Cancer Chemotherapy and Pharmacology (2019) 84:539–550 https://doi.org/10.1007/s00280-019-03871-w ORIGINAL ARTICLE Pharmacokinetic and exposure–response analysis of pertuzumab in patients with HER2‑positive metastatic gastric or gastroesophageal junction cancer Whitney P. Kirschbrown1 · Bei Wang1 · Ihsan Nijem1 · Atsushi Ohtsu2 · Paulo M. Hoff3 · Manish A. Shah4 · Lin Shen5 · Yoon‑Koo Kang6 · Maria Alsina7 · Sandhya Girish1 · Amit Garg1 Received: 18 January 2019 / Accepted: 22 April 2019 / Published online: 10 June 2019 © The Author(s) 2019 Abstract Purpose To characterize the pharmacokinetics (PK) of pertuzumab and trastuzumab in patients with HER2-positive meta- static gastric or gastroesophageal junction cancer in the randomized, double-blind, phase III JACOB study (NCT01774786), and to evaluate the appropriateness of the pertuzumab regimen in these patients. Methods Patients received 840 mg intravenous pertuzumab or placebo plus trastuzumab q3w and chemotherapy. Pertuzumab and trastuzumab were administered until disease progression or unacceptable toxicity. Chemotherapy was administered for up to six cycles or disease progression or unacceptable toxicity. Serum concentrations of pertuzumab and trastuzumab were measured. Pertuzumab PK was characterized across treatment cycles. The impact of anti-drug antibodies (ADAs) on pertuzumab PK and the impact of pertuzumab on trastuzumab PK were assessed. An exploratory exposure–efficacy analysis was also conducted. Results In total, 374 patients in the pertuzumab arm had evaluable PK data. The mean observed pertuzumab steady-state serum trough (minimum) concentration (Cmin,ss) ± standard deviation was 114 ± 51.8 μg/mL. The target pertuzumab Cmin,ss of ≥ 20 μg/mL was reached in 99.3% of patients at Cycle 5 (steady state) and beyond. Greater than 90% of patients were above the PK target right after the first pertuzumab dose. There was no apparent impact of ADAs on pertuzumab PK nor of pertuzumab on trastuzumab PK. There were no differences in overall survival across Cycle 1 pertuzumab (Cmin) or Cycle 5 pertuzumab (Cmin,ss) exposure quartiles. Conclusions Pertuzumab exposure in JACOB was consistent with prior studies in advanced gastric cancer and breast cancer. The 840 mg q3w dose allowed the majority of patients in JACOB to achieve target pertuzumab concentrations and appears to be an appropriate dose selection. Keywords Pertuzumab · Trastuzumab · Pharmacokinetics · HER2-positive · Metastatic gastroesophageal junction · Metastatic gastric cancer Introduction dimerization with other HER family receptors, thereby inhibiting mitogen-activated protein kinase and phospho- Pertuzumab ­( PERJETA ®, F. Hoffmann-La Roche Ltd, inositide 3-kinase signaling pathways and promoting cell- Basel, Switzerland) is a monoclonal antibody (mAb) that growth arrest and apoptosis [1, 2]. inhibits human epidermal growth factor receptor 2 (HER2) Adding pertuzumab to trastuzumab (­Herceptin ®, F. Hoffmann-La Roche Ltd) may provide a more compre- hensive HER2 pathway blockade (vs. trastuzumab alone) Electronic supplementary material The online version of this due to the complementary modes of action of the two article (https​://doi.org/10.1007/s0028​0-019-03871​-w) contains drugs [3]. The addition of pertuzumab to trastuzumab has supplementary material, which is available to authorized users. been shown to improve survival outcomes in patients with * Amit Garg HER2-positive early breast cancer (EBC) and metastatic gargamit1@gmail.com breast cancer (MBC) [4–6]. Although there are differences Extended author information available on the last page of the article in tumor biology between HER2-positive breast cancer 13 Vol.:(0123456789) 540 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 and HER2-positive gastric cancer (GC), it was hypoth- Methods esized that the dual HER2-targeted regimen could also improve survival outcomes in patients with HER2-positive Patients and study design advanced GC (AGC) [7]. The pharmacokinetics (PK) and safety of pertuzumab JACOB was a randomized, multicenter, multinational, plus trastuzumab and chemotherapy in patients with double-blind, placebo-controlled, phase III trial evaluating HER2-positive AGC was initially evaluated in the phase the efficacy and safety of pertuzumab vs. placebo in com- IIa dose-finding JOSHUA study (NCT01461057) [7]. This bination with trastuzumab and chemotherapy. The JACOB study showed that the combination of pertuzumab, tras- study design has been described previously [11]. In brief, tuzumab, and chemotherapy was well tolerated in these patients aged ≥ 18 years with HER2-positive MGC/GEJC, patients, with an adverse event (AE) profile similar to that measurable or evaluable non-measurable disease at base- seen in the ToGA trial (trastuzumab and chemotherapy vs. line, Eastern Cooperative Oncology Group performance chemotherapy alone; NCT01041404) [8]. The PK analysis status of 0 or 1, and baseline left ventricular ejection frac- of JOSHUA found that mean serum trough levels at day tion ≥ 55% were eligible; patients who had received the 43 were 37% lower in patients with HER2-positive AGC, previous therapy with an HER2-targeted drug or previ- compared with patients with HER2-positive MBC when ous systemic chemotherapy for metastatic disease were the same dose was administered in both indications [7, excluded. 9]. PK analyses also showed that the new dosing regi- Enrolled patients were randomized 1:1 to receive per- men of 840 mg every 3 weeks (q3w) resulted in higher tuzumab or placebo (840 mg pertuzumab or placebo given pertuzumab concentrations and increased the probability IV q3w) plus trastuzumab (IV; 8 mg/kg loading dose fol- of patients achieving a target serum trough concentration lowed by 6 mg/kg maintenance dose, q3w) and chemo- (Cmin) of pertuzumab similar to those observed in HER2- therapy (oral capecitabine: 1000 mg/m2, taken twice daily positive EBC and MBC with the previously established on days 1–15, q3w, or 5-fluorouracil: 800 mg/m2/24 h IV 840 mg loading dose followed by 420 mg q3w approved by continuous infusion for 120 h on days 1–5, q3w; plus in those indications, without compromising the overall cisplatin: 80 mg/m2 IV on day 1 only, q3w). Randomiza- safety profile of the regimen [7]. The pertuzumab serum tion was stratified by geographic region (Japan vs. North concentration of > 20 μg/mL was established as the target America/Western Europe/Australia vs. Asia [excluding efficacious exposure (Cmin) based on non-clinical efficacy Japan] vs. South America/Eastern Europe), prior gastrec- models [10]. tomy, and HER2 immunohistochemistry. Pertuzumab (or JACOB (NCT01774786) was a prospective, randomized, placebo) and trastuzumab were given until disease pro- multicenter, multinational, double-blind, placebo-controlled, gression or unacceptable toxicity. Chemotherapy was phase III trial that evaluated the efficacy and safety of the given for six treatment cycles and only discontinued dur- new pertuzumab dosing regimen of 840 mg q3w adminis- ing or before Cycle 6 for progressive disease or unaccep- tered intravenously (IV) plus trastuzumab and chemotherapy table toxicity. (pertuzumab arm), compared with placebo plus trastuzumab The JACOB study was conducted in accordance with and chemotherapy (placebo arm) as first-line therapy in the International Conference on Harmonisation E6 guide- patients with previously untreated HER2-positive metastatic line for Good Clinical Practice (ICH–GCP E6) and the GC or gastroesophageal junction cancer (MGC/GEJC) [11]. principles of the Declaration of Helsinki, or the laws The JACOB study did not meet its primary endpoint of a and regulations of the country in which the research was statistically significant improvement in overall survival [OS; done; whichever provided the greater protection for the hazard ratio 0.84, 95% confidence interval (CI) 0.71–1.00, individual. stratified log-rank p = 0.0565]. Here, we present data from the PK analysis of pertu- zumab and trastuzumab in patients with MGC/GEJC in the PK, anti‑drug antibodies, and HER2 extracellular JACOB study. A secondary objective of the JACOB study domain sampling was to assess the pharmacokinetics of pertuzumab. Explora- tory objectives included an assessment of PK drug–drug PK samples were taken from all patients in the JACOB interactions (DDIs), pharmacokinetics of trastuzumab, the study for the PK analysis of pertuzumab and trastuzumab. impact of anti-drug antibodies (ADAs) on PK, efficacy, and Serum samples were collected pre-dose at Cycles 1, 2, 3, 4, safety, effect of shed HER2 extracellular domain (ECD) on 6, and 8 and post-infusion at Cycles 1, 2, 4, and 8. Cycle 8 PK, and evaluation of the exposure–efficacy relationship, serum samples were collected when patients were on bio- conducted to evaluate the appropriateness of the 840 mg logic therapy (i.e., pertuzumab and/or trastuzumab) alone q3w pertuzumab regimen. 13 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 541 and the last chemotherapy administration was ≥ 6 weeks acceptable accuracy (% difference) and % CV with ranges ago. Two additional serum samples were collected at of − 8.08 to − 1.47% and 3.07 to 8.44%, respectively. post-treatment monitoring visits at 28 (± 7) days and A validated ELISA was used to detect and confirm the 60‒90 days after last study treatment administration (post- presence of ADAs to pertuzumab in human serum. This treatment monitoring visits 1 and 2, respectively) to sup- assay used two conjugated reagents to capture ADAs port interpretation of anti-drug antibody (ADA) with mini- directed against pertuzumab: biotin-conjugated pertuzumab mized potential for drug interference in the ADA assay. and digoxin-conjugated pertuzumab. Bound ADAs to per- These samples were analyzed for pertuzumab (pertuzumab tuzumab were then detected with a mouse anti-digoxin arm only) and trastuzumab (pertuzumab and placebo arms) antibody conjugated with horseradish peroxidase. Using an concentrations by P ­ PD® Laboratories, LLC (Richmond, anti-idiotypic mAb directed against pertuzumab as a posi- VA, USA) using validated assays. tive control, the relative sensitivity was determined to be Based on historical data in EBC and MBC, pertuzumab 3.59 ng/mL in the absence of pertuzumab. In the presence steady state should be reached at day 43, following one load- of 200 μg/mL pertuzumab, the assay can detect 500 ng/mL ing and one maintenance dose [1, 10]. Pertuzumab PK char- of the anti-idiotypic mAb control. acterization in this study was performed using observed data Serum concentrations of HER2 ECD were measured only, not model predicted, which could account for time- using the commercially available ADVIA ­Centaur® Serum dependent clearance. Therefore, the latest cycle of observed HER-2/neu Assay (Siemens Healthcare Diagnostics Inc., data, where most patients remained on study (Cycle 5), was Deerfield, IL, USA), a fully automated two-site sandwich used to designate steady state. immunoassay using direct chemiluminescent technology. Prior pertuzumab population PK (popPK) analyses did The immunoassay has a minimum detectable concentra- not indicate that geographic region or race were covariates tion of 0.5 ng/mL; the assay showed acceptable accuracy on clearance or volume; however, geographic region was (% recovery) and % CV with ranges of 88.7–100.9 and one of the stratification factors of the JACOB study, as dif- 3.2–5.7%, respectively. ferent regions have different screening and early detection practices, which could influence efficacy. Therefore, ensur- Data handling ing no pertuzumab exposure differences in this large global trial was warranted. All patients treated who had at least one documented per- Serum samples to test for the presence of ADAs against tuzumab or trastuzumab administration and at least one pertuzumab were collected from all patients pre-dose at corresponding measurable concentration of pertuzumab or Cycles 1, 3, and 6 and at post-treatment monitoring visits 1 trastuzumab were included in the PK analysis, unless there and 2. The samples were analyzed for ADAs against pertu- were major protocol deviations or information that may have zumab (pertuzumab arm only) by P ­ PD® Laboratories, LLC interfered with PK evaluation (i.e., labeling error, technical using a validated immunoassay. failure in sample analysis). Records were excluded if the Separate blood samples were obtained for the assessment time of drug administration or sample collection was miss- of serum concentration of shed HER2 extracellular domain ing. No imputation of PK values was performed. Observa- (ECD) at pre-dose at Cycles 1, 3, and 6, and during post- tions with missing PK or time values, or those below the treatment monitoring visit 1. The samples were analyzed for lower limit of quantification, were omitted from the analysis. shed HER2 ECD by Covance Laboratories (Indianapolis, IN, USA) using a validated immunoassay. PK assessments Bioanalytical methods Peak (maximum) serum concentrations Cmax (post-dose) and Cmin (pre-dose) from prespecified collection timepoints Serum concentrations of pertuzumab were determined by were summarized using descriptive statistics and graphical a validated enzyme-linked immunosorbent assay (ELISA) assessment. Cmin refers to the concentration at the end of with a minimum quantifiable concentration of 150 ng/mL a dosing interval, and therefore, the pre-dose PK sample [12]. The ELISA assay showed acceptable accuracy (% of a given cycle refers to the previous cycle’s Cmin (i.e., difference) and inter-assay percent coefficient of variation pre-dose Cycle 6 is the Cycle 5 Cmin). To reduce burden (% CV) with ranges of − 8.75 to 3.84% and 3.89–15.3%, for patients and healthcare providers, sparse PK sampling respectively. was used, where only pre-dose and a few post-dose samples Serum concentrations of trastuzumab were determined by were collected. Steady-state concentrations of pertuzumab a validated high-performance liquid chromatography assay (Cmax,ss and Cmin,ss) from JACOB were compared across dif- with tandem mass spectrometry detection (minimum quanti- ferent geographic regions (a prespecified stratification fac- fiable concentration of 100 ng/mL) [13]. This assay showed tor in the study): Japan, North America/Western Europe/ 13 542 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 Australia, Asia (excluding Japan), and South America/ within Cycle 1 and Cycle 5 Cmin quartiles. A log-rank test Eastern Europe. Steady-state concentrations of pertuzumab determined whether statistically significant survival differ- from JACOB were compared with equivalent data obtained ences existed among Cmin quartiles. in the previous studies of women with MBC (CLEOPATRA, NCT00567190) [9] and patients with AGC (JOSHUA) [7]. Assessment of ADAs on PK, safety, and efficacy Similarly, steady-state concentrations of trastuzumab from JACOB were compared with equivalent data obtained in the Incidence of ADAs to pertuzumab was measured and the previous AGC studies (JOSHUA and ToGA) [7, 14]. impact of ADAs on pertuzumab PK, safety, and efficacy was assessed. Analysis of DDIs Assessment of HER2 ECD concentrations Potential effects of pertuzumab on the steady-state PK of trastuzumab were assessed by comparing the arithmetic Serum concentrations of HER2 ECD were measured in both means of serum trastuzumab concentrations at pre-dose treatment arms at baseline. The effect of shed HER2 ECD Cycle 6 (Cycle 5 Cmin,ss) and post-dose (Cmax,ss) in Cycle on pertuzumab PK was assessed. 4 between the pertuzumab and placebo arms. In addition, the 90% CI for the ratio of the geometric means were con- structed. Potential effects of chemotherapy on pertuzumab or Results trastuzumab PK were assessed by comparing the arithmetic means of Cmin in Cycle 5 (with chemotherapy) and Cycle 7 Study population and demographics (without chemotherapy). If the 90% CI of the ratio of arith- metic means was contained within 80–125%, no apparent In JACOB, 780 patients were randomized to receive per- DDI was concluded. tuzumab plus trastuzumab and chemotherapy (n = 388) or placebo plus trastuzumab and chemotherapy (n = 392). Of Exploratory exposure–efficacy analyses these, 374 patients in the pertuzumab arm and 375 patients in the placebo arm had evaluable PK data and were included Observed individual pertuzumab exposures at Cycle 1 and at in this analysis. Patient demographics and disease charac- steady state from patients in the pertuzumab arm were used teristics at baseline in the intention-to-treat population are in the exploratory exposure–efficacy analysis. The efficacy provided in Online Resource 1 [11]. endpoint in the analysis was the primary study endpoint, OS. Patients who had not had an event at the time of data PK analyses analysis were censored at the date they were last known to be event free. The primary exposure metrics used in the Summary statistics of pertuzumab PK in patients treated exposure–efficacy analysis were individual-observed Cycle with pertuzumab are presented in Table 1. Following the 1 Cmin and Cycle 5 Cmin,ss. Individuals who died prior to first IV administration of pertuzumab 840 mg, the mean Cycle 6 were not included in the analysis. Kaplan–Meier observed Cmax ± standard deviation (SD) for pertuzumab curves were generated to determine survival probability was 258 ± 90.3 μg/mL. The mean observed Cmin,ss ± SD of Table 1  Summary of pharmacokinetic parameters Drug substance Cmax Cmin Cmin Cmax Cmin Cmin Cmax (treatment arm) Cycle 1 (μg/mL) Cycle 1 (μg/mL) Cycle 3 (μg/mL) Cycle 4 (μg/mL) Cycle 5 (μg/mL) Cycle 7 (μg/mL) Cycle 8 (μg/mL) Pertuzumab (pertuzumab arm) n 374 349 305 302 274 114 106 Mean (SD) 258 (90.3) 42.4 (24.8) 90.4 (42.4) 341(111) 114 (51.8) 142 (67.9) 371 (127) Trastuzumab (pertuzumab arm) n 372 345 305 304 274 114 115 Mean (SD) 142 (86.8) 15.4 (11.3) 22.9 (12.7) 127 (50.9) 26.3 (14.8) 32.7 (15.0) 130 (50.8) Trastuzumab (placebo arm) n 375 354 300 299 254 92 90 Mean (SD) 139 (58.6) 17.2 (15.4) 24.1 (19.0) 129 (58.1) 29.8 (21.9) 37.4 (20.3) 147 (90.2) Cmax peak (maximum) serum concentration, Cmin serum trough (minimum) concentration, SD standard deviation 13 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 543 Fig. 1  Cross-study comparison of PK for a pertuzumab [7, 9, 11] and post-dose; CLEOPATRA, Cycle 9 post-dose; ToGA, Cycle 5 post- b trastuzumab [7, 11, 14]. Cmax,ss steady-state peak (maximum) serum dose. Red bars = Cmin; blue bars = Cmax; lower and upper ends of concentration, Cmin,ss steady-state serum trough (minimum) concen- each box plot = 25th and 75th percentile exposure value; horizontal tration, PK pharmacokinetic. Cycle numbers for Cmin,ss and Cmax,ss white line = median per group; points = individual PK data. Brackets vary between studies. Cmin,ss: JACOB, Cycle 3 pre-dose; JOSHUA, extending from the ends of the box are drawn to the nearest value, not Cycle 4 pre-dose; CLEOPATRA, Cycle 9 pre-dose; ToGA, Cycle beyond 1.5 times the interquartile range 9 pre-dose. Cmax,ss: JACOB, Cycle 4 post-dose; JOSHUA, Cycle 4 13 544 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 13 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 545 ◂Fig. 2  Drug–drug interaction assessment of the potential effects arm was 101.5% (90% CI 94.8–108.8). Trastuzumab PK of a chemotherapy on pertuzumab exposure and b pertuzumab on parameters were comparable between the pertuzumab and trastuzumab exposure. C Cycle, Chemo chemotherapy, Cmax peak (maximum) serum concentration, Cmin serum trough (minimum) placebo arms, suggesting that adding pertuzumab-to-trastu- concentration, PK pharmacokinetic. Red bars = post-dose; blue zumab treatment did not alter the PK of trastuzumab (Fig. 2). bars = pre-dose; lower and upper ends of each box plot = 25th and For patients in the pertuzumab arm who completed the 75th percentile exposure value; horizontal white line = median per last chemotherapy treatment in Cycle 6 and continued per- group; points = individual PK data. Brackets extending from the ends of the box are drawn to the nearest value, not beyond 1.5-times the tuzumab 840 mg treatment without chemotherapy thereafter, interquartile range the Cycle 7 observed that mean Cmin of pertuzumab was 142 ± 67.9 μg/mL. This was comparable to the pertuzumab PK exposure in Cycle 5 when patients were treated with per- pertuzumab following five subsequent pertuzumab doses at tuzumab in combination with trastuzumab and chemother- 840 mg q3w at Cycle 5 was 114 ± 51.8 μg/mL (Table 1). apy (Fig. 2a), indicating that chemotherapy had no apparent The target Cmin,ss of ≥ 20 μg/mL was reached in 99.3% of impact on pertuzumab PK. patients at Cycle 5 Cmin,ss and beyond. In addition, > 90% of For patients who completed the last chemotherapy treat- patients were above the PK target right after the first pertu- ment in Cycle 6 and continued 6 mg/kg trastuzumab without zumab dose. chemotherapy thereafter, the Cycle 7 observed mean Cmin Summary statistics of trastuzumab PK in patients treated of trastuzumab was 32.7 ± 15.0 μg/mL in the pertuzumab with trastuzumab (pertuzumab and placebo arms) are pre- arm and 37.4 ± 20.3 μg/mL in the placebo arm, which was sented in Table 1. Following the loading dose IV admin- comparable to the PK exposure in Cycle 5 when patients istration of trastuzumab at 8 mg/kg, the mean observed were treated with pertuzumab or placebo in combination Cmax ± SD of trastuzumab was 142 ± 86.8 μg/mL in the with trastuzumab and chemotherapy (Fig. 2b), suggesting pertuzumab arm and 139 ± 58.6 μg/mL in the placebo arm that chemotherapy had no impact on trastuzumab PK. (Table 1). The mean observed Cmin,ss of trastuzumab follow- ing five subsequent trastuzumab maintenance doses of 6 mg/ Exposure–efficacy relationships kg q3w in Cycle 5 was 26.3 ± 14.8 μg/mL in the pertuzumab arm and 29.8 ± 21.9 μg/mL in the placebo arm (Table 1). There were no statistically significant differences in OS across Cycle 1 pertuzumab Cmin or Cycle 5 pertuzumab Comparison of pertuzumab and trastuzumab PK Cmin,ss quartiles (Fig. 3). The median OS in weeks for Cycle data from JACOB with the previous studies in MBC 1 pertuzumab Cmin quartiles were Q1: 43, Q2: 23, Q3: 28, and AGC, and across geographic regions and Q4: 40 (n = 349, p = 0.52) and the median OS in weeks for Cycle 5 pertuzumab Cmin,ss quartiles were Q1: 26, Q2: Pertuzumab Cmin,ss from JACOB was comparable to the pre- 28, Q3: 26, and Q4: 38 (n = 274, p = 0.78). vious studies in MBC (CLEOPATRA; 840 mg loading dose followed by 420 mg maintenance dose q3w) [9] and AGC Immunogenicity (JOSHUA; 840 mg q3w arm) [7] (Fig. 1a). The higher pertu- zumab dose in the JACOB study was reflected in higher Cmax At baseline, 746 patients were evaluable for ADAs (pertu- concentrations at steady state, compared with CLEOPATRA zumab arm, n = 376; placebo arm, n = 370); 21 patients in (mean Cmax ± SD: 341 ± 111 vs. 196 ± 66.3) and consist- the pertuzumab arm and 14 patients in the placebo arm were ent with Cmax in JOSHUA (mean Cmax ± SD [840 mg q3w ADA-positive at baseline (likely reflecting false positives or arm]: 316 ± 134 μg/mL). Steady-state trastuzumab PK from pre-existing ADAs). Of the 347 ADA-evaluable patients in JACOB was comparable to the previous studies in AGC the pertuzumab arm who had at least one post-dose sample (JOSHUA [840 mg q3w arm] [7] and ToGA [14]) (Fig. 1b). available, only two were ADA-positive (0.6%). The pertu- Cmin,ss for pertuzumab across the different geographic zumab Cmin,ss for the two ADA-positive patients—135 ug/ regions are shown in Online Resource 2. Pertuzumab expo- mL and 59.8 ug/mL, respectively—are within the range sure was comparable across geographic regions. observed in the studied population. DDIs Effect of baseline shed HER2 ECD on pertuzumab PK The ratio of arithmetic means of Cmin,ss serum trastuzumab There was considerable inter-patient variability in base- concentrations at Cycle 5 (pre-dose Cycle 6) in the pertu- line HER2 ECD concentrations, as reflected by the broad zumab versus placebo arm was 90.9% (90% CI 81.5–101.5). ranges of individual values (3.5–2086.4 ng/mL) observed The ratio of arithmetic means of Cmax,ss serum trastuzumab in both treatment arms of the study; however, most patients concentrations at Cycle 4 in the pertuzumab versus placebo had baseline HER2 ECD concentrations below 100 ng/ 13 546 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 Fig. 3  Kaplan–Meier plot of overall survival by pertuzumab exposure quartiles at a Cycle 1 (Cmin) and b Cycle 5 (Cmin,ss) mL. In the pertuzumab arm, the mean ± SD and min–max not receive pertuzumab. However, there was utility in baseline ECD concentrations were 42.1 ± 119 ng/mL and conducting a PK analysis to better understand the pertu- 3.5–1350 ng/mL, respectively (n = 361). Similar results were zumab PK–pharmacodynamic relationship and to assess observed in the placebo treatment arm. No apparent rela- the appropriateness of the new pertuzumab dosing regimen tionship was seen between pertuzumab Cmin,ss and baseline selected for this indication. Collection of serum samples HER2 ECD concentrations (Fig. 4). HER2 ECD concentra- from all patients who participated in the JACOB study tions observed in JACOB were comparable to concentrations enabled this extensive characterization of the PK of pertu- observed previously in breast cancer (F. Hoffmann-La Roche zumab in patients with HER2-positive MGC/GEJC. Ltd. Data on file). The appropriateness of an 840 mg q3w dose was assessed in an exposure–efficacy analysis, which showed no statistically significant relationship between pertu- Discussion zumab exposure quartiles and the probability of survival. Our findings, therefore, provide no evidence of additional The JACOB trial did not meet its primary endpoint of efficacy with increasing pertuzumab exposure in this showing a statistically significant improvement in OS in trial. Given the overall JACOB study outcome, which patients who received pertuzumab in addition to trastu- showed that OS was not significantly different between zumab and chemotherapy, compared to those who did treatment and control (median OS 17.5 months [95% CI 13 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 547 Fig. 4  Pertuzumab steady-state exposure vs. HER2 ECD con- centrations. Cmin serum trough (minimum) concentration, ECD extracellular domain, HER2 human epidermal growth factor receptor 2 16.2–19.3] in the pertuzumab arm and 14.2 months [95% ToGA [14]. In addition, in JACOB, trastuzumab Cmin at CI 12.9–15.5] in the control arm; hazard ratio 0.84 [95% Cycles 1 and 5 were similar in both the pertuzumab and CI 0.71–1.00]; p = 0.57), the power to detect a difference placebo treatment arms. between exposure quartiles in the treatment arm would be Pertuzumab was well tolerated in the JACOB study and predictably very small. the safety profile was generally similar between the two In JACOB, 99.3% of patients with HER2-positive treatment groups [11]. No new or unexpected safety events MGC/GEJC achieved target Cmin,ss, the target efficacious were reported, which is important given that the dose of exposure based on the previous non-clinical efficacy mod- 840 mg pertuzumab every 3 weeks was double the main- els [10] when receiving the 840 mg q3w pertuzumab dose. tenance dose currently approved for breast cancer treat- In comparison, pertuzumab PK data from the registrational ment [11]. Given the lack of significant improvements in phase II NeoSphere study (pertuzumab, trastuzumab, efficacy outcomes and that pertuzumab was well tolerated and docetaxel in the neoadjuvant treatment of HER2- at a higher dose, decreasing the pertuzumab dose was not positive EBC; NCT00545688) showed that the 840 mg justified as a viable treatment option, and therefore, no loading dose followed by 420 mg q3w pertuzumab regi- exposure–safety analyses were warranted. men resulted in 94% of patients reaching the target Cmin,ss No DDIs were found in this study, which is as expected of > 20 μg/mL [10]. The exposure–response analysis based on prior data on pertuzumab and trastuzumab PK from NeoSphere also showed that there was no significant [10, 12] and the distinct clearance mechanisms between impact (p = 0.996) on the probability of the primary study mAbs and small-molecule chemotherapy agents [16–18]. endpoint (pathologic complete response in the breast) with Pertuzumab and trastuzumab bind to distinct epitopes of an increase in pertuzumab serum concentration beyond 20 HER2 simultaneously without steric hinderance, providing μg/mL [10]. The findings of this exposure–response analy- an additional rationale for why no DDI was expected [19]. sis were also replicated in the phase III APHINITY study There was no apparent impact of ADAs on pertuzumab (pertuzumab, trastuzumab, and chemotherapy for the adju- PK in patients who were ADA-positive, and no apparent vant treatment of HER2-positive EBC; NCT01358877) impact on safety as neither ADA-positive patient had any [15]. These two studies in patients with breast cancer serious immune-related AEs nor other AEs suggestive of support the selection of ≥ 20 µg/mL as a rational target hypersensitivity. Although data were limited, there was serum trough exposure level of pertuzumab for therapeutic no apparent impact on efficacy, as both patients were efficacy. Therefore, the 840 mg q3w dose, which enabled identified as partial responders (OS: 39 and 15 months; target pertuzumab concentrations to be achieved in the vast progression-free survival: 14 and 11 months). The ADA majority of patients with HER2-positive MGC/GEJC in results from the JACOB study confirm the low immuno- the JACOB study, appears to be an appropriate selection genic potential of pertuzumab, as reported in other indica- from a PK/pharmacodynamic perspective. Furthermore, tions [1]. efficacious exposures of trastuzumab were also obtained Pertuzumab exposure in JACOB was consistent with in JACOB, as evidenced by a consistent % of patients prior studies in AGC (the phase IIa dose-finding study, reaching the target trastuzumab Cmin as in the pivotal trial JOSHUA, which used 840 mg q3w) and breast cancer (e.g., 13 548 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 CLEOPATRA, which used 840 mg loading dose followed hypoalbuminemia [25], which is known to be negatively cor- by 420 mg q3w). This further supports the existence of a related with mAb clearance [9, 14, 21, 23]. Yang et al. [26] significant difference in the clearance of pertuzumab based showed a decrease in murine IgG1 mAb area under the curve on indication. The difference in pertuzumab PK in AGC for 0–14 days from 1368 ± 255 to 594 ± 224 μg/mL/day in a vs. MBC indications identified in JOSHUA led to the deci- murine model of colitis and PLE, providing preclinical support sion to pursue an 840 mg q3w dosing regimen in JACOB of this potential hypothesis. rather than the 840 mg/420 mg q3w regimen approved in the Further studies are needed to identify first-line treatment HER2-positive breast cancer indication [1, 2]. The observed options to improve patient outcomes in HER2-positive AGC difference in clearance in AGC compared with other indi- and to better identify patients who might benefit from dual cations is consistent with other mAbs and antibody–drug anti-HER2 targeted regimens [11]. Using methods beyond conjugates such as trastuzumab, bevacizumab, and ado- the current process for identifying HER2-positive AGC may trastuzumab emtansine [20]. compensate for the intratumoral heterogeneity observed in It should be noted that exposures in JACOB were limited GC, which may have negatively impacted responses to tar- to Cmax and Cmin only, due to the sparse PK sampling sched- geted therapies in this and other trials [27, 28]. Furthermore, ule employed in this large Phase III clinical trial. Within a comprehensive cross-antibody covariate analysis, as well cycle pertuzumab, PK samples collected in the previous as potential in vitro experiments and/or clinical studies, are gastric cancer study (JOSHUA), pooled with the sparse PK warranted to further understand the clearance difference and samples collected in the JACOB study, were deemed suffi- to support appropriate dosing regimens of other biologics in cient to characterize pertuzumab PK in this patient popula- this indication. tion. Here, we rely on Cmax and Cmin only when comparing across geographies or to prior studies such as JOSHUA, Acknowledgements We would like to acknowledge all volunteers, patients, and investigators participating in this study. This study is ToGA, and CLEOPATRA. This comparison gives a reason- sponsored by F. Hoffmann-La Roche Ltd, Basel, Switzerland. Sup- able idea of exposure similarities or differences. Without port for third-party writing assistance for this manuscript, furnished extensive sampling and the fact that post-dose and pre-dose by Rachel Johnson, PhD, of Health Interactions, was provided by F. PK samples were not collected during the same cycles in Hoffmann-La Roche Ltd. JACOB, the onset of steady state was concluded to be Cycle Funding This study is funded by F. Hoffmann-La Roche Ltd, Basel, 4 for Cmax,ss and Cycle 5 for Cmin,ss. Switzerland. A popPK model of trastuzumab showed that no single covariate could explain the PK difference between AGC and Compliance with ethical standards MBC [21]. Hypotheses to explain possible reasons for faster clearance of HER2-targeting molecules in AGC include both Conflict of interest All authors received support for third-party writing target- and non-target-related mechanisms. HER2-target assistance for this manuscript, provided by F. Hoffmann-La Roche Ltd. mechanisms include tumor burden and HER2 shed ECD. Whitney Kirschbrown, Ihsan Nijem, Amit Garg, and Sandhya Girish are employees of Genentech, Inc., and hold stock in Roche Holding Tumor burden was ruled out as a potential mechanism in Ltd. Whitney Kirschbrown is an inventor on a pertuzumab-related fil- previously published popPK analyses [14]. Shed HER2 ECD ing. Bei Wang is an employee of Genentech, Inc. Atsushi Ohtsu has was investigated within the JACOB study, as it was a signifi- received research grants from BMS and honoraria from BMS, Ono, cant covariate for the clearance of trastuzumab or ado-tras- Chugai, and Taiho. Paulo M. Hoff has received research grants from Roche relating to the conduct of this study. Manish A. Shah disclos- tuzumab emtansine in MBC, although in these studies, the es research funding paid to his institution from Boston Biomedical, magnitude of impact was relatively small and not of clinical Merck, and Roche. Lin Shen received non-financial support from relevance [22, 23]. In the JACOB study, baseline shed HER2 Roche Pharmaceuticals Ltd. relating to the conduct of this study and ECD levels were comparable to levels seen previously in a research grant from Hengrui Medicine Co. Ltd, indirectly related to this manuscript. Yoon-Koo Kang has received research grants from breast cancer patients [F. Hoffmann-La Roche Ltd. Data on Roche, Daehwa and LSK Biopharma, and has received personal fees file] and there was no apparent relationship between baseline from Ono, BMS, Novartis, Lilly, Daehwa, and LSK Biopharma. Maria shed HER2 ECD and steady-state pertuzumab Cmin. Alsina has performed an advisory role for BMS, Servier, and MSD, One promising non-target hypothesis is potential gas- received research funding from Merck-Serono, and received speaker fees from MSD, BMS, Lilly, Roche, Amgen, and AstraZeneca. tric protein leakage or protein-losing enteropathy (PLE) in patients with AGC. There is no direct clinical evidence for Ethical standards The study was conducted in accordance with the a relationship between PLE and the PK of mAbs; however, ethical standards established in the 1964 Declaration of Helsinki and its it is conceivable that PLE could be a driving factor in faster later amendments. The study was approved by the institutional review board of each study site, and all patients participating in the study mAb clearance, given that GC is one of the disease states asso- provided informed consent before enrollment. ciated with PLE [24] and patients with PLE often develop 13 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 549 Open Access This article is distributed under the terms of the Crea- 14. Cosson VF, Ng VW, Lehle M, Lum BL (2014) Population phar- tive Commons Attribution 4.0 International License (http://creat​iveco​ macokinetics and exposure–response analyses of trastuzumab in mmons​.org/licen​ses/by/4.0/), which permits unrestricted use, distribu- patients with advanced gastric or gastroesophageal junction can- tion, and reproduction in any medium, provided you give appropriate cer. Cancer Chemother Pharmacol 73:737–747 credit to the original author(s) and the source, provide a link to the 15. Kirschbrown WP, Kagedal M, Wang B et al (2018) Pharmacoki- Creative Commons license, and indicate if changes were made. netic and exploratory exposure-response analysis of pertuzumab in patients with operable HER2-positive early breast cancer. Clin Pharmacol Ther 103(Suppl):S74 16. Keizer RJ, Huitema AD, Schellens JH, Beijnen JH (2010) Clini- References cal pharmacokinetics of therapeutic monoclonal antibodies. Clin Pharmacokinet 49:493–507 1. Genentech Inc (2018) Perjeta™ (pertuzumab). Prescribing Infor- 17. Sonnichsen DS, Relling MV (1994) Clinical pharmacokinetics of mation (USA). https​://www.acces​sdata​.fda.gov/drugs​atfda​_docs/ paclitaxel. Clin Pharmacokinet 27:256–269 label​/2018/12540​9s121​lbl.pdf. Accessed 18 Dec 2018 18. van der Vijgh Wim JF (1991) Clinical pharmacokinetics of car- 2. Roche Registration Ltd (2018) ­Perjeta® (pertuzumab). Summary boplatin. Clin Pharmacokinet 21:242–261 of Product Characteristics. https​://www.ema.europ​a.eu/docum​ 19. Lu D, Burris HA 3rd, Wang B et al (2012) Drug interaction ents/produ​ct-infor​matio​n/perje​ta-epar-produ​ct-infor​matio​n_en-0. potential of trastuzumab emtansine (T-DM1) combined with per- pdf. Accessed 18 Dec 2018 tuzumab in patients with HER2-positive metastatic breast cancer. 3. Harbeck N, Beckmann MW, Rody A et al (2013) HER2 dimeriza- Curr Drug Metab 13:911–922 tion inhibitor pertuzumab—mode of action and clinical data in 20. Chen SC, Kagedal M, Gao Y et al (2017) Population pharmacoki- breast cancer. Breast Care (Basel) 8:49–55 netics of trastuzumab emtansine in previously treated patients with 4. Baselga J, Cortés J, Kim SB et al (2012) Pertuzumab plus trastu- HER2-positive advanced gastric cancer (AGC). Cancer Chem- zumab plus docetaxel for metastatic breast cancer. N Engl J Med other Pharmacol 80:1147–1159 366:109–119 21. Han K, Jin J, Maia M, Lowe J, Sersch MA, Allison DE (2014) 5. Swain SM, Baselga J, Kim SB et al (2015) Pertuzumab, trastu- Lower exposure and faster clearance of bevacizumab in gastric zumab, and docetaxel in HER2-positive metastatic breast cancer. cancer and the impact of patient variables: analysis of individual N Engl J Med 372:724–734 data from AVAGAST phase III trial. AAPS J 16:1056–1063 6. von Minckwitz G, Procter M, de Azambuja E et al (2017) Adju- 22. Bruno R, Washington CB, Lu JF, Lieberman G, Banken L, Klein vant pertuzumab and trastuzumab in early HER2-positive breast P (2005) Population pharmacokinetics of trastuzumab in patients cancer. N Engl J Med 377:122–131 with HER2 + metastatic breast cancer. Cancer Chemother Phar- 7. Kang YK, Rha SY, Tassone P et al (2014) A phase IIa dose-find- macol 56:361–369 ing and safety study of first-line pertuzumab in combination with 23. Lu D, Girish S, Gao Y et al (2014) Population pharmacokinetics trastuzumab, capecitabine and cisplatin in patients with HER2- of trastuzumab emtansine (T-DM1), a HER2-targeted antibody- positive advanced gastric cancer. Br J Cancer 111:660–666 drug conjugate, in patients with HER2-positive metastatic breast 8. Bang YJ, Van Cutsem E, Feyereislova A et al (2010) Trastuzumab cancer: clinical implications of the effect of covariates. Cancer in combination with chemotherapy versus chemotherapy alone for Chemother Pharmacol 74:399–410 treatment of HER2-positive advanced gastric or gastro-oesopha- 24. Nakatani N (1994) Quantitative study of serum protein loss into geal junction cancer (ToGA): a phase 3, open-label, randomised the alimentary tract in patients with gastric cancer [article in Japa- controlled trial. Lancet 376:687–697 nese]. Nihon Shokakibyo Gakkai Zasshi 91:1391–1398 9. Garg A, Quartino A, Li J et al (2014) Population pharmacokinetic 25. Strygler B, Nicar MJ, Santangelo WC, Porter JL, Fordtran JS and covariate analysis of pertuzumab, a HER2-targeted monoclo- (1990) Alpha 1-antitrypsin excretion in stool in normal subjects nal antibody, and evaluation of a fixed, non-weight-based dose in and in patients with gastrointestinal disorders. Gastroenterology patients with a variety of solid tumors. Cancer Chemother Phar- 99:1380–1387 macol 74:819–829 26. Yang Y, Li TR, Balthasar JP (2017) Investigation of the influence 10. Quartino AL, Li H, Jin JY et al (2017) Pharmacokinetic and of protein-losing enteropathy on monoclonal antibody pharma- exposure–response analyses of pertuzumab in combination with cokinetics in mice. AAPS J 19:1791–1803 trastuzumab and docetaxel during neoadjuvant treatment of HER2 27. Alsina M, Gullo I, Carneiro F (2017) Intratumoral heterogeneity early breast cancer. Cancer Chemother Pharmacol 79:353–361 in gastric cancer: a new challenge to face. Ann Oncol 28:912–913 11. Tabernero J, Hoff PM, Shen L et al (2018) Pertuzumab plus tras- 28. Gullo I, Carneiro F, Oliveira C, Almeida GM (2018) Heterogene- tuzumab and chemotherapy for HER2-positive metastatic gastric ity in gastric cancer: from pure morphology to molecular clas- or gastro-oesophageal junction cancer (JACOB): final analysis sifications. Pathobiology 85:50–63 of a double-blind, randomised, placebo-controlled phase 3 study. Lancet Oncol 19:1372–1384 Publisher’s Note Springer Nature remains neutral with regard to 12. Cortés J, Swain SM, Kudaba I et al (2013) Absence of pharma- jurisdictional claims in published maps and institutional affiliations. cokinetic drug-drug interaction of pertuzumab with trastuzumab and docetaxel. Anticancer Drugs 24:1084–1092 13. Liu L, Xu K, Li J et al (2018) Optimizing hybrid LC-MS/MS binding conditions is critical: impact of biotransformation on quantification of trastuzumab. Bioanalysis 10(22):1819–1831 13 550 Cancer Chemotherapy and Pharmacology (2019) 84:539–550 Affiliations Whitney P. Kirschbrown1 · Bei Wang1 · Ihsan Nijem1 · Atsushi Ohtsu2 · Paulo M. Hoff3 · Manish A. Shah4 · Lin Shen5 · Yoon‑Koo Kang6 · Maria Alsina7 · Sandhya Girish1 · Amit Garg1 1 5 Clinical Pharmacology, Genentech Research and Early Key Laboratory of Carcinogenesis and Translational Development, DNA Way, MS463a, South San Francisco, Research (Ministry of Education/Beijing), Department CA 94080, USA of Gastrointestinal Oncology, Peking University Cancer 2 Hospital and Institute, Beijing, China Department of Gastrointestinal Oncology, National Cancer 6 Center Hospital East, Kashiwa, Japan Department of Oncology, Asan Medical Center, University 3 of Ulsan College of Medicine, Seoul, Korea Instituto do Câncer de São Paulo, Faculdade de Medicina da 7 Universidade de São Paulo, São Paulo, Brazil Vall d’Hebron University Hospital and Institute of Oncology 4 (VHIO), Universitat Autònoma de Barcelona, Barcelona, Medical Oncology/Solid Tumor Program, Sandra Spain and Edward Meyer Cancer Center, Weill Cornell Medical College, New York, NY, USA 13