Fig 2: DKK1-controlled inflammation involves SOCS3-restricted nuclear RelA activity.a IL1B mRNA levels in wildtype and DKK1-deficient PC3 cells following exposure to DMSO (−) or XAV939 (+) (5 µM) for 24 h (also see Fig. S3). The two DKK1-deficient conditions shown correspond to two independent siRNAs as in previous figures (n = 5/condition). b Transcript levels of IL1B in wild-type PC3 cells 48 h after siRNA-mediated knockdown of the two Wnt- and DKK1-receptors LRP5, LRP6, or the combination of both (also see Fig. S3) (n = 3). c IL1B mRNA expression in wild-type PC3 cells treated with the proteasome inhibitor bortezomib (BZ) at 100 nM for 6 h (n = 3). d IL1B transcript levels in wild-type PC3 cells treated with the CRM1 inhibitor leptomycin B (LMB) at 25 nM for 6 h (n = 3). e IL1B mRNA expression in DKK1-deficient PC3 cells exposed to the translation inhibitor cycloheximide (CHX) (1 µM) for 6 h. Both DKK1-targeting siRNAs (siRNA#1 and siRNA#2, respectively) are shown. The dashed line corresponds to IL1B expression of DMSO-treated PC3 cells transfected with non-targeting control oligonucleotides (siCtrl) (n = 3). f Normalized enrichment scores (NES) for signaling pathways linked to the pro-inflammatory response in DKK1-deficient vs. DKK1-competent PC3 cells. Results were generated using the RNAseq data from previous figures and CARNIVAL. g Representative SOCS1 and SOCS3 immunoblot from wild-type PC3 cells exposed to DMSO or Cycloheximide (CHX) for 6 h. h Representative immunoblot of subcellular fractions from wild-type PC3 cells treated with leptomycin B (LMB) or DMSO for 2 h. The upper SOCS1 lane is strongly contrast-stretched to aid visualization. i IL1B mRNA levels in DKK1-deficient cells exposed to siRNAs directed against SOCS1 (left panel) or SOCS3 (right panel). Results for two independent DKK1-targeting siRNAs are shown (siRNA#1 and siRNA#2, respectively) and normalized to siCtrl (n = 4/condition). j IL1B protein levels in total cells lysates of DKK1-competent and -deficient PC3 with or without SOCS3 knockdown 24 h following LPS exposure (1 µg/ml) (n = 3). k, l Representative immunoblot of subcellular fractions from wildtype (siCtrl) and DKK1-deficient (siDKK1) PC3 cells. Band intensities were quantified by ImageJ. m Representative immunoblot of subcellular fractions from DKK1-deficient cells transfected with (+) or without (−) siRNA directed against SOCS3 (siSOCS3). n IL1B mRNA levels in DKK1-competent and -deficient PC3 cells transfected with an empty backbone vector (pCMV4) or a plasmid encoding for human RelA (pCMV4-RELA) (n = 4/condition). *p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001. Data were expressed as mean ± SEM [a, j, n) one-way ANOVA with Holm–Sidak’s post hoc test, c–e, i and l two-tailed, unpaired student’s t-test].

Fig 3: Genetic DKK1 variants are linked to boosted cytokine production in humans.a Associations between genetic DKK1 variants and cytokine expression in response to infectious triggers (cytokine quantitative trait loci, cQTLs) from the 500FG study. The numbers shown on the top of the bars correspond to distinct experimental setups used in the original study29 (see Supplementary Data 1). b Associations between genetic DKK1 variants (protein quantitative trait loci; pQTLs) and the expression of various cytokines from an independent human cohort30. Log-transformed p values are shown. c A bipartite network depiction of the DKK1-associated pQTLs and various cytokines/chemokines. Green hexagonal nodes depict pQTLs, whereas cytokines are visualized as pink ovals. Positive associations (DKK1 variants linked to increased cytokine expression) are visualized in red and negative associations (reduced cytokine expression) in blue with color gradations in between. Effect sizes are highlighted by the thickness of the respective arrows. DKK1 variants shared between the two study populations are highlighted in yellow.

Fig 4: Suppression of DKK1 production curtails inflammatory cytokine expression in cancer cells.a Differential expression analysis of RNAseq data from DKK1-deficient and wild-type PC3 cells (siDKK1 and siCtrl, respectively) visualized as a volcano plot. Significantly differentially expressed genes are highlighted in red (n = 3/genotype). b Results from Gene Set Enrichment Analysis (GSEA) visualized as a bubble plot. c Pro-inflammatory cytokine and chemokine PCR array from DKK1-deficient (siDKK1) and DKK1-competent (siCtrl) PC3 cells. The most differentially regulated transcripts are shown (mean of n = 3/genotype is shown). d Quantification of selected cyto-/chemokines by qPCR analysis in DKK1-deficient and -competent PC3 cells. Results for two independent DKK1-targeting siRNAs are shown (referred to as siDKK1#1 and siDKK1#2 and visualized as two separate bars across all figures) (n = 3/genotype). e IL1B and CXCL8 protein levels in total cell lysates and cell culture supernatants from PC3 cells with or without DKK1 knockdown. IL1B remained undetectable in cell culture supernatants (n = 3–5/genotype). f CXCL8 and IL1B mRNA expression in DU-145 prostate and MD-MB-231 breast cancer cells 48 h following siRNA-mediated DKK1 knockdown determined by qPCR. CXCL8 was chosen as a read-out in the former cells as IL1B expression was barely detectable (n = 4/genotype) *p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001. Data were expressed as mean ± SEM. [d, e: one-way ANOVA with Holm–Sidak’s post hoc test, f two-tailed, unpaired student’s t-test].

Fig 5: Genetic Dkk1 deletion ameliorates inflammation and disease trajectories in a mouse model of endotoxemia.a Circulating Il1b levels in C57BL/6J mice 3 h following LPS (5 mg/kg) treatment. Eighteen hours prior to being challenged with LPS, mice received a single injection of a Dkk1-neutralizing antibody (DKN01, 5 mg/kg) or an equivalent dose of isotype control antibody (murine IgG) (n = 5–8/group). b Il1b mRNA expression in primary murine osteoblasts (mOB) differentiated from bone marrow precursors of C57BL/6J mice. Cells were differentiated for 7 days before being exposed to DKN01 (10 µg/ml) or isotype control (murine IgG; 10 µg/ml) for 18 h, followed by LPS treatment (1 µg/ml) for 6 h (n = 3/group). c Il1b mRNA expression in Dkk1 knock-out (isolated from bone marrow precursors of Dkk1−/− mice) or wild-type mOBs (obtained from littermate controls) treated with LPS (1 µg/ml) for 6 h (n = 3/group). d Il1b protein expression in bone lysates of mice with (Dkk1ΔDmp1) or without (littermate controls; referred to as wildtype or “WT”) osteoblast/osteocyte-specific Dkk1 deletion treated with LPS (5 mg/kg) for 3 h. Bones were collected in PBS, flushed to remove the bone marrow, and total protein of the lysate was analyzed for Il1b concentration by ELISA (n = 8–9/group). e Il1b concentration in cell culture supernatants determined by ELISA following 24 h cultivation of the contralateral femur (without the bone marrow) from the same animals (n = 4–5/group). f–i Circulating and splenic tissue levels of Il1b and tumor necrosis factor-alpha (Tnfa) in Dkk1 knock-out mice (Dkk1−/−) or wild-type littermate controls (WT) 3 h following treatment with LPS or water (n = 3–5/group). j Cytokine proteome profiler array of spleen lysates. Membranes precoated with cytokine/chemokine-specific antibodies (in duplicates) were incubated with spleen lysates from Dkk1−/− mice or littermate controls. Differences between selected cyto-/chemokines were analyzed by quantification of dot intensities using ImageJ and expressed as fold-change (lower panel). Bar graphs show results from three biologically independent replicates (= 3 mice/genotype). k, l Circulating alanine aminotransferase (ALAT) and creatinine kinase (CK) levels in Dkk1−/− mice and littermate controls 30 h following LPS administration. Results are expressed as fold-change compared to LPS-treated wild-type controls (n = 5–8/group). m Survival of wildtype (n = 7) and Dkk1−/− mice (n = 10) treated with LPS (15 mg/kg). *p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001. Data were expressed as mean ± SEM. [a–c and f–l one-way ANOVA with Holm–Sidak’s post hoc test, d, e and j–l) two-tailed, unpaired student’s t-test].