Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Apr 11.
Published in final edited form as: Cancer Cell. 2016 Mar 17;29(4):523–535. doi: 10.1016/j.ccell.2016.02.008

Spontaneous hepatocellular carcinoma after the combined deletion of Akt isoforms

Qi Wang 1,*, Wan-Ni Yu 1,*, Xinyu Chen 1, Xiao-ding Peng 1, Sang-Min Jeon 1,**, Morris J Birnbaum 3, Grace Guzman 4, Nissim Hay 1,2,***
PMCID: PMC4921241  NIHMSID: NIHMS761990  PMID: 26996309

Summary

Akt is frequently hyperactivated in human cancers and is targeted for cancer therapy. However, the physiological consequences of systemic Akt isoforms inhibition were not fully explored. We showed that while combined Akt1 and Akt3 deletion in adult mice is tolerated combined Akt1 and Akt2 deletion induced rapid mortality. Akt2−/− mice survived hepatic Akt1 deletion but all develop spontaneous hepatocellular carcinoma (HCC), which is associated with FoxO-dependent liver injury and inflammation. The gene expression signature of HCC bearing livers is similar to aggressive human HCC. Consistently, neither Akt1−/− nor Akt2−/− mice are resistant to diethylnitrosamine-induced hepatocarcinogenesis, and Akt2−/− mice display high incidence of lung metastasis. Thus, in contrast to other cancers, hepatic Akt inhibition induces liver injury that could promote HCC.

Graphical abstract

graphic file with name nihms761990u1.jpg

Introduction

The serine/threonine kinase, Akt, is perhaps the most frequently activated oncoprotein in human cancers. Akt is hyperactivated by multiple mechanisms, such as by mutations that hyperactivate the catalytic subunits of PI3K, mutations that inactivate the tumor suppressor PTEN, activation of growth factor receptors, and activating mutations in Akt isoforms themselves. The hyperactivation of PI3K/Akt signaling and its downstream effector mTORC1 in human cancers have prompted approaches that target this pathway for cancer therapy. Currently there are multiple clinical trials aimed at assessing the efficacy of PI3K/Akt inhibitors in cancer patients (Dienstmann et al., 2014; Fruman and Rommel, 2014). However, pan-PI3K and pan-Akt inhibitors inhibit the activity of the three Akt isoforms. Therefore, it became necessary to understand the consequences of systemic Akt isoforms deletion on the survival and physiology of adult mice.

The three Akt isoforms (Akt1-3), which are encoded by separate genes, share greater than 80% identity at the amino acid level. Gene targeting of the individual Akt isoforms in mice resulted in distinct phenotypes. Therefore, it is important to determine whether systemic Akt inhibition is therapeutic without eliciting adverse physiological consequences. Genetic studies in mice have demonstrated that the Akt1 deletion is well tolerated and can inhibit development of cancer in various mouse models (Chen et al., 2006; Hollander et al., 2011; Ju et al., 2007; Maroulakou et al., 2007; Skeen et al., 2006). By contrast, Akt2 deletion elicits insulin resistance and does not inhibit cancer development in mouse models (Maroulakou et al., 2007; Xu et al., 2012). However, those studies utilized germ line deletion of Akt isoforms, and thus they do not address the effect of Akt isoforms on tumor progression and do not emulate the drug therapy that is administered after tumor detection. Additionally, germ line deletion of both Akt1 and Akt2 is neonatal lethal, and the deletion of both Akt1 and Akt3 is embryonic lethal, while the deletion of both Akt2 and Akt3 does not cause lethality (reviewed in (Hay, 2011)). However, the consequences of systemic combined deletions of Akt isoforms in adult mice, which are more relevant to cancer therapy, are not known.

To address these issues, we deleted Akt isoforms systemically in adult mice. We recently showed that systemic deletion of Akt1 in adult p53−/− mice after tumor onset halted and regressed thymic lymphoma developed in these mice (Yu et al., 2015). In the present studies we determined the consequences of combined Akt isoforms deletion in adult mice.

Results

Combined deletion of Akt1 and Akt2 in adult mice induces rapid mortality

We addressed the effects of individual Akt isoform or combined Akt isoform systemic deficiency in adult mice. We generated Akt1f/f;R26CreERT2 mice, in which a construct expressing the Cre recombinase fused with a mutated estrogen receptor ligand binding domain that can be activated by tamoxifen (CreERT2) was knocked into the Rosa26 (R26) locus. Thus, upon injection of tamoxifen, Cre is activated to systemically delete the floxed allele (Ventura et al., 2007). We have shown that Akt1 can be systemically deleted in these adult mice following injection of tamoxifen and that the deletion did not elicit any overt phenotype but was able to regress thymic lymphoma in Trp53−/− mice (Yu et al., 2015). Akt1f/f;R26CreERT2 mice were crossed with either Akt2−/− or Akt3−/− mice to generate either Akt1f/f;Akt2−/−;R26CreERT2 or Akt1f/f;Akt3−/−;R26CreERT2 mice (Figure 1A). At 2 months of age, the mice received tamoxifen. Surprisingly, the Akt3−/− mice survived the deletion of Akt1 without any overt phenotype (Figure 1B). By contrast, the Akt2−/− mice did not survive the Akt1 systemic deletion and died within 2 weeks or 2 months after tamoxifen injection, depending on the level of Cre recombinase expression (heterozygous vs. homozygous; Figure 1B). The mortality appeared to be independent of the age at which the deletion of Akt1 was induced because we found that 6–8-month-old mice also died with similar kinetics. Immediately after the induction of the systemic deletion of Akt1 in Akt2−/− mice, we observed a transient increase in fed glucose levels that was followed by a decrease in glucose levels and hypoglycemia, depending on the level of Cre recombinase (Figure 1C). A marked decrease in body weight was observed after the deletion of Akt1 in Akt2−/− mice (Figure 1D). To further assess the effect of the combined Akt1 and Akt2 deletion in adult mice, we generated Akt1f/f;Akt2f/f; R26CreERT2 mice in which both Akt1 and Akt2 could be systemically deleted. Tamoxifen injection in Akt1f/f;Akt2f/f;R26CreERT2 mice induced extremely rapid mortality (Figure 1E) and a rapid loss of body weight (Figure S1A). Notably, Akt1f/f;Akt2f/f;R26CreERT2 mice died faster than Akt2−/−;Akt1f/f;R26CreERT2 mice following tamoxifen administration. Thus, the rapid mortality was not due to the germ line deletion of Akt2 and was also observed following acute systemic deletion of both Akt1 and Akt2. Interestingly, we observed a dramatic loss of fat tissue after systemic deletion of both Akt1 and Akt2 (Figure S1B), suggesting the utilization of fat for fatty acid oxidation and possibly a block in adipose differentiation, as observed in newborn mice carrying a germ line deletion of both Akt1 and Akt2 (Peng et al., 2003). We measured the glucose and insulin levels in these mice and found that both were elevated immediately after tamoxifen injection and reduced to below normal levels before death (Figs. 1F and 1G). We also observed increased circulating levels of IL-6, with dramatic elevations prior to death, suggesting severe inflammation possibly as a consequence of tissue damage (Figure 1H). Notably, systemic deletion of both Akt1 and Akt2 in adult mice was conducted in the FVB background, while the Akt2−/−;Akt1f/f;R26CreERT2 mice were in the C57BL/6 background, indicating that the observed phenotype was independent of the mouse strain. Finally, we found that the mortality rates of male and female mice were similar following systemic Akt1 and Akt2 deletion. We concluded that although both Akt1 and Akt3 are required for embryonic development, they are not required for adult viability. By contrast, germ line deletion of both Akt1 and Akt2 is not required for the development of embryos to term but is required for adult viability, and thus, complete ablation of the activity of Akt1 and Akt2 is not tolerated in adult mice. The pan-Akt inhibitor MK2206 has been used in vivo in mice with doses ranging from 60 mg/kg to 300 mg/kg. One of the most commonly used dose, is 120 mg/kg orally administered every two days. When we doubled this dose for a period of 3 weeks, MK2206 administration recapitulated the systemic deletion of Akt1 and Akt2 and reduced body weight (Figure S1C), and markedly induced IL-6 levels (Figure S1D). Although, we do not know the exact cause of death in response to the systemic deletion of Akt1 and Akt2, we hypothesize that the mice could not utilize external nutrients or absorb food; thus, they consumed fat as an energy and nutrient source. However, following the depletion of fat, they become hypoglycemic and die. Consistently, we found that the structure of the villi in the intestine was severely impaired following the induction of the Akt1 and Akt2 deletion (Figure S1E). This phenomenon was likely due to the inhibition of crypt cell proliferation, as measured by Ki67 staining (Figure S1F). Impaired intestinal structure and crypt cell proliferation was also observed after MK2206 administration (Figure S1G). Taken together, these results indicate that systemic inhibition of Akt1 and Akt2 could exert severe toxicity manifested by tissue damage, inflammation, and hypoglycemic shock.

Figure 1. The combined deletion of Akt1 and Akt2 in adult mice induces rapid mortality preceded by body weight loss and hypoglycemia.

Figure 1

Open in a new tab

A. Breeding scheme to generate Akt1f/f;Akt2−/−;R26CreERT2 and Akt1f/f;Akt3−/−;R26CreERT2 mice. B. Kaplan-Meier plot showing the survival of 2-month-old Akt1f/f;Akt2−/−;R26CreERT2+/− (n=20), Akt1f/f;Akt2−/−;R26CreERT2+/+ (n=10) and Akt1f/f;Akt3−/−;R26CreERT2+/+ (n=10) mice after tamoxifen injection for 5 consecutive days. C. Blood glucose levels in fed Akt1f/f;Akt2−/−, Akt1f/f;Akt2−/−R26CreERT2+/−, and Akt1f/f;Akt2−/−;R26CreERT2+/+ mice before and after tamoxifen injection (n=6, ±SEM, p<0.05 vs. before tamoxifen). D. Body weights of Akt1f/f;Akt2−/−, Akt1f/f;Akt2−/−;R26CreERT2+/−, and Akt1f/f;Akt2−/−R26CreERT2+/+ mice after tamoxifen injection (n=6, ±SEM, p<0.05 vs. Akt1f/f;Akt2−/− mice). E. Kaplan-Meier plot showing the survival of 2-month-old Akt1f/fAkt2−/−R26CreERT2+/+ mice after the tamoxifen injection (n=11). F. Glucose levels in Akt1f/f;Akt2−/−;R26CreERT2+/+ mice before and after the start of tamoxifen injection (n=4, ±SEM, p<0.05 tamoxifen vs. corn oil). G. Serum insulin levels in Akt1f/f;Akt2−/−;R26CreERT2+/+ mice before and after the start of tamoxifen injection and before death (n=4, ±SEM, p<0.03 tamoxifen vs. corn oil). H. Serum IL-6 levels in tamoxifen-injected Akt1f/f;Akt2−/−;R26CreERT2+/+ mice (n=4, ±SEM, p<0.03 tamoxifen vs. corn oil). See also Figure S1.

Hepatic deletion of Akt1 in Akt2−/− mice caused severe growth retardation

As shown in Figure 1, systemic deletion of both Akt1 and Akt2 in adult mice induced rapid mortality. However, hepatic deletion of Akt1 by albumin promoter-driven Cre in Akt2−/− mice (Akt1hep−/−;Akt2−/−) was tolerated. The Akt1hep−/−;Akt2−/− mice had higher blood glucose levels than Akt2−/− mice (Figure S2A) that were similar to those previously described for Akt2−/− mice in which Akt1 was deleted by alpha-fetoprotein (Afp) promoter-driven Cre (Lu et al., 2012). The mice also displayed glucose intolerance (Figure S2B). Despite higher glucose levels, the high insulin level detected in Akt2−/− mice was not significantly different after hepatic Akt1 deletion, and thus, Akt1hep−/−;Akt2−/− and Akt2−/− mice were phenotypically similar in terms of the insulin tolerance (Figure S2C, S2D).

Akt1hep−/−Akt2−/− mice displayed marked growth retardation with a decrease in body weight of approximately 40% compared with the mice with other genotypes (Figure S3A). The Akt1hep−/−;Akt2−/− mice had markedly reduced levels of IGF-1 in the blood (Figure S3B), which likely resulted in the growth retardation. The liver is the major organ responsible for generating circulating IGF1, wherein expression is transcriptionally induced primarily by growth hormone (GH) [reviewed in (Herrington et al., 2000; Klammt et al., 2008)]. However, circulating GH levels were not significantly different between Akt1hep−/−;Akt2−/− and Akt2−/− or WT mice (Figure S3C). Nonetheless, Igf1 mRNA levels in the livers of Akt1hep−/−;Akt2−/− mice were markedly lower than those in the livers of Akt2−/− mice (Figure 2A). GH induces IGF1 expression by binding to its cognate receptor on hepatocytes and activating the STAT5 transcription factor, which in turn translocates into the nucleus to induce Igf1 transcription (Leung et al., 2000). We found that STAT5 tyrosine phosphorylation, which is readout for its activity, was severely inhibited in the livers of Akt1hep−/−;Akt2−/− mice (Figure 2B). Thus, in addition to the known effect of GH, Akt activity was also required for STAT5 activation and the expression in the liver. GH receptor (GHR) expression has been reported to be induced by insulin (Leung et al., 2000) and is a potential transcriptional target of FoxO1 in the liver, as determined by Chip-seq (Shin et al., 2012). Hepatic GHR has also been reported to be downregulated in a FoxO1-dependent manner after hepatic deletion of IRS1 and IRS2 (Dong et al., 2008). Because FoxO transcription factors are phosphorylated and inhibited by the insulin-Akt axis, we examined the expression of GHR in the livers of Akt1hep−/−;Akt2−/− mice. Indeed, we found that the mRNA level of GHR was significantly reduced in Akt1hep−/−;Akt2−/− mouse livers compared with Akt2−/− mouse livers (Figure 2C). Thus, similar to the results reported for the hepatic deletion of IRS1 and IRS2 (Dong et al., 2008), GHR expression was suppressed in the absence of hepatic Akt1 and Akt2, likely by FoxO. This phenomenon could explain, at least in part, the requirement for Akt for GH-mediated expression of IGF1 in the liver (Figure 2D). However, we cannot completely exclude the possibility of additional mechanisms by which Akt affects STAT5 phosphorylation (see Discussion).

Figure 2. Hepatic deletion of Akt1 in Akt2−/− mice inhibits growth hormone receptor expression, STAT5 activation, and IGF1 expression, and induces early onset of HCC.

Figure 2

Open in a new tab

A. Quantitative RT-PCR showing hepatic Igf11 mRNA levels in Akt2−/− and Akt1hep−/−;Akt2−/− mice (n=4, ±SEM, p=0.0067). B. Hepatic STAT5 phosphorylation on Tyr694 in Akt2−/− or Akt1hep−/−;Akt2−/− mice. C. Level of growth hormone receptor mRNA in Akt2−/− and Akt1hep−/−;Akt2−/− mouse livers (n=4, ±SEM, p=0.013). D. Model illustrating the role of Akt in the GH/IGF1 axis (see text for details). E. Representative images of the livers of 6-month-old male mice of the indicated genotype. F. Quantification of the number of macroscopic tumors per liver in 6–7-month-old male mice (n=10, ±SEM, p=0.0001 Akt1hep−/−;Akt2−/− mice vs. Akt2−/−). G. Liver sections showing tumor lesions in 4- and 7-month-old Akt1hep−/−Akt2−/− mice (upper panel-low magnification; bottom panel-high magnification). The tumor areas are marked. See also Figure S2 and Figure S3.

Hepatic deletion of Akt1 in Akt2−/− mice induces early onset of spontaneous HCC, liver injury, and inflammation

Further analyses of the Akt1hep−/−;Akt2−/− mice revealed that all the Akt1hep−/−;Akt2−/− male mice developed HCC by 6–7 months of age (8–9 months for females; Figure 2E). This finding was not observed in Akt2−/− mice, even when one allele of Akt1 was deleted in the liver of the Akt2−/− mice (Akt1hep+/−;Akt2−/− mice) or when Akt1 was deleted in the liver of Akt2+/− mice (Akt1hep−/−;Akt2+/− mice) (Figure 2E and Figure S3D). An average of nineteen macroscopic tumors per liver was detected in the Akt1hep−/−;Akt2−/− mice, and no macroscopic tumors were observed in the mice with other genotypes (Figure 2F). Importantly, the hepatocarcinogenesis observed in Akt1hep−/−;Akt2−/− mice was a relatively rapid process compared with most mouse models of hepatocarcinogenesis (Heindryckx et al., 2009), and microscopic neoplastic lesions had been observed in 3–4-month-old male mice (Figure 2G). By comparison, hepatic activation of Akt by hepatic deletion of Pten induces HCC after 75 weeks (Horie et al., 2004). In both the 4-month and 7-month-old mice, two major types of tumors, clear cell and trabecular, were observed (Figure 2G). Akt2 is the major Akt isoform expressed in the liver and its deletion substantially reduced hepatic Akt activity, but this is not sufficient to induce HCC. Since even the deletion of Akt2 together with heterozygous deletion of Akt1 is not sufficient to induce HCC, we concluded that HCC is induced only when hepatic Akt activity is reduced below a threshold level. Considering that Akt is frequently hyperactivated in human cancers, these results were unexpected.

The initiation of and predisposition to liver carcinogenesis is often associated with liver injury and inflammation. Consistently, we found that serum levels of the liver enzymes alanine transaminase (ALT) and aspartate transaminase (AST) were markedly elevated in Akt1hep−/−;Akt2−/− mice (Figure 3A, B), indicating liver damage, which was not observed in either Akt1hep+/−;Akt2−/− or Akt2−/− mice. Moreover, serum IL-6 levels (Figure 3C) and liver TNFα mRNA (Figure 3D) were highly elevated in Akt1hep−/−;Akt2−/− mice compared with Akt2−/− mice, suggesting inflammation. Consistent with the identified inflammation, F4/80 staining revealed numerous liver macrophages (Kupffer cells) in the tumor sections obtained from the livers of 3–4-month and 7–8-month-old Akt1hep−/−;Akt2−/− mice (Figure 3E, F). In addition, plasma cells were frequently observed in the livers of Akt1hep−/−;Akt2−/− mice (Figure S4A). The tumors cells in the livers of Akt1hep−/−;Akt2−/− mice were highly proliferative, as indicated by Ki67 staining (Figure 3G). Interestingly, a high level of apoptosis, as measured by cleaved caspase 3, was observed mostly outside of the tumor lesions in the livers of Akt1hep−/−;Akt2−/− mice, while Ki67 staining was observed mostly inside the tumors (Figure 3G and Figure S4BC). Taken together, these results indicate that the deletion of Akt1 in the livers of Akt2−/− mice induces liver damage followed by inflammation and HCC development (Li et al., 2012).

Figure 3. Hepatic deletion of Akt1 in Akt2−/− mice induces liver injury, high serum levels of IL-6, inflammation, cell death, and cell proliferation.

Figure 3

Open in a new tab

A. Serum AST levels in 3-month and 6-month-old male mice with the indicated genotypes (n=4–6, ±SEM, p<0.05 Akt1hep−/−;Akt2−/− vs. Akt1hep+/−;Akt2−/− or Akt2−/− mice). B. Serum ALT levels in 3-month and 6-month-old male mice with the indicated genotypes (n=4–6, ±SEM, p<0.05 Akt1hep−/−;Akt2−/− vs. Akt1hep+/−;Akt2−/− or Akt2−/− mice). C. Serum IL-6 levels in 4–6-month-old Akt1hep−/−;Akt2−/− and Akt2−/− male mice (n=6, ±SEM, p<0.05 Akt1hep−/−;Akt2−/− vs. Akt2−/− mice). D. TNFα mRNA levels as measured by qRT-PCR in the livers of 4–6-month-old Akt1hep−/−;Akt2−/− and Akt2−/− male mice (n=4, ±SEM, p=0.045). E. Representative images of H&E and F4/80 staining in liver sections from 3–4-month-old Akt1hep−/−;Akt2−/− and Akt2−/− mice (N, non-tumor area; T, tumor area). Scale bar 100 μm, F. Images of H&E and F4/80 staining of liver sections from 7–8-month-old Akt1hep−/−;Akt2−/− and Akt2−/− mice. Scale bar 100 μm. Bar graph shows the quantification of F4/80 staining (n=4, ±SEM, *p=0.024). G. Images of liver sections from 7–8-month-old Akt1hep−/−;Akt2−/− and Akt2−/− male mice showing immunohistochemistry with either anti-Ki67 or anti-cleaved caspase 3 (CC3). Scale bar 100 μm. Bar graphs shows quantifications of the results (n=4, ±SEM; *p=0.031 **p=0.0005). See also Figure S4.

Liver injury and inflammation induced by hepatic deletion of Akt1 and Akt2 is FoxO-dependent

We conducted RNA-seq analyses of the livers of 6-month-old Akt2−/− and Akt1hep−/−;Akt2−/− mice. Although Akt2 was the major isoform expressed in the liver, the deletion of Akt1 in the livers of Akt2−/− mice significantly and markedly induced many FoxO target genes (Figure S5). We confirmed the induction of several FoxO target genes, such as Rbp1, Irs2, Igfbp1, and Sgk1, by qRT-PCR (Figure 4A). We also found that the pro-apoptotic FoxO target genes Fasl and Bcl2l11 (Bim) were significantly induced in Akt1hep−/−;Akt2−/− vs. Akt2−/− livers (Figure 4B), raising the possibility that liver injury and inflammation induced by hepatic deletion of Akt1 and Akt2 were dependent on FoxO. To verify this possibility, we employed Akt1f/f;Akt2f/f and Akt1f/f;Akt2f/f;Foxo1f/f mice (Lu et al., 2012). One-month-old mice were systemically injected with AAV8-Tbg-Cre, which is an adeno-associated virus that expresses Cre driven by the hepatocyte-specific promoter of the thyroxine-binding globulin (Tbg) gene. AAV8-Tbg-GFP was used as a control (Figure 4C). This strategy has been used previously to efficiently delete both Akt1 and Akt2 in hepatocytes (Lu et al., 2012). Hepatic deletion of both Akt1 and Akt2 induced liver injury, as manifested by high levels of ALT, which was not observed after combined hepatic deletion of Akt1, Akt2, and Foxo1 (Figure 4D). Consequently, hepatic deletion of both Akt1 and Akt2 also induced liver inflammation, which was not observed in the absence of FoxO1 (Figure 4E, F). The results indicate that even in adult mice the hepatic deletion of Akt1 and Akt2 elicits liver injury and inflammation, which are pro-tumorigenic. Indeed, these mice eventually develop macroscopic tumors with an average of 2–3 tumors per liver (Figure S6). While AAV8-Tbg-Cre injected Akt1f/f;Akt2f/f mice eventually develop liver tumors Akt1f/f;Akt2f/f;Foxo1f/f mice did not, although some abnormalities were observed (Figure S6A). This is also manifested by the dramatic increase in Ki67 staining in the Akt1hep−/−;Akt2hep−/− tumor bearing livers in comparison with Akt1hep−/−;Akt2hep−/−;Foxo1hep−/− livers (Figure S6B). Consistently Akt1hep−/−;Akt2hep−/− mice display high levels of circulating IL-6, while Akt1hep−/−;Akt2hep−/−;Foxo1hep−/− mice did not (Figure S6C). Taken together, these results show that hepatic deletion of both Akt1 and Akt2 induces chronic liver injury, inflammation, and HCC, which are largely dependent on the activation of FoxO1.

Figure 4. Liver injury and inflammation induced by hepatic deletion of Akt1 and Akt2 are FoxO1-dependent.

Figure 4

Open in a new tab

A. The mRNA level of known FoxO-target genes in the livers of Akt1hep−/−;Akt2−/− compared with Akt2−/− mice was measured by qRT-PCR (n=4 ±SEM, *p<0.05, **p<0.01). (B) Fasl and Bcl2l11 mRNAs in the livers of Akt1hep−/−;Akt2−/−and Akt2−/− mice were measured by qRT-PCR (n=4, ±SEM, p<0.05). (C). Schematic depicting systemic injection of AAV-Tbg-Cre or control AAV-Tbg-GFP for the hepatic deletions of either both Akt1 and Akt2 or Akt1, Akt2, and Foxo1 in Akt1f/f;Akt2f/f and Akt1f/f;Akt2f/f;Foxo1f/f mice, respectively. H. Serum ALT levels after AAV-Tbg-Cre or AAV-Tbg-GFP injection (n=4, ±SEM, *p<0.05). I. Representative liver section images immunostained with anti-F4/80 after deletion of either both Akt1 and Akt2 or Akt1, Akt2, and FoxO1. Scale bar 100 μm. J. Quantification of F4/80 immunostaining (n=4, ±SEM **p<0.005,GFP and Akt1f/f;Akt2f/f vs. Akt1f/f;Akt2f/f;Foxo1f/f mice). See also Figure S5 and Figure S6.

Liver tumors that develop in Akt1hep−/−Akt2−/− mice display relatively high levels of STAT3 phosphorylation and the expression of genes commonly induced in HCC

A high level of IL-6 in response to liver damage is pro-oncogenic in the liver, and it can induce the proliferation of liver cells via STAT3 activation (reviewed in (Taniguchi and Karin, 2014)). Consistently, we detected relatively high levels of STAT3 phosphorylation in the livers of Akt1hep−/−;Akt2−/− mice and, in general, higher STAT3 phosphorylation in tumors compared with non-tumor areas (Figure 5A). We found that apoptosis occurred mostly outside the area of the tumor and proliferation occurred mostly inside the tumor (Figure 3G, and Figure S4B, C). Therefore, we speculated that the dead cells could induce inflammation and higher levels of IL-6, which in turn activated STAT3 in the survived cells that then give rise to tumors. Although we initially observed relatively high Akt1 expression in the tumor bearing livers (Figure S7A), this expression was originated from non-hepatocytes such as hematopoeitc cells, fibroblasts, and endothelial cells, which reside at higher levels in the tumor area. When liver tumor sections were subjected to immune-histo-cytochemistry with Akt1 antibodies only the non-hepatocyte cells within the tumor area were stained positive (Fig 5B). Co-immunostaining with F4/80 and Akt1 clearly showed that the Akt1 positive cells are macrophages (Figure S7B). When hepatocytes derived from the tumors were cultured we found almost no Akt1 expression and Akt phosphorylation in comparison with hepatocytes derived from Akt2−/− liver (Figure 5C and Figure S7C, D). The residual Akt phosphorylation that was observed after cell culturing in some of the samples cannot be derived from Akt3 since both normal hepatocytes and HCC cells do not expresses detectable level of Akt3 (Figure S7D). It is therefore likely that the observed minimal Akt1 expression is derived from a small population of non-hepatocyte cells that were co-cultured with the hepatocytes. Genotyping analysis of the cultured tumor hepatocytes is consistent with this notion (Figure 5D and Figure S7E). This is consistent with previously described hepatic deletion of genes that promote liver cancer and in which residual expression of the deleted genes is due to the present of non-hepatocyte cells (Luedde et al., 2007; Umemura et al., 2014). We therefore, concluded that the hepatic tumor cells survive and proliferate despite Akt ablation possibly through other mutations and/or STAT3 activation. Interestingly, we found that IL-6 was also produced by hepatocytes in the liver, indicating autocrine IL-6 production (Figure 5E), although these hepatocytes were usually located in areas distant from the tumor area (Figure 5F).

Figure 5. Phospho-STAT3, and IL-6 expression in Akt1hep−/−;Akt2−/− liver.

Figure 5

Open in a new tab

A. Immunoblot showing the level of STAT3 phosphorylation in the livers of 6-month-old male mice with the indicated genotypes (NT, non-tumor area; T, tumor area). B. Representative liver tumor section showing immunohistochemistry with Akt1 antibodies. C. Immunoblot showing the expression and phosphorylation of Akt1, Akt2 and STAT3 in isolated hepatocytes derived from Akt1f/f;Akt2−/− liver and isolated hepatocytes derived from tumors in Akt1hep−/−;Akt2−/− liver. D. PCR genotyping showing the floxed and deleted Akt1 in hepatocytes isolated from Akt1f/f;Akt2−/− liver and from tumors in Akt1hep−/−;Akt2−/−liver. NS- indicates non-specific band. E. Representative images of IHC for IL-6 of Akt1hep−/−;Akt2−/− liver sections from different liver samples. Scale bar 100 μm. F. A representative image showing positively stained IL-6 cells and the location of the tumor (marked with a dashed line). Scale bar 500 μm. Inset shows a higher magnification of the IL-6-positive cells. Scale bar 100 μm. See also Figure S7.

Analysis of the RNA-seq data revealed that the expression of many genes that are induced in DEN-mediated HCC and in human HCC (He et al., 2013) are also induced in the livers of Akt1hep−/−;Akt2−/− mice. For instance, genes that are induced in DEN-mediated HCC, such as trefoil factor 3 (Tff3), the serine protease inhibitor Spink1, Ly6d, glypican 3 (Gpc3), alpha-fetoprotein (Afp), and Cd44, are also induced in the livers of Akt1hep−/−;Akt2−/− mice (Figure 6A).

Figure 6. Characterization of Akt1hep−/−;Akt2−/− liver bearing tumors, by gene expression.

Figure 6

Open in a new tab

A. Expression of genes that are frequently induced in mouse and human HCC in Akt1hep−/−;Akt2−/− livers vs. Akt2−/− livers of 6-month old male mice was measured by qRT-PCR (n=4, ±SEM, *p ≤ 0.05, **p<0.005). B. The expression of Fosl2 and associated target genes in the liver of 6-month old male Akt1hep−/−;Akt2−/− and Akt2−/− mice were determined by qRT-PCR (n=4, +/−SEM, p<0.05). C. Expression of Igf2bp1 mRNA in Akt1hep−/−;Akt2−/− liver when compared to Akt2−/− liver (n=3 +/−SEM, p=0.052). D. Expression of Igf2bp3 mRNA in Akt1hep−/−;Akt2−/− liver when compared to Akt2−/− liver (n=4, +/−SEM, **p<0.005). E. Expression of IGF2BP3 protein in Akt1hep−/−;Akt2−/− livers when compared to Akt2−/− livers. See also Figure S8.

Thorgeirsson and colleagues have analyzed gene expression in human HCC and based on this analysis, they found a distinct subtype of aggressive HCC (Lee et al., 2006). This subtype, which is associated with poor prognosis, expresses several embryonic genes and was characterized by the regulation of a network of genes that are associated with the induction of Fos, and FosL2 expression. Interestingly, we found that Fos and Fosl2 expression, as well as of several other genes in the network that are considered Fos and FosL2 or AP-1 targets, were markedly induced in Akt1hep−/−;Akt2−/− livers bearing tumors (Figure S5 and Figure 6B). Finally, in addition to Afp, H19, and Gpc3, we discovered several other fetal genes that were induced in human HCC, such as Scd2 and Bex1 (Figure S8). An increase in Igf2bp1 expression was observed (Figure 6C), but the most profound finding was the induction of Igf2bp3 mRNA and protein in Akt1hep−/−Akt2−/− livers (Figure 6D, E). The IGF2BPs, which include IGF2BP1, IGF2BP2, and IGF2BP3, were originally discovered as RNA binding proteins, which stabilize IGF2 mRNA. However, they also affect mRNA stability and the expression of many other genes (reviewed in (Bell et al., 2013; Lederer et al., 2014)). While IGF2BP2 is widely expressed in all mouse adult tissues, IGF2BP1 and IGF2BP3 are expressed mostly in embryonic tissues and are not expressed in normal adult liver (Lederer et al., 2014). However, a high expression level of IGF2BP1 and IGF2BP3 has been observed in many human cancers, including HCC, in which they have been associated with the induction of cell proliferation, cell survival, and invasiveness (Bell et al., 2013; Lederer et al., 2014). Both IGF2BP1 and IGF2BP3 (also known as IMP3) are highly and frequently expressed in human HCC, and their expression is correlated with aggressiveness, invasiveness, and poor prognosis (Gutschner et al., 2014; Jeng et al., 2008; Wachter et al., 2012). Because of its selective expression in aggressive HCC, IGF2BP3 was suggested as a prognostic marker (Wachter et al., 2012). Thus, based on studies in human HCC the tumors developed in the livers of Akt1hep−/−;Akt2−/− mice are considered aggressive.

All together the results show that the HCC is initiated by cell death of hepatocytes in which both Akt1 and Akt2 were deleted, resulting in liver injury followed by inflammation, which are FoxO1-dependent. The liver injury, inflammation, and high IL-6 level are considered to be prerequisites for HCC development. IL-6 activates STAT3 and promotes compensatory proliferation in hepatocytes that escaped cell death, subsequently promoting tumor development. In addition, based on gene expression, Akt1hep−/−;Akt2−/− liver tumors appeared to be characterized as undifferentiated and of embryonic origin, and the high expression levels of genes, such as IGF2BP1 and, particularly, IGF2BP3, facilitated tumorigenesis. Finally, autocrine secretion of IL-6 could further facilitate HCC development.

It is possible that the high serum insulin level in Akt1hep−/−;Akt2−/− mice also facilitated tumor development. To further determine the role of Akt1 and Akt2 as well as insulin levels in HCC, two weeks old Akt1−/− and Akt2−/− male mice were subjected to intraperitoneal injection of 25 mg/ml diethylnitrosamine (DEN) and hepatic tumor incidence was analyzed 12 months later. Unexpectedly, the incidences of macroscopic tumors in Akt1−/− and Akt2−/− mice are not significantly different from that of WT mice (Figure 7A). Notably, Akt2 is the major Akt isoform expressed in the liver, yet its deletion resulted in not only no resistance to hepatocarcinogenesis but also a rather marked increase in the incidence of metastasis. A high incidence of lung tumors was observed in Akt2−/− mice, wherein 50% of the mice showed macroscopic lung tumors, with a total of 12 lung tumors in 16 mice (Figure 7 B, C). By comparison, only 10% of the WT mice developed macroscopic lung tumors, with a total of 3 tumors in 20 mice. This difference was much more profound when microscopic lesions in the lung were quantified (Figure 7D WT mice. These tumors were derived from primary liver tumors based on pathology (Figure 7E), and they expressed alpha-fetoprotein (Afp), which was not expressed in the lung but only in liver tumors (Figure 7F). Therefore, we concluded that these tumors were metastases that had originated from primary liver tumors.

Figure 7. DEN-induced hepatocarcinogenesis in Akt1−/− and Akt2−/− mice.

Figure 7

Open in a new tab

A. The incidence of macroscopic tumors ≥1 cm in diameter in the livers of Akt1−/−, Akt2−/−, and WT mice, 12 months after exposure to DEN. B. The incidence of macroscopic metastatic tumors in the lungs of Akt1−/−, Akt2−/−, and WT mice 12 months after exposure to DEN (results are presented as the mean ±SEM). C. Images depicting lungs with macroscopic lung tumors. Scale bar, 10 mm. D. The number of microscopic metastatic tumors in the lung of Akt2−/−, and WT mice, 12 months after exposure to DEN. Data are presented as the mean ±SEM. E. H&E stained lung sections from WT, Akt1−/−, and Akt2−/− mice. Tumor is marked in the Akt2−/− lung section. F. RT-PCR analysis of Afp expression in tumors developed in the lung of Akt2−/− mice. G. Representative images of livers bearing tumors and corresponding lungs from DEN treated Akt2−/− and WT mice (scale bars 10 mm). Total number of macroscopic liver tumors and number of tumors ≥1 cm in diameter are shown in parenthesis. Fed glucose and insulin levels in the corresponding mice are shown adjacent to the images. H. Immunoblot showing p-Akt1 in tumors derived from either WT or Akt2−/− livers. I. Immunoblot showing p-ERK in tumors derived from either WT or Akt2−/− livers. Upper panel - total Akt (t-Akt); Middle panel - p-ERK; bottom panel – total ERK (t-ERK).

The major physiological difference between Akt2−/− mice and either Akt1−/− or WT mice is the insulin resistance and high blood insulin level in Akt2−/− mice. We found up to 10-fold higher level of insulin in Akt2−/− mice (Chen et al., 2009). Indeed, the Akt2−/− mice that developed metastases to the lung had extremely high levels of serum insulin compared with the WT mice that did not develop lung metastasis (Figure 7G). The high level of insulin in Akt2−/− mice leads to a compensatory activation of Akt1 (Xu et al., 2012), and Figure 7H), as well as Erk1/2 activation (Figure 7I), that could be pro-metastatic. Since Akt2 is the major Akt isoform expressed in the liver and its deletion reduced about two third of total hepatic Akt activity, these results further support the notion that inhibiting hepatic Akt activity accelerates HCC progression. Thus, the inhibition of hepatic Akt activity induces liver injury and inflammation in a FoxO1-dependent manner, which in turn induces compensatory proliferation via STAT3 activation. Subsequently, additional acquired mutations lead to HCC (Figure 8A). High level of insulin may facilitate the development of HCC. Thus, the deficiency of Akt2 only, which induces high levels of insulin, facilitates HCC progression and metastasis (Figure 8B). Consequently, the results could explain, at least in part, why fatty liver that inhibits hepatic Akt activity facilitates HCC.

Figure 8. Illustration depicting HCC development and progression in Akt1hep−/−;Akt2−/− and Akt2−/− mice.

Figure 8

Open in a new tab

A. The deletion of hepatic Akt1 and Akt2 induces liver damage and apoptosis, which is largely dependent on the activation of FoxO. FoxO induces inflammation and IL-6 and TNFα production by macrophages and plasma cells. IL-6 induces compensatory proliferation by activating STAT3 in the survived hepatocytes. Eventually, these cells give rise to HCC with gene signatures similar to human HCC. Autocrine production of IL-6 by hepatocytes was also observed. The high level of insulin in the mice may facilitate HCC progression. B. Exposure of Akt2−/− mice that display high level of insulin to DEN-induced hepatocarcinogenesis induces high incidence of metastasis to the lung.

Discussion

Because Akt is frequently activated in human cancers and inhibitors of Akt are being used in clinical trials for cancer therapy, it has become necessary to understand the consequences of the deletion of Akt isoforms on the survival and physiology of adult mice. To address this important issue, we systemically deleted Akt isoforms in adult mice. In the present study, we showed that deletion of both Akt1 and Akt3 in adult mice was tolerated, while deletion of both Akt1 and Akt2 elicited rapid mortality. The rapid mortality was preceded by a transient increase followed by a decrease in glucose and insulin levels and subsequent cachexia and hypoglycemia. Our results indicated that the mice developed tissue damage and inflammation. The tissue damage disrupted the villi in the intestine, which impaired food absorption. The mice lost body fat possibly because of its utilization as an alternative source for nutrients.

The most intriguing and unexpected observation that was uncovered in our studies is the early onset of spontaneous hepatocarcinogenesis in Akt1hep−/−Akt2−/− mice with complete penetrance. Akt is perhaps the most frequently activated oncoprotein in human cancers, and activation of Akt was implicated in the genesis of HCC (Moeini et al., 2012; Shearn and Petersen, 2015). Furthermore, hepatic deletion of Pten in mice induces HCC (Horie et al., 2004). Inhibitors of Akt and PI3K are being used for cancer therapy (Dienstmann et al., 2014; Fruman and Rommel, 2014), and the germ line deletion of Akt1 in mice was reported to significantly impair tumor development in several mouse models of cancer (Chen et al., 2006; Hollander et al., 2011; Ju et al., 2007; Maroulakou et al., 2007; Skeen et al., 2006). Moreover, systemic deletion of Akt1 after tumor onset in Trp53−/− mice halts and regresses thymic lymphoma (Yu et al., 2015). By contrast our results showed that hepatic inactivation of Akt1 and Akt2 induces HCC with markedly faster onset than by the activation of Akt through hepatic Pten deletion, which occurs 75 weeks after birth (Horie et al., 2004). Although this observation may constitute an apparent paradox, it is consistent with the role of obesity, and a high fat diet (HFD) in HCC. It has been documented that obesity and a fatty liver can promote human HCC, and in mouse models, HFD has been shown to accelerate HCC (Baffy et al., 2012; Park et al., 2010; Sun and Karin, 2012). Under these conditions, HCC develops despite insulin resistance and the inhibition of hepatic Akt activity. The contribution of HFD and fatty liver to HCC genesis has been attributed to oxidative and endoplasmic reticulum stress (Nakagawa et al., 2014). However, our results suggest that hepatic Akt inhibition alone, as a consequence of obesity, can induce liver damage and inflammation and subsequently HCC, adding a potential mechanism of pro-tumorigenesis in the fatty liver. Notably, the hepatic deletion of Tsc1 in mice, which inhibits hepatic Akt activity, induces spontaneous HCC albeit with slower kinetics compared to what we observed in Akt1hep−/−Akt2−/− mice (Menon et al., 2012). The development of HCC in these mice was attributed to the hyperactivation of mTORC1 (Menon et al., 2012). However, based on our studies, it is possible that hepatic Akt inactivation in this mouse model also contributed to HCC.

HCC development is often preceded by liver injury and inflammation and, consequently, the production of pro-tumorigenic cytokines, in particular IL-6, which induces STAT3 activation in hepatocytes. Consistently, we observed hepatic cell death and liver injury, as well as inflammation and elevated circulating IL-6 levels, in liver tumor-bearing Akt1hep−/−;Akt2−/− mice. The liver injury and inflammation appeared to be largely FoxO1-dependent because the hepatic deletion of both Akt1 and Akt2 induced liver injury and inflammation that were blunted when FoxO1 was also deleted. Indeed, although Akt2 is the major isoform expressed in the liver (Chen et al., 2006), deletion of Akt1 in the liver of Akt2−/− mice further increased FoxO target gene expression. This phenomenon could be explained by the high level of insulin in Akt2−/− mice, which hyperactivates hepatic Akt1 in these mice and therefore causes a reduction in total hepatic Akt activity that is lower than that anticipated for the full activation of hepatic FoxO. Following the deletion of hepatic Akt1 in Akt2−/− mice, the reduction in total Akt activity becomes sufficient to fully activate FoxO. Thus, the high activity of FoxO could induce apoptosis, at least in part, by elevating the expression of the pro-apoptotic genes Fasl and Bcl2l11. Interestingly, apoptosis largely occurred in hepatocytes located outside or distal to the area of the tumor. The dying hepatocytes elicit inflammation by macrophages and other cells, which in turn results in increased IL-6 and other pro-tumorigenic cytokines. IL-6 then activates STAT3 in the surviving hepatocytes that subsequently give rise to the tumors. Although our results support this mechanism we cannot completely exclude other not mutually exclusive mechanisms that contribute to the rapid onset of HCC in Akt1hep−/−;Akt2−/− mice. It is worth noting, for example, that mice with hepatic deletion of Stat5 are more susceptible to hepatocarcinogenesis, and that the hepatic deletion of Stat5 induces STAT3 activity (Cui et al., 2007). Thus, it is possible that STAT5 inactivation in the livers of Akt1hep−/−;Akt2−/− mice contributed to the increased STAT3 activity and hepatocarcinogenesis. Notably, we found that the source of IL-6 in Akt1hep−/−;Akt2−/− livers was not necessarily only macrophages but also certain hepatocytes that begin to express IL-6. Interestingly, the IL-6 positive cells were usually located at a site that was distant from the tumor site. Although the origin of these cells is not clear, it is possible that they evolved as a consequence of the Akt1 and Akt2 deletion to maintain survival via IL-6-dependent and Akt-independent mechanisms. Autocrine secretion of IL-6 has also been observed in human HCC, and a recent study showed that in DEN-induced HCC in mice, certain progenitor cells start expressing IL-6 to promote their malignant progression (He et al., 2013)

The histopathology revealed two types of tumors that were morphologically consistent with clear cell HCC and trabecular HCC in Akt1hep−/−;Akt2−/− livers. Gene expression analysis demonstrated the expression of many genes in the Akt1hep−/−;Akt2−/− liver that are also induced in human HCC and in a mouse model of DEN-induced HCC (He et al., 2013; Lee et al., 2006). The gene expression analysis showed that several embryonic genes that are not normally expressed in the adult liver are induced in the tumor bearing Akt1hep−/−;Akt2−/− liver. Of particular interest is the expression of the network of genes associated with the induction of Fos, Fosl1 and Fosl2, which have been implicated in driving tumorigenesis in an aggressive subtype of human HCC (Lee et al., 2006). Among the embryonic genes that were strongly expressed in the Akt1hep−/−;Akt2−/− liver were Igf2bp1 and Igf2bp3. Both IGF2BP1 and IGF2BP3 are highly and frequently expressed in human HCC, and their expression has been correlated to aggressiveness, invasiveness, and a poor prognosis (Gutschner et al., 2014; Jeng et al., 2008; Wachter et al., 2012). Because of its selective expression in aggressive HCC, IGF2BP3 has been suggested to be a prognostic marker (Wachter et al., 2012). Thus, based on studies in human HCC, the tumors that develop in the livers of Akt1hep−/−;Akt2−/− mice are considered aggressive.

Finally, we found that neither Akt1−/− nor Akt2−/− mice were resistant to DEN-induced HCC. Importantly, Akt2−/− mice displayed a marked increase in the incidence of lung metastases. We attributed the high incidence of metastasis to the very high level of circulating insulin in Akt2−/− mice. As shown previously, the high level of insulin in Akt2−/− mice could hyperactivate other Akt isoforms and potentially other downstream signaling pathways in different tissues (Xu et al., 2012). The activation of these signaling pathways could contribute to the high incidence of metastasis in Akt2−/− mice. However, we cannot completely exclude the possibility that the deletion of Akt2 in other tissues facilitate metastasis in a non-cell autonomous manner.

Taken together, the results presented here provided unexpected conceptual paradigm whereby Akt activity, which is required for cancer development in general, is not required for the initiation of liver cancer, and its inhibition could promote liver cancer and liver cancer progression. Our results suggest that systemic and efficient inhibition of both Akt1 and Akt2 by pan-Akt inhibitors might not be well tolerated. Furthermore, because orally or systemically delivered drugs can accumulate at the highest levels in the liver, they could inhibit the hepatic activity of Akt to a similar extent as that observed for the hepatic deletion of Akt1 and Akt2. Unlike genetic deletion, drug therapy is not usually administered for an unlimited time period. However, our results suggest that the ablation of Akt1 and Akt2 could elicit a pre-disposition to HCC by inducing liver damage and inflammation. Because liver damage and inflammation could occur even after transient inhibition of hepatic Akt activity, it would subsequently increase the risk for liver cancer. This process could be exacerbated by obesity, which also induces insulin resistance and inhibits hepatic Akt activity. When we treated the mice with moderately high dose of the pan-Akt inhibitor MK2206, we recapitulated the consequences of systemic Akt1 and Akt2 deletion. However, the mice recovered after the treatment was stopped. We followed these mice for 6 months, and thus far we did not find liver tumors, although we cannot completely exclude the possibility that liver tumors will eventually develop. The high systemic insulin levels following pan-Akt inhibition or Akt2 inhibition could counteract the benefits of cancer therapy, due to the hyperactivation of other pro-tumorigenic signaling pathways downstream of insulin. Because HCC development is characterized by a long latency period, HCC after treatment with pan-Akt inhibitors might not be manifested early but rather late after treatment.

Experimental Procedures

Mice

Akt1−/−, Akt2−/−, and Akt3−/− mice have been previously described (Chen et al., 2009; Chen et al., 2001; Peng et al., 2003). Akt1f/f mice were generated by Ozgene (Australia), and the generation of Akt1f/f;R26RCreERT2 mice has been described elsewhere (Yu et al., 2015). Akt1f/f;Akt2f/f, and Akt1f/f;Akt2f/f;FoXo1f/f mice have been previously described (Lu et al., 2012). Akt1f/f;Akt2−/−;R26CreERT2 and Akt1f/f;Akt3−/−;R26CreERT2 mice were generated by crossing Akt1f/f;R26CreERT2 with either Akt2−/− or Akt3−/− mice. R26CreERT2 knock-in mice (strain 01XAB) were obtained from Tyler Jacks’ laboratory (Massachusetts Institute of Technology, USA). Akt1hep−/−;Akt2−/− mice were generated by crossing Akt1/f/f;Akt2−/− with the albumin-CRE mouse strain (obtained from the Jackson Laboratory, Bar Harbor, ME, USA). All mice are in the C57Bl6 background. The generation of Akt1f/f;Akt2f/f;R26CreERT2 was performed by crossing Akt1f/f;R26CreERT2 in the FVB background with Akt2f/f mice in the FVB background. Systemic injection of AAV into Akt1f/f;Akt2f/f or Akt1f/f;Akt2f/f;Foxo1f/f mice was conducted as previously described (Lu et al., 2012)Tamoxifen injection for systemic activation of Cre recombinase has been previously described (Yu et al., 2015). All animal experiments were approved by the University of Illinois at Chicago Institutional Animal Care and Use Committee, as required by United States Animal Welfare Act, and the NIH’s policy.

Histopathology, immunocytochemistry, and immunofluorescence

Tissue samples were fixed in 10% formalin overnight and then processed and embedded in paraffin. The sections (5 μm) were stained with hematoxylin and eosin. For immunohistochemistry, the tissue sections were incubated at 95°C in 0.01 M citric acid (pH 6.0) for 20 min, followed by a 30-min cool-down. The sections were then treated with 0.3% hydrogen peroxide to block endogenous peroxides and then washed with PBS and blocked with serum. The samples were incubated with primary antibodies overnight at 4°C, followed by incubation with the biotinylated secondary antibody and avidin-biotin-complexes (ABC). Signals were visualized with diaminobenzidine. Some of the sections were lightly counterstained with hematoxylin. The primary antibodies included anti-Ki67 (cell signaling), anti-F4/80 (Abcam), anti-cleaved caspase 3 (Cell Signaling), anti-IL-6 (Abcam). Biotin-conjugated secondary antibody kits, avidin–biotin complexes, and diaminobenzidine were purchased from Vector Laboratories. Ki67, cleaved caspase-3 and F4/80-positive cells were counted in 10 fields at 200× magnification per slide using ImageJ software (NIH, Bethesda, MD, USA).

For immunofluorescence, slides were blocked with 10% goat serum in PBS for 30min and incubated with rabbit anti-AKT1 (Abcam ab32505, Lot: GR145853-2) and Rat anti-F4/80 (Abcam ab6640, Lot: GR169072-1) at 4°C overnight. After washing in TNT buffer (0.1 M Tris-HCl, pH7.5, 150 mM NaCl, 0.05% Tween 20), slides were incubated with TRITC-conjugated anti-rabbit for Akt1 and FITC anti-rat for F4/80 secondary antibodies, and then stained with 1μg/mL DAPI for 10 min at room temperature. Slides were mounted with mounting medium from Vector Laboratories. Slides were examined and imaged by a Zeiss LSM 700 confocal microscope using the manufacturer’s imaging software ZEN (ZEISS Efficient Navigation).

Other detailed experimental procedures are described in the supplementary information.

Supplementary Material

supplement

Significance.

The results presented here suggest that the development of Akt isoform specific inhibitors should be considered. The liver injury, inflammation, and unexpected spontaneous HCC developed when both hepatic Akt1 and Akt2 are deleted might have implications for cancer therapy. Moreover, while neither Akt1 nor Akt2 are required for hepatocarcinogenesis, the deficiency of Akt2 promotes high incidence of lung metastasis following hepatocarcinogenesis. The results raise the possibility that systemic administration of pan-PI3K/Akt inhibitors might increase the risk for liver injury and inflammation, and possibly HCC, if hepatic Akt activity is markedly decreased. The results suggest that hepatic Akt inhibition in fatty liver could be one mechanism by which obesity promotes HCC and that pan-PI3K/Akt inhibitors could have undesired outcomes especially for obesity-mediated HCC.

Highlights.

  • Early onset of HCC after hepatic deletion of Akt1 in Akt2−/− mice.

  • The HCC is preceded by liver injury and inflammation in FOXO1-dependent manner.

  • Akt2 deficiency promotes lung metastasis after hepatocarcinogenesis.

  • Akt is required for hepatic STAT5 phosphorylation and IGF1 expression

Acknowledgments

These studies were supported by NIH grants R01AG016927 and R01CA090764, by VA Merit Award BX000733 to N.H., and by NIH grant RO1 DK56886 to M.J.B.

Footnotes

Accession numbers

The accession number for the RNA sequencing data reported in this manuscript is GEO: GSE77862.

Author contributions

Q.W., W.N.Y, X.D.P., X.C., and N.H. designed the experiments and analyzed the data. W.N.Y, generated mice and conducted the initial experiments. Q.W. and X.C completed the experiments for publication. X.C. generated the mice with systemic Akt1 and Akt2 deletion and conducted experiments with these mice. X.D.P. conducted the hepatocarcinogenesis experiments. S.M.J. generated the Akt3−/− mice with systemic deletion of Akt1 and analyzed them. G.G. made the pathological evaluations of the tissue sections. M.J.B provides some of the mice for the experiments. Q.W., W.N.Y., and NH wrote the paper.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Baffy G, Brunt EM, Caldwell SH. Hepatocellular carcinoma in non-alcoholic fatty liver disease: an emerging menace. J Hepatol. 2012;56:1384–1391. doi: 10.1016/j.jhep.2011.10.027. [DOI] [PubMed] [Google Scholar]
  2. Bell JL, Wachter K, Muhleck B, Pazaitis N, Kohn M, Lederer M, Huttelmaier S. Insulin-like growth factor 2 mRNA-binding proteins (IGF2BPs): post-transcriptional drivers of cancer progression? Cellular and molecular life sciences: CMLS. 2013;70:2657–2675. doi: 10.1007/s00018-012-1186-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Chen ML, Xu PZ, Peng X, Chen WS, Guzman G, Yang X, Di Cristofano A, Pandolfi PP, Hay N. The deficiency of Akt1 is sufficient to suppress tumor development in Pten+/− mice. Genes Dev. 2006;20:1569–1574. doi: 10.1101/gad.1395006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Chen WS, Peng XD, Wang Y, Xu PZ, Chen ML, Luo Y, Jeon SM, Coleman K, Haschek WM, Bass J, et al. Leptin deficiency and beta-cell dysfunction underlie type 2 diabetes in compound Akt knockout mice. Mol Cell Biol. 2009;29:3151–3162. doi: 10.1128/MCB.01792-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chen WS, Xu PZ, Gottlob K, Chen ML, Sokol K, Shiyanova T, Roninson I, Weng W, Suzuki R, Tobe K, et al. Growth retardation and increased apoptosis in mice with homozygous disruption of the Akt1 gene. Genes Dev. 2001;15:2203–2208. doi: 10.1101/gad.913901. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cui Y, Hosui A, Sun R, Shen K, Gavrilova O, Chen W, Cam MC, Gao B, Robinson GW, Hennighausen L. Loss of signal transducer and activator of transcription 5 leads to hepatosteatosis and impaired liver regeneration. Hepatology. 2007;46:504–513. doi: 10.1002/hep.21713. [DOI] [PubMed] [Google Scholar]
  7. Dienstmann R, Rodon J, Serra V, Tabernero J. Picking the point of inhibition: a comparative review of PI3K/AKT/mTOR pathway inhibitors. Mol Cancer Ther. 2014;13:1021–1031. doi: 10.1158/1535-7163.MCT-13-0639. [DOI] [PubMed] [Google Scholar]
  8. Dong XC, Copps KD, Guo S, Li Y, Kollipara R, DePinho RA, White MF. Inactivation of hepatic Foxo1 by insulin signaling is required for adaptive nutrient homeostasis and endocrine growth regulation. Cell Metab. 2008;8:65–76. doi: 10.1016/j.cmet.2008.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Fruman DA, Rommel C. PI3K and cancer: lessons, challenges and opportunities. Nature reviews Drug discovery. 2014;13:140–156. doi: 10.1038/nrd4204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Gutschner T, Hammerle M, Pazaitis N, Bley N, Fiskin E, Uckelmann H, Heim A, Grobeta M, Hofmann N, Geffers R, et al. Insulin-like growth factor 2 mRNA-binding protein 1 (IGF2BP1) is an important protumorigenic factor in hepatocellular carcinoma. Hepatology. 2014;59:1900–1911. doi: 10.1002/hep.26997. [DOI] [PubMed] [Google Scholar]
  11. He G, Dhar D, Nakagawa H, Font-Burgada J, Ogata H, Jiang Y, Shalapour S, Seki E, Yost SE, Jepsen K, et al. Identification of liver cancer progenitors whose malignant progression depends on autocrine IL-6 signaling. Cell. 2013;155:384–396. doi: 10.1016/j.cell.2013.09.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Heindryckx F, Colle I, Van Vlierberghe H. Experimental mouse models for hepatocellular carcinoma research. International journal of experimental pathology. 2009;90:367–386. doi: 10.1111/j.1365-2613.2009.00656.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Herrington J, Smit LS, Schwartz J, Carter-Su C. The role of STAT proteins in growth hormone signaling. Oncogene. 2000;19:2585–2597. doi: 10.1038/sj.onc.1203526. [DOI] [PubMed] [Google Scholar]
  14. Hollander MC, Maier CR, Hobbs EA, Ashmore AR, Linnoila RI, Dennis PA. Akt1 deletion prevents lung tumorigenesis by mutant K-ras. Oncogene. 2011;30:1812–1821. doi: 10.1038/onc.2010.556. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Horie Y, Suzuki A, Kataoka E, Sasaki T, Hamada K, Sasaki J, Mizuno K, Hasegawa G, Kishimoto H, Iizuka M, et al. Hepatocyte-specific Pten deficiency results in steatohepatitis and hepatocellular carcinomas. J Clin Invest. 2004;113:1774–1783. doi: 10.1172/JCI20513. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Jeng YM, Chang CC, Hu FC, Chou HY, Kao HL, Wang TH, Hsu HC. RNA-binding protein insulin-like growth factor II mRNA-binding protein 3 expression promotes tumor invasion and predicts early recurrence and poor prognosis in hepatocellular carcinoma. Hepatology. 2008;48:1118–1127. doi: 10.1002/hep.22459. [DOI] [PubMed] [Google Scholar]
  17. Ju X, Katiyar S, Wang C, Liu M, Jiao X, Li S, Zhou J, Turner J, Lisanti MP, Russell RG, et al. Akt1 governs breast cancer progression in vivo. Proc Natl Acad Sci U S A. 2007;104:7438–7443. doi: 10.1073/pnas.0605874104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Klammt J, Pfaffle R, Werner H, Kiess W. IGF signaling defects as causes of growth failure and IUGR. Trends Endocrinol Metab. 2008;19:197–205. doi: 10.1016/j.tem.2008.03.003. [DOI] [PubMed] [Google Scholar]
  19. Lederer M, Bley N, Schleifer C, Huttelmaier S. The role of the oncofetal IGF2 mRNA-binding protein 3 (IGF2BP3) in cancer. Semin Cancer Biol. 2014;29:3–12. doi: 10.1016/j.semcancer.2014.07.006. [DOI] [PubMed] [Google Scholar]
  20. Lee JS, Heo J, Libbrecht L, Chu IS, Kaposi-Novak P, Calvisi DF, Mikaelyan A, Roberts LR, Demetris AJ, Sun Z, et al. A novel prognostic subtype of human hepatocellular carcinoma derived from hepatic progenitor cells. Nat Med. 2006;12:410–416. doi: 10.1038/nm1377. [DOI] [PubMed] [Google Scholar]
  21. Leung KC, Doyle N, Ballesteros M, Waters MJ, Ho KK. Insulin regulation of human hepatic growth hormone receptors: divergent effects on biosynthesis and surface translocation. J Clin Endocrinol Metab. 2000;85:4712–4720. doi: 10.1210/jcem.85.12.7017. [DOI] [PubMed] [Google Scholar]
  22. Li Y, Tang ZY, Hou JX. Hepatocellular carcinoma: insight from animal models. Nat Rev Gastroenterol Hepatol. 2012;9:32–43. doi: 10.1038/nrgastro.2011.196. [DOI] [PubMed] [Google Scholar]
  23. Lu M, Wan M, Leavens KF, Chu Q, Monks BR, Fernandez S, Ahima RS, Ueki K, Kahn CR, Birnbaum MJ. Insulin regulates liver metabolism in vivo in the absence of hepatic Akt and Foxo1. Nat Med. 2012;18:388–395. doi: 10.1038/nm.2686. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Luedde T, Beraza N, Kotsikoris V, van Loo G, Nenci A, De Vos R, Roskams T, Trautwein C, Pasparakis M. Deletion of NEMO/IKKgamma in liver parenchymal cells causes steatohepatitis and hepatocellular carcinoma. Cancer Cell. 2007;11:119–132. doi: 10.1016/j.ccr.2006.12.016. [DOI] [PubMed] [Google Scholar]
  25. Maroulakou IG, Oemler W, Naber SP, Tsichlis PN. Akt1 ablation inhibits, whereas Akt2 ablation accelerates, the development of mammary adenocarcinomas in mouse mammary tumor virus (MMTV)-ErbB2/neu and MMTV-polyoma middle T transgenic mice. Cancer Res. 2007;67:167–177. doi: 10.1158/0008-5472.CAN-06-3782. [DOI] [PubMed] [Google Scholar]
  26. Menon S, Yecies JL, Zhang HH, Howell JJ, Nicholatos J, Harputlugil E, Bronson RT, Kwiatkowski DJ, Manning BD. Chronic activation of mTOR complex 1 is sufficient to cause hepatocellular carcinoma in mice. Sci Signal. 2012;5:ra24. doi: 10.1126/scisignal.2002739. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Moeini A, Cornella H, Villanueva A. Emerging signaling pathways in hepatocellular carcinoma. Liver cancer. 2012;1:83–93. doi: 10.1159/000342405. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Nakagawa H, Umemura A, Taniguchi K, Font-Burgada J, Dhar D, Ogata H, Zhong Z, Valasek MA, Seki E, Hidalgo J, et al. ER stress cooperates with hypernutrition to trigger TNF-dependent spontaneous HCC development. Cancer Cell. 2014;26:331–343. doi: 10.1016/j.ccr.2014.07.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Nogueira V, Sundararajan D, Kwan JM, Peng XD, Sarvepalli N, Sonenberg N, Hay N. Akt-dependent Skp2 mRNA translation is required for exiting contact inhibition, oncogenesis, and adipogenesis. EMBO J. 2012;31:1134–1146. doi: 10.1038/emboj.2011.478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Park EJ, Lee JH, Yu GY, He G, Ali SR, Holzer RG, Osterreicher CH, Takahashi H, Karin M. Dietary and genetic obesity promote liver inflammation and tumorigenesis by enhancing IL-6 and TNF expression. Cell. 2010;140:197–208. doi: 10.1016/j.cell.2009.12.052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Peng XD, Xu PZ, Chen ML, Hahn-Windgassen A, Skeen J, Jacobs J, Sundararajan D, Chen WS, Crawford SE, Coleman KG, Hay N. Dwarfism, impaired skin development, skeletal muscle atrophy, delayed bone development, and impeded adipogenesis in mice lacking Akt1 and Akt2. Genes Dev. 2003;17:1352–1365. doi: 10.1101/gad.1089403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Shearn CT, Petersen DR. Understanding the Tumor Suppressor PTEN in Chronic Alcoholism and Hepatocellular Carcinoma. Advances in experimental medicine and biology. 2015;815:173–184. doi: 10.1007/978-3-319-09614-8_10. [DOI] [PubMed] [Google Scholar]
  33. Shin DJ, Joshi P, Hong SH, Mosure K, Shin DG, Osborne TF. Genome-wide analysis of FoxO1 binding in hepatic chromatin: potential involvement of FoxO1 in linking retinoid signaling to hepatic gluconeogenesis. Nucleic acids research. 2012;40:11499–11509. doi: 10.1093/nar/gks932. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Skeen JE, Bhaskar PT, Chen CC, Chen WS, Peng XD, Nogueira V, Hahn-Windgassen A, Kiyokawa H, Hay N. Akt deficiency impairs normal cell proliferation and suppresses oncogenesis in a p53-independent and mTORC1-dependent manner. Cancer Cell. 2006;10:269–280. doi: 10.1016/j.ccr.2006.08.022. [DOI] [PubMed] [Google Scholar]
  35. Sun B, Karin M. Obesity, inflammation, and liver cancer. J Hepatol. 2012;56:704–713. doi: 10.1016/j.jhep.2011.09.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Taniguchi K, Karin M. IL-6 and related cytokines as the critical lynchpins between inflammation and cancer. Seminars in immunology. 2014;26:54–74. doi: 10.1016/j.smim.2014.01.001. [DOI] [PubMed] [Google Scholar]
  37. Umemura A, Park EJ, Taniguchi K, Lee JH, Shalapour S, Valasek MA, Aghajan M, Nakagawa H, Seki E, Hall MN, Karin M. Liver damage, inflammation, and enhanced tumorigenesis after persistent mTORC1 inhibition. Cell Metab. 2014;20:133–144. doi: 10.1016/j.cmet.2014.05.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Ventura A, Kirsch DG, McLaughlin ME, Tuveson DA, Grimm J, Lintault L, Newman J, Reczek EE, Weissleder R, Jacks T. Restoration of p53 function leads to tumour regression in vivo. Nature. 2007;445:661–665. doi: 10.1038/nature05541. [DOI] [PubMed] [Google Scholar]
  39. Wachter DL, Kristiansen G, Soll C, Hellerbrand C, Breuhahn K, Fritzsche F, Agaimy A, Hartmann A, Riener MO. Insulin-like growth factor II mRNA-binding protein 3 (IMP3) expression in hepatocellular carcinoma. A clinicopathological analysis with emphasis on diagnostic value. Histopathology. 2012;60:278–286. doi: 10.1111/j.1365-2559.2011.04091.x. [DOI] [PubMed] [Google Scholar]
  40. Xu PZ, Chen ML, Jeon SM, Peng XD, Hay N. The effect Akt2 deletion on tumor development in Pten(+/−) mice. Oncogene. 2012;31:518–526. doi: 10.1038/onc.2011.243. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Yu WN, Nogueira V, Sobhakumari A, Patra KC, Bhaskar PT, Hay N. Systemic Akt1 Deletion after Tumor Onset in p53(−/−) Mice Increases Lifespan and Regresses Thymic Lymphoma Emulating p53 Restoration. Cell Rep. 2015;12:610–621. doi: 10.1016/j.celrep.2015.06.057. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplement
NIHMS761990-supplement.pdf (8MB, pdf)

RESOURCES