Evans blue nanocarriers visually demarcate margins of invasive gliomas
Aggressive surgical resection is the primary therapy for glioma. However, aggressive resection may compromise functional healthy brain tissue. Currently, there are no objective cues for surgeons to distinguish healthy tissue from tumor and determine tumor borders; surgeons skillfully rely on subjective means such as tactile feedback. This often results in incomplete resection and recurrence. The objective of the present study was to design, develop, and evaluate, in vitro and in vivo, a nanoencapsulated visible dye for intraoperative, visual delineation of tumor margins in an invasive tumor model. Liposomal nanocarriers containing Evans blue dye (nano-EB) were developed, characterized, and tested for safety in vitro and in vivo. 3RT1RT2A glioma cells were implanted into brains of Fischer 344 rats. Nano-EB or EB solution was injected via tail vein into tumor-bearing animals. To assess tumor staining, tissue samples were analyzed visibly and using fluorescence microscopy. Area, perimeter ratios, and Manders overlap coefficients were calculated to quantify extent of staining. Nano-EB clearly marked tumor margins in the invasive tumor model. Area ratio of nano-EB staining to tumor was 0.89 ± 0.05, perimeter ratio was 0.94 ± 0.04, Manders R was 0.51 ± 0.08, and M1 was 0.97 ± 0.06. Microscopic tumor border inspection under high magnification verified that nano-EB did not stain healthy tissue. Nano-EB clearly aids in distinguishing tumor tissue from healthy tissue in an invasive tumor model, while injection of unencapsulated EB results in false identification of healthy tissue as tumor due to diffusion of dye from the tumor into healthy tissue.
KeywordsNanocarriersLiposomesTumor marginTumor borderIntraoperative MRITumor resection
Currently, aggressive surgical resection is used to extend survival for patients suffering from high-grade glioma [1–5]. Even with the currently available treatment options, prognosis is poor, resulting in a mean survival time as low as 11 to 14 months . However, patient life expectancy increases with an increase in extent of resection . Concomitantly, minimizing removal of healthy neural tissue is critical to ensure quality of life. As a consequence, total and complete surgical resection of tumors is rare, and a vast majority of brain tumors recur, often in close proximity to the resection site [7, 8].
A major impediment to achieving greater accuracy in tumor resection stems from the absence of clear visual cues demarcating the tumor margin. Therefore, neurosurgeons currently rely on subjective criteria, such as tactile feedback, to assess whether a particular tissue is tumor or healthy. Presurgical magnetic resonance imaging (MRI) or computed tomography (CT) can provide relevant information to plan the surgical procedure, but does not provide real-time information during resection. Intraoperative histology can provide surgeons with information about the margins, but this method is time consuming. Further, by this methodology, assessment of the nature of the tissue (healthy vs. tumor) can be made only after its removal from the patient. Intraoperative CT or MRI are possible alternatives, providing non-invasive information about the extent of residual tumor, but are cost limited and special training is often needed, making these techniques available only in more advanced surgical environments. Use of fluorescent probes that stain tumors require special equipment for excitation and detection, as well as low light conditions, neither of which are optimal in a surgical environment [9–13]. Thus, a method to demarcate tumor margins without a significant change to surgical conditions is desired.
Ozawa et al. and Orringer et al. have suggested the use of a small molecule visible dye to address this challenge [14, 15]. This is an elegantly simple solution, however, the demarcation is temporally transient and the margin blurs, necessitating multiple dye injections during the course of the surgical procedure. More importantly, the technique stains healthy tissue in invasive tumors, where a strong need for demarcation of tumor exists. A non-invasive tumor such as the 9 L glioma used by Ozawa and others merely “pushes” healthy tissue out of the way as it grows so the tumor margin is already fairly distinguished; the tumor extracellular space and vasculature do not interact with the healthy tissue . Clearly, this is not the case with most invasive tumors; diffusion through extracellular space would cause staining of healthy tissue, as evident from injections of Evans blue (EB) dye in sterile saline. Here we test the hypothesis that a nanoencapsulated dye, due to size-dependent diffusion limitations, would clearly demarcate tumor margins after extravasation via the enhanced permeation and retention effect, even in an invasive cancer model.
In previous studies, our lab has shown that a nanoencapsulated X-ray contrast agent will passively but selectively accumulate at the periphery of the tumor . Due to their relatively large size (∼100–200 nm), these nanoparticles mark the margin stably without diffusing away, moving no more than 50 μm from the vessel from which they exited . Further, liposomal nanocarriers are FDA approved for use in cancer therapy for delivery of chemotherapeutics, such as Doxil®, yielding predictable circulation times and biodistribution, as well as reducing systemic toxicity [19–22]. Here, we have developed a liposomally nanoencapsulated Evans Blue dye (nano-EB) to provide accurate visual cues for the surgeon to intraoperatively delineate the tumor margins without the need for MRI, fluorescence, or other equipment. Demarcation of the tumor margin will allow more accurate and complete removal of tumor tissue, conceivably leading to extended survival times for patients.
Materials and methods
Choosing a dye using perceived luminance
Dyes were evaluated for perceived luminance, which estimates the darkness of a color. Dyes were dissolved in phosphate-buffered saline (PBS) at the same concentration (m/v), dilute enough for all dyes to remain translucent when contained in a 1.5-mL microcentrifuge tube. An image was taken of all the dyes together to ensure consistent lighting conditions. RGB values were measured in three places for each dye using the color picker tool in Photoshop (Adobe, San Jose, CA) and averaged. Perceived luminance was calculated using the following equation: 0.299*R + 0.587*G + 0.114*B .
Preparation of liposomal nanocarriers
Liposomal nanocarriers were prepared using a mixture of 1, 2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC; Genzyme, Cambridge, MA), cholesterol (Sigma-Aldrich, St Louis, MO), and N-(carbonyl-methoxypolyethyleneglycol-2000)-1, 2-distearoyl-sn-glycero-3-phosphoethanolamine (MPEG-2000-DSPE) (Genzyme) in a 55:40:5 molar ratio dissolved in ethanol. Nano-EB particles were made by hydrating the lipid mixture using 75 mg/ml of Evans blue (Alfa Aesar, Ward Hill, MA) in PBS while blank liposomes were made by hydrating with PBS alone. The solution was stirred at 60 °C for 120 min then dialyzed overnight against iso-osmolar phosphate-buffered saline to remove ethanol. The liposomes were resized by placing the suspension in a Branson 1510 bath sonicator (Branson Ultrasonics, Danbury, CT) for 120 min at 60 °C. Nano-EB particles were then run through a Sepharose CL-4B (GE Healthcare, Pittsburgh, PA) size exclusion column chromatography to remove excess, unencapsulated dye. The filtrate was run through a 500 kD MWCO MicroKros diafiltration cartridge (Spectrum Laboratories, Rancho Dominguez, CA) to remove volume gained during chromatography.
Liposomal nanocarrier analysis and characterization
Liposomal nanocarriers were analyzed to obtain size distribution and lipid to dye ratio, as well as for stability in conditions similar to those encountered in vivo. Size distribution of the liposomal nanocarriers was determined using a Zetasizer Nano ZS90 dynamic light scattering particle sizer (Malvern Instruments Ltd, Worcestershire, UK). By using the same instrumentation, the zeta potential of the nano-EB was assessed. The particles were suspended in PBS and three estimations were made. Mean values ± SD were calculated from the values obtained.
For morphological characterization of nano-EB, the liposomes were frozen at −20 °C and then freeze dried using a Lacbconoco FreeZone® freeze dryer. The particles were then processed for scanning electron microscopy (SEM) and images obtained using a Zeiss Ultra-60 Field Emission Scanning Electron Microscope (FE-SEM).
Lipid to dye ratio is presented as amount of dye per unit of lipid (mg/mg). First, the amount of encapsulated dye in the liposome solution was determined by measurement of absorbance of EB. Liposomes were lysed in 0.8 % Triton X100 to release their contents and absorbance was measured at 610 nm using a Synergy HT Plate Reader (BIO-TEK, Winooski, VT) and compared to a standard curve.
Final lipid concentration in the liposomal nanocarrier suspension was obtained by quantifying the fluorescence of a reporter molecule included in the original liposome mixture. β-DPH-HPC (Invitrogen, Carlsbad, CA) was included in the liposome mixture at a concentration of 0.01 mol%. After liposome preparation was complete, lipid and dye in a 100-μL sample were separated via Folch extraction . Dried lipid extract was resuspended in ethanol and fluorescence was measured in a Synergy HT Plate Reader (excitation 360 nm; emission 460 nm). A standard curve of known concentrations of β-DPH-HPC was used to calculate the amount of lipid present in the mixture.
3RT1RT2A (rat glioma cell line, stably expressing green fluorescent protein, GFP) was generously donated by Helen L. Fillmore (Virginia Commonwealth University, Richmond, VA). Cells were cultured in Dulbecco’s Modification of Eagle’s Medium with 4.5 g/L glucose without L-glutamine and sodium pyruvate (Mediatech, Manassas, VA), supplemented with 10 % FBS, 1 % penicillin/streptomycin (Mediatech), 1 % non-essential amino acids (Mediatech), 1 % L-glutamine (HyClone Thermo Scientific, Logan, UT), and 1 mg/ml G418 (Gemini Bio-Products) as selective pressure for the GFP marker. Primary cortical rat astrocytes were obtained and cultured as described previously .
In vitro cell viability
Primary astrocytes, representing normal brain tissue, and 3RT1RT2A cells, representing aggressive glioblastoma, were used to assess viability of cells treated with nano-EB in vivo. Astrocytes were plated at 10,000 cells per well, while 3RT1RT2A cells were plated at 25,000 cells/well, in tissue culture-treated 24-well plates. Cells were treated for 24 h with serum-free media containing either nano-EB, unencapsulated EB in saline, blank liposomes, or sterile saline as a control. EB dose was 0.2 mg/ml in both nano-EB and unencapsulated EB solutions; the dose was chosen under the assumption that a maximum of 2 % of the liposomes delivered to an animal would reach the tumor based on previous work in the lab with similar liposomal nanocarriers . The lipid dose was equal in the two liposomal groups. After 24 h, the treatments were removed and fresh appropriate standard serum containing cell culture medium was added. Cell viability counts were taken at 0, 24, 48, and 72 h using the CCK-8 counting kit (Dojindo, Rockville, MD) as per the manufacturer’s instructions. Cell counts were normalized to the untreated groups at each time point and reported as percent viable cells. Statistical significance was determined using a two-way ANOVA with Tukey’s post test using GraphPad Prism (Graphpad, La Jolla, CA). Data is presented as average ± standard deviation.
Ethical use of animals
All animal protocols were approved by the Georgia Institute of Technology Institutional Animal Care and Use Committee. Fisher 344 male rats 8–10 weeks of age were purchased from Harlan Laboratories. Animals were housed in a temperature-controlled environment with a 12-h light/dark cycle, and were provided food and water for the entire duration of their housing.
In vivo toxicity testing
Non-tumor-bearing male Fisher 344 rats were injected via tail vein with sterile solutions of nano-EB, blank liposomes, or unencapsulated EB dye dissolved in saline. To ensure safety and sterility, all solutions were analyzed for correct pH and osmolarity, as well as tested for endotoxins (Lonza, Walkersville, MD), then filtered through a 0.2-μm sterile filter. Dye dose was 25 mg/kg in the two Evans blue formulations and lipid dose was kept equal in the two liposomal treatment groups. Total body weights and overall condition of the animals were monitored over a period of 4 weeks.
Tumor implantation was adapted from Fillmore and colleagues . Rats were anesthetized initially using 5 % isofluorane and maintained at 2–3 % during the entire procedure. Briefly, animals were placed in a stereotactic frame, an incision was made to expose the skull, and a burr hole was made at the coordinates of 2 mm anterior and 2 mm left lateral from lambda. 3RT1RT2A GFP expressing rat glioma cells were slowly injected 3 mm below the surface of the skull at this position over the course of 3 min (200,000 cells in 10 μL Liebovitz L-15 medium (Invitrogen Life Technologies/GIBCO, Carlsbad, CA)). Tumors were present in all animals that were injected in this way. Tumor-implanted animals typically died by day 12–14; for this reason, animal experiments were constrained to 10 days.
Tumor staining in vivo
On day 8 of tumor growth, animals were injected via tail vein with sterile, non-pyrogenic solutions of 40 mg/kg EB, either as nano-EB or EB in sterile saline solution. Animals were euthanized 48 h later and brains were removed and preserved in a 4 % paraformaldehyde solution.
Visual and histological analysis of tumor staining
Brains of animals were removed and preserved in 4 % paraformaldehyde in PBS. Brains were sliced coronally through the tumor and tumors were visually inspected to confirm presence of blue staining. Tissue was then cryoprotected by submersion in a 30 % w/v solution of sucrose in PBS. Sections of 16 μm thickness were obtained using a Leica CM 300 cryostat (Leica, Bannockburn, IL). Sections were stained with anti-GFP monoclonal antibody conjugated with Alexa Fluor 488 (Invitrogen) to enhance green fluorescence and DAPI (Invitrogen) to indicate nuclei. Microscopic images of the entire tumor (×5 magnification) were obtained and tiled together using an Axiovert 200 M microscope (Zeiss, Thornwood, NY) with a ORCA-ER camera (Hamamatsu Photonics, Bridgewater, NJ), Sedat-ET filter set (Chroma Technology Corp., Bellows Falls, VT), and Volocity acquisition software (PerkinElmer, Walthamm, MA).
Images were imported to ImageJ (NIH) where background subtraction was performed using the included rolling ball background subtraction algorithm. Images were converted to 8-bit and Manders Coefficients plugin was used to find Manders colocalization R and M1 (red/green, indicates how much red staining is also green) . Additionally, the 8-bit images were used to trace the tumor border as well as the EB staining border in the nano-EB-treated samples. The measure function was used to obtain area and perimeter of the tracing. A ratio of red area to green area and red perimeter to green perimeter were calculated. Data is presented as average ± standard deviation.
Higher magnification fluorescence images (×10 magnification) of the tumor border were taken at 0°, 90°, 180°, and 270° around the border of the tumor to visually confirm that there was no healthy tissue staining.
Liposomal nanocarrier fabrication and characterization
Nano-EB particles were developed and characterized, assessing size, drug to lipid ratio, and stability. After processing was complete, the ratio of dye to lipid was 0.08 mg EB to 1 mg lipid. Average hydrodynamic diameter of nanocarriers was 173 nm with a polydispersity index of 0.1. The morphological characteristics of the nano-EB particles as characterized by SEM are depicted in Fig. 1b. Nano-EB particles had a zeta potential of −8.84 ± 0.17. Nano-EB particles were stable at body temperature (37 °C), leaking only 13.3 % (SD = 0.49 %) of their contents over a 48-h period.
In vitro and in vivo toxicity assessment
Tumor staining/visualization in vivo
Border quantification of tumor-bearing animals treated with nano-EB
0.89 ± 0.05
0.94 ± 0.04
0.51 ± 0.08
0.97 ± 0.06
Accurate, complete resection of tumor is essential for successful treatment of gliomas. Currently, during surgery, the surgeon has limited ability to distinguish the margin of the tumor, relying on subjective criteria such as tactile feedback and marginal, often indistinguishable, color differences between healthy and tumor tissue. Intraoperative MRI has been deployed, but it remains an expensive and therefore inaccessible tool in a majority of hospitals. Thus, it would be beneficial to have an objective means to distinguish the tumor margin visually for the neurosurgeons when performing the surgery. Other approaches use fluorescent molecules, but these require extra equipment and training, as well as altered lighting conditions, since the halogen lights typically used in surgical environments often emit radiation in the wavelengths that causes interference with fluorescence detection [9–13]. In response, non-fluorescent dyes have been evaluated, though these compounds are freely injected small molecules and were only examined in non-invasive tumor models to date, which may not adequately represent the use in the clinic [14, 15].
In this study, we demonstrate that freely circulating small molecule compounds, such as Evans blue dye, diffuse out of the tumor and into healthy tissue, complicating their ability to help demarcate the margins between tumor and healthy tissues. On the contrary, intravenous injection of liposomal nanocarriers containing Evans blue dye clearly marked the tumor and its margins, as well as small nodules of invasive cells that had invaded beyond the bulk tumor border. A Manders coefficient of 0.6 or higher is generally accepted as indicating true colocalization in fluorescence microscopy . In this study, the Manders overlap coefficient average is slightly lower than ideal at 0.51 ± 0.08, but this was not unexpected. As Manders coefficients measure overlap, actually, intracellular green fluorescent protein does not overlap with extracellular liposomal Evans blue as little endocytosis of the nanocarriers occurs at 48 h.
Further, more importantly, in the overlap coefficient analysis, the value of M1 (red/green) was 0.97, indicating that nearly all area that was stained red was also stained green. This indicates that nanocarriers were localized specifically to tumor tissue. The area and perimeter ratio analysis further demonstrated that the nanocarriers only stained tumor tissue. Higher magnification border inspection again confirmed that only tumor tissue was stained, with nanocarriers slightly and consistently underestimating the true margin on the order of tens to hundreds of micrometers. This is sufficient, considering surgeons often remove tissue in sections with thicknesses on the order of millimeters, several orders of magnitude greater than the underestimation . It is important to note that healthy tissue was not stained, preventing false positives and removal of healthy tissue by the surgeon which facilitates sparing the healthy tissue. This, in turn, has positive implications for maintaining function.
The amount of leak of dye from the liposomes over the 48-h period after injection is consistent with previously reported clinically used liposomal nanocarrier formulations [33–35]. It should be noted that even with this slight leak, the nano-EB did not stain healthy tissue. Evans blue is a strongly polar molecule; leaked dye quickly binds to blood serum proteins, limiting its diffusion .
Evans blue dye was initially FDA approved for use in blood volume measurement. It was later found to be toxic in repeated use or large doses [29, 30]. However, the negative side effects of molecules are usually reduced and their pharmacokinetics are drastically altered when encapsulated in liposomal nanocarriers [19–22].
In vitro, nano-EB did not have a negative effect on the viability of either the tumor cell line (3RT1RT2A) or its healthy counterpart, primary astrocytes. Intravenous injections of either nano-EB or EB had no observable negative effect on the growth of animals treated, nor did it cause any visible signs of distress in the animals. This was to be expected since the doses used were much lower than those seen as toxic in previous studies [29, 30]. Further, studies using other types of brain tumors are ongoing to test the utility and generality of this methodology for effective demarcation of tumors for aiding the surgical removal of brain cancers.
This study demonstrates that size-mediated diffusion limitation made possible by nanocarriers is critical to maintaining tumor staining without ‘leak’ of dye into normal brain tissue. This methodology thus allows specific staining of the tumor as well as reduces or eliminates false positive staining thereby preventing the removal of healthy tissue. Avoiding the unnecessary removal of healthy tissue allows for preservation of function and thus the quality of life of the patient after resection.
This work was funded by support from the Georgia Cancer Coalition and Ian’s Friends Foundation to RVB.
Conflict of interest
The authors have no financial interests relevant to the technologies or data published in this manuscript.