The heterogeneous intra-tumoral accumulation of liposomes has been linked to an abnormal tumor microenvironment. Herein methods are presented to measure tumor microcirculation by perfusion imaging and elevated interstitial fluid pressure (IFP) using an image-guided robotic system. Measurements are compared to the intra-tumoral accumulation of liposomes, determined using volumetric micro-CT imaging.
The heterogeneous intra-tumoral accumulation of liposomes is a critical determinant of their efficacy. Both the chaotic tumor microcirculation and elevated IFP are linked to the heterogeneous intra-tumoral distribution of nanotechnology-based drug delivery systems such as liposomes. In the present study, the relationship between tumor microcirculation, elevated IFP, and accumulation of nanoparticles was investigated through in vivo experimentation. This was accomplished by evaluation of the tumor microcirculation using dynamic contrast enhanced computed tomography (DCE-CT) and measurement of tumor IFP using a novel image-guided robotic needle placement system connected to the micro-CT scanner. The intra-tumoral accumulation of liposomes was determined by CT image-based assessment of a nanoparticle liposomal formulation that stably encapsulate the contrast agent iohexol (CT-liposomes). CT imaging allowed for co-localization of the spatial distribution of tumor hemodynamics, IFP and CT-liposome accumulation in an individual subcutaneous xenograft mouse model of breast cancer. Measurements led to the discovery that perfusion and plasma volume fraction are strong mediators of the intra-tumoral distribution of liposomes. Furthermore, the results suggest that IFP plays an indirect role in mediating liposome distribution through modulating blood flow.
Measuring the intra-tumoral accumulation of nanoparticle drug delivery systems may provide an important tool to determine if an adequate concentration of cytotoxic drug has been achieved within the tumor. The development of "image-able" liposomal systems allows for non-invasive and quantitative in vivo detection of the drug delivery vehicle using imaging modalities such as positron emission tomography (PET)1, optical fluorescence2, and computed tomography (CT)3,4 and magnetic resonance imaging (MRI)5. Imaging has been used to determine the pharmacokinetics and biodistribution of liposome delivery systems and to reveal the extent of inter-subject and intra-tumoral heterogeneity in nanoparticle accumulation6,7. However, imaging of nanoparticles alone does not identify the biological barriers that have contributed to their poor accumulation and distribution. This knowledge is paramount to the rational development of more efficacious formulations, and strategies to improve intra-tumoral accumulation8. It has been demonstrated that therapeutic strategies can be applied to modulate specific biological barriers resulting in improved nanoparticle transport9. Additionally, nanoparticle formulations have been developed to specifically overcome specific biological transport barrier10. In both scenarios, measurements of biological barriers could be used to guide the use of an appropriate nanoparticle drug delivery strategy.
Tumor microcirculation and elevated IFP are believed to be two key determinants of the intra-tumoral accumulation of nanoparticles, such as liposomes, in solid tumors9,11. However, other barriers that to contribute to poor liposome accumulation include a dense extracellular matrix, impermeable vasculature, and solid tissue pressure12. These barriers are related in a spatio-temporal manner, with abnormal blood flow and elevated interstitial fluid pressure being two important factors driving the initial delivery and extravasation of nanoparticles. As previously discussed, establishing the relationship between the tumor microcirculation, elevated IFP, and the intra-tumoral accumulation of liposomes is imperative for proper interpretation of liposome imaging data. Herein quantitative methods to measure the relationship between the tumor microcirculation, elevated IFP, and nanoparticle accumulation in a solid tumor are presented. This is accomplished by performing co-localized measurements of the intra-tumoral distribution of a CT liposome contrast agent using volumetric CT imaging, tumor microcirculation using dynamic contrast enhanced computed tomography imaging, and tumor IFP using an image-guided robotic needle positioning system, termed the CT-IFP robot13.
All in vivo experiments were performed under a protocol approved by the University Health Network Institutional Animal Care and Use Committee.
1. Animal Model
2. CT-liposome Preparation and Characterization
3. CT Imaging of Tumor Microcirculation and CT-liposome Distribution
NOTE: Follow the manufacturer's instructions for performing a volumetric scan if different software version or equipment is used.
4. Spatial Measurements of Tumor Interstitial Fluid Pressure
The aforementioned protocol should yield CT-liposomes with an encapsulated concentration of iohexol, mean liposome diameter, and zeta potential of 55 mg ml-1, 91.8 ± 0.3 nm and -45.5 ± 2.5 mV, respectively. Figure 1a includes representative DCE-CT imaging results, yielding a time series of volumetric data that show the temporal changes in intra-tumoral accumulation of iohexol. Selecting a ROI within the tumor yields a TIC that can be quantified using tracer kinetic modeling methods to obtain estimates of perfusion, vascular permeability, plasma volume fraction, and interstitial volume fraction (Figure 1b). In this study, a two-compartment tracer kinetic model was used and fit to the measured TIC using a non-linear curve fitting routine implemented in Matlab14. Segmenting the tumor volume into multiple regions of interest of equal size allows for quantification of the spatial distribution of haemodynamic parameters within the tumor volume (Figure 1c). Segmentation can be performed either manually, which is time consuming and difficult, or automatically as performed here using an algorithm that divides the tumor in multiple equal sized ROIs using a spherical coordinate system. The DCE-CT methods provide quantitative estimates of the spatial distribution of perfusion, vascular permeability, plasma volume fraction, and interstitial volume fraction. These parameters were observed to be spatially heterogeneous with higher levels of perfusion, plasma and interstitial volume fractions along the periphery compared to the central tumor volume.
The volumetric CT imaging method reveals the biodistribution and intra-tumoral distribution of CT-liposomes. Figure 2a shows the biodistribution of CT-liposomes at 48 hr post-injection. The agent is still circulating in the vascular system, with substantial uptake observed in the spleen and liver. The intra-tumoral accumulation of CT-liposomes was observed to be heterogeneous, with predominantly peripheral accumulation compared to the center, as denoted by the bright regions within the tumor volume (Figure 2b).
Volumetric CT imaging can be used to track the location of IFP measurements made using the CT-IFP robot setup. Figure 3a shows the placement of the IFP needle within the tumor volume as imaged using high-resolution micro-CT. The needle can clearly be identified within the tumor volume allowing for spatial localization of the IFP measurements within the tumor volume (Figure 3b). It is possible to generate a spatial map of IFP throughout the tumor by performing multiple IFP measurements within the tumor volume. The spatial IFP can then be correlated with the corresponding measurements of tumor microcirculation and CT-liposome accumulation.
Volumetric CT imaging allows for a common frame of reference making it possible to co-localize measurements of haemodynamics, IFP, and CT-liposome accumulation. Figure 4 gives an example of spatially co-localized measurements of CT-liposome accumulation, IFP, perfusion, vascular permeability, plasma volume fraction, and interstitial volume fraction. It was observed that perfusion and the plasma volume fraction was significantly correlated with the intra-tumoral accumulation of CT-liposomes in subcutaneous MDA-MB-231 tumors. Furthermore, the radial distribution of IFP correlated with haemodynamic measurements. These results suggest a complex spatio-temporal relationship exists between the tumor microcirculation, IFP and the intra-tumoral accumulation of liposomes14.
Figure 1: DCE-CT Imaging of the Tumor Microcirculation. (a) A representative series of temporal CT images collected within the tumor volume, depicting the contrast agent kinetics as a function of time. The red contour represents a ROI where the time intensity curve (TIC) is measured. (b) The TIC is fit using a two-compartment tracer kinetic model to yield quantitative estimates of haemodynamic parameters within the ROI. (c) Representative spatial distribution of quantitative haemodynamic parameters in the tumor. Please click here to view a larger version of this figure.
Figure 2: Volumetric CT-Imaging of Liposome Accumulation. (a) A representative 3D volume-rendered image demonstrating the biodistribution of CT-liposomes. (b) Representative axial, coronal, and sagittal slices taken through the center of the tumor showing the intra-tumoral accumulation of CT-liposomes at 48 hr post-injection. Please click here to view a larger version of this figure.
Figure 3: Image Guided IFP Measurements. (a) A representative 3D volume-rendered image of the CT-IFP robot system (green) post-needle insertion into a subcutaneous tumor at 48 hr post-injection of CT-liposomes (orange). (b) A representative CT image of the post-needle insertion. Please click here to view a larger version of this figure.
Figure 4: Co-localized Measurements of Tumor Microcirculation, IFP, and CT-liposome accumulation. Panel showing a representative spatial co-localization of CT-liposome accumulation taken 48 hr post-injection, IFP, perfusion, vascular permeability, plasma volume fraction and interstitial volume fraction. Re-print with permission from14. Please click here to view a larger version of this figure.
The methods for image-based measurement presented herein enable determination of the spatial distribution of tumor microcirculation properties, IFP, and CT-liposome accumulation. Previous attempts to relate these properties have relied on performing bulk measurements across multiple tumor-bearing animals and therefore lack the sensitivity to elucidate mechanisms responsible for heterogeneity in intra-tumoral accumulation that has commonly been observed for nano-sized drug delivery systems15. DCE-CT provides a tool to measure the intra-tumoral variations in properties of the tumor microcirculation, volumetric CT provides an accurate depiction of CT-liposome deposition kinetics, and the CT-IFP robot system provides a tool to perform spatial mapping of IFP in the same animal. Furthermore, DCE-CT imaging is a clinically approved method for measuring tumor hemodynamics in the clinical setting, making the findings of this study potentially clinically translatable.
Given the complexity of the measurements, there are several critical factors to ensure the collection of robust data sets. DCE-CT based quantification of the tumor microcirculation is arguably the most difficult to ensure accurate estimates of tumor haemodynamics. It requires obtaining TICs with high signal to noise ratios (SNR) and employing a robust fitting algorithm to quantify the TICs16,17. Visual inspection of TICs can be used to remove low SNR data from the analysis. Furthermore, if care is not taken then the fitting of high SNR TICs may also lead to erroneous estimates of tumor perfusion, vascular permeability, plasma volume fraction, and interstitial volume fraction16. In order to maximize quantification accuracy a strategy was employed to obtain model independent estimates of the plasma and interstitial volume fractions, which are subsequently used as fixed parameters during the model fit of measured TICs. This method ensures robust estimates of tumor perfusion and vascular permeability are obtained15.
Robust analysis of the intra-tumoral distribution of CT-liposomes requires performing volumetric CT imaging after sufficient accumulation of the agent. From previous studies, the peak tumor accumulation of CT-liposomes occurs between 48 to 72 hr in mouse xenografts3,15. Furthermore a linear relationship exists between CT-liposome concentration and contrast enhancement in CT imaging allowing for simple quantification of the variations in intra-tumoral accumulation of CT-liposomes15.
Accurate measurements of IFP using the needle-based method require good fluid communication between the catheter and the tissue. Furthermore, it is important to only use tumors that have high central tumor IFP (>5 to 10 mmHg), otherwise there will be minimal spatial variations in IFP. Spatial measurements of IFP using the CT-IFP robot systemcan be challenging due to tissue motion caused by needle insertion. Imaging pre- and post-needle placement is crucial for accurately identifying the needle placement; however, it can be difficult to relate the position between subsequent needle placements due to tissue warping between measurements. It was found that randomly choosing needle positions results in significant tissue deformation during needle insertion. As a result, this method provided the least accurate spatial mapping of IFP. Conversely, performing measurements along a linear track across the tumor volume and inserting the needle tangential to the track can improve the spatial accuracy of IFP measurements. Inserting the needle tangential to the track minimizes the effects of tissue deformation along the measurement track direction.
This study demonstrated the ability to measure the spatial distribution of tumor microcirculation, IFP and CT-liposome accumulation in an individual tumor. After mastering these techniques, it is then possible to perform these measurements independently or together to characterize the tumor microenvironment and its effects on drug delivery. Using these methods in the MDA-MB-231 breast xenograft model revealed that perfusion and plasma volume fraction are strong mediators of the intra-tumoral distribution of liposomes14. There was not found to be a strong relationship between IFP and liposome distribution. However, IFP was strongly correlated to measurements of tumor perfusion, suggesting that IFP may play an indirect role in mediating liposome distribution through modulation of blood flow.
The authors have nothing to disclose.
The authors would like to thank Dr. Javed Mahmood for assistance with culturing MDA-MB-231 cells and implanting the MDA-MB-231 xenografts, Linyu Fan for preparing the CT-liposomes. Shawn Stapleton is grateful for funding from the Natural Sciences and Engineering Research Postgraduate Scholarships Program and the Terry Fox Foundation Strategic Initiative for Excellence in Radiation Research for the 21st Century (EIRR21) at CIHR. This study was supported by grants from the Terry Fox New Frontiers Program (020005) and the Canadian Institutes of Health Research (102569).
MDA-MB-231 metastatic breast adenocarcinoma tumor cells | ATCC | HTB-26 | |
Dulbecco's Modified Eagle Medium (DMEM) | Life Technologies | 11965-092 | |
Fetal Bovine Serum (FBS) | Sigma-Aldrich | F1051 | |
HyClone Penicillin-Streptomycin 100x Solution | GE Healthcare Life Sciences | SV30010 | |
Trypsin-EDTA (0.05%), phenol red | ThermoFisher Scientific | 25300-054 | |
1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) | Avanti Lipids Inc., USA | 850355P | |
Cholesterol (CH) | Avanti Lipids Inc., USA | 700000P | |
1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-poly(ethylene glycol) 2000 (DSPE-PEG2000) | Avanti Lipids Inc., USA | 880128P | |
Omnipaque (Iohexol) 300 mg of iodine/mL | GE Healthcare, CA | ||
80 nm pore size Track-Etch polycarbonate membranes | Whatman Inc., USA | ||
200 nm pore size Track-Etch polycarbonate membranes | Whatman Inc., USA | ||
10 mL Lipex Extruder | Nothern Lipids Inc, CA | ||
Dialysis Bag Molecular Weight Cut Off (MWCO) of 8 kDa | Spectrum Labs, USA | ||
750,000 Nomical Molecular Weight Cut Off (NMWC) Tangential flow column | MidGee ultrafiltration cartridge, GE Healthcare, CA | ||
Peristaltic pump | Watson Marlow Inc., USA | ||
UV spectrometer | Helios γ, Spectronic Unicam, USA | ||
90Plus particle size analyzer | Brookhaven, Holtsville, USA | ||
eXplore Locus Ultra micro-CT system | GE Healthcare, CA | Manipulated using CT-Console Software | |
AxRecon GPU-based Reconstruction | Acceleware Corp. CA | ||
27G Catheter SURFLO Winged Infusion Set | Terumo Medical Products, USA | SV*27EL | |
PE20 polyethylyne tubing | Becton Dickinson, USA | 427406 | |
Pen tip 25G × 3.5′′ Whitacre spinal needle | Becton Dickinson, USA | 405140 | IFP needle |
P23XL pressure transducer | Harvard Apparatus, CA | P23XL | |
PowerLab 4/35, Bridge Amp, with LabChart Pro 7.0 | ADInstruments Pty Ltd., USA | PL3504, FE221 | IFP acquisition system and acquisition software |
CT-Sabre Small Animall Intervention system (CT-IFP Robot) | Parallax Innovations, CA | Manipulated using CT-IFP robot Control Software | |
CT-IFP robot alignment software | Custom Matlab software | ||
DCE-CT Analysis Software | Custom Matlab software | ||
Matlab 2013b | Mathworks, USA |