Fig 2: NOTCH3-IC is a master regulator of ADRN-to-MES reprogramming a. Analysis of mRNA signature scores for MES and ADRN cells in a panel of neuroblastoma cell lines and a time series of NOTCH3-IC induction. Control (-dox) SH-SY5Y-NOTCH3-IC cells are depicted in blue; SH-SY5Y cells with induction of NOTCH3-IC (+dox) are shown in red. Time of NOTCH3-IC induction is indicated in days. Dashed circles group cell lines of MES- or ADRN-type as previously determined1. b Z-score mRNA expression of core super-enhancer associated transcription factors specific for ADRN or MES cells. Time of NOTCH3-IC induction (in days) is indicated above. dox, doxycycline. c Western blot analysis showing proteins of ADRN-markers (PHOX2A, PHOX2B, DBH, and TFAP2B), MES-markers (FN1, SNAI2, YAP1, and SOX9) and members of the NOTCH pathway (JAG1, NOTCH1, NOTCH2, NOTCH3, MAML2, and HES1) in SH-SY5Y with inducible overexpression of NOTCH3-IC. Time of NOTCH3-IC induction is indicated above the western blots. Protein lysates from 691-MES and 691-ADRN cells are loaded on the same blot as a reference for MES and ADRN states, respectively. ß-actin was used as loading control. dox, doxycycline. Source data are provided as a Source Data file

Fig 3: NOTCH-IC paralogues induce a reciprocal feed-forward signaling cascade. a SH-SY5Y cells with inducible overexpression of NOTCH3-IC were profiled for gene expression up to 21 days. Z-scores of mRNA expression values are shown for JAG1, NOTCH1, NOTCH2, NOTCH3, MAML2, and HES1. The transcription of NOTCH receptors from endogenous loci is measured using probesets located in the 3'UTRs of NOTCH1 (probeset 218902_at), NOTCH2 (probeset 202443_x_at) and NOTCH3 (probeset 203238_s_at). Notably, the 3'UTR is absent in the NOTCH3-IC transgene. b. Western blot analysis of NOTCH1, NOTCH2, NOTCH3, MAML2, and JAGGED1 proteins in SH-SY5Y with NOTCH1-IC, NOTCH2-ICFLAG or NOTCH3-IC inducible overexpression. ß-actin was used as loading control. Time of induction was 96 h. Source data are provided as a Source Data file. c Western blot analysis of SH-SY5Y cells with (+dox) or without (-dox) NOTCH3-IC expression that were cultured in the presence (+) or the absence (-) of the gamma-secretase inhibitor RO4929097. Protein lysates were analyzed for NOTCH pathway genes (NOTCH1, MAML2, HES1), MES markers (FN1, YAP1, SNAI2) and ADRN markers (PHOX2A, PHOX2B, GATA2, TFAP2B). ß-actin was used as loading control. dox, doxycycline. Source data are provided as a Source Data file. d Analysis of MES and ADRN mRNA signature scores in SH-SY5Y cells with NOTCH3-IC induction that were treated with the gamma-secretase inhibitor RO4929097. Control cells (-dox) treated with DMSO or RO4929097 are shown in dark- and light blue, respectively. NOTCH3-IC expressing cells with DMSO or RO4929097 are shown in red and purple, respectively. Cells were induced with doxycycline and treated with RO4929097 for 14 days. Dashed circles indicate groups of MES- or ADRN-type cell lines as previously determined1

Fig 4: Embryonic corneal defects in E18.5 AP-2ß NCC KO mutant embryos. Coronal sections of E18.5 anterior segments stained using either H&E (A-D) or immunofluorescence for N-cadherin expression (E,F) for controls (A,C,E) or AP-2ß NCC KO mutants (B,D,F). Boxed regions in A and B are shown in greater detail in C and D. Red arrow in F shows that corneal endothelium does not stain for endothelial cell marker N-cadherin in absence of Tfap2b. Blue arrowheads in B and D indicate presence of red blood cells in mutant corneal stroma. CEn, corneal endothelium; aCEn, absent corneal endothelium; CEp, corneal epithelium; CS, corneal stroma; L, lens; Re, retina. Scale bars: 100 µm.

Fig 5: Inactivation of AP-2ß in DCT1s impairs their terminal differentiation and function in the adult.a Lack of AP-2ß function in DCT1s (PvalbCre+Tfap2bfl/fl mice) leads to loss of EGF protein in DCT1s (Pvalb+; yellow arrows), whereas EGF immunoreactivity remains in TALs (NKCC2+; white arrows) that are not targeted by PvalbCre. 16-months-old mice. Scale bars, 50 µm. b Reduced Pvalb protein levels in whole kidney lysates of 12-months-old PvalbCre+Tfap2bfl/fl mice compared to WT littermates. Values show densitometric values for Western blot bands normalized to ß-actin. Size markers are indicated. P value was determined by a two-tailed t test. Graph represents data as mean ± SEM. N = 4 WT mice; n = 5 PvalbCre+Tfap2bfl/fl mice. Source data are provided as a Source Data File. c Immunolabeling shows that PvalbCre+Tfap2bfl/fl mice have normal-appearing TALs (NKCC2) (white arrows). Co-immunolabeling for Pvalb (DCT1 marker) and Calb1 (stronger expressed in CTs than in DCTs) identifies in WT mice CTs (stronger Calb1 staining compared to Pvalb) (blue arrows) and DCT1s (stronger Pvalb staining compared to Calb1) (yellow arrows). In PvalbCre+Tfap2bfl/fl mice immunolabeling for Pvalb is reduced in DCT1s, leading to similar staining intensity for Calb1 and Pvalb in both nephron segments. 16-months-old mice. Scale bars, 100 µm. d Left: EGF immunolabeling is detected in a cytoplasmic localization in TALs (NKCC2+) and DCTs (Pvalb+) in WT mouse kidneys (green arrows). Middle: No or severely diminished EGF immunolabeling is detected in DCTs (Pvalb+ [red]; white arrow) of Six2Cre+Tfap2bfl/WT mice (heterozygous for functional AP-2ß in the kidney), whereas their TALs show EGF immunolabeling (NKCC2+; yellow arrows). Right: Six2Cre+Tfap2bfl/fl mice (null for functional AP-2ß in the kidney) lack DCTs, but their TALs show immunolabeling for EGF (yellow arrow). Scale bars, 100 µm. e Inactivation of Tfap2b in DCT1s (PvalbCre+Tfap2bfl/fl mice) does not lead to an increase in BUN and does not affect body weight. P values were determined by a two-tailed Mann–Whitney test. BUN in mg/dl. Graphs represent data as mean ± SEM. Each dot represents a different mouse. Number of mice per group is indicated in brackets. Source data are provided as a Source Data File. f PvalbCre+Tfap2bfl/fl mice show with progressive age (9-12 months) an increase in 24-hour urine production. P values were determined by a two-tailed Mann–Whitney test. Graphs represent data as mean ± SEM. Each dot represents a different mouse. Number of mice per group is indicated in brackets. Source data are provided as a Source Data File.