CD74-NRG1 Fusions Are Oncogenic In Vivo and Induce Therapeutically Tractable ERBB2:ERBB3 Heterodimerization

Abstract

NRG1 fusions are recurrent somatic genome alterations occurring across several tumor types, including invasive mucinous lung adenocarcinomas and pancreatic ductal adenocarcinomas and are potentially actionable genetic alterations in these cancers. We initially discovered CD74-NRG1 as the first NRG1 fusion in lung adenocarcinomas, and many additional fusion partners have since been identified. Here, we present the first CD74-NRG1 transgenic mouse model and provide evidence that ubiquitous expression of the CD74-NRG1 fusion protein in vivo leads to tumor development at high frequency. Furthermore, we show that ERBB2:ERBB3 heterodimerization is a mechanistic event in transformation by CD74-NRG1 binding physically to ERBB3 and that CD74-NRG1-expressing cells proliferate independent of supplemented NRG1 ligand. Thus, NRG1 gene fusions are recurrent driver oncogenes that cause oncogene dependency. Consistent with these findings, patients with NRG1 fusion-positive cancers respond to therapy targeting the ERBB2:ERBB3 receptors.

Figure 1.

Establishment of a CD74-NRG1 mouse model via homologous recombination at the ROSA26 locus. A, Representation of the primary targeting vector, where the CD74-NRG1 transgene (green) expression is driven by the CAG promotor but inactive due to a STOP cassette (gray). The stop cassette carries a kanamycin selection marker and is flanked by FRT sites (blue triangles), which provide target sites for the Flp recombinase that allow excision of the stop cassette and conditional activation of CD74-NRG1 transgene expression. B, Segregation and transcriptional activity of the CD74-NRG1 transgene in Rosa26^{CD74-NRG1/Flp} mouse embryonic fibroblasts after Flp-mediated recombination. Genomic PCR differentiates a 570-bp ROSA26 wt amplicon and a 380-bp amplicon when the transgene is present; RT-PCR results in a 263-bp amplicon after excision of the stop cassette detects the CD74-NRG1 fusion transcript. C, Immunoblot of the same Rosa26^{CD74-NRG1/Flp} mouse embryonic fibroblasts as in B to confirm protein expression. On top, a loading control stained with anti--HSP90. In the second panel: anti-CD74 staining in green, below an anti-NRG1 stain (red), and at the bottom, both protein determinants multiplexed (merged) resulting in a yellow signal. D, Kaplan–Meier curves illustrating the overall survival in the two transgenic mouse cohorts. Survival is significantly reduced only in the Rosa26^{CD74-NRG1/Flp} cohort (red line) compared with either the control cohort (black line; P = 0.04). Rosa26^CD74-NRG1 mice were used as control, the CD74-NRG1 transgene is present in the Rosa26 locus but is not expressed without the Rosa26^Flp allele (stop cassette not excised in these mice). E, Alternative depiction of D when death in cohorts is recorded only in consequence of tumor development (log-rank P value < 0.001).

Figure 2.

Subcutaneous tumors in Rosa26^{CD74-NRG1/Flp} mice and characterization. A, MRI scans of representative tumors in Rosa26^{CD74-NRG1/Flp} mice scanned on a 3.0T MRI system in a monthly interval. Letters indicate names of organs/tissues (a, spine; b, paravertebral muscle tissue; c, intestine; d, urinary bladder; e, brain; f, ear). Axial T2-weighted images show subcutaneous tumors in three different mice at different locations. The boxes below each primary picture represent magnifications where arrows point to subcutaneous tumor tissue. B, H&E staining of two different subcutaneous tumors from independent Rosa26^{CD74-NRG1/Flp} mice, representative areas are boxed and magnified fourfold below. The white arrow points to a mitotic cell. C, Ki67 staining results of the same tumors as in B and a fourfold magnification of representative areas. Scale bars at the bottom left in B and C represent 39,5 µmol/L. D, Principal component analysis (PCA) of 3′RNA-Seq results comparing normal and tumor tissue of Rosa26^{CD74-NRG1/Flp} mice. Variance stabilizing transformation (VST) was performed on the 1,000 most variably transcribed genes. Each square indicates a tumor sample, each dot represents a normal tissue sample. First principal component (PC1) on the x-axis explains 61% of the variance and is plotted against second principle component (PC2) on the y-axis that accounts for 20% variance among samples. Colors indicate tissue origin of the sample (red = liver, yellow = lung, green = muscle, blue = subcutaneous sarcoma, purple = spleen). E, Transcript levels of Erbb2 and Erbb3 estimated by 3′RNA-Seq in sarcoma tissue compared normal muscle tissue of Rosa26^{CD74-NRG1/Flp} mice are plotted for seven animals. Dots on the y-axis represent gene expression values [log2 (CPM+1)].

Figure 3.

CD74-NRG1 expressing cells are sensitive to ERBB2 and ERBB3 inhibitors. A, Proliferation of BaF3 ERBB2 2YF/ERBB3 wt cells over time after transduction with pBabe-hygro-CD74-NRG1 or pBabe-hygro empty vector (time in days plotted on the x-axis, cell count in cells/mL on the y-axis). The color code to the left of the graph indicates whether the pBabe-hygro empty vector control (e.v.) or the pBabe-hygro-CD74-NRG1 vector were used for cell transduction and whether recombinant NRG1 was supplemented or not. Note that in the absence of NRG1 BaF3 ERBB2 2YF/ERBB3 wt cell proliferation strictly depends on NRG1 (compare green squares to dark gray dots), whereas CD74-NRG1 expression of (ERBB2-YF/ERBB3 wt CD74-NRG1) releases this requirement (compare yellow to red triangle curve). B–D, Viability screening of BaF3 ERBB2 2YF/ERBB3 wt cells treated with afatinib (B), lapatinib (C), and Compound 3 (D) for 96 hours (n = 3). Each graph compares cell viability in nontransduced BaF3 ERBB2 2YF/ERBB3 wt cells (gray), to cells transduced with the empty pBabe-hygro (green) or the pBabe-hygro-CD74-NRG1 vector (red and yellow). Note that CD74-NRG1 cells (yellow curves) proliferate independently of NRG1, whereas BaF3 ERBB2 2YF/ERBB3 wt cells not expressing CD74-NRG1 require exogenous NRG1 supply. In each experimental set-up, every inhibitor efficiently reduces cell proliferation in pBabe-hygro-CD74-NRG1 transduced cells that are not supplemented with NRG1. E, Immunoblot analysis of BaF3 ERBB2 2YF/ERBB3 wt transduced with either empty vector or a construct encoding CD74-NRG1 were treated with afatinib and Compound 3 for effects on downstream activation of HER3. Cells were treated for 6 hours. Inhibitor concentrations are depicted on the top, antibodies used to evaluate downstream activation of ERBB3 are shown on the left side. Actin was used as a loading control.

Figure 4.

Evidence for autocrine and paracrine CD74-NRG1 signaling. A, Schematic representation of cells overexpressing ERBB2, ERBB3, and CD74-NRG1 with different resistance cassettes in all individual cells. (paracrine and autocrine interaction). B and C, Immunoblots of ERBB3 (B) and CD74 (C) co-immunoprecipitations with lysates from NIH-3T3 cells overexpressing combinations of ERBB2, ERBB3, CD74-NRG1, CD74-NRG1_del, or empty vector (ev). D, An experiment in which cells are mixed 1:1 overexpressing ERBB2 and ERBB3 and cells overexpressing CD74-NRG1 with different resistance cassettes in all individual cells (paracrine interaction only). E and F, Immunoblots of ERBB3 (B) and CD74 (C) co-immunoprecipitations with lysates from 1:1 mixed NIH-3T3 cells overexpressing ERBB2 and ERBB3 and cells overexpressing CD74-NRG1 or CD74-NRG1_del.