Key reagents and resource identifiers are provided in Supplementary Table 3.
Human samples
Human HCC TMAs used in this study were obtained with informed patient consent from K.B. as described previously53. In brief, TMAs with formalin-fixed paraffin-embedded (FFPE) tissues (n = 731) contained tumour-free/cirrhotic livers (n = 241), premalignant dysplastic nodules (n = 14) and HCCs (n = 473; with G1 (87), G2 (311), G3 (75))20. Tissue cores had a diameter of 1 mm and slides had a thickness of 1–2 µm. We complied with all relevant ethical regulations. The study was approved by the institutional ethics committee of the Medical Faculty of Heidelberg University (S-206/2005). Liver sections and snap-frozen tissue samples from healthy donors and patients with hepatitis were obtained from M.R. and N.R. with the approved institutional review board (IRB) protocol (2012-293N-MA) from the University Hospital Mannheim; from S. Roth with the approved ethical protocol S-629/2013; from A.W. with the approved application number KEK-ZH-Nr 2013-0382 by the local ethics committee (Kantonale Ethikkommission Zurich) in University Hospital Zurich. Human liver sections involved in spatial biology and IMC analysis were obtained from M. Hofmann, following the Declaration of Helsinki (1975), federal guidelines and local ethics committee regulations (Albert-Ludwigs-University, Freiburg, Germany, 20-1066). Detailed information is provided in the ‘IMC analysis of human samples’ section and Supplementary Table 2.
Mice, diets, and treatments
The nomenclature and a description of ATF6α mouse models are provided in Supplementary Table 1. The nATF6fl/fl (R26-LSL-nATF6-HA) mouse line was obtained by D. Haller15. Alb-cre mice and Atf6fl/fl mice were obtained from The Jackson Laboratory. Pdcd1−/− mice were provided by G. Tiegs and K. Neumann54. Atf6−/− mice were described previously by R.J.K.3. MUP-uPA mice were described previously37. The nATF6fl/fl mice were crossed with Alb-cre mice or intravenously injected with AAV8-cre (Vector Biolabs, VB1724 or VB1743 GFP control, 1E11VG/mouse) to generate hepatocyte-specific nATF6-HA-overexpressing heterozygous mice (TGAlb-cre+ or TGAAV-cre). Hepatocyte-specific nATF6-HA-overexpressing heterozygous mice (TGAlb-cre+) were bred with Pdcd1−/− mice to generate TG:Pdcd1−/− mice. Heterozygous R26-LSL-nATF6-HA mice (nATF6fl/+) were intravenously co-injected with AAV8-cre and AAV8-FBP1 or AAV8-FBP1E98A (plasmids were provided by M.K. and L.G.31; 1E11VG/mouse) to generate hepatocyte-specific FBP1-overexpressing mice (TGAAV-cre/fbp1 or TGAAV-cre/fbp1E98A). Atf6−/− and littermate Atf6+/+ mice were bred with MUP-uPA mice to generate Atf6−/−:MUP-uPA and Atf6+/+:MUP-uPA mice. The Atf6fl/fl mice were crossed with Alb-cre mice to generate hepatocyte-specific Atf6-knockout mice (Atf6ΔHep). All of the mouse lines were either on a pure C57BL/6J genetic background or crossed into it for at least ten generations.
Mice were housed under specific-pathogen-free (SPF) conditions at the German Cancer Research Center (DKFZ) or Sanford Burnham Prebys (SBP) at constant temperature of 20–24 °C and 45–65% humidity under a 12 h–12 h light–dark cycle. All control mice were age, gender and genetic-background matched. Where applicable, littermate controls were used to minimize the variation between mouse strains.
For mice receiving injections, the following protocols were used where applicable. TGAlb-cre+ mice (aged 9 months) were treated with anti-PD-1 antibody (Bioxcell, BE0146) or isotype control (Bioxcell, BE0089) at an initial dose of 500 μg i.p. followed by doses of 200 μg i.p. bi-weekly for 12 weeks, as previously described55. Mice (aged 2 weeks) were i.p. injected once with DEN (Sigma-Aldrich, N0756, 25 mg per kg). GalNAc conjugation to ASOs against Atf6 (GalNac-ASO-Atf6; Gen 2.5 ASO (16-mer 3-10-3): GAATTTTTCAGCAAGG conjugated to GalNAc on the 5′ end; Ionis Pharmaceuticals) or a scrambled nucleotide sequence (GalNac-ASO-scramble; Gen 2.5 ASO (16-mer 3-10-3): CGCCGATAAGGTACAC conjugated to GalNAc on the 5′ end; Ionis Pharmaceuticals) were subcutaneously injected at 2.5 mg per kg once weekly at 4, 9 or 30 weeks of age (see the schematics in the figures). Oncogene NRASG12V plasmid DNA (Addgene, 20205) was administered at 20 µg transposon (NRASG12V) combined with 10 µg transposase (Sleeping Beauty (SB) 100, Addgene, 34879) in 2.5 ml by HDTVi per mouse at 8 weeks of age (see the schematics in the figures). The oncogene KRASG12D (5 µg per mouse, from D.T.), MYC (10 µg per mouse, from D.T.) and sg-P53 (10 µg per mouse in combination with KRASG12D, 20 µg per mouse in combination with MYC, Addgene, 59910) plasmid DNA were delivered together with SB transposase (transposon:transposase, 5:1; from D.T.) in 2 ml saline solution through HDTVi to the mouse liver. For HDTVi experiments, mice aged 8–12 weeks were used (see the schematics in the figures).
Dietary models started after 6 weeks of age (see the schematics in the figures) and included HFD (60% HFD; BioServ F3282 or Research Diets D12492i), CD-HFD (Research Diets D05010402) and WD (Research Diets D1602230i). Cholaemic mice were excluded from dietary experiments56. The i.p. glucose tolerance test and insulin tolerance test were performed as previously described57. The pyruvate tolerance test was performed in 16-h-fasted mice by measuring the blood glucose levels after a 2 g per kg pyruvate i.p. injection. Many treatment regimens, with experimental schemes with timelines shown in the figures, extended data figures and supplementary figures, used previously published reagents and standard experimental techniques55.
Housing and breeding of mice without interventions were performed in accordance with the approved protocols (A-23/17, EP-Z146102, G6/22 and G279/16) in the German Cancer Research Center (DKFZ). Mouse experiments were performed in accordance with German law and the governmental bodies, with approval from the Regierungspräsidium Karlsruhe (DKFZ 332, G6/22, G11/16, G129/16, G279/16, G7/17, G80/17, G70/18, G178/19, G141/19, G132-23 and G97/24) or National Institute of Health (NIH) guidelines of the United States, with approval from the SBP Institutional Animal Care and Use Committee (IACUC, AUF 23-027 (previously 20-030), AUF 23-045 (previously 20-056)). Tumour models used in this study were orthotopic hepatic tumour models; thus, direct calliper-based measurement of tumour size in living mice was not feasible. Animal monitoring and experimental procedures strictly adhered to the termination criteria outlined in the above-mentioned protocols (DKFZ 332, G6/22, G11/16, G129/16, G279/16, G7/17, G80/17, G70/18, G178/19, G141/19, G132-23 and G97/24 in DKFZ; IACUC, AUF 23-027 and AUF 23-045 in SBP). Each mouse was examined daily by trained animal care staff or research personnel. Animals exhibiting signs of distress, morbidity, clinical signs of pain or distress (including but not limited to cachexia, cyanosis, dyspnoea, ascites, or lack of mobility, food and water intake), or any abnormality meeting the predefined termination criteria were promptly euthanized, after which the biological materials were collected. These limits were not exceeded in any of the experiments. Mice that remained clinically normal and did not reach the termination criteria were maintained until the designated experimental endpoint, at which time they were euthanized, and the liver tumours were excised and measured.
Measurements of serum parameters
... continue reading