A Conditional Model System for MYC-Induced HCC
We used the Tet system to conditionally express MYC
in murine hepatocytes (Kistner et al. 1996
; Felsher and Bishop 1999b
). We mated the transgenic line TRE-MYC
(Felsher and Bishop 1999b
), which contains the tetracycline response element adjacent to the human MYC
cDNA, with the transgenic line LAP-tTA
(Kistner et al. 1996
), which contains a liver-specific enhancer that drives the expression of the tetracycline-transactivating protein. Mice possessing both transgenes exhibited increased expression of the MYC
transgene in their hepatocytes (A). Mice possessing either transgene alone did not overexpress MYC
and lacked evidence of morbidity or mortality. Similarly, mice possessing both transgenes that were treated with doxycycline to suppress MYC
transgene expression did not exhibit a phenotype. Thus, we have developed a transgenic model that enables us to conditionally regulate MYC
expression in murine hepatocytes.
MYC Overexpression in Adult Hepatocytes Results in HCC
To investigate if MYC overexpression is sufficient to induce HCC in our model system, we removed doxycycline treatment in adult mice (6–12 weeks of age) transgenic for both TRE-MYC and LAP-tTA. Ninety percent of adult mice overexpressing MYC succumbed to liver tumors with a mean latency of 35 weeks. At necropsy, mice exhibited marked gross enlargement of the liver with multiple tumor masses (B). The normal liver architecture was disrupted by nodular tumors with histological features typical of HCC. Tumors were composed of dysplastic nests of cells with large pleomorphic nuclei, delicate vesicular chromatin, and very prominent nucleoli (C). Tumors could invade into the abdomen and the lung (A–C). These features demonstrated that MYC overexpression in adult mice resulted in HCC.
MYC-Induced Hepatic Tumors Are Invasive and Metastatic
To confirm that these tumors were malignant, we transplanted them subcutaneously into scid mice. Tumors formed in the inoculated mice after an 8–10-week latency (see D). The transplanted tumors displayed identical histology to the primary transgenic tumor (see E versus C). Normal adult hepatocytes failed to induce tumors when inoculated into scid mice. We conclude that MYC overexpression in adult hepatocytes results in the formation of highly malignant liver cancers with features consistent with human HCC.
Developmental State of the Host Influences on Frequency and Latency of Tumor Onset
To determine if the developmental state of host hepatocytes influenced the ability of MYC activation to induce tumorigenesis, we induced MYC in cohorts of different ages (A). Mice that overexpressed MYC during embryonic development of the liver succumbed to neoplasia within 10 d of birth. Mice in which MYC was activated at birth (neonates) succumbed to neoplasia within 8 weeks. Mice in which MYC was induced at 4 weeks or 6–12 weeks of age developed tumors after a mean latency of 15 and 35 weeks, respectively (A). We conclude that the ability of MYC activation to induce tumorigenesis in hepatocytes is inversely correlated with the developmental age of the host.
MYC's Ability to Induce HCC Is Inversely Correlated with the Age of the Host at the Time of MYC Activation
One possible explanation for our results was that the levels of MYC induction were different in embryonic and neonatal versus adult hosts. To address this possibility, we examined total MYC protein levels by Western analysis using a polyclonal antibody that recognizes both the human c-MYC protein and the endogenous murine c-MYC (B). In neonatal and adult mice, MYC protein levels were induced at similar levels (B). In tumors from embryos, neonatal, and adult mice, MYC protein levels increased an additional 5- to 10-fold over the levels observed in nontransgenic and in MYC
-induced nonneoplastic livers. Tumorigenic conversion of hepatocytes was associated, in all age groups of mice, with further increases in the levels of MYC protein (B). We obtained similar results by quantitative PCR analysis of mRNA expression of the human MYC
transgene (C). In neonatal and adult mice, MYC
transgene expression was induced at similar levels. In tumors from embryos, neonatal, and adult mice, the levels of MYC
transgene RNA increased an additional 10-fold over the levels observed in nonmalignant hepatocytes (C). Hence, tumorigenic conversion of hepatocytes was associated, in all age groups of mice, with further increases in the levels of MYC
transgene expression. However, differences in the ability of MYC
activation to initiate tumorigenesis in mice of different ages did not appear to be related to differences in the levels of induction of MYC
transgene expression. The increased levels of MYC expression we observed in tumors likely reflect that the proliferating tumor cells express more abundant levels of the MYC
transgene than normal hepatocytes. This observation is consistent with observations described in other transgenic models in which expression of transgenes is generally higher in tumors than it is in the normal cellular counterparts (Weiss et al. 2003
MYC Activation in Embryonic and Neonatal Hepatocytes Induces Cellular Proliferation and Tumorigenesis
To evaluate how MYC's ability to induce tumorigenesis is influenced by the age of mice, we investigated the initial consequences of MYC activation in hepatocytes during different developmental periods. Mice that overexpressed MYC during embryonic development were born with livers similar in weight and gross architecture to normal age-matched livers, yet exhibited increasing numbers of neoplastic cells from birth through the first week of life associated with progressive abdominal enlargement. At necropsy, abdominal enlargement was associated with marked hepatomegaly with a 5-fold increase in total liver weight (A, and see B below). Although these livers were larger, the gross architecture was preserved (A, MYC ON versus MYC OFF). When we examined the histology of the livers in which MYC was overexpressed during embryogenesis, we found that they resembled liver cancers (B versus C) similar to the MYC-induced HCCs we observed in adult mice (see C versus C). Hence, MYC overexpression appears to induce cellular proliferation in neonatal hepatocytes that progresses rapidly to neoplasia.
MYC Activation during Embryonic Development Induces a Rapid Onset of Neoplasia
MYC Activation in Adult Hepatocytes Causes Cellular Hypertrophy
To determine if MYC was inducing changes in cell cycle transit, we measured the DNA content of isolated nuclei from normal hepatocytes and hepatocytes in which MYC was activated during embryonic development (D and E). Normal neonatal hepatocytes mostly contained 2N DNA content, consistent with most of the cells residing in G1 (D). A minority of hepatocytes exhibited 4N and 2N–4N DNA content, demonstrating that few cells were in G2/M and S phase, respectively. In contrast, upon MYC activation the proportion of neonatal hepatocytes with 2N–4N DNA content substantially increased, suggesting that an increased number of cells were in S phase (E).
In order to confirm that MYC activation caused tumorigenesis by inducing cell proliferation, we performed Ki67 immunofluorescence and DAPI staining in tumors induced by MYC activation during embryonic development and in age-matched nontransgenic livers. Indeed, there was evidence for increased hepatocyte proliferation in the MYC-induced neonatal tumor, as demonstrated by an increase in Ki67-positive cells (G and I versus F and H). We conclude that MYC activation during embryonic development causes neonatal hepatocytes to undergo DNA replication, cell cycle transit, proliferation, and almost immediate neoplastic conversion.
To confirm that MYC activation during embryonic development induced tumorigenesis in neonatal livers, we transplanted neoplastic hepatocytes into scid mice. We found that neoplastic neonatal hepatocytes readily formed tumors, whereas the transplantation of normal neonatal hepatocytes did not form tumors (see D and unpublished data). Therefore, MYC overexpression during embryonic development of the murine liver causes hepatocellular tumorigenesis within the first 10 d of birth (). We conclude that MYC overexpression results in rapid neoplastic conversion of neonatal hepatocytes.
Mice in which MYC was activated at birth exhibited progressive abdominal enlargement during their second and third weeks of life, and they showed signs of tumorigenesis by 18 to 40 d of age. When these mice developed tumors their livers were ten times the normal size, were paler, exhibited a multitude of coalescing tumor nodules, and preserved a normal gross architecture (unpublished data). We did not observe any histological changes in the liver after 10 d of MYC activation (A versus B); however, by 18 d of MYC activation the histology resembled liver cancers (C versus D), similar to the MYC-induced HCCs we observed in adults (D versus C).
MYC Activation at Birth Induces Proliferation of Neonatal Hepatocytes
We performed Ki67 immunofluorescence and DAPI staining in order to determine if MYC overexpression in neonatal livers was inducing hepatocyte proliferation. At 10 d of age the liver is in an active state of proliferation; thus, there was little detectable difference in the number of Ki67-positive cells between the MYC-activated and nontransgenic livers (E and I versus F and J). However, once these livers became neoplastic, there was a great increase in the number of Ki67-positive cells (G and K versus H and L).
MYC Activation in Adult Hepatocytes Induces Cellular Growth, but Not Proliferation
We examined the initial consequences of MYC overexpression in adult hepatocytes. In contrast to the rapid neoplastic conversion we observed in embryonic or neonatal hepatocytes ( and ), MYC overexpression in adult hepatocytes caused a marked cellular growth, accompanied by an even greater relative nuclear growth, as observed by histological analysis (–). The effects of MYC overexpression on the size of adult hepatocytes depended on the duration of MYC activation. After 2 weeks of MYC overexpression, no changes were observed in cell size compared to normal hepatocytes (C and unpublished data). However, after 4–8 weeks of MYC activation, adult hepatocytes exhibited increased cell and nuclear size (B versus A and C). Similar results were observed in over 20 different mice. Similarly, we observed that MYC induces hypertrophy of hepatocytes by flow cytometry analysis (). Further duration of MYC activation did not induce further cell growth, as measured up to 50 weeks of MYC activation (unpublished data). Thus, there may be an absolute limit to the ability of MYC to induce liver growth. MYC activation in adult hepatocytes was not associated with a change in overall liver weight (D). Since the cells were bigger, but the overall weight of the liver did not increase, we infer that the total number of hepatocytes was unchanged or slightly decreased. One possible explanation for these results is that MYC induced apoptosis, as described below.
MYC Activation in Adult Hepatocytes Induces Increased Cell Size and Endoreduplication, and Only Results in Cell Proliferation upon Neoplastic Conversion of Hepatocytes
MYC Activation in Adult Hepatocytes Induces Increased Cell Size
To examine if MYC
activation induced proliferation of adult hepatocytes, we measured Ki67 expression by immunofluorescence. We did not observe increased Ki67 expression when MYC was overexpressed in the adult liver (E and H versus D and G). Only upon neoplastic conversion of hepatocytes was there evidence for increased hepatocyte proliferation (F and I). We conclude that MYC overexpression in adult hepatocytes induces increased nuclear and cell growth, but not cell proliferation. Our observations are consistent with previous reports that MYC
activation induces cell growth (Mateyak et al. 1997
; Iritani and Eisenman 1999
; Johnston et al. 1999
; Grandori et al. 2000
; Kim et al. 2000
MYC activation in adult hepatocytes eventually culminated in tumorigenesis, demonstrating that some adult hepatocytes acquire the ability to undergo cell division. To confirm this, we measured the nuclear and cellular sizes in liver tumors. When we examined the cell size in ten different tumors from adult hosts, we found that in all tumors, the cell size was reduced to below normal and the nuclear size was similar to that of normal hepatocytes (C and C). We were also able to confirm that the cell size of tumor cells was reduced to below normal by FACS forward versus side scatter (unpublished data). We conclude that MYC-induced malignant conversion of adult hepatocytes is associated with the acquired ability to undergo mitotic division.
MYC Overexpression in Adult Hepatocytes Results in Endoreduplication
To further define the consequences of MYC
activation on the cell cycle, we examined the DNA content of isolated nuclei from normal and MYC
-activated adult hepatocytes. We expected that if adult hepatocytes were restrained from undergoing mitotic division, MYC
activation might result in endoreduplication. Age-matched normal hepatocytes exhibited a 2N DNA content consistent with most of the cells residing in G1, and there was no evidence for cells in S or G2/M (J). After MYC
activation for 2 months, we found that almost all nuclei had a 4N, 8N, or 12N DNA content, suggesting that the cells replicated their DNA repeatedly without dividing (K). Almost no cells contained the intermediate DNA content (2N–4N), demonstrating that very few cells were in S phase at any given time. We conclude that MYC
activation induces endoreduplication of the genomes of normal adult hepatocytes. Our results are consistent with reports that MYC overexpression can enforce DNA replication, resulting in endoreduplication in normal cells (Cerni et al. 1986
; Mai et al. 1996
; Chernova et al. 1998
; Felsher et al. 2000
We reasoned that if MYC was causing endoreduplication in adult hepatocytes by arresting cell division and enforcing DNA replication, then upon neoplastic conversion these hepatocytes must acquire the ability to undergo mitotic division and would no longer endoreduplicate. As predicted, tumors did not exhibit evidence for endoreduplication (L). Greater than 70% of the tumor cells contained a 2N–4N DNA content and none of the cells contained greater than 4N DNA content. The majority of tumor cells were in S phase (L). Hence, MYC-induced tumorigenesis in adult hepatocytes is associated with the acquired ability to divide mitotically.
MYC Overexpression in Adult Hepatocytes Does Not Induce Apoptosis
-induced apoptosis is an important mechanism that restrains MYC
from causing tumorigenesis (Evan et al. 1992
; Pelengaris et al. 2000
). We reasoned that MYC may induce cellular hypertrophy, but not an increase in liver mass, because MYC induces apoptosis. Normal neonatal and adult hepatocytes did not undergo apoptosis, as demonstrated by TUNEL assay or DAPI staining (A, B, E, and F). Surprisingly, we could not find evidence that MYC
induced apoptosis in adult hepatocytes by TUNEL assay or DAPI staining after 2, 4, or 8 weeks of MYC
induction prior to tumor formation (G and H and unpublished data). Contrary to what we expected, MYC
activation was only associated with increased apoptosis in the hepatocytes of liver cancers (C, D, I, and J). Hence, apoptosis is not necessarily the mechanism restraining MYC
from causing tumorigenesis, at least in hepatocytes (Pelengaris et al. 2002a
). However, we recognize that MYC could be inducing low levels of apoptosis in hepatocytes, perhaps not easily detected by TUNEL, since apoptotic cells may be rapidly eliminated from the liver through the host reticulo-endothelial system. Such a low level of apoptosis still could explain why in the adult liver hepatocytes become hypertrophic but the liver mass does not increase.
MYC Activation Does Not Induce Apoptosis in Murine Hepatocytes
Loss of p53 Function Cooperates with MYC to Induce Tumorigenesis in Adult Mice
Previously, we have shown that the loss of p53 function is required to permit the cell division of normal mouse and human fibroblasts overexpressing MYC (Felsher et al. 2000
). We speculated that loss of p53 function might be similarly required for MYC
activation to induce cell proliferation and tumorigenesis in hepatocytes. First, we examined if MYC
activation affected p53 protein expression. We found that MYC
activation was associated with an increase in p53 protein levels in adult hepatocytes, as measured by Western analysis (A). Conversely, tumors in adult mice frequently exhibited reduced levels of p53 protein expression (A and unpublished data). Not all tumors exhibited reduced p53 expression. Since p53 mediates its function largely through inducing the transcription of many different genes, we evaluated if tumors exhibited a loss of p53 function by measuring the expression of these target genes. We found by Northern analysis that the p53 target genes, p21
were induced upon MYC
activation in adult livers (C). Conversely, MYC
-induced adult HCCs frequently exhibited reduced or no expression of p53 downstream targets. Notably, a tumor that exhibited high p53 protein levels, tumor 2264, exhibited a loss in expression of p53 target genes. In contrast, tumors arising in neonatal mice expressed p53 protein and exhibited the induction of p53 target genes (B and C). We conclude that the adult, but not neonatal MYC
-induced liver tumors require the loss of p53 function for tumorigenesis. Hence, HCCs that arise in adult versus neonatal hosts appear to occur through genetically distinct mechanisms.
MYC Activation Induces p53 Function, and Loss of p53 Function Is Necessary for MYC to Induce Tumorigenesis in Adult Hepatocytes
To directly address if loss of any of p53's functions accelerates the ability of MYC
to induce HCC in adult mice, we generated transgenic mice that overexpressed MYC
in their hepatocytes in the absence of one p53
allele. We mated LAP-tTA
mice with p53+/−
mice that were in the FVB/N background. We activated MYC
in mice when they were 6 weeks old and monitored them for morbidity. We found that even the loss of a single p53
allele was sufficient to reduce the mean latency of tumor onset in 6-week-old adult mice from 20 weeks to 15 weeks (D). We found by PCR that tumors did not generally inactivate the second allele through deletion (E). Our results extend previous findings that suggest that the lack of p53 function cooperates with MYC to induce HCC (Klocke et al. 2001
). We conclude that even a slight reduction of p53 function greatly facilitates the ability of MYC to induce tumorigenesis in adult hepatocytes.
Partial Hepatectomy Accelerates MYC-Induced Tumorigenesis
Our results suggest that the ability of MYC to induce tumorigenesis in hepatocytes depends on the developmental context. We recognized that an alternative explanation is that MYC induces tumorigenesis more readily in hepatocytes that are already proliferating. Adult hepatocytes are known to undergo rapid proliferation in response to partial hepatectomy resulting in the complete regeneration of the liver within 2 weeks of surgical removal (Michalopoulos and DeFrances 1997
; Kountouras et al. 2001
). We found that MYC
activation in adult mice that have undergone partial hepatectomy exhibited a reduced latency of tumor induction in comparison with adult mice that had not undergone surgery (mean latency of 14 weeks versus 35 weeks). However, this latency of tumorigenesis in adult mice after partial hepatectomy was still up to two magnitudes longer than what was observed when MYC
was activated in embryonic and neonatal mice (<10 d and 4 weeks, respectively) ( and A). In addition, tumors in mice that had undergone partial hepatectomy, unlike tumors arising in neonatal mice, were multifocal, suggesting that tumorigenesis was occurring infrequently (A versus A). Similarly, upon histological analysis, after MYC
was activated for 7 weeks in mice that had undergone partial hepatectomy, mice exhibited many individual foci of HCC (B and C). Finally, in adult mice after partial hepatectomy, but not in neonatal mice, areas of the liver that had not undergone neoplastic conversion clearly exhibited increased cellular hypertrophy, and hence were unable to undergo mitotic division (unpublished data). We conclude that the ability of MYC to induce tumorigenesis in adult hepatocytes is accelerated after partial hepatectomy when adult hepatocytes are proliferating, but other developmentally specific parameters are more important in determining when oncogene activation will induce tumorigenesis than the ability of hepatocytes simply to proliferate.
Partial Hepatectomy Accelerates the Ability of MYC to Induce HCC in Adult Mice