1 Introduction
Rearranged during transfection (RET) is a transmembrane receptor tyrosine kinase. The ligand-activated wildtype RET kinase is involved in embryogenesis, spermatogenesis, neural and neuroendocrine and hematopoietic cell development [
1,
2]. Aberrantly activated RET by mutations or chromosomal rearrangements has been recognized as an oncogenic driver in many types of cancers, such as non-small cell lung cancer (NSCLC), in which RET fusions account for 1%–2% of the cases [
3–
5]. In addition to serving as a primary oncogene, RET fusions are found as acquired oncogenes in EGFR- or KRAS mutant-NSCLC that developed resistance to EGFR protein tyrosine kinase inhibitors (TKIs) or a KRAS
G12C inhibitor [
6–
9].
Similar to other protein tyrosine kinase fusion oncogene-positive NSCLC, RET fusion-positive NSCLC has a low tumor mutation burden (TMB), low PD-L1 expression, and has not benefited from immune checkpoint blockage therapy [
10–
13]. Early trials of TKIs in RET-altered thyroid cancer and NSCLC utilized multikinase TKIs that were developed originally to inhibit other kinases but cross-inhibited RET [
3]. These inhibitors gave low response rates and the responses were not durable in most patients, with a median objective response rate (ORR) of 0–37% and a median progression-free survival (PFS) of 2.3 months [
14]. In the last few years, two potent and selective RET TKIs, pralsetinib and selpercatinib, were developed [
15,
16]. Pralsetinib and selpercatinib gave higher response rates in RET-altered NSCLC and thyroid cancer and the responses were considered durable [
17–
22]. Pralsetinib and selpercatinib were initially approved for the treatment of advanced and metastatic RET-altered NSCLC and thyroid cancer in 2020. Interim reports of on-going basket trials of pralsetinib and selpercatinib in RET fusion-positive solid tumors showed pan-cancer efficacy [
23,
24]. In September 2022, the US Food and Drug Administration (FDA) updated the selpercatinib approval as a tumor-agnostic inhibitor for RET fusion-solid tumors. It is anticipated that pralsetinib will be approved similarly as a tumor-agnostic inhibitor in the near future.
While rapid progress has been made in RET-targeted cancer therapy, many challenges remain. In most patients, pralsetinib or selpercatinib treatment gave incomplete response [
17–
22]. The drug-tolerant, residual tumors will inevitably evolve to pralsetinib- or selpercatinib-resistance, leading to disease progression [
25–
29]. In this review, we summarize the progress in the current knowledge and gaps of oncogenic RET alterations and RET-targeted cancer therapy, and discuss the challenges in RET TKI resistance and the future direction toward a cure.
2 Oncogenic RET alterations
The human RET is encoded by 20 exons located on chromosome 10q11.21 and expressed as two major spliced forms (RET51, RET9) that differ in the C-terminal region [
1,
4]. The longer RET51 has an extra Grb2 SH2 domain binding motif at Y1096 and a stronger transformation activity [
30,
31]. Ligand binding of RET is mediated by one of its four glycosylphosphatidylinositol-linked co-receptors: glial cell-derived neurotrophic factor (GDNF) receptor-α family proteins GFRα1-4, or by the transmembrane co-receptor GDFα-linked protein (GFRAL). The co-receptors dictate the binding specificity of five RET ligands: GDNF, neurturin, artemin, persephin, and differentiation factor 15 (GDF15). Under normal physiological conditions, spatiotemporal expression of these co-receptors and ligands regulate RET dimerization and activation in cells where RET is expressed, which is also subject to spatiotemporal regulation during development. In human cancers, oncogenic RET activation occurs through two genetic mechanisms: single nucleotide variant (SNV) or short insertion/deletion (indel) mutations, and gene fusions through chromosomal rearrangements.
2.1 Oncogenic RET mutations
Germline and somatic RET mutations are associated with multiple endocrine neoplasia type 2 (MEN2). About 95% of MEN2A, which includes familial medullary thyroid carcinoma (FMTC) [
32], have mutations in the RET extracellular region. These mutations are observed most often in the cysteine-rich domain (CRD) that contains two pairs of intramolecular disulfide bonds (C609:C620; C611:C618) and the region connecting the CRD and the transmembrane domain (TM), which contains C630 and C634 that may form another intramolecular disulfide bond (Fig.1 and 1B) [
5,
33]. In addition to single amino acid substitution, short indels, most frequently in the CRD and the CRD-TM linker region, have also been observed in medullary thyroid carcinoma (MTC) [
20,
34]. Mutations in this region that disrupt the intramolecular disulfide bonds are believed to promote intermolecular disulfide bond formation, resulting in ligand-independent RET dimerization and activation. There was experimental evidence that RET
C634R/Y mutants form dimers when expressed in NIH3T3 cells [
35]. In the RET/co-receptor/ligand complex, such as the RET/GFRα1/GDNF complex (PDB id, 6Q2N; Fig.1), the two CDRs of RET are too far apart to form an intermolecular disulfide bond in the presence of GDNF. Thus, the intermolecular disulfide bond(s) formation between the CRD Cys residues may only occur in the absence of the ligand/co-receptor. In other words, the ligands serve as inhibitors for the intermolecular disulfide bond formation and thus may attenuate the oncogenic activation of the mutant RET. More likely, the intermolecular disulfide bond is formed between Cys630 or Cys634 (Fig.1). Alternatively, the CRD mutations may simply induce an active RET conformation without a disulfide bond formation. This possibility remains to be tested. The possible mechanism of a conformational change would also explain other mutations in the RET cadherin-liked domains (CLDs) that are not predicted to induce intermolecular disulfide bonds, if these mutations are proven to activate RET kinase.
In comparison, the more aggressive MEN2B is most often associated with somatic mutations in the cytoplasmic region of RET, which contains the protein tyrosine kinase domain [
1,
3,
33]. The M918T mutation is found in ~90% of sporadic MTC cases. Enzymatic and structural studies of the M918T mutation, which is located in the carboxyl flank of the activation loop, showed that the M918T mutation increased ATP affinity and relieved a cis-autoinhibitory conformation of the activation loop to promote intermolecular autophosphorylation [
36]. Other cytoplasmic mutation sites identified in MEN2B and MEN2A include E768, L790, V804, A883, and S891 (Fig.1). Fig.1 lists pathologic and likely pathogenic RET mutations annotated by the Associated Regional and University Pathologists (ARUP) Laboratories as of October 2022 [
37] and additional RET indel mutations reported in the clinical trial of selpercatinib in thyroid cancer patients [
20].
We previously examined RET variants from 46 612 tumor samples in the curated data sets from cBioportal [
4,
38], which showed 701 RET missense mutations in 413 amino acid residues located throughout the molecule. Therefore, there are many more RET variants in various types of human cancer as reported in the cBioportal or COSMIC [
5] data sets than those that have been functionally annotated. It is a prerequisite to distinguish oncogenic driver RET mutations from passenger mutations and loss-of-function (LOF) mutations to select patients for RET-targeted treatment. While the list of oncogenic RET mutations remains to be compiled, a simple first step is to analyze whether the mutant RET transcript or protein is expressed in those tumor cells.
2.2 Oncogenic RET fusions
Somatic RET fusion oncogenes were first discovered in papillary thyroid carcinoma (PTC) in 1987 using a DNA transfection assay [
39]. RET fusion oncogenes remain the most prevalent in PTC, consisting of 5%–10% of cases [
2]. Nine different RET fusion oncogenes in PTC were initially numerically termed RET/PTC1 to RET/PTC9, but the number of different RET fusions continues to expand [
40]. Another major advance that accelerates the field of RET-targeted cancer therapy was the discovery of RET fusion oncogenes in 1%–2% of NSCLC, mostly lung adenocarcinoma [
31,
41–
44]. Today, over 100 different RET fusion partners have been identified in many different types of solid tumors and blood cancers such as breast cancer, ovarian cancer, prostate cancer, head and neck cancer, pancreatic cancer, gastric cancer, small intestine cancer, colon cancer, cholangiocarcinoma, sarcoma, chronic myelomonocytic leukemia (CMML), and acute myeloid leukemia (AML) [
2–
4,
23,
24,
40,
45] (Table S1). The most common RET fusion partner in NSCLC is kinesin family member 5B (KIF5B) followed by the coiled-coil domain containing 6 (CCDC6) and nuclear receptor coactivator 4 (NCOA4) [
46]. In PTC, the most common RET fusion partner is CCDC6 followed by NCOA4 and KIF5B (Fig.2). The transcripts of RET fusion oncogenes consist of 5′-region derived from the fusion partners and the 3′-region derived from the RET. The breakpoints on the RET gene most often occur in intron 11 and less often in introns 7 and 10, resulting in chimeric proteins containing the entire RET tyrosine kinase domain to the RET C terminus fused with the fusion partners located at the N-terminal portion of the protein (Fig.2).
The fusion partners contribute at least three functions to the RET fusion oncogenes. First, the fusion partners contain protein–protein dimerization domains, such as the coiled-coil domain in CCDC6. This results in ligand-independent dimerization of the RET kinase. Oligomerization is the universal mechanism of protein tyrosine kinase activation. Second, by locating at the 5′ region, the fusion partners control the expression of the RET fusion oncogenes. This allows the RET tyrosine kinase to express in cell types that do not normally express it or increases the level of RET tyrosine kinase expression. For instance, KIF5B-RET fusions allow the RET kinase to express 2×–30× higher level in NSCLC [
42]. Moreover, the expression levels of different RET fusion protein correlate with the aggressiveness of the tumors and the response to treatment. The relatively low levels of CCDC6-RET and NCOA4-RET in PTC correlates with the relatively benign PTC phenotype [
1,
47]. RXDX-105 (CEP-32496) is a VEGFR-sparing multikinase inhibitor with RET inhibiting activity [
48,
49]. In a phase I/Ib clinical trial of RXDX-105 in NSCLC, none of the responders were KIF5B-RET [
50], corresponding to significantly higher levels of KIF5B gene expression than expression of other partner genes (CCDC6, EML4, PARD3) in NSCLC [
2]. RXDX-105 was no longer in clinical development to treat RET-altered cancers. In the selpercatinib basket trial of RET fusion-positive solid tumors [
24], the ORRs for KIF5B-RET, CCDC6-RET, and NCOA4-RET tumors were 25.0% (
n = 4), 66.7% (
n = 6), and 31.3% (
n = 16), respectively. The median duration of responses (DOR) for KIF5B-RET, CCDC6-RET, and NCOA4-RET tumors were 9.2 months, not-reached (NR), and 24.5 months, respectively. Again, the KIF5B-RET fusion had the least favorable response to selpercatinib. However, in the RET fusion-positive NSCLC phase I/II trials of pralsetinib, KIF5B-RET and CCDC6-RET fusion had similar ORR (62% vs. 58%) [
21]. Although it was reported that the KIF5B fusion had enhanced signaling activity compared to CCDC6 and NCOA4 fusions [
51], the higher level of KIF5B expression may contribute to the poorer responses to RXDX-105 and selpercatinib.
The third function of the RET fusion partners is to regulate the subcellular localization and complex formation of the fusion proteins. It was demonstrated that CCDC6 promotes the assembly of membraneless cytoplasmic condensates (phase separation) that serve as the platform for the CCDC6-RET fusion kinase to activate the RAS-ERK1/2 MAP kinase signaling pathway [
52]. In a
Drosophila kinome screen that compared CCDC6-RET and NCOA4-RET, six genes were identified as genetic modifiers of lethality caused by CCDC6-RET, whereas 11 genes were identified as genetic modifiers of lethality caused by NCOA4-RET in
Drosophila [
53], suggesting that signaling complexes of CCDC6-RET and NCOA4-RET are not identical. Interestingly, CCDC6-RET had milder phenotypes than NCOA4-RET in the
Drosophila model [
53], which is in parallel to the better response of CCDC-RET-positive solid tumors than the response of NCOA4-RET-positive solid tumors in the selpercatinib basket trial [
24].
2.3 Detection of RET alterations
Immunohistochemistry (IHC), fluorescent
in situ hybridation (FISH), DNA-polymerase chain reaction (DNA-PCR), reverse transcription-PCR (RT-PCR and RT-qPCR), DNA-based next-gen sequencing (DNA NGS), and RNA NGS methods have been used to detect RET mutations and fusions. IHC, because of its low sensitivity and specificity, is not a recommended screening method for detection of RET alterations by the European Society for Medical Oncology (ESMO) Translational Research and Precision Medicine Working Group [
54]. The advantages, disadvantages, and usage of technologies to detect RET alterations are summarized in Tab.1. FISH (either break-apart or defined fusions) is used to detect RET fusions in tumor tissue specimens. It has single-cell resolution and rapid turnaround time. Many targeted DNA-seq and RNA-seq assays for detecting RET alternations in tissues or cell-free DNA (cfDNA) are commercially available. The Thermo Fisher Scientific’s Oncomine DX Target Test has been approved by the US FDA for detection of RET mutations and fusions as a companion diagnostic (CDx) of pralsetinib and selpercatinib. For precision management of RET-altered cancers, the noninvasive cfDNA targeted sequencing assays offer several advantages, including readily availability of series collection of plasma samples for longitudinal monitoring of tumor response to treatment, and the possibility of detecting heterogenous variants originated from different metastases, querying acquired RET mutations and multiple genes to identify mechanisms of acquired resistance. For monitoring acquired mutations, it is critical to choose a cfDNA assay that covers all RET kinase domain coding exons.
3 Pralsetinib and selpercatinib are first-in-class RET-selective TKIs
Pralsetinib (BLU-667) and selpercatinib (LOXO-292) (Fig.3) are the first two compounds that have been developed to inhibit RET and progressed to regulatory approval for RET-targeted cancer therapy. They are potent RET kinase inhibitors. In
in vitro kinase assay, pralsetinib had a 0.4 nmol/L RET IC
50, which is 10×–28× more potent than that of vandetanib and cabozantinib [
15]. Similarly, selpercatinib was a potent inhibitor of RET kinase with an
in vitro IC
50 of 0.92 nmol/L. While the
in vitro VEGFR2 IC
50 of vandetanib and cabozantinib were 4 and 2 nmol/L, the
in vitro VEGFR2 IC
50 of pralsetinib and selpercatinib were 35 nmol/L and 68 nmol/L. Pralsetinib and selpercatinib are considered highly selective RET inhibitors. A few other protein tyrosine kinases inhibited by pralsetinib are DDR1, JAK1/2, TRKA/C, FLT3, PDGFRB, FGFR1 with > 10-fold less potency (Fig.3) [
15]. Other protein tyrosine kinases inhibited by selpercatinib at higher concentrations are VEGFR1/3.
Multikinase RET inhibitors such as vandetanib, cabozantinib, lenvatinib, and nintedanib are ineffective on RET
V804M/L gatekeeper mutants [
55,
56]. In fact, gatekeeper mutation is the major on-target mechanism of acquired resistance to these multikinase TKIs. In contrast, pralsetinib and selpercatinib display a similar potency on the wild type and V804M/L gatekeeper mutated RET [
15,
27]. Crystal structures of pralsetinib and selpercatinib show that pralsetinib and selpercatinib use an unconventional binding mode to occupy both the front and back clefts of the drug binding pockets by wrapping around the area outside the gate wall formed by the side chain of the invariable K758, thus avoid the interference caused by V804M/L mutations [
27]. In comparison, vandetanib [
57] or nintedanib [
58] inserts into the gate formed by V804 and K758, and thus is subjected to binding disruption when the variable gatekeeper V804 is mutated.
The phase I/II clinical study of pralsetinib in thyroid cancer, NSCLC, and other advanced solid tumors (ARROW, ClinicalTrials.gov ID: NCT03037385) and the phase I/II study of selpercatinib in advanced solid tumors, RET fusion-positive solid tumors, and MTC (LIBRETTO-001, ClinicalTrials.gov ID: NCT03157128) were started in 2017. After phase I dose escalation trials, the pralsetinib dose of 400 mg once a day (QD) and the selpercatinib dose of 160 mg twice a day (BID) were selected for phase II trials.
In the LIBRETTO-001 phase I/II trials of RET-altered NSCLC and thyroid cancer, selpercatinib had 61%–84% ORR in RET fusion-positive NSCLC and PTC, and in RET mutant MTC, depending on whether the patients had prior-treatments, and the median DOR was 18.4–28.6 months (Tab.2) [
18,
20,
22]. In the ARROW phase I/II trials of RET-altered NSCLC and thyroid cancer, pralsetinb gave 59%–89% ORR in RET-altered NSCLC and thyroid cancer and a median DOR of 22.3 months or NR (Tab.2) [
17,
19,
21]. These ORR and DOR were much improved over the previously tested multikinase RET TKIs and were in the similar range as other regulatory approved protein tyrosine kinase-targeted drugs, such as EGFR TKIs in NSCLC [
59–
61]. The most common grade ≥ 3 adverse events of selpercatinib were hypertension (19.7%), hepatotoxicity (8.8%–11.4%), diarrhea (5.0%), and electrocardiogram QT prolonged (4.8%) [
22]. The most common grade ≥ 3 adverse events of pralsetinib were neutropenia (18%), hypertension (10%), increased blood creatine phosphokinase (9%), and lymphopenia (9%). In RET-fusion solid tumors other than NSCLC and thyroid cancer, selpercatinib had an ORR of 43.9% and a DOR of 24.5 months, whereas pralsetinib had an ORR of 57% and a DOR of 11.7%. Furthermore, selpercatinib and pralsetinib are intracranially active [
21,
62]. Brain is a frequent metastasis site of RET-fusion NSCLC [
63]. Based on the initial LIBRETTO-001 and ARROW trial data that showed high ORR and DOR, and acceptable toxicity [
17–
20], the US FDA approved selpercatinib and pralsetinib as the first two RET-targeted TKIs for metastatic RET-fusion positive NSCLC and thyroid cancer, and advanced and metastatic RET mutant MTC in 2020. Following the basket trial [
24], US FDA updated the selpercatinib approval to treat advanced and metastatic RET-fusion solid tumors.
4 Resistance to pralsetinib and selpercatinib
Besides primary resistance, at least two major mechanisms are known to contribute to acquired TKI resistance in targeted cancer therapy (Fig.4). The first mechanism is secondary target mutations that interfere with TKI binding, thus resulting in target-reactivation. For instance, EGFR gatekeeper T790M mutation is the major mechanism of acquired resistance to gefitinib and erlotinib, and RET gatekeeper V804M/L mutations cause resistance to vandetanib. This mechanism is easiest to detect and can be overcome by developing new TKIs that inhibit these mutants. For example, osimertinib was developed to inhibit the gefitinib/erlotinib-resistant EGFR T790M mutant [
59]. The second mechanism is acquisition of alternative driver oncogene, which bypass the need of the targeted kinase for driving tumor growth, such as acquisition of RET fusion in osimertinib-treated EGFR mutant-positive NSCLC [
6,
64]. This mechanism can be alleviated by combination of drugs to co-inhibit both drivers, such as combination of osimertinib and pralsetinib [
6,
64].
To identify RET mutations that caused selpercatinib or pralsetinib resistance in the laboratory, we utilized the RET kinase-dependent BaF3/KIF5B-RET cells that express the oncogenic KIF5B-RET [
65]. In the first approach, we cultured BaF3/KIF5B-RET cells with increasing concentrations of selpercatinib or pralsetinib. In the second approach, we expressed a pool of randomly mutated KIF5B-RET in BaF3 cells and selected cell clones that grew in the presence of selpercatinib or pralsetinib and had active KIF5B-RET kinase. The cDNA from ~100 individual selpercatinib- or pralsetinib-resistant cell clones were sequenced to identify RET kinase domain mutations. The G810C/S mutations located at the RET kinase C-lobe solvent-front site, Y806C/N mutations located at the hinge region, and V738A mutation located at the RET kinase β2 strand were found in the selpercatinib-resistant clones [
27]. In addition to G810S, Y806C/N, and V738A, L730V/I located at the N-lobe solvent-front site (roof), and E732K located near the Gly-rich region were found in the pralsetinib-resistant clones [
66]. Importantly, no gatekeeper mutations were identified. In clinic, RET
V804M/L mutant tumors were resistant to vandetanib but responded to selpercatinib [
27], consistent with the activity of selpercatinib on the RET gatekeeper mutants. We then cross-profiled these mutations and the G810R mutation found in selpercatinib-resistant NSCLC patients. The solvent-front G810C/S/R were strongly resistant to selpercatinib and pralsetinib, in the order of G810R > G810C > G810S, whereas Y806C/N, V738A, and E732K had modest resistance to selpercatinib and pralsetinib [
27,
66]. Interestingly, the L730V/I mutations were resistant to pralsetinib but remained sensitive to selpercatinib. This finding suggests that selpercatinib is a potential option for a second line treatment when pralsetinib-adapted L730V/I mutations caused pralsetinib resistance during the treatment. Several clinical studies that analyzed plasma cfDNA or tumor samples of RET-altered cancer patients who developed selpercatinib resistance identified acquired RET G810C/R/S/V and Y806C/N mutations [
25–
27,
29]. Among these mutations, G810C was observed most often. In pralsetinib-treated patients who acquired resistance, mutations at the G810 and L730 sites were found [
67].
In a case study [
68], a KIF5B-RET-positive lung cancer patient who had prior radiotherapy and chemotherapy responded to selpercatinib but the disease progressed after 10 months. cfDNA analysis did not detect an actionable new lesion. However, next-generation sequencing analysis (Illumina TST-170) of the pre-treatment tumor and resistant tumor biopsy samples at a later time revealed a new KHDRBS1-NTRK3 fusion in the resistant tumor sample. The NTRK3 fusion was not detected in cfDNA because NTRK3 fusion was not in the cfDNA assay panel. Laboratory experiments confirmed that the KHDRBS1-NTRK3 fusion had constitutively active tyrosine kinase activity that could transform BaF3 cells into cytokine-independence, which was inhibited by the Trk tyrosine kinase inhibitor larotrectinib [
68].
A bypass mechanism of resistance in oncogenic tyrosine kinase-targeted NSCLC therapy is MET amplification. Not surprisingly, MET amplification was observed in NSCLC patients treated with selpercatinib [
28,
29,
69]. The MET amplification may be pre-existing and enriched, or only evident post selpercatinib treatment [
28]. In a preclinical experiment, increased MET expression caused selpercatinib resistance [
28]. Other acquired genetic mechanisms of resistance to selpercatinib were KRAS, NRAS, BRAF activating mutations and FGF1 amplification [
29].
The mechanism of primary resistance is less clear. In a study that addressed this issue [
29], it was found that two patients with primary resistance to selpercatinib had KRAS G12D and G12V mutations. In one of these cases, the KRAS mutation and TAF3-RET fusion resided in different tumor cells. In the same study that analyzed RET-altered NSCLC and thyroid cancer, patients with co-mutations in the PI3K pathway, including PIK3CA and PTEN mutations, had a 91% of clinical benefit rate from selpercatinib treatment. Thus, PI3K pathway lesions do not appear to be a major cause of primary resistance. Interestingly, no statistically significant association was found between TMB and intratumor heterogeneity among the RET-altered NSCLC and thyroid cancer cases with the depth or durability of selpercatinib response. While more cases are needed to confirm the result, this may reflect the overall low TMB in RET-altered cancer. Nevertheless, RET fusion-positive patients with concurrent TP53 mutations had shorter median PFS in selpercatinib treatment.
5 Next generation of RET-selective kinase inhibitors
A number of next-generation of RET TKIs have been reported. These new RET TKIs are being developed to address the selpercatinib- and pralsetinib-resistant RET mutations. Those that have progressed to clinical trials are listed in Tab.3. TPX-0046 is the first of these new RET TKIs to reach clinical trial. TPX-0046 is a macrocyclic TKI (Fig.3) that potently inhibits RET and SRC family kinases (SFKs) (
in vitro IC
50: 1 nmol/L). It also cross-inhibits several other kinases with < 10-fold difference in
in vitro IC
50, including receptor tyrosine kinases FGFR1/2, TRKA/B/C, FLT3, and non-receptor tyrosine kinases BMX and JAK2 [
70]. TPX-0046 is effective on RET G810C/S/R/N solvent-front mutations but may be subject to resistance due to the V804M gatekeeper mutant [
71,
72]. LOXO-260 is a new RET-selective TKI developed to target both solvent-front and gatekeeper mutations [
73]. TAS0953/HM06 is another RET-selective TKI reported to be active against RET, RET solvent-front G810, hinge Y806, and gatekeeper V804 mutations, and highly selective against VEGFR2, FLT, PDGFR, FGFR, DDR, and JAK families of protein tyrosine kinases [
74]. An issue with the previous multikinase RET TKIs was believed to be attributed to potent inhibition of VEGFR2. BOS172738 was reported as a potent and selective RET TKI with > 300-fold selectivity against VEGFR2 [
75]. SY-5007 and HA121-28 are two other new RET TKIs in clinical trials. HA121-28 is an EGFR, VEGFR, and RET TKI. The latest to enter clinical trial is APS03118 [
76]. APS03118 has a low nanomolar
in vitro IC
50 on RET solvent-front G810C/R/S and gatekeeper V804M/L/E mutants, and 130-fold selectivity over VEGFR2.
6 Unmet challenges
Although selpercatinib and pralsetinib gave high ORR, the CR was < 10%. The drug-tolerated residual tumor cells are the seeds that give arise to the TKI-resistant tumors. The new generation of RET-selective TKIs can be used to circumvent the selpercatinib- and pralsetinib-adapted RET mutations, and thus prolong the duration of tumor response to the targeted therapy. However, if the residual tumors persist, tertiary RET mutations will inevitably emerge to cause resistance to the secondary TKIs, resulting in tumor progression.
Moreover, the RET-TKI-treated tumors could acquire one or more alternative oncogenes to bypass the requirement of the RET oncokinase for tumor growth. This may occur within the same cancer cells, in different cancer cells within the same tumors, or in different tumors in a patient. Because only limited number of pre-defined, actionable driver oncogenes are included in the cfDNA assays, a novel oncogene could evade detection in the cfDNA assay if the resistant tumor sample was not available immediately at the time of tumor progression [
68]. Thus, it is critical to expand the panel of clinically actionable oncogenes in the non-invasive cfDNA assay in order to manage the disease for a longer duration. Because of the intra-tumor and inter-tumor heterogeneity, and the continuing evolution of the tumors during the TKI treatment, simultaneous action against multiple driver oncogenes may be prohibited due to adverse events.
A key to prevent progression of the responsive tumors to RET TKI-resistant tumors is to eliminate the RET TKI-tolerated persisters before they acquire a key genetic event, either on target or off target that can drive fast tumor growth even in the presence of the RET TKI. The non-genetic or pre-existing genetic events that allow survival of the persisters under the RET TKI pressure may involve the rewiring of oncogenic signaling pathways after the RET kinase activity is inhibited by the TKI, epigenetic changes in the tumor cells, and the influence of the tumor microenvironment (TME). Conceivably, because of intra-tumor and inter-tumor heterogeneity, multiple mechanisms could co-exist in a patient to support the survival of RET TKI-tolerated persisters. It is a daunting task and unlikely to be effective to co-target RET and an identified supporting factor in a patient one at a time. The key to solve this complex problem of heterogenous mechanisms of RET TKI-tolerated persisters is to find the convergent point among different mechanisms that support the RET TKI-tolerated residual tumors, and then co-target RET and this convergent point.
7 Conclusions
RET alterations by mutations or gene fusions occur in diverse cancers from different organs but are most often associated with NSCLC and thyroid cancer. Rapid progress has been made in the last few years with the development of the potent and selective RET TKIs pralsetinib and selpercatinib. Identification of pralsetinib- and selpercatinib-adapted RET kinase domain mutations promoted development of the next-generation of RET TKIs. However, the residual RET TKI-tolerated tumors in most patients remain an unmet challenge that prevents a cure.