The ROS-1 gene plays a major role in the oncogenesis of numerous tumors. ROS-1 rearrangement is found in 0.9–2.6% of non-small-cell lung cancers (NSCLCs), mostly lung adenocarcinomas, with a significantly higher rate of women, non-smokers, and a tendency to a younger age.
1. Introduction
Lung cancer represents the leading cause of cancer deaths worldwide, with more than 1.8 million deaths in 2020
[1]; 85% are non-small-cell lung cancers (NSCLCs) and 25% of them harbor oncogenic alterations that can be targeted by therapy. That is the case for patients whose tumors are positive for proto-oncogene tyrosine-protein kinase-1 (
ROS-1;
c-Ros oncogene-1)-gene fusion. Identification of that translocation makes patients eligible for targeted therapy. Prospective phase I/II trial results have shown the efficacy of crizotinib, that is a tyrosine-kinase inhibitor (TKI) targeting ROS, anaplastic lymphoma kinase protein (ALK) or mesenchymal-to-epithelial transition (MET) protein, and is now recommended as first-line therapy
[2][3][4]. Unfortunately, despite initial responses,
ROS-1-positive NSCLCs develop resistances to crizotinib, allowing tumor progression, notably brain metastases. More recently devised new molecules are active against crizotinib resistances and have good brain penetration.
2. ROS-1 Gene
The
ROS-1 gene was discovered in the 1980s as the product of the avian sarcoma virus RNA UR2 (University of Rochester)
[5]. This gene codes for 2347 amino acids that form a transmembrane protein sharing structural characteristics with the family of insulin receptors and ALK. The ROS-1 protein is composed of an extracellular domain containing a hydrophobic segment allowing transmembrane passage and an intracellular component containing a tyrosine-kinase domain with a terminal carboxyl
[6]. Its physiological role is poorly understood but study results suggest that wild-type ROS is involved during embryonic development as an initiator of signaling events for the differentiation of epithelial tissues
[7]. Early during the 2000s, the proto-oncogene role of ROS was first identified in brain tumors. A microdeletion in chromosome 6q21 is responsible for
ROS-1 fusion with a new fused-in-glioblastoma (
FIG) gene that is responsible for
ROS-1 overexpression and production of signals abnormally activating the tyrosine-kinase pathway, conferring its proto-oncogene role
[8]. In mouse models, FIG–ROS-1 transcript expression induces tumorigenesis and treatment with a small molecule, TAE684, inhibits growth of Baf3 cells overexpressing short and long isoform of
FIG-ROS-1, thereby defining
ROS-1 as an oncogenic “driver”.
The
ROS-1 proto-oncogene role in lung cancer was first reported in 2007 by Rikova et al., who identified two other protein fusion transcripts: a transmembrane solute transporter (SLC34A2) and a type-2 transmembrane protein (CD74)
[9]. Since then, improved sequencing techniques have enabled the discovery of increasing numbers of fusion partners
[10][11], whose proto-oncogene roles in numerous cancers is now clearly established
[12].
ROS-1 plays a major role in the activation of several signaling pathways associated with differentiation, proliferation, cell growth, and survival.
ROS-1 rearrangement, by forming phosphotyrosine-recruitment sites in the terminal tail of ROS, causes kinase-activity deregulation of the protein and abnormal activation of signaling pathways, mediated by tyrosine-phosphatase tumor-suppressor SHP1/SHP2, pro-mitotic protein extracellular signal-regulated kinase (ERK1/2), insulin-receptor substrate (IRS-1), phosphatidylinositol 3-kinase (PI3K) pathway, protein kinase B (AKT), mitogen-activated protein kinases (MAPKs), signal transducer and activator of transcription (STAT3), and VAV3
[13]. However, the ligand binding to its recruitment site initiating ROS-protein activation remains unknown. Moreover, study results suggest that activation of the pathways could depend on the fusion transcript. Thus, in the case of
CD74–ROS-1 fusion, the role played by phosphorylation of the E-Syt-1 (extended synaptotagmin-like) protein could confer the tumor cell with greater metastatic and invasive potential
[14].
3. Epidemiology, Clinical, and Histological Characteristics
ROS-1 rearrangement is present in approximately 0.9–2.6% of NSCLCs
[11][15][16][17][18]. Like ALK rearrangements, it is more frequent in young subjects, women and never-smokers
[11][15][16][17][18]. ROS-1-positive NSCLCs are predominantly lepidic, acinar, or solid adenocarcinomas, with more than 90% expressing thyroid transcription factor-1 (TTF1), diagnosed at an advanced stage (stage III–IV), with a higher frequency of brain metastases
[18][19][20][21][22]. More rarely,
ROS-1 rearrangement is found in other histological subtypes, e.g., squamous-cell carcinomas, pleomorphic carcinomas or large-cell carcinomas
[11][21][23][24][25]. Histological examination of these tumors mainly finds a solid architecture with round nuclei containing macronucleolus, “signet-ring cell”, and a close relationship with adjacent bronchioles
[23]. Imaging shows metastatic lymph-node tropism, often reported with less frequent extrathoracic metastatic sites than for ALK rearranged or epidermal growth factor receptor (EGFR)-mutated NSCLCs
[26]. Miliary forms of lung metastases have also been described
[27]. Several authors reported a heightened thromboembolic risk of ROS-1-rearranged tumors compared with NSCLCs harboring non-rearranged ROS-1
[28], and even rarer cases of thrombotic microangiopathies
[29] and disseminated intravascular coagulation
[30]. Although the underlying mechanisms remain to be elucidated, interaction between extracellular carcinoid mucins secreted into the bloodstream and platelet (P-) and leukocyte (L-) selectins could trigger platelet activation and the formation of microthrombi responsible for thromboembolic events
[31].
4. Molecular Characteristics
4.1. Fusion Partners
ROS-1 rearrangement occurs at a breakpoint in the
ROS gene at the 5′ end of exons 32, 34, 35, or 36, or introns 31 or 33
[12][32]. The most frequently seen fusion partners (
Table 1) are
CD74 (38–54%),
EZR (13%–24%),
SDC4 (9–13%), and
SLC34A2 (5–10%)
[32][33]. However, improved DNA- and RNA-sequencing techniques have enabled identification of numerous fusion partners, including
CCCKC6, TFG, SLMAP, MYO5C, FIG, LIMA1, CLTC, GOPC, ZZCCHC8, CEP72, MLL3, KDELR2, LRIG3, MSN, MPRIP, WNK1, SLC6A17, TMEM106B, FAM135B, TPM3, and
TDP52L1 [34]. The prognostic role of fusion partners is still being debated. The results of several studies showed that the presence of CD74–ROS-1 rearrangement was associated with longer progression-free survival (PFS) and overall survival (OS) than non-
CD74–ROS-1 rearrangement
[34]. However, that association with survival was not found in other studies
[32].
Table 1. Main ROS-1-fusion partners in ROS-1-positive non-small-cell lung cancers.
4.2. Oncogenic Co-Mutations
About 36% of
ROS-1-positive NSCLCs have oncogenic co-mutations
[46]. Their frequencies are lower compared to cancers with non-rearranged ROS-1
[33] and are usually mutually exclusive with Kirsten rat-sarcoma viral oncogene (KRAS), EGFR, and ALK mutations
[47]. However, rare co-mutations with other so-called driver oncogenes have been reported
[48][49]. Lambros et al., reported 15 cases of co-mutations between ROS-1 and EGFR with nine exon-19 deletions, one exon-20 insertion, and five L858R mutations. A first-line EGFR TKI obtained a tumor response or stability for 80% of the patients. ROS-1 rearrangement appeared in one tumor as a resistance mechanism under EGFR-TKI and crizotinib was able to obtain a partial response
[50]. A low percentage of co-mutations involving EGFR was reported in several papers
[46]. Extremely rare ALK and ROS-1 co-mutations have also been described
[48][51]. Given that both of these mutations are sensitive to crizotinib, the treatment choice is less problematic than for other co-mutations. The NSCLCs of all reported patients who had benefited from systemic treatment responded to crizotinib.
KRAS co-mutations can be present at diagnosis or progression. The scarce data available in this setting render evaluation of crizotinib sensitivity difficult. Based on six patients with KRAS–ROS co-mutation at diagnosis, a tumor response was obtained for 1; but not in KRASG13D- or G12V-mutation carriers
[46][47].
ROS-1-positive NSCLCs with MET amplification are extremely rare. Tang et al., explored the genetic co-alterations involving ALK, ROS-1, RET, and MET in a series of 15 patients and found one ROS-1–MET co-mutation but they provided no treatment information
[52]. One reported case of ROS-1–c-MET co-mutations experienced capmatinib (TKI anti-MET) failure followed by a 11 months crizotinib benefit
[53]. In contrast, Zeng et al., described a patient with ROS-1-positive NSCLC and MET amplification whose tumor progressed 1.5 months after starting crizotinib
[54].
Co-mutations between ROS-1–BRAF (v-RAF murine sarcoma viral oncogene homolog B) have been published, but without any information on anti-ROS-1 or anti-BRAF treatments in this context
[22][46]. To the best of knowledge,
ROS-1 co-mutation with rearranged-during-transfection translocation (
RET) or human epidermal growth factor receptor 2 (
HER2) have not been described to date. Other non-targetable co-mutations also have a prognostic role, for example, tumor protein 53 (
TP53), which was associated with shorter survival in one study
[55]. In another study, Zeng et al.
[54] found that patients whose NSCLCs had an exclusive
ROS-1 fusion lived longer without progression than those whose tumors carried a
ROS-1 fusion with co-mutations (15.5 vs. 8.5 months, respectively;
p = 0.0213).