Assessment of chemotherapy resistance changes in human colorectal cancer xenografts in rats based on MRI histogram features

Purpose We investigated the value of magnetic resonance imaging (MRI) histogram features, a non-invasive method, in assessing the changes in chemoresistance of colorectal cancer xenografts in rats. Methods A total of 50 tumor-bearing mice with colorectal cancer were randomly divided into two groups: control group and 5-fluorouracil (5-FU) group. The MRI histogram characteristics and the expression levels of p53 protein and MRP1 were obtained at 24 h, 48 h, 72 h, 120 h, and 168 h after treatment. Results Sixty highly repeatable MRI histogram features were obtained. There were 16 MRI histogram parameters and MRP1 resistance protein differences between groups. At 24 h after treatment, the MRI histogram texture parameters of T2-weighted imaging (T2WI) images (10%, 90%, median, energy, and RootMeanSquared) and D images (10% and Range) were positively correlated with MRP1 (r = 0.925, p = 0.005). At 48 h after treatment, histogram texture parameters of apparent diffusion coefficient (ADC) images (Energy) were positively correlated with the presence of MRP1 resistance protein (r = 0.900, p = 0.037). There was no statistically significant difference between MRI histogram features and p53 protein expression level. Conclusions MRI histogram texture parameters based on T2WI, D, and ADC maps can help to predict the change of 5-FU resistance in colorectal cancer in the early stage and provide important reference significance for clinical treatment.


Introduction
Colorectal cancer (CRC) is one of the most common malignant tumors in humans (1)(2)(3).5-Fluorouracil (5-FU)-based chemotherapy palliative treatment has become an important means of treatment.However, the tumor cells may develop drug resistance quickly and easily in the course of chemotherapy, making the progression-free survival of patients only 8.7-12.3months (4,5).Clinicians need to adjust treatment regimens based on tumor drug resistance, which is an important step in guiding treatment.Currently, the detection of chemotherapy resistance in tumors is mainly performed through invasive laboratory tests, such as biopsy or postoperative pathology.However, due to the differences in sampling sites, the samples cannot fully reflect the drug resistance of tumor cells and are not always well accepted by patients (6).Therefore, it is urgent to find a relatively accurate and comprehensive non-invasive method for drug resistance assessment of colorectal cancer.The nude mouse model of colorectal cancer transplantation can well reflect the pathological characteristics of human colorectal cancer, and it is well controlled in the experiment (7).
The magnetic resonance imaging (MRI) histogram feature is becoming more and more popular among scholars (8,9).It plays an important role in predicting the outcome of colorectal cancer treatment.First, it can detect tissue changes that are not easily detected by the naked eye by quantifying the relationship between gray patterns and pixels in the image, thus making up for the deficiency of traditional image analysis methods.Second, it can well distinguish subtle differences in texture information of images, evaluate tumor heterogeneity, and carry out differential diagnosis, efficacy evaluation, and prognosis prediction of tumors.You J et al. performed preoperative T staging of rectal cancer based on wholetumor texture analysis on MRI (10).Li J et al. found that magnetic resonance diffusion-weighted imaging histogram parameters can be used as biomarkers to predict lymph node metastasis in T3 rectal cancer (11).Importantly, it is a non-invasive test that is easily accepted by patients and can be widely used in clinical practice (12,13).Therefore, the MRI histogram feature may be a good method to detect drug resistance in colorectal cancer.
Multidrug resistance-associated protein 1 (MRP1), a member of multidrug resistance protein MRP, is a glutathione transport pump (GS-X) widely expressed in human tissues.Many studies have found that tumor-associated macrophages (TAMs) in colorectal cancer can confer resistance to 5-FU treatment through MRP1dependent drug efflux mediated by cell-cell interactions (14).p53 gene is located on the short arm of chromosome 17 (17p13) and encodes nucleic acid phosphoprotein (15).Mutant p53 protein can be detected in more than 50% of human tumors and approximately 70% of colorectal cancer cases (16,17).It promotes tumor progression and resistance to therapy, and this mutant form has become the most common prognostic indicator of tumor recurrence and death from cancer (18,19).In this study, we investigated the feasibility of MRI histogram features to predict chemotherapy response in CRC-resistant xenograft mice by analyzing the correlation between MRI histogram features and MRP1 protein and p53 protein expression.

Establishment of animal model
This study used male and female BALB/c nude mice (6-8 weeks old, weighing 20-25 g).All the animals were housed in an environment with a temperature of 22°C ± 1°C, relative humidity of 50% ± 1%, and a light/dark cycle of 12/12 h and had free access to standard water and food (SYXK-2016-0168).All procedures were conducted in accordance with the "Guiding Principles for the Care and Use of Animals" (China) and were approved by the Institutional Animal Care and Use Committee of Guangzhou Medical University (GY2017-007).
Colorectal cancer (SW480) tissue was implanted subcutaneously in the back of nude mice to establish a xenograft model (20).All procedures were performed under the induction of chloral hydrate (4.5% chloral hydrate, 2 mL/100 g body weight, i.p.) to minimize pain.Mice were maintained in a specific pathogen-free (SPF) laboratory.When the tumor diameter was approximately 1.5 cm, the mice were randomly divided into two groups: control group and 5-FU group.The dose of 5-FU and normal saline was as follows: 25 mg in tumor diameter of 0.5-0.9cm, 50 mg in tumor diameter of 1.0-1.4cm, and 75 mg in tumor diameter above 1.5 cm.The route of administration was direct intra-tumoral injection.

Western blotting analysis
Western blotting analysis was used to detect the protein expression of p53 and MRP1, as described in the manufacturer's package (Phototope ® -HRP Western Blot Kit; Cell Signaling Technology, Danvers, MA, USA).Cell extracts were obtained from frozen tumor tissue at −30°C.Immunoblot analysis was performed using anti-p53 monoclonal antibodies (Santa Cruz Biotechnology, Dallas, TX, USA) and MRP1 monoclonal antibodies (Santa Cruz Biotechnology).Subsequent protein detection was performed using an enhanced chemiluminescence (ECL) detection system (Hitachi, Tokyo, Japan).
The band intensities [integrated optical density (IOD)] of protein expression stated above were scanned into the computer and analyzed using Image-Pro Plus 6.0 software.The relative IOD (RIOD) of protein expression was calculated as the IOD of the protein in the control group and therapeutic groups at each time point divided by the corresponding IOD of GAPDH (internal control).

MRI examination
At 24 h, 48 h, 72 h, 120 h, and 168 h after 5-FU treatment, five tumor-bearing mice in each group were randomly selected for MRI examination.MRI scan was performed using Siemens MAGNETOM Verio 3.0T superconducting magnetic resonance imaging system and Shanghai Chenguang 8-channel phased-array animal coil.The MRI sequence included the following: 1) T1weighted imaging (T1WI) (turbo spin echo (TSE)) sequence,

Data analysis
After the MRI examination, the original data were analyzed using the picture archiving and communication system (PACS) system.The intravoxel incoherent motion-diffusion-weighted imaging (IVIM-DWI) was imported into the MITK Diffusion on analysis software in Digital Imaging and Communications in Medicine (DICOM) format, and the bi-exponential model parameter map was obtained, including a true diffusion coefficient map (D map), false diffusion coefficient map (D* map), perfusion coefficient map (f map), DWI map, and apparent diffusion coefficient map (ADC map).

MR histogram texture feature acquisition
The region of interest (ROI) was delineated by three radiologists.The largest tumor layer was delineated as ROI, along the outer edge of the tumor, and an attempt was made to include the entire tumor range (Figures 1A-F).The MRI histogram features of the tumor were obtained using The MR Station software.A total of 108 MRI histogram texture features were obtained.MRP1 between the control group and the 5-FU group.Finally, the correlation of MRI histogram texture parameters with the p53 protein and drug resistance protein at five time points after 5-FU administration was analyzed.p < 0.05 was considered statistically significant.

Results
3.1 MRI histogram texture parameters with high repeatability (ICC ≥ 0.75) Of the 108 MRI histogram texture parameters, 60 had high repeatability and were analyzed.The T2WI maps showed nine MRI histogram texture parameters (ICC = 0.995).The D maps showed 10 histogram parameters (ICC = 0.985).The D* maps showed 10 histogram parameters (ICC = 0.978).The f maps showed 12 histogram parameters (ICC = 0.985).The DWI maps showed 11 histogram parameters (ICC = 0.996).The ADC maps s h o w e d e i g h t hi s t o g r a m p a r a m e t e r s ( I C C = 0 .9 7 8 ) (Figure 2) (Table 1).

Difference analysis of the MRI histogram texture parameters between the 5-FU group and control group (Table 2)
Between the 5-FU group and the control group, there were differences in the six MRI histogram texture parameters of 10%, 90%, Median, Energy, Robust-Mean-Abs-Dev, and RootMeanSquared in T2WI maps, and the difference in Energy was the most obvious (p = 0.001).There were differences in the four MRI histogram texture parameters of 10%, Range, Entropy, and RootMeanSquared in the D maps, and the difference in Energy was the most obvious (p = 0.002).There were differences in the five MRI histogram texture parameters of 10%, Range, Entropy, Energy, Max, and Interquartile in DWI maps, and the difference in Entropy was the most obvious (p = 0.005).There were differences in energy in the D* maps and ADC maps (p = 0.002 and p = 0.034, respectively) (Table 2).

Difference analysis of p53 protein and MRP1 resistance protein between the 5-FU group and the control group (Table 3)
In this study, the differences in the expression levels of p53 protein and MRP1 between the 5-FU group and the control group were detected.The expression of MRP1 was significantly different between the two groups (p = 0.000).The expression levels of p53 protein were not different between the two groups (Table 3).Frontiers in Oncology frontiersin.orgsignificant positive correlation with MRP1 (r = 0.975, p = 0.005).At 48 h after 5-FU administration, the MR histogram texture Energy of the ADC map showed a significant positive correlation with MRP1 (r = 0.900, p = 0.03).There was no correlation between the MR texture parameters and the expression of MRP1 at 72 h, 120 h, and 168 h after 5-FU administration (Figure 3).

Discussion
MRP1 can transport chemotherapeutic drugs coupled with reduced glutathione against the concentration gradient and reduce the concentration of chemotherapeutic drugs in tumor cells, thus allowing tumor cells to develop drug resistance.The high expression of MRP1 is one of the important causes of multidrug resistance in colorectal cancer cells.Therefore, we can effectively predict the generation of drug resistance by detecting the expression of MRP1 in tumor cells.
MRI histogram features is a non-invasive imaging technology.It uses mathematical methods to quantify the grayscale, spatial distribution, structure, and other information of MRI image pixels, allowing to extraction of image features of lesions.It accurately reflects the difference in image voxels, evaluates tumor heterogeneity, and performs differential diagnosis, efficacy evaluation, and prognosis prediction of tumors (21,22).These texture features significantly expand the image information and application value of tumors and have high repeatability (23).Previous studies have also shown that MR histogram features can provide valuable medical imaging information in identifying colorectal cancer malignancy, T staging, and efficacy prediction (24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34).
Based on the above clinical issues and theoretical principles, our study extracted MRI histogram features in MRI multi-parameter  In this study, the time points of 24 h, 48 h, 72 h, 120 h, and 168 h after injection were statistically analyzed.The results showed that there was no correlation between MRI histogram texture and p53 protein over time.There were the following two possible reasons: the first possible reason was that histogram texture can only accurately detect changes in drug resistance over a specific time period, and p53 protein expression was not evident in this time period.The second possible reason was that our study was too short and the correlation between the two had not yet been reflected.reference for the adjustment of therapeutic drugs for patients with colorectal cancer and has positive significance.

1
FIGURE 1Nude mouse transplanted tumors in different MRI sequences (A-F).

5 Conclusion
MRI histogram texture can predict the change of drug resistance in colorectal cancer in the early stage.It can provide a

FIGURE 3
FIGURE 3    Correlation analysis between MRI histogram texture parameters and MRP1 expression level in 5-fluorouracil (5-FU) group at different time points.

TABLE 2
(34)erence analysis of MRI histogram texture parameters between 5-FU group and the control group.DWI, ADC, D, D*, and f) to analyze the relationship between MRI histogram features and the expression levels of MRP1 protein at different time points in human colorectal cancer nude mouse transplanted tumors after 5-FU treatment.The results show that the texture features of the MRI histogram (10%, 90%, Median, Energy, RootMeanSquared, Range, and Energy) are highly correlated with MRP1 expression.They describe the tumor heterogeneity, which reflects high cell density, necrosis, hemorrhage, and myxoid changes within the tumor area.Tumor heterogeneity is an important factor for prognosis.Song et al. found that the texture features of T2WI can provide valuable information for identifying the lymph node infiltration status of rectal cancer(34).In our study, it was found that the MRI histogram features of 10%, 90%, Median, Energy, and RootMeanSquared in T2WI images of colorectal transplanted tumors were highly correlated with MRP1, which could support the study results of Song et al. from another aspect.ADC and D can further reflect tumor tissue proliferation and metabolism, cell density, and blood supply (35-37).In our study, it was found that 10% of D, Range MRI histogram texture features, and Energy MRI histogram texture features in the ADC map were highly correlated with MRP1.However, our study did not find MRI histogram texture features related to MRP1 at 72 h, 120 h, and 168 h after 5-FU treatment.The possible reason for this is that the MRI histogram texture features can only accurately detect changes in drug resistance at a certain point in time, and after the most appropriate detection time, the correlation between the two will be affected by some unknown reasons.This is what we need to find out in our follow-up studies.

TABLE 2 Continued
*Mann-Whitney U test; the median (upper and lower quartiles).

TABLE 3
Difference analysis of p53 protein and MRP1 resistance protein between 5-FU group and the control group.
*Mann-Whitney U test; the median (upper and lower quartiles).