TNFR2 is required for RIP1-dependent cell death in human leukemia

Abstract

Despite major advances in the treatment of patients with acute lymphoblastic leukemia in the last decades, refractory and/or relapsed disease remains a clinical challenge, and relapsed leukemia patients have an exceedingly dismal prognosis. Dysregulation of apoptotic cell death pathways is a leading cause of drug resistance; thus, alternative cell death mechanisms, such as necroptosis, represent an appealing target for the treatment of high-risk malignancies. We and other investigators have shown that activation of receptor interacting protein kinase 1 (RIP1)-dependent apoptosis and necroptosis by second mitochondria derived activator of caspase mimetics (SMs) is an attractive antileukemic strategy not currently exploited by standard chemotherapy. However, the underlying molecular mechanisms that determine sensitivity to SMs have remained elusive. We show that tumor necrosis factor receptor 2 (TNFR2) messenger RNA expression correlates with sensitivity to SMs in primary human leukemia. Functional genetic experiments using clustered regularly interspaced short palindromic repeats/Cas9 demonstrate that TNFR2 and TNFR1, but not the ligand TNF-α, are essential for the response to SMs, revealing a ligand-independent interplay between TNFR1 and TNFR2 in the induction of RIP1-dependent cell death. Further potential TNFR ligands, such as lymphotoxins, were not required for SM sensitivity. Instead, TNFR2 promotes the formation of a RIP1/TNFR1-containing death signaling complex that induces RIP1 phosphorylation and RIP1-dependent apoptosis and necroptosis. Our data reveal an alternative paradigm for TNFR2 function in cell death signaling and provide a rationale to develop strategies for the identification of leukemias with vulnerability to RIP1-dependent cell death for tailored therapeutic interventions.

Graphical abstract

Figure 1.

TNFR2 predicts SM sensitivity. (A) Sensitivity of primary ALL samples in different risk categories, according to risk stratification in the AIEOP-BFM 2000 study, to birinapant. Numbers are cases with IC₅₀ < 100 nM and the total analyzed cases. TNFR2 (TNFRRSF1B) expression from 17 primary samples correlated with the response (IC₅₀) to birinapant (B) and LCL161 (C). TNFR2 (D) and TNFR1 (E) expression by quantitative real-time PCR from 12 responsive PDX samples (birinapant IC₅₀ ≤ 100 nM) and 22 nonresponsive PDX samples (birinapant IC₅₀ > 100 nM). MR/HR, medium risk/high risk; SR, standard risk; VHR, very high risk.

Figure 2.

Validation of association between TNFR2 and sensitivity to birinapant in an independent cohort. (A) Expression of TNFR1 and TNFR2 from the validation cohort (n = 44), with selected patients marked in blue (high TNFR2, n = 5) and red (low TNFR2, n = 5). (B) Response to birinapant for selected patients (n = 10). (C) Birinapant IC₅₀ values. (D) Rescue from birinapant (48 hours, 50 nM) with zVAD (25 μM) and/or Nec-1s (25 μM). All quantifications are mean ± SEM and were derived from 1 experiment in triplicate for the indicated number of primary samples.

Figure 3.

TNFR1 and TNFR2 are required for birinapant response. (A) Schematic diagram of CRISPR in vivo selection. Total engraftment (purple) and TNFR2ko BFP-positive (B) or TNFR1ko mCherry-positive (C) cells over the total engraftment for R-03 PDX cells. Data are representative of 3 mice per experiment. (D) In vitro birinapant response of WT, TNFR2ko, and TNFR1ko R-03 cells treated for 48 hours. (E) In vitro birinapant response after reconstitution of TNFR2 expression in TNFR2ko ALL cells. The same experiments were performed for PDX samples VHR-10 (E-H) and SR-13 (G-J), with 1 mouse per experiment (E-F,H-I). In vitro data (D,H,K) are derived from 3 independent experiments in triplicates; data are mean ± SEM. R, relapse.

Figure 4.

TNF-α is not required for birinapant response. (A) TNF-α expression by quantitative real-time PCR from 12 responsive PDX samples (birinapant IC₅₀ ≤ 100 nM) and 22 nonresponsive PDX samples (birinapant IC₅₀ > 100 nM). (B) Quantification of TNF-α by ELISA from cell lysates in WT and CRISPR-generated TNFko PDXs (3 independent experiments performed in duplicates). (C) Validation of TNFko MSCs by copy number variant quantitative PCR (2 experiments performed in triplicates). (D) Birinapant response curves for WT and TNFko PDX cells cocultured with TNFko MSCs (3 independent experiments performed in triplicates). (E) Gene expression of death receptor ligands and receptors for 17 primary B-ALL samples. (F) Birinapant response curves for PDX samples treated with neutralizing antibodies against lymphotoxin (LT; 0.5 μg/ml), TRAIL (10 ng/mL), and FasL (10 ng/mL) (3 independent experiments performed in triplicates). All quantifications are mean ± SEM. R, relapse.

Figure 5.

TNFR2 promotes recruitment of RIP1 to TNFR1. (A) Endogenous TNFR1 immunoprecipitation (IP) in WT and TNFR2ko R-03 (responder) PDX cells treated with birinapant, as indicated. (B) Lysates and TNFR1 immunoprecipitation of sensitive (R-03, IC50 < 100 nM) and resistant (VHR-01, IC50 > 1000 nM) ALL. (C) WB in nonreducing nondenaturing conditions for TNFR1 monomers (mono), dimers, and trimers (Tri) in responder and nonresponder PDX (left panel), and WT and TNFR2ko R-03 PDX cells (right panel) treated as indicated. (D) Analysis of RIP1 phosphorylation at serine-166 (pRIP1) in WT and TNFR1ko ALL samples. (E) Lysates and caspase-8 (Casp8) IP (ripoptosome) for WT and TNFR2ko (R2ko) R-03 and VHR-10 PDX cells treated with birinapant, as indicated. Data are representative of 3 independent experiments. Bir/bir, birinapant; 10′, 10 minutes; 30′, 30 minutes.