Pancreatic growth occurs after CCK, CCK-induced pancreatitis, and pancreatectomy; the mechanisms involved remain unknown. This study evaluates mitogen-activated protein kinase (MAPK) activation and expression of cell cycle regulatory proteins after pancreatectomy to understand the cellular and molecular mechanisms involved in pancreas regeneration. Rats were killed 1–12 days after pancreatectomy, and p42/p44 MAPK activation, expression of the cyclins D and E, cyclin-dependent kinase (Cdk)-2 activity, retinoblastoma protein (pRb) hyperphosphorylation, and expression of the cyclin kinase inhibitors p15, p21, and p27 were examined. Pancreatic remnants exhibited sustained p42/p44 MAPK activation within 8 h. Cyclins D1 and E showed maximal expression after 2 and 6 days, coinciding with maximal hyperphosphorylation of pRb and Cdk2 activity. The expression of p15 vanished after 12 h, p27 disappeared gradually, and p21 increased early. The p27 complexed with Cdk2 dissociated after 2 days, whereas p21 associated in a reverse fashion. In conclusion, sustained activation of p42/p44 MAPKs and Cdk2 along with overexpression of cyclins D1 and E and reduction of p15 and p27 cyclin inhibitors occurred early after pancreatectomy and are active factors involved in signaling that leads to pancreas regeneration.
- mitogen-activated protein kinase
it is now well accepted that the rat pancreas can be stimulated to grow (42) and that CCK is one of the most potent trophic factors (37). This organ also has a potential for regeneration after partial pancreatectomy (3, 32) and induced pancreatitis (20). The mechanisms involved in the regulation of normal pancreas growth and its regeneration have not yet been fully elucidated. Recently, the existence of complex signaling pathways has been described, and CCK was shown to be among the potent stimuli. Indeed, CCK activations of tyrosine kinases, phosphatidylinositol 3-kinase, and phospholipase D have been reported as early events (38), along with stimulation of the mitogen-activated protein kinase (MAPK) pathway involving Ras through tyrosyl-phosphorylated adaptor proteins (45). Among the short-term responses to CCK are the induction of the expression of c-fos, c-jun, and c-myc, members of the early responsive nuclear oncogene family reported to be associated with cell growth and differentiation (27, 40).
The extracellular factors and intracellular events that induce pancreatic acinar cell proliferation in the regenerating pancreas are incompletely characterized but are an area of active interest. Recent studies reported that pancreatic gene expression was altered during pancreas regeneration. Indeed, expression of genes encoding exocrine enzymes seems to be specifically repressed during acute experimental pancreatitis (19) and after partial pancreatectomy (4), whereas other genes associated with signal transduction, like H-ras, and cell proliferation, like c-myc, displayed increased expression for up to 48 h after pancreatectomy (4). We have also reported that the pancreatic gland overexpresses insulin-like growth factor I mRNA for up to 6 days after pancreatectomy and 9 days after acute pancreatitis (6), along with increases in hepatic growth factor and fibroblast growth factor mRNA after acute pancreatitis (5,7).
Progression through the G1 phase of the cell cycle is orchestrated by complexes containing a G1-specific cyclin and a G1-specific cyclin-dependent kinase (Cdk). Cyclin D family members are found associated with either Cdk4 or Cdk6 in early G1 phase and cyclin E with Cdk2 in late G1(10). Activity of these complexes can be inhibited by the presence of a cyclin-dependent kinase inhibitor (CKI), which belongs to an expanding new family of mammalian cell cycle modulators, so far including p21Cip, p27Kip1, and p57Kip2 in one subfamily and p16INK4A, p15INK4B, and p18INK4C in the other (10, 26,41). The INK4 proteins specifically inhibit Cdk4 and Cdk6, whereas the other family of inhibitors has less specificity, being involved in the inhibition of cyclin D-Cdk 4, cyclin E-Cdk2, cyclin A-Cdk2, and the mitotic complex cyclin B-Cdk1 (41). However, when overexpressed in cells, p21 and p27 cause only a G1 arrest, suggesting that, despite their ability to inhibit the mitotic cyclin B-Cdk1 complex in vitro, they do not act on mitotic Cdks in vivo (10, 26, 37, 41).
With most investigations on the respective roles of the cell cycle regulators such as cyclins, Cdks, and CKIs being performed on embryonic cells or in vitro systems using artificially synchronized dividing cells, we believe the regenerative pancreas offers the possibility of analyzing the cell cycle dynamic controls in naturally synchronized cells present in their natural native environment. This study was therefore undertaken to evaluate activation of the MAPKs p42/p44 and to analyze protein expression of some cell cycle regulatory proteins in the pancreatic remnant after partial pancreatectomy to better understand the cellular mechanisms involved in the early stages of pancreas regeneration.
MATERIALS AND METHODS
The following antibodies were used at the following dilutions: polyclonal cyclin E (1/1,000), Cdk2 (1 μg/tube), p15 and p21 (1/1,000) (Santa Cruz Biotechnologies, Santa Cruz, CA); monoclonal cyclin D1 (1/1,000) (NeoMarkers, Freemont, CA); monoclonal retinoblastoma protein (pRb) (1/200) (Pharmingen, Mississauga, ON, Canada); and monoclonal p27 (1/2,000) (Transduction Laboratories, Mississauga, ON, Canada). Polyclonal p42/p44 active (1/2,000) (Promega, Madison, WI) was purchased for Western blot analysis. Antiserum E1B, which specifically recognizes p42 and p44 MAPKs on Western blots, was a kind gift from Dr. Fergus McKenzie and Dr. Jacques Pousségur (Université de Nice, Nice, France). Histone H1 was purchased from Calbiochem (San Diego, CA); [γ-32P]ATP was from Amersham (Montréal, PQ, Canada). All other materials were obtained from Sigma unless otherwise stated.
Sprague-Dawley rats (100–120 g) from Charles River (St-Constant, PQ, Canada) were fed Purina rat chow ad libitum and kept in a room with controlled temperature and light (20°C, 12:12-h light-dark cycle). All studies were conducted in agreement with the principles and procedures outlined in the Canadian Guidelines for Care and Use of Experimental Animals.
A pancreatectomy of 90% was performed under Fluothane anesthesia according to the procedure of Foglia (14) described by Calvo et al. (6). Sham-operated animals had their pancreata manipulated with a cotton swab. Fed rats, controls, and pancreatectomized animals were killed by decapitation 8 and 12 h and 1, 2, 3, 6, and 12 days after the operation. Their respective numbers appear in the legend of each figure. Their pancreata were rapidly removed and homogenized in RIPA buffer (50 mg tissue/ml). Half of the homogenate was mixed in Laemmli's buffer, boiled 5 min, and frozen until time of the Western blots; the other half was supplemented with protease inhibitors (0.1 mM phenylmethylsulfonyl fluoride, 1 μg/ml pepstatin, 10 μg/ml aprotinin, and 10 μg/ml leupeptin), and the phosphatase inhibitor orthovanadate (0.2 mM) was saved and frozen for the Cdk2 assay. Proteins were determined by the modified Lowry procedure described by Peterson (34).
Fifty micrograms of protein from the Laemmli lysates were loaded and separated through 10% SDS-PAGE for all proteins except for pRb, which was run through a 7.5% SDS gel. After migration, gels were electrotransferred onto intracellulose (Amersham) for 2 h at 50 V. Western blots were blocked for 1 h at 37°C with 10% low-fat milk in PBS-0.1% Tween 20 before the incubation for 2–3 h at room temperature with the specific antibodies and then washed again in PBS-0.1% Tween 20 (4 × 10 min). This was sequentially followed by a 1-h incubation at room temperature with either anti-rabbit (for polyclonal antibodies) or anti-mouse (for monoclonal antibodies) conjugated IgG horseradish peroxydase. Peroxydase activity was revealed using an enhanced chemiluminescence kit from Amersham.
Immunoprecipitation experiments and Cdk2 activity.
The lysate of the RIPA (500 μg) was cleared by centrifugation (10,000g, 10 min) before a 2-h incubation at 4°C with protein A-Sepharose preincubated for 1 h with the Cdk2 antibody (4 μl:1 μg/tube). The immunocomplex was first washed four times with ice-cold buffer and three times with ice-cold kinase buffer (20 mM p-nitrophenyl phosphate, 10 mM MgCl2, 1 mM DTT, in 20 mM HEPES, pH 7.4) before the kinase assay was performed. The reaction was started by incubating the washed immunocomplex at 30°C for 30 min in the presence of 5 μg of histone H1, 2 μCi/assay of [32P]ATP, and 50 μM cold ATP. After 30 min, the reaction was stopped by addition of Laemmli 2× buffer. Radiolabeled histone H1 was separated from the immunocomplex by 10% SDS-PAGE and autoradiographed. For immunoprecipitations followed by Western blotting, lysates were incubated with α-Cdk2 cross-linked to beads for 2 h at 4°C. Beads were pelleted briefly and washed twice with the lysis buffer before electrophoresis. Immunoprecipitated proteins were analyzed by Western blotting and then probed with p27 or p21 antibodies.
Modulation of p42/p44 MAPK activities after pancreatectomy.
Previous studies have demonstrated in rat pancreatic acini that the trophic agent CCK can activate the three MAPK cascades: p42/p44 MAPK, p46/p54 c-Jun NH2-terminal kinase/stress-activated protein kinase, and p38/Hog MAPK (45). Unfortunately, this model of acini was inappropriate to study the biological events preceding mitogenesis and occurrence over hours and days. We thus selected the partial pancreatectomy model in the rat because it gives the opportunity to analyze the cell cycle dynamic of naturally synchronized cells in their normal environment. Because activation of the Ras/MAPK cascade is required to pass the G1 restriction checkpoint to progress into the cell cycle (31), this critical reaction was then first investigated.
As shown in Fig.1 A using a polyclonal antibody specific for the detection of the active phosphorylated p42/p44 isoforms, sham-operated controls exhibited minimal kinase activities throughout the entire time course study. In contrast, pancreatectomy significantly increased p42/p44 activities as early as 8 h after the operation. This activation is sustained for at least 12 days and seems already maximal after the first day up to 6 days. Activity of the p42 isoform is always superior to that of p44 at any given time point. When the data are expressed as active p42/p44 MAPK over total p42/p44 protein levels (Fig.1 B), maximal activity is observed 8 h after pancreatectomy and remained elevated up to 12 days after surgery.
Effects of pancreatectomy on cyclins D1 and E protein expression.
In this study, we elected to determine the expression of cyclin D1 and cyclin E as representative of the events happening in the early and late G1 phases, respectively (22, 35). As seen in Fig.2 A, cyclin D1 is undetectable in the sham control animals, but expression of the protein became easily detectable 8 h after pancreatectomy, with increasing expression up to a maximum after 6 days (more than 12-fold) and a decrease at day 12. Expression of cyclin E is evident in sham control rats and remained constant during the first 24 h after pancreatectomy. Maximal expression of the protein was observed after 2 days with a fourfold induction, sustained up to 6 days, and regressed to control values after 12 days (Fig.2 B).
Effects of pancreatectomy on the pRb phosphorylation and Cdk2 activation.
Among the physiological substrates of the cyclin-Cdk complexes whose phosphorylation is likely to be relevant to the G1/S transition is the gene product pRb. On progression through G1, pRb is phosphorylated by cyclin-Cdk, which causes the release and activation of the E2F/DP transcription factors; this process constitutes a key event for progression to S phase (12,29). As shown in Fig.3 A, pRb remained unaltered for the first 24 h after pancreatectomy; its hyperphosphorylation was maximal after 2 days, remained as such for the next 4 days, and returned to basal control values onday 12 postoperation.
As indicated earlier, cyclin E in association with Cdk2 is required for the G1/S transition (22), and cyclin A, again in a complex with Cdk2, is essential for progression through S phase (16). As shown in Fig. 3 B, immunoprecipitated Cdk2 exhibited a very low basal activity in phosphorylating histone H1 that was extended during the first 24 h after pancreatectomy. However, on days 2–6 postoperation, enormous Cdk2 activity was present in remnant pancreata, activity that was still above control values 12 days after pancreatectomy but at a much slower pace that atdays 2 and6. Notice the parallelism between maximal pRb hyperphosphorylation and Cdk2 activation.
Expression of Cdk inhibitors during pancreas regeneration.
The major roles of the Cdk inhibitors are to mediate cell cycle arrest in response to antimitogenic factors and to ensure that specific cell cycle events do not initiate before others are completed (18, 33). Among these inhibitors are the two structurally related p21 and p27, which form complexes with cyclin/Cdk and possibly block their interaction with substrates (33). Among the second family of inhibitors are p16 and p15, which exclusively target Cdk4 and Cdk6 and prevent their binding to cyclins (17).
As shown in Fig. 4, p21 expression is minimal in sham-operated controls and became maximally expressed (more than 2.5-fold) within the first 12 h after pancreatectomy; its presence remained at a high level through day 6postoperation, and its expression reduced at day 12. In contrast, expression of p27 is quite high in the sham-operated controls, slightly decreased during the first 24 h after operation, decreased by 48% from days 2–6, and increased toward basal values byday 12 (Fig.4 B). The strongest expression of p15 was exhibited in the sham-operated animals; it became undetectable at 12 h after pancreatectomy and reappeared slightly onday 12.
Interaction of p21 and p27 with Cdk2 after pancreatectomy.
We next examined the potential role of p21 and p27 in the regulation of Cdk2 activity in the regenerating pancreas. We immunoprecipitated Cdk2 and subjected the precipitated proteins to Western blot for p21 and p27 (Fig. 5). The amount of p27 immunoprecipitating with Cdk2 inversely correlated with the histone H1-kinase activity (Fig. 3 B and Fig.5 A). This suggests that p27 plays a role in downregulating Cdk2 activity before and after peak of pRb hyperphosphorylation and DNA synthesis. In contrast, p21 mostly associates with Cdk2 on days 1–12 postpancreatectomy (Fig.5 B), suggesting that this inhibitor may be involved in governing progression through the cell cycle (10).
In this study, we determined that regeneration was initiated in the pancreatic remnant after 90% pancreatectomy. Such an initiation is illustrated by sustained long-term activation of the MAPK p42/p44, the timely expression of cyclins D1 and E, both directly involved in cell cycle G1 progression and G1/S transition, respectively, the simultaneous hyperphosphorylation of pRb, and activation of Cdk2. During that regenerative process, we also observed an immediate downregulation of the CKI p15, a progressive decrease in p27, and an early upregulation of p21.
Previous studies reported that regeneration after subtotal pancreatectomy was accompanied by early overexpression of c-myc and H-ras mRNA (4) and elevated mitotic indices at two days (6) up to two weeks (3) after surgery. This latter information indicates that specific reactions involved in the control of the cell cycle progression exhibit sustained activation. We indeed reported significant increases in membrane tyrosine kinase and phospholipase D activities three days after pancreatectomy (38).
Among the signal transduction mechanisms underlying cellular proliferation and differentiation is the MAPK cascade. In this cascade, the MAPK p42/p44 were stimulated within 5 min in isolated pancreatic acini in response to CCK (11), an essential factor in pancreatic regeneration after 60% distal pancreatic resection (32). Our study indicates for the first time in the rat that activation of p42/p44 MAPK is sustained for at least 12 days after pancreatectomy; a previous observation in the pig also indicated that MAPK remained active four weeks after pancreatectomy (13). Such a sustained activation was previously stated as required for continued expression of cyclin D1 in G1 phase in fibroblasts (24, 44). These sustained p42/p44 MAPK activities strongly support their integral involvement in the control of the cell cycle during pancreas regeneration.
The only data available on cyclin-Cdk implication in the pancreas indicate that cyclin D1 protein overexpression occurs in pancreatic endocrine tumors (8) and that expression of a cyclin D1 antisense construct significantly reduced the growth factors' mitogenic capacity in human pancreatic PANC-1 cancer cells (23). In vertebrates, D-type cyclins are considered to be key regulators of G1progression, and their association with Cdk4/Cdk6 is regulated by extracellular signals. In response to pancreatectomy, we already mentioned the early and sustained activation of the p44/p42 MAPKs, and this activation coincides with a strong induction of cyclin D1 protein that exhibited its highest level 2 and 6 days after operation. Almost in parallel to cyclin D1 protein synthesis, we can also observe that of cyclin E, which, in association with Cdk2, is involved in the G1/S transition phase of the cell cycle. Our data also indicate that peak expression of cyclin E protein coincides with maximal activity of its complement Cdk2 and increased DNA synthesis previously observed under identical experimental conditions (6). Also in line with activation of this integrated system of cyclin-Cdk is the hyperphosphorylation and inactivation of pRb, a key event for progression to S phase, leading to the release and activation of the transcription factor E2F to activate E2 site-containing gene promoters (12). This phenomenon of hyperphosphorylation of pRb by cyclin-Cdk complexes (29) is initiated 1 day after pancreatectomy, exhibited its highest state after 2 days, and reached control values after 12 days.
Cell cycle progression is controlled by cyclin-Cdk complexes counterbalanced by Cdk inhibitors. After pancreatectomy, we can appreciate the dynamic of the Cdk inhibitors and their expression seems to follow the predicted patterns, in which p27 has the tendency to accumulate in quiescent cells and decrease in response to mitogenic stimulation (10, 37), whereas levels of p21 are generally low in quiescent cells and increased in response to mitogenic stimulation (10).
Within this Cdk inhibitor dynamic, p21 expression parallels that of the cyclins, and the important feature with regard to this specific inhibitor would be its ratio to cyclin-Cdk within the complex. Our results demonstrate that p21 mostly associates with Cdk2 ondays 2 and6 after pancreatectomy, a time when we can observe the highest Cdk2 activity. Recent studies have demonstrated the existence of such p21 cyclin-Cdk complexes that are largely active (46). It thus appears that p21-containing enzymes can transition between active and inactive states, probably through changes in the stoichiometry of the p21 subunit. The complex is inactive when many p21 molecules associate with it. In our study we have not established such a ratio, but the coupled p21 cyclin-Cdk2 is active because cells are progressing into S phase. A recent study also reported that p21 protein expression is induced 12 and 24 h after hepatectomy in mice (1). Moreover, compared with cogenic wild-type mice, p21 knockout mice demonstrated evidence of markedly accelerated hepatocyte progression through the G1 phase after hepatectomy, with DNA synthesis occurring earlier in p21 knockout mice. The authors suggested that p21 may regulate the rate of progression through the G1 phase of the cell cycle in vivo (14, 41).
Elevated expression of p27 has been previously documented in contact-inhibited or mitogen-deprived cells, and its expression often declines on proliferation stimulation (9, 30, 36, 37). This pattern of expression can be found in normal pancreas and in response to pancreatectomy, because elevated p27 levels were observed in normal functioning pancreas with reduced expression beginning 2 days after operation. Moreover, immunoprecipitation experiments also suggest that p27 may act as an inhibitor of Cdk2 activity in the pre- (0–1 day) and postreplicative phases of pancreas regeneration. The importance of p27 abrogation in suppressing cell quiescence has been recently demonstrated in fibroblasts, as its cDNA antisense expression suppressed their G0 state (9, 37). Conversely, it was demonstrated that targeted disruption of the murine p27 gene enhances growth of the mice (21).
Much less is known about the physiological role of p57Kip2, the last member of the Cip/Kip family of mammalian Cdk inhibitors (25). p57 is a potent inhibitor of G1- and S-phase Cdks (cyclin E-Cdk2, cyclin D-Cdk4, and cyclin A-Cdk2) and, to lesser extent, of the mitotic cyclin B-cdc2. Moreover, in contrast to the widespread expression of p21 and p27 in human tissues, p57 is expressed in a tissue-specific manner in the placenta, skeletal muscle, heart, kidney, and pancreas (25). Our attempts to identify p57 by immunoblot analysis in our system were not successful. Thus we could not exclude a potential role of this cell cycle inhibitor in the regulation of the cell cycle in pancreatic acinar cells after pancreatectomy.
In light of the present findings obtained in the rat, do we foresee any clinical relevance of these data? This is a difficult question because similar studies were never performed in humans, and the few data available on human pancreas growth are not really comparable. Indeed, Friess et al. (15) have recently demonstrated that the human pancreas exhibited tissue hypertrophy in response to a 4-wk feeding of the proteinase inhibitor camostate; gland hyperplasia could not be measured in that study. However, preliminary data from Barry et al. (2) provided no evidence that the human pancreas regenerated after a partial anatomic 50% resection. On the contrary, in a better controlled study, Fiorucci et al. (13) recently indicated that the pig pancreas regenerated after a subtotal (70%) distal pancreatectomy in response to a bombesin treatment. In these pancreatectomized pigs, after four weeks of bombesin, Shc phosphorylation, Shc-Grb2 association, and MAPK activation were observed in freshly prepared acini. We have previously demonstrated similarities between the pig and human pancreata with regards to their CCK-B receptor subtypes (28). Furthermore, acini prepared from pigs (28) and human pancreata (43) exhibited similar secretory responses to CCK and some of its analogs. Although it still remains to be demonstrated, we strongly feel that the human pancreas can regenerate, as did the pig pancreas after resection; we need to find the appropriate therapy.
In this study, we clearly described the dynamic existing between some cell cycle stimulatory and inhibitory regulatory proteins in normal rat regenerating pancreatic gland after subtotal pancreatectomy. The induction of such a process could be under the control of several growth factors, such as insulin-like growth factor-I (6), hepatic growth factor (5), and fibroblast growth factor (7), whose mRNA expressions were all stimulated in response to pancreatic damage.
We acknowledge Pierre Pothier and Christiane Ducharme for secretarial assistance.
Address for reprint requests and other correspondence: J. Morisset, Service de gastro-entérologie, Dépt. de médecine, Faculté de médecine, Université de Sherbrooke, Sherbrooke, Québec, Canada J1H 5N4 (E-mail:).
This study was supported by grants from Natural Sciences and Engineering Research Council of Canada Grant GP 6369 to J. Morisset, Medical Research Council Grant MT 14405 to N. Rivard, and le Ministère de l'Éducation du Québec Grant ER 1092.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 1999 the American Physiological Society