Plinabulin and Peglgrastim in Combination for the Prevention of Chemotherapy-Induced Neutropenia in Patients with Breast Cancer (PROTECTIVE-2): A Randomized Trial

Purpose Plinabulin is a non-granulocyte colony-stimulating factor (G-CSF) novel small molecule with both anticancer and myeloprotective effects. Single-agent plinabulin is myeloprotective in the rst week of the chemotherapy cycle, and peglgrastim in the second week. We assessed the ecacy and safety of the combination of plinabulin and peglgrastim for the prevention of chemotherapy-induced neutropenia (CIN) following chemotherapy. primary objective to establish the recommended Secondary endpoints included absolute neutrophil count (ANC) relative dose intensity (RDI), and incidence of adverse events including neutropenia and bone


Introduction
Neutropenia and febrile neutropenia (FN) after myelotoxic chemotherapy are predictive biomarkers for development of infection-related complications -sepsis, antibiotic use, unplanned hospitalization, centers. Exclusion criteria included concurrent administration of chemotherapy or radiation therapy, active infection, or the use of strong cytochrome P4503A4 inhibitors.
All patients received docetaxel 75 mg/m 2 , doxorubicin 50 mg/m 2 , and cyclophosphamide 500 mg/m 2 (TAC) on Day 1 by intravenous infusion (IV), repeated every three weeks for four cycles. We evaluated patients in two separate, sequential cohorts with parallel randomization schemes. In the monotherapy cohort, patients were randomized to peg lgrastim 6 mg by subcutaneous injection (SQ) on Day 2 or plinabulin IV at either 10, 20, or 30 mg/m 2 , 30 minutes after docetaxel on Day 1, with no Day 2 treatment. (Online Resource Fig. 1a). Based on the results of the single-agent plinabulin study, a superiority objective would not be met. We amended the protocol to evaluate an additional combined therapy cohort: plinabulin 20 mg/m 2 with peg lgrastim 1.5, 3, or 6 mg on Day 2, compared with monotherapy peg lgrastim 6 mg (Online Resource Fig. 1b). Randomization was by the Medidata electronic data capture (EDC) system.
At investigator discretion, doxorubicin could be omitted (that is, TAC converted to TC) during Cycles 2 through 4 or continued for more than four cycles. Docetaxel premedication with corticosteroids and rescue medications to treat FN was pre-speci ed.

Trial Outcomes
The primary objective was to establish the recommended Phase 3 dose (RP3D) of plinabulin based on pharmacokinetic (PK) and pharmacodynamic (PD) analyses. The primary e cacy PD endpoint was DSN (Grade 4; de ned as ANC <0.5 × 10 9 /L). The primary safety endpoint was blood pressure on Day 1 of Cycle 1 within 3.5 hours after the docetaxel infusion. PK of plinabulin, peg lgrastim and the TAC components, and the plinabulin PK and PD for CIN prevention was modeled with a nonlinear mixed effects approach (NONMEM, ICON Development, Ellicott City, MD) using a sequential PK/PD approach for plinabulin CIN prevention. Separate population PK models were developed for plinabulin, docetaxel, doxorubicin, cyclophosphamide, and peg lgrastim (Online Resource Supplemental Material).
Secondary endpoints included the frequency of patients with at least one day of Grade 4 neutropenia, Grade 3 or 4 neutropenia, ANC nadir, change in bone pain between pretreatment during cycle 1, relative dose intensity, safety, and tolerability. Bone pain was evaluated by using the Brief Pain Inventory [19,20] before trial drug infusion on Day 1 compared with Days 1-8 of Cycle 1. Tolerability is de ned as the number of patients who received >85% of the planned chemotherapy dose.
Exploratory endpoints included FN incidence, CD34+ stem cell mobilization, health-related quality of life (QOL), and thrombocytopenia. The European Organization for Research and Treatment of Cancer Quality of Life Questionnaire C30 (EORTCQLQ-C30) was used to evaluate QOL [21][22][23]. Changes from baseline in platelet count were used to investigate the incidence of thrombocytopenia during the trial. The frequency of patients with thrombocytopenia was analyzed for each treatment group in the combined therapy cohort.

Statistical Methods
Patients were strati ed by region: China versus the rest of the world. The peg lgrastim 6 mg monotherapy data served as the comparator arm for both cohorts, and hypotheses testing used a signi cance level of 0.05. In the monotherapy cohort, with an 80% power to detect an event rate ratio of 0.3 and a signi cance level of 0.05, up to 18 patients in each arm were required. In the combined therapy cohort, two sequential tests were made by using the O'Brien-Fleming spending function to determine the test boundaries. During an interim analysis, if a statistically signi cantly superior outcome with respect to DSN was reached when comparing peg lgrastim alone to any of the combination treatments, or an indication that superiority could not be achieved, then no further patients were allocated to the combination arms.
All randomized patients were included in the intent-to-treat (ITT) analysis set, and the safety analysis set included all randomized patients who received at least one dose of any trial medication. Stata v11.0 or later generalized Poisson model procedures were used to assess Cycle 1 DSN. Barnard's method was used to evaluate the difference in proportions of Grade 4 and Grade 3/4 neutropenia between the treatment arms. For the assessment of bone pain, the mean change from pre-dose (Day 1) bone pain score to Days 2 through 8 in Cycle 1 were analyzed by using repeated measures mixed linear model with the predose (Day 1) value and treatment arm as covariates. The method was used to construct point estimates and con dence intervals (CIs). Post-hoc analyses of mean bone pain score by visit and treatment group were performed by using a repeated measure mixed linear model, with the baseline value and treatment arm as covariates. Continuous variables were summarized with counts, means, standard deviations, medians, CIs, minimums, and maximums. Categorical variables were summarized by counts and by the percentage of patients. Descriptive summaries were also provided. Missing or invalid data were not imputed. The PK/PD modeling approach adopted to estimate the population PK of plinabulin and the TAC drugs and to characterize the plinabulin PK and PD in preventing CIN is summarized in Online Resource Supplemental Material.
All patients provided written informed consent. The trial was conducted following International Council for Harmonization Good Clinical Practice guidelines and ethical principles and approved by the relevant Independent Ethics Committee or Institutional Review Board at each site. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request 3 Results One hundred fteen patients were treated, 42 in the monotherapy and 51 in the combined therapy cohorts; 22 patients who were concurrently randomized to peg lgrastim 6 mg were included in both cohorts for comparison purposes. (Fig. 1a, Fig. 1b, Online Resource Fig. 2a, Fig. 2b) Patients were accrued between December 2017 and January 2019 at 9 sites in China and Ukraine. There were no pretreatment demographic differences among the groups. (Online Resource Tables 1a and 1b).
For the primary endpoint -determination of the RP3D -results from all seven treatment arms were evaluated. In the monotherapy cohort, the single agent plinabulin arms were not superior to peg lgrastim ( Table 1a). The ANC recovery curves are illustrated in Fig. 2a, and distribution of the DSN is shown in Online Resource Fig. 3a. Plinabulin's bene t is protecting the ANC during Week 1 after chemotherapy, while peg lgrastim's protective effect is constant during Week 2. With results from the Phase 2 clinical trial (PROTECTIVE-1) also showing Week 1 CIN prevention in plinabulin monotherapy [9], we modi ed the study design to include combinations of plinabulin and peg lgrastim. Plinabulin 20 mg/m 2 with peg lgrastim 6 mg performed best, with a mean DSN in Cycle 1 of 0.69 days vs 0.73 days for peg lgrastim alone (P non-signi cant, Table 1B). The ANC recovery curves for all combination arms are illustrated in Fig. 2b. The DSN distribution for this best performing combination compared with the control is shown in Fig. 2c.
Plinabulin did not have an effect on AMBP, the primary safety objective (see Online Resource Fig. 4).
For the secondary endpoint of frequency of at least one day of Grade 4 neutropenia, 37.5% of patients receiving plinabulin 20 mg/m 2 and peg lgrastim 6 mg combination had at least one day of Grade 4 neutropenia, compared with 59.1% for patients receiving peg lgrastim 6 mg alone (Online Resource Table  4b). For bone pain, patients taking peg lgrastim 6 mg alone experienced a higher number of days of bone pain than did patients taking combined plinabulin 20 mg/m 2 and peg lgrastim 6 mg ( Fig. 2d). For chemotherapy tolerability, 100% of patients in the plinabulin 20 mg/m 2 and peg lgrastim 6 mg arm were able to receive >85% of their planned chemotherapy dose compared with 82% in the peg lgrastim alone arm (Table 2a). The relative dose intensity (RDI) also favored the combination arm (0.998 vs 0.952) as fewer patients in the combination arm were changed to TC by their oncologists (6.25%) than in peg lgrastim alone arm (14.29%) (Table 2b).
For the combined therapy cohort other secondary endpoints favored the combination: FN events (Online Resource Table 5) treatment exposure (Online Resource Table 6) and CD34+ stem cell mobilization (Online Resource Table 7). Combined plinabulin 20 mg/m 2 and peg lgrastim 6 mg did not alter EORTCQLQC30 scores compared with peg lgrastim alone (Online Resource Fig. 6). Reductions in platelet counts were larger with peg lgrastim 6 mg monotherapy than with any of the plinabulin monotherapy treatments (Online Resource Fig. 7a). In the combined therapy cohort, platelet counts decreased with increasing peg lgrastim dose (Online Resource Fig. 7b). In correlation with this, the incidence of all grades of thrombocytopenia increased with increasing peg lgrastim dose (Online Resource Fig. 8).
To investigate use of a xed dose of plinabulin, we performed extensive PK analyses similar to the equivalence modeling for PROTECTIVE1 [9]. The xed dose plinabulin 40 mg is equivalent to plinabulin 20 mg/m 2 (Online Resource Supplemental Material and Online Resource Fig. 9). The RP3D for the next phase of PROTECTIVE-2 clinical trial is plinabulin 40 mg on the day of chemotherapy and peg lgrastim 6 mg on the day after chemotherapy.

Adverse Events
In the monotherapy cohort, all patients experienced at least one treatment-emergent adverse event (TEAE) ( Table 3a). In the peg lgrastim 6 mg group, 21 of 22 patients (95.5%) experienced at least one Adverse Event (AE). The frequency of serious AEs was not substantially different between the plinabulin 20 mg/m 2 cohort (13.3%) and the peg lgrastim 6 mg cohort (9.1%). Two deaths occurred in the monotherapy cohort: one patient in the plinabulin 10 mg/m 2 group died of sepsis, and one patient in the plinabulin 30 mg/m 2 group died of septic shock. In the combination therapy cohort, all patients experienced at least one TEAE (Table 3b). The RP3D combination arm had lower frequencies of Grade 3 (18.8% vs 31.8%, respectively) and Grade 4 (37.5% vs 54.5%) AEs compared with peg lgrastim alone. Serious AEs were inversely correlated with peg lgrastim dose, primarily due to more CINrelated AEs. One patient in the plinabulin 20 mg/m 2 and peg lgrastim 3 mg group died of heart failure during the trial. No infections or sepsis, but one hospitalization, occurred in the plinabulin 20 mg/m 2 and peg lgrastim 6 mg group.
Changes in systolic and diastolic blood pressure were not clinically signi cant; they were transient and of short duration. There were no clinically relevant changes in clinical biochemistry, hematology, or urinalysis parameters in any treatment arm.

Discussion
Single agent plinabulin protects the ANC in the rst week of the chemotherapy cycle, and single agent peg lgrastim protects the ANC in the second week. This observation led to successfully testing the combination. First, we demonstrated that the monotherapy 20 mg/m 2 plinabulin dose performed better than other plinabulin doses but did not provide optimum ANC protection. By protocol amendment, we tested the combination of 20 mg/m 2 plinabulin dose with de-escalating peg lgrastim doses and demonstrated that plinabulin 20 mg/m 2 on Day 1 with peg lgrastim 6 mg on Day 2 was best performing. This result was con rmed with detailed pharmacokinetic and pharmacodynamic modeling. We next modeled plinabulin 40 mg as a xed dose and found PK equivalence to the body surface area-based dose of 20 mg/m 2 , similar to the result in our previous study [9]. These differences are consistent with in vitro observations of a different mechanism of action for plinabulin compared with G-CSF-derived peg lgrastim [8,24]. On the basis of the shape of the clinical neutrophil recovery curves and the in vitro data, we hypothesize that the protection of hematopoietic stem cells from docetaxel-induced damage explains plinabulin's neutrophil protective effects. The dose and schedule for testing in the Phase 3 portion of PROTECTIVE-2 will be plinabulin 40 mg on Day 1 and peg lgrastim 6 mg Day 2.
The combination facilitates the delivery of higher relative dose intensity. Maintenance of chemotherapy dose and dose intensity improves cure in several settings. In early-stage breast cancer, dose-dense chemotherapy improves outcomes and reduces complications [25,26]. In patients with lymphoma treated with curative intent, dose-dense chemotherapy also improves outcomes in older patients [27].
Maintenance of full-dose chemotherapy with G-CSF support in palliative settings is not currently recommended, as costs are not outweighed by the bene ts [28]. With new supportive care agents including G-CSF biosimilars and agents such as plinabulin and trilaciclib with different mechanisms of actions and cost, [29,30], and increases in hospital care costs, the cost-bene t calculus will change. Furthermore, the patient bene t of avoiding dose delay, dose reduction, and dose deferral should be revisited.
Delivery of higher dose-intensity chemotherapy is accompanied by higher chemotherapy complications, including febrile neutropenia (FN) and its sequelae -antibiotic use, unplanned hospitalization, and death. The FN rate for dose dense AC in the initial clinical trial was 3% [25] but is higher in general use, as indicated from observational studies [31], and is considered high FN-risk (i.e., greater than 20%) by the NCCN guideline panel [32]. Other two-drug regimens, such as three-week AC (i.e., non-dose dense), are also high FN-risk, and real-world FN rates range as high as 25% in Asian populations [33]. The two-agent docetaxel cyclophosphamide regimen (TC) is also widely used [34] and is considered by the NCCN panel to be a high neutropenic fever risk. Our results suggest that the FN rates associated with the three-drug combination TAC can by improved with the combination plinabulin and peg lgrastim, and the two-drug combination TC can also have lower FN risk. The Phase 2 results we report are the basis for the Phase 3 clinical trial, which will test the ability of the combination plinabulin and peg lgrastim to lower the FN risk, as well as the clinical consequences of FN.
No new toxicities emerged during our trial with combined plinabulin with peg lgrastim. Hypertension incidence was low and transient during the trial. There was no dosedependent relationship between post plinabulin blood pressure and plinabulin dose, as shown by the PK/PD analysis. Treatment related AEs were of lower severity with the combination compared with single-agent peg lgrastim. No deaths were considered by the investigators to be related to plinabulin or peg lgrastim treatments. We did not detect any ANC-related differences in patients by geographic origin, as has been observed for docetaxel [35][36][37].
An unexpected positive nding in our Phase 2 trial was the mitigation of G-CSF-related bone pain with the combination therapy [38--41]. Patients in the peg lgrastim alone group experienced more bone pain, and more days of bone pain, than did patients taking combined peg lgrastim and plinabulin. No patients in the RP3D arm experienced 5 or more days of bone pain compared with 4 (18%) of patients taking peg lgrastim alone. In addition, patients in all plinabulin arms experienced less thrombocytopenia than patients receiving peg lgrastim, and statistically equivalent quality of life.

Limitations
DSN, when less than 2 days, probably has low predictive value for clinically relevant endpoints such as FN, hospitalization, unplanned care, and death. The small sample sizes in each arm may have yielded an inaccurate DSN for peg lgrastim (0.73 days) compared with 1.2 to 1.8 days and a relatively low incidence of Grade 4 neutropenia for peg lgrastim (59.1%) compared with 83-93% in larger studies [14,15]. The DSN estimate drove the trial size determination. Seemingly discordant results on the secondary and exploratory endpoints of febrile neutropenia and FN clinical sequelae may be a result of the relatively small sample size and relatively small numbers for FN in the control arm. Furthermore, absence of blinding could impact subjective measures, such as patientreported bone pain. We plan to mitigate these potential limitations further in the Phase 3 portion.

Conclusions
Plinabulin recently received breakthrough designation for CIN prevention from the US Food and Drug Administration (FDA) and China National Medical Products Administration (NMPA). Combining plinabulin 20 mg/m 2 and peg lgrastim 6 mg ameliorates neutropenia throughout the chemotherapy cycle, has potentially better CIN prevention and safety pro le than either agent alone, and almost eradicated peg lgrastiminduced bone pain. The con rmatory Phase 3 portion of PROTECTIVE-2 will test superiority of plinabulin 40 mg and peg lgrastim 6 mg combination therapy over peg lgrastim 6 mg alone in earlystage breast cancer patients treated with TAC. The Phase 3 trial excludes patients with bone metastases to allow evaluation of bone pain as an endpoint and has CIN endpoints and formally evaluates quality of life, FN, hospitalization, and infectionrelated complications. agreements with the sponsor regarding con dentiality of all patient data. The academic rst author (Douglas W. Blayney), who had full access to the data, wrote an initial draft of the manuscript. All authors were involved in interpreting, writing, and reviewing subsequent drafts of the manuscript and in making the decision to submit the manuscript for publication. All authors read and approved the nal manuscript.

Data Availability
The datasets generated during and/or analyzed during the current study are not publicly available due to participant privacy concerns but can be accessed from the corresponding author on reasonable request.

Code Availability
Not applicable

Ethics approval
This study was performed in line with the principles of the Declaration of Helsinki and the International Council for Harmonisation Good Clinical Practice (GCP) guidelines. The protocol was approved by the relevant Independent Ethics Committee or Institutional Review Board at each site.

Consent to participate
Informed consent was obtained from all individual participants included in the study.  Median ANC Nadir (x10 9 cell/L) 0.46 1.0 CI, con dence interval.
95% CI and two-sided P-values were calculated by using Poisson Regression model.