Fig 2: Nfkbid is strongly regulated by Roquin on the mRNA and protein level. a–c Immunoblot (a), flow cytometry (b) and quantitative RT-PCR (c) analysis of previously described Roquin targets in CD4+ T cells derived from Rc3h1-2fl/fl; Cd4-Cre-ERT2 (iDKO) mice treated in vitro with 4' OH-tamoxifen. As wild-type (WT) controls either T cells from Rc3h1-2+/+; Cd4-Cre-ERT2 (WT) mice were treated in vitro with 4' OH-tamoxifen or T cells from Rc3h1-2fl/fl; Cd4-Cre-ERT2 mice were left untreated. d Immunoblot analysis of Roquin-1 and Roquin-2 proteins before and after 4’ OH-tamoxifen treatment of MEF cells that contained Rc3h1 and Rc3h2 alleles with loxP-flanked exons and a Cre-ERT2 transgene. e Flow cytometry analysis of GFP in Rc3h1–2fl/fl, Cre-ERT2 MEF cells retrovirally transduced with a reporter construct encoding GFP and human ICOS coding sequence (ICOS–CDS) followed by the full length Nfkbid 3'-UTR (1–559) harboring two CDE, one ADE and one LBE element or truncated versions (scheme) of the 3'-UTR, either treated with 4' OH-tamoxifen (iDKO) or left untreated (WT). Statistical significance was calculated by unpaired two-tailed Student’s t test (b, c); ns = not significant; ***p < 0.001, **p < 0.01. Error bars indicate mean ± SEM. Data are representative of three (a, b, d, e) and seven (c) independent experiments

Fig 3: Nfkbid regulation by Roquin redundantly involves decapping and deadenylation. a Schematic representation of ICOS reporter constructs with or without 5' cap (m7G) and poly(A) tail (polyA). (b) Flow cytometry analysis of ICOS reporter expression in Rc3h1–2-/- MEF cells containing a doxycycline-inducible cassette for re-expression of Roquin-1. Rc3h1–2-/- MEF cells were treated with doxycycline (+dox) for 16–20 h or left untreated (-dox) and electroporated with the different reporter constructs as indicated in (a). c, e, g, i Schematic representation of reporter constructs composed of the ICOS–CDS fused to the 3'-UTR of Nfkbid. The 5' end of each reporter construct either contained a 5' cap (m7G), ARCA (anti-reverse cap), S-D1, S-D2, or BH3 cap analogs (c, e, g, i). The 3' end either contained a poly(A) tail consisting of 146 adenine bases (polyA146) (c, e, g, i) or a polyA146 tail modified by the addition of ten cytosines (C10) to block deadenylation (c, g, i). d, f, h Flow cytometry analysis of ICOS reporter expression in Rc3h1–2-/- MEF cells containing a doxycycline-inducible cassette for re-expression of Roquin-1. Rc3h1–2-/- MEF cells were treated with doxycycline (+dox) for 16–20 h or left untreated (-dox) and electroporated with the different reporter constructs as indicated in (c, e, g). Black lines represent nonelectroporated cells. j Quantification of ICOS reporter expression using flow cytometry in Rc3h1–2-/- MEF cells transfected with Roquin relative to cells without Roquin transfection. These cells contain a doxycycline-inducible shRNA against cnot1 and were treated with doxycycline for 3 days before electroporation with the different reporter constructs as indicated in (i). cNOT1 knock-down (KD) samples represent only GFP positive and therefore shRNA expressing cells. k Immunoblot analysis of knockdown of cNOT1 expression in MEF cells containing a doxycycline-inducible shRNA against cnot1. Only GFP+ cells express the shRNA. A nontargeting (NT) shRNA was used as a control. Statistical significance was determined by unpaired two-tailed Student’s t test (j), ns = not significant; *p < 0.05; **p < 0.01. Error bars represent three technical replicates (mean ± SEM). Data are representative of one (b, k), seven (d), two (h, j), and three (f) independent experiments

Fig 4: High Tact/Treg ratios after prostate GVAX/ipilimumab therapy are associated with significantly longer PFS following mitoxantrone. Kaplan–Meier curves of progression-free survival (PFS) following (A) mitoxantrone or (B) docetaxel treatment for prostate GVAX/ipilimumab-treated patients with high vs. low CD8+ICOS+/Treg ratios. Number of patients and corresponding median PFS for each group is given. Statistical significance of the survival distribution was analyzed by log-rank testing.

Fig 5: Nfkbid regulation requires Roquin to interact with at least three stem–loops. a Nucleotide sequences of the loop structures of wild-type mouse Nfkbid 3'-UTR stem–loops (SL) 1, 2, and 5 (gray) and introduced loop exchange mutations (LE: change of hexa- to tri-loops and vice versa, blue) are depicted. b, c Flow cytometry analysis of GFP in Rc3h1–2fl/fl; Cre-ERT2 MEF cells retrovirally transduced with the ICOS–GFP-Nfkbid 3'-UTR (1–263) reporter containing LE mutations (b) and individual or combined loop mutations (c) in the indicated stem–loops as described in Fig. 3b. MEF cells were either treated with 4' OH-tamoxifen (iDKO) or left untreated (WT). MFI of GFP is shown for WT (red) and iDKO (blue) MEF cells. (d) Flow cytometry analysis of GFP in the ?N-boxB-tethering system in Rc3h1–2-/- MEF cells stably expressing the reverse tetracycline-controlled transactivator rtTA3. MEF cells were first either transduced with a retrovirus containing the ICOS–GFP reporter with the wildtype Nfkbid 3'-UTR (1–263) or mutants containing one (at position of SL1, SL2, or SL5) or three (at the positions of SL1, SL2, and SL5) boxB structures. The cells were additionally transduced with a retrovirus encoding for ?N-p2A-mCherry or ?N-Roquin(K220A, K239A, and R260A)-p2A-mCherry containing a doxycycline-inducible cassette. Overexpression of the ?N-constructs was induced by doxycycline administration for 14 h. MFI of GFP for WT (red) and iDKO (blue) MEF cells are indicated. e Fold-change of GFP MFI levels in cells as described in (d) with doxycycline-induced overexpression of ?N-Roquin (K220A, K239A, and R260A)-p2A-mCherry, normalized to mean of wild-type Nfkbid (1–263). Error bars indicate mean ± SEM. Data are representative of six (b) and three (c–e) independent experiments