Everolimus in treatment of neuroendocrine tumors: efficacy, side-effects, resistance and factors affecting its place in the treatment sequence.

1. Background

Neuroendocrine tumors (NETs) are rare malignancies derived from the neuroendocrine system throughout the body [1]. The annual incidence of NETs is increasing and was 6.98 per 100,000 persons in 2012 reported from Surveillance, Epidemiology, and End Results (SEER) program[2], with most NETs commonly arising from lung and gastroenteropancreatic (GEP) sites[2,3]. NETs comprise a heterogeneous group of neoplasms with a wide spectrum of clinical and malignant behaviors depending on various clinicopathological factors. NETs are usually indolent by nature and slow growing, however, poor prognosis of those in advanced stage have not improved dramatically until recently[2–4], and thus effective therapeutic options are highly warranted.

The phosphatidylinositol-3-kinase (PI3K)/Akt and mammalian target of rapamycin (mTOR) signaling pathway is well known to play a crucial role in controlling cancer cell-cycle and growth[5], and its activation is involved in the pathogenesis of NETs[6]. Upregulation of mTOR signaling components[7–11], downregulation of its upstream negative regulators; phosphatase and tensin homolog deleted on chromosome 10 [PTEN] and tuberous sclerosis complex 2 [TSC2] [12–14], and genomic mutation of mTOR pathway[15,16] in NETs were reported from numerous groups. The anti-proliferative effect of mTOR pathway inhibition was reported in preclinical models of NETs [12,17], and thus it was a promising therapeutic target[6,18]. The efficacy and safety of everolimus, an oral inhibitor of the mTOR pathway[19], in patients with advanced NETs of different origins was examined in a series of phase II/III clinical studies; RADIANT trials (Table 1) [20–23]. Of these, RADIANT-3 trial was a randomized, double-blind, placebo-controlled, phase III trial that demonstrated significant clinical benefit in the largest number of patients with advanced pancreatic NETs (panNET) [22]. Based on this result, the Food and Drug Administration (FDA) approved everolimus as an anti-tumoral drug for the treatment of panNETs with progressive disease on May 5, 2011, and it is now commercially available in >110 countries.

The purpose of this article is to evaluate at present (7 years post initial approval), the clinical utility of everolimus in NET patients. This is not possible without first considering the advances in the diagnosis and treatment of NETs that are currently influencing anti-tumor treatment approaches to advanced NETs, which have occurred since everolimus’ initial approval in 2011 (Table 2). An understanding of these is essential to exploring the current options that are available and the place of everolimus within these for the current treatment of NETs.

Shortly after the approval in May, 2011 of everolimus for treatment of patients with advanced panNETs, sunitinib, an oral multi-targeted tyrosine kinases inhibitor, was shown to have efficacy in patients with advanced panNETs, prolonging progression free survival (PFS) in a phase III trial[24], and it was approved as a standard therapeutic agent in those patients worldwide (Table 2), thus offering an additional treatment choice. Three years later, lanreotide autogel was shown to improve PFS in a phase III trial (CLARINET) in patients with advanced grade 1 or 2 NETs, and it was approved by FDA as an anti-growth therapy in GEP-NET[25](Table 2). Therefore, by 2014, there were three new anti-tumor medications (everolimus, sunitinib, and somatostatin analogues [SSAs]) available to treat patients with advanced panNETs, in addition to older therapies such as chemotherapy. In 2016, the RADIANT-4 trial showed the benefit of everolimus in lung and all gastrointestinal (GI) NETs to extend PFS in patients with advanced disease[23], leading to its additional approval for extrapancreatic NETs. In 2018, peptide receptor radionuclide therapy (PRRT) using ¹⁷⁷ Lutetium (Lu) -DOTATATE was globally approved for the treatment of GEP-NETs[26,27](Table 2). This approval was based on the results of the phase III (NETTER-1) trial in which ¹⁷⁷ Lu-DOTATATE markedly extended PFS in patients with advanced somatostatin-receptor (SSTR) -positive midgut NETs with acceptable safety[26], and on data from 504 patients treated with advanced panNETs and other NETs in Rotterdam[27]. In addition, over the last 7 years, therapeutic options have been developed for the treatment of patients with liver predominant metastatic disease from malignant NETs. These treatments include the increasing use of selective internal radiation therapy (SIRT) with ⁹⁰Yttrium (Y) labeled microspheres to reduce the hepatic tumor burden[28,29](Table 2).

There have also been additional drugs approval to treat the clinical symptoms of functional NETs (F-NETs), such as the carcinoid syndrome or F-panNETs syndromes. In the management of carcinoid syndrome, one new drug became available recently (Table 2). Telotristat ethyl, a peripheral-acting tryptophan hydroxylase inhibitor, which blocks serotonin synthesis, has been shown to cause a significant reduction in bowel movement frequency in a phase III trial (TELESTAR) in patients with carcinoid syndrome[30]. In combination with SSA, it was approved by FDA for the treatment of SSA-refractory carcinoid syndrome in February, 2017 (Table 2). In addition, SSAs such as octreotide and lanreotide have been approved for the treatment of F-NETs, and in a recent phase III trial (ELECT), lanreotide autogel reduced bowel movement frequency in patients with carcinoid syndrome, regardless of primary location (Table 2)[31].

In addition to new treatments for patients with NETs, there was a number of changes in the management of these patients over the past 7 years since the initial approval of everolimus which affect current practices (Table 2). Histopathological classification and TNM staging systems proposed by World Health Organization (WHO), European Neuroendocrine Tumor Society (ENETS) and American Joint Committee on Cancer (AJCC) are now routinely used in clinical practice to predict prognosis and to select optimal treatments. The WHO classification system separates NETs into several groups based on the proliferative index and tumor differentiation assessed by Ki-67 labeling index and mitotic counts, including NET G1, NET G2 and G3 for GEP-NETs[32], and typical carcinoids (TC), atypical carcinoids (AC) and neuroendocrine carcinoma (NEC) for lung NETs[33], respectively (Table 3). The prognostic value of these systems was validated by several studies[34–36], however, heterogeneity remained in the high-grade G3 that showed different prognosis, genetic profiles, and treatment response to platinum-based chemotherapy in different groups of G3 patients[37–45]. To address this issue, the recent 2017 WHO classification for panNETs has divided the G3 group for panNETs into well-differentiated G3 NETs (WDNEC) as a distinct subgroup from poorly-differentiated NEC (PDNEC, Table 2,3)[46]. The use of these classification/grading systems is now mandatory in all recent guidelines, and their widespread use is having an effect on the management of these patients[34–36]. It affects anti-tumor drug use, because cisplatin-based regimens are recommended for G3 NEC and the role of everolimus in NET G3 patients, as well as other therapeutic options, in these patients is still being defined.

Diagnostic modalities, especially molecular imaging, have now been validated[1], and are now routinely used with ⁶⁸Gallium (Ga) -DOTATATE positron emission tomography (PET/CT) imaging receiving FDA approval on June 2016, supported by several studies[47–49] that demonstrated its high sensitivity/specificity for the detection of location and extent of SSTR positive NETs (Table 2). In addition, ¹⁸F-fluorodeoxyglucose (FDG) PET is being increasingly used in combination with ⁶⁸Ga-DOTATATE to define patients with aggressive tumors, usually of high grade (Table 2)[50,51]. These modalities can influence treatment options, because they are better defining the extent of the disease as well as pattern of growth, which can be important in selecting different anti-tumor treatments.

Despite these many recent advances (Table 2), mentioned above, the treatment of advanced NETs still has many unanswered questions. These will be discussed in a separate, later section. Prior to discussion of the current controversies as well as the clinical use of everolimus at present in patients with advanced NETs, the findings from studies of the efficacy/safety of everolimus in these patients will be briefly reviewed.

2. An overview of the current market

As described in depth in the previous background section, since the approval by the FDA of everolimus for its anti-growth effect in patients with advanced panNETs in May 5, 2011, and the subsequent approval for GI and lung NETs in May 26, 2016, there have been numerous changes in the diagnosis and treatment of these diseases which affects the potential use of everolimus at present. These are summarized in Table 2 and includes both the introduction of other anti-tumor agents (sunitinib, lanreotide, PRRT, and liver-directed therapy) and changes in the management of NETs that affect its treatment. These include the demonstration of efficacy of everolimus in controlling the hormone-excess state, as well as changes in tumor imaging and tumor classification which provide prognostic information which can affect the anti-tumor treatment.

3. Introduction to the Compound

As reviewed in the background, numerous in vitro and in vivo studies in animals with NETs, as well as preliminary human studies (Table 4) provide evidence that the PI3K/Akt/mTOR pathway plays an important role in proliferation of NETs. Everolimus (40-O-[hydroxyethyl]-rapamycin), also known as RAD001, is a rapamycin analog that directly inhibits the mTORC1 pathway by binding to intracellular receptor FK506-binding protein (FKBP-12), which subsequently inhibits activation of its downstream mediators such as p70S6K/4E-BP1 (Box 1)[5,52]. Its efficacy in patients with NETs was established in numerous studies (Table 1,3). Before the approval of everolimus for the treatment of panNETs in 2011, it was already approved by FDA for its anti-growth effect in patients with renal cell carcinoma in 2009[53] and TSC-associated subependymal glial cell astrocytoma in 2010[54], as well as for immunosuppression to prevent organ rejection in kidney transplants recipients in 2010[55], and had proven safe/effective in these patients. Subsequent to its approval for patients with advanced NETs, everolimus also received FDA approval for the anti-proliferative treatment of hormone receptor-positive breast cancer (in combination with exemestane)[56] and TSC-associated renal angiomyolipoma[54] in 2012.

4. Pharmacokinetics and Pharmacodynamics

Oral everolimus is rapidly absorbed after oral administration, and reaches a peak serum concentration after 1.3–1.8 hours[57]. Steady-state serum concentrations are reached within a week, and both maximum serum concentration and the area under the curve, increased in a dose-proportional manner[58]. Everolimus is metabolized mainly by cytochrome P450 (CYP) 3A1, 3A5 and 2C8 in the gut and liver, and is a substrate for P-glycoprotein (P-gp). Approximately 98% and 2% of everolimus is excreted in the bile in the form of metabolites and urine, respectively (for detail, see[57]).

5. Efficacy of everolimus

6. Markers predicting efficacy of everolimus

The recent increase in therapeutic options for patients with advanced NETs show variable responses in different patients, and thus have raised the importance of biomarkers in predicting response to treatment that could enable optimal treatment selection and change in therapy[140]. In the RADIANT-1 trial, early response in serum chromogranin A (CgA) or neuron-specific enolase (NSE) levels was a significant predictor for prolonged PFS with everolimus treatment[72]. In addition, elevated baseline CgA and NSE levels were associated with shorter PFS and OS with everolimus treatment[141]. Similarly, sub-analysis of RADIANT-3 trial also showed that high baseline CgA, NSE, placental growth factor (PIGF) or soluble vascular endothelial growth factor receptor 1 (sVEGFR1) levels were associated with poor OS[65], however, baseline CgA levels were not predictive for everolimus treatment effect on OS in the RADIANT-2 trial[59]. Another study genotyped the fibroblast growth factor receptor 4 (FGFR4) using DNA isolated from tissue or blood in panNETs, and found that patients homozygous for FGFR-G388 demonstrated greater tumor shrinkage with everolimus treatment than those harboring only one FGFR4-R388 allele[142]. Although longer PFS and OS were also observed in those homozygous for FGFR4-G388, these were not statistically significant due to small number of patients, which is consistent with the result from another group showing that FGFR4 polymorphism did not have predictive value on everolimus treatment[143].

The efficacy of other markers, specifically the components of mTOR pathway, has been evaluated by several groups. In 17 patients with NETs treated with everolimus and octreotide, high tumor p-Akt levels in pre-treatment (R = 0.476, P = 0.053) and on-treatment (R = 0.604, P = 0.015) tumor biopsy specimens correlated with longer PFS[144]. In an immunohistochemical analysis of bronchial NETs in 21 patients, protein levels of the total and phosphorylated form of mTOR and its component (p70S6K [ribosomal protein S6 kinase, 70 kDa], Akt and ERK 1/2) were upregulated in everolimus sensitive patients compared to everolimus resistant patients[145]. The efficacy of monitoring intra-tumor p-p70S6K [80,146–148] was suggested by other groups, however, the relationship to PFS and tumor response with everolimus were conflicting. Other clinicopathologic factors including histological grade [22,93–95], Ki-67 index[91,94] and hepatic metastasis burden[93] were also reported to affect PFS of everolimus treatment. In addition, the results of imaging modalities, such as ⁸⁹Zirconium (Zr)-bevacizumab PET[149] and perfusion CT[96], in predicting response to everolimus was suggested by several groups. Furthermore, patients experienced stomatitis within 8 weeks of everolimus initiation had longer PFS than those without stomatitis in the RADIANT-3 trial (13.9 vs 8.3 months, hazard ratio [HR] 0.70) and in RADIANT-2 trial (HR 0.87) [150]. Similar to this, occurrence of hypercholesterolemia was also suggested as a possible marker of response[146].

7. Safety of everolimus (Table 6)

Treatment with everolimus is well known to cause a spectrum of adverse events (AEs) in various patients. According to meta-analyses including patients with NETs, treatment with everolimus as well as other mTOR inhibitors, is most commonly associated with an increased risk of stomatitis [150–153], rash[151,154], fatigue[155], infection[156], pulmonary toxicities [152,157,158], hyperglycemia[155,159], anemia[160,161], and thrombocytopenia[160]. The pharmacokinetic analysis of RADIANT-2 trial also revealed that an increased everolimus minimum concentration was associated with higher risk for pulmonary and metabolic events[64]. In addition, the development of severe AEs, such as infection and pneumonitis[162], could result in treatment discontinuation[163] and fatal outcome[162,164]. The frequencies of common AEs associated with everolimus treatment observed in the previous studies were listed in Table 6. The majority of AEs in the different studies were mild to moderate, and treatment-related mortality did not occur in most of the studies (Table 6). Pulmonary toxicity (n=5) and infection/sepsis (n=3) were the most common fatal AEs that caused everolimus-related death[22,23,75,76,93]. Though patient-reported tolerance with everolimus treatment was reported to be poorer than SSA[165], health-related QOL was well maintained in large clinical trials[71,79,166]. These results strongly support the general tolerability of everolimus by most patients in both the clinical trial setting and real-world setting.

However, everolimus-associated AEs can be significant and have led 5–35% of patients to treatment discontinuation (Table 7). In addition, dose reduction and/or interruption are frequently (13–100%) required to control the severity of the AEs (Table 7). Like other targeted agents, the importance of appropriate dose modification for the management of everolimus-associated AEs has been emphasized, especially for the long-term continuation of everolimus[94]. This is in line with the pharmacokinetic analysis including patients with NETs that showed significant correlation between elevated everolimus trough concentration (C_min) and increased risk of toxicity[152,167]. Regarding efficacy, higher C_min were associated with improved tumor size reduction[152], whereas lower C_min increased risk of progression[167]. It was reported that patients with high cumulative dose (>3000 mg) and dose intensity (> 9 mg/day) of everolimus experienced significantly longer OS[92]. Although dose intensities were reported to be well maintained in most previous studies[21,72,73,75,76,85], therapeutic drug monitoring (TDM) may be useful in prevention of developing severe AEs without lowering the efficacy of everolimus[167], in addition to the patients’ education and physicians careful monitoring[168].

8. Mechanism of resistance to everolimus

Despite the high anti-tumor activity of everolimus, disease progression can occur early after everolimus initiation in some patients, and even in those who once experienced tumor response with everolimus treatment. In the previous studies, up to 57% of the patients terminated everolimus treatment due to disease progression (Table 7). Primary resistance, as well as acquired (secondary) resistance to everolimus is thought to play a role, thus understanding and overcoming the mechanisms underlying everolimus resistance may prove important in preventing patients from treatment discontinuation due to disease progression.

mTOR is the core components of 2 complexes, mTORC1 and mTORC2, which have distinct signaling pathways and functions[19,169]. The activation of mTORC1 subsequently phosphorylates its downstream p70S6K and 4E-BP1[5,52]. Everolimus specifically directly inhibits the mTORC1 and its downstream p70S6K/4E-BP1[170,171], resulting in attenuation of protein synthesis as well as cell growth[12,52,147,148]. The PI3K/Akt/mTORC1 pathway is normally regulated by in part from its upstream insulin-like growth factor 1 (IGF-1) receptors with IGF-1-dependent activation of the PI3K/Akt pathway leading to inhibition of signaling due to mTOR-mediated phosphorylation and degradation of insulin receptor substrate-1 (IRS-1) [80,81,172,173].

Suppression of these feedback loops by inhibiting mTORC1 causes over-activation of upstream signaling, including PI3K/Akt[174,175]. Several studies showed that everolimus may only partially inhibit mTORC1 and incompletely block its certain downstream target, especially 4E-BP1[176,177]. The inability of everolimus to block mTORC2 induces upstream Akt phosphorylation[19,170,178,179]. In addition, a genetic mutation in FKBP12 is reported to affect the sensitivity of mTOR inhibitor[180]. Other mechanisms of resistance to mTOR inhibitors have also demonstrated including; activation of mitogen activated protein kinase(MAPK); up-regulation of pro-angiogenic factors; and activation of Ras pathway[92,174,181,182]. All these mechanisms potentially counteract the anti-proliferative effects of mTORC1 inhibitors and lead to drug resistance[80,172,173].

Consistent with this hypothesis, histopathological analyses and preclinical studies reveal the increased activation of Akt in everolimus-sensitive cell lines and patients[80,144,183]. On the other hand, several studies have showed contrary results that high p-Akt levels were significantly associated with poor survival with everolimus treatment[9,12]. In addition, overexpression p70S6K was reported to be associated with shorter PFS[10,146]. Discrepancies between these studies may partly result from the small number of patients, and due to the heterogeneity in patient’s background, especially their genetic profile[8,16,145,184].

Emerging evidence shows that dual inhibition of these pathways, particularly PI3K/Akt and the mTOR pathway, could be a novel therapeutic approach to overcome everolimus resistance [6,178,183–187]. The efficacy and safety of the dual PI3K/mTOR inhibitor, BEZ235, was examined in patients with advanced NETs[86,188,189]. In a phase II trial, 62 patients with mTOR inhibitor-naïve advanced panNETs were randomized to receive BEZ235 (n=31) or everolimus (n=31)[86]. Unfortunately, BEZ235 did not demonstrated superior efficacy compared with everolimus (Table 4). Despite the acceptable tolerability observed in a phase I trial in patients with advanced solid tumors[188], high toxicity of BEZ235 observed in this study led patients to require frequent dose modifications and treatment discontinuations. As a result, the shorter treatment duration of BEZ235 compared to everolimus (22.9 vs 39.4 weeks) might have negative impact on its anti-tumor activity. Poor tolerability of BEZ235 was also reported from a single-arm phase II trial in advanced patients with everolimus-refractory panNETs[189], which has led to its development being stopped. Further studies using other agents with better tolerability are warranted to clarify the effect of pan-PI3K inhibitors against resistance to everolimus.

9. Controversies in everolimus treatment in advanced NETs

10. Regulatory affairs

For the patients with NETs, everolimus (AFINITOR^®) has been approved by the FDA for the treatment of panNETs on May 5, 2011, and for GI and lung NETs on May 26, 2016, respectively. The initial dose of AFINITOR^® is 10 mg, to be taken once daily at the same time every day, either consistently with food or consistently without food. Dose reduction is recommended for patients with hepatic impairment depending on its severity and those who required co-administration of moderate inhibitors of CYP3A4 and/or P-gp. If patients require concomitant use of strong CYP3A4 and P-gp inducers, increase dose using increments of 5 mg or less increments to a maximum of 20 mg/day. Dose reduction is also used in controlling side effects.

11. Conclusion

Everolimus is a key therapeutic agent in the treatment of advanced NETs, which covers the spectrum of this heterogeneous groups of neoplasms, regardless of its primary site, histological grade, functional status, and radiological aggressiveness. Further studies are required to define its efficacy and safety in combination with other agents, mechanisms of resistance, and to better define its place in the current treatment options with the recent release of other anti-tumor agents such as PRRT.

12. Expert opinion

The RADIANT trials have corroborated the basic science results, isolated cell studies and studies in animals showing the anti-growth mechanisms of everolimus on NETs, and established its efficacy in patients with advanced NETs[20–23]. However, there still remain a number of controversies related to everolimus treatment in the management of NETs.

Despite the clinically meaningful prolongation of PFS with everolimus treatment demonstrated in the RADIANT trials, its effect on overall survival has not yet been established[22,23]. Treatment of advanced NETs requires a multidisciplinary approach, thus appropriate timing for everolimus initiation is important to evaluate long-term outcome. The exact order of the use of everolimus in the treatment cascade of patients with advanced NETs is at present unclear and being affected by a number of factors, such as the recent widespread use of WHO classification with its prognostic value and importance in treatment selection; everolimus’s effectiveness in combination with other anti-tumor treatments; the recognition of long-term potential resistance to everolimus and other targeted therapies; side effects of long-term everolimus treatment; and the increasing development of competing therapies. The use of biomarkers that predict the sensitivity and response to each treatment could be useful in selecting optimal therapeutic options, however, their utility in real clinical practice is so far, not clear.

In recent guidelines, everolimus is generally recommended as second-line therapy behind SSAs[201–203,207–209], because of the low toxicity with high tolerability of SSAs treatment[165]. At present, everolimus is one of the most frequently used second line treatments in patients presenting clinically significant tumor burden and disease progression on SSAs. Other competing therapeutic agents including sunitinib for panNETs, PRRT, chemotherapy, and liver directed therapy for panNETs and other NETs, may also be indicated in these settings[201–203,207–209]. Of these, the recent approval of PRRT is most likely to change everolimus’s order in the anti-tumor therapeutic approach. PRRT results in a longer PFS than everolimus, is well-tolerated by patients, a significant percentage of patients show a cytotoxic effect with reduced tumor burden[26,27], and thus it will likely have a significant impact on the use of cytostatic agents including everolimus and sunitinib, especially in patients with aggressive tumors. Severe AE/toxicity of PRRT is reported to be low (1–2%) (leukemia, myeloproliferative effects) [26,27,213]. With PRRT’s general availability and use in all groups of NETs, AE/toxicity will be better defined, and this could play a role in its general use and the position of everolimus in the treatment sequence. It should be noted that most guidelines have not reflected the recent availability of PRRT, shown in the NETTER-1 trial and other recent studies[202,203,209,210].

The underlying mechanisms of primary and/or acquired resistance to everolimus occurring in as many as 17–90% of patients stopping treatment due to loss of efficacy is not entirely clear (Table 7) [80,172,173]. It has a strong impact on the long-term continuation of monotherapy with everolimus, as well as combination and subsequent treatment with other therapeutic agents. Emerging evidences from the preclinical studies have shown a number of likely mechanisms of resistance to everolimus, and several potent agents with or without combination with everolimus have been examined in clinical trials for the purpose of overcoming the resistance to everolimus as well as adding a synergistic anticarcinogenic effect. Some of these show encouraging results, however high toxicity may limit their anti-tumor effect[86,189]. The efficacy of combination therapy with everolimus still needs to be explored more widely in further studies.

It is reported that percentage of patients, which is generally low, stopped everolimus treatment in various regimens owing to AEs (Table 7). Everolimus is generally well tolerated, and most of these AEs/side effects are manageable with dose reduction and interruption. In contrast, decreased dose intensity of everolimus may impair its anti-tumor activity. Balancing the anti-tumor activity and treatment-associated toxicity, as well as maintaining QOL is important in term of long-term continuation. Although it is not required in the real clinical practice, therapeutic drug monitoring may provide important information in this regard[167], as recommended for the use of immunosuppressants after organ transplantation[214]. Development of severe AEs, especially pulmonary toxicities and infections (Table 7)[22,23,75,76,93], can result in fatal outcomes, thus screening of occult infections and pulmonary function is mandatory before everolimus initiation. Adequate control of comorbidities, such as diabetes and hyperlipidemia, could also prevent patients from unnecessary treatment discontinuation. From this viewpoint, one possible strategy for treatment decision-making could be based on patient tolerance, comorbidities, and toxicity profiles. Specifically, everolimus is not indicated in patients with impaired pulmonary function, uncontrolled infections and metabolic disorders[152,155–159].

Although a number of unanswered questions remained in this area, in the near future with more widespread use of PRRT and other treatment options, it is likely new guidelines will become available addressing some of these issues.