Letter | Published:

p38γ is essential for cell cycle progression and liver tumorigenesis

Abstract

The cell cycle is a tightly regulated process that is controlled by the conserved cyclin-dependent kinase (CDK)–cyclin protein complex1. However, control of the G0-to-G1 transition is not completely understood. Here we demonstrate that p38 MAPK gamma (p38γ) acts as a CDK-like kinase and thus cooperates with CDKs, regulating entry into the cell cycle. p38γ shares high sequence homology, inhibition sensitivity and substrate specificity with CDK family members. In mouse hepatocytes, p38γ induces proliferation after partial hepatectomy by promoting the phosphorylation of retinoblastoma tumour suppressor protein at known CDK target residues. Lack of p38γ or treatment with the p38γ inhibitor pirfenidone protects against the chemically induced formation of liver tumours. Furthermore, biopsies of human hepatocellular carcinoma show high expression of p38γ, suggesting that p38γ could be a therapeutic target in the treatment of this disease.

Access optionsAccess options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Data availability

The datasets supporting the findings of this study are available within the paper and its Supplementary Information. Source Data (gels and graphs) for Figs. 14 and Extended Data Figs. 110 are provided with the online version of the paper. There is no restriction on data availability.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    Malumbres, M. Cyclin-dependent kinases. Genome Biol. 15, 122 (2014).

  2. 2.

    Varjosalo, M. et al. The protein interaction landscape of the human CMGC kinase group. Cell Rep. 3, 1306–1320 (2013).

  3. 3.

    Malumbres, M. & Barbacid, M. Cell cycle, CDKs and cancer: a changing paradigm. Nat. Rev. Cancer 9, 153–166 (2009).

  4. 4.

    Canhoto, A. J., Chestukhin, A., Litovchick, L. & DeCaprio, J. A. Phosphorylation of the retinoblastoma-related protein p130 in growth-arrested cells. Oncogene 19, 5116–5122 (2000).

  5. 5.

    Mayhew, C. N. et al. Liver-specific pRB loss results in ectopic cell cycle entry and aberrant ploidy. Cancer Res. 65, 4568–4577 (2005).

  6. 6.

    Gubern, A. et al. The N-terminal phosphorylation of RB by p38 bypasses its inactivation by CDKs and prevents proliferation in cancer cells. Mol. Cell 64, 25–36 (2016).

  7. 7.

    Sabio, G. et al. p38γ regulates the localisation of SAP97 in the cytoskeleton by modulating its interaction with GKAP. EMBO J. 24, 1134–1145 (2005).

  8. 8.

    González-Terán, B. et al. p38γ and δ promote heart hypertrophy by targeting the mTOR-inhibitory protein DEPTOR for degradation. Nat. Commun. 7, 10477 (2016).

  9. 9.

    González-Terán, B. et al. Eukaryotic elongation factor 2 controls TNF-α translation in LPS-induced hepatitis. J. Clin. Invest. 123, 164–178 (2013).

  10. 10.

    Lundberg, A. S. & Weinberg, R. A. Functional inactivation of the retinoblastoma protein requires sequential modification by at least two distinct cyclin-cdk complexes. Mol. Cell. Biol. 18, 753–761 (1998).

  11. 11.

    Hu, W. et al. Concurrent deletion of cyclin E1 and cyclin-dependent kinase 2 in hepatocytes inhibits DNA replication and liver regeneration in mice. Hepatology 59, 651–660 (2014).

  12. 12.

    Narasimha, A. M. et al. Cyclin D activates the Rb tumor suppressor by mono-phosphorylation. eLife 3, e02872 (2014).

  13. 13.

    Diril, M. K. et al. Cyclin-dependent kinase 1 (Cdk1) is essential for cell division and suppression of DNA re-replication but not for liver regeneration. Proc. Natl Acad. Sci. USA 109, 3826–3831 (2012).

  14. 14.

    Diskin, R., Askari, N., Capone, R., Engelberg, D. & Livnah, O. Active mutants of the human p38α mitogen-activated protein kinase. J. Biol. Chem. 279, 47040–47049 (2004).

  15. 15.

    Giacinti, C. & Giordano, A. RB and cell cycle progression. Oncogene 25, 5220–5227 (2006).

  16. 16.

    Repetto, M. V. et al. CDK and MAPK synergistically regulate signaling dynamics via a shared multi-site phosphorylation region on the scaffold protein Ste5. Mol. Cell 69, 938–952.e6 (2018).

  17. 17.

    Manieri, E. & Sabio, G. Stress kinases in the modulation of metabolism and energy balance. J. Mol. Endocrinol. 55, R11–R22 (2015).

  18. 18.

    Hui, L. et al. p38α suppresses normal and cancer cell proliferation by antagonizing the JNK-c-Jun pathway. Nat. Genet. 39, 741–749 (2007).

  19. 19.

    Matesanz, N. et al. p38α blocks brown adipose tissue thermogenesis through p38δ inhibition. PLoS Biol. 16, e2004455 (2018).

  20. 20.

    González-Terán, B. et al. p38γ and p38δ reprogram liver metabolism by modulating neutrophil infiltration. EMBO J. 35, 536–552 (2016).

  21. 21.

    Postic, C. & Magnuson, M. A. DNA excision in liver by an albumin-Cre transgene occurs progressively with age. Genesis 26, 149–150 (2000).

  22. 22.

    Trakala, M. et al. Functional reprogramming of polyploidization in megakaryocytes. Dev. Cell 32, 155–167 (2015).

  23. 23.

    Cubero, F. J. et al. Haematopoietic cell-derived Jnk1 is crucial for chronic inflammation and carcinogenesis in an experimental model of liver injury. J. Hepatol. 62, 140–149 (2015).

  24. 24.

    Zheng, K., Cubero, F. J. & Nevzorova, Y. A. c-MYC—making liver sick: role of c-MYC in hepatic cell function, homeostasis and disease. Genes (Basel) 8, 123 (2017).

  25. 25.

    Askari, N. et al. Hyperactive variants of p38α induce, whereas hyperactive variants of p38γ suppress, activating protein 1-mediated transcription. J. Biol. Chem. 282, 91–99 (2007).

  26. 26.

    Miao, C. H. et al. Inclusion of the hepatic locus control region, an intron, and untranslated region increases and stabilizes hepatic factor IX gene expression in vivo but not in vitro. Mol. Ther. 1, 522–532, (2000).

  27. 27.

    Case, D. A. et al. AMBER v.16 (University of California, San Francisco, 2017).

  28. 28.

    Hamelberg, D., Mongan, J. & McCammon, J. A. Accelerated molecular dynamics: a promising and efficient simulation method for biomolecules. J. Chem. Phys. 120, 11919–11929 (2004).

  29. 29.

    Hamelberg, D., de Oliveira, C. A. & McCammon, J. A. Sampling of slow diffusive conformational transitions with accelerated molecular dynamics. J. Chem. Phys. 127, 155102 (2007).

Download references

Acknowledgements

We thank S. Bartlett for English editing, D. Engelberg for the constitutively active mutants, the Division of Signal Transduction Therapy for recombinant proteins, and CNIC Advanced Imaging and Vector Units for technical support. G.S. (RYC-2009-04972), F.J.C. (RYC-2014-15242), and Y.A.N. (RYC-2015-17438) are investigators of the Ramón y Cajal Program. E.M. and M.T. were awarded La Caixa fellowships and R.R.-B. was a fellow of the Fundación Ramón Areces-UAM and FPU. B.G.-T. is a fellow of the FPI Severo Ochoa CNIC program (SVP-2013-067639). F.J.C. is a Gilead Liver Research Scholar. This work was funded by grants supported in part by funds from the European Regional Development Fund: the European Union’s Seventh Framework Programme (FP7/2007-2013) ERC 260464, EFSD/Lilly European Diabetes Research Programme Dr Sabio, 2017 Leonardo Grant for Researchers and Cultural Creators, BBVA Foundation (Investigadores-BBVA-2017) IN[17]_BBM_BAS_0066, MINECO-FEDER SAF2016-79126-R, and Comunidad de Madrid IMMUNOTHERCAN-CM S2010/BMD-2326 and B2017/BMD-3733 to G.S.; Juan de la Cierva and MINECO SAF2014-61233-JIN to A.T.-L.; the European Community for MSCA-IF-2014-EF-661160-MetAccembly grant to F.F.; Spanish MINECO CTQ2014-59212-P, European Community for CIG project (PCIG14-GA-2013-630978), and European Research Council (ERC) under the European Union’s Horizon 2020 (ERC-2015-StG-679001-NetMoDEzyme) to S.O.; the German Research Foundation (SFB/TRR57/P04 and DFG NE 2128/2-1) and MINECO SAF2017-87919R to Y.A.N.; EXOHEP-CM S2017/BMD-3727 and the COST Action CA17112, MINECO SAF2016-78711, and the AMMF Cholangiocarcinoma Charity 2018/117 to F.J.C.; MINECO (SAF2015-69920-R co-funded by ERDF-EU), the Consolider-Ingenio 2010 Programme (SAF2014-57791-REDC), Excellence Network CellSYS (BFU2014-52125-REDT), and the iLUNG Programme (B2017/BMD-3884) from the Comunidad de Madrid to M. Malumbres; MINECO SAF2015-67077-R and SAF2017-89901-R to J.B.; MINECO (BIO2015-67580-P), Carlos III Institute of Health-Fondo de Investigación Sanitaria (ProteoRed PRB3, IPT17/0019 - ISCIII-SGEFI/ERDF), Fundación La Marató and ‘La Caixa’ Banking Foundation (HR17-00247) to J.V.; ISCIII and FEDER PI16/01548 and Junta de Castilla y León GRS 1362/A/16 and INT/M/17/17 to M. Marcos; Junta de Castilla y León GRS 1356/A/16 and GRS 1587/A/17 to J.L.-T.; and MCNU (SAF2017-84494-C2-1-R) to J.R.-C. The CNIC is supported by the Ministerio de Ciencia, Innovación y Universidades (MCNU) and the Pro CNIC Foundation, and is a Severo Ochoa Center of Excellence (SEV-2015-0505).

Author information

G.S. conceived and supervised this project. G.S. and A.T.-L. designed and developed the hypothesis. E.M., L.L.-V., M.L. and A.T.-L. performed experiments using DEN, carbon tetrachloride, streptozotocin and dextran sodium sulfate, and A.T.-L. analysed the data. B.G.-T., A.T.-L. and A.M. performed partial hepatectomies. A.M., A.M.S., R.R.-B. and A.T.-L. prepared Fig. 2a. A.T.-L., H.M. and B.C. performed cell experiments. A.P.-C. performed S10b. E.R., A.P.-C. and A.C. carried out immunostaining experiments and A.T.-L. analysed the data. M. Malumbres, M.T., A.M., A.M.S., V.M.-R. and A.T.-L. performed CDK1/2 knockout experiments and immunohistochemistry. M. Marcos, L.H.-C., O.B., J.L.T. and N.M. performed the analysis of human samples. Y.A.N., F.J.C. and R.R.-B. developed genetic HCC models. S.O. and F.F. carried out molecular dynamics simulations. J.A.B. and M.R.-M. generated the AAVs. J.B. performed heuristic three-dimensional analysis. J.A.L. and J.V. performed and analysed the proteomic experiments. J.R.-C. carried out MRI experiments and A.T.-L. analysed the data. D.M.-S. generated the phylogenetic tree. A.T.-L. performed the remainder of the experiments. A.T.-L. and G.S. wrote the manuscript with input from all authors.

Competing interests

The authors declare no competing interests.

Correspondence to Guadalupe Sabio.

Extended data figures and tables

  1. Extended Data Fig. 1 Similarities between p38 MAPK proteins and CDKs.

    a, Phylogenetic tree of murine CMGC group kinases. The CDK family is boxed in orange, and the p38 family in green. b, c, Plot of the distance (in Å) between RO3306 and the ATP-binding site along the 10 aMD simulations of CDK1 and p38γ (b), and p38α and CDK2 (c), together with representative RO3306-bound conformations. Spontaneous binding of RO3306 occurs in 5/10 simulations with p38γ, 2/10 with CDK1, 0/10 with p38α, and 1/10 with CDK2. d, Comparison of spontaneous RO3306-binding events observed in 500 ns of aMD simulation time for p38γ and p38δ. e, Representative binding pathway obtained from aMD simulations of RO3306 with p38γ (top) and CDK1 (bottom). f, p38γ phosphorylation sites detected in Rb by an in vitro kinase assay followed by mass spectrometry.

  2. Extended Data Fig. 2 Mass spectrometric analysis of in vitro phosphorylation of Rb by active p38γ or by CDK2–Cyclin A, and the expression of CDK and cyclin mRNA in AlbCre-p38γ mice.

    a, In an in vitro kinase assay, recombinant human Rb protein (2 μg) was incubated alone or in the presence of p38γ or CDK2–cyclin A kinases (1 μg) and 0.2 mM of cold ATP for 60 min. Interpreted MS/MS spectra demonstrating the phosphorylation of the indicated sites in Rb (in lower-case letters). The table shows the total spectral counts of the peptides where each phosphosite was identified. The data are representative of at least three independent experiments. No phosphopeptides were identified in the negative control without kinase. b, Quantitative PCR performed on liver extracts from AlbCre and AlbCre-p38γ mice. Expression was normalized to Gapdh. Data are shown as mean ± s.e.m. n = 15 mice for Cdk1, Cdk2, Cdk4 and Cdk6, and n = 6 mice for cyclins A1 (Ccna1), D1 (Ccnd1) and E1 (Ccne1). *P < 0.05; **P < 0.01; ***P < 0.001. Comparisons were made by one-way ANOVA coupled to Bonferroni’s post-tests. Source data

  3. Extended Data Fig. 3 Expression of active p38γ in hepatocytes reverts liver proliferation in AlbCre-p38γ mice.

    AlbCre control mice, AlbCre-p38γ mice and AlbCre-p38γ mice infected with AAV expressing active p38γ (AlbCre-p38γ AAVp38γ*) were subjected to 70% PHx or a sham procedure. a, Gained liver mass, liver weight and liver:body mass ratio were measured 15 days after PHx and expressed as mean ± s.e.m. Gained liver mass: AlbCre mice, n = 8; and AlbCre-p38γ mice, n = 5; Liver:body weight: AlbCre mice, n = 9; AlbCre-p38γ mice, n = 5; Liver weight: AlbCre mice, n = 9; AlbCre-p38γ mice, n = 5. Comparisons were performed using a two-sided Student’s t-test; *P < 0.05; **P < 0.01. b, BrdU incorporation quantified by cytometry. Data are mean ± s.e.m. n = 3. Comparisons were performed using a one-way ANOVA coupled to Bonferroni’s post-tests; ***P < 0.001; *P < 0.05; **P < 0.01. c, pRb Ser795 immunostaining in livers 48 h after PHx. Left, representative images. Scale bar, 100 μm. Right, quantification. Data are mean ± s.e.m. n = 5 mice. Comparisons were made by one-way ANOVA coupled to Bonferroni’s post-tests; ***P < 0.001. d, Immunoblot analysis of liver extracts with antibodies against phospho-Rb S807/S811, Rb, p38γ and vinculin (as a loading control). Each lane corresponds to a different mouse. The data are representative of at least three independent experiments. e, Liver:tibia length ratio, expressed as mean ± s.e.m. n = 6 mice. Comparisons were made by one-way ANOVA coupled to Bonferroni’s post-tests; *P < 0.05; **P < 0.01. f, Hepatocyte proliferation analysed by Ki67 immunostaining 48 h after PHx. Left, representative images. Scale bar, 100 μm. Right, quantification of Ki67-positive cells, shown as mean ± s.e.m. Comparisons were performed using a one-way ANOVA coupled to a Kruskal–Wallis post-test. g, Wild-type mice infected with AAV expressing active p38γ (AAVp38γ*) or active p38α (AAVp38α*) were subjected to 70% PHx. Rb phosphorylation at the specified residues was assessed by western blot 48 h after PHx. Anti haemagglutinin (HA)-tag antibody was used as control of liver infection by AAVp38γ* and AAVp38α*; each lane corresponds to a different mouse. The data are representative of at least three independent experiments. h, Left, hepatocyte proliferation 48 h after PHx was studied by Ki67 immunostaining (top) or BrdU incorporation (bottom) in immunohistological liver sections. Scale bar, 100 μm. Right, quantification of Ki67- and BrdU-positive cells, shown as mean ± s.e.m. n = 5 counted areas from AlbCre mice: 0 h, n = 4; 48 h, n = 4; AlbCre-p38γ mice: 0 h, n = 3; 48 h, n = 3; AlbCre-p38γ AAVp38γ* mice: 0 h, n = 3; 48 h, n = 3; n = 5–25 counted areas from wild-type AAVp38γ* mice: n = 4; and wild-type AAVp38α* mice: n = 2. Comparisons were performed using a two-sided Student’s t-test; ***P < 0.001; *P < 0.05; **P < 0.01. Source data

  4. Extended Data Fig. 4 p38δ partially compensates for the lack of p38γ.

    a, Kaplan–Meier analysis of survival in PHx-treated AlbCre and AlbCre-p38γ mice. n = 20 mice per genotype. Mantel–Cox log-rank tests were used. b, Rb phosphorylation in the liver was studied in AlbCre and AlbCre-p38γ mice by western blotting with the indicated antibody 48 h, 60 h and 72 h after PHx; each lane corresponds to a different mouse. The data are representative of at least three independent experiments. c, p38δ expression was studied by qPCR at different time points after PHx and its change in expression was represented. AlbCre mice: n = 7 and AlbCre-p38γ mice: n = 6. Comparisons were performed using a two-sided Student’s t-test. dh, AlbCre control mice and AlbCre-p38γδ mice were subjected to 70% PHx or a sham procedure and were analysed after 48 h, 60 h or 72 h. d, Kaplan–Meier analysis of survival. A Mantel–Cox log-rank test was used. e, Immunoblot analysis of liver extracts from AlbCre-p38γδ mice with antibodies against phospho-Rb S807/S811, Rb, p38γ and vinculin (loading control); each lane corresponds to a different mouse. The data are representative of at least three independent experiments. f, g, Hepatocyte proliferation was analysed by BrdU incorporation (f; two-sided Student’s t-test; **P < 0.05) and Ki67 immunostaining (g; two-sided Student’s t-test with Welch’s correction; *P < 0.01) 48 h after PHx in AlbCre-p38γ (n = 3) and AlbCre-p38γδ (n = 3) mice. h, Hepatocyte proliferation was analysed by BrdU incorporation 60 h and 72 h after PHx in AlbCre mice (60 h, n = 9; 72 h, n = 7), AlbCre-p38γ mice (60 h and 72 h, n = 6) and AlbCre-p38γδ mice (60 h and 72 h, n = 3). Comparisons were performed using a one-way ANOVA coupled to Bonferroni’s post-tests; ***P < 0.001. Left, representative images. Scale bar, 50 μm. Right, quantification of Ki67- and BrdU-positive cells. Data are mean ± s.e.m. n = 5 counted areas from the specified number of mice. Source data

  5. Extended Data Fig. 5 Reduced epithelial proliferation and Rb phosphorylation in the absence of p38γ after treatment with dextran sodium sulfate.

    Wild-type and p38γ knockout mice were treated for 6 days with dextran sodium sulfate (DSS) administered in drinking water. a, Representative images showing the shortening of the colon after DSS treatment. b, Immunohistochemical staining (left) and BrdU quantification (right) in colon tissue sections of DSS-treated wild-type and p38γ knockout mice. c, Immunohistochemical staining (left) and phospho-Rb S795 quantification (right). Quantification is shown as mean ± s.e.m. n = 5–10 fields from wild-type control (H2O) mice: n = 5; wild-type DSS-treated mice: n = 7; p38γ KO control mice: n = 5; p38γ KO DSS-treated mice: n = 9. In b, c, comparisons were made using a one-way ANOVA coupled with Bonferroni’s multiple comparison test; **P < 0.01; ***P < 0.001. Scale bar, 100 μm. d, Immunoblot analysis of Rb phosphorylation in the intestine, detected with the indicated antibody; each lane corresponds to a different mouse. The data are representative of at least three independent experiments. Source data

  6. Extended Data Fig. 6 p38γ and CDKs cooperate in the induction of Rb phosphorylation and liver proliferation.

    a, In vitro kinase assay, in which phosphorylation sites identified by mass spectrometry are underlined. b, Immunoblot analysis of CDK2 in p38γ immunoprecipitates from the liver of AlbCre-p38γ mice with or without infection with AAV expressing active p38γ (AAVp38γ*). c, Immunoblot analysis of Rb expression in the liver in wild-type and CDK1/2 KO mice (AAV2/8-Cre-infected) in steady state. d, BrdU immunostaining analysis after PHx. Quantification is shown as mean ± s.e.m. n = 5 fields: 0 h, n = 4; 2 h, n = 5; 8 h, n = 5; 12 h, n = 5; 24 h, n = 5; 36 h, n = 7; 48 h, n = 14; 60 h, n = 6; 72 h, n = 4 mice. Comparison was performed using a one-way ANOVA coupled with Bonferroni’s multiple comparison test. e, Immunoblot analysis in livers from AlbCre and AlbCre-p38γ mice. f, Immunoprecipitation–immunoblot analysis of the interaction of CDK2 with wild-type and nonphosphorylatable Rb in HEK-293T cells transfected with human HA-Rb wild-type or HA-Rb ΔCDK (nonphosphorylatable by CDKs). g–m, Wild-type mice were injected with lentivirus containing shScramble control or short hairpin RNA (shRNA) targeting CDK1/2 (shCDK1/2) or CDK4/6 (shCDK4/6) with or without AAV expressing active p38γ (AAVp38γ*). Mice were subjected to PHx or to a sham procedure. g, k, l, Immunoblot analysis. Hepatocyte proliferation was analysed by Ki67 immunostaining 48 h after PHx. Scale bars, 50 μm. h, i, m, Hepatocyte proliferation was analysed 48 h after PHx. h, Ki67 immunostaining. Scale bar, 100 μm. i, Ki67-positive cell quantification is shown as mean ± s.e.m. n = 5–10 counted areas from wild-type mice: 0 h, n = 2; 48 h, n = 3; shCDK1/2 mice: 0 h, n = 3; 48 h, n = 4; shCDK1/2 mice: 0 h, n = 4; 48 h, n = 4; AAVp38γ* mice: 0 h, n = 4; 48 h, n = 5. One-way ANOVA coupled with Bonferroni’s multiple comparison test; ***P < 0.001. m, n = 5 counted areas from n = 5 mice). Scale bar, 100 μm. Comparisons were made by two-sided Student’s t-test; ***P < 0.001. In the western blots, each lane corresponds to a different mouse and is representative of at least three independent experiments. Source data

  7. Extended Data Fig. 7 p38γ expression is increased in human HCC.

    a, Percentage of patients with HCC with mutations in p38γ, CDK1 and CDK2 and the number of HCC mutations in p38γ, CDK1 and CDK2. Data were obtained from the International Cancer Genome Consortium (data from 17 July 2017). b, Expression of p38γ in human primary hepatocytes and in human HCC cell lines (Huh7, HepG2, Snu449 and Snu398) and another type of cancer cells (HTB77). c, Immunoblot analysis of phospho-p38γ and total p38γ in liver extracts from mice lacking IKKγ in the liver (LIKKγKO) and control littermates (WT; left) and from c-Myc transgenic mice (cMYCtg) and wild-type counterparts (right). p38γ phosphorylation was detected only in mice lacking IKKγ specifically in the liver or overexpressing c-Myc. Vinculin served as a loading control. In the western blots each lane corresponds to a different mouse and is representative of at least three independent experiments. d, Immunohistochemical staining of p38γ in human liver with HCC. Negative control, p38γ knockout mice; positive control, p38γ knockout mice infected with human AAVp38γ. The chart shows stratification of p38γ expression in human liver samples as no expression, low, medium, and high expression (n = 46 patients with HCC and n = 11 healthy patients). Source data

  8. Extended Data Fig. 8 p38γ expression is necessary for Snu398 and HepG2 proliferation.

    a, Immunoblot analysis of Snu398 cells treated with lentiviral particles containing two p38γ-targeting shRNAs (B5 or G1) or shScramble. Representative western blot of at least three independent experiments. b, Growth of Snu398 cells infected with shp38γ or shScramble. Cells were plated and cultured for 2 days in medium supplemented with different serum concentrations. Relative cell numbers were measured by crystal violet staining. Data are mean ± s.e.m. (n = 12). Comparisons were made by two-way ANOVA; ***P < 0.001. c, Colony-formation assay of Snu398 cells infected with shp38γ or shScramble. Representative images are shown. d, Cells were grown in DMEM with 10% serum. The number of colonies with >10 cells was counted after 15 days. Data are mean ± s.e.m. (n = 3) and are representative of results from three independent experiments. Comparisons were made by two-sided Student’s t-test; **P < 0.01; ***P < 0.001. e, Immunoblot analysis in HepG2 cells treated with lentiviral particles containing an shRNA against p38γ or Scramble control. Representative western blot of at least three independent experiments. f, Growth of HepG2 cells infected with shp38γ or shScramble. Cells were plated and cultured for 2 days in medium supplemented with different serum concentrations. Relative cell numbers were measured by crystal violet staining. Data are mean ± s.e.m. (n = 3) and are representative of results from three independent experiments. Comparisons were made by two-way ANOVA; ***P < 0.001. g, Left, colony-formation assay of HepG2 cells infected with shp38γ or shScramble. Representative images are shown. Cells were grown in DMEM with 10% serum. Right, the number of colonies with >10 cells was counted after 15 days. Data are mean ± s.e.m. (n = 3). Comparisons were made by two-sided Student’s t-test; ***P < 0.001. h, Soft agar assay of HepG2 cells infected with shp38γ or shScramble. Cells were grown in DMEM with 10% serum. The number of colonies with >10 cells was counted after 20 days. Data are mean ± s.e.m. (n = 12). Comparisons were made by Mann–Whitney U-test; ***P < 0.001. Source data

  9. Extended Data Fig. 9 AlbCre-p38γ mice are protected against carbon tetrachloride-induced liver damage and HCC induced by type 1 diabetes.

    a, Immunoprecipitation–immunoblot analysis of phosphorylated and total p38γ in liver extracts from AlbCre control mice, showing p38γ activation upon acute treatment with DEN (100 mg kg−1 for 2 h). Liver lysates (2 mg) were immunoprecipitated with anti-p38γ antibody followed by immunoblotting as indicated. b, Immunoblot analysis of Rb S807/S811 phosphorylation and proliferating cell nuclear antigen content in livers of AlbCre mice and AlbCre-p38γ mice 1 month after injection of DEN. Vinculin is shown as a loading control. c, Representative images of Ki67 immunohistochemistry in livers of AlbCre and AlbCre-p38γ mice 8 months after injection of DEN. Scale bar, 500 μm. Data are mean ± s.e.m. n = 10 fields in n = 7 mice. Two-sided Student’s t-test; ***P < 0.001. d, e, AlbCre and AlbCre-p38γ mice were injected with 2 ml kg−1 of carbon tetrachloride (v/v) in 20% corn oil, three times per week for 14 weeks. All mice were fed a high-fat diet. d, Representative images of liver tumours (left) and quantification of tumour size (right). Comparisons were made by a two-tailed Student’s t-test with Welch’s correction; ***P < 0.001. e, Immunohistochemical staining and Ki67 quantification of liver tissue sections. Comparisons were performed with a two-sided Student’s t-test; ***P < 0.001. Scale bar, 100 μm. Quantification is shown as mean ± s.e.m. n = 5 fields from n = 9 mice. f, g, Streptozotocin was subcutaneously injected (60 mg g−1) into AlbCre and AlbCre-p38γ mice at P1.5. All mice were fed a high-fat diet and histopathologically assessed at 27 weeks of age. f, Immunohistochemical staining for phosphor-Rb S795 of liver tissue sections. AlbCre mice: n = 5; AlbCre-p38γ mice: n = 5. Comparisons were performed using a two-tailed Student’s t-test with Welch’s correction; ***P < 0.001. Scale bar, 100 μm. g, Ki67 staining on liver tissue sections. AlbCre mice, n = 4; AlbCre-p38γ mice, n = 5. Quantification is shown as mean ± s.e.m. n = 2–6. Scale bar, 100 μm. Comparisons were performed using a two-sided Student’s t-test;**P < 0.01. Source data

  10. Extended Data Fig. 10 p38γ deletion or inhibition protects against DEN-induced HCC.

    a, Top, representative conformations of p38γ and the ATP-binding site of p38γ, both with the inhibitor pirfenidone bound (in purple), extracted from the molecular dynamics simulations. Middle, the activation loop of p38γ is shown in teal and other relevant ATP-binding site residues are shown in light blue. Bottom, plot of distance (in Å) between the oxygen of the pirfenidone carbonyl group and the amide backbone of p38γ Met112 during the molecular dynamics simulations for the five replicas (shown in different colours). Short distances indicate binding of pirfenidone in the ATP-binding site. Spontaneous binding of pirfenidone to p38γ was observed in one out of five simulations. b, Western blot of phosphorylated Rb. 10 μM BIRB796 or pirfenidone were added 30 min before the kinase assay. Representative western blot of at least three independent experiments. c, Wild-type mice were untreated (control) or treated with pirfenidone for 10 weeks and blood concentrations of the following selected parameters were assayed: alanine aminotransferase (ALT), as a readout of hepatic injury (comparisons were made by two-sided Student’s t-test); aspartate aminotransferase (AST), as a readout of hepatic and cardiac injury (comparisons were made by two-sided Student’s t-test); total bilirubin, as a readout of hepatic injury (comparisons were made by Mann–Whitney U-test); alkaline phosphatase, as a readout of hepatic and cardiac injury (comparisons were made by Mann–Whitney U-test); creatine kinase (CK) and creatinine, as a readout of cardiac and renal injury (comparisons were made by Mann–Whitney U-test). All data are mean ± s.d. (n = 10 mice). d, Immunoblot analysis of p38γ in liver and tumour samples after AAV-Cre infection. Representative western blot of at least three independent experiments. e, Number of tumours and tumour size as analysed at the end of the experiment. Data are mean ± s.e.m. n = 10 untreated and n = 20 cre-treated mice. Comparisons were made by two-sided Student’s t-test with Welch’s correction; *P < 0.05; ***P < 0.001. f, Representative contrast-enhanced MRI results from mice 7 months after DEN injection with or without CRE-mediated p38γ deletion. The figure shows axial slices extracted from the 3D volume dataset. Arrowheads mark typical liver tumours. g, AAV-Cre-mediated deletion of p38γ protects against streptozotocin-induced HCC. Left, increased liver damage was found after streptozotocin treatment. Comparisons were made by Student’s t-test; ***P < 0.001. Right, tumour size as analysed at the end of the experiment. Comparisons were made by one-way ANOVA coupled to Bonferroni’s post-test; *P < 0.05. Data are mean ± s.e.m. n = 10 untreated and n = 20 AAV-Cre-treated mice. In the western blots, each lane corresponds to a different mouse. Source data

Supplementary information

  1. Supplementary Information

    This file contains Supplementary Table 1 and 2, Supplementary Video legends and Supplementary Figure 1

  2. Reporting Summary

  3. Video 1

    RO3306 binding to CDK1

  4. Video 2

    RO3306 binding to p38γ

Source data

  1. Source Data Fig. 1

  2. Source Data Fig. 2

  3. Source Data Fig. 3

  4. Source Data Fig. 4

  5. Source Data Extended Data Fig. 2

  6. Source Data Extended Data Fig. 3

  7. Source Data Extended Data Fig. 4

  8. Source Data Extended Data Fig. 5

  9. Source Data Extended Data Fig. 6

  10. Source Data Extended Data Fig. 7

  11. Source Data Extended Data Fig. 8

  12. Source Data Extended Data Fig. 9

  13. Source Data Extended Data Fig. 10

Rights and permissions

To obtain permission to re-use content from this article visit RightsLink.

About this article

Publication history

  • Received

  • Accepted

  • Published

  • Issue Date

DOI

https://doi.org/10.1038/s41586-019-1112-8

Fig. 1: p38γ phosphorylates Rb and promotes liver proliferation after PHx.
Fig. 2: p38γ compensates for the loss of CDK1 or CDK2.
Fig. 3: p38γ drives the development of HCC.
Fig. 4: Pirfenidone inhibits p38γ activity and protects against DEN-induced HCC.
Extended Data Fig. 1: Similarities between p38 MAPK proteins and CDKs.
Extended Data Fig. 2: Mass spectrometric analysis of in vitro phosphorylation of Rb by active p38γ or by CDK2–Cyclin A, and the expression of CDK and cyclin mRNA in AlbCre-p38γ mice.
Extended Data Fig. 3: Expression of active p38γ in hepatocytes reverts liver proliferation in AlbCre-p38γ mice.
Extended Data Fig. 4: p38δ partially compensates for the lack of p38γ.
Extended Data Fig. 5: Reduced epithelial proliferation and Rb phosphorylation in the absence of p38γ after treatment with dextran sodium sulfate.
Extended Data Fig. 6: p38γ and CDKs cooperate in the induction of Rb phosphorylation and liver proliferation.
Extended Data Fig. 7: p38γ expression is increased in human HCC.
Extended Data Fig. 8: p38γ expression is necessary for Snu398 and HepG2 proliferation.
Extended Data Fig. 9: AlbCre-p38γ mice are protected against carbon tetrachloride-induced liver damage and HCC induced by type 1 diabetes.
Extended Data Fig. 10: p38γ deletion or inhibition protects against DEN-induced HCC.

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.