Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Hepatocytes direct the formation of a pro-metastatic niche in the liver

Abstract

The liver is the most common site of metastatic disease1. Although this metastatic tropism may reflect the mechanical trapping of circulating tumour cells, liver metastasis is also dependent, at least in part, on the formation of a ‘pro-metastatic’ niche that supports the spread of tumour cells to the liver2,3. The mechanisms that direct the formation of this niche are poorly understood. Here we show that hepatocytes coordinate myeloid cell accumulation and fibrosis within the liver and, in doing so, increase the susceptibility of the liver to metastatic seeding and outgrowth. During early pancreatic tumorigenesis in mice, hepatocytes show activation of signal transducer and activator of transcription 3 (STAT3) signalling and increased production of serum amyloid A1 and A2 (referred to collectively as SAA). Overexpression of SAA by hepatocytes also occurs in patients with pancreatic and colorectal cancers that have metastasized to the liver, and many patients with locally advanced and metastatic disease show increases in circulating SAA. Activation of STAT3 in hepatocytes and the subsequent production of SAA depend on the release of interleukin 6 (IL-6) into the circulation by non-malignant cells. Genetic ablation or blockade of components of IL-6–STAT3–SAA signalling prevents the establishment of a pro-metastatic niche and inhibits liver metastasis. Our data identify an intercellular network underpinned by hepatocytes that forms the basis of a pro-metastatic niche in the liver, and identify new therapeutic targets.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Fig. 1: Primary PDAC development induces a pro-metastatic niche in the liver.
Fig. 2: IL-6 is necessary for the establishment of a pro-metastatic niche in the liver.
Fig. 3: STAT3 signalling in hepatocytes orchestrates the formation of a pro-metastatic niche in the liver.
Fig. 4: SAA is a critical determinant of liver metastasis.

Data availability

QuantSeq 3′ mRNA sequencing data have been deposited in the Gene Expression Omnibus (GEO) under accession number GSE109480. Source Data are provided for all figures and extended data figures. All data are available from the corresponding author upon reasonable request.

References

  1. 1.

    Hidalgo, M. Pancreatic cancer. N. Engl. J. Med. 362, 1605–1617 (2010).

    CAS  Article  Google Scholar 

  2. 2.

    Becker, A. et al. Extracellular vesicles in cancer: cell-to-cell mediators of metastasis. Cancer Cell 30, 836–848 (2016).

    CAS  Article  Google Scholar 

  3. 3.

    Sceneay, J., Smyth, M. J. & Möller, A. The pre-metastatic niche: finding common ground. Cancer Metastasis Rev. 32, 449–464 (2013).

    CAS  Article  Google Scholar 

  4. 4.

    Hingorani, S. R. et al. Trp53R 172H and Kras G12D cooperate to promote chromosomal instability and widely metastatic pancreatic ductal adenocarcinoma in mice. Cancer Cell 7, 469–483 (2005).

    CAS  Article  Google Scholar 

  5. 5.

    Lee, J. W., Komar, C. A., Bengsch, F., Graham, K. & Beatty, G. L. Genetically engineered mouse models of pancreatic cancer: the KPC model (LSL-Kras G12D/+ ;LSL-Trp53 R172H/+ ;Pdx-1-Cre), its variants, and their application in immuno-oncology drug discovery. Curr. Protoc. Pharmacol. 73, 14.39.1–14.39.20 (2016).

    Article  Google Scholar 

  6. 6.

    Rhim, A. D. et al. EMT and dissemination precede pancreatic tumor formation. Cell 148, 349–361 (2012).

    CAS  Article  Google Scholar 

  7. 7.

    Grünwald, B. et al. Pancreatic premalignant lesions secrete tissue inhibitor of metalloproteinases-1, which activates hepatic stellate cells via CD63 signaling to create a premetastatic niche in the liver. Gastroenterology 151, 1011–1024.e7 (2016).

    Article  Google Scholar 

  8. 8.

    Seubert, B. et al. Tissue inhibitor of metalloproteinases (TIMP)-1 creates a premetastatic niche in the liver through SDF-1/CXCR4-dependent neutrophil recruitment in mice. Hepatology 61, 238–248 (2015).

    CAS  Article  Google Scholar 

  9. 9.

    Hoshino, A. et al. Tumour exosome integrins determine organotropic metastasis. Nature 527, 329–335 (2015).

    ADS  CAS  Article  Google Scholar 

  10. 10.

    Costa-Silva, B. et al. Pancreatic cancer exosomes initiate pre-metastatic niche formation in the liver. Nat. Cell Biol. 17, 816–826 (2015).

    CAS  Article  Google Scholar 

  11. 11.

    Kowanetz, M. et al. Granulocyte-colony stimulating factor promotes lung metastasis through mobilization of Ly6G+Ly6C+ granulocytes. Proc. Natl Acad. Sci. USA 107, 21248–21255 (2010).

    ADS  CAS  Article  Google Scholar 

  12. 12.

    Hiratsuka, S. et al. The S100A8-serum amyloid A3-TLR4 paracrine cascade establishes a pre-metastatic phase. Nat. Cell Biol. 10, 1349–1355 (2008).

    CAS  Article  Google Scholar 

  13. 13.

    Hiratsuka, S., Watanabe, A., Aburatani, H. & Maru, Y. Tumour-mediated upregulation of chemoattractants and recruitment of myeloid cells predetermines lung metastasis. Nat. Cell Biol. 8, 1369–1375 (2006).

    CAS  Article  Google Scholar 

  14. 14.

    Schmidt-Arras, D. & Rose-John, S. IL-6 pathway in the liver: From physiopathology to therapy. J. Hepatol. 64, 1403–1415 (2016).

    CAS  Article  Google Scholar 

  15. 15.

    Long, K. B. et al. IL6 receptor blockade enhances chemotherapy efficacy in pancreatic ductal adenocarcinoma. Mol. Cancer Ther. 16, 1898–1908 (2017).

    CAS  Article  Google Scholar 

  16. 16.

    Lesina, M. et al. Stat3/Socs3 activation by IL-6 transsignaling promotes progression of pancreatic intraepithelial neoplasia and development of pancreatic cancer. Cancer Cell 19, 456–469 (2011).

    CAS  Article  Google Scholar 

  17. 17.

    Fukuda, A. et al. Stat3 and MMP7 contribute to pancreatic ductal adenocarcinoma initiation and progression. Cancer Cell 19, 441–455 (2011).

    CAS  Article  Google Scholar 

  18. 18.

    Corcoran, R. B. et al. STAT3 plays a critical role in KRAS-induced pancreatic tumorigenesis. Cancer Res. 71, 5020–5029 (2011).

    CAS  Article  Google Scholar 

  19. 19.

    Eckhardt, E. R. et al. Intestinal epithelial serum amyloid A modulates bacterial growth in vitro and pro-inflammatory responses in mouse experimental colitis. BMC Gastroenterol. 10, 133 (2010).

    Article  Google Scholar 

  20. 20.

    Urieli-Shoval, S., Linke, R. P. & Matzner, Y. Expression and function of serum amyloid A, a major acute-phase protein, in normal and disease states. Curr. Opin. Hematol. 7, 64–69 (2000).

    CAS  Article  Google Scholar 

  21. 21.

    Notta, F. et al. A renewed model of pancreatic cancer evolution based on genomic rearrangement patterns. Nature 538, 378–382 (2016).

    ADS  CAS  Article  Google Scholar 

  22. 22.

    Yachida, S. et al. Distant metastasis occurs late during the genetic evolution of pancreatic cancer. Nature 467, 1114–1117 (2010).

    ADS  CAS  Article  Google Scholar 

  23. 23.

    Reichert, M. et al. Regulation of epithelial plasticity determines metastatic organotropism in pancreatic cancer. Dev. Cell 45, 696–711.e8 (2018).

    CAS  Article  Google Scholar 

  24. 24.

    de Beer, M. C. et al. Impact of serum amyloid A on high density lipoprotein composition and levels. J. Lipid Res. 51, 3117–3125 (2010).

    Article  Google Scholar 

  25. 25.

    Liang, W. S. et al. Genome-wide characterization of pancreatic adenocarcinoma patients using next generation sequencing. PLoS One 7, e43192 (2012).

    ADS  CAS  Article  Google Scholar 

  26. 26.

    Long, K. B. et al. IFNγ and CCL2 cooperate to redirect tumor-infiltrating monocytes to degrade fibrosis and enhance chemotherapy efficacy in pancreatic carcinoma. Cancer Discov. 6, 400–413 (2016).

    CAS  Article  Google Scholar 

  27. 27.

    Soares, K. C. et al. A preclinical murine model of hepatic metastases. J. Vis. Exp. 91, 51677 (2014).

    Google Scholar 

  28. 28.

    Ma, Y., Gao, M., Sun, H. & Liu, D. Interleukin-6 gene transfer reverses body weight gain and fatty liver in obese mice. Biochim. Biophys. Acta 1852, 1001–1011 (2015).

    CAS  Article  Google Scholar 

  29. 29.

    Tseng, W., Leong, X. & Engleman, E. Orthotopic mouse model of colorectal cancer. J. Vis. Exp. 10, 484 (2007).

    Google Scholar 

  30. 30.

    Klaas, M. et al. The alterations in the extracellular matrix composition guide the repair of damaged liver tissue. Sci. Rep. 6, 27398 (2016).

    ADS  CAS  Article  Google Scholar 

  31. 31.

    Webb, N. R. et al. Deficiency of endogenous acute-phase serum amyloid A protects apoE−/− mice from angiotensin II-induced abdominal aortic aneurysm formation. Arterioscler. Thromb. Vasc. Biol. 35, 1156–1165 (2015).

    CAS  Article  Google Scholar 

  32. 32.

    Sun, Y. et al. Sudan black B reduces autofluorescence in murine renal tissue. Arch. Pathol. Lab. Med. 135, 1335–1342 (2011).

    CAS  Article  Google Scholar 

  33. 33.

    Kalbasi, A. et al. Tumor-derived CCL2 mediates resistance to radiotherapy in pancreatic ductal adenocarcinoma. Clin. Cancer Res. 23, 137–148 (2017).

    CAS  Article  Google Scholar 

  34. 34.

    Untergasser, A. et al. Primer3—new capabilities and interfaces. Nucleic Acid Res. 40, e115 (2012).

    CAS  Article  Google Scholar 

  35. 35.

    Bindea, G. et al. ClueGO: a Cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks. Bioinformatics 25, 1091–1093 (2009).

    CAS  Article  Google Scholar 

  36. 36.

    Bindea, G., Galon, J. & Mlecnik, B. CluePedia Cytoscape plugin: pathway insights using integrated experimental and in silico data. Bioinformatics 29, 661–663 (2013).

    CAS  Article  Google Scholar 

  37. 37.

    Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13, 2498–2504 (2003).

    CAS  Article  Google Scholar 

  38. 38.

    The Gene Ontology Consortium. Expansion of the Gene Ontology knowledgebase and resources. Nucleic Acids Res. 45, D331–D338 (2017).

    Article  Google Scholar 

  39. 39.

    Ashburner, M. et al. Gene ontology: tool for the unification of biology. Nat. Genet. 25, 25–29 (2000).

    CAS  Article  Google Scholar 

  40. 40.

    Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl Acad. Sci. USA 102, 15545–15550 (2005).

    ADS  CAS  Article  Google Scholar 

Download references

Acknowledgements

We thank members of the Zaret laboratory for assistance with isolation and culture of primary hepatocytes and members of the Genomics Facility (Wistar Institute) and the Molecular Pathology and Imaging Core (University of Pennsylvania) for technical support; E. J. Wherry, D. J. Powell, J. R. Conejo-Garcia, I. E. Brodsky, and E. L. Stone for discussions and advice; and the Mayo Clinic Arizona for provision of liver tissue sections collected from patients with pancreatic cancer. This work was supported by National Institutes of Health grants F30 CA196106 (J.W.L.), T32 HL007439 (J.W.L.), T32 CA009140 (M.L.S.), R01 CA197916 (G.L.B.), R01 HL134731 (N.R.W. and F.C.d.B.), the University of Pennsylvania Molecular Pathology and Imaging Core of the Center for Molecular Studies in Digestive and Liver Diseases grant P30 DK050306, the 2015 Pancreatic Cancer Action Network-AACR Career Development Award 15-20-25-BEAT supported by an anonymous foundation (G.L.B.), the 2017 Stand Up to Cancer (SU2C) Innovative Research Grant SU2C-AACR-IRG 13-17 (G.L.B.), a Research Support Grant from the University of Kentucky Office of the Vice President for Research (N.R.W.), the American Surgical Association Foundation Fellowship (P.M.P.), the University of Pennsylvania Pancreatic Cancer Research Center (E.L.C.), and the Abramson Cancer Center Translational Centers of Excellence (E.L.C.).

Reviewer information

Nature thanks Tim Greten, Anirban Maitra, Robert Schwabe and the other anonymous reviewer(s) for their contribution to the peer review of this work.

Author information

Affiliations

Authors

Contributions

Experiments and data analysis were performed by J.W.L., M.L.S., P.M.P., S.K.T., C.A.K., D.D., W.L.G., X.H., A.J., M.C.d.B, F.C.d.B, N.R.W., and G.L.B; generation of mouse pancreatic tumour cell lines by W.L.G. and G.L.B.; immunofluorescence and immunohistochemistry by J.W.L., M.L.S., J.H.L., D.D., W.L.G., and A.J.; RNA in situ hybridization by J.W.L. and X.H.; tumour cell culture by J.W.L., M.L.S., and C.A.K.; animal studies by J.W.L., M.L.S., S.K.T., C.A.K., K.G., and W.L.G.; flow cytometry by J.W.L., M.L.S., and S.K.T.; QuantSeq 3′ mRNA sequencing and data analysis by J.W.L.; quantitative PCR by J.W.L., J.H.L., and X.H.; cytokine bead array by J.W.L., M.L.S., and C.A.K.; primary hepatocyte studies by J.W.L.; enzyme-linked immunosorbent assays by J.W.L. and M.C.d.B.; hydrodynamic injection studies by J.W.L, D.L., and M.G.; parabiotic joining by J.W.L. and P.M.P. M.C.d.B, F.C.d.B., and N.R.W. provided Saa−/− mice; D.X. and A.S. provided liver specimens from healthy donors; M.J.B. and R.K.R. provided liver specimens from patients with PDAC; T.A.B., A.L.C., K.S.M., J.C.T., S.S.Y., M.H.O., C.A., and E.L.C. provided plasma samples; M.G. established the hydrodynamic injection procedure; D.L. designed and prepared the IL-6 expression vector ; P.M.P., M.G., D.L., E.L.C., M.C.d.B., F.C.d.B., and N.R.W. provided experimental advice; J.W.L. and G.L.B designed the study; and J.W.L., M.L.S., and G.L.B. prepared and wrote the manuscript.

Corresponding author

Correspondence to Gregory L. Beatty.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

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

Extended data figures and tables

Extended Data Fig. 1 Primary PDAC development induces myeloid cell accumulation and fibrosis within the liver.

a, Gating strategy for identification of F4/80+, Ly6G+, CD3+, and CD19+ cells. Representative images from flow cytometric analysis of cells isolated from the liver of a tumour-bearing KPC mouse are shown. b, Quantification of immune cells in the liver by flow cytometry. Numbers in parentheses indicate the number (n) of mice. Data pooled from four experiments. c, Representative Sirius red staining of the liver (n = 5 for all groups) viewed using standard light microscopy (top) and polarized light (bottom). d, mRNA levels of Fn1, Col1a1, and Des in the liver (n = 6 for all groups). Data pooled from two experiments (c, d). ei, Wild-type mice were orthotopically injected with PBS or PDAC cells and analysed on day 20. e, Quantification of immune cells in the liver by flow cytometry (n = 5 for both groups). f, t-distributed stochastic neighbour embedding (t-SNE) 2D plots of immune cells analysed in e. g, Quantification of myeloid cells in the liver (n = 7 for PBS and n = 6 for PDAC). h, i, Images and quantification of fibronectin and COL1 in the liver (n = 6 for both groups). Data representative of at least two independent experiments (ei). j, Study design for k and l (n = 4 for mice injected with PBS; n = 4 and 3 for mice injected with PDAC cells and then intraperitoneally injected with PBS and clodronate-encapsulated liposomes (CEL), respectively). k, Images and quantification of F4/80+ cells in the liver. l, Quantification of COL1 and fibronectin in the liver. Data representative of one experiment (jl). Scale bars, 100 μm (c) and 50 μm (other panels). Statistical significance calculated using two-tailed unpaired Student’s t-test (d), one-way ANOVA with Dunnett’s test (k, l), and two-tailed Mann–Whitney test (other panels). NS, not significant. Data represented as mean ± s.d., except b and e, which are shown as box plots (centre line, median; box limits, upper and lower quartiles; whiskers, maximum and minimum values). Source data

Extended Data Fig. 2 Primary PDAC development enhances liver susceptibility to metastatic colonization.

a, b, Control mice (n = 14) and NTB KPC mice (n = 10) were intrasplenically injected with PDAC–YFP cells, and the liver was analysed after 10 days. a, Images of the liver showing metastatic lesions (yellow, CK19) and Ki-67 (purple). Scale bars, 4 mm (left) and 200 μm (right). b, Quantification of lesions (left) and Ki-67+ tumour cells (right). Data pooled from three experiments (a, b). c, d, n = 9 and 13 for control mice and n = 6 and 13 for NTB KPC mice intrasplenically injected with PBS and PDAC (PDA.69), respectively. c, Images of the liver (top) and metastatic lesions in the liver (stained with PDX1). Scale bars, 1 cm (top) and 50 μm (bottom). d, Liver weights and mRNA levels of Pdx1 in the liver relative to Gapdh. Data pooled from five experiments (c, d). e, Study design for f and g. Wild-type mice were injected with PBS or PDAC cells and then injected with PDAC–YFP cells on day 10. The liver was removed 2 h (f, n = 4 for both groups) or 24 h (g, n = 5 for both groups) after intraportal injection. f, g, Images and quantification of tumour cells in the liver. Scale bars, 50 μm. Data representative of one experiment (eg). h, Study design for il (n = 4 for both groups). i, Images of the liver and flow cytometric analysis. Scale bars, 1 cm. j, Quantification of PDAC–YFP cells. k, l, Images of the liver showing metastatic lesions (yellow, CK19) and Ki-67 (purple). Scale bars, 4 mm (k) and 200 μm (l). Data representative of at least three independent experiments (hl). m, Study design for n and o (n = 5 for both groups). n, Gating strategy for identification of T cell subsets. Images from flow cytometric analysis of cells isolated from the liver of a wild-type mouse are shown. EM, effector memory. o, Quantification of CD4+ T cell subsets (top) and CD8+ T cell subsets (bottom) in the liver. Data representative of two independent experiments (mo). ps, Wild-type mice were orthotopically injected with PBS or PDAC cells (n = 4 for both groups). One group received anti-CD4 and anti-CD8 antibodies on days 8 and 13. Both groups were intraportally injected with PDAC–YFP cells on day 10. p, Flow cytometric analysis of peripheral blood and quantification of CD4+ and CD8+ T cells. q, Images of the liver, flow cytometric analysis, and quantification of PDAC–YFP cells. Scale bars, 1 cm. r, Images of the liver showing metastatic lesions (yellow, CK19), Ki-67 (purple, top), and CD4+ cells (purple, bottom) and CD8+ cells (brown, bottom). Scale bars, 200 μm (top) and 50 μm (bottom). s, Quantification of lesions and Ki-67+ tumour cells (top) and CD4+ cells and CD8+ cells (bottom). Data representative of one experiment (ps). Statistical significance calculated using one-way ANOVA with Dunnett’s test (d) or two-tailed Mann–Whitney test (other panels). NS, not significant. Data represented as mean ± s.d. Source data

Extended Data Fig. 3 Primary PDAC development induces expression of myeloid chemoattractants and activates STAT3 signalling in the liver.

a, Study design for be (n = 5 for both groups for bd and n = 10 and 9 for control mice and NTB KPC mice, respectively, for e). b, Heat map showing differentially expressed genes in the liver. c, Enriched biological processes in the liver of NTB KPC mice. Left, significance; right, number of genes in each group. d, FPKM values for chemoattractant genes in the liver. e, mRNA levels of chemoattractant genes in the liver. Data representative of one experiment (ae). f, Enrichment of IL-6–JAK–STAT3 signalling genes in the liver (n = 5 for control mice and NTB KPC mice). FDR, false discovery rate; NES, normalized enrichment score. Data representative of one experiment. g, Left, images showing hepatocytes (stained for albumin) and pSTAT3. Right, percentage of hepatocytes that were pSTAT3+ in control mice (n = 4), NTB KPC mice (n = 5), and TB KPC mice (n = 5). h, Left, images of F4/80+ cells and pSTAT3. Right, percentage of F4/80+ cells that were pSTAT3+ in control mice (n = 9), NTB KPC mice (n = 8), and TB KPC mice (n = 5). i, Images of F4/80+ cells and pSTAT1 in the liver of control mice (n = 9), NTB KPC mice (n = 8), and TB KPC mice (n = 5). Data pooled from two experiments (gi). j, k, Wild-type mice were orthotopically injected with PBS (n = 7) or PDAC cells (n = 6). Left, images of hepatocytes (j), F4/80+ cells (k), and pSTAT3. Right, percentage of hepatocytes (j) and F4/80+ cells (k) that are pSTAT3+. Scale bars, 50 μm. Statistical significance calculated using ClueGO35 (c), two-tailed unpaired Student’s t-test (d), GSEA40 (f), one-way ANOVA with Dunnett’s test (g, h), and two-tailed Mann–Whitney test (other panels). NS, not significant. Data represented as mean ± s.d., except d, e, which are shown as box plots (centre line, median; box limits, upper and lower quartiles; whiskers, max and min values). Source data

Extended Data Fig. 4 IL-6 promotes the formation of a pro-metastatic niche in the liver.

ag, n = 5 and 6 for Il6+/+ mice and n = 4 and 5 for Il6−/− mice orthotopically injected with PBS and PDAC cells, respectively. a, Images of pSTAT3+ cells, myeloid cells, and fibronectin. Arrows indicate Ly6G+ cells. b, c, Images and quantification of COL1. d, Images of sinusoids (brown, stained with CD31) in the liver. e, mRNA levels of Lcn2, S100a8, S100a9, Ccl6, Cxcl1, Fn1, Col1a1, and Des in the liver. f, Images of pancreas and primary tumour stained with CD31 (brown), CK19 (yellow), and Ki-67 (purple). g, Quantification of the weight of pancreas or primary tumour (left), number of Ki-67+ tumour cells (middle), and vascular area (right). Data representative of one (bd, Fn1, Col1a1, and Des in eg) or two independent experiments (a, all other genes in e). h, Study design for ik (n = 5 for all groups). All groups were injected with PDAC–YFP cells on day 10. i, j, Images of the liver showing metastatic lesions (yellow, CK19) and Ki-67 (purple). Scale bars, 4 mm (i) and 200 μm (j). k, Quantification of lesions (left) and Ki-67+ tumour cells (right). Data representative of one experiment (hk). l, Study design for mo (n = 4 and 5 for mice injected with PBS and treated with isotype control or anti-IL-6R antibodies, respectively; n = 7 and 8 for mice injected with PDAC cells and treated with isotype control or anti-IL-6R antibodies, respectively, unless indicated otherwise below). m, n, Images and quantification of pSTAT3+ cells, myeloid cells, and fibronectin. For fibronectin, n = 7 for mice injected with PDAC cells and treated with anti-IL-6R antibodies. Arrows indicate Ly6G+ cells. o, mRNA levels of Saa1 and Saa2. p, Study design for qs (n = 7 and 8 for mice injected with PBS and treated with isotype control and anti-IL-6R antibodies, respectively; n = 7 for mice injected with PDAC cells and treated with isotype control and anti-IL-6R antibodies). All groups were injected with PDAC–YFP cells on day 10. q, r, Images of the liver and flow cytometric analysis. Scale bars, 1 cm. s, Quantification of PDAC–YFP cells. Data representative of two independent experiments (ls). Scale bars, 50 μm unless indicated otherwise. Statistical significance calculated using one-way ANOVA with Dunnett’s test. NS, not significant. Data represented as mean ± s.d. Source data

Extended Data Fig. 5 Non-malignant cells are the predominant source of IL-6.

a, Study design for be (n = 5 and 6 for Il6+/+ mice injected with PBS and PDAC cells, respectively; n = 4 and 5 for Il6−/− mice injected with PBS and PDAC cells, respectively). SN, supernatant. b, Concentration of IL-6 in pancreas supernatant and serum collected from indicated sites in Il6+/+ mice injected with PBS or PDAC cells. c, Concentration of IL-6 in pancreatic tumour supernatant and serum collected from indicated sites in Il6−/− mice injected with PBS or PDAC cells. Solid lines indicate data points from individual mice, and dashed lines indicate the lower limit of detection (b, c). d, Images of CK19 and Il6 mRNA in the pancreas and primary tumour. e, Images of Il6 mRNA in the liver and lung of Il6+/+ mice injected with PDAC cells. Data representative of two independent experiments (ae). f, Study design for g and h (n = 4 for Il6+/+ mice injected with PDAC cells and n = 5 for human samples). g, Images of α-SMA, CD31, CK19, and Il6 mRNA in perivascular cells (top), stromal cells (bottom, left), endothelial cells (bottom, middle) and malignant cells (bottom, right) present within the mouse primary tumour. h, Images of α-SMA (yellow), CD31 (yellow), CK19 (yellow), and IL6 mRNA (brown) in perivascular cells (top, left), stromal cells (top, right), endothelial cells (bottom, left) and malignant cells (bottom, right) present within the human primary tumour. Data representative of one experiment (fh). i, Study design for j and k (n = 5). j, k, Representative images of α-SMA, CK19, and Il6 mRNA detected in PanIN (j) and invasive PDAC (k). Data representative of one experiment (ik). Scale bars, 50 μm. Statistical significance calculated using two-tailed Wilcoxon test. NS, not significant. Dagger, blood vessel; double dagger, PanIN lesion. Data from individual mice shown in b, c. Source data

Extended Data Fig. 6 STAT3 signalling in hepatocytes promotes the formation of a pro-metastatic niche in the liver.

a, b, Representative images and quantification of pSTAT3+ hepatocytes (n = 3 technical replicates per condition) treated with or without IL-6 (a) or anti-IL-6R (b). Arrows indicate pSTAT3+ hepatocytes. SN, pancreatic tumour supernatant. Data representative of two independent experiments. For cf and n, n = 4 for Stat3flox/flox mice and n = 8 and 7 for Stat3flox/flox Alb-cre mice orthotopically injected with PBS or PDAC cells, respectively. c, mRNA levels of Fn1 in the liver. d, Images of sinusoids (brown, stained for CD31) in the liver. e, Images of pancreas and primary tumour stained for CD31 (brown), CK19 (yellow), and Ki-67 (purple). f, Quantification of the weight of pancreas or primary tumour (left), number of Ki-67+ tumour cells (middle), and vascular area (right). Data representative of one experiment (cf). g, Study design for hl (n = 4 and 5 for Stat3flox/flox mice and Stat3flox/flox Alb-cre mice, respectively). All groups were injected with PDAC–YFP cells on day 10. h, Flow cytometric analysis. i, Quantification of PDAC–YFP cells. j, k, Images of the liver showing metastatic lesions (yellow, CK19) and Ki-67 (purple). Scale bars, 4 mm (j) and 200 μm (k). l, Quantification of lesions (left) and Ki-67+ tumour cells (right). Data representative of one experiment (gl). m, Images of SAA detected by immunohistochemistry (brown, n = 5 for wild-type mice orthotopically injected with PBS or PDAC cells). Dashed lines and asterisks indicate sinusoids and hepatocytes, respectively. n, Images of Saa1 and Saa2 mRNA (brown) detected by RNA in situ hybridization. Data representative of one experiment (m, n). Scale bars, 50 μm unless indicated otherwise. Statistical significance calculated using one-way ANOVA with Dunnett’s test. NS, not significant. Data represented as mean ± s.d. Source data

Extended Data Fig. 7 SAA promotes the formation of a pro-metastatic niche in the liver.

a, Concentration of circulating SAA in healthy donors (n = 69), patients with locally advanced PDAC (n = 28), and patients with liver metastases (n = 43). Data represented as a box plot (centre line, median; box limits, upper and lower quartiles; whiskers, max and min values). b, Images of SAA (yellow) and pSTAT3 (purple) in the liver of healthy donors and patients with PDAC with liver metastases. Dashed lines and asterisks indicate sinusoids and hepatocytes, respectively. c, Kaplan–Meier survival curve for patients with PDAC with liver metastases who had low (<250 μg ml–1, black, n = 21) or high (>250 μg ml–1, red, n = 22) levels of circulating SAA. d, Concentration of circulating SAA in patients with locally advanced NSCLC (n = 8) and patients with NSCLC with liver metastases (n = 13). Data shown as a box plot (centre line, median; box limits, upper and lower quartiles; whiskers, max and min values). e, Images of SAA (brown) in the liver of CRC patients with liver metastases. Dashed lines and asterisks indicate sinusoids and hepatocytes, respectively. Data representative of one experiment (ae). f, g, n = 5 and 4 for Saa+/+ mice and n = 4 and 5 for Saa−/− mice orthotopically injected with PBS or CRC cells (MC-38), respectively. f, Quantification of myeloid cells, fibronectin, and COL1. g, Images of fibronectin (left) and COL1 (right). Data representative of one experiment (f, g). ho, n = 5 for all groups unless indicated otherwise. h, Images of pSTAT3+ cells, myeloid cells, and fibronectin. i, j, Images and quantification of COL1. k, Images of sinusoids (brown, CD31) in the liver. l, mRNA levels of Lcn2, S100a8, S100a9, Ccl6, Cxcl1, Fn1, Col1a1, and Des in the liver. m, mRNA levels of Saa1 and Saa2. n, Images of pancreas and primary tumour stained for CD31 (brown), CK19 (yellow), and Ki-67 (purple). o, Quantification of the weight of pancreas or primary tumour (left), number of Ki-67+ tumour cells (middle), and vascular area (right). For weight, n = 4 for Saa+/+ mice injected with PDAC cells. Data representative of one (ik, Fn1, Col1a1, and Des in l, n, o) or two independent experiments (h, all other genes in l, m). p, Study design for qs (n = 8 and 5 for Saa+/+ mice and n = 5 and 7 for Saa−/− mice injected with PBS and PDAC cells, respectively). All groups were injected with PDAC–YFP cells on day 10. q, r, Images of the liver showing metastatic lesions (yellow, CK19) and Ki-67 (purple). Scale bars, 4 mm (q) and 200 μm (r). s, Quantification of lesions (left) and Ki-67+ tumour cells (right). Data representative of one experiment (ps). Scale bars, 50 μm unless indicated otherwise. Statistical significance calculated using two-sided Mann–Whitey test (a, d), Mantel–Cox test (c), and one-way ANOVA with Dunnett’s test (other panels). NS, not significant. Data represented as mean ± s.d. unless indicated otherwise. Source data

Extended Data Fig. 8 SAA is a downstream mediator of IL-6 signalling that drives myeloid cell accumulation and fibrosis within the liver.

a, Study design for bd (n = 5 for all groups, except n = 4 for Saa−/− mice). b, Concentration of IL-6 in the serum collected from control mice (left), Il6−/− mice (middle), and Saa−/− mice (right) on indicated days. c, d, Images and quantification of pSTAT3+ cells, myeloid cells, fibronectin, and COL1. Arrows indicate Ly6G+ cells. Data representative of one experiment (ad). e, Study design for fh (n = 5 for mice injected with pLIVE-vector; n = 6 for mice injected with pLIVE-IL-6). f, Concentration of IL-6 in the serum. g, Images of the liver showing metastatic lesions (yellow, CK19) and Ki-67 (purple). h, Quantification of lesions (left) and Ki-67+ tumour cells (right). Data representative of one experiment (eh). i, Study design for j (n = 5 for all groups, except n = 4 for Saa−/− mice injected with CCl4). j, Images and quantification of myeloid cells, fibronectin, and COL1. Arrows indicate Ly6G+ cells. Data representative of one experiment (i, j). Dashed lines indicate the lower limit of detection (b, f). Scale bars, 50 μm. Statistical significance calculated using two-tailed Mann–Whitney test (b, f, h) or one-way ANOVA with Dunnett’s test (other panels). NS, not significant. Data represented as mean + s.d. or mean ± s.d. Source data

Extended Data Fig. 9 IL-6–STAT3–SAA signalling axis does not affect expression of MIF and TIMP1.

a, Study design for be (n = 5 and 4 for mice injected with PBS and PDAC cells). b, c, Images of CK19 (yellow) and TIMP1 (purple) in the pancreas, primary tumour, and liver. Data representative of at least three independent experiments (ac). d, FPKM values of genes in the liver of control mice (n = 5) and NTB KPC mice (n = 5) obtained from QuantSeq 3′ mRNA sequencing. Data represented as box plots (centre line, median; box limits, upper and lower quartiles; whiskers, max and min values). e, mRNA levels of Timp1 in the liver of control mice (n = 6), NTB KPC mice (n = 7), and tumour-bearing KPC mice (n = 6) relative to Actb. Data representative of one experiment (d, e). f, j, n, Study designs for gi, km and oq, respectively. For f, n = 5 for Il6+/+ mice injected with PBS and n = 5 or 6 for Il6+/+ mice injected with PDAC cells. n = 4 and 5 for Il6−/− mice injected with PBS and PDAC cells, respectively. For j, n = 5 for all groups, except n = 4 or 5 for Saa+/+ injected with PDAC cells. For n, n = 4 for Stat3flox/flox mice and n = 8 and 7 for Stat3flox/flox Alb-cre mice injected with PBS and PDAC cells, respectively. g, k, o, mRNA levels of Mif and Timp1 in the indicated organs relative to Actb. h, l, p, Images of CK19 (yellow) and TIMP1 (purple) in the pancreas and primary tumour. i, m, q, Concentration of TIMP1 in the serum. Data representative of one experiment (fq). Scale bars, 50 μm. Statistical significance calculated using one-way ANOVA with Dunnett’s test (i, m, q) or two-tailed Mann–Whitney test (other panels). NS, not significant. ND, not detected. Data represented as mean ± s.d. unless indicated otherwise. Source data

Extended Data Fig. 10 Primary PDAC development induces a systemic response that promotes the formation of a pro-metastatic niche in the liver.

a, Study design for bg (n = 4 and 8 for CD45.2 mice injected with PBS (group 1) and PDAC cells (group 2), respectively). b, Assessment of chimaerism in parabiotically joined mice. Flow cytometric analysis of peripheral blood gated on CD45.1+ and CD45.2+ cells as a percentage of CD45+ cells. c, Dorsal (top) and ventral (bottom) views of parabiotically joined mice. Dagger, site of laparotomy for orthotopic injection; double dagger, skin suture for parabiotic joining. d, Images and quantification of myeloid cells and fibronectin. Arrows indicate Ly6G+ cells. e, mRNA levels of Saa. f, Concentration of IL-6 in pancreas supernatant and serum. Solid lines indicate data points from individual mice. Dashed lines indicate the lower limit of detection. g, Concentration of circulating SAA. Data representative of one (f, g) or two independent experiments (ae). h, i, n = 5 for all groups. h, Images of myeloid cells and fibronectin in the lung of control mice, NTB KPC mice, and tumour-bearing KPC mice. i, Quantification of myeloid cells and fibronectin. Data representative of one experiment (h, i). j, Study design for km. k, n = 4 and 5 for Il6+/+ mice and n = 5 and 4 for Il6−/− mice injected with PBS and PDAC cells, respectively. For l, n = 5 for all groups. For m, n = 4 for Stat3flox/flox mice and n = 8 and 7 for Stat3flox/flox Alb-cre mice injected with PBS and PDAC cells, respectively. km, Quantification of myeloid cells and fibronectin. Data representative of one experiment (jm). n, Study design for o (n = 4 for all groups of mice). All groups of mice were injected with PDAC–YFP cells on day 10. o, Quantification of PDAC–YFP cells. Data representative of one experiment (n, o). p, Conceptual model of generation of pro-metastatic niche in the liver. Scale bars, 50 μm. Statistical significance calculated using two-sided Mann–Whitney test (d, e, g), two-sided Wilcoxon test (f), and one-way ANOVA with Dunnett’s test (other panels). NS, not significant. Data represented as mean ± s.d., except d, e, g, which are shown as box plots (centre line, median; box limits, upper and lower quartiles; whiskers, max and min values). Data from individual mice are shown in f. Source data

Supplementary information

Supplementary Information

This file contains Supplementary Data 1 and Supplementary Tables 1-4.

Reporting Summary

Source data

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Lee, J.W., Stone, M.L., Porrett, P.M. et al. Hepatocytes direct the formation of a pro-metastatic niche in the liver. Nature 567, 249–252 (2019). https://doi.org/10.1038/s41586-019-1004-y

Download citation

Further reading

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.

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing