Monitoring EPR Effect Dynamics during Nanotaxane Treatment with Theranostic Polymeric Micelles

Abstract Cancer nanomedicines rely on the enhanced permeability and retention (EPR) effect for efficient target site accumulation. The EPR effect, however, is highly heterogeneous among different tumor types and cancer patients and its extent is expected to dynamically change during the course of nanochemotherapy. Here the authors set out to longitudinally study the dynamics of the EPR effect upon single‐ and double‐dose nanotherapy with fluorophore‐labeled and paclitaxel‐loaded polymeric micelles. Using computed tomography‐fluorescence molecular tomography imaging, it is shown that the extent of nanomedicine tumor accumulation is predictive for therapy outcome. It is also shown that the interindividual heterogeneity in EPR‐based tumor accumulation significantly increases during treatment, especially for more efficient double‐dose nanotaxane therapy. Furthermore, for double‐dose micelle therapy, tumor accumulation significantly increased over time, from 7% injected dose per gram (ID g–1) upon the first administration to 15% ID g–1 upon the fifth administration, contributing to more efficient inhibition of tumor growth. These findings shed light on the dynamics of the EPR effect during nanomedicine treatment and they exemplify the importance of using imaging in nanomedicine treatment prediction and clinical translation.


SUPPORTING FIGURES
: GPC chromatograms of free Cy7 and Cy7-labeled mPEG-b-p(HPMAm-Bz) after purification by dialysis, confirming stable dye conjugation. Significantly reduced Ki67 signal was found for both micelles groups as compared to the control group, but not for free drug. In addition, tumors in mice treated with double-dose micelles showed significantly less proliferation than free PTX. Data are presented as average ± SD (n = 4 per group). Statistical differences were analyzed via unpaired, parametric oneway ANOVA with Tukey correction. P values: * < 0.05, *** < 0.001. C: Representative fluorescence microscopy images showing Caspase-3 + (Casp-3, red) apoptotic cells in control, PTX-15, M-PTX-15 and M-PTX-30 groups. Scale bar = 60 µm. D: Treatment with micellar PTX (but not free PTX) induced a significantly higher degree of apoptosis as compared to the control group. Data are presented as average ± SD (n = 4 per group). Statistical differences were analyzed via unpaired, non-parametric one-way ANOVA with Dunn's correction. P values: * < 0.05, ** < 0.01.  Comparable healthy tissues levels were observed for all healthy organs apart from spleen on day 17 (which was also higher for M-PTX-30). Data are presented as average ± SD (n = 5 per group). Statistical significance between different time points within each organ was analyzed via two-way ANOVA with Tukey correction. P values: * < 0.05, ** < 0.01, *** < 0.001.

Figure S5: Macroscopic ex vivo analysis of micelle accumulation and distribution. A-B:
Quantification of ex vivo fluorescence units (f.u.) of Cy7-labeled PTX-micelles in tumors showed significantly higher accumulation of the double-dosed micelles. C-D: Fluorescence was also quantified in 100 µm-thick tumor sections. Higher accumulation and higher intertumor variability was observed in the M-PTX-30 group as compared to the M-PTX-15 group. E-F: Cy7-labeled micelles showed a more favorable tumor distribution pattern upon treatment with the double dose, as evidenced by the relatively deeper penetration of the micelles into the core of the tumor. Data are presented as average ± SD (n = 5 per group). Unpaired, non-parametric two-tailed t-tests were performed to compare the ex vivo fluorescence images. * < 0.05.

Figure S6: Microscopic ex vivo analysis of micelle accumulation and distribution. A:
Representative whole-tumor fluorescence microscopy image showing perfused blood vessels (lectin) and micelles in tumor tissue. As in Figure S5, tumors were divided into two regions, considering the inner 75% as the core and the outer 25% as the periphery. B-C: Representative zoom-in images of tumor core and periphery for the single-and double-dose treatment groups, exemplifying the accumulation and distribution patterns of the micelles. D-E: Microscopically, micelles in the M-PTX-15 group accumulated preferentially in the periphery. Conversely, micelles in the M-PTX-30 group displayed a better distribution, with relatively higher levels in the tumor core, but they also presented with higher variability. F-G: For both treatment groups, significantly higher amounts of micelles were found to have extravasated out of the blood vessels, with approx. 80% of the fluorescence signals detected outside of the vessel lumen. Data are presented as average ± SD(n = 4 per group). Unpaired, parametric, two-tailed t-tests were performed to compare the micelles in the core versus periphery of both groups as well as the amount of intravascular vs. extravasated micelles. P values: **** < 0.0001. . Panels C, D, F: unpaired, parametric one-way ANOVA with Tukey correction was performed to analyze statistical differences between the groups. Panel E: statistical differences were analyzed via unpaired, non-parametric one-way ANOVA with Dunn's correction. P values: * < 0.05, ** < 0.01, *** < 0.001. G-J: In terms of differences in tumor vasculature features in core vs. periphery, we observe that particularly for the double-dose micelles treated group, vessel perfusion, functionality and maturity were increased in the periphery vs. in the core of tumors. Data are presented as average ± SD (n = 4 per core, n = 4 per periphery). Statistical differences were analyzed via two-way ANOVA and Bonferroni correction. P values: * < 0.05, ** < 0.01.

Microscopy
Microscopy images of tumor sections were acquired using an AxioImager M2 microscopy system equipped with an AxioCamMRm Rev.3 camera (Carl Zeiss AG, Jena, Germany). A magnification of 10x was used for acquiring overview images, 20x was used for acquiring images for quatification. The overview images were spatially divided into "core" (the inner 75% of the lesion) and "periphery" (the outer 25%). We always acquired 4 representative images per region and per tumor for quantification. Images in the bright-field channel were acquired for the H&E-stained tumor sections. Two-photon laser scanning microscopy (TPLSM) images to assess collagen content (via second harmonic generation (SHG) imaging) were obtained based on 100 µm-thick, unstained and water-immersed cryosections.

Microscopy image analysis
The area fraction (AF %) was quantified for the fluorescence signals associated with the Cy7-labeled micelles, Ki67, Cap-3, DAPI and Collagen I. This was done using the AxioVision TPLSM images were analyzed using the LAS X 3D software (Leica Microsystems, Germany). All microscopy values are presented as average ± standard deviation (n = 4 per group), and all according statistical analyses were performed using GraphPadPrims 9.

In vivo CT-FMT analysis micelle biodistribution
3D CT-FMT reconstructions were segmented and the fluorescence distribution of Cy7labeled micelles was quantified, converted into percentage of the injected dose (%ID) and normalized to comparable tissue volumes (in cm 3 ) for tumor, liver, spleen, kidneys, lungs and heart for accumulation on days 3, 7, 10, 14 and 17 after the initiation of micelle-based nanotaxane therapy. The values for % ID cm -3 were averaged per group (n = 5) and plotted as average ± standard deviation for each organ and each time point. Statistically significant differences in the % ID cm -3 values for organs at different time points were analyzed via twoway ANOVA with Tukey correction, using GraphPadPrism 9 software.

Ex vivo FRI analysis of micelle tumor targeting
2D fluorescence reflectance images (FRI) at an excitation wavelength of 750 nm were acquired for harvested whole tumors as well as for 100 µm-thick tumor tissue cryosections using the VisEn 2500 FMT device (PerkinElmer, USA). Fluorescence units (f.u.) of whole tumors and tumor sections were obtained from the segmentation of the images using Imalytics Preclinical 2.0. The segmented tumor sections were eroded based on the surface areas of the whole tumor tissue slices, in order to compare micelle localization in the tumor periphery (i.e. the outer 25% surface area) to micelle accumulation in the tumor core (i.e. the inner 75% surface area). Data were presented as average ± standard deviation (n = 4 per group). To assess statistical significance, unpaired, parametric, two-tailed t-tests were performed using the GraphPadPrism 9 software.

Ex vivo micro-CT analysis
To 3D visualize and quantify perfused tumor blood vessels upon taxane-based (nano)therapy, 1 mouse per group was intracardially perfused with MicrofilMV-112 (Flow Tech, USA) prior to sacrificing the mice. After Microfil polymerization, tumors were resected and fixed in 10% formalin/PBS. The SkyScan 1272 micro-CT system (SkyScan, Belgium) was employed to acquire high-resolution ex vivo scans of the fixed tumor tissue specimens.
Micro-CT scanning layers were subsequently reconstructed into 3D images via employing Feldkamp-type filtered back-projection. The vessels were segmented in a stack of bidimensional images using the Imalytics Preclinical 2.0 software. The extent of functional tumor vascularization was calculated as the percentage of the CT-segmented vessel volume versus the respective CT-segmented total tumor volume.