HCV infection-associated hepatocellular carcinoma in humanized mice

Inactivation of tumor suppressor and amplification of oncogene

Loss of tumor suppressor proteins and induction of oncogenic proteins are the initiating
events that promote deregulation of molecular pathways underlying the development
of hepatocellular carcinoma 12], 13]. We investigated whether HCV infection associated HCC is characterized by the loss
of tumor suppressor proteins. Of particular interest is the deregulation of PTEN tumor
suppressor due to its critical role in hepatocellular carcinoma 14]–16]. Viral role in PTEN depletion was suggested in an earlier report which showed that
HCV-derived small non-coding RNA restricted nuclear translocation of PTEN protein
by down-regulating Transportin 2 expression in HCV-infected human hepatocytes 9]. Here we investigated whether the loss of PTEN defines HCC progression in a humanized
mouse model of HCV infection.

Immunoblots of PTEN protein from liver tumors (T) of thirteen hepatocyte-engrafted
and HCV-infected MUP-uPA/SCID/bg mice and twelve control (C) mice that were engrafted
with human hepatocytes but not infected with HCV are shown Fig. 1. We also examined liver tissues from seven MUP-uPA/SCID/bg mice that were engrafted
with human hepatocytes and infected with HCV but remained HCC-negative (N). We observed
a consistent decline of both nuclear (Fig. 1a and c) and cytoplasmic (Fig. 1b) PTEN protein in HCV-infected HCC. Interestingly, PTEN protein in HCV-infected but
HCC-negative liver (N) also declined to similar extent (Fig. 1d), suggesting that loss of PTEN may be necessary but insufficient to promote HCC.

Fig. 1. Tumor suppressor proteins: Liver tissues from control animals (MUP-uPA/SCID/bg mice
engrafted with human hepatocytes but not infected with HCV), HCC-negative mice (engrafted
and HCV infected mice that did not develop HCC), and HCC positive mice (engrafted
and HCV infected) were examined by Western blotting with monoclonal antibodies against
PTEN, Phospho-Akt or Cyclin D1. Panels (a) and (c) show representative Western blots of nuclear protein fraction (with Lamin B1 as
loading control); and panel (b) is representative Western blot of corresponding cytoplasmic fraction (with GAPDH
as loading control). Liver tissues were homogenized with RIPA buffer, and 30 microgram
total proteins per lane were resolved by SDS-PAGE, and immunoblotted with antibodies
as described before 11]. Panel (d) represents quantitative assessment (fold change) of nuclear or cytoplasmic PTEN
(PTEN-N, PTEN-C); cytoplasmic p-Akt or nuclear Cyclin-D1 from the liver tissues. Quantitation
was based on 12 controls, 7 HCC negative and 13 HCC positive liver tissues, analyzed
in three independent SDS-PAGE runs. The Western blots were normalized to the internal
“loading” controls (Lamin-B1 for nuclear and GADH for cytoplasmic fractions) (*p??0.05)

The next issue is whether the PTEN produced in chimeric mouse liver is functional.
PTEN is a dual specificity phosphatase with lipid and protein phosphatase activities
17]–19]. Cytoplasmic lipid phosphatase activity of PTEN is a central negative regulator of
phosphatidylinositol-3-kinase (PI3K) signal cascade for cell growth and proliferation.
Loss of PTEN is associated with increased phosphorylation of Akt, a proto-oncogene
with key roles in cell survival and cell proliferation in many types of cancer 15], 16]. We observed increased phosphor-Akt levels in HCV-infected liver tumor (Fig. 1a, b and d), consistent with the role of PTEN as a tumor suppressor.

Nucleus-localized PTEN protein has an essential role in cell cycle homeostasis and
genomic stability 17], 19], 20]. We observed a reduced level of nuclear PTEN that corresponded with increased Cyclin
D1 in HCV infection-associated liver tumors (Fig. 1c). The observed increase in Cyclin D1 in liver tumors is consistent with enhanced
proliferation of HCV-infected hepatocytes 10]. Interestingly, we observed similar increase in Cyclin D1 of HCV-infected but HCC-negative
liver (Fig. 1c and d), suggesting that deregulation of the cell cycle is an early event resulting from
PTEN insufficiency of HCV-infected cells.

Induction of c-Myc oncoprotein

c-Myc is a constitutively induced transcription activator in a broad range of human
cancers 21], 22]. We observed increased c-Myc protein levels in HCV-infected liver tumors compared
to the control (Fig. 2a). By contrast, induction of c-Myc in HCC-negative liver was modest (Fig. 2d), suggesting that induction of c-Myc oncoprotein is a relatively late event in the
development of HCV-infection associated HCC.

Fig. 2. Oncoproteins: Western blots of the control, HCC-negative and liver tumor tissues (as
is shown in Fig. 1) were probed with antibodies against c-Myc, DLC-1 or p21 proteins (panels a, b and b). Panel (d) represents quantitative analysis (based on the loading controls) of liver tissues
from uninfected control, HCC negative and HCC positive mice (as in Fig. 1)

Down-regulation of DLC-1 tumor suppressor

DLC1 (Deleted in Liver Cancer 1) gene maps to chromosome 8 (8p21.3-22), a region that
is frequently deleted in solid tumors. DLC1 encodes GTPase-activating protein (GAP)
that acts as a negative regulator of the Rho family GTPases 23]. Loss of DLC1 is considered an independent marker of hepatocellular carcinoma 4], 6]. We observed the loss of DLC1 protein in HCV-infected liver tumors and less so in
HCV-infected but HCC-negative liver tissue (Fig. 2b and d).

P21

HCV structural and non-structural proteins have been shown to interact with and modulate
transcriptional regulatory activity of p53 tumor suppressor protein 24]–26]. Modulation of transcriptional activity of p53 by viral proteins distinguishes HCV
from other RNA viruses in its ability to interfere with the p53 function 13]. Therefore, it was important to determine if HCV infection of humanized mice modulated
p53 to promote HCC. We assessed the modulation of p53 function in HCV-infected chimeric
mice on the basis of p21 expression, a direct target of p53 transcriptional regulatory
function. We observed a marked decline of p21 protein in HCV-infected liver tumor
(Fig. 2c and d), and less so in HCC-negative mice (N, Fig. 2d), suggesting that HCC progression is correlated with the loss of function of p53
tumor suppressor.

Inflammatory response

Persistent viral infection is an underlying cause of inflammation-induced cancer,
including HCC. More than 90 % of HCCs arise in the context of hepatic injury and inflammation.
Inflammation-associated oncogenic response is mediated by STAT proteins; in particular,
activated STAT3 27], 28]. To ascertain if HCV infection-associated HCC in humanized mice mimics the natural
inflammatory response, we assayed activated STAT3 levels in the liver tumors and in
HCC-negative as compared to the uninfected control mice. As shown (Fig. 3), there is a marked induction of activated (phosphorylated) STAT3 in HCV infection-associated
liver tumors as compared to HCV-infected but HCC-negative liver.

Fig. 3. Inflammatory response markers: Total proteins from liver tissues of 7 uninfected controls, 8 HCV infected HCC negative
and 7 HCV infected HCC positive mice were analyzed by Western blotting. (Fig. 3a). Representative Western blots of two controls (C), three HCC (T) and three HCC negative
(N) liver are shown. Numbers underneath is relative values normalized to ?-Actin loading
control. (Fig. 3b): Quantitative assessment of ?-Catenin, STAT-3 and IL-6R (from 7 uninfected control,
8 HCC negative and 7 HCC positive mice) was based on B-Actin internal controls analyzed
by three independent SDS-PAGE runs

Next, we investigated whether the activation of STAT3 in liver tumors was coordinately
regulated with other inflammatory response proteins, notably interleukin-6 (IL-6)
and IL-6R 28]–30], in HCV-infected chimeric mice. As shown in Fig. 3, we observed a parallel increase in STAT3 and IL-6R in HCV-infected HCC as compared
with control liver, suggesting coordinate regulation of inflammatory response molecules
in tumor development.

Signal transduction pathways, in particular Wnt signaling, are frequently mutated in liver cancer. Disruption of Wnt signaling results in ?-Catenin stabilization and translocation to the nucleus where
it activates cell survival and cell proliferation genes 12]. Activation of ?-Catenin in HCC as contrasted with the surrounding (normal) liver
tissue was recently reported from HCV-infected patients 31]. Considering the significance of ?-Catenin activation and Wnt signaling in human
cancer, we considered it important to investigate the activated state of these tumorigenic
proteins in HCV infection-associated HCC in humanized mice. Results illustrated in
Fig. 3a, b show coordinate regulation of inflammatory response proteins and deregulation of
signal transduction pathways (activated ?-Catenin, IL-6R and STAT3) in HCV-infected
HCC.

MicroRNA markers of HCC

MicroRNAs can function as tumor suppressors or oncogenes (oncomiRs) 32]. Altered expression levels of miRNAs have been reported in a number of human cancers
8], 32]–34]. In this study we sought to identify miRNAs that would serve as distinguishing markers
of HCV infection-associated HCC.

MicroRNA 141 (miR-141) is induced in HCV-infected human primary hepatocytes. Importantly,
miR-141 directly targets DLC1 tumor suppressor protein expression 10], attesting to its role as bona fide oncomiR. Here we compared expression levels of
miR-141 along with other known oncomiRs (miR-21 and miR-221) in HCV infection-induced
HCC (Fig. 4). Results suggest that expression of miR-141 and oncomiRs miR-21 and miR-221 that
target cell cycle inhibitors 34], 35] is coordinately induced in HCV infection-associated HCC.

Fig. 4. Altered MicroRNA expression: Changes in microRNA expression in Controls, HCC negative and Liver Tumor tissues
were analyzed by RT-PCR. Total RNA was prepared by Trizol procedure and equal RNA
amounts were analyzed by qRT-PCR. The data represents similar number of Controls,
HCC negative and HCC positive liver tissues (as in Fig. 1), analyzed in triplicates

Tumor suppressor microRNA 26a (miR-26a) is depleted in liver tumor as compared to
the surrounding normal liver tissues 36]. We then investigated whether the HCV-infected HCC in our humanized mouse model down
regulated miR-26a tumor suppressor, resembling the loss of miR-26a in human HCC. We
observed a marked reduction of miR-26a in liver tumors of human hepatocyte-engrafted
HCV-infected mice compared to controls (Fig. 4). Our results suggest that induction of oncomiRs (miR-141, miR-21 and miR-221) and
down-regulation of tumor suppressor miRNA (miR-26a) constitute distinguishing markers
of HCV infection-associated HCC.

Altered expression levels of miRNAs have also been reported from liver tumors of heterogeneous
origin. To identify changes in miRNA expression that distinguish HCV infection-associated
HCC, we examined selected examples of miRNAs that have been reported from studies
of various liver tumors. Increase in miR-181 levels was reported in AFP-positive tumors
and in embryonic liver cells 37]. Our results by contrast, showed a decline of miR-181 in HCV infected-HCC (Fig. 5), suggesting that miR-181 is less reliable indicator of HCV infection-associated
liver tumors.

Fig. 5. HCV non-specific microRNAs: Total RNA from eight HCV infected and four uninfected human hepatocyte engrafted MUP-uPA/SCID/Bg mice were analyzed by qRT-PCR.
Relative values of microRNAs are shown as fold change compared to the uninfected control
liver

Contribution of miR-122 in HCV replication has been studied in a number of cell culture
systems and in animal models; however, its role in HCV infection-related HCC in humans
is less clear 38], 39]. A recent report 40] suggests that phenotypic effects of miR-122 depletion may be a consequence of its
sequestration by HCV genomic RNA. Our analysis of miR-122 shows an insignificant change
in HCV infection-associated HCC of human hepatocyte-engrafted mice (Fig. 5).

MicroRNA 23a (miR-23a) has been shown to target genes involved in gluconeogenesis
at different stages of hepatocarcinogenesis in mouse liver as well as in primary human
HCC 41], 42]. In our studies, amplification of miR-23a appears to be a less specific indicator
of HCV infection-induced HCC of humanized mice. As shown (in Fig. 5), liver tumors of HCV-infected MUP-uPA/SCID/bg mice show a relatively modest change
in miR-23a levels.

Our studies suggest that amplification of oncomiRs (miR-21, miR-221 and miR-141) and
the loss of tumor suppressor miR-26a are specific indicators of HCV infection-associated
HCC. By contrast, altered expression of miR-181, miR-122 and miR-23a appears to be
a less specific indicator of HCV infection-associated HCC.