Advertisement
Trends in Cancer
This journal offers authors two options (open access or subscription) to publish research

NRG1 and NRG2 fusion positive solid tumor malignancies: a paradigm of ligand-fusion oncogenesis

  • Misako Nagasaka
    Affiliations
    Chao Family Comprehensive Cancer Center, Orange, CA 92868, USA

    University of California Irvine Susan and Henry Samueli College of Health Sciences, University of California Irvine School of Medicine, Department of Medicine, Division of Hematology-Oncology, Orange, CA 92868, USA

    Division of Neurology, Department of Medicine, St. Marianna University School of Medicine, Kawasaki, Kanagawa, Japan
    Search for articles by this author
  • Sai-Hong Ignatius Ou
    Correspondence
    Correspondence:
    Affiliations
    Chao Family Comprehensive Cancer Center, Orange, CA 92868, USA

    University of California Irvine Susan and Henry Samueli College of Health Sciences, University of California Irvine School of Medicine, Department of Medicine, Division of Hematology-Oncology, Orange, CA 92868, USA
    Search for articles by this author
Open AccessPublished:January 04, 2022DOI:https://doi.org/10.1016/j.trecan.2021.11.003

      Highlights

      • A quarter of a century (1997–2021) has elapsed from the first report of neregulin-1 fusion (NRG1+) in a breast cancer cell line, to the first report from an international prospective trial demonstrating preliminary clinical efficacy of an anti-HER2/HER3 bispecific monoclonal antibody in NRG1 solid tumors. These results provided the first proof of principle that NRG1 fusions represent a novel class of ‘ligand-fusion’-driven oncogenesis.
      • Two isoforms (α and β) of the EGF-like domain within NRG1 ligand exist and are generated by differential exon splicing confer differential binding affinities of NRG1 to HER receptor family members with the β-isoform having higher affinity to HER3 than the α-isoform. The ratio of β-/α- isoforms potentially could modulate response and resistance to treatment.
      • NRG1 fusions have been identified in multiple solid organs but are extremely rare, with an incidence of ~0.5%. RNA NGS is likely the optimal method to detect these rare fusions and ensure they are transcribed and in-frame. Immunohistochemistry of HER3 phosphorylation may be used as an initial screening method to detect NRG1+ solid tumors. Histology [invasive mucinous adenocarcinoma (IMA) for NSCLC] or co-genomic alterations (KRAS-wild type) can also augment initial screening. Several screening strategies have been proposed.
      • Recently, NRG2 fusion positive solid tumors have been identified at an even lower frequency than NRG1+ solid tumors. It remains to be determined whether NRG2 fusions are oncogenic, as few compounds exist to target the NRG2/HER4 axis.
      Neuregulins (NRGs) are a family of six related physiological ligands all containing a receptor-binding epidermal growth factor (EGF)-like domain that mediate their binding to cellular receptors. Neuregulin-1 (NRG1) is the main physiological ligand to HER3. NRG1 fusion (NRG1+) was first reported in a breast cancer cell line and NRG2 fusions have recently been identified in solid tumors. It is postulated that NRG1 fusions, through mostly transmembrane fusion partners, result in NRG1 being concentrated in proximity to HER3, leading to its constitutive activation and oncogenesis. Recently, a monoclonal antibody that disrupts the binding of NRG1 to HER3 and HER3/HER2 heterodimerization has resulted in NRG1+ tumor shrinkage, suggesting that ‘ligand-fusion’ may be a novel mechanism of oncogenesis.

      Keywords

      Introduction

      The discovery of chromosomal rearrangements involving receptor tyrosine kinases (RTKs) across different cancer types has stimulated interest in oncogenic fusions as potential therapeutic targets. RTK fusions as exemplified by ALK, ROS1, RET, FGFR, and NTRK involve chromosomal rearrangements that result in the formation of chimeric constituitive fusion protein kinases capable of oncogenic transformation. Multiple tyrosine kinase inhibitors (TKIs) have now been approved for treatment against various solid malignancies harboring these RTK fusions [
      • Shaw A.T.
      • et al.
      Tyrosine kinase gene rearrangements in epithelial malignancies.
      ,
      • Schram A.M.
      • et al.
      Fusions in solid tumours: diagnostic strategies, targeted therapy, and acquired resistance.
      ]. There are 58 human RTKs with multiple ligands interacting with these RTKs to effect cellular signaling pathways [
      • Blume-Jensen P.
      • Hunter T.
      Oncogenic kinase signaling.
      ].
      Though initially described in a breast cancer cell line [
      • Schaefer G.
      • et al.
      Gamma-heregulin: a novel heregulin isoform that is an autocrine growth factor for the human breast cancer cell line, MDA-MB-175.
      ], neuregulin-1 (NRG1) fusions (NRG1+) have recently been identified in many solid tumors at an extremely rare frequency (~0.1% to 0.3%) [
      • Jonna S.
      • et al.
      Detection of NRG1 gene fusions in solid tumors.
      ]. NRG1 fusions are postulated to result in an artificially high concentration of NRG1 near its primary receptor human epidermal growth factor receptor 3 (HER3). Since most fusion partners of NRG1 fusions contain a transmembrane domain, these chimeric NRG1 fusion proteins essentially 'trap' NRG1 ligand with its intact epidermal growth factor (EGF)-like RTK-binding domain in the local cellular milieu resulting in a 'super-biological' level of ligand in close proximity to HER3 leading to functionally constitutive activation of the HER3 receptor pathway [
      • Fernandez-Cuesta L.
      • Thomas R.K.
      Molecular pathways: targeting NRG1 fusions in lung cancer.
      ,
      • Dimou A.
      • Camidge D.R.
      Detection of NRG1 fusions in solid tumors: rare gold?.
      ]. Subsequently NRG1 breakpoints found in breast, pancreatic, and squamous cell carcinoma of the lung further suggest that NRG1 rearrangements play a role in oncogenesis and are not mere chromosomal artifacts of multigeneration passaged tumor cell lines. Indeed, years later clinical activity of monoclonal antibodies that disrupt the binding of NRG1 to HER3 and/or heterodimerization of HER3/HER2 have resulted in shrinkage of tumors harboring NRG1 fusions [
      • Schram A.M.
      • et al.
      Efficacy and safety of zenocutuzumab in advanced pancreas cancer and other solid tumors harboring NRG1 fusions.
      ]. See Figure I of Box 1. While large-scale global clinical trials of these monoclonal antibodies against HER3 in NRG1 fusions are ongoing, NRG1 fusion is emerging as a novel actionable molecular alteration in solid tumors. Indirectly supporting this conclusion is the recent discovery of an even rarer neuregulin-2 fusion (NRG2+) [
      • Kohsaka S.
      • et al.
      Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
      ,
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ], where the normal physiological function of NRG2 remains to be fully elucidated.
      Discovery of NRG1 and NRG2 fusions in solid tumors
      The first evidence that NRG1 fusions could serve as a driver mutation in solid malignancies was reported in 1997, when NRG1 containing unique amino acids in the 5′ portion of NRG1 (named at the time as γ-heregulin) was identified from a breast cancer cell line (MDA-MB-175) [
      • Schaefer G.
      • et al.
      Gamma-heregulin: a novel heregulin isoform that is an autocrine growth factor for the human breast cancer cell line, MDA-MB-175.
      ] (Figure I). This discovery was followed by the identification of DOC4-NRG1 (now TENM4-NRG1) as the novel NRG1 fusion variant from the MDA-MB-175 cell line 2 years later [
      • Liu X.
      • et al.
      Gamma-heregulin: a fusion gene of DOC-4 and neuregulin-1 derived from a chromosome translocation.
      ]. This DOC4-NRG1 fusion was generated by chromosomal translocation between 5′ fusion partner gene DOC4 and 3′ NRG1 gene that contains the EGF-like domain identical to the mechanism in generating the RTK fusions [
      • Shaw A.T.
      • et al.
      Tyrosine kinase gene rearrangements in epithelial malignancies.
      ,
      • Schram A.M.
      • et al.
      Fusions in solid tumours: diagnostic strategies, targeted therapy, and acquired resistance.
      ]. In-depth whole transcriptome sequencing (WTS) in 2021 indicated that the DOC-NRG1 fusion variant harbored a more complex rearrangement as PPP6R3-TENM4-NRG1 (PPP6R3-DOC4-NRG1) [
      • Howarth K.D.
      • et al.
      NRG1 fusions in breast cancer.
      ]. In 2003, the chromosomal breakpoints spreading over 1.1 Mb of NRG1 were identified in four additional breast cancer cell lines (ZR-75-1, HCC1937, SUM-52, UACC-812) and two pancreatic cancer cell lines (PaTu I and SUIT-2), providing further evidence that NRG1 breakpoints could be a recurring phenomenon in solid tumors [
      • Adelaide J.
      • et al.
      A recurrent chromosome translocation breakpoint in breast and pancreatic cancer cell lines targets the neuregulin/NRG1 gene.
      ]. One year later in 2004, a survey of tumor cell samples extended the finding that NRG1 breakpoints were identified in tumor samples of breast, pancreatic, and squamous cell carcinoma of the lung, further cementing evidence that NRG1 rearrangements play a role in oncogenesis and are not mere chromosomal artifacts of multigeneration passaged tumor cell lines [
      • Huang H.E.
      • et al.
      A recurrent chromosome breakpoint in breast cancer at the NRG1/neuregulin 1/heregulin gene.
      ]. Then came an approximate 10-year gap in the literature on NRG1 fusion in malignancies from 2004 until 2014, when five groups independently and with various approaches identified NRG1 fusions in NSCLC [
      • Fernandez-Cuesta L.
      • et al.
      CD74-NRG1 fusions in lung adenocarcinoma.
      ,
      • Nakaoku T.
      • et al.
      Druggable oncogene fusions in invasive mucinous lung adenocarcinoma.
      ,
      • Gow C.H.
      • et al.
      Multidriver mutation analysis in pulmonary mucinous adenocarcinoma in Taiwan: identification of a rare CD74-NRG1 translocation case.
      ,
      • Dhanasekaran S.M.
      • et al.
      Transcriptome meta-analysis of lung cancer reveals recurrent aberrations in NRG1 and Hippo pathway genes.
      ,
      • Zheng Z.
      • et al.
      Anchored multiplex PCR for targeted next-generation sequencing.
      ]. Since then, there has been a rapid increase in the knowledge of biology of NRG1 fusions, especially in NSCLC [
      • Nagasaka M.
      • Ou S.I.
      Neuregulin 1 fusion-positive NSCLC.
      ,
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ,
      • Drilon A.
      • et al.
      Response to ERBB3-directed targeted therapy in NRG1-rearranged cancers.
      ] and pancreatic ductal adenocarcinoma (PDAC) [
      • Drilon A.
      • et al.
      Response to ERBB3-directed targeted therapy in NRG1-rearranged cancers.
      ] (Figure I).
      In 2020, a single case of CD74-NRG2α was reported from whole transcriptome profiling of Japanese patients with lung adenocarcinoma who were light/never-smokers (Figure I) [
      • Kohsaka S.
      • et al.
      Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
      ]. Although CD74-NRG2α transcript was highly overexpressed, only HER4 was phosphorylated by CD74-NRGα, as detected by IHC, consistent with HER4 as the primary receptor for NRG2. In contrast, in the three cases of CD74-NRG1, all four members of the HER family (EGFR, HER2, HER3, HER4) were phosphorylated, indicating further that NRG1/HER3 is more active than the NRG2/HER4 signaling pathway. Subsequently, a pan-tumor survey was performed in approximately 54 000 solid tumors and identified seven NRG2 fusions among NSCLC (2/9600, 0.021%), endometrial (2/2600, 0.065%), ovarian (1/5030, 0.020%), and prostate cancer (1/1600, 0.065%) and adenocarcinoma of unknown primary by WTS [
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ] (Figure I). The same F11R-NRG2 fusion identified in NSCLC was also found at 0.5% (1/200) in cases of invasive mucinous adenocarcinoma (IMA) [
      • Chang J.C.
      • et al.
      Comprehensive molecular and clinicopathologic analysis of 200 pulmonary invasive mucinous adenocarcinomas identifies distinct characteristics of molecular subtypes.
      ]. A total of seven NRG1 fusions were identified in the same study, indicating NRG2 fusions are even rarer than NRG1 fusions [
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ,
      • Chang J.C.
      • et al.
      Comprehensive molecular and clinicopathologic analysis of 200 pulmonary invasive mucinous adenocarcinomas identifies distinct characteristics of molecular subtypes.
      ]. All known NRG2 fusions identified to date contain the EGFR-like α-isoform domain, a theoretically lower binding affinity ligand than the β-isoform of NRG2 [
      • Kohsaka S.
      • et al.
      Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
      ,
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ].
      Figure I
      Figure ITimeline of the milestones of NRG1 and NRG2 fusions in solid malignancies.
      Abbreviations: HER, human epidermal growth factor receptor; mAb, monoclonal antibody; NRG, neuregulin; NSCLC, non-small cell lung cancer; SqCC, squamous cell carcinoma; wt, wild type.
      Here, we review the biology of NRG1 and NRG2 pertaining to NRG1 and NRG2 fusions, summarize known clinicopathologic and molecular characteristics of specific NRG1+ and NRG2+ solid tumors, propose screening strategies for NRG1+ and NRG2+ solid tumors with considerations given to resource constraints in many countries, and summarize currently ongoing clinical trials in NRG1+ and NRG2+ solid tumors.

      Functions of NRG1 and NRG2

      NRG family

      NRGs are a complex family of structurally related cellular growth factors encoded by six closely related genes that are involved primarily in the development of the nervous and cardiovascular systems [
      • Mei L.
      • Nave K.A.
      Neuregulin-ERBB signaling in the nervous system and neuropsychiatric diseases.
      ,
      • Holmes W.E.
      • et al.
      Identification of heregulin, a specific activator of p185erbB2.
      ,
      • Wen D.
      • et al.
      Neu differentiation factor: a transmembrane glycoprotein containing an EGF domain and an immunoglobulin homology unit.
      ,
      • Goodearl A.D.
      • et al.
      Purification of multiple forms of glial growth factor.
      ,
      • Marchionni M.A.
      • et al.
      Glial growth factors are alternatively spliced erbB2 ligands expressed in the nervous system.
      ,
      • Falls D.L.
      • et al.
      ARIA, a protein that stimulates acetylcholine receptor synthesis, is a member of the neu ligand family.
      ,
      • Ho W.H.
      • et al.
      Sensory and motor neuron–derived factor.
      ,
      • Busfield S.J.
      • et al.
      Characterization of a neuregulin-related gene, Don-1, that is highly expressed in restricted regions of the cerebellum and hippocampus.
      ,
      • Chang H.
      • et al.
      Ligands for ErbB-family receptors encoded by a neuregulin-like gene.
      ,
      • Carraway 3rd, K.L.
      • et al.
      Neuregulin-2, a new ligand of ErbB3/ErbB4-receptor tyrosine kinases.
      ,
      • Higashiyama S.
      • et al.
      A novel brain-derived member of the epidermal growth factor family that interacts with ErbB3 and ErbB4.
      ,
      • Zhang D.
      • et al.
      Neuregulin-3 (NRG3): a novel neural tissue-enriched protein that binds and activates ErbB4.
      ,
      • Harari D.
      • et al.
      Neuregulin-4: a novel growth factor that acts through the ErbB-4 receptor tyrosine kinase.
      ,
      • Kanemoto N.
      • et al.
      Expression of TMEFF1 mRNA in the mouse central nervous system: precise examination and comparative studies of TMEFF1 and TMEFF2.
      ,
      • Kinugasa Y.
      • et al.
      Neuroglycan C, a novel member of the neuregulin family.
      ] (Table 1). Importantly, all six NRG family members share an epidermal growth factor (EGF)-like domain of about 65 amino acids, which is responsible for binding to the HER RTK family members (EGFR, HER2, HER3, and HER4) (Figure 1A ). The EGF-like domain is encoded by two exons alternatively spliced together [
      • Falls D.L.
      Neuregulins: functions, forms, and signaling strategies.
      ].
      Table 1Characteristics of NRG gene and canonical protein isoform
      NRG geneNameOther namesChromosomal locationCharacterized transcripts
      As reported in the Ensembl database (https://www.ensembl.org/index.html).
      Total/Coding Exons
      Coding exon for the canonical form of the protein.
      TranscriptCCDS ID
      CCDS database (https://www.ncbi.nlm.nih.gov/projects/CCDS/CcdsBrowse.cgi).
      Protein nameUniProtKD identifier
      As described in the UniProt Knowledgebase (https://www.uniprot.org).
      Amino acidsGenomic span (coding exons), kb
      National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov).
      NRG1Neuregulin-1-Heregulin (HRG) [
      • Holmes W.E.
      • et al.
      Identification of heregulin, a specific activator of p185erbB2.
      ]

      -Neu differentiation factor (NDF) [
      • Wen D.
      • et al.
      Neu differentiation factor: a transmembrane glycoprotein containing an EGF domain and an immunoglobulin homology unit.
      ]

      -Glial growth factor (GGF) [
      • Goodearl A.D.
      • et al.
      Purification of multiple forms of glial growth factor.
      ,
      • Marchionni M.A.
      • et al.
      Glial growth factors are alternatively spliced erbB2 ligands expressed in the nervous system.
      ]

      -Acetylcholine receptor-inducing activity (ARIA) [
      • Falls D.L.
      • et al.
      ARIA, a protein that stimulates acetylcholine receptor synthesis, is a member of the neu ligand family.
      ]

      -Sensory and motor neuron-derived differentiation -factor (SMDF) [
      • Ho W.H.
      • et al.
      Sensory and motor neuron–derived factor.
      ]
      8p124719/12NM_013957.5CCDS6084.1Neuregulin 1Q022976371216.4
      NRG2Neuregulin-2-Divergent of neuregulin 1 (Don-1) [
      • Busfield S.J.
      • et al.
      Characterization of a neuregulin-related gene, Don-1, that is highly expressed in restricted regions of the cerebellum and hippocampus.
      ]

      -Neuregulin-2 [
      • Chang H.
      • et al.
      Ligands for ErbB-family receptors encoded by a neuregulin-like gene.
      ,
      • Carraway 3rd, K.L.
      • et al.
      Neuregulin-2, a new ligand of ErbB3/ErbB4-receptor tyrosine kinases.
      ]

      -Neural- and thymus-derived activator for ErbB kinases (NTAK) [
      • Higashiyama S.
      • et al.
      A novel brain-derived member of the epidermal growth factor family that interacts with ErbB3 and ErbB4.
      ]
      5q13.2915/10NM_004883.3CCDS4217.1Neuregulin 2Q14511850196.5
      NRG3Neuregulin-3-Neuregulin-3 [
      • Zhang D.
      • et al.
      Neuregulin-3 (NRG3): a novel neural tissue-enriched protein that binds and activates ErbB4.
      ]
      10q23.1921/9NM_001010848.4CCDS31233.1Neuregulin 3P569756961112.0
      NRG4Neuregulin-4-Heregulin-4 (HRG4) [
      • Harari D.
      • et al.
      Neuregulin-4: a novel growth factor that acts through the ErbB-4 receptor tyrosine kinase.
      ]
      15q24.21614/5NM_138573.4CCDS10288.1Neuregulin 4Q8WWG1115123.8
      NRG5-Tomoregulin-1 [
      • Mei L.
      • Nave K.A.
      Neuregulin-ERBB signaling in the nervous system and neuropsychiatric diseases.
      ]

      -transmembrane protein with EGF like and two follistatin-like domains 1 (TMEFF1) [
      • Kanemoto N.
      • et al.
      Expression of TMEFF1 mRNA in the mouse central nervous system: precise examination and comparative studies of TMEFF1 and TMEFF2.
      ]
      9q31.1110/10NM_003692.5CCDS6750.1tomoregulin-1 or TMEFF1Q8IYR6380104.4
      NRG6-Chondroitin sulfate proteoglycan 5 (CSPG5) [
      • Kinugasa Y.
      • et al.
      Neuroglycan C, a novel member of the neuregulin family.
      ]

      -Neuroglycan C

      -Chicken acidic leucine-rich EGF-like domain-containing brain protein (CALEB) [
      • Mei L.
      • Nave K.A.
      Neuregulin-ERBB signaling in the nervous system and neuropsychiatric diseases.
      ]
      3p21.3156/5NM_006574.4CCDS2757.1CSPG5, Neuroglycan C, chicken acidic leucine-rich EGF-like domain-containing brain protein (CALEB)O9519653918.6
      a As reported in the Ensembl database (https://www.ensembl.org/index.html).
      b Coding exon for the canonical form of the protein.
      d As described in the UniProt Knowledgebase (https://www.uniprot.org).
      e National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov).
      Figure 1
      Figure 1Structure of the epidermal growth factor (EGF)-like domains of neuregulin (NRG) family members and NRG1 gene.
      (A) Alignment of the amino acids of EGF-like domains of the six family members of NRG and EGF. The full length EGF-like domain of all six NRG family members encoded by two exons are presented and labeled in the diagram. The underlined amino acid represents the start of a new exon. NRG1 is represented by NRG1-β. NRG2 is represented by NRG2-α. The conserved amino acid residues are labeled in red and aligned. The amino acid sequence of EGF is from CCDS3689.1 of the Consensus Coding Sequencing (CCDS) database (https://www.ncbi.nlm.nih.gov/projects/CCDS/CcdsBrowse.cgi). (B) Alignment of amino acids of the α- and β-isoforms of EGF-like domain of NRG1 and NRG2 with reference to EGF. The full-length EGF-like domains for NRG1 and NRG2 (encoded by two exons) are presented in the diagram. A total of 63 amino acids are encoded by the two exons that encode the NRG1 EGF-like domain. A total of 67 amino acids are encoded by the two exons that encode the NRG2 EGF-like domain. The underlined amino acid represents the start of a new exon. NRG1 is represented by NRG1-β, which is the dominant isoform among NRG1 fusion. NRG2 is represented by NRG2-α, which is the dominant NRG2 fusion isoform. EGF is from CCDS3689.1 of the CCDS database. The conserved amino acid residues are labeled in red and aligned. References for the amino acid sequences are from the CCDS project. CCDS database. (C) Structure and pattern of splicing of exons of type I, II, and III NRG1. Modified from reference [
      • Falls D.L.
      Neuregulins: functions, forms, and signaling strategies.
      ], with permission. Abbreviations: CRD, cysteine-rich domain; CTc, cytoplasmic tail domain C terminal of the EGF-like domain; TMc, transmembrane domain C terminal of the EGF-like domain; TMn, transmembrane domain N terminal of the EGF-like domain.
      In NRG1 and NRG2, like all six NRG family members, the EGF-like domain is encoded by a common ‘core’ exon encoding the 5′-end of the EGF-like domain spliced together with one of the two different exons that encodes the 3′-end of the EGF-like domain generating an α- and β-isoform of NRG1 and NRG2, respectively [
      • Falls D.L.
      Neuregulins: functions, forms, and signaling strategies.
      ] (Figure 1B). The critical function of the EGF-like domain of NRG1 is illustrated when NRG1 knockout mice are generated by disruption of the EGF-like domain (neuregulinδEGF-LacZ). The mice are embryonically lethal with malformation of the heart and the nervous system (cranial nerves and myelination defect) as all the NRG1 isoforms will be nonfunctional [
      • Meyer D.
      • Birchmeier C.
      Multiple essential functions of neuregulin in development.
      ].
      Both the α- and β-isoforms of NRG1 and NRG2 have different binding affinities to their primary receptors, HER3 and HER4, respectively, with the β-isoform having a higher binding affinity than the α-isoform [
      • Pinkas-Kramarski R.
      • et al.
      ErbB tyrosine kinases and the two neuregulin families constitute a ligand-receptor network.
      ,
      • Crovello C.S.
      • et al.
      Differential signaling by the epidermal growth factor-like growth factors neuregulin-1 and neuregulin-2.
      ,
      • Jones J.T.
      • et al.
      Binding specificities and affinities of EGF domains for ErbB receptors.
      ,
      • Sweeny C.
      • et al.
      Ligand discrimination in signaling through an Erb4 receptor homodimer.
      ,
      • Sweeny C.
      • et al.
      Growth factor specific signaling pathway stimulation and gene expression mediated by ErB receptors.
      ]. NRG1α and NRG1β exist physiologically in the human body but the ratio is unknown [
      • Falls D.L.
      Neuregulins: functions, forms, and signaling strategies.
      ]. The difference in binding affinities may also contribute to the relative importance of NRG1β over NRG1α isoform. NRG1β knockout mice were embryonically lethal due to significant maldevelopment of the heart and nervous system [
      • Meyer D.
      • Birchmeier C.
      Multiple essential functions of neuregulin in development.
      ,
      • Kramer R.
      • et al.
      Neuregulins with an Ig-like domain are essential for mouse myocardial and neuronal development.
      ], while NRG1α knockout mice survive to adulthood with only pronounced defects in mammary gland lobuloalveolar development [
      • Li L.
      • et al.
      The breast proto-oncogene, HRG-alpha regulates epithelial proliferation and lobuloalveolar development in the mouse mammary gland.
      ].

      Normal physiological functions of NRG1 and NRG2

      The prototypic NRG1 gene consist of tissue-specific N terminal exons, followed by immunoglobulin-like (Ig-like) domains, and then a common EGF-like domain. However, multiple isoforms of NRG1 protein ligand are generated by alternate splicing [
      • Falls D.L.
      Neuregulins: functions, forms, and signaling strategies.
      ] (Figure 1C). In addition to the EGF-like domain, NRG1 knockout mice generated by the disruption of the Ig-like domain (neuregulinIg) are also embryonically lethal, with malformation of the heart and cranial nerves [
      • Kramer R.
      • et al.
      Neuregulins with an Ig-like domain are essential for mouse myocardial and neuronal development.
      ]. It is also important to note that the NRG1 gene spans a large region of chromosome with large intronic regions [
      • Steinthorsdottir V.
      • et al.
      Multiple novel transcription initiation sites for NRG1.
      ], which has implications for screening strategies of these NRG1 fusions, especially when using DNA next generation sequencing (NGS) as it requires the use of primers hybridizing on the intron regions to capture the translocation breakpoints [
      • Steinthorsdottir V.
      • et al.
      Multiple novel transcription initiation sites for NRG1.
      ,
      • Nagasaka M.
      • Ou S.I.
      Neuregulin 1 fusion-positive NSCLC.
      ]. The mice knockout experiments demonstrated that NRG1 plays a pivotal role in the establishment of the cardiovascular and nervous system in the perinatal period [
      • Meyer D.
      • Birchmeier C.
      Multiple essential functions of neuregulin in development.
      ,
      • Kramer R.
      • et al.
      Neuregulins with an Ig-like domain are essential for mouse myocardial and neuronal development.
      ,
      • Li L.
      • et al.
      The breast proto-oncogene, HRG-alpha regulates epithelial proliferation and lobuloalveolar development in the mouse mammary gland.
      ]. NRG1 as a ligand for HER family members is involved in the development of cardiomyocytes, which has safety implications with therapeutic inhibition of NRG1 [
      • Bersell K.
      • et al.
      Neuregulin1/ErbB4 signaling induces cardiomyocyte proliferation and repair of heart injury.
      ]. Recombinant NRG1 protein is being investigated for the treatment of congestive heart failure [
      • Gao R.
      • et al.
      A phase II, randomized, double-blind, multicenter, based on standard therapy, placebo-controlled study of the efficacy and safety of recombinant human neuregulin-1 in patients with chronic heart failure.
      ,
      • De Keulenaer G.W.
      • et al.
      Mechanisms of the multitasking endothelial protein NRG-1 as a compensatory factor during chronic heart failure.
      ]. Conversely, treatment targeting NRG1 fusion should continually assess for evidence of congestive heart failure like HER2 antibody treatment in HER2+ breast cancer [
      • Bouwer N.I.
      • et al.
      Cardiac monitoring in HER2-positive patients on trastuzumab treatment: a review and implications for clinical practice.
      ].
      The primary receptor for NRG2 is HER4 [
      • Carraway 3rd, K.L.
      • et al.
      Neuregulin-2, a new ligand of ErbB3/ErbB4-receptor tyrosine kinases.
      ,
      • Jones J.T.
      • et al.
      Binding specificities and affinities of EGF domains for ErbB receptors.
      ,
      • Sweeny C.
      • et al.
      Ligand discrimination in signaling through an Erb4 receptor homodimer.
      ,
      • Sweeny C.
      • et al.
      Growth factor specific signaling pathway stimulation and gene expression mediated by ErB receptors.
      ]. The normal physiological function of NRG2 is more subtle, playing a complimentary role to NRG1. Following birth, homozygous NRG2 null mice displayed initial severe growth retardation and increased morbidity. Once these NRG2 null mice survived the weaning period (40 days), their growth caught up and they were indistinguishable from their littermates by 4 months. After up to 18 months of monitoring, there was no difference in mortality between NRG2 null and their normal littermates [
      • Britto J.M.
      • et al.
      Generation and characterization of neuregulin-2-deficient mice.
      ]. Subsequent knockout experiments indicated that NRG2 may be involved in regional development in the central nervous system and that the loss of NRG2 may lead animals to display behavioral characteristics of human psychiatric disease [
      • Yan L.
      • et al.
      Neuregulin-2 ablation results in dopamine dysregulation and severe behavioral phenotypes relevant to psychiatric disorders.
      ]. Thus, normal function of NRG2 is subtle and remains to be fully elucidated and the biological significance of NRG2 fusions in solid tumors is to be determined [
      • Kohsaka S.
      • et al.
      Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
      ,
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ].

      Distribution of NRG1+ and NRG2+ solid malignancies

      The overall incidence of NRG1 and NRG2 fusions is very rare [
      • Jonna S.
      • et al.
      Detection of NRG1 gene fusions in solid tumors.
      ,
      • Kohsaka S.
      • et al.
      Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
      ,
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ,
      • Wang X.Z.
      • et al.
      gamma-heregulin is the product of a chromosomal translocation fusing the DOC4 and HGL/NRG1 genes in the MDA- MB-175 breast cancer cell line.
      ,
      • Liu X.
      • et al.
      Gamma-heregulin: a fusion gene of DOC-4 and neuregulin-1 derived from a chromosome translocation.
      ,
      • Howarth K.D.
      • et al.
      NRG1 fusions in breast cancer.
      ,
      • Adelaide J.
      • et al.
      A recurrent chromosome translocation breakpoint in breast and pancreatic cancer cell lines targets the neuregulin/NRG1 gene.
      ,
      • Huang H.E.
      • et al.
      A recurrent chromosome breakpoint in breast cancer at the NRG1/neuregulin 1/heregulin gene.
      ,
      • Fernandez-Cuesta L.
      • et al.
      CD74-NRG1 fusions in lung adenocarcinoma.
      ,
      • Nakaoku T.
      • et al.
      Druggable oncogene fusions in invasive mucinous lung adenocarcinoma.
      ,
      • Gow C.H.
      • et al.
      Multidriver mutation analysis in pulmonary mucinous adenocarcinoma in Taiwan: identification of a rare CD74-NRG1 translocation case.
      ,
      • Dhanasekaran S.M.
      • et al.
      Transcriptome meta-analysis of lung cancer reveals recurrent aberrations in NRG1 and Hippo pathway genes.
      ,
      • Zheng Z.
      • et al.
      Anchored multiplex PCR for targeted next-generation sequencing.
      ,
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ,
      • Drilon A.
      • et al.
      Response to ERBB3-directed targeted therapy in NRG1-rearranged cancers.
      ,
      • Chang J.C.
      • et al.
      Comprehensive molecular and clinicopathologic analysis of 200 pulmonary invasive mucinous adenocarcinomas identifies distinct characteristics of molecular subtypes.
      ,
      • Heining C.
      • et al.
      NRG1 fusions in KRAS wild-type pancreatic cancer.
      ,
      • Jones M.R.
      • et al.
      NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.
      ,
      • Laskin J.
      • et al.
      NRG1 fusion-driven tumors: biology, detection, and the therapeutic role of afatinib and other ErbB-targeting agents.
      ]. In many tumor types, only limited numbers (or even just one) of NRG1 or NRG2 fusion variant have been identified. By percentage, there is no organ dominance of the presence of NRG1 or NRG2 fusions [
      • Jonna S.
      • et al.
      Detection of NRG1 gene fusions in solid tumors.
      ,
      • Kohsaka S.
      • et al.
      Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
      ,
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ,
      • Wang X.Z.
      • et al.
      gamma-heregulin is the product of a chromosomal translocation fusing the DOC4 and HGL/NRG1 genes in the MDA- MB-175 breast cancer cell line.
      ,
      • Liu X.
      • et al.
      Gamma-heregulin: a fusion gene of DOC-4 and neuregulin-1 derived from a chromosome translocation.
      ,
      • Howarth K.D.
      • et al.
      NRG1 fusions in breast cancer.
      ,
      • Adelaide J.
      • et al.
      A recurrent chromosome translocation breakpoint in breast and pancreatic cancer cell lines targets the neuregulin/NRG1 gene.
      ,
      • Huang H.E.
      • et al.
      A recurrent chromosome breakpoint in breast cancer at the NRG1/neuregulin 1/heregulin gene.
      ,
      • Fernandez-Cuesta L.
      • et al.
      CD74-NRG1 fusions in lung adenocarcinoma.
      ,
      • Nakaoku T.
      • et al.
      Druggable oncogene fusions in invasive mucinous lung adenocarcinoma.
      ,
      • Gow C.H.
      • et al.
      Multidriver mutation analysis in pulmonary mucinous adenocarcinoma in Taiwan: identification of a rare CD74-NRG1 translocation case.
      ,
      • Dhanasekaran S.M.
      • et al.
      Transcriptome meta-analysis of lung cancer reveals recurrent aberrations in NRG1 and Hippo pathway genes.
      ,
      • Zheng Z.
      • et al.
      Anchored multiplex PCR for targeted next-generation sequencing.
      ,
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ,
      • Drilon A.
      • et al.
      Response to ERBB3-directed targeted therapy in NRG1-rearranged cancers.
      ,
      • Chang J.C.
      • et al.
      Comprehensive molecular and clinicopathologic analysis of 200 pulmonary invasive mucinous adenocarcinomas identifies distinct characteristics of molecular subtypes.
      ,
      • Heining C.
      • et al.
      NRG1 fusions in KRAS wild-type pancreatic cancer.
      ,
      • Jones M.R.
      • et al.
      NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.
      ,
      • Laskin J.
      • et al.
      NRG1 fusion-driven tumors: biology, detection, and the therapeutic role of afatinib and other ErbB-targeting agents.
      ] in contrast to another rare RTK fusion, tropomyosin receptor kinase (TRK), encoded by the neurotrophic receptor tyrosine kinase gene (NRTK), where it is concentrated in certain pediatric and adult tumors [
      • Cocco E.
      • et al.
      NTRK fusion-positive cancers and TRK inhibitor therapy.
      ]. The incidence of NRG2α fusions are even rarer and are five to ten times lower in frequency then NRG1 fusions [
      • Kohsaka S.
      • et al.
      Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
      ,
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ,
      • Chang J.C.
      • et al.
      Comprehensive molecular and clinicopathologic analysis of 200 pulmonary invasive mucinous adenocarcinomas identifies distinct characteristics of molecular subtypes.
      ] (Box 1). Of the approximate 6400 cases of colon cancer and 5400 cases of breast cancer profiled by whole transcriptome sequencing (WTS), no NRG2 fusions were identified [
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ].
      Given the differential binding affinities between the α- and β-isoform of NRG1, it is logical to speculate the α/β ratio may be important in the oncogenic process, with a high β/α ratio exhibiting a more aggressive clinical course and more refractoriness to treatment and where a switch to a higher β/α ratio is a form of dynamic acquired resistance during treatment. Nevertheless, the ability to detect and report the α- and β-isoform is an important and necessary future goal to better understand NRG1 and NRG2 fusions.

      Methods and strategies to detect NRG1 fusions

      Given the diversity of NRG fusions (in terms of fusion partners and tumor or origin), coupled with their extreme rarity, and the further need to distinguish the α- and β-isoform of NRG fusions for future translational research and clinical care, the method used to detect NRG fusions should be highly technically advanced with maximum sensitivity. However, the detection method should also be cost-effective, taking into consideration limited resource availability of many countries, due to the need to detect multiple more common driver mutations in many tumor types. Therefore, it is necessary to incorporate knowledge of NRG1 fusion in each individual tumor type to optimize screening. Thus, an ‘all size fits all’ strategy may not be the most effective as different tumor-specific clinicians are involved. As there is more published literature on the clinicopathologic and molecular characteristics of NRG1+ non-small cell lung cancer (NSCLC) and pancreatic adenocarcinoma, we have focused on these tumor types when proposing a screening strategy. Below are the various detecting methods for NRG1 fusions. The advantages and disadvantages of various molecular testing methods are summarized in Table 2.
      Table 2Pros and cons of various molecular testing methods
      ProsCons
      RNA-based NGSCan capture transcribed products, allowing identification of in-frame transcripts even in genes with multiple splice variants and transcriptional start sites.

      Has detected the most fusion partners for NRG1+ NSCLC. The most common method NRG1 and NRG2 fusions were detected.
      RNA of sufficient quality and quantity may be difficult to obtain, as RNA is labile and sample collection, fixation, and storage conditions may affect the degradation of extracted RNA.
      1. WTSComprehensive method to detect RTK or ligand-fusions.

      Does not require a priori knowledge of the gene fusion partner in anchored-multiplex PCR targeted RNA sequencing.
      Detection is limited to the primers designed. For example, the commercially available PCR targeted RNA sequencing does not detect NRG2 fusions as primers were not designed to bait NRG2 gene.
      2. NanostringUtilizing hybridization, alignment, and digital counting, Nanostring can efficiently determine the level of expression of all the exons in the gene of interest.

      Can detect complex rearrangements.
      The assay may not be easily optimized if the quality of the RNA is variable.

      Any unknown 5′ fusion partners will not be identified.
      3. RT-PCRRNA isolated from tumor samples is reverse transcribed to cDNA and amplification is performed using primers upstream and downstream of the expected breakpoint to detect the fusion transcripts.

      Efficient where the 5′ and 3′ partner genes and breakpoints are highly recurrent.
      Novel rearrangements cannot be detected.

      Potential chance of missing rare NRG fusion partners.
      DNA-based NGSAllows for massively parallel sequencing and provides information on multiple targets in a high throughput and cost-effective manner.

      Can be performed on small and limited tumor specimens.
      Unable to determine whether a putative fusion would be expressed or functional (in-frame) because in many cases the chromosomal translocation occurred in intronic regions and cryptic splicing may lead to an ‘out-of-frame’ RNA transcript.
      1. Hybrid captureDNA primer probes are designed to cover the intronic regions in which most breakpoints occur.

      Can detect novel fusion partners and determine the exact breakpoints.
      Genomic coverage may not be comprehensive.

      Sensitivity for fusion detection can be reduced if intronic sequences are large and difficult to completely cover (i.e., NRG1 gene).

      RNA-based NGS

      RNA-based NGS approaches have the advantage of providing evidence that a putative fusion is expressed and in-frame, as not all the RNA transcripts are in-frame. In genes that have multiple splice variants and transcriptional start sites, RNA-based sequencing methods can capture the transcribed products that could have undergone alternative splicing, which will not be captured by DNA sequencing. DNA-based sequencing can only predict a transcript but cannot guarantee that an in-frame messenger RNA product will be produced due to potential cryptic transcription start sites. Additionally, the frequency of the NRG1 fusions can be calculated from the numbers of junctional reads, including the ratio of β/α-isoforms. Limitations to RNA NGS are, as with other RNA-based assays, primarily due to the ability to obtain RNA of sufficient quality and quantity from clinical samples to meet the assay requirements, particularly in formalin-fixed, paraffin-embedded tissues.

      WTS and anchor multiple PCR targeted RNA sequencing

      WTS is the most comprehensive method to detect fusions (RTK or ligand). Almost all the current literature on NRG1 and on NRG2 fusions were detected by WTS [
      • Jonna S.
      • et al.
      Detection of NRG1 gene fusions in solid tumors.
      ,
      • Kohsaka S.
      • et al.
      Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
      ,
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ,
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ,
      • Chang J.C.
      • et al.
      Comprehensive molecular and clinicopathologic analysis of 200 pulmonary invasive mucinous adenocarcinomas identifies distinct characteristics of molecular subtypes.
      ]. The advantages of WTS are that actual transcribed messenger RNA is sequenced and the detection is not dependent on the initial adapter ligation step that facilitates priming without a priori knowledge of the gene fusion partner in anchored-multiplex PCR (AMP) targeted RNA sequencing [
      • Zheng Z.
      • et al.
      Anchored multiplex PCR for targeted next-generation sequencing.
      ]. The AMP targeted RNA sequencing test is available commercially. However, it does not detect NRG2 fusions as the current primers were not designed to bait NRG2. Thus, while AMP does not require the a priori knowledge of the fusion partners, without primers specifically baiting NRG2, AMP cannot reliably detect NRG2 fusions, which is one of its disadvantages compared with WTS.

      Differential exon expression (3′/5′ NRG1 ratio; Nanostring)

      Nanostring uses a methodology of hybridization, alignment, and digital counting to determine the level of expression of all the exons in the gene of interest. Probes that target 5′ and 3′ gene regions can be designed to determine the relative expression [
      • Vaughn C.P.
      • et al.
      Simultaneous detection of lung fusions using a multiplex RT-PCR next generation sequencing-based approach: a multi-institutional research study.
      ]. To increase the sensitivity and specificity of the detection method, breakpoint-specific probes (i.e., 5-fusion partner and 3′ NRG) can also be used to confirm the actual translocation [
      • Jones M.R.
      • et al.
      NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.
      ]. Another advantage of this methods is that it can detect complex rearrangements, as has been reported in NRG1+ pancreatic cancer [
      • Jones M.R.
      • et al.
      NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.
      ]. In several cases of NRG1+ pancreatic cancer, both the 5′ and 3′ NRG1 were rearranged, leaving only the expression of the EGF-like domain detectable in the NRG1 fusion. The assay may not be easily optimized if the quality of the RNA is variable and a disadvantage of this technique is that any unknown 5′ fusion partners will not be identified. However, it will be expensive to purchase all the primers to published fusion partner exons of NRG1 and NRG2 (Figure 2), especially if screening of NRG fusions is done in a multiplex manner.
      Figure 2
      Figure 2Distribution of NRG1 and NRG2 fusion variants in various organs.
      WHSC1L1-NRG1 is the NRG1 fusion variant identified in a NRG1+ soft tissue sarcoma of the extremity/trunk. Out-of-frame (non-functional) variants are denoted by **. From references [
      • Jonna S.
      • et al.
      Detection of NRG1 gene fusions in solid tumors.
      ,
      • Kohsaka S.
      • et al.
      Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
      ,
      • Ou S.I.
      • et al.
      Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
      ,
      • Nagasaka M.
      • Ou S.I.
      Neuregulin 1 fusion-positive NSCLC.
      ,
      • Howarth K.D.
      • et al.
      NRG1 fusions in breast cancer.
      ,
      • Fernandez-Cuesta L.
      • et al.
      CD74-NRG1 fusions in lung adenocarcinoma.
      • Jones M.R.
      • et al.
      NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.
      ]. Abbreviations: NRG, neuregulin.

      RT-PCR

      In RT-PCR approaches to detect fusion transcripts, RNA isolated from tumor samples is reverse transcribed to cDNA and amplification is performed using primers to exonic sequences upstream and downstream of the expected breakpoint. RT-PCR for fusion detection is best utilized for scenarios in which the 5′ and 3′ partner genes, and the breakpoints for each, are highly recurrent, such as EML4-ALK, CD74-ROS1, or KIF5B-RET [
      • Li W.
      • et al.
      Intergenic breakpoints identified by DNA sequencing confound targetable kinase fusion detection in NSCLC.
      ]. Novel rearrangements will not be detected by this methodology. Thus, given the tremendous number of published fusion partners to NRG1 and NRG2 fusion positive solid tumors (Figure 2) and given the rarity of NRG fusions, the potential chance of missing many NRG fusions make RT-PCR less attractive as an option and is not included into our proposed screening strategies.

      DNA-based NGS

      DNA NGS is a commonly employed sequencing method in both tumor and plasma genotyping. It allows for massively parallel sequencing, provides information on multiple targets in a high throughput and cost-effective manner, and can be performed on small and limited tumor specimens. Hybrid capture, where a certain intronic hotspot region for translocation with or without whole exome sequencing (WES), is the most employed sequencing strategy in DNA-based NGS. WES alone could not identify fusions as this method cannot capture the translocation events as they occur within the intronic regions keeping each exon intact even if DNA rearrangement has occurred. The sensitivity for detection of fusions using hybrid capture can be reduced if intronic sequences are large and difficult to completely cover, such as in the case of NRG1 gene [
      • Steinthorsdottir V.
      • et al.
      Multiple novel transcription initiation sites for NRG1.
      ,
      • Nagasaka M.
      • Ou S.I.
      Neuregulin 1 fusion-positive NSCLC.
      ]. Furthermore, because many of the chromosomal translocations occur in intronic regions, there may be cryptic exons that are included in final mRNA transcripts that may have out-of-frame transcripts (premature stop codons). Hence, DNA NGS cannot reliably predict whether a putative NRG1 fusion would be expressed as a functional NRG1 protein [
      • Li W.
      • et al.
      Potential unreliability of uncommon ALK, ROS1, and RET genomic breakpoints in predicting the efficacy of targeted therapy in NSCLC.
      ,
      • Trombetta D.
      • et al.
      Frequent NRG1 fusions in Caucasian pulmonary mucinous adenocarcinoma predicted by Phospho-ErbB3 expression.
      ].

      Immunohistochemistry (IHC)

      Immunohistochemical detection of HER3 phosphorylation (p-HER3) has been utilized as a functional screen for rare NRG1 fusions. In one report, IHC for p-HER3 had a sensitivity of 100% (all five CD74-NRG1 were p-HER3 IHC+) among invasive mucinous adenocarcinoma (IMA) and 97.5% specificity (six IHC+/241 IHC– among adenocarcinoma of the lung) in detecting NRG1 fusions among adenocarcinoma of the lung [
      • Fernandez-Cuesta L.
      • et al.
      CD74-NRG1 fusions in lung adenocarcinoma.
      ]. A study in which p-HER3 IHC was performed on 85 cases of lung cancer (51 IMA and 34 non-IMA) and correlated with fluorescent in situ hybridization (FISH) for NRG1 rearrangements, showed a sensitivity of 94% (16 cases positive by p-HER3/17 cases positive by FISH) and a specificity of 97% (two cases positive by p-HER3/66 cases negative by FISH) [
      • Duruisseaux M.
      • et al.
      NRG1 fusion in a French cohort of invasive mucinous lung adenocarcinoma.
      ]. Limitations to IHC testing includes the potential for false negative results due to sample fixation conditions. Potential sources of false positivity are expression of tissue-specific NRG1 isoforms and endogenous expression of NRG1 in tissues such as neural crest-derived tissues. Although for such a rare tumor, false positivity (sensitivity) is preferred over false negativity (specificity). While further studies on additional tumors and tissue types would be necessary to determine the feasibility of p-HER3 screening by IHC across tumor types, we incorporated p-HER3 IHC as part of the initial screening strategy due to relative simplicity and low cost of IHC.

      FISH

      Break-apart FISH to detect fusions is a commonly used methodology and is one of the FDA-approved methods to detect ALK rearrangements. FISH has been used to detect NRG1 fusions in NSCLC [
      • Duruisseaux M.
      • et al.
      NRG1 fusion in a French cohort of invasive mucinous lung adenocarcinoma.
      ,
      • Cha Y.J.
      • Shim H.S.
      Biology of invasive mucinous adenocarcinoma of the lung.
      ] and has also validated the presence of NRG1 fusions in breast tumors [
      • Huang H.E.
      • et al.
      A recurrent chromosome breakpoint in breast cancer at the NRG1/neuregulin 1/heregulin gene.
      ]. However, out of the 15 NRG1 FISH positive invasive mucinous adenocarcinoma (IMA) cases, only four underwent RNA sequencing and confirmed the presence of CD74-NRG1 fusion. The NRG1 fusion transcript status of the remaining 11 NRG1 FISH positive cases were not tested further by RNA sequencing to determine if they were truly NRG1 fusions or not [
      • Duruisseaux M.
      • et al.
      NRG1 fusion in a French cohort of invasive mucinous lung adenocarcinoma.
      ]. Furthermore, unlike the FISH testing in ALK fusion, a cut-off value for scoring NRG1 fusion as positive has not yet been vigorously studied and validated. As such, the ‘conventional wisdom’ of 15% break-apart signals as FISH positivity from the ALK detection standard is empirically applied as NRG1 FISH positivity criterion but will require a vigorous validation study before NRG1 FISH can be applied universally and unequivocally. Additionally, limitations of FISH testing include the potential for false negative cases that may occur due to fixation conditions, the inability to detect the fusion partner, whether a fusion transcript is expressed, and any co-occurring mutations. Moreover, break-apart FISH did not identify two out of three KRAS wild type NRG1+ pancreatic adenocarcinoma due to the frequent complex rearrangement of NRG1 in pancreatic adenocarcinoma [
      • Jones M.R.
      • et al.
      NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.
      ]. Thus, currently we do not propose FISH as part of the screening or diagnostic process for NRG1 due to its high cost and unreliable sensitivity and specificity.

      Clinicopathologic characteristics and screening strategies of NRG fusion positive solid malignancies

      NRG1+ NSCLC

      The three dominant fusion partners of NRG1+ NSCLC among 99 NRG1+ NSCLC patients were CD74, SLC3A2, and SDC4 [
      • Jonna S.
      • et al.
      Detection of NRG1 gene fusions in solid tumors.
      ,
      • Nagasaka M.
      • Ou S.I.
      Neuregulin 1 fusion-positive NSCLC.
      ,
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ]. IMA accounted for more than half of the histology of all NRG1+ NSCLC [
      • Nagasaka M.
      • Ou S.I.
      Neuregulin 1 fusion-positive NSCLC.
      ,
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ]. In the only large-scale NRG1+ NSCLC registry to date, the median age of diagnosis for the 110 NRG1+ NSCLC patients included was 64 years. Slightly more than half of the NRG1+ NSCLC patients were never-smokers (57%). The majority of NRG1+ NSCLC had IMA as histology (57%) and 6% were nonadenocarcinoma [
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ]. The majority of NRG1+ NSCLC did not express PD-L1, followed by low PD-L1 expression. High PD-L1 expression (⩾50%) was rare [
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ]. The median tumor mutation burden (TMB) was extremely low at 0.9 (mutations/megabase), even lower than ALK or ROS1 fusions, potentially indicating that NRG fusion-driven solid malignancies resulted from this single molecular event.
      As IMA comprises about 57–61% of NRG1+ NSCLC [
      • Nagasaka M.
      • Ou S.I.
      Neuregulin 1 fusion-positive NSCLC.
      ,
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ], the first decision point of our proposed screening strategy is to identify if the histology of the NSCLC is IMA or not. If IMA histology is confirmed, we recommend going directly to NGS when feasible, as the percentage of IMA among NSCLC is low (~5%) and IMA harbor other driver mutations [
      • Chang J.C.
      • et al.
      Comprehensive molecular and clinicopathologic analysis of 200 pulmonary invasive mucinous adenocarcinomas identifies distinct characteristics of molecular subtypes.
      ,
      • McWilliams R.R.
      • et al.
      Risk factors for early-onset and very-early-onset pancreatic adenocarcinoma: a pancreatic cancer case-control consortium (PanC4) analysis.
      ]. While NRG1 fusion accounts for ~7% of all NSCLC with IMA histology [
      • Chang J.C.
      • et al.
      Comprehensive molecular and clinicopathologic analysis of 200 pulmonary invasive mucinous adenocarcinomas identifies distinct characteristics of molecular subtypes.
      ,
      • McWilliams R.R.
      • et al.
      Risk factors for early-onset and very-early-onset pancreatic adenocarcinoma: a pancreatic cancer case-control consortium (PanC4) analysis.
      ], KRAS mutations are present in the range of 62–76%, especially KRAS G12D [
      • Chang J.C.
      • et al.
      Comprehensive molecular and clinicopathologic analysis of 200 pulmonary invasive mucinous adenocarcinomas identifies distinct characteristics of molecular subtypes.
      ,
      • McWilliams R.R.
      • et al.
      Risk factors for early-onset and very-early-onset pancreatic adenocarcinoma: a pancreatic cancer case-control consortium (PanC4) analysis.
      ]. Thus NGS on IMA is cost-effective as it will very likely identify actionable driver mutations in IMA. Another cost-effective way is to perform RNA NGS in NSCLC is with KRAS-wild type IMA histology. A limitation of this histology-based screening approach is that IMA is usually diagnosed from a large sample from resected tumor samples or large core biopsy, which may not be readily available using the current diagnostic approach to NSCLC. Thus, if the histological diagnosis of IMA is not definitive and NGS is not routinely available, our strategy recommends screening by phospho-HER3 (p-HER3) IHC (Figure 3A ). This screening method will avoid NGS being performed on many adenocarcinomas of the lung in situations with resource constraints.
      Figure 3
      Figure 3Proposed screening strategies for NRG1+ solid tumors.
      (A) Proposed screening strategies for NRG1+ non-small cell lung cancer (NSCLC). The identification of invasive mucinous adenocarcinoma (IMA) is best performed with a surgical sample rather than with a core biopsy. When the histological diagnosis of IMA is not definitive or the diagnosis is not IMA and next generation sequencing (NGS) is not routinely available, we recommend screening by phospho-HER3 (p-HER3) immunohistochemistry (IHC). This will avoid performing NGS on many adenocarcinomas of the lung in situations with resource constraints. (B) Proposed screening strategies for NRG1+ pancreatic ductal adenocarcinoma (PDAC). NRG1+ PDAC is essentially KRAS-wild type and the current recommendation is to screen for BRCA mutations in PDAC, which is usually done in panel, and KRAS mutation status can be identified. It is recommended that NGS is performed on KRAS-wild type, early onset, or phosphor-HER3 positive PDAC. (C) Proposed screening algorithm for NRG1+ solid tumors. As the publication on NRG1+ breast adenocarcinoma demonstrated the false positivity rate of DNA NGS in screening for NRG1+ breast cancer, we recommended RNA NGS for our screening whenever feasible.

      NRG1+ pancreatic ductal adenocarcinoma (PDAC)

      While the overall incidence of NRG1+ PDAC is estimated to be 0.48% [
      • Jonna S.
      • et al.
      Detection of NRG1 gene fusions in solid tumors.
      ], the incidence of NRG1+ PDAC seemed to be enriched in KRAS-wild-type PDAC [
      • Heining C.
      • et al.
      NRG1 fusions in KRAS wild-type pancreatic cancer.
      ,
      • Jones M.R.
      • et al.
      NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.
      ]. There were much fewer NRG1+ PDAC cases described in the literature. Among the 12 cases of NRG1+ PDAC, ATP1b1 is the dominant fusion partner [
      • Heining C.
      • et al.
      NRG1 fusions in KRAS wild-type pancreatic cancer.
      ,
      • Jones M.R.
      • et al.
      NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.
      ]. Even more importantly, NRG1 fusions in PDAC seem to be complex fusions, with translocation involving the 5′ and 3′ end of NRG1. In one NRG1+ fusion variant (APP-NRG1), only the EGF-like domain of NRG1 (exons 6 and 7) was inserted between exon 15 and 16 of the APP gene [
      • Jones M.R.
      • et al.
      NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.
      ].
      Given the lethality of advanced PDAC and the frequent paucity of tumor tissue obtained at the time of diagnosis (e.g., fine needle aspiration from endoscopic retrograde cholangiopancreatography), we recommend all PDAC undergo at least DNA NGS as the first comprehensive molecular profiling step. For KRAS-wild type patients or early onset pancreatic cancer (patients <60 years old without risk factors such as obesity, smoking, diabetes mellitus, excessive alcohol use, pancreatitis, and family history of pancreatic cancer) [
      • Prentice L.M.
      • et al.
      NRG1 gene rearrangements in clinical breast cancer: identification of an adjacent novel amplicon associated with poor prognosis.
      ], we recommend RNA NGS to screen for NRG1 fusions. With the entrance of multiple covalent KRAS G12C inhibitors in clinical trials, and in the future other non-KRAS G12C isoform inhibitors, it is anticipated that more clinicians will also screen for KRAS mutations (Figure 3B).

      NRG1+ breast cancer

      Although a recent report initially identified four cases of NRG1+ breast cancer out of 571 cases by WTS (WRN-NRG1, FAM91A1-NRG1, ARHGEF39-NRG1, and ZNF704-NRG1), further analysis of the RNA sequencing data identified only FAM91A1-NRG1 and ZNF704-NRG1 as truly in-frame, resulting in a true incidence of 0.35% (2/571) [
      • Howarth K.D.
      • et al.
      NRG1 fusions in breast cancer.
      ]. Furthermore, the same study identified 20 out of the first 250 (8%) breast cancer carcinoma samples to have a detectable DNA breakpoint within NRG1 but no RNA transcript was detected. Thus, RNA NGS remained the gold standard in identifying transcribed and in-frame NRG1 fusion transcript. One study identified that NRG1 gene rearrangement portends a poor prognosis in breast cancer but given the rarity of NRG1+ breast adenocarcinoma and improvement in diagnostic methods of NRG1 fusions, this result requires further investigation and confirmation [
      • Cadranel J.
      • et al.
      Therapeutic potential of afatinib in NRG1 fusion-driven solid tumors: a case series.
      ].

      Screening for other NRG1+ solid tumors

      Even less is known about other NRG1+ solid tumors beside NRG1+ NSCLC and NRG1+ PDAC. Without knowledge of a predominant histology, a well-defined molecular subgroup, or hormonal status associated with a particular tumor-specific NRG1 fusion, we propose a dichotomy approach of either upfront RNA NGS or screening phospho-HER3 IHC as a screening strategy (Figure 3C).

      Potential treatment strategies of NRG1+ solid malignancies

      Chemotherapy and immune checkpoint inhibitors (ICI)

      Platinum-based chemotherapy and immunotherapy in combination is widely used as first-line treatment of advanced NSCLC. Nevertheless, the efficacy of platinum-based chemotherapy in NRG1+ NSCLC is modest, with an overall response rate (ORR) of about <15% and a median progression-free survival (PFS) between 4 and <6 months. Thus, chemotherapy seems to have cytostatic effects on NRG1+ NSCLC patients. Given low PD-L1 expression and low TMB, efficacy of single agent ICI was also limited, with ORR <15% and median of 4 months, although this was shown with a limited number of patients [
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ]. No other large-scale registration study has investigated the efficacy of chemotherapy and immunotherapy in other non-NSCLC NRG1+ solid tumors.
      It is hypothesized that NRG1 fusions promote oncogenesis by being in proximity and in high concentration to HER3, leading to a functionally constitutively active HER3, disrupting the NRG1/HER3 ligand–receptor binding and HER3/HER2 heterodimerization. Thus, the logistic approach currently being adopted to target NRG1+ solid tumors can be divided into either pan-HER TKIs, as HER3 can also hetero-dimerize with EGFR and HER4, or monoclonal antibodies that block NRG1 binding to HER3 and potentially disrupt HER3/HER2 hetero-dimerization.

      Pan-HER TKIs

      One of the rationales to target NRG1 fusion is through the HER3/HER2 signaling pathway. However, HER3 does not have a kinase domain and its primary heterodimerization partner is HER2, which currently can only be inhibited by pan-HER inhibitors.

      Afatinib

      Many case reports describe the efficacies of afatinib in NRG1+ tumors [
      • Nagasaka M.
      • Ou S.I.
      Neuregulin 1 fusion-positive NSCLC.
      ,
      • Laskin J.
      • et al.
      NRG1 fusion-driven tumors: biology, detection, and the therapeutic role of afatinib and other ErbB-targeting agents.
      ,
      • Estrada-Bernal A.
      • et al.
      Tarloxotinib Is a hypoxia-activated pan-HER kinase inhibitor active against a broad range of HER-family oncogenes.
      ]. The response to afatinib is heterogeneous, with partial response, stable disease, and progressive disease all been reported. From the global NRG1+ NSCLC registry database, ORR to afatinib was about 25%, with progressive disease being the most common response [
      • Drilon A.
      • et al.
      Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
      ]. Currently, afatinib is being investigated in the TAPUR trial (Group 18: NRG1) (NCT02693535)i and in Germany (NCT04410653)ii.

      Tarloxotinib

      Tarloxotinib is a hypoxia-activated prodrug that is converted to and released as a potent irreversible pan-ErbB TKI (tarloxotinib-E) by STEP4, a membrane reductase, under pathophysiological hypoxia present in solid tumors. Tarloxotinib-E inhibits growth of DOC4-NRG1 breast cancer cell line (MA-MBA-175) at GI50 at 0.3 nM, while the GI50 for afatinib was 1.2 nM [
      • Tirunagaru V.G.
      • et al.
      Tarloxotinib exhibits potent activity in NRG1 fusion and rearranged cancers.
      ]. Additionally, tarloxotinib-E inhibited CLU-NRG1 ovarian xenograft, while afatinib did not inhibit this xenograft growth at all [
      • Geuijen C.A.W.
      • et al.
      Unbiased combinatorial screening identifies a bispecific IgG1 that potently inhibits HER3 signaling via HER2-guided ligand blockade.
      ]. Tarloxotinib is given as a weekly infusion, which may be inconvenient to some patients, and there is a potential for developing QTc prolongation. A prospective trial attempted to investigate EGFR and HER2 alterations, including NRG1 fusions, has been terminated and tarloxotinib is currently no longer in clinical development (NCT03805841)iii.

      Anti HER2/HER3 monoclonal antibodies

      Zenocutuzumab (MCLA-128)

      Zenocutuzumab is a novel bispecific HER2/HER3 antibody of the IgG1 class that utilizes the ‘dock-and-block’ mechanism. Zenocutuzumab binds to domain I of HER2 on top of HER2 domain with a KD of around 3.2 + 0.5 nM, which is like trastuzumab. The other epitope of zenocutuzumab interferes with the binding of NRG1 to HER3 by steric hindrance and, together, zenocutuzumab inhibits the interaction of HER2-HER3 heterodimer formation [
      • de Vries Schultink A.H.M.
      • et al.
      Translational PK-PD modeling analysis of MCLA-128, a HER2/HER3 bispecific monoclonal antibody, to predict clinical efficacious exposure and dose.
      ]. Zenocutuzumab can inhibit HER2/HER3 signaling, independent of the expression of HER2, due to its binding to the domain 1 of HER2. Furthermore, given its specificity in inhibiting HER2/HER3 dimerization, it has no additive toxicity on cardiomyocytes, which is mediated by HER2/HER4 [
      • Alì G.
      • et al.
      Analysis of fusion genes by NanoString system: a role in lung cytology?.
      ]. In vitro modeling indicated that flat doses ≥360 mg every 3 weeks are expected to be efficacious in human, based on receptor occupancies and pharmacokinetic-pharmacodynamic model simulations [
      • Schoeberl B.
      • et al.
      Systems biology driving drug development: from design to the clinical testing of the anti-ErbB3 antibody seribantumab (MM-121).
      ]. An ongoing prospective trial of zenocutuzumab, amended to investigate its clinical efficacy in NRG1+ tumor, was initiated around January 2019 (NCT02912949)iv. On January 7, 2021, the US FDA granted ‘fast track’ designation of zenocutuzumab for NRG1+ cancers based on preliminary efficacy datav. Updated preliminary data based on 47 evaluable NRG1+ patients (25 NRG1+ NSCLC, 12 NRG1+ pancreatic cancer, ten other NRG1+ patients) was presented at the American Society of Clinical Oncology annual meeting 2021. In the overall NRG1+ population, the confirmed ORR was 29% among the 45 NRG1+ patients enrolled. Importantly, though, with limited number of patients, there seemed to be differential response to zenocutuzumab by tumor origin: PDAC ORR = 42% (5/12); NSCLC ORR = 24% (6/24), and others (non-NSCLC, non-PDAC) ORR= 22% (2/9). Zenocutuzumab was well tolerated with grade 3 adverse events being reported in ≤5% of patients across all cohorts. There was a notable lack of cardiotoxicity and severe gastrointestinal or skin toxicity [
      • Schram A.M.
      • et al.
      Efficacy and safety of zenocutuzumab in advanced pancreas cancer and other solid tumors harboring NRG1 fusions.
      ].

      Seribantumab (MM-121, FTN-001), anti HER3 antibody

      Seribantumab is a monoclonal IgG2 antibody against HER3 [
      • Odintsov I.
      • et al.
      The anti-HER3 monoclonal antibody seribantumab effectively inhibits growth of patient-derived and isogenic cell line and xenograft models with oncogenic NRG1 fusions.
      ]. To date, seribantumab has been tested in three randomized Phase II trials of metastatic cancer: in combination with paclitaxel versus paclitaxel alone in platinum-resistant/refractory ovarian cancer (NCT01447706)vi; in combination with exemestane versus exemestane plus placebo in ER/PR+ HER2− breast cancer (NCT01151046)vii; and in combination with erlotinib versus erlotinib alone in EGFR wild-type NSCLC (NCT00994123)viii. Results from all three randomized trials have been published and while the addition of seribantumab did not improve upon the PFS, which was the primary endpoints of all three trials, subgroup analysis indicated that patients whose tumor had detectable or high expression of NRG1 and/or low HER2 level (implying a lower level of receptor occupancy/activation) may have more benefit from the addition of seribantumab [
      • Odintsov I.
      • et al.
      The anti-HER3 monoclonal antibody seribantumab effectively inhibits growth of patient-derived and isogenic cell line and xenograft models with oncogenic NRG1 fusions.
      ].
      Recent preclinical data demonstrated that seribantumab could inhibit NRG1-stimulated growth of patient-derived NRG1+ breast (MDA-MB-175-VII, DOC4-NRG1) and NRG1+ lung (LUAD-0061AS3, SLC3A2-NRG1 fusion) tumors by reducing phosphorylation of all four HER family members (EGFR, HER2, HER3, HER4) and known downstream signaling molecules of AKT and ERK1/2, leading to apoptosis [
      • Odintsov I.
      • et al.
      The anti-HER3 monoclonal antibody seribantumab effectively inhibits growth of patient-derived and isogenic cell line and xenograft models with oncogenic NRG1 fusions.
      ]. Furthermore, seribantumab induced tumor reduction by 50–100% in NRG1+ PDX tumors bearing LUAD-0061AS3 (SLC34A2-NRG1) and ovarian tumor OV-10-0050 (CUL-NRG1). These results together led to initiation of a Phase II trial of seribantumab in NRG1+ tumors (NCT04383210)ix, CRESTONE: Clinical Study of Response to Seribantumab in Tumors with Neuregulin-1 (NRG1) Fusions.

      Potential resistance mechanisms to current treatment approaches

      There is extensive crosstalk among NRG1 and HER receptors. Besides binding to HER3, NRG1 can also bind to HER4 and EGFR, albeit at a lower affinity, while simultaneously HER3 can form heterodimers with HER4 and EGFR in addition to HER2. In three cases of NRG1+ NSCLC where the phosphorylation of the HER family members were examined, all four HER family receptors were phosphorylated [
      • Kohsaka S.
      • et al.
      Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
      ]. Thus, targeting a specific pair of ligand–receptor axis may not be sufficient. A further potential resistance mechanism is overexpression of the HER family receptors, especially HER2 or HER3, or both. Thus, once single agent activity of monoclonal antibodies or pan-HER TKIs is established, a combination approach aiming for maximum inhibition of the HER family signaling will need to be investigated.

      Concluding remarks and future perspectives

      Although NRG1 fusion was first reported in 1997 in a breast cancer cell line, it was not until the mid-2010s when the significance of NRG1 fusions started to gain appreciation. In 2021, we saw the preliminary clinical efficacy of zenocutuzumab, an anti-HER2/HER3 bispecific antibody, in solid tumors harboring NRG1 fusions. These results affirmed that NRG1 fusion constitutes a rare, yet novel category of ‘ligand-fusion’ as actionable oncogenic mutations. While we await the maturation of these clinical trial results, the most critical step would be broadening the effort to identify patients with these rare NRG1 fusions so that they can be enrolled into these clinical trials in a timely manner. RNA NGS, especially WTS, is the best method to detect these transcribed and in-frame NRG1 fusion variants. Phosphorylation of HER3 may be used as an initial screening test for NRG1 fusions if RNA NGS is not widely available. We here propose practical screening strategies based on histology (i.e., IMA for NSCLC) and co-genomic alterations (KRAS-wild type) that we hope aid in identifying patients with NRG fusions. In addition, advances in sequencing and screening methods may also reveal new fusions in other NRG family members. Despite the identification of existing fusions partners, it will be necessary to investigate the molecular and cellular regulation underpinning NRG1 fusions to create efficacious therapies targeting these fusions. As ligands are well known to crosstalk within the HER family of receptors, will monotherapy against NRG1 fusions be effective or will combination therapies with HER inhibitors be necessary to mount an appropriate antitumor response (see Outstanding questions)?
      What are the benchmarks that the US FDA will use to approve therapy targeting NRG1+ solid tumors given the near nonexistent knowledge of the natural history of these NRG1+ tumors?
      Is single agent enough to target NRG1 fusions? Given the extensive crosstalk between the ligand and the HER family of receptors, would combination of pan-HER TKIs and monoclonal HER3 antibodies be synergistic in improving clinical efficacy?
      Will the likely dynamic changes in the β/α ratio of NRG1 fusions during the course of treatment lead to 'tumor cell plasticity' and confer resistances to NRG1 fusion targeted therapy?
      How does alternative splicing control the generation of NRG1 β- versus α- variants?
      For fusion partners in NRG1+ solid tumors that are not transmembrane proteins, what are the postulated mechanism(s) of activation, since NRG1 will not be localized to the HER3 receptors through fusion partners?
      Are NRG2 fusions oncogenic and, if so, what is the best approach to target these fusions, as currently there are no known drugs that target the NRG2/HER4 axis?
      Are there more novel NRG fusions within the other four NRG families that remain to be discovered? Is the ligand-fusion oncogenic concept to be extended beyond NRGs, for example, will fusions in hepatocyte growth factor (HGF), which serves as a ligand to the MET RTK, be discovered?

      Acknowledgments

      This study did not receive any funding from agencies in the public, commercial, or not-for-profit sectors.

      Declaration of interests

      M.N. has received honorarium from Astra Zeneca, Caris Life Sciences, Daiichi Sankyo, Takeda, Novartis, EMD Serono, Blueprint Medicines, and Tempus. S.H.I.O. has received honorarium as speaker bureau of Roche/Genentech, Pfizer, consulting fees from, Pfizer, Roche/Genentech, Astra Zeneca, Takeda/ARIAD, Daiichi Sankyo, Jassen/JNJ, is on the Scientific Advisory Board of Elevation Oncology, Inc., is a former member of the Scientific Advisory board of Turning Point Therapeutics, has stock ownership in Turning Point Therapeutics, Inc., is a current member of the Scientific Advisory board of Elevation Oncology, and has stock ownership in Elevation Oncology.

      Resources

      References

        • Shaw A.T.
        • et al.
        Tyrosine kinase gene rearrangements in epithelial malignancies.
        Nat. Rev. Cancer. 2013; 13: 772-787
        • Schram A.M.
        • et al.
        Fusions in solid tumours: diagnostic strategies, targeted therapy, and acquired resistance.
        Nat. Rev. Clin. Oncol. 2017; 14: 735-748
        • Blume-Jensen P.
        • Hunter T.
        Oncogenic kinase signaling.
        Nature. 2001; 411: 355-365
        • Schaefer G.
        • et al.
        Gamma-heregulin: a novel heregulin isoform that is an autocrine growth factor for the human breast cancer cell line, MDA-MB-175.
        Oncogene. 1997; 15: 1385-1394
        • Jonna S.
        • et al.
        Detection of NRG1 gene fusions in solid tumors.
        Clin. Cancer Res. 2019; 25: 4966-4972
        • Fernandez-Cuesta L.
        • Thomas R.K.
        Molecular pathways: targeting NRG1 fusions in lung cancer.
        Clin. Cancer Res. 2015; 21: 1989-1994
        • Dimou A.
        • Camidge D.R.
        Detection of NRG1 fusions in solid tumors: rare gold?.
        Clin. Cancer Res. 2019; 25: 4865-4867
        • Schram A.M.
        • et al.
        Efficacy and safety of zenocutuzumab in advanced pancreas cancer and other solid tumors harboring NRG1 fusions.
        J. Clin. Oncol. 2021; 39: 3003
        • Kohsaka S.
        • et al.
        Identification of novel CD74-NRG2α fusion from comprehensive profiling of lung adenocarcinoma in Japanese never or light smokers.
        J. Thorac. Oncol. 2020; 15: 948-961
        • Ou S.I.
        • et al.
        Identification of novel CDH1-NRG2α and F11R-NRG2α fusions in NSCLC plus additional novel NRG2α fusions in other solid tumors by whole transcriptome sequencing.
        JTO Clin. Res. Rep. 2020; 2100132
        • Mei L.
        • Nave K.A.
        Neuregulin-ERBB signaling in the nervous system and neuropsychiatric diseases.
        Neuron. 2015; 83: 27-49
        • Holmes W.E.
        • et al.
        Identification of heregulin, a specific activator of p185erbB2.
        Science. 1992; 256: 1205-1210
        • Wen D.
        • et al.
        Neu differentiation factor: a transmembrane glycoprotein containing an EGF domain and an immunoglobulin homology unit.
        Cell. 1992; 69: 559-572
        • Goodearl A.D.
        • et al.
        Purification of multiple forms of glial growth factor.
        J. Biol. Chem. 1993; 268: 18095-18102
        • Marchionni M.A.
        • et al.
        Glial growth factors are alternatively spliced erbB2 ligands expressed in the nervous system.
        Nature. 1993; 362: 312-318
        • Falls D.L.
        • et al.
        ARIA, a protein that stimulates acetylcholine receptor synthesis, is a member of the neu ligand family.
        Cell. 1993; 72: 801-815
        • Ho W.H.
        • et al.
        Sensory and motor neuron–derived factor.
        J. Biol. Chem. 1995; 270: 14523-14532
        • Busfield S.J.
        • et al.
        Characterization of a neuregulin-related gene, Don-1, that is highly expressed in restricted regions of the cerebellum and hippocampus.
        Mol. Cell. Biol. 1997; 17: 4007-4014
        • Chang H.
        • et al.
        Ligands for ErbB-family receptors encoded by a neuregulin-like gene.
        Nature. 1997; 387: 509-512
        • Carraway 3rd, K.L.
        • et al.
        Neuregulin-2, a new ligand of ErbB3/ErbB4-receptor tyrosine kinases.
        Nature. 1997; 387: 512-516
        • Higashiyama S.
        • et al.
        A novel brain-derived member of the epidermal growth factor family that interacts with ErbB3 and ErbB4.
        J. Biochem. 1997; 122: 675-680
        • Zhang D.
        • et al.
        Neuregulin-3 (NRG3): a novel neural tissue-enriched protein that binds and activates ErbB4.
        Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 9562-9567
        • Harari D.
        • et al.
        Neuregulin-4: a novel growth factor that acts through the ErbB-4 receptor tyrosine kinase.
        Oncogene. 1999; 18: 2681-2689
        • Kanemoto N.
        • et al.
        Expression of TMEFF1 mRNA in the mouse central nervous system: precise examination and comparative studies of TMEFF1 and TMEFF2.
        Brain Res. Mol. Brain Res. 2001; 86: 48-55
        • Kinugasa Y.
        • et al.
        Neuroglycan C, a novel member of the neuregulin family.
        Biochem. Biophys. Res. Commun. 2004; 321: 1045-1049
        • Falls D.L.
        Neuregulins: functions, forms, and signaling strategies.
        Exp. Cell Res. 2003; 284: 14-30
        • Meyer D.
        • Birchmeier C.
        Multiple essential functions of neuregulin in development.
        Nature. 1995; 378: 386-390
        • Pinkas-Kramarski R.
        • et al.
        ErbB tyrosine kinases and the two neuregulin families constitute a ligand-receptor network.
        Mol. Cell. Biol. 1998; 18: 6090-6101
        • Crovello C.S.
        • et al.
        Differential signaling by the epidermal growth factor-like growth factors neuregulin-1 and neuregulin-2.
        J. Biol. Chem. 1998; 273: 26954-26961
        • Jones J.T.
        • et al.
        Binding specificities and affinities of EGF domains for ErbB receptors.
        FEBS Lett. 1999; 447: 227-231
        • Sweeny C.
        • et al.
        Ligand discrimination in signaling through an Erb4 receptor homodimer.
        J. Biol. Chem. 2000; 275: 19803-19807
        • Sweeny C.
        • et al.
        Growth factor specific signaling pathway stimulation and gene expression mediated by ErB receptors.
        J. Biol. Chem. 2001; 276: 22685-22698
        • Kramer R.
        • et al.
        Neuregulins with an Ig-like domain are essential for mouse myocardial and neuronal development.
        Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4833-4838
        • Li L.
        • et al.
        The breast proto-oncogene, HRG-alpha regulates epithelial proliferation and lobuloalveolar development in the mouse mammary gland.
        Oncogene. 2002; 21: 4900-4907
        • Steinthorsdottir V.
        • et al.
        Multiple novel transcription initiation sites for NRG1.
        Gene. 2004; 342: 97-105
        • Nagasaka M.
        • Ou S.I.
        Neuregulin 1 fusion-positive NSCLC.
        J. Thorac. Oncol. 2019; 14: 1354-1359
        • Bersell K.
        • et al.
        Neuregulin1/ErbB4 signaling induces cardiomyocyte proliferation and repair of heart injury.
        Cell. 2009; 138: 257-270
        • Gao R.
        • et al.
        A phase II, randomized, double-blind, multicenter, based on standard therapy, placebo-controlled study of the efficacy and safety of recombinant human neuregulin-1 in patients with chronic heart failure.
        J. Am. Coll. Cardiol. 2010; 55: 1907-1914
        • De Keulenaer G.W.
        • et al.
        Mechanisms of the multitasking endothelial protein NRG-1 as a compensatory factor during chronic heart failure.
        Circ. Heart Fail. 2019; 12e006288
        • Bouwer N.I.
        • et al.
        Cardiac monitoring in HER2-positive patients on trastuzumab treatment: a review and implications for clinical practice.
        Breast. 2020; 52: 33-44
        • Britto J.M.
        • et al.
        Generation and characterization of neuregulin-2-deficient mice.
        Mol. Cell. Biol. 2004; 24: 8221-8226
        • Yan L.
        • et al.
        Neuregulin-2 ablation results in dopamine dysregulation and severe behavioral phenotypes relevant to psychiatric disorders.
        Mol. Psychiatry. 2018; 23: 1233-1243
        • Wang X.Z.
        • et al.
        gamma-heregulin is the product of a chromosomal translocation fusing the DOC4 and HGL/NRG1 genes in the MDA- MB-175 breast cancer cell line.
        Oncogene. 1999; 18: 5718-5721
        • Liu X.
        • et al.
        Gamma-heregulin: a fusion gene of DOC-4 and neuregulin-1 derived from a chromosome translocation.
        Oncogene. 1999; 18: 7110-7114
        • Howarth K.D.
        • et al.
        NRG1 fusions in breast cancer.
        Breast Cancer Res. 2021; 23: 3
        • Adelaide J.
        • et al.
        A recurrent chromosome translocation breakpoint in breast and pancreatic cancer cell lines targets the neuregulin/NRG1 gene.
        Genes Chromosom. Cancer. 2003; 37: 333-345
        • Huang H.E.
        • et al.
        A recurrent chromosome breakpoint in breast cancer at the NRG1/neuregulin 1/heregulin gene.
        Cancer Res. 2004; 64: 6840-6844
        • Fernandez-Cuesta L.
        • et al.
        CD74-NRG1 fusions in lung adenocarcinoma.
        Cancer Discov. 2014; 4: 415-422
        • Nakaoku T.
        • et al.
        Druggable oncogene fusions in invasive mucinous lung adenocarcinoma.
        Clin. Cancer Res. 2014; 20: 3087-3093
        • Gow C.H.
        • et al.
        Multidriver mutation analysis in pulmonary mucinous adenocarcinoma in Taiwan: identification of a rare CD74-NRG1 translocation case.
        Med. Oncol. 2014; 31: 34
        • Dhanasekaran S.M.
        • et al.
        Transcriptome meta-analysis of lung cancer reveals recurrent aberrations in NRG1 and Hippo pathway genes.
        Nat. Commun. 2014; 5: 5893
        • Zheng Z.
        • et al.
        Anchored multiplex PCR for targeted next-generation sequencing.
        Nat. Med. 2014; 20: 1479-1484
        • Drilon A.
        • et al.
        Clinicopathologic features and response to therapy of NRG1 fusion-driven lung cancers: the eNRGy1 global multicenter registry.
        J. Clin. Oncol. 2021; 39: 2791-2802
        • Drilon A.
        • et al.
        Response to ERBB3-directed targeted therapy in NRG1-rearranged cancers.
        Cancer Discov. 2018; 8: 686-695
        • Chang J.C.
        • et al.
        Comprehensive molecular and clinicopathologic analysis of 200 pulmonary invasive mucinous adenocarcinomas identifies distinct characteristics of molecular subtypes.
        Clin. Cancer Res. 2021; 27: 4066-4076
        • Heining C.
        • et al.
        NRG1 fusions in KRAS wild-type pancreatic cancer.
        Cancer Discov. 2018; 8: 1087-1095
        • Jones M.R.
        • et al.
        NRG1 gene fusions are recurrent, clinically actionable gene rearrangements in KRAS wild-type pancreatic ductal adenocarcinoma.
        Clin. Cancer Res. 2019; 25: 4674-4681
        • Laskin J.
        • et al.
        NRG1 fusion-driven tumors: biology, detection, and the therapeutic role of afatinib and other ErbB-targeting agents.
        Ann. Oncol. 2020; 31: 1693-1703
        • Cocco E.
        • et al.
        NTRK fusion-positive cancers and TRK inhibitor therapy.
        Nat. Rev. Clin. Oncol. 2018; 15: 731-734
        • Alì G.
        • et al.
        Analysis of fusion genes by NanoString system: a role in lung cytology?.
        Arch. Pathol. Lab. Med. 2018; 142: 480-489
        • Vaughn C.P.
        • et al.
        Simultaneous detection of lung fusions using a multiplex RT-PCR next generation sequencing-based approach: a multi-institutional research study.
        BMC Cancer. 2018; 18: 828
        • Li W.
        • et al.
        Intergenic breakpoints identified by DNA sequencing confound targetable kinase fusion detection in NSCLC.
        J. Thorac. Oncol. 2020; 15: 1223-1231
        • Li W.
        • et al.
        Potential unreliability of uncommon ALK, ROS1, and RET genomic breakpoints in predicting the efficacy of targeted therapy in NSCLC.
        J. Thorac. Oncol. 2020; 16: 404-418
        • Trombetta D.
        • et al.
        Frequent NRG1 fusions in Caucasian pulmonary mucinous adenocarcinoma predicted by Phospho-ErbB3 expression.
        Oncotarget. 2018; 9: 9661-9671
        • Duruisseaux M.
        • et al.
        NRG1 fusion in a French cohort of invasive mucinous lung adenocarcinoma.
        Cancer Med. 2016; 5: 3579-3585
        • Cha Y.J.
        • Shim H.S.
        Biology of invasive mucinous adenocarcinoma of the lung.
        Transl. Lung Cancer Res. 2017; 6: 508-512
        • McWilliams R.R.
        • et al.
        Risk factors for early-onset and very-early-onset pancreatic adenocarcinoma: a pancreatic cancer case-control consortium (PanC4) analysis.
        Pancreas. 2016; 45: 311-316
        • Prentice L.M.
        • et al.
        NRG1 gene rearrangements in clinical breast cancer: identification of an adjacent novel amplicon associated with poor prognosis.
        Oncogene. 2005; 24: 7281-7289
        • Cadranel J.
        • et al.
        Therapeutic potential of afatinib in NRG1 fusion-driven solid tumors: a case series.
        Oncologist. 2021; 26: 7-16
        • Estrada-Bernal A.
        • et al.
        Tarloxotinib Is a hypoxia-activated pan-HER kinase inhibitor active against a broad range of HER-family oncogenes.
        Clin. Cancer Res. 2021; 27: 1463-1475
        • Tirunagaru V.G.
        • et al.
        Tarloxotinib exhibits potent activity in NRG1 fusion and rearranged cancers.
        Cancer Res. 2019; 79: 2202
        • Geuijen C.A.W.
        • et al.
        Unbiased combinatorial screening identifies a bispecific IgG1 that potently inhibits HER3 signaling via HER2-guided ligand blockade.
        Cancer Cell. 2018; 33: 922-936
        • de Vries Schultink A.H.M.
        • et al.
        Translational PK-PD modeling analysis of MCLA-128, a HER2/HER3 bispecific monoclonal antibody, to predict clinical efficacious exposure and dose.
        Investig. New Drugs. 2018; 36: 1006-1015
        • Schoeberl B.
        • et al.
        Systems biology driving drug development: from design to the clinical testing of the anti-ErbB3 antibody seribantumab (MM-121).
        NPJ Syst. Biol. Appl. 2017; 3: 16034
        • Odintsov I.
        • et al.
        The anti-HER3 monoclonal antibody seribantumab effectively inhibits growth of patient-derived and isogenic cell line and xenograft models with oncogenic NRG1 fusions.
        Clin. Cancer Res. 2021; 7: 3154-3166