Hepatic IGFBP1 is a prosurvival factor that binds to BAK, protects the liver from apoptosis, and antagonizes the proapoptotic actions of p53 at mitochondria

Abstract

Liver is generally refractory to apoptosis induced by the p53 tumor suppressor protein, but the molecular basis remains poorly understood. Here we show that p53 transcriptional activation leads to enhanced expression of hepatic IGFBP1 (insulin-like growth factor-binding protein-1). Exhibiting a previously unanticipated role, a portion of intracellular IGFBP1 protein localizes to mitochondria where it binds to the proapoptotic protein BAK and hinders BAK activation and apoptosis induction. Interestingly, in many cells and tissues p53 also has a direct apoptotic function at mitochondria that includes BAK binding and activation. When IGFBP1 is in a complex with BAK, formation of a proapoptotic p53/BAK complex and apoptosis induction are impaired, both in cultured cells and in liver. In contrast, livers of IGFBP1-deficient mice exhibit spontaneous apoptosis that is accompanied by p53 mitochondrial accumulation and evidence of BAK oligomerization. These data support the importance of BAK as a mediator of p53's mitochondrial function. The results also identify IGFBP1 as a negative regulator of the BAK-dependent pathway of apoptosis, whose expression integrates the transcriptional and mitochondrial functions of the p53 tumor suppressor protein.

Figure 1.

IGFBP1 is a p53-induced BAK-binding protein. (A) Whole-cell extracts (WCE) prepared from HepG2 cells before or after 10 μM Nutlin-3 treatment were immunoprecipitated using an anti-BAK antibody. The excised band of ∼30 kDa shown in the Coomassie gel contained the following IGFBP1 peptide sequences (**): IPGSPEIR, NGFYHSR, ALHVTNIK, ALHVTNIKK, RIPGSPEIR, AQETSGEEISK, IELYRVVESLAK, and ALPGEQQPLHALTR. (B) Western blot analysis shows induction of IGFBP1 protein, but not BCL-xL or MCL1, in Nutlin-3-treated HepG2 cells. (C) IP-Western blot (WB) analysis of WCE from Nutlin-3-treated HepG2 cells shows that BAK interacts with IGFBP1, but not p53. (D) HepG2 cells were examined using immunofluorescent microscopy for IGFBP1 protein (red) and BAK protein (green). The merged image provides evidence of colocalization of these proteins (yellow). Bars, 20 μm. (E) HepG2 cells were treated with 10 μM Nutlin-3 for the indicated times, protein extracts were prepared, and Western blot analysis was used to determine expression of the proteins indicated. (F) Nutlin-3-treated HepG2 (wild-type p53) or Huh7 (mutant p53) hepatoma cells were examined for the proteins or RNA indicated. (G) Western blots of protein extracts prepared following Nutlin-3 treatment of the parental HepG2 cells and two independent clones (E6-1 and E6-2) that stably express the HPV-E6 protein.

Figure 2.

IGFBP1 expression negatively correlates with mitochondrial accumulation of p53 and induction of apoptosis in cultured cells. (A) HepG2 cells were treated for 24 h with cisplatin (3.5–50 μM), protein extracts prepared, and Western blot analysis was used to examine expression of the proteins indicated. (B) Cytosolic (Cyto) and enriched mitochondrial (Mito) fractions isolated from HepG2 cells were examined by Western blot for the presence of the proteins indicated following treatment of the cells with low-dose (L-Cis, 3.5 μM) or high-dose (H-Cis, 50 μM) cisplatin (Cis). As controls for the integrity of the preparations, the extracts were probed for the mitochondrial proteins GRP75 and BAK. (C) IP-Western blot (WB) analysis reveals that IGFBP1 interacts with BAK, but not BCL-xL, in HepG2 cells after exposure to L-Cis (3.5 μM). (D) IP-Western blot analysis reveals that p53, but not IGFBP1, coimmunoprecipitates with BAK following H-Cis (50 μM) treatment of HepG2 cells. In contrast, cisplatin did not alter the BCL-xL/BAK complex. (E) Samples in D were examined by immunoblotting for the proteins indicated. (F) Apoptosis in HepG2 cells untreated or treated with either L-Cis or H-Cis was analyzed by staining for Annexin V followed by flow cytometry. (G) HepG2 cells were treated with 10 μM Nutlin-3 or 10 μg/mL α-aminitin for 24 h, protein extracts were prepared, and Western blot analysis was used to examine expression of the proteins indicated. (H) HepG2 cells were treated with 10 μg/mL α-aminitin for 24 or 48 h, and Western blot analysis was used to examine protein expression. (I) HepG2 cells untreated or treated with 10 μg/mL α-aminitin for 24 or 48 h were analyzed by staining for Annexin V. (J) Total (T), cytosolic (C), and mitochondrial (M) fractions were isolated from HepG2 cells before and after treatment with 10 μg/mL α-amanitin. Samples were examined by immunoblotting for the proteins indicated. (K) IP-Western blot analysis reveals that endogenous p53 interacts with BAK after exposure to α-amanitin for the indicated times. In contrast, no interaction between cytochrome c (Cyt c) and BAK was noted.

Figure 3.

Interaction between p53 and BAK is required for apoptosis induction. (A) HepG2 cells were either untreated or pretreated with PFT-μ (20 μM) for 1 h, followed by the addition of 50 μM cisplatin (H-Cis) for 22 h; Western blot analysis examined the proteins indicated. (B) Mitochondrial (Mito) fractions were isolated from HepG2 cells that were either untreated or pretreated with PFT-μ (20 μM) for 1 h, followed by the addition of 50 μM cisplatin (H-Cis) for 22 h. Western blots analyzed the proteins indicated. (C) IP-Western blot (WB) analysis confirms the presence of a p53/BAK complex in whole-cell extracts (WCE) prepared from cisplatin-treated HepG2 cells indicated in A. (D) Samples in A were examined by immunoblotting for the proteins indicated. The cleaved products of caspase-3 and caspase-8 are marked with arrows. (E) WCE prepared from primary MEFs (wild type or BAK^−/−) either untreated or treated with cisplatin (50 μM, 15 h) were examined for the proteins indicated. The cleaved products of caspase-3 and PARP are marked with arrows. (F) TUNEL staining of primary MEFs (wild type or BAK^−/−), either before or after cisplatin (50 μM, 15 h) treatment. The brown staining indicates positive TUNEL staining. Bars, 50 μm. (G) Quantification of TUNEL-positive cells in MEFs from F. Average of eight random 20× fields ± SD. (H) Mito fractions prepared from the same samples indicated in E were examined by immunoblotting for the proteins indicated. (I) IP-Western blot analysis confirms the presence of an p53/BAK complex in WCE prepared from the cisplatin-treated wild-type MEFs indicated in E.

Figure 4.

Exogenous IGFBP1 expression interferes with mitochondrial accumulation of p53, p53/BAK interaction, and p53-mediated BAK oligomerization. (A) Western blot (WB) analysis shows cleavage of caspase-8, caspase-3, and PARP in whold-cell extracts (WCE) prepared from U2OS treated with cisplatin (50 μM, 24 h). Note these antibodies detect both intact (top) and cleaved (bottom) forms of these proteins. IP-Western blots show a p53/BAK complex in U2OS treated with cisplatin (50 μM, 24 h). (B) IP-Western blot analysis reveals that IGFBP1 interacts with BAK in U2OS stably expressing exogenous IGFBP1 (U-BP1 cells). (C) Immunofluorescence analysis of U-BP1 cells for IGFBP1 protein (red) and BAK protein (green) before and after cisplatin (50 μM, 8 h). The merged image indicates colocalization of IGFBP1 with BAK (yellow). Bars, 20 μm. (D) Western blots of the indicated proteins in WCE or mitochondrial fractions (Mito) prepared from untreated or cisplatin-treated (50 μM, 15 h or 24 h) U2OS and U-BP1 cells. Note cleaved PARP is not detected in U-BP1 cells until 24 h following high dose of cisplatin, which blocks transcription (Cullinane et al. 1999; Jung and Lippard 2006). (E) Western blots of the indicated proteins in Mito fractions prepared from HepG2 cells, untreated or treated with Nutlin-3 (10 μM, 24 h). (F) Mito prepared from HepG2 cells untreated or treated with 10 μM of Nutlin-3 were incubated with 50 ng of purified p53 recombinant protein followed by IP-Western analyses. (G) The same samples (20 μg) indicated in E were incubated with 0, 5, 25, or 100 pmol of purified p53 recombinant protein. The reactions were cross-linked with 5 mM BMH (the uncleavable protein cross-linker 1,6-bismaleimidohexane) followed by Western analysis using an anti-BAK antibody (a conformation-specific antibody directed to the N terminus).

Figure 5.

IGFBP1 binds to BAK in mouse liver. (A) Western blots of whole-cell extracts (WCE) or culture media (Media) prepared following 10 μM Nutlin-3 or 3.5 μM cisplatin (L-Cis) treatment of the HepG2 cells. (B) Western blots show that IGFBP1 protein is induced and present in both Sera and mitochondrial (Mito) fractions of IGFBP1^+/+ (BP1^+/+) mice after overnight fasting, or after a 90-min treatment with each of the following: EGF, Fas mAb, or cisplatin. The cisplatin results are from two different mice. (C) IP-Western blots confirm the presence of an IGFBP1/BAK complex in liver WCE prepared from the same samples indicated in B. The results shown are representative of the analysis of at least three mice per genotype. (D) Western blot of proteins from WCE prepared from untreated IGFBP1^+/+ (BP1^+/+) and IGFBP1^−/− (BP1^−/−) livers. (E) Western blot of proteins from Mito fractions prepared from untreated BP1^+/+ and BP1^−/− livers. (F) IP-Western blot analysis confirms the presence of an p53/BAK complex in WCE prepared from untreated BP1^−/− liver indicated in D. (G) Proteins from liver WCE of control (untreated) BP1^+/+ and BP1^−/− mice were examined by Western blotting for the proteins indicated. (H) Proteins from liver Mito of control (untreated) BP1^+/+ and BP1^−/− mice were cross-linked with 5 mM BMH followed by Western analysis using an anti-BAK antibody. In addition to the BAK monomer (*), BAK aggregates (**) were detected in untreated IGFBP1-deficient livers.

Figure 6.

Increased injury and apoptosis after acute cisplatin treatment in IGFBP1^−/− livers. (A) Total (T), cytosolic (C), and mitochondrial (M) liver extracts were examined by Western blot for the proteins indicated, following a 90-min treatment of the animals with cisplatin or Fas mAb. The internal control BCL-xL is present in both cytosolic and mitochondrial fractions. (B–G) Hematoxylin and eosin (H&E) staining of control BP1^+/+ liver (B), BP1^+/+ liver 3 h after cisplatin (Cis) injection (C), control BP1^−/− liver (D), BP1^−/− liver 1 h after cisplatin challenge (E), BP1^−/− liver 3 h after cisplatin exposure (F), and BP1^−/− liver 3 h after cisplatin treatment (G). F is enlarged in G to show greatly enhanced hepatocellular injury and prominent histological features of apoptosis (i.e., cellular condensation, hypereosinophilic cytoplasm, and aggregation of chromatin at the nuclear membrane) in the BP1^−/− livers 3 h following acute cisplatin exposure. Bars, 100 μm. (H) Immunohistologic staining of control BP1^+/+ liver, control BP1^−/− liver, BP1^+/+ liver 3 h after cisplatin injection, and BP1^−/− liver 3 h after cisplatin challenge using anti-caspase-3 p17 antibody. The brown staining indicates positive anti-caspase-3 p17 staining. Arrowheads in control BP1^−/− liver indicate positive anti-caspase-3 p17 staining. Bars, 100 μm. (I) Western blot of proteins from mitochondrial (Mito) fractions prepared from control and cisplatin-treated BP1^+/+ and BP1^−/− livers. (J) IP-Western blot analysis reveals that Mito p53 interacts with BAK in both untreated and cisplatin-treated BP1^−/− livers. In contrast, formation of a BCL-xL/BAK complex is found in all cohorts before and 90 min following intraperitoneal cisplatin injection. The results shown are representative of the analysis of at least three mice per genotype.

Figure 7.

Markedly attenuated α-amanitin-mediated hepatic injury and apoptosis in p53^−/− or BAK^−/− livers. (A) Whole-cell extracts (WCE) were prepared from wild-type (WT) livers and examined by Western blot analyses for the proteins indicated, following a 90-min treatment of the animals with cisplatin (Cis) or 24 h α-amanitin (α-am). (B) H&E staining of control wild-type (WT) liver, control IGFBP1^−/− (BP1^−/−) liver, wild-type liver 24 h after α-amanitin treatment, and BP1^−/− liver 24 h after α-amanitin treatment. Bars, 100 μm. (C) Apoptosis assayed in the indicated livers by TUNEL or anti-caspase-3 p17 staining. The brown staining indicates positive TUNEL or anti-caspase-3 p17 staining. Arrowheads in control BP1^−/− liver indicate positive TUNEL staining. Bars, 100 μm. (D) Western blot of proteins from mitochondrial (Mito) fractions prepared from control and α-amanitin-treated wild-type livers. (E) H&E staining of control BAK^−/− liver, BAK^−/− liver 24 h after α-amanitin treatment, and α-amanitin-treated p53^−/− liver. Negligible parenchymal injury and focal hemorrhage were noted in the α-amanitin-treated BAK^−/− or p53^−/− livers relative to the corresponding wild-type or BP1^−/− livers (shown in B). Bars, 100 μm. (F) Markedly reduced positive TUNEL or anti-caspase-3 p17 staining noted in BAK^−/− liver 24 h after α-amanitin treatment. Bars, 100 μm. (G) TUNEL or anti-caspase-3 p17 staining in p53^−/− livers 24 h after α-amanitin treatment. Note the generally intact parenchyma in either the p53^−/− (E) or BAK^−/− (E) livers. Like the α-amanitin-treated BAK^−/− livers (F), trace amounts of positive TUNEL or anti-caspase-3 p17 staining were noted in p53^−/− liver 24 h after α-amanitin treatment, relative to the corresponding wild-type or BP1^−/− livers (C). Bars, 100 μm. (H) Western blot of proteins from Mito fractions prepared from control wild-type, control BAK^−/−, α-amanitin-treated wild-type, and α-amanitin-treated BAK^−/− livers. The results shown are representative of the analysis of at least three mice per genotype.