Mechanobiology June 26th - June 2nd 2016

Mechanobiology: June 26th  - June 2nd 2016

Nrf2 participates in regulating maternal hepatic adaptations to pregnancy
Yuhong Zou, Min Hu, Qi Bao, Jefferson Y. Chan, Guoli Dai

Summary

Pregnancy induces widespread adaptive responses in maternal organ systems including the liver. The maternal liver exhibits significant growth by increasing the number and size of hepatocytes, by largely unknown mechanisms. Nrf2 mediates cellular defense against oxidative stress and inflammation and also regulates liver regeneration. To determine whether Nrf2 is involved in the regulation of maternal hepatic adaptations to pregnancy, we assessed the proliferation and size of maternal hepatocytes and the associated molecular events in wild-type and Nrf2-null mice at various stages of gestation. We found that wild-type maternal hepatocytes underwent proliferation and size reduction during the first half, and size increase without overt replication during the second half, of pregnancy. Although pregnancy decreased Nrf2 activity in the maternal liver, Nrf2 deficiency caused a delay in maternal hepatocyte proliferation, concomitant with dysregulation of the activation of Cyclin D1, E1, and, more significantly, A2. Remarkably, as a result of Nrf2 absence, the maternal hepatocytes were largely prevented from reducing their sizes during the first half of pregnancy, which was associated with an increase in mTOR activation. During the second half of pregnancy, maternal hepatocytes of both genotypes showed continuous volume increase accompanied by persistent activation of mTOR. However, the lack of Nrf2 resulted in dysregulation of the activation of the mTOR upstream regulator AKT1 and the mTOR target p70SK6 and thus disruption of the AKT1/mTOR/p70S6K pathway, which is known to control cell size. This suggests an mTOR-dependent and AKT1- and p70S6K-independent compensatory mechanism when Nrf2 is deficient. In summary, our study demonstrates that Nrf2 is required for normal maternal hepatic adjustments to pregnancy by ensuring proper regulation of the number and size of maternal hepatocytes.

Introduction

Pregnancy is characterized by a series of coordinated physiological adjustments in major maternal organ systems to meet the metabolic demands of the development and growth of the placenta and fetus (Smith et al., 1998; Nielsen et al., 1999; Audus et al., 2002; Shingo et al., 2003; Sweeney et al., 2006; Bustamante et al., 2008; Huang et al., 2009; Kim et al., 2010). Among the maternal adaptive responses to pregnancy is the enlargement of the maternal liver, which has previously been described in several reports (Kennaway and Kennaway, 1944; Kennedy et al., 1958; Smith, 1975; Mesbah and Baldwin, 1983; Hollister et al., 1987; Dickmann et al., 2008). Recently, significant insights have been gained into this pregnancy-dependent phenomenon. Our work and that of others have demonstrated that the maternal liver responds to pregnancy by marked growth driven by maternal hepatocyte hyperplasia and/or hypertrophy (Bustamante et al., 2010; Milona et al., 2010; Dai et al., 2011) concomitant with changes in hepatic gene expression profiles (Bustamante et al., 2010). However, the mechanisms mediating the maternal hepatic adaptations to pregnancy remain elusive.

Nuclear factor erythroid 2-related factor 2 (Nrf2) belongs to a family of transcription factors containing a basic leucine zipper (bZip) region and is expressed predominantly in metabolic organs, such as the liver (Chan et al., 1996). Nrf2 regulates the constitutive and/or inducible expression of a battery of genes encoding the components of cellular defense systems against oxidative stress and inflammation, as thoroughly reviewed by several groups recently (Osburn and Kensler, 2008; Hu et al., 2010; Klaassen and Reisman, 2010; Villeneuve et al., 2010; Wakabayashi et al., 2010a; Martín-Montalvo et al., 2011; Taguchi et al., 2011; Bataille and Manautou, 2012; Copple, 2012). Moreover, Nrf2 was shown to modulate the hepatic regenerative response to liver mass loss (Beyer et al., 2008; Wakabayashi et al., 2010b). Our previous study indicated that pregnancy-induced maternal liver growth and liver regeneration show both similarities and differences in molecular mechanisms (Dai et al., 2011). Thus, the aim of this study was to determine whether Nrf2 participates in mediating maternal hepatic adaptive responses to pregnancy. We found that Nrf2 is indeed involved in the regulation of maternal liver adaptations to the physiological stressor.

Results

Nrf2 deficiency leads to abnormal maternal liver growth response to pregnancy

To determine whether Nrf2 regulates the maternal liver growth in response to pregnancy, we first evaluated the changes in maternal liver-to-body weight ratio at different stages of pregnancy in wild-type and Nrf2-null mice (Fig. 1A). In a non-pregnant state, Nrf2-null mice showed significantly lower liver-to-body weight ratios than their wild-type littermates. The observation indicates that Nrf2 deficiency leads to a smaller liver in female mice, similar to what has been reported in male mice (Beyer et al., 2008; Huang et al., 2010). During pregnancy, maternal livers displayed significant increases in maternal liver-to-body weight ratio in comparison with the non-pregnant state in both genotypes of mice. However, when compared between wild-type and Nrf2-null mice, the genotype-dependent difference in liver-to-body weight ratio in non-pregnant mice was narrowed at certain stages of pregnancy, especially on gestation day 10. The data suggested genotype and gestation stage-dependent changes in maternal liver growth. Thus, we next evaluated the magnitudes of maternal liver enlargement during the course of gestation in each genotype of mice, using percent change in liver weight as compared with the non-pregnant state as a parameter (Fig. 1B). The result showed that Nrf2-null mice exhibited significantly higher magnitudes of maternal liver enlargement than wild-type mice on gestation days 8, 10 and 18. Hence, Nrf2 is required for normal adjustment of maternal liver size, especially during the first half of pregnancy. The finding demonstrates that Nrf2 plays a regulatory role in maternal hepatic growth response to pregnancy.

Fig. 1.

Maternal liver growth during pregnancy in wild-type (Nrf2+/+) and Nrf2-null (Nrf2−/−) mice. Maternal livers were collected from non-pregnant (NP) and pregnant (gestation days 8, 10, 11, 13, 15 and 18) wild-type and Nrf2-null mice and weighed. Maternal liver-to-body weight ratios (A) and the percentage changes in maternal liver weight as compared with the non-pregnant state (B) are presented. Data are expressed as the means ± s.d. (n = 3–5). ‘a’ and ‘b’ indicate P<0.05 compared with non-pregnant wild-type and Nrf2-null mice, respectively. *P<0.05, between wild-type and Nrf2-null mice.

Pregnancy downregulates maternal hepatic Nrf2 activity

Pregnancy may affect the functional state of maternal hepatic Nrf2. To test this hypothesis, we first measured Nrf2 mRNA expression in the maternal livers of wild-type mice at various stages of pregnancy by quantitative real-time polymerase chain reaction (qRT-PCR) (Fig. 2A). The result showed that pregnancy did not increase mRNA levels of maternal hepatic Nrf2 at any of the time points selected. Instead, a significant downregulation in Nrf2 mRNA expression was observed on gestation days 8 and 11. Next, we assessed Nrf2 protein levels in the nuclei of the maternal liver cells by western blotting (Fig. 2B). Marked changes in maternal hepatic nuclear Nrf2 protein were not observed throughout pregnancy. NAD(P)H:quinone oxidoreductase 1 (NQO-1) is a typical target gene of Nrf2 (Chan et al., 2001; Ishii et al., 2002; Aleksunes et al., 2006). Even under pathological conditions, the mRNA and protein expression and enzyme activity of NQO-1 are regulated by Nrf2 (Aleksunes et al., 2006). Therefore, estimating Nrf2 activity with the NQO-1 level is reliable. As pregnancy advanced, the gene expression of maternal hepatic NQO-1 was decreased on gestation day 8, returned to pre-pregnancy levels on gestation day 10, and subsequently gradually decreased until term (Fig. 2C). Aldehyde oxidase 1 (AOX1) is another direct target gene of Nrf2 (Maeda et al., 2012). Maternal hepatic AOX1 exhibited an mRNA expression pattern similar to that of NQO-1 (Fig. 2D). Remarkably, pregnancy-dependent downregulation of AOX1 on gestation days 8, 11, 13 and 15 was more evident than that of NQO-1. The observation suggests that AOX1 may be a marker that is more sensitive than NQO-1 in monitoring Nrf2 activity. NQO-1 protein expression followed a similar pattern as its mRNA expression in wild-type, but was dysregulated in Nrf2-null, maternal livers (Fig. 2E). The results indicate that maternal hepatic Nrf2 activity is gestational stage dependent; it is temporarily decreased on gestation day 8 during the first half of pregnancy and progressively reduced during the second half of pregnancy. Generally speaking, pregnancy decreases maternal liver Nrf2 activity.

Fig. 2.

Assessment of maternal hepatic Nrf2 functional state during pregnancy. (A) Nrf2 mRNA expression in the maternal liver during pregnancy. Total RNA was prepared from liver tissues of non-pregnant (NP) and pregnant (gestation days 8, 10, 11, 13, 15 and 18) wild-type mice. Hepatic mRNA levels of Nrf2 were measured by qRT-PCR and are expressed as the mean fold changes relative to NP controls (±s.d.; n = 3 for each group). *P<0.05 compared with NP controls. (B) Liver nuclear Nrf2 protein assay. Liver nuclear extracts were prepared from NP and pregnant (gestation days 8, 10, 11, 13, 15 and 18) wild-type (Nrf2+/+) mice and NP Nrf2-null (Nrf2−/−) mice. Nuclear proteins were quantified and subjected to western blotting analysis using a Nrf2 antibody. Lamin B1 protein levels were used as loading controls. Each lane shows a sample prepared from an individual mouse. (C,D) Expression of Nrf2 target genes Nqo-1 and Aox1 in the maternal liver during pregnancy. Total RNA samples from the experiment described in A were used for measuring hepatic mRNA expression of NQO-1 (C) and AOX1 (D) by qRT-PCR, which is expressed as the mean fold changes relative to NP controls (±s.d.; n = 3 for each group). *P<0.05, compared with NP controls. (E) Expression of hepatic NQO-1 protein during pregnancy. Liver tissues from NP and pregnant (gestation days 8, 10, 11, 13, 15 and 18) wild-type and Nrf2-null mice were homogenized. Western blotting was performed on tissue homogenates using a NQO-1 antibody. GAPDH protein levels were used as loading controls.

Nrf2 absence causes a delay in maternal hepatocyte proliferation during pregnancy

Nrf2 may be associated with maternal hepatocyte proliferative response to pregnancy. To confirm this, we performed Ki-67 immunostaining (Fig. 3A) and quantified proliferating hepatocytes (Fig. 3B) in maternal liver sections prepared from both genotypes of mice at various stages of pregnancy. In wild-type mice, maternal hepatocytes exhibited robust proliferative response during the first half of pregnancy. In Nrf2-null mice, proliferating maternal hepatocytes were less on gestation day 10 but more on gestation day 11 compared with that in wild-type mice. Apparently, maternal hepatocyte replication was delayed when Nrf2 was absent. The data indicate that Nrf2 participates in the regulation of maternal hepatocyte proliferation induced by pregnancy.

Fig. 3.

Maternal hepatocyte proliferation during pregnancy. Livers were collected from non-pregnant (NP) and pregnant (gestation days 8, 10, 11, 13, 15 and 18) wild-type (Nrf2+/+) and Nrf2-null (Nrf2−/−) mice, fixed in formalin, and embedded in paraffin. Liver sections were prepared, and Ki-67 immunostaining was performed. The nuclei of Ki67-positive cells were stained dark brown. (A) Representative liver sections of NP and pregnant (gestation days 8, 10, 11 and 13) wild-type and Nrf2-null mice were immunohistochemically stained for Ki-67. (B) Ki67-positive hepatocytes were counted (400× magnification), and the results are shown as the means ± s.d. per field (n = 3 to 5). ‘a’ and ‘b’ indicate P<0.05 compared with NP wild-type and Nrf2-null mice, respectively. *P<0.05, between wild-type and Nrf2-null mice.

To gain insight into how Nrf2 is associated with maternal hepatocyte replication, protein expression of several cell cycle components were assessed in both genotypes of mice throughout gestation (Fig. 4A). Maternal hepatic Cyclin D1 protein expression displayed a genotype-independent pattern, persisting during the first half and then declining during the second half of pregnancy. However, genotype-dependent variations in Cyclin D1 protein levels were observed. More evidently, Nrf2 deficiency led to higher expression of Cyclin D1 protein on gestation day 8. As pregnancy progressed, Cyclin E1 exhibited gradually increased protein expression in wild-type maternal livers, beginning on gestation day 10, whereas its activation was not observed until gestation day 11 in Nrf2-null maternal livers. Thus, Nrf2 absence caused a delay in maternal hepatic Cyclin E1 activation. Strikingly, Nrf2 null mutation exerted potent effects on hepatic Cyclin A2. First, Nrf2-null mice expressed a higher level of hepatic Cyclin A2 protein than wild-type mice in the non-pregnant state. Second, Cyclin A2 protein was activated on gestation days 8, 10 and 13 in wild-type maternal livers, whereas its activation lasted throughout the course of pregnancy when Nrf2 was lacking. Furthermore, on gestation day 8, the maternal liver expressed markedly higher Cyclin A2 protein in Nrf2-null mice compared with their wild-type littermates. Taken together, Nrf2 plays a negative role in maternal hepatic Cyclin A2 protein expression in both non-pregnant and pregnant mice.

Fig. 4.

Expression of hepatic Cyclins during pregnancy. (A) Protein expression of hepatic Cyclins D1, E1, and A2 during pregnancy. Livers were collected from non-pregnant (NP) and pregnant (gestation days 8, 10, 11, 13, 15 and 18) wild-type (Nrf2+/+) and Nrf2-null (Nrf2−/−) mice. Western blotting was performed in liver homogenates using antibodies against the proteins indicated. GAPDH protein levels were used as loading controls. (B) mRNA expression of hepatic Cyclin A2 during pregnancy. Total RNA was prepared from liver tissues of NP and pregnant (gestation days 8, 10, 11, 13, 15 and 18) wild-type and Nrf2-null mice. Hepatic Cyclin A2 mRNA levels were measured by qRT-PCR and are expressed as the mean fold changes relative to NP controls (±s.d.; n = 3 for each group). ‘a’ and ‘b’ indicate P<0.05 compared with NP wild-type+ and Nrf2-null mice, respectively. *P<0.05, between wild-type and Nrf2-null mice.

To determine whether Nrf2 regulates Cyclin A2 expression at the mRNA level, maternal hepatic transcripts of Cyclin A2 were quantified by qRT-PCR at the same time points as its protein expression (Fig. 4B). Wild-type mice showed a gestation-dependent pattern of Cyclin A2 mRNA expression, similar to its protein expression. Nrf2 deficiency resulted in disruption in the pregnancy-dependent regulation of Cyclin A2 gene expression on and prior to gestation day 13, indicating that Nrf2 exerted an effect on Cyclin A2 mRNA. However, after gestation day 13, Cyclin A2 mRNA expression did not differ between the two genotypes of maternal livers, whereas Cyclin A2 protein levels were much higher due to Nrf2 absence (Fig. 4A). Collectively, our data demonstrate that Nrf2 is essential for normal inhibition of Cyclin A2 at both transcriptional and posttranscriptional levels in the maternal liver during pregnancy.

Lack of Nrf2 leads to failure of maternal hepatocytes to decrease in size during the first half of pregnancy

We also evaluated whether Nrf2 modulates pregnancy-induced maternal hepatocyte hypertrophy. Hepatocyte densities were measured in mice of both genotypes in the non-pregnant state and at different stages of pregnancy (Fig. 5A). In non-pregnant mice, the numbers of hepatocytes per microscope field did not differ significantly between the two genotypes. Wild-type mice displayed gestation stage-dependent changes in maternal hepatocyte density. During the first half of pregnancy (gestation days 8 and 10), there were significant increases in the numbers of maternal hepatocytes per microscopic field compared with the non-pregnant state. In contrast, maternal hepatocyte density decreased during the second half of pregnancy. Thus, in response to pregnancy, the sizes of maternal hepatocytes initially decreased and subsequently increased. However, when Nrf2 was deficient, the sizes of maternal hepatocytes were not significantly reduced on gestation days 8 and 10 but were still increased after gestation day 11 relative to the non-pregnant state. β-catenin immunofluorescent staining of maternal liver sections showed Nrf2-dependent dynamic changes in the volumes of maternal hepatocytes during pregnancy (Fig. 5B). Therefore, due to Nrf2 absence, the maternal hepatocytes were prevented from reducing their sizes during the first half of pregnancy.

Fig. 5.

Dynamic changes in maternal hepatocyte density during pregnancy. (A) Livers were collected from non-pregnant (NP) and pregnant (gestation days 8, 10, 11, 13, 15 and 18) wild-type (Nrf2+/+) and Nrf2-null (Nrf2−/−) mice. Liver sections were prepared and subjected to Hematoxylin and Eosin staining. Hepatocytes were counted (400× magnification) using Image-Pro Plus software. The results are shown as the means ± s.d. per field (n = 3–5). ‘a’ and ‘b’ indicate P<0.05, compared with NP wild-type and Nrf2-null mice, respectively. *P<0.05, between wild-type and Nrf2-null mice. (B) Immunofluorescent staining of β-catenin in liver sections. Formalin-fixed and paraffin-embedded liver sections prepared from NP and pregnant (gestation days 10 and 18) wild-type and Nrf2-null mice were subjected to immunostaining with β-catenin primary antibody and DyLighttm594-conjugated IgG fraction monoclonal mouse anti-rabbit IgG.

Nrf2 genetic deletion results in dysregulation of pregnancy-dependent activation of AKT1, mTOR, p70S6K and 4E-BP1

It has been shown that the AKT/mTOR (mammalian target of rapamycin) signaling pathway and its downstream targets p70 ribosomal S6 kinase (p70S6K) and eukaryotic initiation factor 4E-binding protein 1 (4E-BP1) control the size of hepatocytes (Haga et al., 2005; Mullany et al., 2007; Haga et al., 2009; Gielchinsky et al., 2010). Thus, the expression and activation (phosphorylation) of the four proteins were analyzed by western blotting on maternal liver lysates prepared from both genotypes of non-pregnant and pregnant mice (Fig. 6).

Fig. 6.

Protein expression and/or phosphorylation of hepatic AKT1, mTOR, p70S6K, 4E-BP1 and LC3B. Livers were collected from non-pregnant (NP) and pregnant (gestation days 8, 10, 11, 13, 15 and 18) wild-type (Nrf2+/+) and Nrf2-null (Nrf2−/−) mice. Western blotting was performed on liver homogenates using antibodies against the proteins indicated. GAPDH protein levels were used as loading controls.

In the maternal livers of wild-type mice, pregnancy activated all four proteins, as shown in Fig. 6. Notably, each protein exhibited distinct gestation stage-dependent phosphorylation patterns. AKT1 was highly phosphorylated at ser473 on gestation day 15 and at ser308 from gestation day 8 to day 15. Pregnancy-dependent activation of mTOR began on gestation day 10 and persisted until term. Increases in p70S6K phosphorylation were observed throughout pregnancy and were most robust on gestation days 8, 10 and 15. The phosphorylation of 4E-BP1 was increased during the first half of pregnancy (from gestation day 8 to 11) and then decreased during the second half of pregnancy. During the entire course of gestation, mTOR phosphorylation was neither fully correlated with AKT1 activation upstream nor correlated with p70S6K and 4E-BP1 activation downstream. However, AKT1, mTOR, and p70S6K were all activated with a similar phosphorylation pattern from gestation day 10 to 15. This observation suggests that AKT1, mTOR, and p70S6K form a functional pathway within the time frame of pregnancy that is responsible for the size increases in maternal hepatocytes during the second half of pregnancy.

Strikingly, the lack of Nrf2 let to dysregulation of the phosphorylation of all four proteins in both non-pregnant and pregnant states (Fig. 6). In the livers of non-pregnant mice, genetic mutation of Nrf2 resulted in hyperphosphorylation of AKT1, mTOR and p70S6K and hypophosphorylation of 4E-BP1. In maternal livers of pregnant mice, absence of Nrf2 caused alterations in pregnancy-dependent activation patterns of these four proteins. Owing to the lack of Nrf2, AKT1 was abnormally hyperphosphorylated at ser473 on gestation days 11 and 18, and AKT1 phosphorylation at ser308 was delayed to gestation day 11 and largely prevented on gestation day 15. When Nrf2 was absent, mTOR phosphorylation levels were increased on several gestation days (gestation days 8, 11 and 13). In Nrf2-null maternal livers, p70S6K was not activated until gestation day 18 and 4E-BP1 activation was delayed for 2 days (gestation days 8, 10, and 11 in wild-type versus gestation days 10, 11 and 13 in Nrf2-null mice). The correlation of AKT1, mTOR and p70S6K activation from gestation day 10 to 15 was disrupted because of Nrf2 genetic deletion. Collectively, the data demonstrate that Nrf2 is crucial for the regulation of the activities of these four important signaling molecules in both non-pregnant and pregnant states.

The above findings indicate that the functional state of Nrf2 affects mTOR signaling, a critical modulator of autophagy (Russell et al., 2011). In addition, it has been shown that autophagy deficiency leads to persistent activation of Nrf2 (Ni et al., 2012). These observations suggested potential reciprocal interactions between Nrf2 and autophagy. Thus, we examined hepatic expression of LC3B protein, a typical marker for autophagy, in non-pregnant and pregnant wild-type and Nrf2-null mice. As a result, LC3B-II protein was only weakly detected in the livers of non-pregnant and pregnant Nrf2-null mice (Fig. 6). The data suggest that pregnancy may not exert a notable impact on hepatic autophagy process and that Nrf2 absence causes a minor increase in hepatic autophagy.

Discussion

The current results demonstrate that Nrf2 is an important regulator that ensures normal maternal hepatic adaptations to pregnancy. Our data generated from wild-type mice show that maternal hepatocytes respond to pregnancy via two cellular pathways: the proliferation pathway and the size adjustment pathway. During the first half of pregnancy, maternal hepatocytes replicate to increase their total number and simultaneously reduce their sizes. To our knowledge, for the first time, we have discovered a phenomenon of cell volume reduction under a physiological condition. This might be a cellular mechanism preventing the maternal liver from overgrowth in accordance with the extent of development and growth of the placenta and fetus. As pregnancy progresses to the second half of pregnancy, marked growth of the placenta and fetus occurs. To accommodate the increasing demand of metabolism, maternal hepatocytes stop replication and, at the same time, increase their sizes. Thus, the maternal liver adapts to pregnancy via controlling the number and size of hepatocytes. One question is how these pregnancy-dependent events are regulated. Our results show that Nrf2 is involved in modulating the two cellular pathways during the first half of pregnancy. This is indicated by the dysregulation in the numbers of proliferating maternal hepatocytes and, more strikingly, by the largely prevented size reduction of maternal hepatocytes during the first half of pregnancy in Nrf2-null mice (Figs 3, 5). Therefore, Nrf2 participates in regulating the maternal hepatic adaptive response to pregnancy.

It is intriguing that pregnancy-induced metabolic stress does not increase and even reduces the activity of Nrf2 in the maternal liver. To our knowledge, for the first time, we have identified a physiological circumstance in which Nrf2 activity needs to be downregulated. We previously reported that pregnancy does not activate maternal hepatic pregnant X receptor (PXR) or constitutive androstane receptor (CAR) (Dai et al., 2011). These two nuclear receptors are critical mediators of metabolism and can be activated by a variety of xenobiotics and endobiotics, including naturally occurring steroid hormones (Kliewer et al., 1998; Dai and Wan, 2005; Stanley et al., 2006). It is not clear how the activation of PXR and CAR can be suppressed during exposure to the high concentrations of steroid hormones that are required for establishing and maintaining pregnancy. However, it may not be surprising to observe the downregulation of Nrf2 activity by pregnancy because Lubahn's group has elucidated that estrogen receptor alpha and estrogen-related receptor beta possess potent inhibitory effects on Nrf2 activity by interacting with Nrf2 (Ansell et al., 2004; Ansell et al., 2005; Zhou et al., 2007). The current results reveal the importance of Nrf2 in pregnancy-associated biological events, including cell proliferation, cell size adjustment, and the activation of several critical signaling molecules. These findings suggest that the Nrf2 activity level might play regulatory roles and needs to be tightly controlled during pregnancy.

Our results demonstrate the function of Nrf2 in regulating cell proliferation under a physiological condition (pregnancy). It has been shown that Nrf2 plays an important role in the hepatic proliferative response to liver mass deficit (Beyer et al., 2008). In addition, Reddy's group has reported that Nrf2 deficiency causes a defect in lung type II cell proliferation (Reddy et al., 2007). Moreover, Nrf2 was shown to be highly activated in certain cancer cells, promoting cancer cell proliferation (Ohta et al., 2008; Shibata et al., 2008a; Shibata et al., 2008b; Singh et al., 2008; Homma et al., 2009; Yamadori et al., 2012). However, how Nrf2 regulates cell proliferation is not clear. In this paper, we linked Nrf2 to the cell cycle components Cyclins D1, E1 and A2 and 4E-BP1 and p70S6K (Figs 4, 6), which are known to promote cell cycle progression and cell proliferation (Fingar et al., 2004; Ohanna et al., 2005; Dowling et al., 2007). When Nrf2 is deficient, upregulation of maternal hepatic Cyclins A2 and D1 may counteract the inhibitory effects on hepatocyte proliferation caused by delayed activation of Cyclin E1 and 4E-BP1 and inhibited activation of p70S6K. This might explain why the lack of Nrf2 leads to only mild delay of maternal hepatocyte proliferation. How Nrf2 regulates these molecules requires further investigation.

Our results reveal a novel function of Nrf2: regulation of cell size. Without Nrf2, the maternal hepatocytes were not able to reduce their sizes during the first half of pregnancy (Fig. 5). That is why Nrf2-null mice showed a greater extent of maternal liver growth during the period. Few genes are capable of modulating cell volume (Russell et al., 2011). Organ growth requires coordinated regulation of both cell number and cell size (Conlon and Raff, 1999). Cell cycle control has been intensively investigated, yet considerably less attention has been paid to the mechanisms controlling cell growth, especially in mammalian systems (Fingar et al., 2002). In this paper, we have identified Nrf2 as a novel cell size regulator. Further studies are needed to determine how Nrf2 is required for maternal hepatocytes to reduce their sizes.

Our study demonstrates that Nrf2 is a critical regulator of the mTOR signaling pathway. mTOR is a known signaling molecule capable of controlling the sizes of cells (Kim et al., 2002). Activation of mTOR increases cell volume by signaling downstream to two independent targets, p70S6K and 4E-BP1 (Fingar et al., 2002). It has been shown that the AKT/mTORC1 pathway controls hyperplasia/hypertrophy switching during maternal liver regeneration in aged pregnant mice (Gielchinsky et al., 2010). We found that when the size of maternal hepatocytes was decreased on gestation day 8, maternal hepatic mTOR was not activated (Figs 5, 6). In contrast, Nrf2 deficiency caused hyperphosphorylation of mTOR on day 8 (Fig. 6). This hyperphosphorylation may contribute, at least in part, to maternal hepatocytes being prevented from reducing their volumes due to the lack of Nrf2. On gestation day 10, when the size reduction of maternal hepatocytes was completed, the AKT1/mTOR/70S6K pathway was activated and remained functional until term, leading to continuous increases in the size of maternal hepatocytes during the second half of pregnancy (Figs 5, 6). Nrf2 is indispensable to maintain the function of this signaling pathway because Nrf2 absence caused disruption of the phosphorylation cascade formed by these three molecules. However, Nrf2-null maternal hepatocytes were still capable of increasing their sizes, with highly activated mTOR that was disconnected from AKT1 and p70S6K during the second half of pregnancy (Figs 5, 6). Therefore, an unknown compensatory mechanism that is mTOR dependent but AKT1 and p70S6K independent may exist. Our studies strongly suggest that Nrf2 is a negative regulator of mTOR signaling. First, Nrf2 deficiency causes hyperphosphorylation of mTOR in the non-pregnant state (Fig. 6). Second, a progressive decrease in Nrf2 activity is correlated with a gradual increase in mTOR phosphorylation in the maternal liver during the second half of pregnancy (Figs 2, 6). Furthermore, the lack of Nrf2 leads to higher mTOR phosphorylation in the maternal liver on several gestation days (gestation days 8, 11 and 13; Fig. 6). The association of Nrf2 with mTOR activity is important to reveal novel biological functions of Nrf2 because mTOR signaling plays critical roles in a variety of biological processes, including energy sensing, development, growth, and aging (Russell et al., 2011).

Materials and Methods

Mice

Wild-type and Nrf2-null female mice (3 months old) in a C57BL6/129SV mixed background were used (Chan et al., 1996). They were housed in plastic cages at 22±1°C on a 12-hour/12-hour light/dark cycle with the light on from 06:00 am to 06:00 pm. Timed pregnancies were generated, and the presence of a copulatory plug in the vagina was designated as gestation day 1. Non-pregnant and pregnant mice were weighed and sacrificed. Maternal livers were dissected and weighed. Liver tissues were fixed in formalin and then embedded in paraffin for histological analysis or snap-frozen in liquid nitrogen for total RNA isolation and protein analysis. Litter sizes for pregnancies ranged from 7–10. All of the animal experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Protocols for the care and use of animals were approved by the Indiana University-Purdue University Indianapolis Animal Care and Use Committee.

Immunohistochemistry

Formalin-fixed and paraffin-embedded liver sections were subjected to Ki-67 immunostaining for visualizing and counting proliferating hepatocytes. Ki67-postive hepatocytes were counted in five randomly chosen microscope fields per section at 200× magnification. A primary antibody against Ki-67 (Thermo Scientific, Fremont, CA) was used for immunostaining according to the manufacturer's instructions.

Hepatocyte density measurement

Formalin-fixed and paraffin-embedded liver sections were stained with Hematoxylin and Eosin. Hepatocytes were counted with Image-Pro Plus software (Media Cybernetics, MD) in at least four microscope fields at 400× magnification for each sample. Formalin-fixed and paraffin-embedded liver sections were also subjected to β-catenin immunostaining with β-catenin primary antibody (Cell Signaling Technology, Danvers, MA) and DyLighttm594-conjugated IgG fraction monoclonal mouse anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) to visualize the changes in the sizes of maternal hepatocytes.

Liver nuclear extract preparation

Liver nuclear extracts were prepared from liver tissues with the NE-PER Nuclear Extraction Kit according to the manufacturer's instructions (Pierce Biotechnology, Rockford, IL). Bio-Rad protein quantification reagents were used to measure protein concentrations (Bio-Rad Laboratories, Hercules, CA).

Western blot analysis

Liver homogenates (10 or 30 µg protein per sample) or liver nuclear extracts (5 µg protein per sample) were separated by PAGE under reducing conditions. Proteins from the gels were electrophoretically transferred to polyvinylidene difluoride membrane. Antibodies against Cyclin D1, Cyclin E1, Lamin B1, mTOR, p70S6K, 4E-BP1 and LC3B (Cell Signaling Technology, Danvers, MA), Cyclin A2 and NQO-1 (Epitomic, Burlingame, CA), and glyceraldehyde 3-phosphate dehydrogenase (GADPH) (Santa Cruz Biotechnology, Santa Cruz, CA) were used as probes. A rat monoclonal Nrf2 antibody was kindly provided by Dr Ken Itoh (Hirosaki University, Japan). Immune complexes were detected using the enhanced chemiluminescence system (Pierce, Rockford, IL).

Quantitative real-time polymerase chain reaction

Total RNA was isolated from frozen liver tissue using TRIzol reagent according to the manufacturer's protocol (Invitrogen, Carlsbad, CA). cDNAs were synthesized from total RNA (1 µg) from each sample using a Verso cDNA Kit (Thermo Scientific, Rockford, IL), diluted 4 times with water, and subjected to qRT-PCR to quantify mRNA levels. TaqMan Universal PCR Master Mix and the primers and TaqMan MGB probes of mouse Nrf2 (Mm00477786_m1), NQO-1(Mm01253561_m1), AOX1 (Mm01255332_ml), Cyclin A2 (Mm00438063_m1), and β-actin (Mm00607939_s1) were purchased from Applied Biosystems (Foster City, CA). The amplification reactions were carried out with the ABI Prism 7900 sequence detection system (Applied Biosystems, Foster City, CA) with initial hold steps (50°C for 2 minutes followed by 95°C for 10 minutes) and 40 cycles of a two-step PCR (92°C for 15 seconds and 60°C for 1 minute). The comparative CT method was used for relative quantification of the amount of mRNA in each sample normalized to β-actin transcript levels.

Statistical analysis

Data are shown as the mean ± standard deviation (s.d.). Statistical analysis was performed using a one-way analysis of variance. Comparisons of means were determined by post-hoc analysis. A P-value of <0.05 was considered to be statistically significant.

Footnotes

  • Author contributions

    Y.Z., J.C. and G.D. conceived and designed the experiments; Y.Z., M.H., and Q.B. performed the experiments; Y.Z. and G.D. analyzed the data and wrote the manuscript.

  • Funding

    This work was supported by the National Institute of Diabetes and Digestive and Kidney Diseases [grant number 7RO1DK07596 to G.D.]. Deposited in PMC for release after 12 months.

  • Accepted January 8, 2013.

References

View Abstract