Paclitaxel and Carboplatin in Combination with Low-intensity Pulsed Ultrasound for Glioblastoma

Abstract

Purpose: We recently reported on clinical trials for patients with recurrent glioblastoma where low-intensity pulsed ultrasound and microbubbles (LIPU/MB) improved paclitaxel or carboplatin delivery into the brain. Here, we report variable local tumor control with paclitaxel at the maximal/target dose in our phase I trial (NCT04528680). To address this, we investigated the combination of paclitaxel with carboplatin in preclinical glioma models.

Experimental design: We performed MRI-based analysis to evaluate disease control in patients from our trial. We studied the cytotoxicity of paclitaxel and carboplatin against 11 human glioma lines as monotherapy and in combination at concentrations derived from human intraoperative studies. Synergy was assessed with the Loewe model and the survival benefit evaluated in two xenografts. We examined the effects on cell cycle progression, DNA damage, and apoptosis.

Results: Patients treated with paclitaxel and LIPU/MB exhibited variable local tumor control, which correlated with overall survival. We observed limited cross-resistance to paclitaxel and carboplatin in glioma lines, with almost a third of them being exclusively susceptible to one drug. This combination led to susceptibility of 81% of lines and synergy in 55% of them. The combination proved more efficacious in two intracranial xenografts when administered with LIPU/MB, leading to complementary effects on cell cycle arrest.

Conclusions: Combining paclitaxel and carboplatin in gliomas may be more efficacious than monotherapy, as in other cancers, due to synergy and independent susceptibility to each drug. These results form the basis for an ongoing phase II trial (NCT04528680) where we investigate this combination with LIPU/MB.

Figure 1.. Patients with glioblastoma exhibit variable response to paclitaxel.

(A) Coronal T1-MRI with contrast images for patients 110 and 109 from our NCT04528680 clinical trial, acquired at different time points after surgery for resection of the recurrent tumor and implantation of the ultrasound device. A region of interest (ROI) is delineated in green and illustrates the sonicated region. (B) Graphs representing the volume of enhancement within and outside the ROI for the patients shown in (A). (C) Pie chart showing the proportion of patients enrolled in the NCT04528680 who had tumor control while on LIPU/MB and ABX therapy. (D) Graphs showing the change (left) and rate of change (right) in the enhancing volume recorded within the ROI for patients that received the highest ABX dose (260 mg/m²) in the NCT04528680 clinical trial. (E) Kaplan-Meier survival curves showing the progression-free (left) and overall (right) survival of patients that received the highest Abraxane dose (260 mg/m²) in our phase I clinical trial (NCT04528680). Patients were dichotomized into two groups: patients who had tumor growth on treatment (red), and patients whose tumors were controlled (blue). We illustrate the number at risk below the survival curves.

Figure 2.. Glioma cell lines exhibit non-overlapping susceptibility to paclitaxel and carboplatin.

(A) Schematic representation of the average concentration of PTX and CBDCA recorded within the tumor (red), peritumoral sonicated (green) and peritumoral non-sonicated (beige) brain regions of patients on the NCT04528680 and NCT03744026 clinical trials when the drugs were delivered with LIPU/MB. For PTX, we analyzed 40 non-sonicated, 41 sonicated, and 21 intratumoral biopsies. For CBDCA, we analyzed 25 non-sonicated, 23 sonicated, and 20 intratumoral biopsies. (B) Violin plot illustrating relation between the concentrations of PTX achieved within the peritumoral sonicated, peritumoral nonsonicated and tumor regions of patients and the PTX concentrations needed to achieve a 50% inhibition (IC50) of glioma cell lines, as reported in the Genomic of Drug Sensitivity in Cancer Project (GDSC) database. Intratumoral drug concentrations could potentially be inflated by possible drug contamination, which we did not account for. (C) Scatter plot showing the correlation between the Emax of PTX and CBDCA in the 11 cell lines tested (r = −0.13, p = 0.71). (D) Scatter plot showing the correlation between the IC₅₀ values of glioma cell lines for PTX and Cisplatin, as obtained from the GDSC database (r=0.19, p=0.29). (E, F) Pie charts showing the overlap in the proportion of cells categorized as susceptible or resistant to therapy when using the median Emax or IC₅₀. We show that there is little overlap in the cells categorized as resistant to the drugs as monotherapy. (A, Created with BioRender.com)

Figure 3.. Paclitaxel and carboplatin exhibit synergy in vitro and in vivo when delivered with LIPU/MB.

(A) Pie chart showing the proportion of glioma cell lines that showed synergy by the Loewe model. (B, C) 3D Surface plots showing the Loewe score for two of the cell lines that showed the highest synergy (B) and two cell lines that showed the highest antagonism (C) at clinically relevant PTX and CBDCA concentrations. The x and z axes illustrate the concentrations of paclitaxel and carboplatin used, and the y axis illustrates the Loewe synergy/antagonism score. To the right of the 3D surface plots, we illustrate on bar plots the viability of cells treated with PTX, CBDCA, or the combination of both drugs at the concentrations that showed the highest synergy (B) or antagonism (C). (D, E) Schematic representations of the treatment regimen followed for MES83 bearing mice (D) and GBM6 bearing mice (E). Below the schematics, we illustrate the Kaplan-Meier survival curves for these mice treated with LIPU/MB along with Abraxane monotherapy, carboplatin monotherapy, or a combination of both drugs. The tables illustrate the median survival for these mice. *, p<0.05; **, p<0.01; ***, p<0.001; ****, p<0.0001).

Figure 4.. Paclitaxel and carboplatin combination increases DNA damage and potentiates cell cycle arrest and apoptosis.

(A) Results of flow cytometric analysis of synchronized GBM6 and MES83 cells that received PTX, CBDCA, or a combination of both drugs for 24 hours. On the right, we show the proportion of cells arrested at each of the SubG1, G0/G1, S, and G2/M phases. (B) Representative images of γH2AX immunofluorescence staining of MES83 cells treated with PTX and CBDCA as monotherapy or combination therapy and acquired at a 20x magnification (top). The bar graph on the right shows the quantification of γH2AX positive cells per 1000 cells and demonstrates that combination therapy results in DNA damage in a larger number of cells compared to monotherapy. Below, we show 3D reconstructions of z-stack images acquired at 60x magnification on a confocal microscope, illustrating the extent of DNA damage (as demonstrated by the γH2AX foci) within γH2AX positive cells. The bar graph on the right shows the number of γH2AX foci within γH2AX positive cells across the treatment groups, and demonstrates that combination therapy enhances DNA damage within γH2AX positive cells compared to monotherapy. (C) Coronal brain sections of GBM6-bearing mice treated with LIPU/MB and paclitaxel (n = 4), CBDCA (n = 3), or a combination of both drugs (n = 3), as well as control mice (n = 3). Sections are stained for γH2AX and shown at 1.5x (top) and 30x magnification (bottom). On the right, the bar graph illustrates the density of cells staining positive for γH2AX within the tumor (D) Western Blot with densitometry analysis showing the expression of cleaved caspase 3 at 48 hours of treatment with monotherapy and combination therapy in two distinct cell lines (MES83 and GBM6).