| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
FJ
EXPRESS SUMMARY ARTICLE The Full-length version of this article is also available, published online February 16, 2006 as doi:10.1096/fj.05-4704fje. |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,,,1
* Inserm U680 and
Inserm U515, Hôpital Saint-Antoine, Paris, France;
Université Pierre et Marie Curie, Faculté de Médecine Saint-Antoine, Paris, France;
Service d’Anatomopathologie, Hôpital Saint-Antoine, Paris, France; and
|| CNRS UMR7148, Collège de France, Paris, France
1Correspondence: Inserm U515, Hôpital St-Antoine, Paris 75571, France. E-mail: holzenberger{at}st-antoine.inserm.fr
SPECIFIC AIMS
Recently, a critical role for GH in liver regeneration was demonstrated, but it remained unclear whether GH acted directly on hepatocytes, or indirectly via insulin-like growth factor-I (IGF-I) and its cognate receptor IGF-1R. To test whether IGF-1R regulates liver regeneration, we performed partial hepatectomy in liver-specific IGF type 1 receptor knockout (LIGFREKO) mice and examined hepatocyte proliferation, cyclin expression, and major pathways downstream of IGF-1R during liver regeneration.
PRINCIPAL FINDINGS
1. Efficient inactivation of Igf1r gene in hepatocytes of LIGFREKO mice
To investigate the contribution of IGF-1R in hepatocyte proliferation in vivo, we developed a mouse model deprived of Igf1r gene in hepatocytes using the Cre-lox system. Mice carrying a floxed Igf1r allele (Igf1rflox) were mated with AlfpCre transgenic mice that express the Cre recombinase in hepatocytes. Igf1rflox/flox;AlfpCre+/0 mice were obtained in an F1 129/B6 background. Igf1rflox/flox littermates that had not inherited AlfpCre were used as controls. In LIGFREKO mice, the absence of Igf1r gene did not affect postnatal development or adult liver morphology or histology. Instead, it significantly reduced levels of hepatocyte proliferation during liver regeneration.
2. Liver regeneration is altered in LIGFREKO mice
We performed partial hepatectomy in male and female control and LIGFREKO mice. After surgery, the animals were killed at the times indicated and the regenerating livers harvested for histology, Ki67 immunohistochemistry (to determine hepatocyte proliferation), and extraction of protein and DNA. Mean liver resection in these experiments was 58 ± 2%. In total, 67 animals (15 females, 52 males) underwent partial hepatectomy; perioperative mortality was 9% and postoperative mortality 4%. Liver regeneration was examined in males and females 40 h after surgery (T40), at the time of peak DNA synthesis. Protein expression was analyzed in another cohort at T28.
All preoperative (T0) livers contained very few Ki67+ hepatocytes. The prevalence of Ki67+ hepatocytes markedly increased by two orders of magnitude in both male groups during liver regeneration, but at T40 livers from LIGFREKO males contained 52% fewer Ki67+ hepatocytes than livers from control males (144±33 vs. 303±49; P<0.05, Mann-Whitney test) (Fig. 1
A). However, when we analyzed regeneration in males at a later time point (T48), the difference between genotypes disappeared. This indicates that IGF-1R is required for normal liver regeneration during peak DNA synthesis and its absence delays, but does not prevent regeneration. None of the studied mice showed evidence of liver necrosis. In contrast to males, hepatocyte proliferation in females was very similar in LIGFREKO and controls at T40 (80±42 vs. 76±32, NS) (Fig. 1B
). Thus, the number of Ki67+ hepatocytes was much higher in livers from control males when compared with those of control females (+300%; P<0.02). This sex-related difference was attenuated in the LIGFREKO group and did not reach statistical significance there.
|
3. Induction of cyclin D1 and cyclin A is impaired in regenerating LIGFREKO livers
We analyzed cyclins D1 and A, two reliable markers of liver regeneration, by Western blot. Cyclin D1 levels were very similar in LIGFREKO and control mice at T0. However, by T40, they had increased 8-fold in control livers but only 5-fold in LIGFREKO males (Fig. 2
A). This deficient induction in LIGFREKO mice resulted in cyclin D1 levels being 46% lower than in controls (P<0.05), consistent with the reduction in hepatocyte proliferation at T40. Similarly, cyclin A levels were 50% lower in LIGFREKO than in control livers (P<0.001) (Fig. 2B
). Thus, the impaired regeneration of LIGFREKO livers is associated with deficiencies in the induction of cyclin D1 and cyclin A synthesis, suggesting that these proteins are potential targets of IGF-1R signaling. In female LIGFREKO and control mice, cyclin D1 levels increased by a factor of only 3 between T0 to T40. This increase is markedly smaller than that for any of the male groups, consistent with the levels of hepatocyte proliferation measured in each of these groups. In addition, there was no significant difference in cyclin D1 levels at T0 or T40 between female LIGFREKO and control mice (not shown).
|
4. IRS-1 is deregulated in regenerating LIGFREKO liver
In control livers, IRS-1 increased strongly from T0 to T28 (+74%, P<0.001), and decreased to almost normal levels from T28 to T40. In marked contrast, IRS-1 in LIGFREKO livers was already high at T0, showed no further increase at T28, and had fallen by T40. Thus, the peak in IRS-1 induction observed at T28 in controls was absent in LIGFREKO livers. No significant difference in tyrosine-phosphorylation of IRS-1 was observed between quiescent and regenerating livers from both groups at T28. IRS-2 increased significantly in control livers at T28 but, in contrast to IRS-1, LIGFREKO and control livers had similar IRS-2 overproduction profiles. Levels of Shc did not vary between T0 and T28, and showed no genotype-dependent differences.
5. ERK activation is diminished in regenerating LIGFREKO liver
Downstream from IGF-1R, ERK activation is an essential step in the promotion of cell proliferation. Total ERK1 levels neither varied over time nor depended on genotype, but phosphorylated ERK displayed clear differences between LIGFREKO and controls. In control livers, phosphorylated ERK were similar at T0 and T28, and decreased strongly thereafter (P<0.002). In contrast, phosphorylated ERK levels in LIGFREKO livers had already begun to decrease between T0 and T28 (P<0.01) and continued to decrease thereafter (P<0.02), so that LIGFREKO livers contained significantly less activated ERK at T28 than controls. These results suggest that IRS-1 and ERKs are involved in IGF-1R-dependent liver regeneration in males.
CONCLUSIONS AND SIGNIFICANCE
Igf1r inactivation in LIGFREKO livers markedly altered hepatocyte proliferation after partial hepatectomy at T40, but not at T48. This is compatible with the notion that the disruption of Igf1r delayed the progression from G1 to S phase and that the decrease in hepatocyte replication was transient. Mice lacking the ligand IGF-I specifically in the liver also display reduced hepatocyte proliferation during regeneration. However, those mice were also knockout for ALS, and the relative contributions of ALS and hepatic IGF-I to hepatocyte renewal could therefore not be distinguished. Based on those and our present findings, we may conclude that IGF-I binding to IGF-1R on hepatocytes is important for liver regeneration (Fig. 3
). Quiescent hepatocytes express little IGF-1R, but their very high levels of IGF-I production, together with conspicuous increases in IGF-1R expression during regeneration, seem to be sufficient for significant IGF action. The activation of IGF-1R signaling pathways may also enhance the action of other mitogens. Consistent with this, IGF-I has been identified as a co-mitogen for HGF in hepatocellular carcinomas, and EGF induction of ERK in rat hepatocytes requires IGF-1R transactivation by EGFR.
|
We found a defective liver regeneration in LIGFREKO males, but not in females. Moreover, control females had considerably fewer Ki67+ hepatocytes than males. Thus, the sex-related difference in hepatocyte regeneration disappeared in mice lacking hepatic IGF-1R, suggesting that sex-dimorphism in liver regeneration may involve signal transduction via the IGF-1R. Peak DNA synthesis during liver regeneration in rodents has been reported to depend on sex, with females having a delayed DNA synthesis peak. It therefore remains possible that IGF-1R plays a role in females liver regeneration, but at a time point later than T40.
Studies in liver cells derived from mice with knockouts for insulin receptor and IRS have shown that IRS-1 is the main effector of proliferative signals from IGF-1R, whereas IRS-2 is essential for mediating responses to insulin. In addition, transgene-mediated overproduction of IRS-1 in the liver increases the number of hepatocytes. We observed changes in the abundance pattern of IRS-1, but not in that of IRS-2, in livers from LIGFREKO males in the transition from quiescent to regenerative states. This differential regulation suggests that the production of IRS-1, but not that of IRS-2, is controlled by IGF-1R-dependent pathways. These data support the idea that IRS-1 is a major substrate of IGF-1R in the liver.
Concomitantly with the reduction of hepatocyte proliferation in LIGFREKO males, we observed a significant decrease in cyclin D1 and A induction. IGF-I has been reported to stimulate the proliferation of other types of cell by promoting G1 to S phase progression and cyclin D1 induction. Our data indicate that cyclin D1 may be a target of IGF-1R-dependent pathways in the liver. In vitro, IGF-I increases cyclin D1 production by stimulating ERK1/2- and PI-3K/Akt-dependent mechanisms. Although we observed no modulation of Akt activity at T28, the marked decrease in ERK activity in LIGFREKO livers may be involved in decreasing hepatocyte proliferation.
Comparing the liver regeneration phenotypes of LIGFREKO with IGFBP-1–/– mice we not only found similarities (delayed regeneration) but also differences (IGFBP-1–/– show significant liver necrosis upon partial hepatectomy). The alterations in intracellular signaling events observed in both models also overlap (cyclin A diminished or delayed) and differ (cyclin D1 close to normal in IGFBP-1–/–). Thus, it seems possible that part of the effects of IGFBP-1 on liver regeneration depends on signaling via IGF-1R (Fig. 3)
.
Our findings, together with previous work, indicate that intact GH/IGF-I/IGF-1R signaling pathways are required for normal liver regeneration. In this process, IRS-1, rather than IRS-2, seems to be responsible for transduction downstream from IGF-1R in the regenerating liver, possibly via the activation of ERK, cyclin D1, and cyclin A.
FOOTNOTES
To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.05-4704fje;
This article has been cited by other articles:
|
|
 |
|
  M. Wang, M. Chen, G. Zheng, B. Dillard, M. Tallarico, Z. Ortiz, and A.-X. Holterman Transcriptional activation by growth hormone of HNF-6-regulated hepatic genes, a potential mechanism for improved liver repair during biliary injury in mice Am J Physiol Gastrointest Liver Physiol, August 1, 2008; 295(2): G357 - G366. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Natarajan, B. Wagner, and M. Sibilia The EGF receptor is required for efficient liver regeneration PNAS, October 23, 2007; 104(43): 17081 - 17086. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Masumoto, C. Tateno, A. Tachibana, R. Utoh, Y. Morikawa, T. Shimada, H. Momisako, T. Itamoto, T. Asahara, and K. Yoshizato GH enhances proliferation of human hepatocytes grafted into immunodeficient mice with damaged liver J. Endocrinol., September 1, 2007; 194(3): 529 - 537. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P Froment, M Vigier, D Negre, I Fontaine, J Beghelli, F L Cosset, M Holzenberger, and P Durand Inactivation of the IGF-I receptor gene in primary Sertoli cells highlights the autocrine effects of IGF-I J. Endocrinol., September 1, 2007; 194(3): 557 - 568. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, EGF, HGF, and IL-6. Interesting cooperative effects between IGF-1R and EGF or HGF have been proposed recently. Downstream of IGF-1R, the signaling seems to occur preferentially through IRS-1 and ERK pathways. Cyclins D1 and A may be important signaling targets that could subsequently control the hepatocyte cell cycle during regeneration.