIR/Akt2^−/− mice, however, the majority of the tumors grew at a slower rate than the mammary tumors of MTB-IGFIR mice.

To determine whether tumor proliferation was affected in either the MTB-IGFIR/Akt1^−/− or MTB-IGFIR/Akt2^−/− mice Ki67 immunohistochemistry was performed. It was observed that tumor cell proliferation in the MTB-IGFIR/Akt1^−/− tumors was reduced approximately 55% compared to MTB-IGFIR tumors while proliferation rates in the MTB-IGFIR/Akt2^−/− tumors were reduced approximately 20% (Figure 4). The data was however quite variable and thus neither result were statistically significant.

To confirm protein levels in the mammary tumors, western blotting was performed for IGF-IR, Akt1, Akt2, phosphorylated Akt (pAkt), Erk1/2, phosphorylated Erk1/2 (pErk1/2), Stat3 and phosphorylated Stat3 (pStat3). As shown in Figure 5, mammary tumors with normal levels of Akt1 and Akt2 had similar levels of the IGF-IR as mammary tumors null for Akt1 or null for Akt2 (Figure 5A,B). The levels of Akt1 were undetectable in MTB-IGFIR/Akt1^−/− mammary tumors while the levels of Akt2 were similar in MTB-IGFIR mammary tumors, and mammary tumors from MTB-IGFIR/Akt1^−/− mice. Similarly, the levels of Akt2 were undetectable in MTB-IGFIR/Akt2^−/− mice while Akt1 levels were similar in MTB-IGFIR mammary tumors and mammary tumors from MTB-IGFIR/Akt2^−/− mice. The levels of Akt3 could be detected at very low but similar levels in the mammary tumors from three different genotypes (data not shown). Despite the loss of Akt1 or Akt2, the mammary tumors from MTB-IGFIR/Akt1^−/− and MTB-IGFIR/Akt2^−/− mice had levels of phosphorylated Akt similar to the mammary tumors that developed in MTB-IGFIR mice. The levels of some of the signalling molecules downstream of phosphorylated Akt were also found at similar levels in all genotypes except for higher levels of phosphorylated Erk1/2 associated with MTB-IGFIR/Akt2^−/− tumors (Figure 5A,B).

Since Akt1 and Akt2 have been implicated in epithelial to mesenchymal transition in some mammary models [30, 35, 36], immunohistochemical analysis of cytokeratins 5, 8, 14 and 18 was performed to determine whether the type of mammary tumors that developed in MTB-IGFIR/Akt1^−/− or MTB-IGFIR/Akt2^−/− mice differed from the tumors that developed in MTB-IGFIR mice. The majority of the mammary tumors in each genotype contained cells that expressed moderate or high levels of cytokeratin 8 (Figure 6A-C) and/or 18 (data not shown) indicating these cells were luminal in nature. However, some tumors contained clusters of cytokeratin 5 and 14 positive cells and these clusters were independent of Akt genotype (Figure 6D-I). This mixture of tumor with luminal and basal characteristics was observed in the original characterization of the MTB-IGFIR mammary tumors [16].

Since Akt is one of the most hyperactive kinases in cancer [17] it is an attractive therapeutic target. However, there are three Akt isoforms and our understanding about the functions of the individual Akt isoforms in normal and cancerous tissue is incomplete. In our study we found that loss of either Akt1 or Akt2 produced a small but significant delay in tumor onset and a small but significant decrease in tumor growth rate.

Delayed mammary tumor onset and reduced tumor growth rate observed in the MTB-IGFIR/Akt1^−/− mice was consistent with the findings of Maroulakou et al. [32] which showed that loss of Akt1 significantly delayed tumor onset and reduced tumor growth in MMTV-neu and MMTV-PyMT transgenic mice. However, loss of Akt1 in the MMTV-neu and MMTV-PyMT transgenic mice had a much more dramatic impact on tumor onset than that observed in the MTB-IGFIR transgenic mice. Maroulakou et al. [32] also observed that loss of Akt2 enhanced mammary tumorigenesis, a result not observed in the MTB-IGFIR/Akt2^−/− mice. In addition, Maroulakou et al. [32] observed that loss of Akt1 reduced lung metastasis in MMTV-PyMT but not MMTV-neu transgenic mice. Neither loss of Akt1 nor Akt2 significantly affected metastasis in the MTB-IGFIR transgenic mice however it should be noted that decreased metastatic rate in the MTB-IGFIR transgenic mice may be difficult to detect as these tumors inherently have a relatively low metastatic rate.

There are several differences in the studies involving the MMTV-neu or MMTV-PyMT transgenics and our MTB-IGFIR transgenics. First, the transgene used to produce the mammary tumors was different. In our case, the IGF-IR drove mammary tumorigenesis while in the study by Maroulakou et al. [32], either ErbB2 or the polyoma virus middle T antigen were used. Although IGF-IR and ErbB2 are both tyrosine kinase receptors that initiate similar signalling cascades, tumors induced by these two transgenes are distinct. Using gene expression profiles, we previously showed that our MTB-IGFIR mammary tumors did not cluster with the MMTV-neu or MMTV-PyMT mammary tumors [37]. The MTB-IGFIR mammary tumors clustered most closely with human basal tumors [37] while the MMTV-neu mammary tumors have been reported to cluster more closely with human luminal breast cancers [38]. The MMTV-PyMT transgenics, like the MMTV-neu transgenics, clustered most closely with human luminal breast cancers [38].

Another potential significant difference was the strain of the mice. The study by Maroulakou et al. [32] did not specify the mouse strains used however Akt1^−/− and Akt2^−/− mice are typically in a C57BL/6 background while MMTV-neu and MMTV-PyMT transgenic mice are typically in a FVB background. Therefore, it is unclear whether this study was performed on a mixed strain background. In our study, Akt1^−/− and Akt2^−/− mice were backcrossed 7 times into a FVB background and mated with MTB-IGFIR transgenic mice which are on a FVB background. Therefore, our studies were performed on an essentially pure genetic background. It should be noted that we also attempted to backcross our MTB-IGFIR transgenic mice into a C57BL/6 background however even on the first backcross, these mice failed to form mammary tumors (unpublished observations).

In our study, alternative signaling pathways appeared to compensate for the loss of Akt1 or Akt2. This was evident by the fact that Akt phosphorylation did not decrease in mice lacking either Akt1 or Akt2. Since the total level of Akt phosphorylation did not diminish despite the absence of Akt1 or Akt2, it appears that the tumor cells compensated by increasing the phosphorylation of the remaining Akt isoforms. An increase in Erk phosphorylation was also observed in the MTB-IGFIR/Akt2^−/− tumors compared to MTB-IGFIR tumors. Erk1/2 signaling promotes survival and may have contributed to tumor cell survival in the absence of Akt2. Compensation between these pathways has been observed in mammary tumors in previously published studies [39, 40]. It is also possible that signaling pathways not evaluated in this study may have compensated for the absence of Akt1 or Akt2. The paper by Maroulakou et al. [32] did not evaluate the activation of intracellular signaling molecules in their tumors so comparisons to our study cannot be made.

Despite previous reports implicating Akt in EMT [30, 35, 36] we found no evidence that loss of Akt1 or Akt2 affected tumor cell morphology in the MTB-IGFIR transgenic mice. However, since EMT was not a focus of this study this property was not extensively examined in the current study. EMT is a process whereby epithelial cells convert to a mesenchymal morphology and express genes typically associated with mesenchymal cells (reviewed in [41‐44]). EMT is thought to increase the invasiveness of breast cancers [45‐49].

Although loss of Akt1 or Akt2 significantly inhibited mammary tumor onset and growth rates these effects were very modest. It is probable that the loss of Akt1 or Akt2 was compensated for by the remaining Akt isoforms and/or alternative signaling pathways in the tumor cells. Compensation has important clinical implications in that tumors are unlikely to significantly respond to therapeutic agents targeting individual signaling pathways or individual protein isoforms within signaling pathways. Future studies using Akt inhibitors for the treatment of breast cancer should employ Akt inhibitors that target both Akt1 and Akt2 in combination with inhibitors that target additional signaling pathways such as the MAPK signaling pathway.