Real-world pharmacokinetics and pharmacodynamics of everolimus in metastatic breast cancer

Purpose This study investigated the relationship between the pharmacokinetics and pharmacodynamics of everolimus in patients with metastatic breast cancer (mBC) in real-world practice.Methods Twenty-two patients with mBC treated with everolimus plus exemestane were enrolled. Blood everolimus concentrations were measured at outpatient visits. The inhibition of the mammalian target of rapamycin (mTOR) activity in peripheral blood mononuclear cells (PBMCs) was examined. The efficacy and safety endpoints were progression-free survival (PFS) and the cumulative incidence of dose-limiting toxicities (DLTs), respectively. Results Blood samples were obtained from 19 consenting patients. Everolimus did not completely inhibit mTOR activity in PBMCs at therapeutic concentrations (~ 56 % maximal inhibition). The most common adverse event was stomatitis (any grade 77 %). The trough concentration (Ctrough) was significantly higher in patients experiencing DLTs than in those without any DLTs (P = 0.030). The optimal Ctrough cutoff predicting DLT development was 17.3 ng/mL. The cumulative incidence of DLTs was significantly higher in patients with Ctrough ≥17.3 ng/mL than in other patients (sub-hazard ratio 4.87, 95 % confidence interval [CI] 1.53–15.5; P = 0.007). Furthermore, the median PFS was numerically longer in patients who maintained a steady-state Ctrough below the threshold than in those who did not (327 days [95 % CI 103–355 days] vs. 194 days [95 % CI 45 days–not estimable]; P = 0.35). Conclusions The suggested upper threshold for the therapeutic window of everolimus Ctrough was 17.3 ng/mL. Pharmacokinetically guided dosing may improve the efficacy and safety of everolimus for mBC, warranting further investigation in a larger study. Clinical trial registry: Not applicable.


Introduction
Everolimus, an orally active inhibitor of the mammalian target of rapamycin (mTOR), has been used as an immunosuppressive drug in solid organ transplantation [1]. Based on its antiangiogenic and antitumor activity, everolimus has been approved as single-agent therapy for advanced renal cell carcinoma (RCC), neuroendocrine tumors (NET) of pancreatic, gastrointestinal, or lung origin, and tuberous sclerosis complex (TSC)-associated subependymal giant cell astrocytoma (SEGA) as well as TSC-associated renal angiomyolipoma [2][3][4][5]. Everolimus is also used in combination with exemestane to treat hormone receptor-positive, human epidermal growth factor receptor 2 (HER2)-negative metastatic breast cancer (mBC) after progression on nonsteroidal aromatase inhibitors [6].
Everolimus has a narrow therapeutic index and shows significant inter-individual pharmacokinetic variability when used as an immunosuppressant; therefore, routine therapeutic drug monitoring (TDM) is recommended to maintain a target blood everolimus trough concentration (C trough ) between 3 and 8 ng/mL [7,8]. On the other hand, for the treatment of TSC-associated SEGA only, the dose is adjusted based on TDM to attain a C trough of 5-15 ng/mL [9]. The approved dosage of everolimus for other oncological indications is 10 mg orally once daily. This fixed dosing schedule is associated with clinically relevant toxicities, including severe stomatitis and non-infectious pneumonitis, which often necessitates dose modifications.
Regarding pharmacodynamics, everolimus binds to the intracellular FK506-binding protein 1A (FKBP1A) with two-fold less affinity than tacrolimus (FK506) [19]. This complex inhibits the kinase activity of mTOR, which inactivates the downstream p70S6 kinase (S6K) and increases the inhibitory binding of eukaryotic translation initiation factor 4E (eIF4E)-binding protein 1 (4E-BP1) to eIF4E; together these actions inhibit cell growth and metabolism as well as cell proliferation and survival [20]. It has been shown that the degree of inhibition of p70S6 kinase is identical in peripheral blood mononuclear cells (PBMCs) and simultaneously collected tumor tissue in cancer patients receiving CCI-779, another mTOR inhibitor, suggesting that PBMCs are a valid surrogate tissue for the pharmacodynamic monitoring of mTOR inhibitors in vivo [21].
The objectives of this study were as follows: (1) to characterize the pharmacokinetics and pharmacodynamics of everolimus in patients with mBC; (2) to investigate the potential impact of pharmacogenetics on everolimus exposure; and (3) to clarify the relationship between everolimus exposure and clinical outcomes in real-world practice.

Study design and patients
We conducted a single-center, prospective cohort study in the context of routine clinical practice rather than a clinical trial, with the aim of optimizing everolimus dosing through pharmacokinetic and pharmacodynamic analysis in patients with mBC. The protocol was approved by the institutional ethics committee of Asahikawa Medical University (#14085). Written informed consent was obtained from each patient prior to participation in the study. Consecutive patients with mBC who started everolimus plus exemestane between June 2014 and December 2019 were prospectively enrolled.
The baseline characteristics of the patients are summarized in Table 1.

Treatment
Everolimus was administered orally at 10 mg once daily in combination with exemestane (25 mg orally once daily). When severe adverse events were present, the everolimus dose was temporarily reduced or interrupted, followed by resumption at a reduced dose. Everolimus treatment was continued until the occurrence of disease progression, unacceptable toxicity, or patient refusal.

Blood samples
In the outpatient setting, we obtained blood samples immediately before each patient's morning dose to measure individual everolimus C trough levels (i.e., C 0 ). Otherwise, blood samples were routinely collected from remnant blood specimens after laboratory testing at each visit. These nontrough blood samples, taken within approximately 4 h after the morning dose, were used to measure individual everolimus peak concentrations (C peak ). Blood samples for the concentrations C 1 , C 2 , C 3 , and C 4 were defined as those obtained at 1 h ± 30 min, 2 h ± 30 min, 3 h ± 30 min, and 4 h ± 30 min post-dose, ECOG Eastern Cooperative Oncology Group respectively. Additionally, blood samples taken in the postabsorption phase at 12 ± 2 h and 24 ± 2 h post-dose were used to determine C 12 and C 24 , respectively.

Pharmacokinetic assessment
Whole blood samples (150 µL) were deproteinized with 450 µL of methanol/0.2 M ZnSO 4 (70/30, v/v) containing the internal standard ascomycin. After vortex and centrifugation, the supernatant was treated with solid-phase extraction using Oasis HLB cartridges (Waters, Tokyo, Japan). The eluent was evaporated to dryness and reconstituted in 200 µL of 50 % acetonitrile. After filtration, 50 µL was automatically injected into a liquid chromatography-tandem mass spectrometry (LC-MS/MS) system. Chromatographic separation was performed with a C18 column heated at 65°C and a flow rate of 0.2 mL/min (lower limit of quantification 1 ng/mL, run time 3.5 min). The isocratic mobile phase consisted of a mixture of 5 % 10 mM ammonium acetate and 0.1 % acetic acid in water and 95 % 10 mM ammonium acetate and 0.1 % acetic acid in acetonitrile. Analyses were performed in the multiple reaction monitoring mode at ion transitions m/z 976.4 → 909.3 for everolimus and m/z 810.4 → 756.7 for ascomycin. Interand intra-assay accuracies were ±10 % with a precision (coefficient of variation) below 5 %.

mTOR activity in PBMCs
PBMCs were isolated from whole blood (approximately 2 mL) by Ficoll-Hypaque density gradient centrifugation. Cell counting and viability assessment were performed using the TC20 counter (Bio-Rad, Tokyo, Japan). PBMCs with viability exceeding 90 % were used for the subsequent mTOR assay. After centrifugation at 400 x g for 2 min at 4°C, the supernatant was discarded and the cell pellets were resuspended in 100 µL of ice-cold hypotonic lysis buffer containing protease and phosphatase inhibitor cocktails (Nacalai, Kyoto, Japan), and then stored at − 80°C until analysis. The mTOR activity in the extracts of PBMCs was directly measured using the Kinase-linked immunosorbent assay (K-LISA) mTOR activity kit (CBA055, Calbiochem, USA) by assessing phosphorylation of the specific mTOR substrate p70S6 kinase at Thr 389 , according to the manufacturer's protocol and a previous study [23]. A recombinant human mTOR enzyme (1362-end) with a specific activity of 186 U/mg (Lot #2052551-D, Millipore, UK) was used for calibration (standard series, 0-250 ng). The protein concentration of each lysate was determined using the Coomassie Brilliant Blue R-250 Staining Solution (Bio-Rad, Tokyo, Japan) for normalization of mTOR activity. Finally, the mTOR activity in PBMCs was calculated relative to the baseline activity before the start of therapy in each patient.

Outcomes
The efficacy endpoints were progression-free survival (PFS) and overall survival (OS). For safety assessment, all adverse events were graded according to the Common Terminology Criteria for Adverse Events version 4.03. The cumulative incidence of dose-limiting toxicities (DLTs) leading to treatment discontinuation and dose interruption/reduction was estimated by adjusting for competing risks (e.g., death or treatment discontinuation due to disease progression) using the Fine and Gray model [24]. The data cutoff date was March 31, 2020.

Statistical analyses
The statistical significance of differences in non-parametric values between two groups was analyzed with the Mann-Whitney U test. A receiver operating characteristic (ROC) curve was constructed, and the area under the ROC curve (AUC ROC ) was calculated to estimate the optimal cutoff value of everolimus C trough for predicting development of DLTs. The median PFS and OS were estimated using the Kaplan-Meier method, and the difference between two groups was examined using the Gehan-Breslow-Wilcoxon test. A twosided P < 0.05 was considered statistically significant. All statistical analyses were performed using STATA software, version 16 (StataCorp LLC, Texas, USA).

Patients and everolimus dose
Twenty-two patients with mBC who were treated with everolimus plus exemestane were enrolled ( Table 1). The sample included one male patient. One female patient with advanced age (80 y) started everolimus treatment at 5 mg/day to minimize its toxic effects. Dose modifications of everolimus, specifically interruption and reduction due to DLTs, were performed in seven (32 %) patients (n = 6 and 1, respectively). Everolimus treatment was discontinued because of disease progression (n = 10), adverse events (n = 8), surgery (n = 1), and patient refusal (n = 1). At the data cutoff date, everolimus therapy was ongoing in two patients. The median follow-up period was 494 days (range 102-1067 days).

Everolimus pharmacokinetics/pharmacogenetics and pharmacodynamics
Pharmacokinetic and pharmacogenetic data were available from 19 consenting patients, two of whom did not provide blood samples for determining everolimus C peak levels. Figure 1a shows the distribution of blood everolimus concentrations at each time point during a 24-h dosing interval at steady state in patients receiving 10 mg/day of everolimus (total no. of observations 92). Overall, the maximum concentration of approximately 60 ng/mL appeared to be reached at 3 h post-dose, which was comparable with previous findings [25]. Large interindividual variability in C trough was observed (8.3-36.8 ng/mL). Furthermore, everolimus showed considerable fluctuations in absorption, particularly within 2 h postdose (C 1 7.8-107 ng/mL, C 2 6.7-85.3 ng/mL). Among four patients receiving 5 mg/day of everolimus due to toxicity, the median C peak between 1 and 2 h post-dose at steady state was 10 ng/mL (range 3.0-28.5 ng/mL). This was lower than the median values predicted based on C 1 and C 2 at a dosage of 10 mg/day (20.7 and 27.1 ng/mL, respectively), suggesting nonlinear absorption. We calculated the average C trough and C peak with all the respective measurements available at steady state in each patient on 10 mg/day. No CYP3A4*22 alleles were detected in the cohort, which was consistent with a previous observation in the Japanese population [26]. Additionally, the CYP3A5 6986G > A and ABCG2 421 C > A polymorphisms did not influence C trough or C peak (Supplementary Fig. S1).
Pharmacodynamic data on mTOR activity in PBMCs were obtained in 11 patients with evaluable samples (total no. of observations 60). A significant correlation was observed between blood everolimus concentration and mTOR activity in PBMCs relative to baseline (R 2 = 0.65, P < 0.0001; Fig. 1b). Notably, everolimus did not completely inhibit mTOR activity in PBMCs at therapeutic concentrations at any point during the course of everolimus treatment (~56 % maximal inhibition).

Everolimus toxicity and its relationship to drug exposure
The most frequently observed adverse event of any grade was stomatitis (77 %, Table 2). No grade 4 or 5 toxicities were reported in this study. The most common grade 3 adverse events were stomatitis and anorexia (18 % each), followed by fatigue (14 %). Grade 3 hyperglycemia was observed in one patient, whereas there were no cases of hyperlipidemia. DLTs leading to treatment discontinuation included pneumonitis (n = 4), stomatitis, fatigue, urticaria, and muscle pain (one patient each). One patient reported grade 3 dysgeusia requiring dose reduction. Dose interruption was required in patients with grade 3 non-hematological toxicities, including stomatitis and anorexia (n = 2 each) as well as herpes zoster and diarrhea (one patient each).
The average C trough at steady state did not differ in patients with and without any grade of pneumonitis (median 9.  Notably, the average C trough was significantly higher in patients with DLTs leading to treatment discontinuation and dose interruption/reduction than in those without any DLTs (P = 0.030; Fig. 2a). The optimal C trough cutoff to predict development of DLTs was 17.3 ng/mL (AUC ROC 0.79, 95 % confidence interval [CI] 0.58-1.00; P = 0.031). Furthermore, as shown in Fig. 2b, the cumulative incidence of DLTs was significantly higher in patients with an average C trough ≥17.3 ng/mL than in other patients (sub-hazard ratio 4.87, 95 % CI 1.53-15.5; P = 0.007).

Discussion
There is accumulating evidence that shows a positive relationship between everolimus exposure and clinical outcomes in oncology settings [27]. The current study provides real-world evidence to support an association between everolimus C trough and toxicity in patients with mBC. Deppenweiler et al. [28] previously reported that everolimus C trough levels > 26.3 ng/ mL were associated with a four-fold increased risk of adverse events in patients with a variety of cancers, with a particularly high risk in those with breast cancer. In the present study, we determined that the optimal cutoff value of everolimus C trough  for predicting the development of DLTs was 17.3 ng/mL (Fig. 2), a finding comparable to that of a previous study that identified 19.2 ng/mL as the toxicity threshold in patients with mBC [29]. We also found that the average C trough at steady state did not differ between patients with and without any grade of pneumonitis. This observation is consistent with those in earlier studies demonstrating no correlation between everolimus C trough and the occurrence of pneumonitis [30,31]. Therefore, monitoring of blood everolimus concentrations may not help identify patients at increased risk of pneumonitis. Further development of reliable biomarkers in addition to serum KL-6 levels is necessary to improve the prediction of mTOR inhibitor-induced pneumonitis [31,32].
Regarding the exposure-efficacy relationship, no significant difference in median PFS was reported between patients with everolimus C trough >12.6 ng/mL versus < 12.6 ng/mL who were also treated with exemestane [29]. Similarly, everolimus C trough level was not a significant predictor of long-term efficacy in our patient cohort (Fig. 3b). These findings suggest that higher everolimus exposure will not translate into survival benefit, probably because of decreased tolerability due to toxicities such as severe stomatitis and anorexia (Table 2). Furthermore, the lack of survival benefit despite increased everolimus exposure may be explained by the fact that patients also received exemestane, which might have independently improved efficacy regardless of individual exposure to everolimus.
Overall, as shown in Fig. 1a, everolimus demonstrated variable absorption and marked variation in elimination. Although the number of patients was small, we found no indication that CYP3A5 6986G > A or ABCG2 421 C > A polymorphisms influenced everolimus exposure ( Supplementary Fig. S1), which is consistent with findings in previous studies [14][15][16][17][18]. Therefore, there may be no need to individualize the dose based on genetic polymorphisms in CYP3A5 and ABCG2. Further investigations that consider non-genetic factors (e.g., food ingredients) [33] should be conducted to clarify the mechanisms underlying the variable, potentially nonlinear pharmacokinetic characteristics of everolimus in mBC.
In the present study, we directly measured mTOR activity in PBMCs as a marker of everolimus pharmacodynamics. Interestingly, the mTOR activity in PBMCs was not completely inhibited by everolimus at therapeutic concentrations (Fig. 1b). In a study of renal transplant recipients, Dekter et al. [23] showed similar results regarding the partial inhibition of mTOR activity by everolimus in PBMCs in vivo and in vitro. These observations may be explained by intracellular immunophilin FKBP1A potentially limiting mTOR inhibition by everolimus at high drug concentrations; this occurs with tacrolimus, which binds with the same immunophilin and incompletely inhibits calcineurin phosphatase activity at saturable concentrations [34].
Despite the fact that everolimus almost completely inhibits phosphorylation of the downstream S6 protein, partial inhibition of 4E-BP1 phosphorylation was previously observed in tumor samples and PBMCs from patients treated with everolimus [35,36]. Taking these findings and our results into account, partial rather than complete inhibition of the mTOR pathway by everolimus may be potentially beneficial for preventing life-threatening toxicities and/or recovering early from adverse drug reactions through dose reduction/ interruption in patients receiving everolimus. Actually, no grade 4 or higher adverse events occurred in our patient cohort. On the other hand, given the fact that everolimus only partially inhibits mTOR, everolimus monotherapy might not exert adequate antitumor activity. Thus, in patients with advanced RCC who have already undergone one prior anti- Progression-free survival rate 12 10 7 5 0 0 0 angiogenic therapy, it seems reasonable to administer lenvatinib, a multikinase inhibitor, in combination with everolimus in order to increase response rates and survival [37]. Male breast cancer is a rare disease, accounting for less than 1 % of all breast cancers [38]. In the present study, we enrolled one male patient with hormone receptor-positive, HER2-negative mBC who was treated with everolimus plus exemestane. Ten days after the start of therapy, the patient visited the emergency room because of severe fatigue, leading to immediate discontinuation of everolimus treatment. On that day, the blood everolimus concentration more than 24 h after the last dose was 18.4 ng/mL, which was above the toxicity threshold identified in this study (17.3 ng/mL). The patient was subsequently treated with exemestane alone followed by high-dose toremifene, which was finally discontinued due to disease progression 5.5 months after he began receiving everolimus plus exemestane. Among patients with breast cancer, men had a numerically higher incidence of adverse events than women during everolimus treatment in real-world practice [39]. Potential gender differences in the exposure-toxicity relationship for everolimus should be investigated in a large population-based study that includes male patients with mBC.
Limitations of this study include its small sample size, lack of robust pharmacokinetic-pharmacodynamic profiling, and single-center design. Despite these limitations, the present study demonstrated that increased exposure to everolimus was associated with an increased risk of DLTs among patients with mBC. Through TDM, pharmacokinetically guided dosing of everolimus in routine clinical practice may help avoid the development of DLTs and improve efficacy as well as adherence to treatment in patients who receive everolimus for mBC.

Conclusions
The present study indicates that everolimus may exert antitumor activity without completely inhibiting mTOR activity in patients with mBC. Furthermore, our results suggest that the potential upper threshold for the therapeutic window of everolimus C trough is 17.3 ng/mL, warranting further prospective investigation in a larger patient population.