A typical obstacle to cancer therapy is the limited distribution of low molecular weight anticancer drugs within the carcinoma tissue. In experimental carcinoma, imatinib (STI571) increases efficacy of synchronized chemotherapy, reduces tumor interstitial fluid pressure, and increases interstitial fluid volume. STI571 also increases the water-perfusable fraction in metastases from human colorectal adenocarcinomas. Because the mechanism(s) behind these effects have not been fully elucidated, we investigated the hypothesis that STI571 alters specific properties of the stromal extracellular matrix. We analyzed STI571-treated human colorectal KAT-4/HT-29 experimental carcinomas, known to have a well-developed stromal compartment, for solute exchange and glycosaminoglycan content, as well as collagen content, structure, and synthesis. MRI of STI571-treated KAT-4/HT-29 experimental carcinomas showed a significantly increased efficacy in dynamic exchanges of solutes between tumor interstitium and blood. This effect was paralleled by a distinct change of the stromal collagen network architecture, manifested by a decreased average collagen fibril diameter, and increased collagen turnover. The glycosaminoglycan content was unchanged. Furthermore, the apparent effects on the stromal cellular composition were limited to a reduction in an NG2-positive stromal cell population. The current data support the hypothesis that the collagen network architecture influences the dynamic exchanges of solutes between blood and carcinoma tissue. It is conceivable that STI571 reprograms distinct nonvascular stromal cells to produce a looser extracellular matrix, ultimately improving transport characteristics for traditional chemotherapeutic agents. Mol Cancer Ther; 15(10); 2455–64. ©2016 AACR.

Carcinomas are characterized by a deranged vasculature with aberrant blood flow and leaky blood vessels, leucocyte infiltrates, and a dense and often a fibrotic stromal extracellular matrix (ECM). In addition, a pathologically elevated interstitial fluid pressure (IFP) characterizes both human and experimental carcinomas. These stromal properties contribute to the limited distribution of blood-borne drugs within carcinoma tissue (1–3).

Several agents that lower IFP in experimental carcinomas increase the interstitial fluid volume in parallel to an increase of the uptake and efficacy of small molecular weight chemotherapeutic agents (1, 4–6). Previous studies have revealed a correlation between IFP and the properties of the collagen network in stroma such as collagen fibril density, fibril structure, or network architecture in experimental carcinomas (7–11). In normal loose connective tissues, IFP is actively controlled by connective tissue cells that exert a tension on the collagen/microfibrillar network, thereby restraining the swelling of the underhydrated glycosaminoglycan (GAG)/proteoglycan ground substance (12). Increases in GAG/proteoglycan ground substance will increase IFP provided there is a collagen network that resists the tendency toward increased swelling. Such an example was demonstrated in pancreatic ductal adenocarcinomas (PDAC) where high concentrations of both collagen type I and the GAG hyaluronan correlated with a particularly high IFP and solid stress in the stroma (13). In a therapeutic experiment, enzymatic degradation of hyaluronan lowered IFP, restored blood flow, and improved efficacy of gemcitabine (13). Another study with an experimental model of PDAC showed that inhibition of the Hedgehog cellular signaling pathway reduced stromal fibrosis, reduced collagen content, and increased blood flow and efficacy of gemcitabine (14).

STI571 is a low-molecular weight tyrosine kinase inhibitor originally developed for the treatment of chronic myeloid leukemia by blocking the kinase activity of the pathogenic Bcr-Abl fusion protein (15, 16). STI571 also selectively inhibits the intrinsic receptor tyrosine kinases in discoidin domain receptors, c-Kit, colony-stimulating factor-1 receptor (CSF-1R), and PDGF receptor-α and -β (16, 17). STI571 lowers tumor IFP in several models of experimental carcinoma (including the KAT-4/HT-29 model) without affecting tumor growth (18). Furthermore, STI571 treatment of the ECM-rich KAT-4/HT-29 carcinoma model increases the extracellular fluid (ECV) volume from 34% to 42% of total tissue water (TTW), without significantly altering plasma volume (about 0.9% of the TTW; ref. 19). TTW is the sum of intracellular, ECV, and plasma. Because the ECV is largely contained within the interstitial (stromal) ECM-compartment (12), the data suggest that treatment with STI571 increases the volume of the ECM in the carcinoma and that this compartment amounts to about 25% of the carcinoma volume. These preclinical data agree with the findings that treatment of patients with STI571 increases the water-perfusable tumor fraction in metastases from colorectal adenocarcinoma (20). In the current study, we used MRI to investigate how STI571 affects the uptake of low-molecular weight contrast agents. We also analyzed the stromal ECM to investigate potential ECM-related mechanisms that could influence the solute exchange between blood and tumor tissue.

Tumor model

KAT-4 (ATCC) carcinoma cells were cultured in DMEM or RPMI1640 (Invitrogen) supplemented with 10 % FBS (Saveen-Werner) and penicillin + streptomycin (SVA). Cells were harvested with trypsin (SVA) and washed two times in PBS (SVA) before making the final cell suspension in PBS. Fifty microliters of cell suspension containing 2 × 106 KAT-4 was injected subcutaneously into the left flank of 8-week-old female Fox Chase SCID mice (CB.17, M&B). The mice were housed at the animal facilities of Uppsala and Lund Biomedical Centers. All manipulations of mice were performed under isoflurane (Abbot Scandinavia) anesthesia and the experiments were approved by the Ethical committees for animal experiments in Lund and in Uppsala, Sweden. The origin of KAT-4 has been debated—first described to be originating from human thyroid carcinoma (21), KAT-4 was later shown to resemble HT-29 human colon carcinoma (22). KAT-4 cells, whose identity was investigated by short tandem repeat loci analyses (IdentiCell), were used. KAT-4, as expected (22), matched with HT-29 although alleles D13S317:12 and TH01:9 were absent. The cell line has not undergone epithelial-to-mesenchymal transition (23).

Treatment of animals

Tumor size was measured externally with a caliper. Tumors that had reached volumes between 0.2 and 0.4 mL were randomized into two groups. A treatment group received STI571 (kindly donated by Novartis) 100 mg/kg/d dissolved in PBS and a vehicle group received only PBS. In both groups the substances were administered orally via gavage for 4 consecutive days.

MRI

Animals were anesthetized with 3.5% isoflurane in mixture of 200 mL/minute oxygen and 200 mL/minute nitrous oxide and maintained at 1.5% to 2% isoflurane. Inside the magnet, the respiratory rate of the animal was monitored and the body temperature was maintained at 37°C using warm air (SA Instruments Inc.). Contrast-enhanced MR imaging was performed with a 9.4 T MR scanner (Agilent Inc.) equipped with a 6-cm inner diameter gradient system having a maximum gradient strength of 1,000 mT/m. Reference scans were acquired to ensure the correct animal position inside the magnet and a total of 23 slices were defined with a gap of 0.25 mm to image the whole tumor area. The animal was taken out of the magnet and 100 μL of Omniscan (Gd-DTPA-BMA; GE Healthcare) was injected intraperitoneally with care taken not to move the position of the animal. With the animal placed back in the magnet, contrast-enhanced MR images were acquired using T1-weighted gradient echo sequence (TR: 100 ms, TE: 2 ms, number of averages: 8, resolution in plane 117 μm × 117 μm, slice thickness 0.5 mm). Total scan time was 3 minutes and scans were started with 10-minute interval. Animals were scanned until a clear decrease of the contrast enhancement could be observed in the connective tissue. Typically, the animals were scanned 60 to 80 minutes after injection. Mice were euthanized after the measurements.

MRI data analysis

Analysis was carried out using in-house written Matlab scripts (MathWorks). The first images after injection when no contrast enhancement could be detected were used as reference. The carcinoma tissue was manually delineated on all slices and marked as region of interest (ROI). The tumor volume was calculated by multiplying the total amount of pixels in the ROI in all slices. For all pixels originating from the tumor tissue, the change of signal was plotted as a function of time. To reduce processing time, images were down-sampled with a factor of 2 before curve fitting. The data were filtered using principal component analysis (PCA; ref. 24) in which the 2 first components, representing more than 95% of the variation of the data, were used to reconstruct the data. Quantitative model fitting of this kind of data exists (25), but in this study a more qualitative description of the data was chosen. We considered only the pixels with a signal change greater than three SDs of the noise in the image after PCA. Thus, around 10% of the total number of pixels was not considered due to lack of any significant signal enhancement. The change in the signal of a pixel inside the tumor was fitted to three model line shapes; a linear enhancement, a monoexponential enhancement (wash-in), and a biexponential enhancement (wash-in/wash-out). The most appropriate curve describing the pixel enhancement was chosen from the best correlation coefficient R2 calculated after the fit. Maximum enhancement, time constants for wash-in/wash-out, tumor volume and the distribution of the type of signal enhancement are reported after the analyses. Results are reported in histograms for all pixels in the STI571-treated and the control group. To compare the distribution for every graph we performed a Mann–Whitney test (nonparametric, unknown distribution).

Hydroxyproline determinations

KAT-4/HT-29 carcinomas were hydrolyzed in 6 mol/L HCl for 4 hours at 120°C at a pressure of 2 atmospheres. Hydroxyproline content in the hydrolysates was determined essentially as described previously(26).

Disaccharide fingerprint

KAT-4/HT-29 carcinoma tissues from STI571-treated and control tumors were lyophilized. GAG preparation, lyase treatment, fluorescence disaccharide labeling and separation were performed according to ref. 27. Briefly, tumors were protease and DNAase digested, and GAGs were purified on anion-exchange chromatography. GAGs were roughly estimated by the carbazol method. Then, 500 ng of GAGs was subjected to chondroitinase ABC (Sigma) degradation or degradation with a mixture of heparinases (in house preparation, purified from E. coli, stably singularly transfected with the pET-15b vector containing heparinase I, or vector pET-19b containing heparinase II or III, as provided by Prof. Jian Liu (University of North Carolina, Chapel Hill, NC). Fluorophore-labeling of the resulting disaccharides was performed by 2-aminoacridone (AMAC, Sigma). Pre-column AMAC-labeled disaccharides were analyzed with HPLC-fluorescence as described previously (27). Quantification was done by comparison to known weight of mock-treated standard disaccharides (Iduron).

Electron microscopy

For transmission electron microscopy (TEM) analysis, KAT-4/HT-29 carcinoma from control and STI571-treated mice were fixed in 0.15 mol/L sodium cacodylate–buffered 2.5% glutaraldehyde, post-fixed in 0.15 mol/L sodium cacodylate-buffered 1% osmium tetraoxide, dehydrated in graded ethanol series, impregnated in acetone, and embedded in epoxy resin. Ultra-thin sections were examined in a Philips CM-10 electron microscope (Philips) and the micrographs were quantified with ImageJ software (NIH, Bethesda, MD). For scanning EM analysis, KAT-4 tumors from PBS and STI571-treated mice were processed by alkali maceration as described previously (28) and analyzed in a Philips 515 electron microscope.

Ex vivo collagen synthesis

Tumors were excised and cut into 300- to 700-μm slices with a vibratome (Bannockburn). Samples were weighed (average wet weight 169 ± 20 mg, of all 18 samples) and kept on ice until they were put in collagen synthesis media (MEM proline-glycine-free (Invitrogen) + 0.284 mmol/L ascorbic acid (Sigma) + 0.001 mmol/L FeSO4 (Sigma) + 1 mg/mL BSA (Sigma) + 0.01 mmol/L GM6001 (matrix metalloproteinase inhibitor, Chemicon) together with 50 μCi/mL of 14C-labeled Proline and Glycine (PerkinElmer) at a Proline to Glycine ratio of 1:6. The slices were incubated at 37°C in a humidified incubator for 6 hours. Thereafter, the samples were dried in a vacuum centrifuge. Collagen was solubilized in 0.5 mol/L HAc containing 3 mg/mL pepsin (Sigma) pH 2.5 for 24 hours at 4°C. The samples were centrifuged 17,000 × g for 30 minutes at 4°C and supernatants collected. The supernatants were mixed with 2.8 mol/L NaCl to obtain a final concentration of 0.7 mol/L NaCl, in which native triple helical collagen was precipitated. Precipitated collagen was washed two times in 0.5 mol/L HAc containing 0.7 mol/L NaCl. Dry pellets were stored frozen, pending further analysis. Each pellet was solubilized in equal amounts of 2× sample buffer [0.2 mol/L Tris-HCl pH 8.8 (Sigma), 18% Glycerol (Merck), 0.01 % bromphenol blue (Merck), 4% SDS (Merck), 10% 2-Mercaptoethanol (BDH)] and heated at 100°C for 10 minutes. The samples were run on 6 % SDS-PAGE gels after which the gels were stained by Coomassie blue and dried. The dried gels were exposed to an imaging plate (Fuji) and developed in BAS-2500 Bioimaging analyzer (Fuji). Band intensities were normalized to the individual wet tissue weight.

RT-PCR and real-time qPCR

Total RNA was extracted from tumors (five biological replicates) using TRIzol reagent (Thermo Fisher). Five-hundred nanograms RNA was used for reversed transcription using Superscript VILO (Thermo Fisher). Real-time qPCR was performed with TaqMan probes listed in Supplementary Table, using an Applied Biosystems 7300 detection system. Gene expression was normalized to Actb transcript.

Immunofluorescence

Tumors were snap-frozen in isopentane at −80°C, embedded in OCT (Sakura), and sectioned. Frozen 6-μm sections were fixed in 4% paraformaldehyde (Merck) or acetone (Sigma), blocked in 40% serum [20% goat serum (Serotec) and 20% horse serum (SVA, Sweden) or pig serum from Chemicon (Temacula)] and incubated with the following primary antibodies: monoclonal rat anti-mouse CD31 clones Mec 13.3 and 390 (BD Biosciences), monoclonal rat anti-mouse F4/80 (Serotec), polyclonal rabbit anti-mouse NG2 (Chemicon), mouse anti-α-smooth muscle actin (α-SMA) clone 1A4 FITC-conjugated (Sigma), and rat anti-Reticular Fibroblast Marker (RFM; Cederlane Laboratories). The following secondary antibodies were used: FITC-conjugated goat anti-rat IgG and Texas Red-conjugated goat anti-rabbit (Vector Laboratories). Exchange of primary and secondary antibodies for mouse or rabbit normal IgG or PBS was performed in all combinations necessary to establish the specificity of the observed staining. DAPI (4′,6-diamidino-2-phenylindole) from Sigma was used for nuclear staining. Images were retrieved with a Nikon Eclipse 90i microscope (Nikon Instruments). Image analysis and pixel quantification were performed with Photoshop (Adobe) and ImageJ (NIH, Bethesda, MD) software, respectively. The mean percentage of pixels per image was calculated from several images per investigated tumor. Colocalization data is presented as the percentage of CD31-positive pixels that colocalized with NG2, and vice versa.

STI571 treatment increased exchange between plasma and interstitial fluid in KAT-4/HT-29 experimental carcinoma

MRI was used to determine wash-in and wash-out characteristics in KAT-4/HT-29 carcinomas. The average volumes of the investigated KAT-4/HT-29 carcinomas as determined by MRI were not different between the STI571-treated (0.14 ± 0.12 mL, n = 9) and the control group (0.15 ± 0.08 mL, n = 8). The distribution of the low-molecular weight MRI contrast agent Omniscan within the extracellular fluid of experimental carcinomas was followed during a 60- to 80-minute time course per tumor. Omniscan quickly and freely distributes to the extracellular space where its elimination tends to be rapid and complete (29). The contrast signal confidently represents extracellular fluid as uptake by cultured cells has been shown to be ineffective and requires prolonged incubation times (30). Dynamic contrast–enhanced MRI data were recorded and analyzed for each individual pixel. Contrast enhancement in the pixels followed time courses which fit best either with biexponential, monoexponential, or linear enhancement curves as exemplified in Fig. 1. The total number of pixels having significant signal enhancement from all STI571-treated (n = 9) and control (n = 8) tumors, respectively, were pooled and compared. Thus, a total of approximately 33,000 pixels from the STI571-treated group and approximately 25,000 from the control group were examined. A majority of the pixels in both groups followed a biexponential enhancement curve (Fig. 1A), around 11% to 12% followed a monoexponential (Fig. 1B), and less than 7.5% followed a linear enhancement curve (Fig. 1C). In these distributions, no significant differences were recorded between the treatment and control group. On the other hand, STI571-treated carcinomas displayed significant increases in time constants in pixels following monoexponential wash-in enhancement curves (Fig. 2A) and biexponential wash-in (Fig. 2B) and wash-out (Fig. 2C) enhancement curves (P < 0.001 in Fig. 2A–C). Furthermore, the pixels best described by monoexponential uptake enhancement curves revealed a significant shift toward higher signal amplitudes (Fig. 2D, P < 0.001), indicating an increased uptake after STI571 treatment. When biexponential wash-out time constants per tumor were compared between the groups the STI571-treated tumors displayed a significantly higher average (Fig. 2F, P = 0.047). Together, the data show that STI571 treatment increased the dynamic exchange of solutes between blood and carcinoma tissue.

Figure 1.

Examples of the three different types of time courses for uptake of contrast agent. A, biexponential uptake. B, monoexponential uptake. C, linear uptake; y-axes, contrast enhancement that is proportional to contrast concentration (a.u., arbitrary units); x-axes, time (minutes). Data are obtained from three individual pixels. Points are acquired data points and lines show calculated best-fit curves.

Figure 1.

Examples of the three different types of time courses for uptake of contrast agent. A, biexponential uptake. B, monoexponential uptake. C, linear uptake; y-axes, contrast enhancement that is proportional to contrast concentration (a.u., arbitrary units); x-axes, time (minutes). Data are obtained from three individual pixels. Points are acquired data points and lines show calculated best-fit curves.

Close modal
Figure 2.

STI571 increases dynamic exchange between the blood and the tumor interstitium as measured by MRI. Histograms in AE show number of pixels in STI571-treated (red) and control tumors (blue) distributed according to their time constants calculated from the equations of the fitted curves. A, monoexponential wash-in. B, biexponential wash-in. C, biexponential wash-out. D, distribution of maximum amplitudes in individual pixels taking up contrast according the monoexponential kinetics. E, distribution of maximum amplitudes in individual pixels taking up contrast according the biexponential kinetics. F, average time constants in STI571-treated tumors (n = 9) and control tumors (n = 8); error bars, SEM. *, P < 0.001 by the Student t test. G, heatmap illustrating wash-out dynamics in PBS- and STI571-treated tumors, scale goes from blue (low time constant) to yellow (high time constant).

Figure 2.

STI571 increases dynamic exchange between the blood and the tumor interstitium as measured by MRI. Histograms in AE show number of pixels in STI571-treated (red) and control tumors (blue) distributed according to their time constants calculated from the equations of the fitted curves. A, monoexponential wash-in. B, biexponential wash-in. C, biexponential wash-out. D, distribution of maximum amplitudes in individual pixels taking up contrast according the monoexponential kinetics. E, distribution of maximum amplitudes in individual pixels taking up contrast according the biexponential kinetics. F, average time constants in STI571-treated tumors (n = 9) and control tumors (n = 8); error bars, SEM. *, P < 0.001 by the Student t test. G, heatmap illustrating wash-out dynamics in PBS- and STI571-treated tumors, scale goes from blue (low time constant) to yellow (high time constant).

Close modal

STI571 treatment decreased collagen fibril diameter

Electron microscopy was used to assess the overall structure of the collagen network in STI571-treated and control KAT-4/HT-29 tumors. TEM revealed a difference in collagen fibril diameter and morphology: Collagen fibrils in STI571-treated tumors were more heterogeneous in diameter and had distorted surface morphology, whereas fibrils in PBS-treated tumors were more uniform and had a smooth surface (Fig. 3A). Quantification revealed a Gaussian distribution of collagen fibril diameters in control tumors, whereas a skewed distribution toward thinner fibril diameters was apparent in STI571-treated tumors (Fig. 3B). This was also reflected in the average collagen fibril diameter, which was 35 nm in STI571-treated mice and 45 nm in control carcinoma (Fig. 3C; P < 0.029). These results show that STI571 alters the collagen fibril thickness and morphology, indicating an effect on cell signaling-regulated collagen fibril assembly.

Figure 3.

Ultrastructure of stromal carcinoma collagen fibers. A, micrographs obtained by TEM of KAT-4/HT-29 carcinomas from STI571 and control mice, depicting morphology of collagen fibrils. Note, the higher abundance of smaller fibrils in tumor from STI571-treated mice. B, histogram showing distribution of collagen fibril diameters in STI571 (black bars) and control carcinomas. The two populations differed significantly when tested with the Kolmogrov–Smirnov test (P < 0.0001). C, average fibril diameters from individual carcinoma in the STI571-treated (n = 8) and control (n = 8) groups. **, P < 0.05 by the Student t test; error bars, SD.

Figure 3.

Ultrastructure of stromal carcinoma collagen fibers. A, micrographs obtained by TEM of KAT-4/HT-29 carcinomas from STI571 and control mice, depicting morphology of collagen fibrils. Note, the higher abundance of smaller fibrils in tumor from STI571-treated mice. B, histogram showing distribution of collagen fibril diameters in STI571 (black bars) and control carcinomas. The two populations differed significantly when tested with the Kolmogrov–Smirnov test (P < 0.0001). C, average fibril diameters from individual carcinoma in the STI571-treated (n = 8) and control (n = 8) groups. **, P < 0.05 by the Student t test; error bars, SD.

Close modal

No effect of STI571 on the amount or structure of GAGs in tumors.

GAGs were extracted and purified from tumors. Chondroitin sulfate (CS)/dermatan sulfate (DS), hyaluronan (HA), or heparan sulfate (HS) fluorescent-labeled disaccharides were obtained after chondroitinase ABC or a mixture of heparinases treatment, respectively, and quantified after HPLC separation. No differences were seen in the amounts of CS/DS, HA, or HS between STI571-treated or control tumors (coefficient of variations ranged from 9% to 19%; Fig. 4A). In KAT-4/HT-29 carcinomas, the high level of HS, representing 33% of total GAGs, most likely reflects a high number of HS-expressing carcinoma cells (31). No differences in the structure of the CS/DS or HS chains between STI571-treated or control tumors could be detected (Table 1).

Figure 4.

GAGs, collagen content, and synthesis in KAT-4/HT-29 carcinomas. A, GAGs were extracted from lyophilized STI571-treated carcinomas (n = 5) and from control carcinomas (n = 4), purified, and treated with chondroitinase ABC or a mixture of heparinase I, II, III to quantitatively degrade CS/DS and HA, or HS, respectively. The obtained disaccharides were fluorescence labeled, separated by HPLC, and quantified using defined disaccharide standards. B, hydroxyproline concentration in tumors from STI571-treated for 4 days (n = 7) and PBS control (n = 11); error bars, SD. C, quantification of metabolically labeled collagen α1(I)-chains from autoradiographs. Collagen was extracted by pepsin-treatment from tissue sections incubated ex vivo for 6 hours. Data are obtained from 9 STI571- and 9 PBS-treated carcinomas. Band intensities were normalized to tissue wet weight at harvest. *, P < 0.03 by the Student t test.

Figure 4.

GAGs, collagen content, and synthesis in KAT-4/HT-29 carcinomas. A, GAGs were extracted from lyophilized STI571-treated carcinomas (n = 5) and from control carcinomas (n = 4), purified, and treated with chondroitinase ABC or a mixture of heparinase I, II, III to quantitatively degrade CS/DS and HA, or HS, respectively. The obtained disaccharides were fluorescence labeled, separated by HPLC, and quantified using defined disaccharide standards. B, hydroxyproline concentration in tumors from STI571-treated for 4 days (n = 7) and PBS control (n = 11); error bars, SD. C, quantification of metabolically labeled collagen α1(I)-chains from autoradiographs. Collagen was extracted by pepsin-treatment from tissue sections incubated ex vivo for 6 hours. Data are obtained from 9 STI571- and 9 PBS-treated carcinomas. Band intensities were normalized to tissue wet weight at harvest. *, P < 0.03 by the Student t test.

Close modal
Table 1.

Compositional analysis of CS/DS and HS chains

PBSSTI571
CS/DS 
 ΔUA-GalNAc (ΔO) 16.1 ± 1.0 15.7 ± 1.4 
 ΔUA-GalNAc-4S (ΔA) 80.2 ± 1.1 80.4 ± 1.4 
 ΔUA-GalNAc-6S (ΔC) 1.8 ± 0.1 1.9 ± 0.2 
 ΔUA-2S-GalNAc-4S (ΔB) 1.9 ± 0.2 2.0 ± 0.2 
 ΔUA-2S-GalNAc-6S (ΔD) <0.1 <0.1 
 ΔUA-GalNAc-4S,6S (ΔE) 0.0 0.0 
HS 
 ΔUA-GlcNAc 54.3 ± 0.2 54.9 ± 0.3 
 ΔUA-GlcNS 21.4 ± 0.1 21.6 ± 0.1 
 ΔUA-GlcNAc-6S 7.1 ± 0.1 7.0 ± 0.3 
 ΔUA,2S-GlcNAc 2.8 ± 0.1 2.7 ± 0.1 
 ΔUA-GlcNS-6S 3.0 ± 0.0 2.9 ± 0.2 
 ΔUA-2S-GlcNS 7.7 ± 0.2 7.5 ± 0.2 
 ΔUA,2S-GlcNAc,6S 1.3 ± 0.2 1.3 ± 0.1 
 ΔUA-2S-GlcNS-6S 2.5 ± 0.2 2.2 ± 0.0 
PBSSTI571
CS/DS 
 ΔUA-GalNAc (ΔO) 16.1 ± 1.0 15.7 ± 1.4 
 ΔUA-GalNAc-4S (ΔA) 80.2 ± 1.1 80.4 ± 1.4 
 ΔUA-GalNAc-6S (ΔC) 1.8 ± 0.1 1.9 ± 0.2 
 ΔUA-2S-GalNAc-4S (ΔB) 1.9 ± 0.2 2.0 ± 0.2 
 ΔUA-2S-GalNAc-6S (ΔD) <0.1 <0.1 
 ΔUA-GalNAc-4S,6S (ΔE) 0.0 0.0 
HS 
 ΔUA-GlcNAc 54.3 ± 0.2 54.9 ± 0.3 
 ΔUA-GlcNS 21.4 ± 0.1 21.6 ± 0.1 
 ΔUA-GlcNAc-6S 7.1 ± 0.1 7.0 ± 0.3 
 ΔUA,2S-GlcNAc 2.8 ± 0.1 2.7 ± 0.1 
 ΔUA-GlcNS-6S 3.0 ± 0.0 2.9 ± 0.2 
 ΔUA-2S-GlcNS 7.7 ± 0.2 7.5 ± 0.2 
 ΔUA,2S-GlcNAc,6S 1.3 ± 0.2 1.3 ± 0.1 
 ΔUA-2S-GlcNS-6S 2.5 ± 0.2 2.2 ± 0.0 

NOTE: Data are expressed as the mole percentage of the disaccharide units (mean ± SD of duplicates).

STI571 treatment did not significantly affect total collagen content.

The observed decrease in fibril diameter in KAT-4/HT-29 tumors from STI571-treated mice could reflect a decreased collagen content. Hydroxyproline measurements indicated a tendency to this effect in STI571-treated tumors; the decrease, however, was not statistically significant (Fig. 4B). A similar trend but no significant difference in hydroxyproline content was observed in carcinomas grown in animals treated with STI571 for 8 days compared with vehicle-treated tumors (1.13 ± 0.52 mg/g wet weight, n = 4, P > 0.2 vs. pooled controls).

STI571 treatment significantly increased collagen type I synthesis, but did not affect collagen gene expression

The KAT-4/HT-29 carcinoma is rich in ECM and has a well-developed stromal compartment (8, 19, 32). Treatment with STI571 had no discernable effect on the gross appearance of the collagenous matrix as judged from stainings with Sirius red (Fig. 5A). Gene expression of collagens, small leucine-rich proteoglycans, and enzymes involved in posttranslational modifications of collagens varied largely in KAT-4/HT-29 carcinomas (Supplementary Table S1). STI571 treatment had no significant effect on expression levels of most transcripts, except for a significantly increased expression of the Lox and Loxl2 genes. Next, we investigated whether STI571-treatment affected collagen synthesis. Metabolic labeling ex vivo for 6 hours of freshly isolated thick sections from STI571- and vehicle-treated carcinomas, revealed a significantly increased de novo synthesis of pepsin-resistant collagen type I in STI571-treated carcinoma (Fig. 4C and Supplementary Fig. S1). The amounts of pepsin-resistant collagen from the sections did not differ between STI571- and vehicle-treated carcinomas when normalized to tissue wet weights (Supplementary Fig. S1).

Figure 5.

Effects of STI571 on stroma characteristics in KAT-4/HT-29 carcinoma. A, gross morphology showing Sirius red staining of thick sections (30 μm) showing an abundant collagen matrix in the stroma of KAT-4/HT-29 with no discernable effect by STI571 treatment. B, epifluorescence images of tumors from STI571-treated and control mice, depicting CD31-positive endothelium (red) and NG2-positive (green) mural pericytes and stromal cells. C, each of the investigated markers were quantified and displayed as a percentage of the whole image area containing positive staining. C, quantification of vascular coverage by NG2-positive areas. Data are shown as the percentage of NG2- or CD31-positive pixels that colocalized with the other marker, CD31 or NG2, respectively, in tumors from STI571- and PBS-treated tumors. *, P < 0.05 by the Mann–Whitney test; error bars, SEM; scale bars, 50 μm in A and B.

Figure 5.

Effects of STI571 on stroma characteristics in KAT-4/HT-29 carcinoma. A, gross morphology showing Sirius red staining of thick sections (30 μm) showing an abundant collagen matrix in the stroma of KAT-4/HT-29 with no discernable effect by STI571 treatment. B, epifluorescence images of tumors from STI571-treated and control mice, depicting CD31-positive endothelium (red) and NG2-positive (green) mural pericytes and stromal cells. C, each of the investigated markers were quantified and displayed as a percentage of the whole image area containing positive staining. C, quantification of vascular coverage by NG2-positive areas. Data are shown as the percentage of NG2- or CD31-positive pixels that colocalized with the other marker, CD31 or NG2, respectively, in tumors from STI571- and PBS-treated tumors. *, P < 0.05 by the Mann–Whitney test; error bars, SEM; scale bars, 50 μm in A and B.

Close modal

STI571 treatment did not affect endothelial staining and pericyte coverage, but reduced an NG2-positive extravascular cell population in the tumor stroma

Sections from KAT-4/HT-29 carcinomas grown in STI571-treated and control mice were stained for the endothelial marker CD31, α-SMA, RFM, macrophages (F4/80) and the pericyte marker NG2 (Fig. 5A). STI571 treatment had no apparent effect on the number of cells expressing RFM and α-SMA (Fig. 5B), that is, cells that could be both mural pericytes (33) and stromal myofibroblasts (34, 35). STI571 induced a decreased number of F4/80-positive macrophages (Fig. 5B); this reduction, however, was not statistically significant. Total NG2 staining was significantly decreased (P < 0.05) by STI571 treatment (Fig. 5A and B). Quantification of pixels representing blood vessels (CD31) and NG2-positive pericytes showed that the ratio between colocalized CD31 and NG2 versus CD31 alone was unaltered in STI571-treated tumors, that is, no difference in vascular NG2 positive pericyte-coverage was detected (Fig. 5A and C). Thus, it can be inferred that the reduction in NG2 staining reflects a decrease of extravascular NG2-positive cells or a reduced NG2 expression by such cells.

By taking advantage of dynamic MRI to follow wash-in and wash-out characteristics, we show that treatment of colorectal KAT-4/HT-29 experimental carcinomas with STI571 (Imatinib/Gleevec) resulted in a significant change of the contrast enhancement dynamics—a faster wash-in and a faster wash-out of contrast agent was observed. This provides a strong indication that STI571 treatment induces an augmented exchange between plasma and the carcinoma extracellular fluid volume. In our study, all pixels were analyzed individually. The physical process of contrast enhancement is complicated and the areas close to the vasculature are likely to show biexponential enhancement in our model, that is, wash-in and -out dynamics, whereas the pixels further away from the blood vessels should tend to follow a monoexponential uptake curve, that is, wash-in flow. The contrast agent may diffuse from one pixel to another over time and a linear uptake is likely due to diffusion to areas with monoexponential uptake profiles (36) Consequently, in the comparison between the different uptake curves between control and STI571-treated groups, a change in the biexponential time constants will directly influence the diffusion into the monoexponential areas. Despite the faster wash-out of contrast agent in the STI571-treated animals, a faster monoexponential uptake was evident. This finding strongly indicates that STI71 improves transport of solutes through the stroma interstitium, thereby making it more accessible for blood stream solutes. Our current data, therefore, provide one plausible contributing mechanism for the enhanced efficacy of taxol (18) and uptake and efficiency of Epothilone B (37) after STI571 treatment of KAT-4/HT-29 experimental carcinomas.

STI571 reduces IFP and increases the interstitial fluid volume in KAT-4/HT-29 carcinomas (18, 19). A potential mechanism could involve a reduction of fibroblast tensile forces applied on the collagen/microfibrillar network (1, 18). This is in analogy with what has been described for conditions in normal loose connective tissues in which PDGF normalizes an anaphylaxis-induced lowering of the IFP (38). Our current data suggest an alternative or complementary mechanism that involves an STI571-induced altered collagen network structure with decreased collagen fibril diameters. This potentially results in a more flexible collagen network that allows for the expansion of the interstitial fluid volume and reduction of IFP. Collagenous fibers and the GAG/proteoglycan ground substance constitute the two major macromolecular complexes in any connective tissue and play a decisive role in the efficacy of fluid and solute transport through tissues (12, 39). The finding that the amounts of hyaluronan, chondroitin/dermatan sulfate, and heparan sulfate per unit carcinoma dry weight were unaffected by STI571 treatment, together with the data on an expanded interstitial fluid volume (19), suggest that the GAG/proteoglycan ground substance was diluted by taking up fluid. The collagen content, measured as the total amount of hydroxyproline per unit carcinoma wet weight was not significantly affected by STI571 but the marginal decrease that was recorded fits well with an increase in the interstitial fluid volume. In conclusion, our data suggest that STI571 reprograms stromal connective tissue cells, altering the collagen network allowing for an under-hydrated GAG/proteoglycan ground substance to swell, which in turn reduces the barrier for fluid and solute transport through the carcinoma, as inferred from the present MRI data.

Our data also suggest that the mechanism by which STI571 induces a change in the collagen network involves an increased turnover of collagen type I, indicated by an increased collagen synthesis rate combined with an unchanged total hydroxyproline content. The latter probably is due to an increased degradation of collagen that offsets the increased synthesis. Collagen transcription was not significantly altered in spite of the increased collagen synthesis. This is in agreement with previous data emphasizing the importance of posttranscriptional regulation of collagen type I synthesis (23, 40). An increased turnover of the collagen and impaired maturation of the collagen fibrils is a plausible mechanism for the observed decrease in average collagen fibril diameter. Inhibition of TGF-β1/β3 in KAT-4/HT-29 downregulates the small-leucine rich repeat proteoglycan fibromodulin and the effects on carcinoma IFP and collagen fibrillar network recorded after TGF-β1/β3 inhibition is mimicked in KAT-4/HT-29 carcinomas grown in fibromodulin-deficient mice (8). Fibromodulin regulates collagen fibril assembly in dense connective tissues (41) and is upregulated and plays a pathogenic role in several models of fibrosis (8, 42–44). STI571 showed a trend to downregulate fibromodulin mRNA in KAT-4/HT-29 carcinoma but, due to a large variance, the changes did not reach statistical significance. It is possible that an inability to mature the collagen fibrils results from disturbances in the enzyme arrays responsible for cross-link formation. Recently, we showed that fibromodulin interacts with Lox and can direct collagen cross-link formation (45). A decreased expression of fibromodulin can result in an impairment of fibril maturation even if Lox is increased, as it was observed in the current study in which the expression of Lox and Loxl2 were increased.

STI571 had no effect on the density of CD31-positive vessels in KAT-4/HT-29 carcinoma, consistent with our earlier findings (19). Furthermore, our data show that the pericyte coverage of CD31-positive vessels was not affected by STI571 treatment. Although STI571-treatment does not affect plasma volume in KAT-4/HT-29 carcinomas but has been shown to reduce blood vessel perimeters (19), it cannot be excluded that STI571 increases blood flow through the tumor. Available data have indeed shown that modulation of the ECM-composition, notably the major ECM-constituents collagen and/or glycosaminoglycans, influence solid stress in carcinoma and thereby compression of blood vessels (13, 14, 46). A less dense and more flexible collagen network would enable the opening of vessels and increase blood flow. It is, thus, possible that STI571 by altering the collagen ultrastructure has a dual function that increases drug distribution within carcinoma, by improving transport through the stroma and, potentially, tumor blood flow.

STI571 inhibits several tyrosine kinases, including PDGF receptors, with a proposed pathogenic role in fibrosis (47). Previously, we showed that a PDGF B–specific aptamer mimicked some of the effects of STI571 on IFP and increased sensitivity to taxol in KAT-4/HT-29 experimental carcinomas (18). No reports, however, have previously shown the effects of this agent on collagen network structure reported on herein. Another target of STI571, ABL is activated down-stream of activated PDGF receptors (48). ABL has been ascribed a role in fibrotic reactions acting upstream of early growth response factor 1 (Egr-1; refs. 49, 50). Furthermore, STI571 attenuates the severity of disease in animal models of fibrosis and is currently under clinical testing for the treatment of fibrotic diseases (51). Inhibition of TGF-β1/β3 in KAT-4/HT-29 experimental carcinomas reduces IFP, fibril diameters of stromal collagen and expression of a subset of inflammation-related genes (8, 32). TGF-β stimulates ABL by a non–Smad-dependent pathway (50, 52). It is, thus, possible that the mechanism(s) by which STI571 and TGF-β1/β3 inhibitors exert their effects in KAT-4/HT-29 carcinomas involve inhibition of c-Abl and its downstream target Egr-1.

There was no reduction in the stromal cell populations that stained positive for RFM and α-SMA, or both. The α-SMA–expressing cells in the tumor stroma are often referred to as myofibroblasts, and are regarded as the major collagen type I-synthesizing cells in tumors (34). Our findings that STI571 had no major effect on total hydroxyproline (collagen) content or the number of α-SMA–expressing in STI571-treated tumors are in agreement with that α-SMA–expressing cells are the major collagen-producing cells. It is, however, possible that STI571 altered the expression of collagen assembly regulators by α-SMA–expressing cells resulting in the production of a less dense collagen network. In contrast, STI571 treatment reduced the number of NG2-expressing cells. The amount of these cells colocalizing with CD31-positive cells was, however, unchanged, inferring a decrease in extravascular NG2-expressing cells. Our previous work has shown that overexpression of PDGF-B in murine skin induces the expansion of NG2-positive stromal cells and that pericytes are the likely precursors for these cells (35). It is possible that treatment with STI571, which inhibits PDGF receptors, hampers the expansion of the NG2-positive stromal cell population and that these cells could have a role in modifying the ECM.

Together, our data suggest that STI571 (Imatinib/Gleevec) improves the distribution of low-molecular weight compounds into carcinoma interstitium by altering the collagen ultrastructure, but not the overall content of collagen or glycosaminoglycans. Drugs that specifically alter collagen ultrastructure should be important tools in improving the uptake, and thereby efficacy, of commonly used chemotherapeutics.

No potential conflicts of interest were disclosed.

Conception and design: P.O. Olsson, R. in 't Zandt, T. Friman, M. Maccarana, Å. Oldberg, K. Rubin, S. Kalamajski

Development of methodology: P.O. Olsson, R. Gustafsson, R. in't Zandt, T. Friman, M. Maccarana, K. Rubin

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P.O. Olsson, R. Gustafsson, R. in't Zandt, M. Maccarana, E. Tykesson, K. Rubin, S. Kalamajski

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P.O. Olsson, R. Gustafsson, R. in 't Zandt, T. Friman, M. Maccarana, K. Rubin, S. Kalamajski

Writing, review, and/or revision of the manuscript: P.O. Olsson, R. Gustafsson, R. in 't Zandt, T. Friman, M. Maccarana, Å. Oldberg, K. Rubin, S. Kalamajski

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Gustafsson, K. Rubin

Study supervision: R. Gustafsson, K. Rubin

Lund University Bioimaging Center (LBIC) at Lund University is gratefully acknowledged for providing experimental resources.

This study was supported by funds from the Swedish Cancer Society (to K. Rubin and M. Maccarana), the Swedish Research Council (to K. Rubin), the Alfred Österlund Foundation (to K. Rubin, Å. Oldberg, and S. Kalamajski), the Koch Foundation (to K. Rubin), the Crafoord Foundation (to S. Kalamajski), the Magnus Bergvall Foundation (to S. Kalamajski), the Åke Wiberg Foundation (to S. Kalamajski), and Uppsala and Lund Universities.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Heldin
CH
,
Rubin
K
,
Pietras
K
,
Östman
A
. 
High interstitial fluid pressure—an obstacle in cancer therapy
.
Nat Rev Cancer
2004
;
4
:
806
13
.
2.
Jain
RK
. 
Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy
.
Science
2005
;
307
:
58
62
.
3.
Minchinton
AI
,
Tannock
IF
. 
Drug penetration in solid tumours
.
Nat Rev Cancer
2006
;
6
:
583
92
.
4.
Salnikov
AV
,
Iversen
VV
,
Koisti
M
,
Sundberg
C
,
Johansson
L
,
Stuhr
LB
, et al
Lowering of tumor interstitial fluid pressure specifically augments efficacy of chemotherapy
.
FASEB J
2003
;
17
:
1756
8
.
5.
Bertino
P
,
Piccardi
F
,
Porta
C
,
Favoni
R
,
Cilli
M
,
Mutti
L
, et al
Imatinib mesylate enhances therapeutic effects of gemcitabine in human malignant mesothelioma xenografts
.
Clin Cancer Res
2008
;
14
:
541
8
.
6.
Fan
Y
,
Du
W
,
He
B
,
Fu
F
,
Yuan
L
,
Wu
H
, et al
The reduction of tumor interstitial fluid pressure by liposomal imatinib and its effect on combination therapy with liposomal doxorubicin
.
Biomaterials
2013
;
34
:
2277
88
.
7.
Eikenes
L
,
Bruland
OS
,
Brekken
C
,
Davies Cde
L
. 
Collagenase increases the transcapillary pressure gradient and improves the uptake and distribution of monoclonal antibodies in human osteosarcoma xenografts
.
Cancer Res
2004
;
64
:
4768
73
.
8.
Oldberg
Å
,
Kalamajski
S
,
Salnikov
AV
,
Stuhr
L
,
Morgelin
M
,
Reed
RK
, et al
Collagen-binding proteoglycan fibromodulin can determine stroma matrix structure and fluid balance in experimental carcinoma
.
Proc Natl Acad Sci U S A
2007
;
104
:
13966
71
.
9.
Gade
TP
,
Buchanan
IM
,
Motley
MW
,
Mazaheri
Y
,
Spees
WM
,
Koutcher
JA
. 
Imaging intratumoral convection: pressure-dependent enhancement in chemotherapeutic delivery to solid tumors
.
Clin Cancer Res
2009
;
15
:
247
55
.
10.
Friman
T
,
Gustafsson
R
,
Stuhr
LB
,
Chidiac
J
,
Heldin
NE
,
Reed
RK
, et al
Increased fibrosis and interstitial fluid pressure in two different types of syngeneic murine carcinoma grown in integrin β3-subunit deficient mice
.
PLoS ONE
2011
;
7
:
e34082
.
11.
Torosean
S
,
Flynn
B
,
Axelsson
J
,
Gunn
J
,
Samkoe
KS
,
Hasan
T
, et al
Nanoparticle uptake in tumors is mediated by the interplay of vascular and collagen density with interstitial pressure
.
Nanomedicine
2013
;
9
:
151
8
.
12.
Reed
RK
,
Rubin
K
. 
Transcapillary exchange: role and importance of the interstitial fluid pressure and the extracellular matrix
.
Cardiovasc Res
2010
;
87
:
211
7
.
13.
Provenzano
PP
,
Cuevas
C
,
Chang
AE
,
Goel
VK
,
Von Hoff
DD
,
Hingorani
SR
. 
Enzymatic targeting of the stroma ablates physical barriers to treatment of pancreatic ductal adenocarcinoma
.
Cancer Cell
2012
;
21
:
418
29
.
14.
Olive
KP
,
Jacobetz
MA
,
Davidson
CJ
,
Gopinathan
A
,
McIntyre
D
,
Honess
D
, et al
Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer
.
Science
2009
;
324
:
1457
61
.
15.
Buchdunger
E
,
Zimmermann
J
,
Mett
H
,
Meyer
T
,
Muller
M
,
Druker
BJ
, et al
Inhibition of the Abl protein-tyrosine kinase in vitro and in vivo by a 2-phenylaminopyrimidine derivative
.
Cancer Res
1996
;
56
:
100
4
.
16.
Buchdunger
E
,
Cioffi
CL
,
Law
N
,
Stover
D
,
Ohno-Jones
S
,
Druker
BJ
, et al
Abl protein-tyrosine kinase inhibitor STI571 inhibits in vitro signal transduction mediated by c-kit and platelet-derived growth factor receptors
.
J Pharmacol Exp Ther
2000
;
295
:
139
45
.
17.
Manley
PW
,
Stiefl
N
,
Cowan-Jacob
SW
,
Kaufman
S
,
Mestan
J
,
Wartmann
M
, et al
Structural resemblances and comparisons of the relative pharmacological properties of imatinib and nilotinib
.
Bioorg Med Chem
2010
;
18
:
6977
86
.
18.
Pietras
K
,
Rubin
K
,
Sjöblom
T
,
Buchdunger
E
,
Sjöquist
M
,
Heldin
CH
, et al
Inhibition of PDGF receptor signaling in tumor stroma enhances antitumor effect of chemotherapy
.
Cancer Res
2002
;
62
:
5476
84
.
19.
Klosowska-Wardega
A
,
Hasumi
Y
,
Burmakin
M
,
Åhgren
A
,
Stuhr
L
,
Moen
I
, et al
Combined anti-angiogenic therapy targeting PDGF and VEGF receptors lowers the interstitial fluid pressure in a murine experimental carcinoma
.
PLoS ONE
2009
;
4
:
e8149
.
20.
Lubberink
M
,
Golla
SS
,
Jonasson
M
,
Rubin
K
,
Glimelius
B
,
Sorensen
J
, et al
15O-Water PET Study of the effect of imatinib, a selective platelet-derived growth factor receptor inhibitor, Versus Anakinra, an IL-1R antagonist, on water-perfusable tissue fraction in colorectal cancer metastases
.
J Nucl Med
2015
;
56
:
1144
9
.
21.
Ain
KB
,
Taylor
KD
. 
Somatostatin analogs affect proliferation of human thyroid carcinoma cell lines invitro
.
J Clin Endocrinol Metab
1994
;
78
:
1097
102
.
22.
Schweppe
RE
,
Klopper
JP
,
Korch
C
,
Pugazhenthi
U
,
Benezra
M
,
Knauf
JA
, et al
Deoxyribonucleic acid profiling analysis of 40 human thyroid cancer cell lines reveals cross-contamination resulting in cell line redundancy and misidentification
.
J Clin Endocrinol Metab
2008
;
93
:
4331
41
.
23.
Dahlman
T
,
Lammerts
E
,
Wik
M
,
Bergström
D
,
Grimelius
L
,
Westermark
K
, et al
Fibrosis in undifferentiated (anaplastic) thyroid carcinomas: evidence for a dual action of tumour cells in collagen type I synthesis
.
J Pahol
2000
;
191
:
376
86
.
24.
Jolliffe
IT
.
Principal component analysis
. 2 ed.
New York, NY
:
Springer-Verlag
; 
2002
.
25.
Hassid
Y
,
Eyal
E
,
Margalit
R
,
Furman-Haran
E
,
Degani
H
. 
Non-invasive imaging of barriers to drug delivery in tumors
.
Microvasc Res
2008
;
76
:
94
103
.
26.
Berg
RA
. 
Determination of 3- and 4-hydroxyproline
.
Methods Enzymol
1982
;
82
:
372
98
.
27.
Stachtea
XN
,
Tykesson
E
,
van Kuppevelt
TH
,
Feinstein
R
,
Malmstrom
A
,
Reijmers
RM
, et al
Dermatan sulfate-free mice display embryological defects and are neonatal lethal despite normal lymphoid and non-lymphoid organogenesis
.
PLoS ONE
2015
;
10
:
e0140279
.
28.
Ohtani
O
,
Ushiki
T
,
Taguchi
T
,
Kikuta
A
. 
Collagen fibrillar networks as skeletal frameworks: a demonstration by cell-maceration/scanning electron microscope method
.
Arch Histol Cytol
1988
;
51
:
249
61
.
29.
Aime
S
,
Caravan
P
. 
Biodistribution of gadolinium-based contrast agents, including gadolinium deposition
.
J Magn Reson Imaging
2009
;
30
:
1259
67
.
30.
Di Gregorio
E
,
Gianolio
E
,
Stefania
R
,
Barutello
G
,
Digilio
G
,
Aime
S
. 
On the fate of MRI Gd-based contrast agents in cells. Evidence for extensive degradation of linear complexes upon endosomal internalization
.
Anal Chem
2013
;
85
:
5627
31
.
31.
Li
G
,
Li
L
,
Tian
F
,
Zhang
L
,
Xue
C
,
Linhardt
RJ
. 
Glycosaminoglycanomics of cultured cells using a rapid and sensitive LC-MS/MS approach
.
ACS Chem Biol
2015
;
10
:
1303
10
.
32.
Salnikov
AV
,
Roswall
P
,
Sundberg
C
,
Gardner
H
,
Heldin
NE
,
Rubin
K
. 
Inhibition of TGF-β modulates macrophages and vessel maturation in parallel to a lowering of interstitial fluid pressure in experimental carcinoma
.
Lab Invest
2005
;
85
:
512
21
.
33.
Sundberg
C
,
Kowanetz
M
,
Brown
LF
,
Detmar
M
,
Dvorak
HF
. 
Stable expression of angiopoietin-1 and other markers by cultured pericytes: phenotypic similarities to a subpopulation of cells in maturing vessels during later stages of angiogenesis invivo
.
Lab Invest
2002
;
82
:
387
401
.
34.
Hinz
B
,
Phan
SH
,
Thannickal
VJ
,
Galli
A
,
Bochaton-Piallat
ML
,
Gabbiani
G
. 
The myofibroblast: one function, multiple origins
.
Am J Pathol
2007
;
170
:
1807
16
.
35.
Rodriguez
A
,
Friman
T
,
Kowanetz
M
,
van Wieringen
T
,
Gustafsson
R
,
Sundberg
C
. 
Phenotypical differences in connective tissue cells emerging from microvascular pericytes in response to overexpression of PDGF-B and TGF-β1 in normal skin invivo
.
Am J Pathol
2013
;
182
:
2132
46
.
36.
van der Sanden
BP
,
in 't Zandt
HJ
,
Hoofd
L
,
de Graaf
RA
,
Nicolay
K
,
Rijken
PF
, et al
Global HDO uptake in human glioma xenografts is related to the perfused capillary distribution
.
Magn Reson Med
1999
;
42
:
479
89
.
37.
Pietras
K
,
Stumm
M
,
Hubert
M
,
Buchdunger
E
,
Rubin
K
,
Heldin
CH
, et al
STI571 enhances the therapeutic index of epothilone B by a tumor-selective increase of drug uptake
.
Clin Cancer Res
2003
;
9
:
3779
87
.
38.
Rodt
,
Åhlen
K
,
Berg
A
,
Rubin
K
,
Reed
RK
. 
A novel physiological function for platelet-derived growth factor-BB in rat dermis
.
J Physiol
1996
;
495
:
193
200
.
39.
Levick
JR
. 
Flow through interstitium and other fibrous matrices
.
Q J Exp Physiol
1987
;
72
:
409
37
.
40.
Ivarsson
M
,
McWhirter
A
,
Borg
TK
,
Rubin
K
. 
Type I collagen synthesis in cultured human fibroblasts: regulation by cell spreading, platelet-derived growth factor and interactions with collagen fibers
.
Matrix Biol
1998
;
16
:
409
25
.
41.
Kalamajski
S
,
Oldberg
Å
. 
The role of small leucine-rich proteoglycans in collagen fibrillogenesis
.
Matrix Biol
2010
;
29
:
248
53
.
42.
Mormone
E
,
Lu
Y
,
Ge
X
,
Fiel
MI
,
Nieto
N
. 
Fibromodulin, an oxidative stress-sensitive proteoglycan, regulates the fibrogenic response to liver injury in mice
.
Gastroenterology
2012
;
142
:
612
21
.
43.
Shami
A
,
Gustafsson
R
,
Kalamajski
S
,
Krams
R
,
Segers
D
,
Rauch
U
, et al
Fibromodulin deficiency reduces low-density lipoprotein accumulation in atherosclerotic plaques in apolipoprotein E-null mice
.
Arterioscler Thromb Vasc Biol
2013
;
33
:
354
61
.
44.
Rydell-Törmänen
K
,
Andreasson
K
,
Hesselstrand
R
,
Westergren-Thorsson
G
. 
Absence of fibromodulin affects matrix composition, collagen deposition and cell turnover in healthy and fibrotic lung parenchyma
.
Sci Rep
2014
;
4
:
6383
.
45.
Kalamajski
S
,
Bihan
D
,
Bonna
A
,
Rubin
K
,
Farndale
RW
. 
Fibromodulin interacts with collagen cross-linking sites and activates lysyl oxidase
.
J Biol Chem
2016
;
291
:
7951
60
.
46.
Chauhan
VP
,
Martin
JD
,
Liu
H
,
Lacorre
DA
,
Jain
SR
,
Kozin
SV
, et al
Angiotensin inhibition enhances drug delivery and potentiates chemotherapy by decompressing tumour blood vessels
.
Nat Commun
2013
;
4
:
2516
.
47.
Trojanowska
M
. 
Role of PDGF in fibrotic diseases and systemic sclerosis
.
Rheumatology
2008
;
47
:
v2
4
.
48.
Srinivasan
D
,
Kaetzel
DM
,
Plattner
R
. 
Reciprocal regulation of Abl and receptor tyrosine kinases
.
Cell Signal
2009
;
21
:
1143
50
.
49.
Karimizadeh
E
,
Motamed
N
,
Mahmoudi
M
,
Jafarinejad-Farsangi
S
,
Jamshidi
A
,
Faridani
H
, et al
Attenuation of fibrosis with selective inhibition of c-Abl by siRNA in systemic sclerosis dermal fibroblasts
.
Arch Dermatol Res
2015
;
307
:
135
42
.
50.
Bhattacharyya
S
,
Ishida
W
,
Wu
M
,
Wilkes
M
,
Mori
Y
,
Hinchcliff
M
, et al
A non-Smad mechanism of fibroblast activation by transforming growth factor-β via c-Abl and Egr-1: selective modulation by imatinib mesylate
.
Oncogene
2009
;
28
:
1285
97
.
51.
Grimminger
F
,
Schermuly
RT
,
Ghofrani
HA
. 
Targeting non-malignant disorders with tyrosine kinase inhibitors
.
Nat Rev Drug Discov
2010
;
9
:
956
70
.
52.
Bhattacharyya
S
,
Fang
F
,
Tourtellotte
W
,
Varga
J
. 
Egr-1: new conductor for the tissue repair orchestra directs harmony (regeneration) or cacophony (fibrosis)
.
J Pathol
2013
;
229
:
286
97
.