Brain Interstitial Fluid Drainage Alterations in Glioma-Bearing Rats

Real time imaging and measurement of the drug distribution in the brain interstitial space (ISS) are able to determine the effeicency of local brain drug delivery to treatment gliomas. In the present study, we used a tracer-based magnetic resonance imaging (MRI) method to quantitatively analyze the effects of glioma growth on ISF drainage. Sprague-Dawley rats were randomly divided into six groups (n = 6). C6 glioma cells were implanted into either the caudate nucleus or thalamus of rats and then were examined 10 or 20 days after implantation. The two control groups were treated with vehicle. A tracer was injected into the caudate nucleus of control rats and rats with gliomas growing in the thalamus for 10 or 20 days. The tracer was similarly injected into the thalamus of control rats and rats implanted with gliomas in the caudate nucleus. The diffusion and clearance parameters of the tracer were calculated using tracer-based MRI techniques. We found that glioma implanted in the caudate nucleus significantly decreased the speed of the ISF flow in thalamus. With the growth of the glioma in thalamus, the drainage route of the brain ISF flow was altered in the caudate nucleus, but the speed of the flow was not significantly changed. These findings indicate that tracer-based MRI is a promising technique for optimizing the interstitial administration of therapeutics aimed at treating brain gliomas.

based on a full understanding of the interstitial fluid drainage changes in the brain. We are not only concerned about changes of interstitial fluid inside the tumor, but also concerned about the interstitial fluid drainage outside the tumor. In our previous studies, we have found that the interstitial fluid drainage in the tumor slows down due to tumor growth [7].
In this study, we used the tracer-based MRI to determine the effects of tumor growth on ISF drainage in deep brain regions by implanting C6 glioma cells into two regions of the rat brain and measuring changes in ISF drainage outside the tumors.

Brain glioma model
All animal experiments were performed following the approval of the Ethics Committee of Peking University Health Science Center (Approval No. LA 2012-016).
Male Sprague-Dawley rats (weight, 250-300 g) were randomly divided into the following six groups (n=6). Group A: tracer injected into the caudate nucleus of normal rats; Group B: tracer injected into the thalamus of normal rats; Group C: tracer injected into the caudate nucleus after C6 cells had been implanted in the thalamus for 10 days; Group D: tracer injected into the caudate nucleus after C6 cells had been implanted in the thalamus for 20 days; Group E: tracer injected into the thalamus of rats with a glioma that had been grown in the caudate nucleus for 10 days; Group F: tracer injected into the thalamus of rats with a glioma that had been grown in caudate nucleus for 20 day.
Each experimental rat was anesthetized with intraperitoneal injection of pentobarbital sodium (50 mg/kg). To implant C-6 glioma cells, the rats' skin was incised and a small a hole (1.0 mm diameter) was drilled into the skull above the target implantation region: for the caudate nucleus, the coordinates were −0.8 mm, +2 mm, and 6 mm ventral; for the thalamus, +2.5 mm, +2.5 mm and 6.5 mm ventral. The C6 cells (10 µl, 1 × 10 5 cells) were injected into the brain with a Hamilton microsyringe at a rate of 1 µL/min.

Tumor detection
MRI examination was performed using a 3.0 T MRI scanner (Siemens, Germany). T2-weighted imaging (T2WI) sequence was used to detect the tumors at 10 and 20 days after implantation with the parameters as follows: repetition time, 3600 ms; echo time, 91 ms; flip angle, 120°; field of view, 80 mm; matrix, 256 × 256; slice thickness, 2 mm; three averages.
Gd-DTPA (Magnevist, Bayer, Germany), diluted with double-distilled water to 10 mmol/L, was microinjected (2 µL, 0.2 L/min) into the caudate nucleus of rats in groups A, C, and D and into the thalamus of rats in groups B, E, and F. After the injection, the needle remained in place for 5 min before it was slowly extracted. A continuous T1WI-MPRG sequence was performed until the appearance of the images had nearly recovered to those acquired at T0.

Calculation of interstitial fluid flow parameters
The calculation of interstitial fluid flow parameters used in this study has been applied in: Transportation in the interstitial space of the brain can be regulated by neuronal excitation [8], the calculation is as below: Substances in the extracellular space (ECS) follow the diffusion equation where the diffusion coefficients of a given molecule in free media and brain ECS are D and D*, respectively, C is related with time and position, indicating the actual concentration in ECS,  2 and  represents the spatial derivatives in the appropriate coordinate system, v is the flow rate of the ISF in the ECS, f(C) denotes the clearance rate of probe in ECS, Q is the concentration of the diffusion source in ECS, and α is the volume fraction of the ECS within the brain. Because Gd-DTPA was not administered continuously, it was taken to be injected at the initial time point, and the value of Q was negligible in this study. The region of interest was as close as possible to the injection site (2 mm), and thus the quality of diffusion was assumed to have had no effect. It was also assumed that the ISF in the local ECS did not flow in a straight line. The following formula was used to calculate the diffusion parameters: Where ΔSI m (r,ti) represents the radial range of diffusion measured at time ti. The diffusion parameters D and k' were measured using the time model and the measured curve. Tortuosity (λ) = (D / D*) 1/2 . The model's predicted profiles and measured profiles at a given time point were minimized to find the effective diffusion coefficient (D*), free diffusion coefficient (D), rate constant (k') and tortuosity (λ). The half time (t 1/2 ) was determined using linear fit.

Immunohistochemical analysis
After perfusion with 4% paraformaldehyde, brain tissues were removed and embedded in paraffin. Coronal sections were sliced to a thickness of 5 μm and then deparaffinized. Heat-mediated antigen retrieval was performed at pH 6.0. Sections were incubated overnight at 4°C with a primary antibody against tenascin C (rabbit anti-rat; 1:300; Abcam, Cambridge, UK), washed, and then incubated with a secondary antibody for 30 min at room temperature. The sections were stained with diaminobenzidine and counterstained with hematoxylin.

Data and statistical analyses
Image analysis was performed using Matlab imaging software (MathWorks, Natick, MA, USA). Tumor volumes were calculated by the equation V ≈ 4/3 × π × (0.5) 3 × (abc), as a, b and c represent the length, width and height of the tumor, respectively. Statistical analyses were conducted using APSS 19.0 and Excel 2010 (Microsoft Co., Redmond, WA, USA). One-way analysis of variance was used to examine differences among multiple groups, and values of P less than 0.05 were considered to be significantly significant.

Glioma implanted in rat brain assessed with MRI
Tumor formation and growth were observed with T2weighted imaging (T2WI) 10 and 20 days after the C6 glioma cells were implanted into either the thalamus or caudate nucleus of rats (Fig. 1). Significant increases in tumor volume were demonstrated in all implanted groups as shown in Figure 1.

ISF flow changes in the caudate nucleus after tumor implantation in the thalamus
Injection of Gd-DTPA into the caudate nucleus resulted in local MRI signal hyperintensities ( Fig. 2 and 3). With the diffusion of Gd-DTPA into the ECS, the high-signal intensity region gradually expanded. Significant differences were demonstrated in the effective diffusion coefficient (D*)，with D* increased in tumor groups: group C, (6.91 ± 4.24) × 10 -5 mm 2 /s; group D, (10.39 ± 8.33) × 10 -5 mm 2 /s); P < 0.05 compared with control group A, (33.33 ± 3.46) ×10 -5 mm 2 /s. Significant differences were also demonstrated in the clearance rate coefficient (k') between groups A and group D (P < 0.05), but not between groups A and C, or C and D: group A, (0.65 ± 1.00) × 10 -5 mm 2 /s; group C, (2.00 ± 2.36) × 10 -5 mm 2 /s; group D, (3.04 ± 2.01) × 10 -5 mm 2 /s. There were no differences in t 1/2 among the three groups: group A, 1.47 ± 0.21 h; group C, 1.24 ± 0.21 h; group D, 1.39 ± 0.21 h; P > 0.05. However, groups A, C, and D showed significantly different tortuosity values: group A, 1.32 ± 0.34; group C, 3.03 ± 0.91; group D, 2.65 ± 0.94; P < 0.05. We also observed pathway changes in the ISF flow among groups A, C, and D, which the magnetic tracer in group D reach the edge of cortex at the early stage of diffusion (Fig. 3), and the magnetic tracer in groups C hardly get into the cortex (Fig. 2B).
Because figure 2 and 3 share the same control group, the distribution pattern of Gd-DTPA in the control group have been included in figure 2 (Fig. 2A).

ISF flow alterations in the thalamus after tumor implantation in caudate nucleus
The MR images of the Gd-DTPA distribution in the thalamus of groups B, E and F are shown in Figure 4. With the diffusion of Gd-DTPA into the ECS, the high-signal intensity region gradually expanded but remained confined to the adjacent region. Significant differences were demonstrated in the effective diffusion coefficients (D*) among groups B, E, and F: group B, (33.72 ± 3.87) × 10 -5 mm 2 /s; group E, (17.33 ± 11.27) × 10 -5 mm 2 /s; group F, (5.26 ± 1.06) × 10 -5 mm 2 /s; P < 0.05). Significant differences in the clearance rate coefficients (k') were also demonstrated, and the values for k' in the tumor groups (groups E and F) were increased compared with that in the normal group: group B, (1.55 ± 0.78) × 10 -5 mm 2 /s; group E, (2.80 ± 2.60) × 10 -5 mm 2 /s; group F, (2.92 ± 0.86) × 10 -5 mm 2 /s; p < 0.05. We also found significant increases in t1/2 between the tumor groups (10 and 20 days after implantation) and the normal group (group B, 0.81 ± 0.03 h; group E, 1.47 ± 0.24 h; group F, 1.46 ± 0.55 h; p < 0.05) as well as significant increases in tortuosity (λ) in the tumor group 10 and 20 days after the implantation (group E, 2.76 ± 1.07; group F, 3.13 ± 0.33) compared with that in the normal group (group B, 1.23 ± 0.64; p < 0.05). However, pathway changes for the ISF flow in the thalamus were not observed after tumor implantation in the caudate nucleus, with alterations observed only in the stated parameters.

Immunohistochemical analysis of an extracellular matrix molecule in gliomas implanted in the thalamus for 10 and 20 days
Immunohistochemical analysis was performed to detect tenascin C changes in the extracellular matrix of gliomas implanted in the thalamus of the rat brain for 10 and 20 days. No accumulation was demonstrated in the ISS outsides the glioma, while, larger amounts of tenascin C accumulated in the extracellular matrix of the 20-day group (Fig. 5C) compared with the 10-day group (Fig.  5B).

DISCUSSION
In the present study, we demonstrated that the ISF flow in regions adjacent to a glioma was altered with the growth of the tumor in the deep brain. Glioma in the caudate nucleus caused a significant decrease in the speed of the ISF flow in the thalamus. As the glioma grew in the thalamus, the drainage route for the ISF flow was altered in the caudate nucleus, but the speed of that flow was not significantly altered. We also found that the ISF flow did not transport molecules from the interstitial space (ISS) outside the glioma to the peripheral region of the glioma. The tracer distribution of ISF and flow bypassed the glioma, as it appeared as a high-intensity signal that wrapped around the glioma.  Early studies reported that some of the matrix components excessively expressed with tumor growth and caused the tortuosity of the brain ECS were increased inside the tumor [9,10]. Our immunohistochemical results demonstrated that the expression of tenascin C increased with the growth of the glioma. The pressure inside the tumor is greater than that in the surrounding tissue, and the interstitial fluid inside the tumor flows from the centerline to the edge [11,12]. In the present study, we found that diffusion in the brain ECS was strictly limited to inside the glioma [7,13], and the diffusion in the brain ISS surrounding the tumor was similarly restricted, with significant changes in the ISF flow rate and flow route. This effect became more pronounced with increasing tumor volume. The different drainage route of the brain ISF in the caudate nucleus was demonstrated when the tumor was implanted in the thalamus. The drainage route was damaged at 10 days and presented as an inability of the ISF to drain into the cortex. At 20 days, the mass effect of the tumor squeezed and distorted the caudate nucleus; the ISF drainage route was demonstrated to be through narrow strips, and the brain ISF again drained to the parenchymal margin of cortex. However, we also found that the ISF flow is also related to the location of the tumor. In a previous study, we have showed that the interstitial spaces within the caudate nucleus and thalamus have different diffusion and flow parameters.
Although the mechanisms underpinning the regionalization of the ISF drainage in the deep brain will require further study, our results provided valuable insights for convection-enhanced delivery (CED) treatments of glioma. CED, which utilizes a convective pattern rather than simple diffusion, directs small molecules or macromolecules into the ISS, where they are distributed to the target [14]. Theoretically CED can maximize the brain concentration of the treatment drug with minimal systemic exposure and toxicity [15]. It has been found that the concentration of the drug injected into the brain by CED can reach more than a thousand times that following traditional drug administration, and the region in which the drug is applied can be relatively vast [16][17][18][19]. However, the use of CED has been limited in clinical practice because of the difficulty in monitoring and controlling the concentration and distribution of the injected drugs [20].
In summary, our results demonstrated that tracerbased MRI provides a real-time in vivo method for monitoring alterations in deep brain ISF drainage. Our findings also indicated that placement of CED administration should be designed based on the divisions of the brain ISS and should take into account potential changes in the brain ISS and ISF flow based on the location of the tumor as well as on the changes in the clearance rate and drainage pathways outside the tumor as the volume of the tumor grows. The tracer-based MRI will play a more and more important role in the optimizing the interstitial administration of therapeutics aimed at treating brain gliomas.