Troponin I Overexpression Inhibits Tumor Growth, Perfusion, and Vascularization of Morris Hepatoma

Cell Culture and Generation of Recombinant Cells

MH3924A cells were cultured (37°C, 95% air/5% CO₂) in RPMI 1640 medium (Gibco BRL) supplemented with 292 mg/L glutamine and 20% fetal calf serum (FCS).

TnI Gene Transfer, Northern Blots, and Western Blots

Stable cell lines were generated after lipofection of MH3924A cells with a human TnI2 (pGT60hTNNI2; InvivoGen) expression vector by hygromycin selection (425 μg/mL) for 4 wk. RNA was isolated using Trizol reagent (Roche), and messenger RNA (mRNA) content was assessed by Northern blots using radioactive labeled probes for TnI and β-actin. Secretion of TnI protein into cell culture media was verified by Western blots. For this purpose conditioned medium was produced by culture of 10⁶ transfected MH3924A cells in Optimem medium (Life Technologies) for 48 h. TnI protein was purified by ammonium sulfate precipitation. After gel electrophoresis, protein was blotted to a polyvinylidene difluoride membrane (Bio-Rad Laboratories) and immunostained using monoclonal mouse anti-TnI (2 μg/mL; Acris) antibodies and ECL Western blotting detection reagent (Amersham Biosciences).

Measurement of Apoptotic/Proliferating Cells

Human umbilical vein endothelial cells (HUVECs) were isolated as described (7,8) and cultured in M199 medium supplemented with 4 ng/mL basic fibroblast growth factor, 20% FCS, 292 mg/L glutamine, 100,000 IE/L penicillin, and 100 mg/L streptomycin. After 3 d of coculture with wild-type (WT) or TnI-MH3924A in a “without-contact” system (pore size, 3 μm; BD Biosciences), proliferation of HUVECs was determined by methyl-³H-thymidine (³H-TdR) (Amersham-Buchler) incorporation after precipitation with 0.5 mol/L perchloric acid/1 mol/L NaOH, and apoptosis was measured in unfixed cells by YO-PRO-1 staining as described earlier (9). Coculture experiments were done in triplicate in 2 independent experiments.

Animal Studies

All animal studies were performed in compliance with the national laws relating to the conduct of animal experimentation. The in vitro doubling time of the cell lines was 16.95 h for WT-MH3924A and 18.7 h for hTnI-MH3924A cells. After subcutaneous inoculation of WT or TnI cells (2 × 10⁶ cells/animal) into the right thigh of ACI or RNU rats (Charles River Laboratories), tumor growth was measured using calipers.

Perfusion and blood volume in tumors with diameters between 10 and 13 mm (8−19 d after transplantation for WT-MH3924A [11.85 ± 3.62 d; median, 10 d] and 14−28 d for TnI-MH3924A [19.83 ± 5.26 d; median, 21 d]) were examined after intravenous bolus injection of 70−150 MBq H₂¹⁵O in 0.3 mL and 5−10 MBq ⁶⁸Ga-labeled albumin (same animals and session; albumin PET scan performed after the activity of H₂¹⁵O decayed). The size of tumors showed no statistically significant difference between the 2 tumor types with 660.5 ± 199.4 μL (mean ± SEM; range, 448.7−904.3 μL; median, 523.4 μL) for WT-MH3924A and 717.5 ± 115.6 μL (mean ± SEM; range, 605.8−904.3 μL; median, 696.5 μL). All PET studies were done as dynamic measurements (20 × 3 s, 6 × 10 s, 4 × 15 s, 6 × 30 s [¹⁵O-labeled water] and 12 × 10 s, 3 × 20 s, 4 × 30 s, 10 × 60 s [⁶⁸Ga-DOTA-albumin]) in 2-dimensional mode using a matrix of 256 × 256 on an ECAT HR+ (Siemens/CTI) scanner (pixel size, 2.277 × 2.777 × 2.425 mm; transaxial resolution, 4.3 mm). A transmission scan was done for 10 min before tracer administration with 3 rotating germanium pin sources to obtain cross-sections for attenuation correction. After iterative reconstruction with the space-alternating generalized expectation maximization method ([SAGE] 0 subsets, 20 iterations) applying median root prior correction, the dynamic PET data were evaluated after definition of volumes of interest (VOIs) as described earlier using the PMOD software package (7). Time–activity curves were created using VOIs, which consist of several regions of interest (ROIs) over the target area. For the input function, we determined the mean value of the VOI data obtained from the heart. ROIs were defined at 3−6 s after bolus injection in at least 3 consecutive slices (2.4-mm thickness) surrounded by 2 slices showing the heart in each direction. The ROI size (mean ± SEM) for the input function was 59.6 ± 25.1 voxels (WT-MH3924A) and 54.8 ± 24.9 voxels (TnI-MH3924A) (P = 0.63); the ROI size for the tumors was 52.4 ± 23.3 voxels (WT-MH3924A) and 61.0 ± 34.4 voxels (TnI-MH3924A) (P = 0.71). This approach was based on the findings of Ohtake et al., who demonstrated that the input function can be retrieved from image data with acceptable accuracy (10). Heart weight and heart volume of a 250-g rat are 1.0 g and 1.2 mL, respectively (11). Furthermore, the size of the heart was determined in the rats showing a median value of 795 μL (range, 605–904 μL). This results in an estimated recovery error of 10%, which can be neglected for the purposes of this study. The ⁶⁸Ga-DOTA-albumin uptake was expressed as the mean standardized uptake value (SUV) between 10 and 15 min after injection when the time–activity curve reaches a plateau (8).

In a further experiment an additional group of ACI rats bearing WT- (n = 5) or TnI- (n = 6) expressing hepatomas was examined by dynamic contrast-enhanced (DCE) MRI. Gadolinium-diethylenetriaminepentaacetic acid (Gd-DTPA; Omniscan, Nycomed GmbH), diluted in 0.9% NaCl (total volume, 600 μL), was used in a dosage of 0.1 mmol/kg and injected into the tail vein as a bolus within 5 s after the fourth measurement. MRI investigations were performed on a 1.5-T whole-body MRI system (Magnetom Vision; Siemens) using a custom-made radiofrequency coil for radiofrequency excitation and signal reception (6). To optimize the available signal-to-noise ratio, manual tuning and matching of the coil's resonance circuitry were performed individually for each animal. Animals were examined with transversal T1-weighted and T2-weighted pulse sequences followed by the T1-weighted DCE MRI scan. T1-weighted spin-echo MRI was performed using the following parameters: repetition time (TR) = 600 ms, echo time (TE) = 14 ms, 1 acquisition, acquisition time (TA) = 3 min 53 s, field of view = 50 × 37.5 mm, matrix = 128 × 106, slice thickness = 1 mm, resulting voxel size = 1.0 × 0.39 × 0.35 mm³; parameters of the transversal T2-weighted turbo spin-echo sequence were TR = 4,000 ms, TE = 96 ms, 1 acquisition, echo train length = 7, TA = 1 min 49 s, field of view = 50 × 37.5 mm, matrix = 140 × 256, slice thickness = 2 mm, voxel size = 2.0 × 0.36 × 0.15 mm³. T1-weighted DCE MRI was conducted with a gradient-echo sequence (FLASH): TR = 100 ms, TE = 6.5 ms, flip angle α = 90°, 1 acquisition, TA = 8 s, field of view = 50 × 37.5 mm, matrix = 64 × 48, slice thickness = 2 mm, resulting voxel size = 2.0 × 0.78 × 0.78 mm³. Six transversal slices were placed, 4 covering the tumor and 2 at the position of the kidneys. Sixty measurements were performed during a measurement time of 8 min (6).

Analysis of dynamic data containing the amplitude (A) and the exchange rate constant (k_ep) was based on a pharmacokinetic 2-compartment model (12). The signal–time curve was fitted using a Levenberg–Marquart algorithm. Amplitude (A) and the k_ep value were calculated for each pixel of the MR images. Exchange of infused contrast agent between the first compartment representing the intravascular space and the second compartment representing the extravascular extracellular space is described by the rate constant k_ep, which depends on perfusion and vessel permeability. The parameter amplitude (A) represents the relative change of signal intensity after contrast media injection and is physiologically a parameter of tissue blood and interstitial volume. ROIs were drawn manually covering the complete tumor at its maximum extension. To obtain precise local tumor perfusion data, additional ROIs were determined in the tumor center and in the periphery at the tumor base (6). Color-coded parameter maps of A and k_ep values were calculated and overlaid on morphologic T1-weighted MR images.

Tissue Preparation and Immunohistochemistry

To obtain quantitative data on tissue properties, which may influence tumor perfusion, an immunohistochemistry analysis of proliferation, apoptosis, necrosis, immune reactions, and vessel density was done. Tumors obtained from the animals immediately after the PET perfusion studies were fixed with IHC zinc fixative (BD Biosciences), 4% paraformaldehyde (and embedded in paraffin), or shock-frozen in liquid nitrogen–cooled isopentane, and kept in a freezer at −70°C. Immunohistochemistry of tumors was performed on 6-μm cryostat sections according to procedures as described (7–9). For immunohistochemical investigations, the monoclonal antibodies mouse anti-α-actin (1:100; Roche), mouse antirat-CD31 (1:500; BD Bisosciences), mouse antiproliferating cell nuclear antigen ([PCNA] 1:500; DakoCytomation), mouse antirat CD11b (1:500; BD Biosciences), rabbit antimouse-cyclooxygenase-2 ([COX-2] 1:300; Cayman Chemical Co.), monoclonal mouse antip53 (1:50; Roche), and mouse antihuman-macrophage migration inhibitory factor (MIF) (1:50; R&D Systems GmbH) were applied. Negative controls were performed using control immunoglobulin or omission of the primary antibody.

Apoptosis was measured on cross-sections of paraffin-embedded tumors by the TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling) technique using a commercial kit (Oncor).

Morphometric Analysis

Immunoreactive cells were recorded by video camera (Olympus HCC-3600 P high gain) and quantified using a computer-assisted image analysis system (VIBAM 0.0-VFG 1-frame grabber) as described earlier (7). For quantifying tumor vascularization, we measured microvessel density and the mean number of vessels in angiogenic hot spots (7). Morphometric analyses were performed twice by 2 independent observers (interobserver variability, <5%). Cell density (number/mm²) or percentage of positive cells was quantified by analyzing 5 ROIs. Immunoreactive areas were quantified by analyzing 5 hot spots per tumor.

Gene Arrays

Aliquots of WT- and TnI-MH3924A tumors (excised immediately after the PET perfusion studies) obtained from the tumor periphery were mashed in Trizol to extract RNA as described (7,8). Furthermore, mRNA extraction was done from the WT- and TnI-expressing cell line. Gene chip expression arrays were performed using The Rat Genome U34 Set (Affimetrix Inc.). RNAs of WT- (n = 6) and TnI-MH3924A (n = 3) tumors were pooled and evaluated on separate gene chips as described (7,8). After normalizing values to β-actin expression, ratios of genetically modified versus WT tumors were calculated. Ratios of modified tumors versus WT tumors were determined and values above 2 and below 0.5 were interpreted as significant. The complete experiment was repeated twice, revealing a high correlation of signals in both experiments (r = 0.99 [tumor]; r = 0.96 [cells]).

Statistical Methods

Results are presented as mean ± SEM. Statistical analyses were performed by the Mann–Whitney U test using the SIGMASTAT program (Jandel Scientific). P ≤ 0.05 was considered statistically significant.

Inhibition of Proliferation and Induction of Apoptosis in Coculture Experiments with HUVECs

To verify TnI expression, Northern and Western blots were performed showing increased TnI mRNA content in transfected MH3924A cells (Fig. 1A). In addition, a protein with a molecular weight of 18 kDa was identified in cell culture medium of TnI-expressing cells (Fig. 1B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1. 

Expression of TnI in clones of MH3924A. (A) Northern blot. (B) Western blot. Representative blots are shown. α-Tubulin or β-actin was used as internal control. bp = base pairs.

After incubation with ³H-TdR and perchloric acid extraction, HUVECs cocultured with WT-MH3924A revealed a 30% increase in proliferation in comparison with HUVECs cultured alone. Coculture of HUVECs with TnI-MH3924A cells significantly decreased endothelial cell proliferation by 30% compared with coculture with WT-MH3924A cells (Fig. 2A). In comparison with HUVECs cocultured with WT-MH3924A, coculture with TnI-MH3924A cells resulted in a 2.6-fold increased percentage of apoptotic HUVECs (Fig. 2B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2. 

Proliferation and apoptosis of HUVECs cocultured with WT-MH3924A or TnI-MH3924A cells. (A) Proliferation. (B) Apoptosis. Values are mean ± SEM. *P < 0.05 vs. WT-MH3924A.

Inhibition of Tumor Growth and Proliferation and Induction of Apoptosis in TnI-MH3924A Tumors

At 36 d after inoculation in ACI rats, in vivo growth of TnI-MH3924A tumors was reduced by 90% (Fig. 3). In cell-mixture experiments, mixtures of 10% and 90% TnI-overexpressing MH3924A cells and corresponding parts of WT-MH3924 cells were inoculated. The 90% TnI tumor group revealed 60% tumor growth inhibition 36 d after inoculation, whereas tumors with 10% TnI-MH3924A cells demonstrated no significant growth inhibition. To exclude immunologic effects as possible causes of growth inhibition, additional experiments were done on athymic (nude) RNU rats. Eighteen days after inoculation, 60% suppression of tumor growth was noted in RNU rats bearing TnI-overexpressing tumors (1,450 ± 160 μL; n = 3) in comparison with WT-MH3924A (3,600 ± 400 μL; n = 5).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3. 

Tumor growth of WT-MH3924A and TnI-MH3924 tumors until 40 d after inoculation. Values are mean ± SEM. ***P < 0.001; **P < 0.01; *P < 0.05 vs. WT-MH3924A.

Immunohistochemistry data were in concert with the coculture experiments showing 70% increased apoptosis, whereas the proliferation rate was significantly decreased (25%; Table 1). Furthermore, the necrotic area was significantly increased in TnI-expressing tumors. Analysis of anti-p53 staining revealed no significant differences. Furthermore, in TnI-MH3924A tumors the percentage of CD11b−immunoreactive cells showed a significant 54% reduction (Table 1). As a marker for inflammation, we used COX-2 antibodies and found that in TnI-MH3924A tumors the percentage of COX-2−immunoreactive cells was significantly reduced to 2% (Table 1). MIF staining showed a significant 24% and 25% increase in the periphery and the center, respectively.

Perfusion and Vascularization in TnI-MH3924A: PET, MRI, and Immunohistomorphometry

Because we expected differences in tissue perfusion or blood volume in TnI-expressing tumors, dynamic PET measurements with H₂¹⁵O and ⁶⁸Ga-DOTA-albumin as well as MRI studies were performed (Figs. 4 and 5). Pharmacokinetic analysis of H₂¹⁵O PET data revealed that tumor perfusion, represented as K₁ (in mL × mL tissue⁻¹ × min⁻¹), showed a mean decrease of 24% (P = 0.14) in TnI-MH3924A (0.32 ± 0.05) in comparison with WT-MH3924A (0.43 ± 0.04; Table 2). However, the flow of tracer from tumor tissue back to the blood vessels, represented as k2, was significantly decreased by 66% (P = 0.002) in TnI-MH3924A (0.25 ± 0.07) compared with WT-MH3924A (0.73 ± 0.09; Table 2). Additionally, the fractional volume of distribution (DV) significantly increased in TnI-MH3924A (1.69 ± 0.40; P = 0.022) compared with WT-MH3924A (0.65 ± 0.11; Table 2), whereas the vascular fraction (VB) was not significantly different between WT (0.02 ± 0.01) and genetically modified (0.016 ± 0.007) tumors (Table 2). Furthermore, blood-pool PET using ⁶⁸Ga-DOTA-albumin for selective measurement of the perfused intravascular space demonstrated no difference between TnI-expressing (0.48 ± 0.03) and WT tumors (0.48 ± 0.06) (Table 2).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4. 

H₂¹⁵O PET. Representative images and time–activity curves measured in WT-MH3924A (A and C) and TnI-MH3924A (B and D) tumor; 2-dimensional acquisition. Results of modeling analysis are shown in Table 2. T = tumor.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5. 

Functional MRI. Representative T2-weighted images of WT-MH3924A (A) and TnI-MH3924A (B) tumor. Corresponding parametric images for A and k_ep are shown in C and D. Range of signal intensity for k_ep and A is indicated at bottom.

In line with the results from PET, the vascular tumor volume described by amplitude was not significantly different between TnI and WT MH3924A tumors in the DCE MRI experiments. ROIs of the tumor periphery revealed significantly increased amplitudes in comparison to total tumor values, whereas amplitudes of the tumor center were significantly decreased (Table 2). The rate constant k_ep, highly influenced by the curve gradient and by vessel permeability, was significantly decreased by 52% in the periphery of TnI-MH3924A tumors and by 56% in the total tumor area. k_ep values of the tumor center were reduced in TnI-MH3924A as well as WT-MH3924A, however without significant differences (Table 2).

In general, microvessel density in both tumor groups was 2−3 times higher in the periphery in comparison with the center, as indicated by the MRI experiments. In contrast to the functional measurements of tumor blood volume, microvessel density was decreased by 45% in the periphery of TnI-MH3924A compared with WT-MH3924A tumors (Table 1). Furthermore, the mean number of blood vessels was significantly decreased in TnI-MH3924A (137.70 ± 15.95 per mm²) in comparison with WT-MH3924A (222.02 ± 25.58 per mm²). Tumor sections were also stained with α-actin, which is expressed in mural cells of arterioles and venules but not in capillaries. The α-actin−immunoreactive area was similar in the periphery as well as in the center of TnI- and WT-MH3924A tumors (Table 1), which is in accordance with our functional blood volume measurements. However, it was twice as high in the circumference in comparison with the center (Table 1).

Effects of hTnI Gene Transfer on Expression of Tumor Genes: Gene Arrays

To assess the general effects of TnI gene transfer on MH3924A cells, the in vitro and in vivo gene expression pattern was studied using gene arrays. Because transfer of selection markers such as the hygromycin gene and the selection procedure may cause changes in the genetic program, we analyzed the expression pattern in the tumor cell lines as well as in explanted tumors of these cell lines. In general, we found changes in expression of genes related to apoptosis, signal transduction/stress, or angiogenesis (Table 3), with clear differences between in vitro and in vivo patterns. This indicates that the observed effects are caused by the microenvironment in vivo.