BRAF and KRAS mutations in metastatic colorectal cancer: future perspectives for personalized therapy

Abstract

Colorectal cancer (CRC) is one of the most commonly diagnosed cancers worldwide and 30% of patients with CRC experience metastasis. Patients with metastatic colorectal cancer (mCRC) have a 5-year overall survival rate of <10%. V-raf murine sarcoma viral oncogene homolog B1 (BRAF) and V-Ki-ras2 Kirsten ratsarcoma viral oncogene homolog (KRAS) mutations are mostly studied in mCRC, as clinical trials found that first-line chemotherapy with anti-epidermal growth factor receptor agent confers limited efficacy for mCRC. Treatment decisions for early-stage mCRC do not consider BRAF or KRAS mutations, given the dramatically poor prognosis conferred by these mutations in clinical trials. Thus, it is necessary to identify patients with mCRC harboring BRAF or KRAS mutations to formulate rational therapeutic strategies to improve prognosis and survival. BRAF and KRAS mutations occur in ∼10% and ∼44% of patients with mCRC, respectively. Although the survival rate of patients with mCRC has improved in recent years, the response and prognosis of patients with the aforementioned mutations are still poor. There is a substantial unmet need for prospective personalized therapies for patients with BRAF- or KRAS-mutant mCRC. In this review, we focus on BRAF and KRAS mutations to understand the mechanisms underlying resistance and improving the response rate, outcomes, and prognosis of patients with mCRC bearing these mutations and to discuss prospective personalized therapies for BRAF- and KRAS-mutant mCRC.

metastatic colorectal cancer, epidermal growth factor receptor, KRAS mutation, BRAF mutation, personalized therapy

Introduction

Colorectal cancer (CRC) is the third most commonly occurring cancer in men and the second most common cancer in women. More than 1.8 million new cases were diagnosed in 2018. Approximately 20% of new colorectal cancer cases are metastatic at the time of diagnosis and another 20% of cases develop into metastatic colorectal cancer (mCRC), which has a significantly lower survival rate [1, 2]. In recent decades, monoclonal antibodies (mAbs) targeting vascular endothelial growth factor (VEGF) and the epidermal growth factor receptor (EGFR) have been used as the first-line treatments for mCRC [3, 4]. Studies indicate that, in the first-line treatment setting, addition of an anti-VEGF antibody therapy improves the median overall survival (OS) [5]. However, unlike VEGF inhibitors, the efficacy of anti-EGFR agents, such as cetuximab and panitumumab, is limited to patients with wild-type (wt) V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) tumors without the V-raf murine sarcoma viral oncogene homolog B1 (BRAF) mutation. Given the poor prognosis of BRAF- and KRAS-mutant mCRC, here, we focus on the BRAF and KRAS mutations, which exhibit two consensus subtypes.

KRAS, NRAS, and HRAS are some of the most well-studied rat sarcoma virus (RAS) subfamily proteins because of their significant role in cancer [6]. Of these RAS mutations, KRAS mutations (85%) are observed most frequently, followed by NRAS (15%) and HRAS (1%). KRAS mutations occur in approximately 44% of mCRC, with the majority observed in codons 12 and 13 of exon 2 (80% are G12D, G12V, G12C, G12A, and G13D) and less commonly in codon 61 of exon 3 (5% are Q61H, Q61L, and Q61R) and codon 146 of exon 4 (2% are A146T and A146V) [7]. Analyses of clinical-trial data strongly suggest that KRAS codon 12 or 13 mutations are a major predictive biomarker for resistance to anti-EGFR therapy in patients with mCRC [8–11]. These sites play a part in the activation of v-Raf murine sarcoma viral oncogene (RAF) kinases and mutations in these sites cause constitutive KRAS activation (without ligand-receptor binding), producing resistance to anti-EGFR agents [11, 12].

NRAS, the RAS homolog, has identical effector binding domains to KRAS. NRAS protein is part of the mitogen-activated protein kinase (MAPK) cascade [13, 14]. Membrane-bound NRAS proteins (189 amino acids) have GTPase activity and play a role in cell-signal transduction. Mutations in NRAS may result in structural activation and increased affinity for downstream effectors of the MAPK-signaling pathway—a key event in many cancers. Hence, NRAS mutations in codons 12, 13, 61, and 146 yield similar effects to KRAS activation [15–17]. NRAS mutations also have predictive value in anti-EGFR treatment resistance. Randomized clinical trials have shown that NRAS mutations in exons 2, 3, and 4 can be predictive markers of a lack of clinical benefit of anti-EGFR treatment when given in combination with chemotherapy in the first-line setting [14–16]. Patients bearing wt-RAS CRC treated with anti-EGFR therapy showed higher response rates and OS than those with RAS mutations (KRAS and NRAS exons 2, 3, and 4) [18–20]. Patients with mCRC with mutations in NRAS exons 2, 3, and 4 showed a poor overall response rate (ORR) and an adverse effect on progression-free survival (PFS) and OS [18, 21, 22]. In another study, patients with NRAS mutations developed more distant metastases and the 3-year risk in patients with NRAS mutations was 40.0% compared with 12.2% in patients with wt NRAS [23], indicating the significance of NRAS mutations in patients with CRC.

BRAF belongs to the RAF family of kinases, which also includes ARAF and CRAF [24]. BRAF mutations are less frequent than KRAS mutations in mCRC: ∼10% of patients with mCRC bear mutations in BRAF. A single substitution missense mutation in V600 comprises approximately 90% of these BRAF mutations and nearly 90% of V600 mutations involve substitution to glutamic acid (V600E) [25, 26]. V600 is necessary for BRAF to remain inactive in the absence of an activation signal from RAS. The V600E substitution results in constitutive MAPK phosphorylation and subsequent RAF-mitogen-activated protein kinase (MEK)-extracellular signal-regulated kinase (ERK) signal transduction [27–29]. BRAF mutations are associated with poor prognosis in mCRC and mortality is nearly triple that of patients with wt-BRAF [30, 31].

Over the past two decades, the molecular characterization of mCRC has been revolutionized by the implementation of routine KRAS- and BRAF-mutation testing. Reportedly, BRAF and KRAS mutations are associated with a very poor prognosis. In terms of treatment, patients with these mutations show poor response to anti-EGFR treatment [32, 33]. Studies analysing the KRAS- and BRAF-mutation status of patients with mCRC found that patients with these mutations had lower PFS and OS rates than those with wt-KRAS and BRAF. Therefore, identification of BRAF or KRAS mutations can play an important role in improving the response rate and survival in mCRC treatment.

Although, with the advancement in systemic chemotherapy, targeted therapies, and surgical techniques, the survival of patients with mCRC has been improved in the past two decades, a few patients show poor response to treatment and continue to have a poor prognosis; particularly, patients with BRAF and KRAS mutations show a poor response and inferior survival. Thus, this review will focus on recent advances in personalized therapies for BRAF- and KRAS-mutant mCRC. Here, we discuss the mechanisms of BRAF and KRAS mutations that cause resistance to targeted therapies and propose future perspectives for personalized therapy for these patients.

KRAS and BRAF mutations as prognostic and predictive biomarkers in mCRC treatment

mAbs that target EGFR can achieve antitumor efficacy and are used as part of the standard treatment for mCRC. However, in clinical trials comparing therapeutic strategies plus anti-EGFR agents cetuximab or panitumumab, the outcomes are not satisfactory [5]. As shown in Table 1, PFS and OS were reduced in patients treated with an anti-EGFR agent compared with those in patients treated without this agent. In the CAIRO2 trial, patients received either capecitabine–bevacizumab (CB) or capecitabine–bevacizumab–cetuximab (CBC). The addition of cetuximab significantly decreased the median PFS and OS [34]. The PACCE study, a phase III trial, evaluated the efficacy of oxaliplatin–bevacizumab (OB) and irinotecan–bevacizumab (IB), with or without panitumumab. Both PFS and OS improved following OB/IB treatment than following OBP/IBP treatment [35]. Both these phase III studies (CAIRO2 and PACCE) indicate that the addition of an anti-EGFR antibody resulted in an inferior outcome.

Table 1.

Effect of treatment with anti-EGFR agents for patients with mCRC harboring BRAF or KRAS mutations

Trial	Treatment	PFS (95% CI, months)	OS (95% CI, months)
CAIRO2	CB	10.7 (9.7–12.3)	20.3 (17.8–24.7)
	CBC	9.4 (8.4–10.5)	19.4 (17.5–21.4)
PACCE	OB	11.4 (10.5–11.9)	24.5 (20.4–24.5)
	OBP	10.0 (8.9–11.0)	19.4 (18.4–20.8)
	IB	11.7 (9.0–13.2)	20.5 (19.8 to NE)
	IBP	10.1 (8.2–13.7)	20.7 (17.8 to NE)

CB, capecitabine–bevacizumab; CBC, capecitabine–bevacizumab–cetuximab; OB, oxaliplatin-based chemotherapy and bevacizumab; OBP, oxaliplatin-based chemotherapy, bevacizumab, and panitumumab; IB, irinotecan-based chemotherapy and bevacizumab; IBP, irinotecan-based chemotherapy, bevacizumab, and panitumumab; NE, not estimable.

Given these unexpected results and the results of studies at the preclinical and clinical levels, which suggest that the presence of mutated KRAS or BRAF is a predictive marker of anti-EGFR resistance [29, 36–45], we analysed the KRAS- and BRAF-mutant status and corresponding PFS and OS of each treatment group across several trials (Table 2). In these trials, the PFS and OS are lower in patients with KRAS and BRAF mutations than in patients with wt-KRAS and BRAF when undergoing anti-EGFR treatment. [34, 35, 46–48]. The effects of treatment without an anti-EGFR agent are greater than those with an anti-EGFR agent in patients with KRAS and BRAF mutations. In addition, one study in particular reported that, out of 79 patients with CRC, 11 had BRAF mutations and none of these 11 patients responded to anti-EGFR treatment [37].

Table 2.

Clinical trials of the effect of treatment with anti-EGFR agents for patients with mCRC harboring BRAF or KRAS mutations

Name of study	Mutated gene type	Phase of clinical trial	No. of patients	Therapeutic strategy	Patients with KRAS/ BRAF mutations		PFS (months)				OS (months)
					No.	%	Wild-type		Mutant		Wild-type		Mutant
							With anti-EGFR	Without anti-EGFR	With anti-EGFR	Without anti-EGFR	With anti-EGFR	Without anti-EGFR	With anti-EGFR	Without anti-EGFR
PACCE	KRAS	III	664	OB/OBP	260	39	9.8	11.5	10.4	11.0	20.7	24.5	19.3	19.3
			201	IB/IBP	86	43	10.0	12.5	8.3	11.9	NE	19.8	17.8	20.5
CAIRO2		III	520	CB/CBC	206	40	10.5	10.6	8.1	12.5	21.8	22.4	17.2	24.9
CRYSTAL		III	1,063	FOLFIRI, cetuximab	397	37	9.9	8.4	7.4	7.7	23.5	20.0	16.2	16.7
PRIME		III	1,096	FOLFOX4, panitumumab	440	40	9.6	8.0	7.3	8.8	23.9	19.7	15.5	19.3
CRYSTAL	BRAF	III	519	FOLFIRI, cetuximab	45	8.7	10.4	12.2	6.6	5.9	21.5	24.6	15.2	15.0
TAILOR		III	393	FOLFOX4, cetuximab	52	13	9.3	7.5	2.0	3.8	–	–	–	–

NE, not estimable; OB, oxaliplatin-based chemotherapy, bevacizumab; OBP, oxaliplatin-based chemotherapy, bevacizumab, panitumumab; IB, irinotecan-based chemotherapy, bevacizumab; IBP, irinotecan-based chemotherapy, bevacizumab, panitumumab; CB, capecitabine, oxaliplatin, bevacizumab; CBC, capecitabine, oxaliplatin, bevacizumab, cetuximab; FOLFIRI, irinotecan, infusional fluorouracil, leucovorin; FOLFOX4, infusional fluorouracil, leucovorin, and oxaliplatin.

Several clinical trials demonstrated that mutated KRAS and BRAF genes are predictive markers of outcomes in mCRC treatment (Table 3). In addition to the trials discussed above, the table includes several trials that indicate the predictive effect of BRAF and KRAS mutations [18, 31, 34, 46, 48–52]. These trials clearly demonstrated that the mutation statuses of BRAF and KRAS were predictors of outcome for mCRC patients when undergoing anti-EGFR treatment. Patients harboring BRAF or KRAS mutations showed poor prognosis and lower survival in mCRC.

Table 3.

BRAF and KRAS mutations as prognostic factors in clinical studies of first-line treatment for mCRC

Name of study	Phase of clinical trial	Therapeutic strategy	Key outcomes in patients	Prognostic finding
MRC COIN	III	OFC	Median OS, 8.8 vs 20.1 months for patients with wild-type BRAF (P < 0.001); median OS, 14 vs 20.1 months for patients with wild-type KRAS (P < 0.001)	KRAS and BRAF mutations have a strong prognostic effect
CRYSTAL	III	FOLFIRI, cetuximab	Median OS, 8.8 vs 20.1 months for patients with wild-type BRAF (P < 0.001); median OS, 14 vs 20.1 months for patients with wild-type KRAS (P < 0.001)	KRAS and BRAF mutations are important indicators of poor prognosis
PRIME	III	FOLFOX4, panitumumab	In the mutant KRAS stratum, PFS was significantly reduced in the panitumumab–FOLFOX4 arm vs the FOLFOX4 arm; median OS, 15.5 vs 19.3 months, respectively (P = 0.068)	KRAS testing is important for patients with mCRC
CAIRO2	III	CBC	Median PFS, 8.1 and 12.5 months in patients with KRAS-mutant vs 10.5 and 10.6 months in patients with KRAS wild-type tumors in two arms	Mutation status of the KRAS gene was a predictor of outcome in the cetuximab group
			Median OS, 17.2 and 24.9 months in patients with KRAS-mutant tumors vs 21.8 and 22.4 months in patients with KRAS wild-type tumors with CB and CBC, respectively
CAIRO2	III	CBC	Median PFS, 5.9 and 6.6 months in patients with BRAF-mutant tumors vs 12.2 and 10.4 months in patients with BRAF wild-type tumors in two arms	BRAF mutation was a negative prognostic marker in patients with mCRC
			Median OS, 15.0 and 15.2 months in BRAF-mutant vs 24.6 and 21.5 months in BRAF wild-type tumors in two arms
NORDIC-VII	III	Nordic FLOX, cetuximab	ORRs, 20% in patients with BRAF-mutant tumors vs 50% in those with BRAF wild-type tumors (P < 0.01)	Presence of BRAF mutations was a strong negative prognostic factor
TAILOR	III	FOLFOX-4, cetuximab	Median PFS, 2.0 in patients with BRAF-mutant tumors vs 9.3 months in patients with BRAF wild-type tumors	Mutation status of the BRAF gene was a predictor of outcome in the cetuximab group
Pooled analysis of CAIRO, CAIRO2, COIN, and FOCUS phase 3 studies	III	CBC	Median PFS, 6.2 in BRAF-mutant vs 7.7 months in BRAF wild-type tumors	BRAF-mutant status was a biomarker conferring poor prognosis in mCRC
			Median OS, 11.4 in BRAF-mutant vs 17.2 months in BRAF wild-type tumors

BRAF, V-raf murine sarcoma viral oncogene homolog B1; KRAS, V-Ki-ras2 Kirsten ratsarcoma viral oncogene homolog; OFC, oxaliplatin, fluoropyrimidine, cetuximab; FOLFIRI, irinotecan, infusional fluorouracil, leucovorin; FOLFOX4, infusional fluorouracil, leucovorin, and oxaliplatin; CBC, capecitabine, oxaliplatin, bevacizumab, cetuximab.

Based on the clinical-trial data shown in Figure 1 [37], we can conclude that patients with mCRC bearing BRAF and KRAS mutations have poor prognosis with anti-EGFR therapies, which make KRAS and BRAF mutations prognostic and predictive biomarkers in anti-EGFR combined mCRC-treatment regimens [37]. In addition, some studies report that the BRAF mutation and the RAS mutation are mutually exclusive in mCRC [29, 53, 54]. None of the patients bearing the KRAS mutation carried the associated BRAF mutation in these trials. In other words, the BRAF mutation occurred exclusively in wt-KRAS mCRC.

Figure 1.

KRAS and BRAF mutations are prognostic and predictive biomarkers in anti-EGFR combined regimens for mCRC treatment. KRAS and BRAF mutations are associated with poor response to treatment with anti-EGFR. In the sample set, none of the patients harboring the KRAS mutation carried the associated BRAF mutation, indicating their exclusive relationship in mCRC. The data in the figure are from a clinical trial mentioned in the corresponding part of the text [37].

The mechanism of anti-EGFR-therapy resistance in BRAF- and KRAS-mutant mCRC

KRAS, NRAS, and BRAF are the most-studied mutant genes in mCRC and current data indicate that mutation status is associated with resistance to single-agent cetuximab or panitumumab in patients with mCRC [21, 55]. Randomized studies have shown that patients with mutations in KRAS exons 3 and 4 or NRAS exons 2, 3, and 4 showed decreased response rates and OS when therapy was combined with an anti-EGFR treatment [18–20]. The BRAF mutation is also a strong indicator of poor prognosis in mCRC; the addition of an anti-EGFR agent had a detrimental effect on survival in BRAF-mutant patients with mCRC [18, 56, 57]. Only 10% of patients with chemotherapy–refractory mCRC achieve a positive response to anti-EGFR agents as a single agent [58, 59]. It seems that molecular alterations in the nodes of the EGFR-signaling pathway contribute to primary resistance to anti-EGFR treatment. Given that gene-expression signatures that correspond to KRAS, BRAF, and phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA)-activating mutations have predictive value on the efficacy of anti-EGFR therapy, downstream components of KRAS, BRAF, and PIK3CA play important roles in anti-EGFR resistance [60–62].

Normally, EGF selectively binds to EGFR and triggers the receptor to form a dimer that activates receptor tyrosine kinase (RTK) activity and RAS (Figure 2). RAS is a membrane-bound GTPase that cycles between an inactive guanosine diphosphate (GDP)-bound form and an active guanosine triphosphate (GTP)-bound form. As a binary cell switch, RAS protein is activated by extracellular stimulation [63–66]. KRAS is located on the inner surface of the cell membrane and it transmits signals from activated transmembrane receptor EGFR to effectors in the MAPK and type I phosphatidylinositol 3-kinase (PI3K)/v-aktmurinethymomaviral oncogene (AKT)-signaling pathway in the cytoplasm [10, 67]. On one side, active GTP-bound RAS binds to the three closely related RAF proteins, CRAF1, BRAF, and ARAF, causing RAF to be relocated to the plasma membrane, triggering engagement of the pathway [68, 69]. Activated RAF changes the phosphorylation status and activates MEK1 and MEK2. MEK are capable of phosphorylating and activating the MAPK extracellular signal-regulated kinases 1 and 2 (ERK1 and ERK2) and regulate the activity of several transcription factors that induce the expression of multiple genes involved in cell survival and proliferation [70–72]. The other effector pathway is the PI3K/AKT pathway. RAS can interact directly with the catalytic subunit of PI3Ks. PI3K phosphorylates phosphatidylinositol 4,5-bisphosphate to produce phosphatidylinositol 3,4,5-trisphosphate that can bind to a large number of proteins as a second messenger through the pleckstrin homology domain and other domains such as AKT [73, 74].

Figure 2.

The downstream signaling pathway of the EGFR mainly. EGF selectively binds to EGFR and triggers the receptor to form a dimer that activates RAS. RAS transmits signals from activated transmembrane receptor EGFR to effectors in the MAPK- andPI3K/AKT-signaling pathway in the cytoplasm. Active GTP-bound RAS triggers RAF to engage in signal transmission. Activated RAF changes the phosphorylation status and activates MEK1 and MEK2. MEK are capable of phosphorylating and activating the MAPK extracellular signal ERK1 and ERK2, and regulate cell survival and proliferation. The other effector pathway is the PI3K/AKT pathway. RAS can interact directly with the catalytic subunit of PI3Ks and activate it; activated PI3Ks trigger AKT participate in signal transmission.

Constitutive activation of the MAPK-signaling pathway is a major cause of resistance to anti-EGFR therapies in KRAS- and BRAF-mutant mCRC. Mutations of the RAS genes themselves, most frequently KRAS mutations, contribute to aberrant signaling through the RAS pathways. This weakens the GTPase-activating proteins (GAPs) and increases the intrinsic catalytic rate of GTPase and prevents the hydrolysis of GTP to GDP, causing an accumulation of active GTP-bound RAS. Accumulation of GTP-bound RAS leads to constitutive KRAS activity, triggering downstream signal transduction and promoting survival and proliferation of the cancer cell. Almost all RAS activation in tumors is caused by mutations in codons 12, 13, and 61 [7, 75–79]. As a downstream signaling protein of RAS, RAF kinases normally play an important role in the EGFR-mediated MAPK pathway [13, 24]. Some (4%–15%) of mCRC tumors bear BRAF mutations and most of them are V600E amino-acid substitutions [11, 25]. As a central amino acid in the kinase domain, V600 is required to maintain RAF in its inactive conformation without RAS activation. Mutations in BRAF V600 can constitutively activate the MAPK pathway, regardless of whether RAS is activated or not, resulting in anti-EGFR resistance [54]. This receptor-independent pathway activation makes cancerous cells unresponsive to anti-EGFR molecules.

Furthermore, alterations in additional nodes of the EGFR pathway also appear to confer primary anti-EGFR resistance. Studies have demonstrated that mutations in PIK3CA exon 20 and phosphatase and tensin homolog (PTEN) alterations frequently coexist with RAS mutations. These alterations are associated with resistance to anti-EGFR mAbs [62, 80]. Amplification of KRAS and the MET receptor have been shown to bypass EGFR signaling and activate the EGFR pathway [81–83]. MET signaling seems to have a synergistic effect on the EGFR pathway in promoting the growth of tumor cells [84]. Moreover, the extensive crosstalk among the ERBB (also known as HER) family of receptors can up-regulate parallel pathways as a compensatory adaptive mechanism when EGFR is inhibited. EGFR can form heterodimers with the receptor tyrosine-protein kinase erbB-3 (HER3) and active the downstream PI3K- and MAPK-signaling pathways [85–87].

Subsequent studies found that some patients with wt-KRAS and wt-BRAF mCRC initially respond to anti-EGFR therapies; a few patients experience disease progression eventually. This is indicative of acquired resistance to anti-EGFR therapy. The most common mechanisms that drive secondary resistance to anti-EGFR agents are genetic alterations that contribute to primary resistance: KRAS, NRAS, BRAF, and PIK3CA mutations. These mutations were found in biopsies of patients who experienced relapse, although these mutations were not observed in patients at the beginning of treatment [88–91]. Biopsies confirmed that the number of mutant alleles increased under drug exposure and became undetectable after drug withdrawal [92]. Another mechanism of secondary-resistance acquisition is the prevention of drug binding to the receptor by acquired EGFR extracellular domain mutations (exon 12). S492R, a particular acquired EGFR mutation, is associated with secondary resistance to cetuximab in mCRC. The S492R mutation within the EGFR ectodomain perturbs the conformational state of the kinase drug-binding sites and the resulting bulky side-chain structure at this position could interfere with cetuximab binding [93–95]. ERBB2 gene upregulation has also been proposed as a factor of secondary resistance to anti-EGFR therapies in wt-RAS and wt-BRAF mCRC. Activated HER2 signaling in cancer cells can promote HER2 transcription or upregulate the HER2/HER3 ligand heregulin, thus conferring resistance to anti-EGFR therapies [96, 97].

Emerging therapies targeting BRAF-mutant mCRC

Combining BRAF and EGFR inhibitors

BRAF mutations commonly occur in melanoma and the BRAF inhibitor vemurafenib elicits a good response in 60%–80% of patients with BRAF-mutant melanoma in clinical trials [98, 99]. Trials testing the efficacy of this BRAF inhibitor in the treatment of BRAF-mutation-bearing patients with mCRC indicated that vemurafenib gives a poor prognosis for this type of patient [100, 101]. Further investigation revealed that the resistance is induced by reactivation of the CRAF-mediated MAPK signal. As shown in Figure 3, when BRAF–MEK–ERK signaling is interrupted by the BRAF inhibitor, the signaling output from mutant BRAF is blocked, causing transient suppression of ERK activation. ERK-dependent negative feedback activates EGFR, causing RAF protein dimer formation and reactivation of the CRAF-mediated MAPK signaling, which can bypass the BRAF node [27, 67, 98]. On the one hand, this mechanism explains why the single BRAF inhibitor has an antitumor effect in melanoma but not in mCRC. Melanoma cells express low levels of EGFR, so they are not influenced by this feedback activation and therefore demonstrate a good response to the single-agent BRAF inhibitor [98]. In contrast, these findings suggest that there is a strong dependency on MAPK signaling in mCRC and adequate inhibition of this pathway may improve response to treatment. A combination of BRAF inhibitors and anti-EGFR mAbs has been shown to produce sustained suppression of MAPK signaling and overcome EGFR-driven resistance in vitro and in xenograft models [102–105]. This indicates that doublet-targeted therapy may overcome the feedback effect. The various combination therapies are shown in Table 4. A combination of panitumumab (EGFR inhibitor) and vemurafenib (BRAF inhibitor) treatment in BRAF-mutant mCRC demonstrated antitumor efficacy: PFS was 3.2 months [95% confidence interval (CI), 1.6–5.3] and OS was 7.6 months (95% CI, 2.1–not reached). Compared with patients on single-agent vemurafenib or panitumumab, patients undergoing double therapy displayed decreased incidence and severity of acneiform rash (40% of grade 1 and 13% of grade 2), maculopapular rash (13%), palmar-plantar erythrodysesthesia syndrome (7%), papilloma (7%), and cutaneous squamous cell carcinoma/keratoacanthoma (0%), likely due to the opposing effects of vemurafenib (activation) and panitumumab (inhibition) on ERK signaling in epidermal keratinocytes [106]. A clinical trial evaluating dabrafenib plus panitumumab reported similar results: the 20 patients enrolled had a median PFS of 3.5 months (95% CI, 2.8–5.8 months) and OS of 13.2 months (95% CI, 6.7–22.0 months) [107]. This therapy was well tolerated and the majority of adverse events were grade 1 or 2. The most common adverse events of any grade were acneiform dermatitis (60%), nausea (50%), fatigue (50%), and diarrhea (45%) [107]. Other combinations of BRAF inhibitors and EGFR inhibitors are also in clinical trial; the PFS for encorafenib plus cetuximab was 3.7 months. The most common treatment-related adverse events were as follows: all grades [fatigue (46%), infusion-related reaction (35%)], and grade 3/4 [fatigue (8%), hypophosphatemia (15%)] [108]. Compared with the PFS of patients treated with singlet BRAF inhibitor (2.1 months, 95% CI, 0.4–11.6), the PFS of patients treated with combined BRAF and EGFR inhibitors significantly improved [109].

Figure 3.

The resistance mechanism of BRAF inhibitors. BRAF inhibition interrupts the RAF–MEK–ERK signaling, releases ERK-dependent negative feedback on the EGFR, and subsequently activates EGFR, formats the RAF protein dimer, then reactivates the CRAF-mediated MAPK signal.

Table 4.

Advanced combining therapies on BRAF-mutated mCRC

Therapeutic strategy	Agents investigated	PFS in the experiment group (months)	PFS in the control group (months)
Combining BRAF and EGFR inhibitors	Vemurafenib + panitumumab	3.2	∼2 months (standard therapies); 2.1 (singlet BRAF inhibitor)
	Dabrafenib + panitumumab	3.5
	Encorafenib + cetuximab	3.7
Combining BRAF and MEK inhibitors	Dabrafenib + trametinib	3.5
Combining EGFR and MEK inhibitors	Panitumumab + trametinib	2.6
Combining BRAF, EGFR, and MEK inhibitors	Dabrafenib + trametinib + panitumumab	4.2	2.6/3.5
	Encorafenib + cetuximab + binimetinib	4.3	1.5

Combining BRAF inhibitors and MEK inhibitors

Studies revealed that BRAF inhibitors can interrupt the signaling output from mutant BRAF and reactivate the CRAF-mediated MAPK signal, bypassing the BRAF node to transmit signaling to MEK. Therefore, MEK inhibitors may suppress the CRAF-mediated MAPK pathway. Mature strategies combining the BRAF inhibitor dabrafenib and MEK inhibitor trametinib have been successful in BRAF-mutant melanoma and BRAF-mutant non-small-cell lung cancer (NSCLC), and significantly improved outcomes compared with those patients treated with dabrafenib alone [2, 110–114]. The doublet combination of dabrafenib and trametinib for BRAF-mutant mCRC has also been trialed in clinic. Combined dabrafenib and trametinib contribute to an improved response rate (median PFS, 3.5 months; 95% CI, 3.4–4.0 months). In terms of safety, 17 adverse events (40%) led to a dose reduction, 25 adverse events (58%) led to a dose interruption, and four patients (9%) discontinued treatment due to an adverse event [115]. Another clinical trial of the same drug combination reported similar results. At data cut-off, 3-year PFS rate was 22% for the doublet-treatment group and 12% for the singlet-treatment group (HR, 0.71; 95% CI, 0.57–0.88), and the 3-year OS rate was 44% and 32%, respectively (HR, 0.75; 95% CI, 0.58–0.96). The incidence of adverse events was higher (>10% difference, any grade) in the doublet-treatment group than in the singlet-treatment group: pyrexia (59% vs 33%), chills (32% vs 17%), diarrhea (31% vs 17%), vomiting (26% vs 15%), and peripheral edema (22% vs 9%) [116]. Combining BRAF and MEK inhibitors has similar efficacy to combining BRAF and EGFR inhibitors, and results in improved PFS compared with singlet BRAF inhibitors.

Combining EGFR inhibitors and MEK inhibitors

Combined MEK and EGFR inhibition with trametinib and/or panitumumab for BRAF-mutant mCRC has also been trialed in clinic. No patient in the trametinib-plus-panitumumab-treatment group achieved complete or partial response and 55% were stable (T + P, n = 31), whereas, in the dabrafenib-plus-panitumumab-treatment group (D + P, n = 20), 10% had a confirmed complete or partial response and 80% were stable. The triplet-treatment group (D + P plus trametinib, n = 91) resulted in a confirmed complete response or partial response in 21% of patients, stable disease in 65%, and an overall disease-control rate of 86%. The median PFS in the trametinib-plus-panitumumab group was 2.6 months (95% CI, 1.4–2.8 months) compared with 3.5 months (95% CI, 2.8–5.8 months) in the D + P group and 4.2 months (95% CI, 4.0–5.6 months) in the D + P plus trametinib group. These results suggest that combination therapy of BRAF and EGFR inhibitors may not be a priority in BRAF-mutant mCRC [107, 117].

The triplet chemotherapy of BRAF, EGFR, and MEK inhibitors

A clinical trial compared the combination of dabrafenib and panitumumab with or without trametinib in 120 patients with BRAF-mutant mCRC. PFS was 4.2 months (95% CI, 4.0–5.6 months) in the triplet-treatment group and 3.5 months (95% CI, 2.8–5.8 months) in the treatment group without trametinib [107]. A large phase III trial involving only patients with BRAF–V600E-mutated mCRC was recently reported. A total of 665 patients were divided into three equal groups: triple-therapy group (encorafenib, binimetinib, and cetuximab), double-therapy group (encorafenib and cetuximab), and a control group (the investigators’ choice of either cetuximab and irinotecan or cetuximab and FOLFIRI). The therapeutic effect was significant; the median OS was 9.0 months (95% CI, 8.0–11.4) in the triple-therapy group, 8.4 months (95% CI, 7.5–11.0) in the double-therapy group, and 5.4 months (95% CI, 4.8–6.6) in the control group. Additionally, the risk of death was significantly lower (by 48%) in the triple-therapy group than in the control group (HR, 0.52; 95% CI, 0.39–0.70; P < 0.001). The PFS was significantly longer in both the triple-therapy group and the double-therapy group than in the control group: the median PFS was 4.3 months (95% CI, 4.1–5.2) in the triple-therapy group, 4.2 months (95% CI, 3.7–5.4) in the double-therapy group, and 1.5 months (95% CI, 1.5–1.7) in the control group [118]. Importantly, no increase in skin toxicity or fatigue was observed with the triple combination and no new safety signals were noted.

Standard therapies have limited benefit for BRAF-mutation-bearing patients with mCRC with a PFS of approximately 2 months and an OS of 4–6 months [118]. Double and triple therapies can both prolong PFS; triple therapy combining BRAF, EGFR, and MEK inhibitors remarkably delivers a much longer PFS than the double therapy (Table 4).

Combination of MEK inhibitors and ERK inhibitors

Studies have evaluated molecular alterations in the MAPK pathway after treatment with a BRAF/EGFR inhibitor or BRAF/MEK inhibitor in BRAF-mutant mCRC [29]. The mechanism of resistance involved KRAS mutation, KRAS upregulation, BRAF upregulation, and MEK1 mutation. The use of MEK inhibitors triggers KRAS mutation, KRAS upregulation, and BRAF upregulation, thus increasing the pathway flux and promoting MEK hyperactivation that can overcome the efficacy of MEK inhibitors [38, 81]. Despite upstream MAPK-pathway alterations, ERK inhibitors still have the ability to suppress the reactive MAPK pathway [119, 120]. Studies on the effects of the ERK-inhibitor combination strategy on a BRAF-mutant mCRC cell line suggest that the combination of ERK inhibitors and MEK inhibitors is more effective than single treatment [119, 120]. Although ERK inhibitors are at the early stages of clinical trial, they are expected to play an important role in the advanced therapeutic regimen for BRAF-mutant mCRC [121].

Emerging therapies targeting KRAS-mutant mCRC

KRAS mutations occur in 35%–45% of mCRC, causing therapy resistance and poor prognosis. KRAS is regarded as an ‘undruggable’ oncoprotein that cannot be targeted pharmacologically, so we discuss recent progress in new directions such as the parallel inhibition of the PI3K/AKT and MAPK pathways, directly targeting mutant KRAS, targeting KRAS-membrane association, exploring the KRAS-regulated metabolic pathway, as well as immunotherapy and KRAS synthetic lethal interactions [72, 122–124].

Combination of PI3K inhibitor and MEK inhibitor in preclinical trial

Although KRAS cannot be directly targeted, we can target downstream RAS and parallel signaling pathways. PI3K functions downstream to RTKs, which are regulated by RAS kinases. MEK and PI3K are parallel signaling pathways. Inhibition of the MEK pathway activates the PI3K pathway, and PI3K activation mediates resistance to MEK inhibition [125]. A study tested whether combining MEK and PI3L inhibitors can produce a stronger response in a KRAS-mutant cell line. The study assessed a combination of an MEK1/2 inhibitor AZD6244 and a PI3K inhibitor BEZ235 in vivo and measured the tumor volumes of tumor-bearing mice. Treatment with cetuximab for 3 weeks did not inhibit tumor growth; treatment with AZD6244 or BEZ235 alone caused a ∼50% reduction in tumor growth. Notably, the combined treatment of AZD6244 and BEZ235 completely inhibited tumor growth at the end of the 3-week treatment period. Importantly, there were no signs of weight loss or other acute or delayed toxicity symptoms in either the single-agent- or combination-treatment group [126]. This preclinical trial provides convincing evidence that combined PI3K inhibitors and MEK inhibitors enhance antitumor efficacy against KRAS-mutant cells, which strongly supports the development of clinical trials for this patient population.

Directly targeting KRAS-mutant mCRC

RAS is a membrane-bound GTPase that cycles between an inactive GDP-bound form and an active GTP-bound form. RAS protein is activated by extracellular stimulation and is converted to active GTP-bound RAS. KRAS G12C is one of the most common KRAS mutants in cancer, present in 10%–20% of all KRAS G12 mutations. Biochemical analysis showed that KRAS G12 preferentially binds to the inactive GDP-bound form of RAS, impairing SOS-catalysed nucleotide exchange and decreasing the affinity of RAS for GTP, resulting in decreased survival and increased apoptosis of tumor cells containing the G12C oncogene [127–129].

AMG 510, the first KRAS G12C inhibitor in clinical development, has undergone preclinical trial [130]. AMG 510 treatment reduced tumor size and, in some cases, even eradicated the tumor in a KRAS-mutant murine model. Moreover, AMG 510 has a synergistic effect with chemotherapy, immunotherapy, and other targeted inhibitors such as the MEK inhibitor. In the NCI-H358 tumor-cell line carrying the KRAS G12C mutation, AMG 510 had synergistic effects with a MEK inhibitor. In vivo, the combination of low-dose AMG 510 and MEK inhibitor significantly improved the antitumor effect compared with single-drug therapy in tumor-bearing mice. Tumor growth was inhibited in singlet-treatment groups compared with the control group, and the tumor volume tended to decrease in the combination-treatment group. These data suggest that the combination of AMG 510 with inhibitors targeting the MAPK-signaling pathway may eliminate bypass or residual signaling, which may in turn enhance efficacy or prevent drug resistance. When AMG 510 is combined with carboplatin, a standard chemotherapy for lung cancer in vivo, the combined treatment significantly reduced tumor volume compared with single-drug therapy. Additionally, tumor-bearing mice were treated with suboptimal doses of AMG 510 and/or PD-1 inhibitors, which can upregulate T-cell activity and enhance the T-cell recognition of tumor cells. The results showed that AMG 510 single-drug treatment caused the complete disappearance of the tumor in only 1 of 10 mice and PD-1-inhibitor monotherapy had a similar effect, whereas the combination therapy resulted in the complete disappearance of tumors in 9 of 10 mice. This indicates that AMG 510 has potential as a combination therapeutic agent in KRAS-mutant mCRC.

Targeting G-quadruplex (G4) structures of the KRAS gene

G4 structures are formed in guanine-rich nucleic-acid sequences and play a key role in tumor development through the regulation of a wide range of oncogenes, tumor-suppressor genes, and somatic copy-number alterations [131, 132]. There are three G4 motifs in the human KRAS promoter: G4-proximal, G4-middle, and G4-distal. In particular, the G4-middle structure may play an important part in the development of targeted therapeutics [133, 134]. EMICORON, a novel synthetic G4 ligand, has shown promising antitumor efficacy against CRC tumors. In HCT-116 CRC cells, KRAS mRNA and protein expression were downregulated after EMICORON treatment, suggesting that EMICORON can downregulate both KRAS mRNA and protein expression by targeting the G4 structure. In vivo, EMICORON was tested on a panel of patient-derived xenografts (PDXs) bearing KRAS mutations. In all three PDXs, the tumor volume of mice treated with EMICORON was decreased compared with that of untreated mice. In addition, no adverse effects were observed in the treated mice. EMICORON can also improve the efficacy of chemotherapy: CRC PDXs treated with EMICORON plus FOLFIRI showed better response rates and survival than the FOLFIRI-alone-treatment group [135, 136]. These results suggest that EMICORON has therapeutic potential for KRAS-mutant mCRC.

Potential therapeutic strategies in KRAS-mutant mCRC

Targeting KRAS-membrane association

KRAS located on the inner leaflet of the plasma membrane transmits signals from activated EGFR transmembrane receptors to downstream cytoplasmic effectors. KRAS realizes its oncogenic activity through this membrane association. Therefore, inhibiting KRAS-membrane association seems a logical therapy for KRAS-mutant cancers [137]. Phosphodiesterase-6δ (PDEδ) is a prenyl-binding protein that enriches RAS at the plasma membrane, thus augmenting RAS signaling [138]. A small-molecule inhibitor, deltarasin, prevents PDEδ from binding to KRAS and impairs KRAS localization to the endomembrane. Deltazinone 1, an improved PDEδ inhibitor, showed high selectivity and less nonspecific cytotoxicity than deltarasin and, at the same time, exhibited high correlation with the phenotypic effect of PDEδ knockdown in a set of human pancreatic-cancer-cell lines [139, 140]. However, further chemical optimization is required, as its chemical properties are unstable in vivo.

Targeting the KRAS-regulated metabolic pathway

The infinite proliferation of tumor cells requires metabolic reprogramming to produce biomass. KRAS-mutant tumors also develop the metabolic pathways to obtain substrates from both extracellular and intracellular sources to meet their metabolic need [141, 142]. KRAS-mutant CRC tumors show high expression of glycolytic and glutamic metabolic proteins [143]. CB-839—an inhibitor targeting the glutaminase (GLS) that contributes to the conversion of glutamine to glutamate—is currently being evaluated in phase I and II clinical trials for the treatment of various cancers [144]. In addition, KRAS-mutant cells show glyceraldehyde 3-phosphate dehydrogenase (GAPDH) metabolic vulnerability compared with wt-KRAS cells and vitamin C selectively kills KRAS-mutant CRC cells by inhibiting GAPDH [145].

Immunotherapy and synthetic lethal interactions in KRAS-mutant cancer

Immunotherapy is a key focus in cancer treatment currently, especially immune-checkpoint-blocking antibodies. Immune-checkpoint-blocking antibodies are profoundly changing the therapeutic strategies of various cancers. Specifically, the anti-PD-1/PD-L1 antibody has shown significant antitumor activity by restoring the T-cell antitumor response. PD-L1 is present in some normal cells as well as cancer cells. When PD-1 is bound to PD-L1, T-cells do not attack the cell. Some cancer cells contain large amounts of PD-L1, allowing them to escape immune attack. mAbs targeting PD-1 or PD-L1 can block this binding and prevent cancer cells from escaping T-cell attack, thereby restoring the normal immune response to cancer cells. Anti-PD-1/PD-L1 antibody treatment is now approved in various tumor types [146]. KRAS-mutant NSCLC was effectively treated with the PD-1 antibody pembrolizumab combined with the MEK inhibitor trametinib [147]. Clinical trials demonstrated that the PD-1 antibody treatment is more effective in patients with KRAS-mutant NSCLC than in patients with wt-KRAS NSCLC [147, 148]. This is likely because KRAS mutations impede DNA repair, especially mismatch repair (MMR), which supports the notion that MMR deficiency is a favorable factor for PD-1 blockade [149]. This result suggests that immunotherapy is a promising potential treatment strategy for KRAS-mutant mCRC.

Synthetic lethal interactions in KRAS-mutant cancer have also been widely studied. Synthetic lethality defines the interaction between two co-essential genes such that inhibiting both genes rather than either single gene can result in cell death [150]. There are several stage I and II clinical trials assessing the synthetic lethality of MEK inhibitors combined with BCL2, AKT, or SHP2 in KRAS-mutant cancer. These treatments produce antitumor efficacy by interrupting the pathways required for the survival of KRAS-mutant cells, including co-operating signaling, transcriptional regulation, and the maintenance of genomic stability [151–153]. We found that most of the synthetic lethality studies in KRAS-mutant cancer are focused on specific cancer cell lines or tumor models; however, as KRAS-mutant cancers in clinical trials are highly heterogeneous, further studies should aim at systematically examining a spectrum of KRAS-mutant cancers.

Combination of curcumin and regorafenib

Curcumin is a polyphenol, which is the main yellow pigment in turmeric. Pharmacologically, it has an anticancer effect in human clinical trials by promoting apoptosis and inducing autophagy [154]. Additionally, in an ongoing phase III clinical trial, curcumin has been demonstrated to be a safe and effective treatment for patients with CRC [155]. Regorafenib is a multityrosine kinase inhibitor that mainly inhibits antigenic and oncogenic kinases, such as VEGFR, PDGFR, FGFR, and BRAF. Two randomized phase III clinical studies (CORRECT and CONCUR) evaluated the efficacy of regorafenib in mCRC treatment and the results indicate that regorafenib is a promising treatment option for patients with mCRC who had poor prognosis [156–158].

Another study found that curcumin had a similar antitumor effect to MEK inhibitor U0126 and that this effect was enhanced when curcumin was used synergistically with regorafenib in the KRAS-mutant CRC cell line HCT 116. Inhibiting MEK conferred the synthetic lethal interaction with RAF1 inhibition in a new concept for combination-drug treatments. It may be possible to overcome the limited therapeutic effect of current treatments for KRAS-mutant mCRC by developing doublet-combination therapies targeting the RAF–MEK pathway. Given that curcumin has the potential to inhibit MEK and regorafenib is a RAF-targeting agent, we propose that the combination of curcumin and regorafenib is a promising therapeutic strategy for KRAS-mutant mCRC [159, 160].

Conclusions

Personalized therapy enables patients to receive maximum treatment efficacy with minimum side effects, ensuring the best possible quality of life and maximum survival of patients. The current approved first-line combination of chemotherapy and anti-EGFR agent treatment that has been highly effective in patients with mCRC clinically has little benefit and poor prognosis when applied to BRAF- and KRAS-mutation-bearing patients with mCRC. Analysis of the BRAF- or KRAS-mutant occurrence in no-response-to-anti-EGFR-agents patients and possible resistance mechanisms suggest that BRAF- and KRAS-mutant statuses are critical prognostic and predictive biomarkers in mCRC treatment. Utilizing this biomarker status to select patients and formulate personalized therapeutic strategies according to oncogene status promises to improve patient outcomes. The addition of an inhibitor of the specific BRAF mutation provides additional treatment options that include EGFR inhibitors and increases the likelihood of treatment-delayed disease progression. Combining inhibitors that target specific node kinases such as EGFR inhibitors, BRAF inhibitors, MEK inhibitors, and ERK inhibitors provides promising clinical treatment to delay disease progression and prolong survival times. While no mature inhibitor-targeted therapies for KRAS-mutant mCRC have been currently adopted into clinical treatment, data from preclinical and clinical trials indicate potential combination-treatment options for KRAS-mutant mCRC patients such as curcumin plus regorafenib or PI3K inhibitor plus MEK inhibitor.

Authors’ contributions

Z.N.L. and L.Z. wrote the manuscript, L.F.Y. contributed to the manuscript, and M.J.W. reviewed the final versions. All authors read and confirmed the final version of the manuscript.

Funding

This work was supported by National Natural Science Foundation of China [No. U1608281]; National Natural Science Foundation of China [81903658]; Liaoning Province Scientific Research Foundation [JC2019032]; Liaoning Revitalization Talents Program [No. XLYC1807201]; and Shenyang S&T Projects [19–109-4–09].

Conflicts of interest

None declared.

Author notes

Zi-Nan Li and Lin Zhao contributed equally to this work.