Research ArticleDRUG DEVELOPMENT

The nuclear translocation of the kinases p38 and JNK promotes inflammation-induced cancer

See allHide authors and affiliations

Sci. Signal.  10 Apr 2018:
Vol. 11, Issue 525, eaao3428
DOI: 10.1126/scisignal.aao3428

A peptide to treat inflammatory disease

Chronic inflammation promotes the development of tumors in various tissues, notably in the gastrointestinal tract. Inflammation induces proliferation and feedforward inflammatory cytokine production in cells typically through the kinases p38 and JNK, which are associated with various diseases. Maik-Rachline et al. developed a peptide that blocked the activities of p38 and JNK specifically in the nucleus. The peptide, called PERY, prevented the development of inflammation-associated colon cancer in mice. Furthermore, without detectable side effects, the peptide protected mice from chemical-induced colitis that models irritable bowel syndrome. These findings suggest that this peptide might be used to treat colitis in patients and prevent the development of colitis-induced colon cancer.

Abstract

The stimulated nuclear translocation of signaling proteins, such as MAPKs, is a necessity for the initiation and regulation of their physiological functions. Previously, we determined that nuclear translocation of the MAPKs p38 and JNK involves binding to heterodimers comprising importin 3 and either importin 7 or importin 9. Here, we identified the importin-binding region in p38 and JNK and developed a myristoylated peptide targeting this site that we called PERY. The PERY peptide specifically blocked the interaction of p38 and JNK with the importins, restricted their nuclear translocation, and inhibited phosphorylation of their nuclear (but not cytoplasmic) substrates. Through these effects, the PERY peptide reduced the proliferation of several (but not all) cancer cell lines in culture and inhibited the growth of a human breast cancer xenograft in mice. In addition, the PERY peptide substantially inhibited inflammation in mice, as manifested in models of colitis and colitis-associated colon cancer. The PERY peptide more effectively prevented colon cancer development than did a commercial p38 inhibitor. In vivo analysis further suggested that this effect was mediated by PERY peptide–induced prevention of the nuclear translocation of p38 in macrophages. Together, these results support the use of the nuclear translocation of p38 and JNK as a novel drug target to treat various cancers and inflammation-induced diseases.

INTRODUCTION

Stimulated translocation of signaling proteins across the nuclear envelope is an essential signaling mediator, which may directly affect gene expression and consequently induce and regulate various cellular processes. Despite being a highly important process, the molecular mechanisms driving the stimulated nuclear translocation of signaling proteins are not completely understood. The classical translocation mechanism uses the nuclear localization signal (NLS)–mediated binding with importin-α (Impα) and Impβ (1). Some signaling proteins such as nuclear factor κB (NF-κB) (2) and extracellular signal–regulated kinase 5 (ERK5) (3) seem to use this translocation mechanism upon various stimulations. However, most other signaling proteins do not contain the canonical NLS and translocate to the nucleus by other Impα/β-independent mechanisms. Many of these proteins seem to use members of the distinct group of β-like importins for their nuclear translocation (4). For example, the translocation of transcription factor SMAD4 requires the β-like importin members Imp7 and Imp8 (5), c-Fos is shuttled by Imp2 (6), and the vitamin D receptor is escorted to the nucleus by Imp4 (7).

Another group of signaling proteins that require β-like importins for stimulated translocation are the mitogen-activated protein kinases (MAPKs) (8). This translocation is important for the physiological and pathological functions of these kinases, particularly for inducing ERK-dependent proliferation (9, 10). We have identified two distinct nuclear translocation mechanisms: one for ERK1/2 (ERK) (11) and another for each of the MAPK isoforms p38α/β and c-Jun N-terminal kinase 1/2 (p38 and JNK) (12). For ERK, we showed that the transition from cytoplasmic anchoring protein-associated localization in resting state (13) to stimulated nuclear ERK is the result of its sequential phosphorylation by MAPK/ERK kinase (MEK) and casein kinase II (CKII). Phosphorylation by MEK not only causes the activation of ERK but also exposes a unique nuclear translocation sequence (NTS). This enables the phosphorylation of two serine residues (SPS) within this region that induces ERK binding to Imp7. This latter carrier then escorts the kinase to the nucleus via the nuclear pores, where it phosphorylates a large number of targets, mainly transcription factors (4, 14, 15). We, as well as others, have found that the nuclear translocation mechanism described above is specific for ERK, MEK, GLI family zinc finger 1 (GLI1), protein tyrosine kinase 2 (Pyk2), and early growth response protein 1 (Egr1) (11, 1618), but distinct from the mechanisms used by other MAPKs, such as p38 and JNK (12).

Because the nuclear translocation of ERK is mainly important for regulating proliferation (9, 19), we then investigated whether prevention of its translocation can inhibit cancer growth. For this purpose, we developed an NTS-derived myristoylated phosphomimetic peptide (EPE peptide), which blocks the interaction of Imp7 and ERK, and consequently the nuclear translocation of ERK (20). This peptide prevented the growth of several cancer-derived cell lines but not of nontransformed immortalized cells. Moreover, in xenograft models, the peptide inhibited the growth of several cancers, obtaining a much better effect than PLX4032 in preventing melanoma recurrence. Notably, this peptide serves as a proof of concept for using the nuclear translocation of ERK as a drug target to combat various cancers.

The other translocation mechanism involved in p38 and JNK accumulation in the nucleus is somewhat different from that of ERK. Similar to ERK, p38 and JNK are retained in the cytoplasm by binding to anchoring proteins. Upon stimulation, the anchoring proteins are phosphorylated to facilitate their detachment from the MAPKs, and translocation of these MAPKs is mediated by Imp3, Imp7, and Imp9. Thus, upon stimulation, p38 and JNK bind to either Imp7 or Imp9, whereas Imp3 joins to any of these formed dimers after its stimulation-induced phosphorylation. This forms heterotrimers, composed of either Imp3/Imp7/MAPK or Imp3/Imp9/MAPK, which move to the nuclear envelope where Imp3 remains, whereas Imp7/9 escorts the MAPKs into the nucleus (12).

Here, we took advantage of the strategy used for inhibiting ERK translocation, to inhibit p38 and JNK nuclear translocation. We were interested in determining whether the inhibition of nuclear translocation affects several pathologies and could be used as a therapeutic strategy to treat cancer. For this purpose, we identified the interaction motif of p38 and JNK with Imp7 and Imp9 and designed a peptide (PERY peptide) targeting this specific site. We further demonstrated that the PERY peptide prevented p38 and JNK interaction with Imp7 and Imp9, thereby inhibiting their nuclear translocation upon stimulation. It is well established that p38α/β pathways play a central role in the regulation of cellular responses to stress as well as the induction and progression of inflammation-related diseases and inflammation-induced cancer (2124) and the initiation and progression of several cancers (25, 26). We therefore tested the therapeutic potential of the peptide in cancer as well as in inflammation-induced cancer models. We demonstrated a significant anti-proliferative effect of the PERY peptide on several breast cancer cells and a marked reduction in growth of PERY peptide–treated breast cancer xenografts. Moreover, using a model of inflammation-associated dextran sulfate sodium (DSS)–induced acute colitis, we showed that the PERY peptide significantly reduces inflammation and intestinal damage. Finally, we applied the peptide to a model of combined azoxymethane (AOM)/DSS colitis–associated colon cancer and showed that systemic treatment with the PERY peptide significantly prevents colon cancer. Treatment with the PERY peptide significantly reduced tumor load and maintained healthier colon histology, even better than colons of mice treated with commercial p38 inhibitor. The effect of the PERY peptide was mediated, at least in part, by prevention of the nuclear translocation of p38 in macrophages. Our results serve as a proof of concept for the therapeutic potential of preventing p38 and JNK nuclear translocation in treating cancers and inflammation-related diseases.

RESULTS

p38α interacts with β-like importins through its N terminus

We have previously demonstrated that the β-like importins Imp3, Imp7, and Imp9 mediate the stimulated nuclear translocation of the MAPKs p38 and JNK. However, the interaction sites of p38 and JNK with Imp7/9 were distinct from those responsible for binding of ERK1/2 interaction with Imp7. We were therefore interested in identifying the interaction motif(s) of p38α/β and JNK1/2 with Imp7 and Imp9. For that purpose, we used p38α as a prototype, deleting regions in its C or N terminus (Fig. 1A), and first assessed their subcellular localization in HeLa cells. As expected, the overexpressed wild-type p38α was localized in the nucleus of both resting and stressed (anisomycin-treated) cells (Fig. 1, B and C). Deletion of 40 amino acids from the C terminus of p38α (ΔC40) resulted mainly in nuclear and cytoplasmic staining for p38 in resting cells, but its abundance in the nucleus still increased after stimulation. Deletion of the N-terminal 20 amino acids of p38α (ΔN20) did not affect the subcellular localization, whereas deletion of N-terminal 30 amino acids (ΔN30) significantly prevented p38α’s nuclear localization in both resting and stimulated cells (Fig. 1, B and C). These results indicate that although the C terminus may affect p38α localization in resting cells, the stimulation-related translocation sequence lies within amino acids 20 to 30 at the N terminus of the protein.

Fig. 1 Identification of the nuclear translocation signal in p38α.

(A) Schematic representation of truncation mutants of p38α. N, N terminus; KD, kinase domain; C, C terminus. (B and C) Representative images (B) and quantification (C) of fluorescence microscopy of p38 staining in HeLa cells transfected with green fluorescent protein (GFP)–tagged wild-type (WT) or truncated (ΔC40, ΔN20, or ΔN30) p38α, serum-starved for 16 hours, and then untreated (NT) or stimulated with anisomycin (Anis; 1 μg/ml) for 15 min. Scale bar, 15 μm. Data are the means ± SE. Percentage of cells with mostly nuclear (N, gray), all over [nuclear and cytoplasmic (NC); black], or mostly cytoplasmic (C, white) staining in at least three fields with >100 cells per field. **P < 0.01 by two-sample test for equality of proportions. (D and E) Representative blots (D) and quantification (E) of the CoIP of Imp7 or Imp9 with antibody to GFP in lysates from HeLa cells transfected with GFP-tagged WT or truncation-mutant (ΔN7, ΔN20, or ΔN30) p38α and treated as described in (B). Data are means ± SE. **P < 0.01 by two-way analysis of variance (ANOVA) followed by Tukey post-tests. (F) Sequence alignment of the PERY peptide with p38α/β and JNK1/2. “*”, identical amino acids (red); “:”, similar amino acids (green). (G and H) Representative blots (G) and quantification (H) of the CoIP of Imp7 with p38α antibody in lysates from HeLa cells that were serum-starved and preincubated with SCR peptide or peptides composed of residues 21 to 29, 15 to 29, or 21 to 34 of p38α (10 μM for 2 hours), stimulated with anisomycin (1 μg/ml for 15 min) or untreated (NT). Data are means ± SE; **P < 0.01 by two-way ANOVA followed by Dunnett’s post-tests. All data are representative of at least three independent experiments. A.U., arbitrary units.

Furthermore, we examined the effect of deleting the N-terminal 7 (ΔN7), 20 (ΔN20), or 30 (ΔN30) amino acids on the interaction of p38α with Imp7 and Imp9. By coimmunoprecipitation (CoIP), we detected some interaction between Imp7 or Imp9 and wild-type p38α, which were increased upon stimulation with anisomycin, in HeLa cells (Fig. 1, D and E). The interaction of the three mutants with Imp7 or Imp9 was not significantly changed in resting cells. However, an increased interaction was seen for ΔN7 and more so ΔN20 p38 mutants upon anisomycin stimulation, whereas the anisomycin-responsive interaction of the ΔN30 p38 was markedly decreased compared to that of wild-type p38. These results indicate that the interaction site with importins lies within N-terminal residues 20 to 30 of p38α. The increased interaction of ΔN7 and more so ΔN20 is probably the result of better exposure of the binding site that lies just after the deleted sequence. To confirm that the lack of importin interaction was not related to impairment of the mutant’s activity, we immunoprecipitated the overexpressed ΔN20 and ΔN30 p38 mutants from HeLa cells and determined their kinase activity in vitro by their ability to phosphorylate the p38 substrate myocyte enhancer factor 2A (MEF2A). Both mutants similarly phosphorylated MEF2A in response to stimulation with anisomycin, although to a lesser extent than did wild-type p38α (fig. S1, A and B). These results further indicate that the site of p38 interaction with β-like importins, which mediates the stimulated nuclear translocation, lies within N-terminal 20 to 30 amino acids of p38α. Notably, the N-terminal region is highly homologous to p38β, JNK1, and JNK2 (Fig. 1F), but is distinct from that of ERK1/2 (~40% sequence homology). This indicates that the nuclear translocation signals of these MAPKs are distinct, explaining the difference in the mechanism of nuclear translocation between ERK and p38 and JNK (11, 12).

PERY peptide interferes with the p38α-Imp7/9 interaction

In a previous report (20), we used a myristoylated peptide to compete for the interaction of Imp7 with the NTS of ERK1/2. This specific sequence lies within the loop region of the kinase insert domain of ERK1/2. This peptide was proven effective in preventing the nuclear translocation of ERK1/2 and thereby prevented proliferation and xenograft growth of several transformed cell lines. Because the Imp7/9 interaction sequence with p38α lies within an exposed loop as well (fig. S1, C and D), we thought to use the same strategy and designed a competing peptide for this region. Thus, we first synthesized an N-terminally myristoylated 9–amino acid peptide based on residues 21 to 29 of p38α. This peptide did not inhibit the Imp7-p38α interaction in HeLa cells upon anisomycin stimulation, but actually increased it for unknown reasons (Fig. 1, G and H). According to our experience, such a short myristoylated peptide might be unstable and its degraded products may stimulate undesired effects. Therefore, we designed longer myristoylated peptides for the same purpose that are much more stable. One of these peptides, based on the sequence of residues 15 to 29, showed an elevation in the nonstimulated Imp7-p38α interaction, without a significant change upon stimulation. The second, based on the sequence of residues 21 to 34, had some effect on the basal interaction, but significantly abolished the stimulated one (Fig. 1, G and H). These results indicate that peptide based on the sequence within residues 21 to 29 competes with the Imp7-p38α interaction when located within longer sequences. However, despite the significant effect of the myristoylated 21–34 peptide, we were worried that it might be unspecific, as its last four amino acids (GSGA) contained the beginning of the kinase domain, which is similar in all kinases (27). To avoid such lack of specificity, we substituted the last four residues with Ala, forming a peptide whose sequence is unique to p38α and has a significant similarity only to JNK. This myristoylated peptide, which contains the interaction site within a longer, more stable sequence, was termed PERY peptide and was used in the rest of our experiments. To ensure the penetration of the peptide into the cytoplasm, where it should act, we conjugated biotin to its C terminus. Using this peptide, we found that it is already detected in the cytoplasm of HeLa cells 2 hours after incubation and is retained there for at least 24 hours (fig. S2A).

The PERY peptide inhibits the interaction with importins and importin-mediated nuclear translocation of p38α/β and JNK1/2

We next examined whether the PERY peptide affects the interaction between Imp7/9 and JNK1/2 or p38α/β. For that purpose, HeLa cells were preincubated with the peptide, and the ability of p38 and JNK to interact with Imp7/9 was examined using CoIP. As expected, no significant interactions of these MAPKs with Imp7/9 were detected in basal state. However, a significant increase in anisomycin-induced interaction was detected in the control samples preincubated with dimethyl sulfoxide (DMSO) or scrambled (SCR) peptide. This stimulation-dependent interaction was strongly inhibited in samples preincubated with the inhibitory PERY peptide (Fig. 2, A and B). A similar observation was demonstrated using proximity ligation assay (PLA), which is a useful method to detect interaction between endogenous proteins. Anisomycin stimulation of HeLa cells preincubated with the SCR peptide significantly increased the interaction between Imp7 and p38α, whereas this interaction was significantly abolished when cells were treated with the PERY peptide (Fig. 2C). Because the PERY peptide so markedly inhibited the Imp7/9-p38/JNK interactions, we investigated whether this effect was unrelated to the kinase activity of p38. We therefore immunoprecipitated overexpressed wild-type p38α from HeLa cells and determined its kinase activity toward MEF2A in vitro in the presence of the PERY or SCR control peptides. p38α phosphorylated MEF2A in basal state and, after anisomycin stimulation, to a similar extent in the presence of the PERY or SCR peptides (Fig. 2D).

Fig. 2 The PERY peptide inhibits JNK1/2 and p38α/β interaction with Imp7/9.

(A and B) Representative blots (A) and quantification (B) of the CoIP of Imp7 or Imp9 with antibody to p38α, p38β, JNK1, or JNK2 in lysates from HeLa cells that were serum-starved [0.1% fetal bovine serum (FBS), 16 hours], preincubated with PERY or SCR peptides or DMSO (10 μM, 12 hours), and then stimulated with anisomycin (1 μg/ml, 15 min) or left untreated (NT). Data are means ± SE; *P < 0.05 and **P < 0.01 by one-way ANOVA. (C) Representative images (left) and quantification (right) of PLA analysis using anti–general p38 and Imp7 antibodies of HeLa cells that were serum-starved (0.1% FBS, 16 hours), preincubated with PERY or SCR peptides (10 μM each for 2 hours), and then stimulated with anisomycin (1 μg/ml, 15 min) or left untreated (NT). Scale bar, 10 μm. Data are means ± SE. **P < 0.01 by paired t test. (D) Representative blots (left) and quantification (right) of p38 in vitro kinase assay using MEF2A as a substrate. HeLa cells were transfected with GFP-p38α followed by starvation (16 hours, 0.1% FBS) and subjected to anisomycin stimulation (1 μg/ml, 15 min, +) or left untreated (NT, −). The cells were lysed and the extracts were subjected to IP using anti-GFP antibody. The immunoprecipitated kinase was then subjected to in vitro kinase assay using MEF2A in the presence of PERY or SCR peptide (10 μM). Data are means ± SE. All data are representative of at least three independent experiments.

Given that the PERY peptide competed with the interaction of p38 and JNK with Imp7/9, we speculated that it should affect the nuclear translocation of the kinases as well. To examine this, HeLa cells were preincubated with the PERY peptide or control SCR peptide and then stimulated. Subsequently, the subcellular localization of the endogenous p38α and JNK1 was determined using specific antibodies to these isoforms. Preincubation with the PERY peptide, but not the SCR peptide, inhibited the stimulation-dependent nuclear translocation of both kinases, verifying the importance of the N-terminal region of these MAPKs in the regulation of their subcellular localization (Fig. 3A). Similar effects were observed with the endogenous JNK2 or p38β isoforms when these two isoforms were stained (fig. S2, B and C). The translocation-prevention ability of the PERY peptide was specific to p38 and JNK, because it did not affect the subcellular localization of other signaling kinases such as ERK or AKT (Fig. 3, B and C). We also demonstrated the ability of the PERY peptide to inhibit p38 nuclear translocation by cell fractionation. For that purpose, HeLa cells were preincubated with the PERY peptide or control SCR peptide and, after anisomycin stimulation, were subjected to cellular fractionation. As reported for ERK (28), the large abundance of the kinases in the cytoplasm was caused by leakage via the nuclear pores (Fig. 3D), and this fractionation system only allows the detection of localization changes in the nucleus. Upon stimulation, p38 shifted to the nuclear fraction in the SCR peptide–treated cells, and this was reduced in cells treated with the PERY peptide (Fig. 3D). Together, the staining and fractionation identified residues 20 to 29 as a nuclear translocation signal, specific for p38α/β and JNK1/2, which mediates the subcellular localization of these MAPKs.

Fig. 3 The PERY peptide inhibits the stimulation-dependent nuclear translocation of JNKs and p38s but not ERK or AKT.

(A to C) Representative images (A and B) and quantification (C) of fluorescence microscopy of p38α and JNK1 antibodies (A) or ERK and AKT antibodies (B) in HeLa cells that were preincubated with the PERY or SCR peptides (10 μM each, 2 hours) and then treated with anisomycin (1 μg/ml, 15 min) or left untreated (NT). The nuclei were detected using 4′,6-diamino-2-phenylindole (DAPI). Scale bar, 20 μm. Data are means ± SE. Percentage of cells with mostly nuclear (N, gray), all over [nuclear and cytoplasmic (NC); black], or mostly cytoplasmic (C, white) staining in at least three fields with >100 cells per field. **P < 0.01 by two-sample test for equality of proportions. (D) Representative blots (upper) and quantification (bottom) of p38 nuclear fractions of HeLa cells, which were treated as described in (A) and were further subjected to subcellular fractionation as described in Materials and Methods. A control cytoplasmic fraction (C) and treated nuclear fractions (N) were subjected to Western blot analysis. Data are means ± SE. *P < 0.05 by one-way ANOVA. All data are representative of at least three independent experiments.

The PERY peptide specifically inhibits phosphorylation of nuclear but not cytoplasmic p38 and JNK targets

To eliminate any possibility of the peptide halting the kinase domain of p38 or that it nonspecifically interferes with other signaling components, we determined the phosphorylation state of nuclear and cytoplasmic targets of p38 and JNK. Thus, HeLa cells were preincubated with the inhibitory peptide or the SCR peptide followed by anisomycin stimulation, and the phosphorylation of the transcription factors c-Jun, MEF2A, activating transcription factor 2 (ATF2), and the p38 substrate MAPK-activated protein kinase 2 (MK2) was determined. As expected, anisomycin induced a strong phosphorylation increase of all transcription factors and substrates tested, as seen in cells treated with the SCR peptide (Fig. 4A). However, treatment with the PERY peptide significantly reduced the enhanced phosphorylation of the nuclear transcription factors c-Jun and MEF2A, confirming the lack of active p38 and JNK in the nucleus. However, phosphorylation of the p38 targets ATF2 (29) and MK2 (30) that are localized mostly in the cytoplasm was not inhibited by the PERY peptide and, in some cases, even increased compared to the control. This suggests that the peptide does not affect p38 and JNK activation, but rather inhibits their nuclear translocation. Therefore, the amount of active p38 and JNK in the nucleus is reduced, leading to a reduced activation of nuclear targets and to a possible hyperactivation of the cytoplasmic ones.

Fig. 4 The effect of the PERY peptide on downstream targets, apoptosis, and MAPK activation.

(A) Representative blots (left) and quantification (right) of phosphorylation levels of nuclear and cytoplasmic p38 targets of HeLa cells that were serum-starved (0.1% FBS, 16 hours), preincubated with PERY or SCR peptides (10 μM each for 2 hours), and then stimulated with anisomycin (1 μg/ml) for the indicated times (15, 30, or 60 min) or left untreated (−). p, phosphorylated; g, general. Data are means ± SE; *P < 0.05 and **P < 0.01 by one-way ANOVA. (B) Representative blots (left) and quantification (right) of cleaved caspase 3 (cl-Casp3) of PC3 or αT3-1 cells that were serum-starved (0.1% FBS, 16 hours), pretreated with PERY or SCR peptides (10 μM for 2 hours), and stimulated with TPA (250 nM, +) or left untreated (NT, −) for 48 hours. Data are means ± SE; *P < 0.05 by one-way ANOVA. (C) Representative blots (left) and quantification (right) of MAPK and AKT phosphorylation levels of HeLa cells that were serum-starved (0.1% FBS, 16 hours), preincubated with PERY or SCR peptides (10 μM each for 2 hours), and then stimulated with anisomycin (1 μg/ml) and TPA (250 nM) for the indicated times (15, 30, or 60 min) or left untreated (−). Data are means ± SE. All data are representative of at least three independent experiments.

Because the PERY peptide affects the nuclear translocation of p38 and JNK, we then examined whether this inhibition affects JNK-mediated nuclear activities. We have previously shown that stimulation of prostate cancer cells results in a JNK-mediated apoptosis (31), which likely requires both cytoplasmic and nuclear functions of the MAPK (32, 33). We found that the PERY peptide significantly reduced tetradecanoyl phorbol acetate (TPA)–induced apoptosis in αT3-1 cells and, to some extent, also in PC3 cells (Fig. 4B), indicating that the peptide affects JNK-mediated nuclear processes. Next, we examined whether the peptide may have any nonspecific effects on upstream and unrelated signaling pathways. For this purpose, we examined the effect of the peptide on the stress (anisomycin)– and mitogenic (TPA)–induced phosphorylation of ERK, p38, JNK, and AKT in HeLa cells. A significant increase in the phosphorylation of all signaling compounds was observed upon anisomycin stimulation, whereas only some stimulated phosphorylation (especially of ERK1/2) was detected upon TPA treatment (Fig. 4C). Preincubation of the cells with the PERY peptide did not significantly change the stimulated phosphorylation of any of the examined components. Together, our results indicate that the PERY peptide specifically inhibits the nuclear translocation of p38 and JNK, and thereby impairs the phosphorylation and activation of their nuclear targets. Consequently, the peptide can interfere with p38- and JNK-dependent cellular processes without affecting parallel or downstream signaling pathways in the cytoplasm.

The PERY peptide reduces proliferation of breast and melanoma cancer cells

Although p38 is not involved in proliferation in most cell lines, recent evidence demonstrated that p38α/β is involved in the regulation of cell proliferation, differentiation, and migration in several cancer cell lines (24, 25, 34). We were therefore interested in the effect of the PERY peptide on proliferation and viability of various cell lines. As we had previously demonstrated, an inhibitory effect on cell proliferation by the optimal concentration of 10 μM for the EPE peptide (20), we tested this concentration for the PERY peptide as compared to the SCR peptide. The PERY peptide markedly reduced the proliferation of MDA-MB-231 and other breast cancer cell lines (T47D, AU565, and, to a lesser extent, BT549 and HCC70) (Fig. 5A). We also observed a significant inhibition in cell viability for the A2352 melanoma cell line, whereas other melanoma-, pancreatic cancer–, colon cancer–, prostate cancer–, and lung cancer–derived cell lines tested were not affected. We further compared the effect of the PERY peptide to that of the commercial p38 inhibitors SB203580 and PH797804 (35, 36). Overall, we observed a similar effect of the PERY peptide to that of these inhibitors. In the breast cancer and melanoma PERY peptide–responsive cells, we saw a similar reduction in cell proliferation for the PERY peptide–treated cells and for the cells treated with the PH797804 inhibitor. One exception was observed in the MDA-MB-231 cells, in which the PERY peptide completely abolished cell proliferation, whereas the effect of the inhibitor was not so pronounced (Fig. 5, A and B). In HB2 and ASPC1, in which treatment with the p38 inhibitor reduced cell proliferation, there was no effect of the PERY peptide, which might be due to indirect properties of the reagents [such as faster clearance of the drug or peptide as reported for the EPE peptide (37)]. The other cancer cell lines were affected neither by the PERY peptide nor by the inhibitors. Thus, our results demonstrate that the PERY peptide is as effective as and sometimes even better than the commercial p38 inhibitor in inhibiting proliferation of breast and melanoma cancer cells.

Fig. 5 The PERY peptide reduces proliferation of breast and melanoma cancer cells and prevents the growth of human breast cancer xenografts.

(A) Proliferation of various cancer cell lines in the presence of PERY peptide, SCR peptides, or p38 inhibitor. Twenty-two cancer cell lines were treated with either SCR peptide, PERY peptide, SB203580 inhibitor (MDA-MB-231, AU565, and A2185), or PH797804 inhibitor (p38 inhibitor, all other cell lines), all at a concentration of 10 μM. Viable cells were quantified as the fold change of the initial cell number by methylene blue at 72 hours (BT549, HCC70, MIAPaca, ASPC1, PC3, and LNCaP) or 96 hours (all other cell lines) after cell seeding. Data are means ± SEM. *P < 0.05 and **P < 0.01 by two-way ANOVA followed by Dunnett’s post-tests. (B) Representatives time course of cell proliferation of MDA-MB-231, A2352, T47D, and MCF7 cells that were treated with the PERY peptide, SCR peptide, or p38 commercial inhibitor (SB203580 or PH797804) all at a concentration of 10 μM, and their proliferation was compared to DMSO control (0.1%) or no treatment at 1% fetal calf serum (FCS). Quantification of viable cells was detected as above, and the graphs present the kinetic of cell growth at the indicated times. All experiments were repeated three times in triplicate. The results were presented as a fold change of the initial cell number obtained from three independent experiments and represent means ± SEM. *P < 0.05 and **P < 0.01 by two-way ANOVA followed by Dunnett’s post-tests. (C) Growth of human breast cancer xenografts of severe combined immunodeficient mice that were inoculated subcutaneously with MDA-MB-231 cells. Upon establishment of tumors, mice were treated intravenously with PERY or SCR peptides (15 mg/kg) or by gavage administration with PH797804 (10 mg/kg), three times a week. Tumor size was recorded at the same time using a caliper, and the volumes were calculated accordingly. Data are means ± SEM, and experiments were reproduced two times; n = 7 mice per group. **P < 0.01 by one-way ANOVA followed by Tukey post-tests.

To verify that the effect of the PERY peptide on proliferation is related to the inhibition of nuclear translocation of p38, we followed its effect on two responsive cells (MDA-MB-231 and A2352) as compared to two unresponsive cells (MCF7 and HCT116). We first determined the effect on the nuclear translocation of endogenous p38α in these cell lines using fluorescent staining. The PERY peptide inhibited the stimulation-dependent nuclear translocation of p38α as compared to cells treated with the SCR peptide in all cells tested (fig. S3, A and B). We then determined whether the peptide has any nonspecific effects on upstream and unrelated signaling pathways. For this purpose, we examined the effect of the peptide on stress (anisomycin)–induced phosphorylation of p38, JNK, and ERK in MDA-MB-231, MCF7, and HCT116 cells. As expected, a significant elevation in the phosphorylation of p38 and JNK was observed upon anisomycin stimulation, whereas elevation in ERK activity was less pronounced (fig. S4, A and B), and this occurred regardless of the PERY or SCR peptides. Together, our results indicate that the lack of effect on proliferation is not due to a lack of effect on nuclear p38 translocation but is likely to be due to the intrinsic property of p38 activity that does not regulate proliferation in most cells (8). A similar effect was also found with the EPE peptide that inhibited the nuclear translocation of ERK in almost all cells, but its effects on cellular proliferation varied dramatically between distinct lines (20).

We further challenged the effect of the peptide on the growth of human breast cancer tumors in a xenograft model. Thus, we inoculated MDA-MB-231 cells into CD-1 nude mice and allowed tumor xenografts to form and reach the size of ~50 to 100 mm3. Treatments were initiated either by a systemic administration into the tail vein of the mice (PERY and SCR peptides groups, 15 mg/kg) three times a week for a period of a month or by gavage (PH797804 group, 10 mg/kg). We found that treatment with the PERY peptide significantly reduced the breast cancer xenograft growth as compared to treatment with the SCR peptide (Fig. 5C). A similar inhibition of tumor growth was observed for mice that were treated with the PH797804 commercial inhibitor. We did not observe any signs of discomfort, change of weight, other toxicity-related side effects, or changes in size or structure of internal organs in response to the treatments. The significant effects of the PERY peptide on cell proliferation and tumor xenograft model support the potential use of nuclear translocation inhibition of p38 and JNK as a treatment for certain cancers.

The PERY peptide protects mice from DSS-induced colitis

The kinase p38 plays a role in the regulation of cellular responses to stress, including inflammation and tissue homeostasis (38). Several studies have demonstrated the involvement of p38α/β (21, 22, 24) and JNK1/2 (39) cascades in the induction and progression of inflammation-related diseases. It was also shown that these effects are induced, at least in part, at the nucleus (40). We therefore aimed to determine the effect of the PERY peptide on the DSS-induced colitis inflammation model. For that purpose, C57BL mice were treated with the PERY or SCR peptides (15 mg/kg) every other day during the 7-day period of DSS administration. Ten days after DSS administration, mice treated with the SCR peptide or DMSO lost significantly more weight (25%) as compared to the PERY peptide–treated group (10%, Fig. 6A). This suggested that the PERY peptide–treated group had reduced inflammation and intestinal damage because body weight loss reflects the severity of the DSS-induced colitis. Endoscopy analysis performed on day 8 after DSS administration revealed extensively inflamed colons on the SCR peptide– and DMSO-treated colons (Fig. 6B). This was apparent by granular mucosa colons accompanied by the presence of fibrin and diarrhea. On the contrary, the PERY peptide–treated colons showed transparent healthier colons (Fig. 6B). Further histological analysis of these colons revealed a similar observation. Although nontreated colons had highly ordered crypts with no evidence of inflammation in the mucosa or submucosa regions, the SCR peptide– and DMSO-treated colons were highly inflamed, accompanied by massive areas of complete crypt loss and erosions along the colon (Fig. 6C). The PERY peptide–treated group showed mostly ordered colons with normal crypts, although evidence of erosion could be seen, mostly in the distal part of the colon where the largest amount of DSS damage is generally observed. Nonetheless, we observed that even in these regions, the PERY peptide–treated group presented normal crypt histology, indicating that the damage there was relatively minimal (Fig. 6C). Quantification of colon inflammation using histological grading of the colon showed a significantly lower severity inflammation score for the PERY peptide–treated group as compared to SCR peptide–treated mice (Fig. 6D). Finally, highly inflamed colons tend to be shorter in length because of changes in the crypt organization (41). When we compared the colons of the treated mice, which were removed from a precise region, we observed that the SCR peptide– and DMSO-treated colons were significantly shorter than the PERY peptide–treated colons (Fig. 6E). Together, these results indicate that the PERY peptide exhibits protective features against DSS-induced colitis and, therefore, may be a good tool in the prevention of inflammation-related diseases.

Fig. 6 The PERY peptide protects mice from DSS-induced colitis.

C57BL/6 male mice (8 weeks, Harlan) were pretreated by intravenous injections with PERY or SCR peptides (15 mg/kg) or DMSO. The following day, mice were treated with 1.5% dextran sodium sulfate (DSS) in drinking water for 7 days followed by 3 days of normal drinking water. During DSS administration, mice were treated every other day (total of five intravenous injections). Control groups included nontreated mice (no DSS, no treatment) and mice receiving PERY peptide with normal drinking water (PERY only, no DSS). Mice were then euthanized and analyzed as described in Materials and Methods. Colitis was evaluated using (A) weight loss measurements, (B) in vivo endoscopy, and (C) hematoxylin and eosin (H&E) histological sections of the colons (top) and their magnifications (bottom). Scale bar, 2 mm. (D) Inflammation score based on the histological sections. (E) Measurements of colon length. n = 8 mice per group for all groups. The experiment was repeated three times. Data are means ± SEM; *P < 0.05 and **P < 0.01 by one-way ANOVA followed by Tukey post-tests.

The PERY peptide reduces the incidence of AOM/DSS-induced colon cancer

It is well established that colorectal cancer is often associated with chronic inflammatory bowel disease (4244). Therefore, using a protocol that combines the carcinogen AOM with the DSS-induced colitis [AOM/DSS cancer model (45, 46)], we next investigated whether the PERY peptide can be used to prevent inflammation-induced cancer (fig. S5). Initially, we aimed to determine whether the PERY peptide has any effect in this AOM/DSS-induced cancer model and, if so, in which concentration. Therefore, we treated ICR mice with various doses of the PERY peptide (1.5, 5, and 15 mg/kg) and compared their effects to those of the SCR peptide or DMSO treatments. Mice treated with the PERY peptide (5 or 15 mg/kg) demonstrated a better overall well-being response compared to mice treated with a lower dose of the PERY peptide, SCR peptide, or DMSO control (fig. S6A). Endoscopy visual analysis (fig. S6B) along with colon index severity grading during the endoscopy revealed a significantly better colon state for these PERY peptide–treated mice. This was apparent by a lower endoscopic score (fig. S6C) and very few small tumors (fig. S6D) in the PERY peptide–treated mice as compared to the SCR peptide or DMSO control. Histological analysis demonstrated the appearance of large massive tumors in the SCR peptide and DMSO control colons and, to some extent, in the colons of animals treated with the PERY peptide (1.5 mg/kg). In addition, a large percentage of these colons were highly inflamed with complete disruption of the crypts and loss of the surface endothelial cells (fig. S6E). On the other hand, histological analysis of the PERY peptide–treated colons demonstrated highly ordered crypts with only minor inflamed sections mainly in the distal area. The EPE peptide that prevents the nuclear translocation of ERK1/2 (20) had no effect on any of the parameters tested, indicating that the tumor formation upon inflammation is a p38- or JNK-dependent, but not an ERK-dependent, process.

Next, we further explored the effect of the PERY peptide in the AOM/DSS-induced cancer model using a slightly different experimental design. In this experiment, we used the PERY peptide at 15 mg/kg, because that dose was the most effective at reducing tumor load (fig. S6D). In addition, we compared the PERY peptide effect with that of the commercial p38 inhibitor SB203580 that was previously used to improve the response in DSS-induced colitis model (47). We also examined the effect of the PERY peptide by itself (PERY only, no AOM, no DSS) to exclude the possibility that the peptide can cause any side effects such as weight gain. Because our experiment (in fig. S6) showed a prophylaxis potential when administered from the first DSS cycle, we also determined its effect when administered after symptoms have already developed, at the beginning of the second DSS cycle [herein referred to as “PERY(2nd cycle)”]. Notably, mice treated with the PERY peptide or the commercial SB203580 inhibitor did not show any significant weight change along the course of the DSS treatment as compared to nontreated mice or mice treated with the PERY peptide only (Fig. 7A). This was unlike the effect of the SCR peptide that showed significant weight loss along the course of the experiment at each of the DSS cycles as compared to the PERY peptide–treated group. This was similar to the weight loss observed in the group that received the PERY peptide only at the beginning of the second DSS cycle (Fig. 7A).

Fig. 7 PERY peptide reduces the incidence of AOM/DSS-induced colon cancer.

Ten-week-old ICR mice (Harlan) were injected once intraperitoneally with AOM (10 mg/kg) followed by DSS administration as described in Materials and Methods. Mice were treated with PERY or SCR peptides (15 mg/kg intravenously for each group) or with the commercial p38 inflammatory inhibitor SB203580 (15 mg/kg, gavage). Additional group was treated with the PERY peptide (15 mg/kg intravenously), but starting only from the second DSS cycle [PERY(2nd cycle)]. The abovementioned groups received AOM/DSS treatment (A/D). Control groups included nontreated mice (no DSS, no treatment) and mice receiving PERY peptide with normal drinking water (PERY only, no DSS). Mice were then euthanized and analyzed as described in Materials and Methods. Colitis and cancer severity were evaluated using (A) weight loss measurements and (B) in vivo endoscopy and (C) by determining tumor load by adding up the average diameters of all tumors for each animal. (D) Counting the average tumor number in each treatment based on tumor size. (E) H&E histological section analysis of the entire colons’ field (×1), magnification of the squared field in the upper panel (×4), or other colons’ magnifications (×10). Scale bar, 2 mm. n = 10 per group. The results were reproduced twice and are means ± SEM; *P < 0.05 and **P < 0.01 by one-way ANOVA followed by Tukey post-tests.

We then proceeded to evaluate the induced tumors in the treated animals. Endoscopic visual analysis 75 days after AOM administration confirmed the presence of tumors in mice in the SCR peptide treatment and PERY(2nd cycle) treatment groups, and surprisingly also in SB203580-treated mice (Fig. 7B). In contrast, PERY peptide–treated mice had only a few small colon tumors. Quantification of the average tumor numbers and size was further carried out on day 85 and revealed a significantly lower tumor load (Fig. 7C) and tumor number (Fig. 7D) in the PERY peptide–treated group as compared to the SCR peptide. Although mice treated with SB203580 and PERY(2nd cycle) had tumors in their colons, they still presented a significant lower overall tumor load as compared to the SCR peptide–treated group (fig. S7). Overall, the most significant tumorigenic colons were observed in the SCR peptide–treated mice. This was apparent by the higher tumor number, higher portions of tumors with large diameters, and thereby the highest tumor load. Histological analysis revealed that tumors in the SCR peptide– and SB203580-treated mice developed mainly in the distal to mid parts (Fig. 7E and fig. S7), highly resembling tumors in human colorectal cancer. The colons of PERY peptide–treated mice displayed relatively normal crypt structure with ordered layers and had very few tumors and inflamed areas (Fig. 7E and fig. S7).

The PERY peptide inhibits the nuclear translocation of p38 in macrophages

To determine whether the beneficial effects of the PERY peptide were due to the prevention of p38 nuclear translocation, we stained these colon sections with a p38α antibody. A specific staining for p38α was observed along the colon sections, mainly in the mucosa layer including the surface epithelium and the epithelium of the colon crypts; however, they were all cytoplasmic in nature for all treatments (Fig. 8A). On the other hand, a significant difference in p38 staining between the PERY peptide– and SCR peptide–treated sections was observed in the lamina propria layer between the crypts. In this area, p38 staining of the SCR peptide–treated section was mainly nuclear, whereas at the same region p38 staining of the PERY peptide–treated colons was mainly cytoplasmic (Fig. 8, A to C). Thus, the stressful DSS treatment stimulated p38 nuclear translocation, which was significantly reduced by administration of the PERY peptide. Because the initial staining of p38 in the nucleus resembled the distribution of macrophages, we undertook to prove this localization by co-staining the colon sections with p38α and a specific macrophage marker, MAC2 (Fig. 8D). Within the co-stained areas of the lamina propria layer, we detected a clear difference between the appearances of p38 staining in the PERY peptide as compared to the SCR peptide–treated colons. The majority of p38 staining within the MAC2-positive cells treated with the PERY peptide was localized mainly in the cytoplasm or all over the cells. On the contrary, the majority of p38 in the SCR peptide–treated MAC2-positive cells was in the nucleus (Fig. 8D). The PERY peptide also inhibited p38α translocation to the nucleus in the anisomycin- or tumor necrosis factor–α (TNFα)–stimulated macrophage cell line J774 (fig. S8). Overall, our data indicate that in DSS-treated mice, the PERY peptide inhibits the nuclear translocation of p38 in macrophages.

Fig. 8 The PERY peptide inhibits the stimulation-dependent nuclear translocation of p38 after DSS administration in AOM/DSS-induced colons.

(A) Representative images of AOM/DSS-induced colon paraffin sections obtained from SCR peptide– or PERY peptide–treated mice and stained for p38 and DAPI. The upper panels (I) show the entire colon field containing the colon crypts (CC) and the lamina propria (LP). The lower panels (II) contain magnification of the lamina propria highlighted in the upper panel by a square. Scale bars, 2 mm (A, B, and D). (B) The same enlarged sections from p38-stained PERY peptide– and SCR peptide–treated mice from (A) where the nucleus is indicated. Arrows are white (PERY) or black (SCR) to emphasize the nature of p38 staining in these sections. (C) Quantification of (A) counting at least three fields with >50 cells per field. Bars represent the average percentage of cells with mostly nuclear (N, gray), all over [nuclear and cytoplasmic (NC); black], or mostly cytoplasmic (C, white) staining. Data are means ± SE; **P < 0.01 by two-sample test for equality of proportions. (D) Representative images of AOM/DSS-induced colon paraffin sections obtained from PERY peptide– or SCR peptide–treated mice and co-stained for p38, MAC2, and Hoechst. The upper panels (I) show the entire colon field containing the colon crypts (CC) and the lamina propria (LP). The lower panels (II) contain magnification of the lamina propria highlighted in the upper panel by a square to enlarge the co-stained area. The results are representative of two repeat experiments.

DISCUSSION

The p38 and JNK MAPKs are often activated by environmental stresses and inflammatory cytokines; therefore, they are also known as stress-activated protein kinases. They have an essential role in stress response and participate in the induction and regulation of inflammation, tissue homeostasis, and other fundamental processes (34). Because they are crucial components in the regulation of these processes, their deregulation is often involved in the induction of inflammatory diseases and cancer (38). It was shown that the p38α/β cascades play a central role in the induction and progression of inflammation-related diseases (48, 49). In addition, p38α and p38β were shown to play important roles in chronic inflammation–related diseases and inflammation-induced cancer (24, 50, 51). Although p38 is usually not involved in the induction of proliferation as known for the ERK cascade, it has been associated with cancer initiation and progression in several cancer types (35). There is evidence to support the involvement of p38α in the regulation of cell proliferation, migration, and invasion of several cancer cell types (24). Thus, several mouse models and patients’ specimen analysis implicated p38 MAPK signaling in breast cancer. For example, a reduced Erbb2-driven breast tumorigenesis was previously demonstrated in Wip1 knockout (KO) mice, which lack the phosphatase that regulates p38 activity, and was correlated with higher p38 MAPK activation (52). Contrary to this tumor-suppressive role of p38α, the LY2228820, a potent and selective adenosine 5′-triphosphate (ATP)–competitive inhibitor of p38α, significantly reduced tumor growth in several in vivo xenografts models, among which are breast cancer, melanoma, non–small cell lung cancer, and ovarian cancer (25). Moreover, tumor samples from patients diagnosed with invasive breast cancer demonstrated positive phosphorylated p38 expression levels, which correlated with clinical pathological markers (53). Here, we show that the nuclear translocation of p38 plays a role in the induction of some, but not all, of these cancers. Because the peptide exhibited no toxic effect, we suggest that inhibition of this process might serve as a beneficial treatment of these cancers along with fewer side effects than the currently used p38 inhibitors.

The link between cancer and inflammation was made following the initial observations that tumors often arose at the site of chronic inflammation and that biopsy samples from patients showed significant levels of inflammatory cells (54). Additional evidence from a wide variety of epidemiological studies, molecular studies, and genetically modified mice presented findings that led to the general link between inflammation and cancer (55). Because p38 was shown to be a major contributor to chronic inflammation, several recent studies have demonstrated its involvement in inflammation-associated colon cancer (21). In particular, it was shown that p38α plays a role in colitis-associated tumor formation by preventing colon epithelial damage and inflammation (56). However, once tumors are formed, p38α supported tumorigenesis by elevating proliferation rates and inhibiting apoptosis, indicating a dual role of p38α in this process. However, this effect is not limited to p38α, as other p38s, as well as JNK isoforms, have been implicated in either inflammation or inflammation-induced tumorigenesis (57). In particular, recent evidence indicates that in mouse colon tissues, intestinal epithelial cell–specific KO of p38γ attenuated inflammatory responses, decreased proinflammatory cytokine expression, and inhibited colitis-associated colon tumorigenesis (23). On a similar note, del Reino et al. (22) demonstrated the role of p38γ and p38δ in colon cancer associated with colitis using the AOM/DSS model. They showed that deficiency of p38γ/p38δ in mice significantly decreased tumor formation, proinflammatory cytokine, and chemokine production. In all these studies, the effect was associated with immune cells such as macrophages—an observation that was demonstrated in our study as well.

Because p38 MAPK plays important roles in cellular responses to environmental stress, much effort was invested to develop p38 MAPK inhibitors for the treatment of inflammatory diseases. Several inhibitors of p38α/β have been developed and tested in clinical trials (50), but the relatively high toxicity of the drugs excluded their use in any inflammation-related diseases. Therefore, we aimed to develop inhibitors of p38α/β nuclear translocation, containing fewer side effects and less toxicity than the currently developed ones. We were interested in comparing the effect of the PERY peptide to the commercial p38 inhibitor SB203580, which was already shown to have partial beneficial effects when used with the DSS-induced colitis model (47). In our experiment regime, the overall well-being of animals treated with SB203580 was similar to the PERY peptide–treated group. The weight loss was minimal (such as with the PERY peptide–treated group), and just by animal observation during the experimental course, no significant symptoms such as diarrhea or bleeding were observed. Surprisingly, only upon endoscopy and termination of the experiment were substantial differences between the PERY peptide– and the SB203580-treated animals observed; the average tumor number and tumor load were significantly higher in the SB203580-treated animals than in the PERY peptide–treated animals. This was also apparent in the histology analysis, not only by the tumors observed but also by the higher ratio of colons with inflamed regions and complete crypt and surface epithelium loss. A similar phenomenon was observed for the PERY(2nd cycle)–treated group. Both treatment regimens were still significantly better than the SCR peptide, albeit at a lower confidence level (P < 0.05) when compared to the PERY peptide. Here, we demonstrated a therapeutic effect of inhibiting the nuclear translocation of p38 using the PERY peptide that demonstrated a much better effect than the commercial inhibitor SB203580. Therefore, the inhibition of the nuclear translocation should be efficient and have fewer side effects not only for breast cancer or melanoma but also in inflammation-associated cancers.

In summary, we used a similar approach here to the one we previously published for inhibiting the nuclear translocation of ERK to eliminate the growth of mutated BRAF- or RAS-associated cancers (20). Thus, we have demonstrated a therapeutic potential of the PERY peptide, which reduced the growth of breast cancer cells, breast xenograft tumors, and a melanoma cell line. It also significantly impeded inflammation in a colitis model and reduced tumor load in the colitis-induced colon cancer model. The effect on colitis-induced colon cancer was mediated, at least in part, by preventing the nuclear p38 translocation in macrophages. The beneficial effects were at least as good and, under certain conditions, even better than effects obtained by the commercial p38 inhibitors SB203580 and PH797804. Together, the cancer and inflammation models used here support the use of inhibiting the nuclear translocation of MAPKs as a novel drug target for cancer.

MATERIALS AND METHODS

Reagents and antibodies

TPA, anisomycin, polyethylenimine (PEI), DAPI, AOM, and TNFα were obtained from Sigma-Aldrich. Protein A/G beads were purchased from Santa Cruz Biotechnology. DSS (molecular weight, 36,000 to 50,000) was purchased from MP Biomedicals. SB203580 and PH797804 inhibitors were from Selleckchem. Antibodies to Imp3 and Imp7 were from Abnova; antibody to Imp9 was from Novus; antibody to GFP was from Roche Diagnostics GmbH; antibodies to JNK1, JNK2, p38α, p38β, MAC2, and pATF2 were from Santa Cruz Biotechnology; antibodies to doubly phosphorylated ERK1/2 (pTEY-ERK), general ERK (gERK), doubly phosphorylated JNK (pJNK), general JNK1/2 (gJNK), doubly phosphorylated p38 (pp38), general p38 (gp38), phosphorylated AKT (pAKT-S473), and general AKT (gAKT) were from Sigma Israel; antibodies to JNK2, p38β, cleaved caspase 3, and MK2 were from Cell Signaling Technology; and antibodies to general and phosphorylated c-Jun and MEF2A as well as recombinant MEF2A were purchased from Abcam. Secondary antibody conjugates, including light chain–specific secondary antibodies, were purchased from Jackson ImmunoResearch. All antibodies were used according to their manufacturer’s recommendations.

Peptides

The peptides used were as follows: (i) peptide aa21–29, PERYQNLSP; (ii) peptide aa15–29, KTIAEVPERYQNLSP; (iii) peptide aa21–34, KPERYQNLSPVGSGA; (iv) PERY peptide, KPERYQNLSPVAAAA; and (v) SCR peptide, KPARYSANELPQAVA. Each of the peptides was conjugated in its N-terminal to myristic acid and C-terminal amidated. To study the rate of absorption, we used fluorescein isothiocyanate fluorochrome–conjugated peptide. All peptides were purchased from GenScript. The peptides were >85% pure and kept at 100 mM DMSO stock solution at −20°C.

DNA constructs and mutations

GFP-JNK1/2 and p38α/β were cloned in pEGFP-C1 (Clontech). JNK1/2 and p38α/β sequences were amplified from HeLa cell complementary DNA (cDNA) and flanked by Eco RI/Bam HI for JNK2 and p38β, and with Xho I/Bam HI for JNK1 and p38α. GST-JNK1/2 and p38α/β were cloned in pGEX-4T-1 vector (GE Healthcare), flanked by Spe I/Not I restriction sites. Deletion mutation of N or C terminus of GFP-p38α was performed using specific primers to the N-terminal deletion mutation or Δ40 C-terminal deletion mutation. Imp3, Imp7, and Imp9 were cloned in pEGFP-C1. Imp7 and Imp9 were amplified from HeLa cell cDNA using specific primers flanked by Bam HI/Sal I for Imp7 and Xho I/Sal I for Imp9 restriction sites. Imp3 was acquired from Forchheimer repository plasmid collection and amplified using specific primers flanked by Xho I/Eco RI restriction sites. GST-Imp3, GST-Imp9, and His-Imp7 were gifts from O. Livnah (Hebrew University, Jerusalem, Israel).

Cell culture and transfection

HeLa, αT3, MDA-MB-231, MCF7, T47D, HCT116, and J774 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 2 mM l-glutamine, 1% penicillin-streptomycin, and 10% FBS. PC3 and A2352 cells were cultured in RPMI supplemented with 2 mM l-glutamine, 1% penicillin-streptomycin, and 10% FBS. Other cell lines were grown according to the American Type Culture Collection recommendations. Transfections were done in HeLa cells using PEI (Sigma-Aldrich). Briefly, cells were grown to 50% confluence in 10% FBS and were transfected with DNA construct using PEI. Twenty-four hours after transfection, cells were washed and further grown under their corresponding conditions.

IP and CoIP

Cells were grown to 70% confluence, serum-starved (0.1% FBS for 16 hours), and then stimulated or treated. Cell extracts were produced as previously described and incubated for 2 hours (4°C, with rotation) with A/G agarose beads (Santa Cruz Biotechnology) prelinked with specific antibodies (60 min, 23°C). For IP, the bound A/G beads were washed once with radioimmunoprecipitation assay buffer [137 mM NaCl, 20 mM tris (pH 7.4), 10% (v/v) glycerol, 1% Triton X-100, 0.5% (v/v) deoxycholate, 0.1% (w/v) SDS, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 20 mM leupeptin] and then washed twice with 0.5 M LiCl and twice with buffer A. For Co-IP assays, the bound A/G beads were washed three times with ice-cold CoIP washing buffer [20 mM Hepes (pH 7.4), 20 mM MgCl2, 2 mM EGTA, 150 mM NaCl, and 0.1% Triton X-100]. The immunoprecipitated beads were then resuspended with 1.5× sample buffer and boiled; the resolved proteins were analyzed by Western blotting (described below).

In vitro kinase assay

Transfected GFP-tagged wild-type p38α, ΔN20, or ΔN30 N-terminal deletion mutants from treated or nontreated cells were immunoprecipitated by protein A/G beads and were used as kinases. The beads were suspended in reaction mixture [10 mM MgCl2, 1.5 mM dithiothreitol (DTT), 25 mM β-glycerophosphate (pH 7.3), 0.05 mM sodium vanadate, 1.25 mM EGTA, 10 μM calmidazolium, and bovine serum albumin (BSA; 0.83 mg/ml)] and 100 μM ATP. Recombinant MEF2A (0.7 μg) was added to the reaction at a final volume of 30 μl and incubated for 20 min at 30°C with shaking. The reaction was terminated by adding sample buffer, and the phosphorylated proteins were resolved on SDS–polyacrylamide gel electrophoresis and subjected to Western blot analysis with the indicated antibodies. In another experimental setting, the activity of wild-type p38α was detected in the presence of the PERY or SCR peptides (10 μM) under the same conditions.

Cell extraction and Western blotting

Cells were grown to 70% confluence and serum-starved (0.1% FBS for 16 hours). After treatments, cells were rinsed twice with ice-cold phosphate-buffered saline (PBS) and twice with buffer A [50 mM β-glycerophosphate (pH 7.3), 1.5 mM EGTA, 1 mM EDTA, 1 mM DTT, and 0.1 mM sodium vanadate]. The cells were then scraped into buffer H [50 mM β-glycerophosphate (pH 7.3), 1.5 mM EGTA, 1 mM EDTA, 1 mM DTT, 0.1 mM sodium vanadate, aprotinin (10 μg/ml), leupeptin (10 μg/ml), and pepstatin A (2 μg/ml)], sonicated (50 W, 2 × 7 s), and centrifuged (20,000g, 15 min, 4°C), and the supernatants were further analyzed by Western blotting. The blots were then incubated with horseradish peroxidase–conjugated antibodies to mouse or rabbit immunoglobulin and were developed with chemiluminescence (ECL).

Subcellular fractionation

Subcellular fractionation was performed essentially as described previously (11). Briefly, harvested cells were suspended in 200 μl of buffer H containing 1% NP-40. The lysates were mixed vigorously and centrifuged immediately to yield supernatants containing the cytoplasmic fraction. Nuclear proteins were extracted by resuspending the nuclear pellets in 200 μl of extraction buffer, holding on ice (5 min), briefly sonicating (twice for 5 s; 40 W, 4°C), vigorously mixing, and then spinning down (15,000g for 15 min). Both cytoplasmic and nuclear fractions were subjected to Western blot analysis with the indicated antibodies.

Immunofluorescence microscopy

Cells were fixed in 3% paraformaldehyde in PBS (20 min, 23°C) and incubated with 2% BSA in PBS (15 min, 23°C), followed by permeabilization with Triton X-100 (0.1% in PBS, 5 min, 23°C). The fixed cells were then incubated with the primary antibodies (60 min, 23°C), washed three times with PBS, and incubated with rhodamine-conjugated secondary antibody (60 min, 23°C) and DAPI. Slides were visualized by using either a fluorescence microscope (×40 magnification; Olympus BX51) or a spinning disc confocal microscope (×60 magnification; Cell Observer SD; Zeiss). Background correction and contrast adjustment of raw data images were performed using Photoshop (Adobe).

For the mice colon sections, paraffin sections were deparaffinized and rehydrated. Antigen retrieval was performed in tris-EDTA (pH 9) for p38α stain or in citric acid (pH 6) for MAC2 or co-stain. After preincubation with 20% normal horse serum and 0.2% Triton X-100, slides were incubated with rabbit anti-p38α (1:100) and/or rat anti-MAC2 (1:100) antibodies at 4°C overnight. To enhance the signal, secondary antibodies, biotinylated anti-rabbit or anti-rat (1:100, Jackson ImmunoResearch), were added for 90 min, followed by Cy2- or Cy3-conjugated streptavidin (1:200, Jackson ImmunoResearch). Sections were counterstained with Hoechst 33,258 (Molecular Probes) for nuclear labeling.

Proximity ligation assay

Protein-protein interactions were detected by using a Duolink PLA kit (Olink Bioscience) (58) according to the manufacturer’s protocol. Briefly, after fixing and permeabilization, cells were incubated with antibodies to Imp7 and p38α (1 hour, 23°C), washed [0.01 M tris-HCl (pH 7.4), 0.15 M NaCl, and 0.05% Tween 20], and then incubated with specific probes (1 hour, 37°C), followed by DAPI staining to visualize nuclei, followed by a wash [0.2 M tris-HCl (pH 7.5), 0.15 M NaCl]. The signal was visualized using spinning disc confocal microscopy. The number of PLA events was counted automatically in ImageJ, using the “analyze particles” feature. Each field was counted for nuclei and the number of PLA events. Then, the average number of events was calculated per nucleus. More than 100 cells were counted per treatment. Background correction, contrast adjustment, and quantification were performed using Photoshop and ImageJ software.

Viability and proliferation assay

Cells were seeded onto 12- or 24-well plates in their appropriate 10% FCS–containing medium and incubated overnight to adhere. The following day, medium was replaced to the appropriate 1% FCS containing the desired treatment(s). These included 1% FCS, DMSO, SCR peptide, PERY peptide, SB203580, or PH797804 commercial inhibitors, all at the final concentration of 10 μM. Fresh medium containing the same reagents was replaced every day. The number of viable cells was measured by methylene blue assay at 72 or 96 hours after cell seeding. Shortly, cells were fixed with 4% formaldehyde for 2 hours at 23°C, washed once with 0.1 M borate buffer (pH 8.5), and stained with 1% methylene blue in 0.1 M borate buffer. Accessed color was extensively washed, and stain was extracted by adding 0.1 M HCl and examined at 595 nm. For time-course experiments, viable cells were measured at 0, 24, 48, 72, and 96 hours after cell seeding.

Breast cancer xenograft model

All animal experiments were approved by the Animal Care and Use Committee of the Weizmann Institute of Science. Female CD-1 nude mice (Harlan), 6 weeks old, were inoculated subcutaneously into the flank region with 4 × 106 MDA-MB-231 cells in 150 μl of PBS. Tumors were allowed to develop to the size of ~50 to 100 mm3, and then the animals were randomly allocated to different treatment groups. The SCR and PERY peptides were administered by intravenous injection into the tail vein (15 mg/kg, dissolved in PBS, 100 μl per mouse, three times a week). The p38 commercial inhibitor PH797804 was dissolved in 0.5% methylcellulose + 0.025% Tween 20 in PBS and administered by gavage (10 mg/kg) at the same time schedule. Tumor dimensions were measured with a caliper, from which tumor volume was calculated. On day 32 after cell inoculation, mice were sacrificed and tumors were excised and fixed in 4% formaldehyde.

DSS-induced colitis model

Eight-week-old C57BL mice (Harlan) were pretreated with the PERY peptide or SCR control peptide (15 mg/kg for each group, n = 8 mice). Control group was treated with DMSO (n = 8). Treatments were administered by intravenous injection into the tail vein. The next day, mice were provided drinking water containing 1.5% DSS for 7 consecutive days and then normal drinking water for an additional 3 days. During the course of DSS administration, mice were given the PERY or SCR peptide treatment every other day for a total of five intravenous injections, of which one injection was given the day before DSS administration, and the following four injections were given every other day thereafter. Additional control groups were untreated mice (no DSS, no peptide) as well as mice treated with the PERY peptide but with normal drinking water for the entire experiment (PERY only, no DSS). Mice were monitored every other day for body weight, diarrhea, and bleeding. At the end of the experiment, we performed endoscopy imaging of the colons, during which endoscopic score was given to each colon (59). Mice were then sacrificed, the colons were consistently removed at a defined area (from the rectum to the cecum), and their lengths were measured with a ruler. Colons were washed with ice-cold PBS, opened longitudinally, rolled like a Swiss roll, fixed with 4% paraformaldehyde for 48 hours, and embedded in paraffin blocks. Then, 5-μm paraffin-embedded tissue slides were stained with H&E and examined by a light microscope to generate an inflammation severity score as described previously (60). Briefly, a score was given to each colon from a scale of 0 to 12, grading the percentage of inflammatory area, amount of inflammation (severity), depth of inflammation (the layer of the colon that was affected), and degree of ulceration/regeneration. The higher the score, the more inflamed was the colon. A blinded inspection of the tissues was performed by an expert pathologist from the Weizmann Institute of Science.

AOM/DSS-induced colorectal cancer model

Colorectal cancer associated with colitis was induced using the AOM/DSS model. Ten-week-old ICR mice (Harlan) were injected once intraperitoneally with AOM (10 mg/kg). Two days after injection, mice were provided drinking water containing 1.5% DSS for 8 consecutive days and then normal drinking water for 14 days. This DSS treatment was repeated for two additional cycles. During the course of the experiment, mice were treated with the PERY peptide or SCR control peptide (15 mg/kg for each group). Another group was treated with the commercial p38 inhibitor SB203580 (15 mg/kg), and an additional group was treated with the PERY peptide (15 mg/kg) but starting only from the second DSS cycle [PERY(2nd cycle), n = 10 for each treatment]. The treatment was given at each course of the DSS administration in a total of five intravenous injections, of which one injection was given a day before DSS administration, and the following four injections were given every other day thereafter. Mice were monitored for body weight, diarrhea, and bleeding as well as endoscopy imaging during the course of the experiment. On day 85, mice were sacrificed and the colons were removed and further analyzed as described above.

Statistical analysis

Data are presented as means ± SE or means ± SEM. Western blotting data were analyzed by a two-way ANOVA if the batch effect was significant, or a one-way ANOVA otherwise. Immunohistochemistry staining data were analyzed by two-sample test for equality of proportions for the nuclear fraction only. PLA data were analyzed with paired t tests. Data from in vivo experiments were analyzed by one-way ANOVA followed by post hoc Tukey or Dunnett’s tests. P < 0.05 was considered statistically significant.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/11/525/eaao3428/DC1

Fig. S1. Characterization of the N-terminal p38α domain.

Fig. S2. The intracellular distribution of the PERY peptide and its ability to inhibit the stimulation-dependent nuclear translocation of p38β and JNK2.

Fig. S3. The PERY peptide inhibits the stimulation-dependent nuclear translocation of p38 in several cancer cell lines.

Fig. S4. The PERY peptide does not affect MAPK phosphorylation upon stimulation in several cancer cell lines.

Fig. S5. Experimental design for the AOM/DSS-induced cancer model.

Fig. S6. The PERY peptide reduces the incidence of AOM/DSS-induced colon cancer.

Fig. S7. The PERY peptide reduces tumor load in the AOM/DSS colon cancer model.

Fig. S8. The PERY peptide inhibits the stimulation-dependent nuclear translocation of p38α in macrophages.

REFERENCES AND NOTES

Acknowledgments: We thank N. Nevo, O. Brenner, R. Eilam-Altstadter, and R. Rotkopf (Weizmann Institute of Science, Rehovot, Israel); O. Livnah (Hebrew University, Jerusalem, Israel); and M. Jarchow (Julius-Maximilian University of Wuerzburg, Wuerzburg, Germany) for their help in various ways throughout the study. Funding: This study was supported by grants from Israel Science Foundation and United States - Israel Binational Science Foundation (to R.S.). R.S. is an incumbent of the Yale S. Lewine and Ella Miller Lewine professorial chair for cancer research. Author contributions: G.M.-R. performed the in vitro and in vivo experiments and wrote the article. E.Z. performed some of the in vitro and in vivo experiments and T.H. performed some of the in vitro assays. J.B. supervised the study, and R.S. supervised the study and wrote the article. Competing interests: The authors declare that they have no competing interests. Data and materials availability: G.M.-R., T.H., E.Z., and R.S. have a patent on the use of inhibitory peptides for the treatment of inflammatory diseases (U.S. Patent 9,714268, 2017). All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Navigate This Article