Method and application for measuring the flux rate of water molecules across cell membranes, method and system for measuring magnetic resonance imaging markers of gliomas - Patents.com

Abstract

The present invention discloses a method for measuring the flux rate of water molecules across a cell membrane, which includes the steps of setting magnetic resonance imaging parameters and measuring the noise level during scanning, optimizing and resetting the flip angle by Monte Carlo simulation, scanning quantitative T1 magnetic resonance imaging, scanning dynamic contrast-enhanced magnetic resonance imaging, injecting contrast agent, and performing full shutter speed model SSM. full Perform pixel analysis for each pixel point in the tumor region using io and obtaining k io The present invention provides a method for evaluating AQP4 expression levels as a magnetic resonance imaging marker for gliomas for non-disease diagnostic purposes. io The present invention discloses the application of k as a magnetic resonance imaging marker of glioma in the preparation of a product for predicting the sensitivity of glioma to radiotherapy and chemotherapy treatment, as well as a method and system for its measurement. io The application, method or system described above provides non-invasive and quantitative measurement and imaging of AQP4 expression levels in gliomas.

Description

[Technical field]

[0001] The present invention relates to the technical field of magnetic resonance imaging, in particular to a method and application for measuring the flux rate of water molecules across cell membranes, and to a method for measuring magnetic resonance imaging markers of gliomas. [Background technology]

[0002] Glioma is the most common central nervous system (CNS) disease in adults, accounting for approximately 60%-70% of primary brain tumors. Recently, many studies have shown that AQP4 (aquaporin 4) plays an important role in glioma cell migration, proliferation, and peritumoral edema, and is an important biomarker for glioma prognosis. Compared with normal brain astrocytes, AQP4 expression in gliomas is significantly increased, and its expression location is redistributed from the edge of astrocytes (perivascular) to the entire cell membrane. Studies have shown that AQP4 expression in various types of glioma cells responds differently to treatment with the anticancer drug temozolomide (TMZ). In short, AQP4 expression level and location changes are one of the early indicators of glioma transformation and treatment resistance.

[0003] At present, traditional biopsy results are the only standard for characterizing AQP4 expression levels in vivo. However, biopsy sampling can only be performed at a single or limited points, making it impossible to obtain the distribution and dynamic information of AQP4 in gliomas, and is an invasive method with a certain degree of risk. Because gliomas are very heterogeneous, one-sided information with sparse sampling may be accompanied by misjudged information. Considering the crucial role of AQP4 expression level and location changes in the fate of gliomas, it is necessary to develop a non-invasive quantitative imaging technique with high spatial resolution of AQP4 expression in vivo. For example, Chinese Patent Application No. CN106683081A provides a radiomics-based non-destructive prediction method and prediction system for glioma molecular marker IDH1. For example, Chinese Patent Application No. CN107169497B provides a method for extracting tumor imaging markers based on gene imaging. CN107169497B uses genetic imaging and combines the advantages of imaging and molecular techniques to invent a non-invasive and highly interpretable biomarker extraction method, which extracts high-dimensional quantitative image features from tumor CT and associates them with corresponding tumor gene expression patterns, and assumes that certain quantitative image features reflect certain gene expression patterns of tumors and can be used as tumor prognostic markers. The ultimate goal is to extract non-invasive and biologically interpretable imaging markers.

[0004] However, the expression concentration of AQP4 in brain tissue is very low (≦[8.65±0.80]ng / ml), so existing molecular imaging techniques such as magnetic resonance spectroscopy (MRS) and chemical exchange saturation transfer (CEST) cannot quantify AQP4 expression. In addition, exogenous contrast agents targeting specific molecules can also be used for MRI molecular imaging. However, currently, no such probes for molecular imaging of AQP4 exist, and even if they were available, the development and approval process of related drugs would be time-consuming and expensive.

[0005] In normal brain astrocytes, AQP4 is the major water channel expressed in the central nervous system and is distributed mainly in the perivascular caudal apex of astrocytes. However, in gliomas, AQP4 is upregulated and redistributed from astrocyte terminals (perivascular) to the entire cell membrane. Aberrant expression of AQP4 aids in glioma brain invasion and is one of the earliest indicators of glioma carcinogenesis. More importantly, AQP4 has been shown to be not only a potential therapeutic target but also a highly sensitive prognostic biomarker for human glioma migration, progression, edema and treatment resistance. Summary of the Invention

[0006] The objective of the present invention is to determine the rate of water flux k across the cell membrane. io The aim of the study is to provide a method for measuring k io The present invention significantly improves the accuracy of k in magnetic resonance imaging markers for gliomas. io The present invention further provides an application for measuring AQP4 expression levels in gliomas. The present invention further provides a method and system for measuring magnetic resonance imaging markers of gliomas, thereby realizing non-invasive and quantitative measurement and imaging of AQP4 expression levels in gliomas. The present invention also ... io This invention provides an application of the present invention, which is useful for predicting treatment response and facilitating precise treatment.

[0007] The present invention uses the following technical solutions:

[0008] The flux rate of water molecules across the cell membrane, k io A method for measuring (1) setting dynamic contrast-enhanced magnetic resonance imaging parameters to measure a noise level during a dynamic contrast-enhanced magnetic resonance scan; (2) optimizing a flip angle in the dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) acquisition parameters by Monte Carlo simulation to reset the dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) flip angle; (3) scanning quantitative T1 magnetic resonance imaging; (4) scanning dynamic contrast-enhanced magnetic resonance imaging and injecting a contrast agent; (5) Full Shutter Speed ​​Model SSM full A pixel analysis was performed for each pixel point in the tumor region using io (The steady-state flux coefficient k io and obtaining the information about the network (also referred to as the network).

[0009] In step (2), the measured vascular contrast agent concentration C p Randomly select one or more datasets from the list, then simulate the real tissue and scanning parameter situations to synthesize DCE-MRI data under different flip angles, and obtain the time series signal of the synthesized DCE-MRI data. JPEG2025505359000002.jpg9146 is the full shutter speed model SSM full Based on the noise level generated by and estimated by the DCE-MRI experiment, JPEG2025505359000003.jpg9146 with white noise of the same noise level added, full shutter speed model SSM full Noise was added using A nonlinear least-squares fitting was performed on JPEG2025505359000004.jpg10155, and the above steps were repeated multiple times under each flip angle. The steady-state flow rate coefficient k of water molecules from the cell to the interstitium was calculated by fitting the steady-state flow rate coefficient k of water molecules from the cell to the interstitium. io Finally, k ioThe flip angle for which the fitting result is closest to the predetermined value of the simulation and the variance is the smallest is the optimal flip angle. It is preferable to repeat the process 100 times or more. Specifically, the full shutter speed model SSM full Please refer to CN201910621579.6.

[0010] In step (2), the full shutter speed model SSM full divides water molecules into three compartments, blood vessels (b), interstitium (o), and intracellular space (i), and divides the water exchange process into two, including blood-interstitium water exchange and cell-cell interstitium water exchange, while the water exchange between blood and intracellular space is negligible. Based on the Gd contrast agent (CA) (Gd-DTPA, gadopentetate dimeglumine), the (MAGNEVIST) in this invention is regarded as an extracellular agent that distributes only in blood vessels and interstitium. The contrast agent concentration [CAo] (T) in the interstitial space is calculated according to the Kety-Schmidt type rate law:

number

[0011] Preferably, in step (2), under ultra-high magnetic fields (>3T), additional quantitative measurements of the spatial distribution of the actual flip angles are performed to optimize the flip angles, and in step (3), quantitative T1 magnetic resonance imaging uses short repeated time series of multiple flip angles.

[0012] Preferably, in step (5), an automatic shutter speed analysis method is used to obtain the vascular leakage rate Ktrans of the contrast agent in the tumor region, and further, SSM full Model fitting with Ktrans > 0.01 min -1 The steady-state water flux coefficient k from the cell to the interstitium at each pixel point was calculated using the tumor region. io Specifically, you can refer to CN201910621579.6.

[0013] The method provided by the present invention comprises: io This is a method for quantitatively measuring the above.

[0014] The present invention provides a method for evaluating the AQP4 expression level by measuring the water flux rate k across the cell membrane as a magnetic resonance imaging marker for gliomas. io We further provide an application of io The method of obtaining is as above. io The present invention is not limited to quantitative measurement of k as a magnetic resonance imaging marker for glioma. io The application of can be used in scientific research.

[0015] Preferably, the flux rate of water molecules across the cell membrane, k io The linear relationship between AQP4 expression level and AQP4 cell positivity rate = (k io -A) / B, In the formula, A is 0.1 to 0.2 seconds. -1 , B is 13.07 to 15.04 seconds -1 It is.

[0016] The present invention further provides a method for measuring the flux rate of water molecules across a cell membrane for non-disease diagnostic purposes as a magnetic resonance imaging marker for glioma to assess AQP4 expression levels, said method comprising: (1) The steady-state flow rate coefficient k of tissue water molecules from cells to the interstitium io Quantitatively measuring the (2) stereotactically taking a biopsy tissue using a biopsy planning system and quantifying the AQP4 expression level in the tissue; (3)k io and establishing a linear relationship between the two based on the expression levels of AQP4. (4) The steady-state flow rate coefficient k of water molecules from cells to the interstitium in the tissue being measured io and obtaining the AQP4 expression level of the tissue to be measured based on the linear relationship in step (3).

[0017] Preferably, the measurement method includes: (1) setting dynamic contrast-enhanced magnetic resonance imaging parameters to measure a noise level during a dynamic contrast-enhanced magnetic resonance scan; (2) optimizing a flip angle in the dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) acquisition parameters by Monte Carlo simulation to reset the dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) flip angle; (3) scanning quantitative T1 magnetic resonance imaging; (4) scanning dynamic contrast-enhanced magnetic resonance imaging and injecting a contrast agent; (5) Full Shutter Speed ​​Model SSM full A pixel analysis was performed for each pixel point in the tumor region using the steady-state flow rate coefficient k of water molecules from the cell to the interstitium at each pixel point. io and obtaining (6) stereotactically taking a biopsy tissue using a biopsy planning system based on dynamic contrast-enhanced magnetic resonance imaging and quantifying the AQP4 expression level in the tissue; (7)k io A linear regression analysis was performed on the AQP4 expression level and k io obtaining a linear equation for (8) Repeat steps (3) to (5) and calculate k according to the linear equation in step (7). io The image is converted into an AQP4 expression level image to achieve intratumoral AQP4 expression imaging.

[0018] Preferably, in step (6), AQP4 immunohistochemistry photographs of the biopsy tissue are obtained to quantify AQP4 expression levels in the tissue.

[0019] Preferably, in step (7), the AQP4 expression level and k io The linear relationship between AQP4 cell positivity rate = (k io -A) / B, In the formula, A is 0.1 to 0.2 seconds. -1 , B is 13.07 to 15.04 seconds -1 It is.

[0020] Preferably, in step (8), the spatial distribution map of AQP4 expression levels can be color-coded to improve visibility.

[0021] The present invention further provides the application of the flux rate of water molecules across cell membranes as a magnetic resonance imaging marker of gliomas in the preparation of a product for predicting the sensitivity of gliomas to radiotherapy and chemotherapy.

[0022] Here, the drug used in the above-mentioned radiotherapy and chemotherapy is temozolomide.

[0023] The present invention further provides a system for measuring a glioma magnetic resonance imaging marker, the system comprising: The steady-state flow rate coefficient k of tissue water molecules from cells to the interstitium io an image extraction and processing module for quantitatively measuring k obtained by the image processing module io and a post-processing module for establishing a linear relationship between the AQP4 expression level of the biopsied tissue. The image extraction module and the image processing module are performed on the tissue to be measured. io and a prediction module for predicting the AQP4 expression level of the tissue to be measured based on the linear relationship of the post-processing module.

[0024] Preferably, the measurement system comprises: a pre-processing module for setting dynamic contrast-enhanced magnetic resonance imaging parameters, measuring a noise level during dynamic contrast-enhanced magnetic resonance scanning, and optimizing a flip angle in the dynamic contrast-enhanced magnetic resonance imaging DCE-MRI acquisition parameters through Monte Carlo simulation to re-set the flip angle of the dynamic contrast-enhanced magnetic resonance imaging; an image extraction module for scanning quantitative T1 magnetic resonance imaging, scanning dynamic contrast-enhanced magnetic resonance imaging, and injecting a contrast agent; Full Shutter Speed ​​Model SSM full A pixel analysis was performed for each pixel point in the tumor region using the steady-state flow rate coefficient k of water molecules from the cell to the interstitium at each pixel point. io an image processing module for acquiring the k obtained by the image processing module io and a post-processing module for establishing a linear relationship between the AQP4 expression level of the biopsied tissue. The image extraction module and the image processing module are performed on the tissue to be measured. io and a prediction module for predicting the AQP4 expression level of the tissue to be measured based on the linear relationship of the post-processing module.

[0025] Here, the repetition time in dynamic contrast-enhanced magnetic resonance imaging parameters remains at or near the minimum.

[0026] Here, the noise level of dynamic contrast-enhanced magnetic resonance imaging is obtained by scanning a normal healthy subject.

[0027] Here, the present invention optimizes the flip angle in dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) acquisition parameters by Monte Carlo simulation to make it the most sensitive to detecting the process of water molecules crossing the cell membrane.

[0028] Here, the present invention rapidly injects the contrast agent at the 8th frame after the start of scanning. The resonance contrast agent is a T1 contrast agent, and the recommended dosage is 15-20 ml of saline immediately after injection, and the recommended injection rate is 2 ml.

[0029] Here, based on dynamic contrast-enhanced magnetic resonance imaging, combined with clinical factors, optimal biopsy point coordinates are obtained, and biopsy tissue is stereotaxically collected. Here, the tumor area can be manually outlined by a professional doctor or staff, or automatically obtained by artificial intelligence methods. The number of biopsy tissue samples is 45, but if the number of samples is further increased, k io This will improve the accuracy of characterization of AQP4 levels by

[0030] In summary, compared with the prior art, the present invention has the following technical advantages:

[0031] 1. The present invention provides a magnetic resonance imaging marker of AQP4 expression in gliomas, which is a steady-state flux coefficient of water molecules from cells to the interstitium, k io We proposed the relationship between AQP4 expression level and k io We prove that there is a highly linear relationship between

[0032] 2. This embodiment is a steady-state flow rate coefficient k of water molecules from the cell to the interstitium. io A dynamic contrast-enhanced magnetic resonance imaging method for quantitatively measuring k is provided, and a dynamic contrast-enhanced magnetic resonance imaging parameter is optimized by Monte Carlo simulation, and the k of the imaging method is obtained. io Greatly improves measurement accuracy.

[0033] 3. The present invention realizes quantitative imaging of AQP4 expression in glioma tumors and can measure the heterogeneity of AQP4 expression in glioma tumors. The AQP4 expression level can be used in scientific research such as research on the therapeutic effects of some anticancer drugs and new treatments, research on the relationship between AQP4 expression level and pathological processes such as glioma migration and proliferation, and research on the relationship between AQP4 expression level and glioma treatment resistance. [Brief description of the drawings]

[0034] [Figure 1] 1 is a schematic diagram of the aquaporin 4 magnetic resonance imaging marker designed in the present invention. The steady-state flow rate coefficient kio of water molecules from the cell to the interstitium is a magnetic resonance imaging marker of aquaporin 4 expression in glioma. [Diagram 2] 1 is a flow chart of the method for measuring a magnetic resonance imaging marker for glioma provided by the present invention. [Diagram 3] 1 is a flowchart of a Monte Carlo simulation. [Figure 4] Optimization of flip angle parameters for dynamic contrast-enhanced magnetic resonance imaging using Monte Carlo simulation. [Diagram 5] This is the result of linear regression analysis of AQP4 expression levels and kio mean values ​​of 45 glioma biopsy tissues. [Figure 6] Quantitative imaging results of AQP4 expression levels in tumors from a patient with low-grade glioma (left) and a patient with high-grade glioma (right) prepared according to the present embodiment. [Figure 7]kio can accurately detect the dynamic expression of AQP4 in C6 cell line during TMZ treatment. [Figure 8] The results show the changes in kio, AQP4 and other parameters of C6 cell line after TMZ treatment. [Figure 9] Confocal results of AQP4 expression and distribution in C6 cell line after TMZ treatment. [Figure 10] The biomarker kio is to accurately track the dynamic regulation of AQP4 in U87MG cell line during the proliferation cycle. [Figure 11] kiomap obtained from water-exchange DCE-MRI precisely revealing AQP4 heterogeneity in vivo. [Figure 12] The kio can accurately indicate the intratumoral AQP4 expression level in each rat glioma model. [Figure 13] 1 shows the linear correlation observed between kio and AQP4 expression in a C6 glioma rat orthotopic model. [Figure 14] Effects of specific pharmacological inhibition of AQP4 on kio in a subcutaneous rat glioma model. [Figure 15] FIG. 11. KIO map obtained from water-exchange DCE-MRI revealing intratumoral AQP4 distribution in human glioma. [Figure 16] Average statistical data for the special case and sample of glioma patients. [Figure 17] Low kio(AQP4) is a phenotype of treatment-resistant glioma. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0035] In order to make the objectives, technical solutions and advantages of the present invention clearer, the following clearly and completely describes the technical solutions of the present invention with reference to the drawings.

[0036] FIG. 1 shows a schematic diagram of the aquaporin 4 magnetic resonance imaging marker designed in the present invention, which shows the steady-state flow rate coefficient k of water molecules from the cell to the interstitium.io The principle behind k is that it is a magnetic resonance imaging marker of aquaporin 4 (AQP4) expression in gliomas. One AQP4 molecule transports approximately 0.24 pL of water molecules per second across the cell membrane, but the cell volume of a glioma cell is approximately 10 pL, so k io can characterize and amplify the AQP4 signal in MRI and thus serve as a linear biomarker of cell membrane AQP4 expression levels.

[0037] As shown in FIG. 2, the method for measuring a magnetic resonance imaging marker for glioma provided by the present invention includes the following steps:

[0038] Step 1, set the dynamic contrast-enhanced magnetic resonance imaging parameters to measure the noise level during dynamic contrast-enhanced magnetic resonance scanning, and the specific steps are as follows:

[0039] (1) Adjust the DCE-MRI sequence parameters of the scanning device and set the repetition time (TR) to the shortest or near-shortest value (e.g., 3 ms).

[0040] (2) Adjust the repetition number of the DCE-MRI sequence and operate a normal subject for 10 min to statistically analyze the noise level.

[0041] Step 2, as shown in Figure 3, optimize the flip angle (FA) in the dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) acquisition parameters by Monte Carlo simulation to be the most sensitive for detecting the process of water molecules crossing the cell membrane.

[0042] (1) A series of contrast agent concentration time curves C from previous dynamic contrast-enhanced magnetic resonance imaging data p (t) or artificial simulation C pSelect (t) and specify the initial physiological parameters. trans = 0.01 (per minute), mole fraction of water in blood vessels p b = 0.05, mole fraction of interstitial water p o = 0.2, the steady-state flow rate coefficient k of water molecules from the cell to the interstitium io = 3 Hz, the steady-state flow rate coefficient k of water molecules from the blood vessels to the interstitium bo =3 Hz.

[0043] (2) Parameter to be optimized - Set the adjustment range of the flip angle to 1 to 40 degrees.

[0044] (3) Set the simulated dynamic contrast-enhanced magnetic resonance imaging parameters to be the same as the actual dynamic contrast-enhanced magnetic resonance imaging parameters (except for the flip angle).

[0045] (4) DCE-MRI time series signal JPEG2025505359000006.jpg9140 is generated according to the three-compartment model (this model is the full shutter speed model SSM described in the patent (patent number ZL201910621579.6) full Refer to the model.

[0046] (5) Random white noise is added to the signal, depending on the noise level estimated by the DCE-MRI experiment. White noise of the same noise level was added to JPEG2025505359000007.jpg10142, and the full shutter speed model SSM of the patent (patent number ZL201910621579.6) was used. full Noise was added using Perform nonlinear least squares fitting on JPEG2025505359000008.jpg10142.

[0047] (6) Repeat steps (3) to (5) 100 times, and for each iteration, the steady-state flow rate coefficient k ioCalculate the standard deviation and median.

[0048] (7) Repeat steps (3) to (6) until all combinations of scanning parameters are completed. io The flip angle for which the fitting result is closest to the preset value of the simulation and the variance is the smallest is the optimal flip angle.

[0049] Step 3, follow step 2, k io The flip angle that minimizes the standard deviation of and brings the median closest to the set value is selected (see Figure 4), and the flip angle for dynamic contrast-enhanced magnetic resonance imaging is reset.

[0050] Step 4, perform quantitative T1 magnetic resonance imaging scan, parameters are set as follows:

[0051] Field of view (FOV): (340 mm)2, layer thickness: 1.5 mm, 80 layers, pixel size 0.8 × 0.8 × 1.5 mm3, echo time (TE) / TR = 2.46 ms / 5.93 ms, flip angle (FA), 2° / 14°, bandwidth, 450 Hz / pixel.

[0052] Step 5, scanning dynamic contrast-enhanced magnetic resonance imaging, and injecting contrast agent quickly at the 8th frame after the start of scanning; The DCE-MRI data acquisition parameters using the 3D CAIPIRINHA-Dixon-TWIST were as follows: FOV = 340 × 340 × 120 mm 3 ,FA=10°, bandwidth, 1090Hz / pixel, TR, 6 ms, TE, 1.3 ms.,The contrast agent was rapidly injected on the 8th frame after the start of the scan, and immediately after the injection, 15-20 ml of saline was injected at the recommended injection rate of 2 ml / s.

[0053] Step 6, Full Shutter Speed ​​Model SSM full A pixel analysis was performed for each pixel point in the tumor region using the steady-state flow rate coefficient k of water molecules from the cell to the interstitium at each pixel point.io The detailed steps are as follows:

[0054] (1) Based on the phenomenon that contrast enhancement in tumor regions is higher than that in normal tissues, tumor regions can be manually contoured by specialized doctors or personnel, or automatically obtained by artificial intelligence methods.

[0055] (2) Obtain the vascular contrast agent concentration time signal (AIF) of a living body from dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) time-series tumor area data, and (3) According to the vascular contrast agent concentration-time signal in step (2), a full shutter speed model (SSM) is applied to the DCE-MRI time series signal at each pixel point. full A nonlinear least-squares sum fitting of the SSM for each pixel point in the tumor region was performed. full Obtain the DCE-MRI signal fitting results of the model respectively. SSM full After fitting the model, the distribution diagram of the five groups of physiological parameters is generated, and the five groups of physiological parameters are K trans , the mole fraction of water in blood vessels p b , the mole fraction of interstitial water p o , the steady-state flow rate coefficient of water molecules from the blood vessels to the interstitium k bo , and the steady-state flux coefficient k of water molecules from the cell to the interstitium io Includes.

[0056] (4) Perform error analysis on kio and kbo in step (3) and set the value for [0 seconds -1 20 seconds -1 ] 95% confidence interval within the interval, or 5 seconds -1 Only pixel point results with the lower limit of the 95% confidence interval exceeding io and k bo Parameter distribution diagram and K trans , p b and p o Generate a distribution plot of the parameters.

[0057] The above steps 1 to 6 are a dynamic contrast-enhanced magnetic resonance imaging method for quantitatively measuring kio, which is io This makes it possible to achieve the first object of the present invention, which is to significantly improve the accuracy of the measurement.

[0058] Step 7: Use the biopsy planning system to identify the k io Obtain optimal biopsy point coordinates based on images combined with clinical factors, and the detailed steps are as follows:

[0059] (1) One to two days before stereotactic biopsy, all patients had a stereotactic frame attached to their head and underwent a structural MRI scan before the biopsy.

[0060] (2) k for the entire tumor area, calculated as described in step 6 io The map is registered and fused with the structural image and input into a stereotactic biopsy planning system for stereotactic biopsy planning.

[0061] (3) In the stereotactic surgical planning system, three-dimensional image reconstruction and surgical simulation are realized by calculating the target coordinates and the trajectory approach angle, and the coordinate approach point and needle insertion angle of the target in the cerebral cortical region are determined. In each patient, under the most favorable clinical factors, it is necessary to design the biopsy trajectory to avoid access through, for example, the sulcus, cortical artery, venous structure or ventricle. Different k io Select multiple ROIs with values.

[0062] Step 8: stereotactically taking biopsy tissue, obtaining AQP4 immunohistochemistry images of the biopsy tissue, and quantifying the AQP4 expression level of the tissue; (1) Tissue samples from all biopsy sites were surgically taken by three brain surgeons according to the designated trajectory plan of the stereotactic biopsy entrance and target points in step 7. Biopsies were performed under local or general anesthesia, with a biopsy needle (inner diameter 2.0 mm, lateral cutting window 10 mm) carefully and gently inserted into the target site and rotated clockwise (0°, 90°, 180°, 270°) to take biopsy tissue samples.

[0063] (2) Stereotactic biopsy is performed to obtain specimens of 5-10 mm in length. Samples that need to be subjected to immunohistochemistry to obtain AQP4 expression distribution are fixed and embedded, and AQP4 immunohistochemical staining is performed according to conventional immunohistochemistry steps.

[0064] (3) High-resolution images of whole-section AQP4 immunohistochemistry are scanned using a microscope.

[0065] (4) The digital processing process of AQP4 immunohistochemical scanning image is as follows: First, calculate the ratio (AQP4mean) between the sum of AQP4 gray values ​​and the number of cell nuclei in the whole slice. Then, in the immunohistochemical slice, select the region (ROI) with the strongest AQP4 staining intensity (i.e., the area with the highest gray value in the photo) as the fully positive expression region of AQP4. Calculate the ratio (AQP4max) between the sum of AQP4 gray values ​​and the number of cell nuclei in the region, and finally, the AQP4 expression level (i.e., the positive rate) of the biopsy point is (AQP4mean / AQP4max)×100%.

[0066] Step 9: Repeat steps 4 to 8 to obtain dynamic contrast-enhanced magnetic resonance images and stereotactic biopsy tissues of multiple glioma patients.

[0067] Step 10. AQP4 expression levels and k in all biopsies io Linear regression analysis was performed on the mean values ​​to determine the relationship between AQP4 expression level and k io The linear equation (i.e., AQP4 cell positivity rate = (k io -0.2 seconds -1 ) / 14.1 seconds -1 The specific steps are as follows:

[0068] 1. First, the k of each stereotactic biopsy point of 19 patients (6 cases WHO I-II, 13 cases WHO III-IV, of which 10 cases had recurrent glioma) and 45 biopsy point samples were ioCalculate the magnetic resonance parameters including values ​​and the corresponding immunohistochemistry results (i.e., AQP4 positivity rate) one by one.

[0069] 2. Next, the linear fitting coefficients for each magnetic resonance parameter and AQP4 expression level are evaluated by linear regression analysis. The optimal linear correlation parameter k io and obtain the linear equation and 95% confidence interval of AQP4 expression level.

[0070] 3. The final results are shown in Figure 5, which shows the relationship between AQP4 expression levels and k io A linear equation, i.e., AQP4 positive rate = (k io -0.2 seconds -1 ) / 14.1 seconds -1 It is.

[0071] In addition, the method can also use a machine learning algorithm to introduce multiple magnetic resonance parameters to establish an AQP4 positivity rate prediction model instead of linear regression analysis, thereby further improving the prediction accuracy.

[0072] Step 11: According to the above linear equation, k for each pixel point in the tumor io The map is converted into an AQP4 expression level image, realizing non-invasive imaging of intratumoral AQP4 expression.

[0073] Step 12: For new glioma patients, steps 4-6 and 11 are repeated without performing stereotactic biopsy, and the spatial distribution map of AQP4 expression in the patient's tumor can only be obtained by magnetic resonance scanning and data analysis. Figure 6 shows the quantitative imaging results of intratumoral AQP4 expression levels in a patient with low-grade glioma (bottom left) and a patient with high-grade glioma (bottom right), as well as the corresponding thin-section enhanced T1-weighted magnetic resonance images (top).

[0074] The magnetic resonance imaging marker measurement system for glioma provided by the present invention comprises: A pre-processing module for performing steps 1, 2 and 3; an image extraction module for performing steps 4 and 5; an image processing module for performing step 6; a post-processing module for combining steps 7, 8, and 9 to perform steps 10 and 11; and a prediction module for performing step 12.

[0075] Evaluation of AQP4 expression levels and its use as a magnetic resonance imaging marker in gliomas in the preparation of products for predicting the sensitivity of gliomas to radiotherapy and chemotherapy io In order to further verify the application of the present invention, the following experimental method is used.

[0076] Cell culture and establishment of TMZ-resistant cell model The present invention uses glioma cell lines C6 and U87MG from the American Type Culture Collection (ATCC, HTB14™). Taking C6 cell culture as an example, the method is as follows: C6 cells are cultured in Dulbecco's Modified Eagle Medium (DMEM, Sigma-Aldrich, D6429-500ML), which contains 10% fetal bovine serum (FBS, Biological Industries, 04-002-1A) and 1% penicillin and streptomycin double antibody (P / S, Gibco, Thermo Fisher Scientific, 10378016), which is a complete medium. The culture environment is a humidified incubator with a temperature of 37°C and 5% CO2+air. The medium is replaced twice a week, and the cells are passaged in the logarithmic growth phase (phase II).

[0077] C6 cells are treated with TMZ (50 μM) in complete medium containing 0.1% DMSO (control group) or 50 μM TMZ in 0.1% DMSO. Cellular magnetic resonance parameters k io and physiological information (such as cell morphology, AQP4 expression, cell activity, and cell migration distance) are obtained on the 3rd and 7th days of TMZ (50 μM) treatment, respectively.

[0078] For primary cell culture, cut glioma biopsy tissue into 1 mm*3 pieces and mix with 10 ml of trypsin-EDTA (Gibco, 25200056) at 37 °C for 10-15 min until most of the pieces are digested into a single cell suspension. Then, culture the cell suspension using the same method as for the glioma cell lines (flow cytometry results in Figure 8).

[0079] Cell samples for MRI studies and protocols for responding to specific inhibitors In order to obtain a cell sample for MRI, in the present invention, U87MG (0.5 to 1 × 10 5 ) and C6(1-2×10 6 ) Cell lines are washed with DPBS and dissociated by adding 1 mL of 0.25% trypsin-EDTA solution to a 60 mm Petri dish, incubated for 0.5-1.0 min, and then resuspended in 2 mL of PBS (or PBS with 5 mM gadoteridol). The cell samples are centrifuged at 150 g for 5 min at 4 °C, and then the samples are resuspended in MR-compatible glass sample tubes containing 250-300 μL of PBS (or PBS + 5 mM gadoteridol). The sample tubes are then centrifuged at 300 g for 2 min at 4 °C before performing MRI. For AQP4-specific inhibition experiments, the PBS in the above steps should be replaced with PBS + inhibitor, i.e., the cell samples are preincubated for 15 min in PBS or PBS with 6.4 μM TGN020 (Axon, CAS 51987-99-6) before detection (c and d in Figure 7, a in Figure 8).

[0080] cell k io A benchtop MRI system for measuring in vitro cell cultures A 0.5T benchtop MRI measurement system for in vitro cell culture (Pure Devices GmbH, Germany) includes the following: an MRI system mounted on a vibration-isolated table, an autonomous NMR tube with live cell pellet from the incubator, 5% CO2 + 95% O2 and PBS supernatant to ensure high viability, and an attached optical fiber for temperature monitoring. All MRI measurements are performed at room temperature (23.5 + / - 1 °C).

[0081] Prior to measuring water exchange DCE-MRI, diffusion-weighted imaging (DWI) was performed to identify the location of cell layers with low apparent diffusivity, using the following parameters: single layer acquisition, slice thickness 5 mm, FOV = 12.8 × 12.8 mm. 2 , matrix size 64 × 64, average value of 16 pieces, and 10 seconds / mm 2 and 2000 sec / mm 2 Water-exchange DCE-MRI uses an inversion-recovery-prepared turbine spin-echo (IR-TSE) sequence and the Gd-based contrast agent gadoteridol (Prohance™, Bracco Diagnostics, Princeton, NJ). The scan parameters used for water-exchange DCE-MRI were: echo time (TE) 3 ms, turbo factor = 16, FOV = 12.8 × 12.8 mm. 2 , matrix size 32 × 32. For water-exchange DCE-MRI, two contrast agent (CA) concentrations (0 mM and 5 mM) are used. For [CA] = 5 mM, 13 IR delays (10 ms, 30 ms, 50 ms, 70 ms, 90 ms, 150 ms, 200 ms, 400 ms, 600 ms, 800 ms, 1 s, 5 s, 5 s) are varied with a repetition time TR (TR = IR delay + 5 s) and one repetition of each IR delay. For [CA] = 0 mM, the longest IR delay is extended to 10 s and at the same time the TR is extended (TR = IR delay + 10 s), thus ensuring full recovery of the longitudinal magnetization for each TR. The scan times for single-layer acquisition for [CA] = 0 mM and 5 mM are 10 min and 5.5 min, respectively.

[0082] The IR-TSE signal for each IR delay (M) is obtained from the average signal in the cellular ROI. The signal is then subtracted and normalized by the equilibrium magnetization (M0, the M with the longest IR delay). Define JPEG2025505359000009.jpg8139 to be the normalized signal analyzed by the water exchange model.

number

[0083] Here, a two-site exchange (2SX) SS model is used, and in brief, the MR signal is considered to be the sum of signals from two water sites, intracellular water and extracellular water, with each site having a similar longitudinal recovery rate R 1 It is assumed that p i and p o are the molar fractions of water molecules in the intracellular and extracellular spaces, respectively, and p i +p o = 1. Next, the normalized IR-TSE signal JPEG2025505359000011.jpg8145 can be expressed as a double exponential function,

number

[0084] Here, R 1、sm and R 1、lar are small R1 and large R1, respectively, and p sm is R 1、sm The three parameters R 1、sm , R 1、lar and p sm is the intracellular water efflux rate constant (k io ), R1 of intracellular water (R 1i ), and CA-dependent extracellular water R1 (R 1o ) are determined by three physical parameters. For each case (e.g., C6 on days 0, 3, and 7 of TMZ), at least three samples are measured using IR-TS as baseline data when [CA]=0 mM. R 1i is previously determined by fitting the IR-TSE signals acquired at two different CA concentrations to more than three samples.

[0085] Immunofluorescence staining, visualization and quantification of AQP4 in vitro Immediately after MRI testing, cell line samples are fixed with 4% paraformaldehyde (PFA) for 20 min at room temperature and then stored in 0.5% PFA (4 °C) until downstream molecular biology experiments. For immunofluorescence (IF) testing, (1) block with 10% goat (Beyotime, C0265) for 1 h at room temperature, (2) incubate with AQP4 primary antibody (1:200) overnight at 4 °C, followed by incubation with AQP4 secondary antibody (1:500) for 1 h at room temperature, (3) stain with DAPI (4',6-diamidino-2-phenylindole, Sigma-Aldrich, 1:1000) for 5 min at room temperature, and (4) wash three times with PBS. Fluorescence quantification is achieved using a microvolume fluorescence spectrophotometer (DFX, Denovix, Wilmington). Data are collected from two fluorescence channels. One characterizes DAPI with excitation / emission wavelengths of 375 nm / 435-485 nm, and the other is collected in the blue (excitation / emission 470 nm / 514-567 nm) channel. All measurements use blank controls to avoid false positives. Both measurements are performed independently by two researchers before processing the DCE-MRI results (right side of Figure 7a).

[0086] Characterization of cell viability upon TMZ treatment On the third and seventh day of TMZ or DMSO treatment, 1% CCK-8 (Cell Counting Kit-8, Beyotime, C0037) is added to the 96-well plate for cell viability assay. After 1 hour of incubation, the supernatant is used for optical density test by microspectrophotometer (Ultrafine Photometer, NanoDrop 2000, Thermo, Waltham) (e in FIG. 7).

[0087] Assessment of cell migration Culture the cells into a cell monolayer and use a standard 200 µl pipette tip to draw a strip-shaped scratch 300-500 µm wide from the bottom of the cell culture dish. Then, after 24 h of incubation under the corresponding conditions, fix the samples with 4% PFA for 30 min and stain with DAPI. For cell migration calculations, only calculate the cells in the scratch area (Figure 3b).

[0088] Methods for classification of fast cycling cell (FCC) and slow cycling cell (SCC) markers The present invention uses a cell tracer method to distinguish FCC from SCC. Here, OG (Oregon Green 488 Carboxylic Acid Diacetate, succinimidyl ester) and CTV (CellTrace™ Violet reagent, ThermoFisher Scientific, Invitrogen C34557) are used to distinguish FCC (low OG or CTV intensity) from SCC (high OG or CTV intensity, e.g., CTV fluorescence intensity >104 in C6 cell line and human glioma primary cultured cells, respectively). The principle is that during the cell culture process, FCC and SCC cell populations are identified and separated based on their different abilities to retain OG or CTV within the same period. The specific method is as follows: First, the cells are suspended in PBS containing 25 μM OG and incubated for 10 minutes (in a cell culture incubator). After incubation, the cell samples are washed with DMEM to remove residual dye and returned to the medium until microscopic imaging. Before fluorescence imaging, samples are fixed with 4% PFA for 30 min and stained with DAPI. The percentage of SCC is defined as the percentage of OG-positive cells among DAPI-positive cells. For primary human glioma cells, they are stained with 5 μM CTV (same method as OG) for 3 days. Then, the fluorescence intensity statistics and counting of cells are measured simultaneously by fluorescence flow cytometry (FACS) (LSRFortessa X-20, BD, USA).

[0089] In addition, the present invention also uses EdU (5-ethynyl-2'-deoxyuridine), a biomarker that marks the subtype of proliferating cells by marking newly synthesized DNA, to mark FCC (high EdU) and SCC (low EdU). The specific method is as follows: Obtain a C6 cell suspension and incubate with 50μM EdU solution (including Cell-LightTM EdU Apollo567, Ribobio C10338-1, EdU solution and Apoll fluorescent solution) at 37℃ for 2 hours, and mark AQP4 according to the above IF method. After AQP4 staining, the sample is incubated in Apollo fluorescent solution for 30 minutes to fluorescently mark EdU, and then transferred for FACS testing.

[0090] The cell morphology and fluorescence images of the present invention are taken using a fluorescent inverted microscope system (cellSensV1.13, Olympus, Japan) or a confocal laser scanning microscope (fv1200, Olympus, Japan).

[0091] Two types of glioma rat models and magnetic resonance measurements The experimental protocols for all animal studies are approved by the Zhejiang University Animal Experimentation Committee. Adult (7-8 weeks old) male Sprague Dawley (SD) rats are obtained from the Zhejiang University Experimental Animal Center. For the glioma cell transduction process, rats are anesthetized with 2% (v / v) isoflurane (R500IE, RWD Life Science Co., Ltd.) and then inoculated with 5 × 10 6 100 µl of PBS containing 100 C6 cells and 1% antibiotics are slowly injected subcutaneously into the right leg. 7-9 days after tumor implantation, animals are subjected to 7T MRI examination. After MRI, tumors are rapidly excised and fixed in 4% PFA for 24 h and stored in 0.5% PFA until IHC.

[0092] For the orthotopic glioma model, inject 10 μl of the C6 cell suspension (0.5 × 10 cells) into the right caudate putamen of the brain using a microinjection needle. 5In addition to injecting 1000 cells / microliter, 4-5 μl / rat, 2 μl / min), the injection coordinates are located in the brain region 0.8 mm from the anterior arcuate suture, 2 mm to the right of the sagittal suture, and 4.5 mm deep. Two weeks after tumor inoculation, animals are subjected to 9.4T-MRI examination, fixed in 4% PFA for 24 h immediately after MRI, and preserved in 0.5% PFA until IHC. The tumor volume of all animal models was 4000 mm 3 not exceed.

[0093] Specific pharmacological inhibition of AQP4 in a rat glioma model Here, we use TGN020 to inhibit AQP4 in a rat glioma model. For the TGN020 group, each animal is intravenously administered TGN020 (3 mg / kg, 4 ml / kg body weight) in the tail vein 15 min prior to water-exchanged DCE-MRI. To increase solubility, TGN020 is dispersed in saline (0.9% NaCl) at 37 °C with repeated sonication and vortexing to promote dissolution prior to intravenous injection. One day prior to TGN020 treatment, the same animals are treated with saline (4 ml / kg) and then subjected to water-exchanged DCE-MRI data acquisition. In the control group, animals are treated with the same volume of saline on both the first and second days. Here, a post-injection flush is required after a small volume injection to ensure safety and therapeutic efficacy in the rats.

[0094] Histology and immunohistochemistry IHC is performed on rat gliomas using paraffin-embedded thin sections. Briefly, rats are euthanized with 5% isoflurane immediately after MRI experiments, and tumor tissues are excised and fixed in 4% PFA for 24 h. Then, tissue thin sections (approximately 4 μm thick) are cut along the MRI scanning plane after the tumor tissues are carefully registered to the MRI data. AQP4-IHC is as follows: (1) incubation with anti-AQP4 rabbit polyclonal antibody overnight at 4 °C, and then (2) incubation with goat secondary antibody HRP (horseradish peroxidase) at room temperature for 1 h. Hematoxylin is also used to perform cell nuclei staining. Finally, the tissue thin sections are thoroughly scanned at 20x magnification by a microscope slide scanning system (VS120, Olympus, Japan).

[0095] The process of quantitative analysis of immunohistochemical profiles is as follows. First, the images are stained and separated with histological imaging by ImageJ (open source Fiji v1.53c, plugin: Color deconvolution), and further analysis is performed using MATLAB® 2018 (MathWorks, Natick, MA, USA) to remove background and quantify the number of nuclei and AQP4 staining (grayscale intensity). Then, the average AQP4 grayscale intensity of all cells in each slide is calculated. At the same time, several ROIs showing the highest AQP4 expression in all slides are manually selected by two experienced pathologists and are classified as 100% AQP4 positive (AQP4 + Finally, the AQP4 in the ROI is considered to be + % mean grayscale density of AQP4 for each slide + The percentage of AQP4 is further calculated as the mean AQP4 greyscale density of the slice. A double-blind principle is followed for IHC and AQP4 quantification.

[0096] Selection of region of interest (ROI) in rat glioma model. Since most subcutaneous glioma tumors exhibit a ring shape with high AQP4 expression, concentric ring ROIs are used to divide the MRI tumor area and the corresponding tissue image into six ROIs. First, the minimum circumscribing rectangle of the entire tumor is obtained, and the center coordinates (XR, YR), length LR, and width WR of the rectangle are determined. Then, a series of concentric elliptical curves are automatically drawn by programming to divide the tumor into six ROIs. The major axis (am) and minor axis (bm) of the concentric elliptical curves are calculated as follows:

number

[0097] where m is the sequence number of each concentric elliptic curve (from the exterior (m=1) to the interior (m=6)), and q=0.75.

[0098] Water exchange DCE-MR and stereotactic biopsy of human brain gliomas (step 7, using a biopsy planning system to identify intratumoral k io Obtain optimal biopsy point coordinates based on images in combination with clinical factors).

[0099] The samples obtained in stereotactic biopsy are usually small. Nevertheless, the present invention incorporates some modifications in the practice of processing the specimens to accommodate the full range of histological, immunocytochemical and ultrastructural studies. For specimens 5-10 mm long, the thin tip is removed from the biopsy needle to avoid drying and transferred to the laboratory in a glass bottle submerged in saline. The biopsy specimen is then divided into several small samples for different purposes, for example, one can be selected for cryopreservation and another fixed in glutaraldehyde for electron microscopy. The remaining specimen is fixed, embedded and processed for routine IHC and HE staining. This method allows special staining procedures and IHC studies to be performed on comparable consecutive samples. Each sample is a specific k ioAfter tissues were histologically identified as glioma by a neuropathologist, AQP4 and ZEB1 staining and image digital analysis were performed following the same IHC protocol as in the rat glioma model.

[0100] For IEM, immunogold-silver labeling was used to detect AQP4, the steps of which are as follows. (1) Biopsy tissues were rapidly fixed in IEM buffer (a PBS solution containing 0.2% glutaraldehyde and 2% paraformaldehyde) and left at room temperature for 2 hours, followed by incubation in PBS containing 50 mM glycine for 15 minutes at room temperature; (2) samples were blocked and permeabilized for 40 minutes in PBS containing 5% goat serum, 1% Triton, and 1% fish collagen; (3) samples and primary antibodies were prepared in PBS containing 1% fish collagen and left at 4°C overnight; (4) post-embedding immunogold labeling was performed using gold-conjugated secondary antibodies in PBS containing 1% goat serum and 1% Triton and left at 4°C overnight; (5) silver enhancement was performed in the dark using the HQ Silver Enhancement Assay Kit (Nanoprobes, 2012-45ML) to observe AQP4 immunoreactivity; and (6) samples were washed several times with deionized water before and after the silver enhancement step. (7) Immunolabeled specimens are fixed in PBS containing 0.2% OSO4 for 2 hours, stained with 0.5% uranyl acetate for 1 hour, dehydrated in graded ethanol, and then flattened and embedded in Epon812 medium. Ultrathin sections (70 nm thick) are then observed under a Jeol-1200 electron microscope (JEOL Ltd., Tokyo, Japan). The above process follows the rules of double-blind experiments.

[0101] The above experiment verifies the following:

[0102] k io Accurately detected the dynamic expression of AQP4 in C6 cell line during temozolomide (TMZ) treatment To evaluate the ability of water-exchange DCE-MRI to detect dynamic changes in AQP4 expression levels during TMZ treatment of glioma, the present invention uses the above TMZ incubation to induce C6 cells to construct a glioma treatment model. As shown in Figure 7a, confocal fluorescence images show that: As TMZ treatment time increases, AQP4 expression continues to decrease, and on the 7th day of TMZ treatment, the expression level of AQP4 almost completely disappears, and AQP4 is only detected in some cell membrane uropods (Figure 7a, Figure 8a, b, c). Using a microfluorescence spectrophotometer (QFX, Denovix, Wilmington), the AQP4 expression level, i.e., the fluorescence intensity (rfu), is further quantified, and it is decreased by 42.5% (p=0.0013) and 85.9% (p<0.0001) on the 3rd and 7th days of TMZ treatment (Figure 7a, right), while no significant change is observed in the control group (Figure 9b). As expected, cell migration (scratch assay, Fig. 7b), proliferation rate (Fig. 9e), Ki-67 expression level (Fig. 9d), and cell viability and proliferation index obtained by CCK-8 (Cell Count Kit-8)-induced cell migration (Fig. 7e) all gradually decreased with increasing TMZ treatment time.

[0103] In the control group, C6 cell line was treated with k io = 10.9 ± 0.7 s (n = 10, Fig. 7c), and k io On the third day of TMZ treatment, the mean time to treatment was 6.7 ± 0.8 seconds. -1 (37.3% decrease, p=0.0008, n=6) and by day 7, it had rapidly decreased to 5.8±0.2 seconds. -1 (45.5% decrease, p<0.0001, n=7) (c in FIG. 7). io The magnitude of the decrease (10.9-5.8=4.9 seconds -1 ) corresponds well to the magnitude of the 85% decrease in AQP4 expression, i.e., 85% AQP4 modulation k io =85%*5.8 seconds -1 =4.9 seconds 1 , which is k during TMZ treatment ioThis indicates that the dynamic changes of k are mainly specifically controlled by the AQP4 pathway. This explanation was further confirmed by the results of glial cell AQP4 specific inhibition using TGN020 with 7 days of TMZ, which showed that on the 7th day of TMZ, AQP4 was hardly expressed, and TGN020 could not inhibit k io It can be seen that the amount of β-K is not able to be further reduced (d in FIG. 7).

[0104] Here, a specific description of FIGS. 7 to 9 is given below.

[0105] Figure 7 shows the k io a) Dynamic expression of AQP4 in C6 cell line during TMZ treatment can be accurately detected. a) Images (confocal microscope) of expression and distribution of AQP4 (red) after TMZ treatment of C6 cells. On the right side, quantitative rfu results (measured by microfluorescence spectrophotometer) of AQP4 and DAPI of C6 cell line after TMZ treatment, n=(4, 3, 3) from left to right. Scale: 25 μm. b) Measurement results of migration ability of C6 cells at different time points of TMZ treatment, where cell nuclei marked with DAPI (4',6-diamidino-2-phenylindole) are shown in green in the scratch area. The bar graph on the right side shows the sum of migration distance of all cells in the scratch area. (n=(13, 4, 7) from left to right), normalized to the result of the control group. Scale: 200 μm. c) Quantitative rfu results of AQP4 and DAPI of C6 cell line after TMZ treatment, n=(4, 3, 3) from left to right. Scale: 25 μm. io Dynamic change trend of n = (10, 6, 7), (G-I: the numbers at the bottom represent control, day 3 of TMZ, and day 7 of TMZ from left to right). d: On the 7th day of TMZ treatment of C6 cells, the specific inhibition of AQP4 by TGN020 (marked as "+" and without TGN020 marked as "-") further increased k io e Calculation of cell viability by CCK-8 optical density (OD) (n=(4, 3, 3). Data are shown as mean ± SEM. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ns not significant.

[0106] Figure 8 shows the kappa variability of the C6 cell line after TMZ treatment. io , AQP4 and other parameters changes, in a-f, a is the k between the treatment group and the control group io a) Change in AQP4 between treated and control groups, n=(4, 6, 4, 6), p=0.0099, p=0.0006; b) Change in AQP4 between treated and control groups, AQP4(rfu) / DAPI(rfu), n=(3, 5, 3, 3), p=0.0328, p<0.0001; c) Migration distance (normalized by control) between treated and control groups, n=(5, 4, 8, 8), p=0.0836, p=0.0001; d) Ki-67(rfu) / DAPI(rfu) between treated and control groups, n=(3, 6, 3, 6), p=0.1041, p=0.0133; e) Cell proliferation rate between treated and control groups. n=(6, 15, 3, 34, 13), p=0.0505, p=0.0001. f: Percentage of SCC (OG+ cells / total cells). n=(4, 3, 3, 3). Here, in the control group, C6 cell line is incubated with DMSO only. e: Results of AQP4 KO group are also shown. Data are shown as mean ± standard deviation. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ns, not significant.

[0107] Figure 9 shows the confocal results of AQP4 expression and distribution in C6 cell line after TMZ treatment. a is a confocal microscope image of AQP4 and DAPI cell morphology in C6 cell line. Scale: 50 μm. b and c are the results of stereotactic analysis of AQP4 expression level on the white dotted line in a, i.e., the results of quantitative analysis of dynamic changes of AQP4 and cell nucleus fluorescence intensity at different positions far from the cell nucleus core. The results show that 7 days after TMZ, the AQP4 expression level decreases, while the highest peak of AQP4 expression is far from the cell nucleus, and the fluorescence of the cell nucleus appears partially bright and irregular.

[0108] quantitative k io Accurately predicted the dynamic regulation of AQP4 during the cell proliferation cycle Biomarkers to detect dynamic AQP4 regulation associated with glioma cell proliferation io To evaluate the ability of the present invention, we performed SS-DCE-MRI measurements at different phases of the U87MG growth curve, and the cells exhibit different proliferation states, as well as different cell densities and morphologies in the cell proliferation curve (Fig. 10a). A typical S-shaped curve of cell growth and proliferation was observed, which can be divided into: (I) the adaptation phase (0-48 h), in which cells do not divide but adapt to the culture conditions; (II) the logarithmic (log) proliferation phase (72-96 h), in which cells proliferate rapidly and the cell number increases exponentially; and (III) the stationary and death phase (starting at 120 h), in which the overall cell proliferation slows down. AQP4 expression levels peak in phase II, decrease by 60.4% in phase I (p<0.0001), and decrease by 66.0% in phase III (p<0.05) (Fig. 10b,c). Interestingly, but not surprisingly, k io shows a similar trend to the growth cycle. io also peaked in logarithmic phase II (6.7 ± 0.8 s -1 , n = 27, smaller than previous results using 0.1% DMSO), Phase I (5.04 ± 0.68 seconds -1 , p<0.0001) and Phase III (4.6±0.9 sec -1 , p<0.0001) (d in FIG. 10). In the first stage, the inhibition by TGN020 was io (e in FIG. 10), which indicates that AQP4-regulated k io Most of the activity was lost in the first step, which explains why the values ​​are smaller than those of the control group.

[0109] Figure 10 shows the biomarker k ioFigure 1 shows that the dynamic regulation of AQP4 in U87MG cell line during the growth cycle is accurately tracked. a: Morphology of U87MG cell line in different growth phases, including adaptation phase I, logarithmic phase II, stationary phase and decline phase III. Scale = 100 μm. b: Typical microscopy results of the changes of AQP4 (red) in U87MG cell line. (c, d) AQP4 expression and k in three cell cycle stages. io e shows that AQP4 inhibition by TGN020 (n=6) was significantly higher than that in Phase I. io In c-e, bar heights and error bar widths indicate the mean and standard deviation, respectively; *p<0.05, ***p<0.001; ns not significant; in c, p=0.7767; in d, p=0.2207; in e, p=0.4031. In c-e, data points (e.g., dot charts) are overlaid on corresponding boxes; two-tailed unpaired t-tests f using data from U87MG and C6 cell lines. io The aim of the study was to perform a correlation analysis between expression and AQP4 expression. The same symbols are used for each cell line (triangles for C6 and points for U87MG), the solid line reflects the linear regression, and the area between the two dashed lines reflects the 95% confidence interval.

[0110] Biomarker k io was linearly correlated with AQP4 expression in C6 and U87M cell lines. Quantifying AQP4 expression io To further evaluate the ability of k io A direct correlation analysis is performed between k and AQP4 expression (for rfu value and detection method, see AQP4 fluorescence spectrophotometer detection). io showed a significant positive correlation with AQP4 (correlation coefficient R = 0.92, p = 0.01), which indicates that k io We show that this assay can accurately represent the dynamic expression of AQP4 even in glioma cell types.

[0111] In vivo k obtained from water-exchange DCE-MRIio The map precisely reveals intra- and intertumor AQP4 heterogeneity in a rat glioma model To further demonstrate the accuracy of water-exchange DCE-MRI in detecting AQP4 in vivo, we establish a glioma animal model by subcutaneously implanting the C6 cell line in the right leg of Sprague Dawley (SD) rats. The in vivo water-exchange DCE-MRI is achieved using a clinically used Gd-based CA (gadopentetate dimethylamine, Guangzhou, China). io In order to better estimate k, the present invention uses a method of injecting CA twice, which reduces k compared to the conventional method of injecting contrast agent once. io It improves the accuracy of the estimation by a factor of 10. Note that a steady-state multi-gradient echo (MGE) sequence is used to overcome potential T2* artifacts caused by contrast agents. In addition, in combination with numerical simulation, the MGE sequence parameters are optimized (e.g., optimal parameters: repetition time (TR) = 100 ms, flip angle (FA) = 20°). Finally, a careful error analysis is performed on the SS model fitting to identify k with large fitting errors. io In short, these steps remove k within the physiological range ([0 s-1, 10 s-1]). io Ensure the accuracy of water-exchange DCE-MRI in estimating .

[0112] As expected, the in vivo SS model fit was excellent for most tumor voxels. io The figure (d in Fig. 11) and the AQP4 immunohistochemistry (IHC) staining (e, f in Fig. 11) show similar spatial distributions. Note that in most animals, we observed strong intratumoral heterogeneity in AQP4 expression, with high expression at the tumor margins (infiltrated areas) and low expression in the tumor center (e, f in Fig. 11). io The map shows a high k band at the tumor edge of the tissue slice shown in the schematic. io , low k in the tumor center io, which show the same expression pattern as AQP4. To verify the accuracy of the water-exchange DCE-MRI quantitative assessment of AQP4 expression measurement in vivo, in combination with the ring-like distribution of AQP4 in most glioma subcutaneous model animals, a series of concentric ring-shaped ROIs are used to divide the tumor slice into six regions (Methods, Fig. 11 d, e). Then, the average k io value and AQP4 positivity (AQP4 + ) fraction. The results are compared with the k fraction obtained from water-exchanged DCE-MRI. io The figure shows the AQP4 of each animal. + The results show that the correlation coefficients of the serotonin levels and the serotonin levels were strong (i in Fig. 11, correlation coefficients R>0.80 for each animal, Fig. 12), with an average R=0.82 (p<0.0001, g in the figure, n=10). io Linear regression analysis (Fig. 11g) further demonstrated that AQP4-related kappa-like expression can be accurately detected in gliomas in vivo. io 10.5 seconds -1 whereas non-AQP4 pathway-mediated k io takes just 0.4 seconds -1 The results show that AQP4 in each ROI (60 ROIs in total) + Percentage of cells (AQP4 + %) follows the following linear relationship:

number

[0113] As shown in Figure 11, the k obtained by water-exchange DCE-MRI io The maps can pinpoint the heterogeneity of AQP4 in vivo. a) T2-weighted TSE image of a glioma tumor xenograft, and the red and blue rectangles indicate the FOV in b and d, respectively. b) K-weighted TSE image overlaid on the T2-weighted TSE image. trans Fig. 3c shows that the optimization of MRI parameters (FA = 20 and TR = 100 ms) ensures water-exchange DCE-MRI within the physiological range ([0 s-1, 10 s-1]). The dotted line indicates the input kio (ground truth). The points and error bars represent the estimated k io , n = (100, 100) represents the mean and standard deviation. d is the mean of the k io (c) and (d) are maps of the IHC results from the same slice d, overlaid (left) and not overlaid (right) on the T2-weighted image. Here, AQP4 and nuclei are marked with different colors. For each animal, six contour lines are used to divide the tumor slice into six concentric ring ROIs. (d) is the magnified IHC from the tumor ring pixels (top of f) and the tumor core pixels (bottom of f), the positions of which are shown in (d,e). From left to right, they are the combined image, nuclei, and AQP4. (g) is the ROI average k for the 60 ROIs. io The linear correlation observed between tumor size and AQP4 positivity (6 ROIs per animal, n = 10, data from the same animal are shown with the same symbols). h is the ... data from the same animal are shown with the same symbols). io and the AQP4-positive rate (n=10, each animal uses the same symbol as in g). In g and h, the solid line reflects the linear regression, and the area between the two dashed lines reflects the 95% confidence interval. i is the k of each animal. io Proportion and AQP4 + Statistics of correlation coefficients in intratumor correlation analysis between proportions (n=10, black points). Bar heights and error bar widths represent the mean and standard deviation, respectively. ****p<0.0001, two-tailed t-test of Z-transformed correlation coefficients.

[0114] Mean k across tumors per animal io and when analyzed at the intertumor level using AQP4 expression, k io This linear relationship between and AQP4 expression still exists (h in FIG. 11) and k in FIG. io Combined with, the intratumoral AQP4 expression level of each rat glioma model can be accurately displayed (good correlation), which is k io We show that this method can accurately measure intra- and inter-tumor AQP4 heterogeneity.

[0115] As shown in Figure 2, k io can accurately represent the intratumoral AQP4 expression level of each rat glioma model, and a~j are the k in each experimental animal. io Value and AQP4 + A significant linear correlation between the percentage of AQP4 and the mean k of the six regions for each rat was obtained. Here, considering the ring-like distribution of AQP4, a series of concentric elliptical ROIs were used to divide each tumor slice (including tissue slices and MRI scan sections) into six concentric ring-like regions. io values ​​and their corresponding six AQP4 + Correlation analysis is performed using percentages. For k, in the TGN020-specific inhibition control group (Figure 12), the mean k across tumors was io showed no significant change between the two days of saline injections. Paired t-test, ns not significant, p=0.9588. Bar heights and error bar widths represent the mean and standard deviation of the mean, respectively. n=4.

[0116] The present invention also constructs an orthotopic glioma model by transplanting the C6 cell line into the right caudate putamen of SD rats, and obtains water exchange DCE-MRI data. The tumor size of the orthotopic glioma model (a, b in Figure 13) is much smaller than that of the subcutaneous glioma model, and the spatial resolution of MRI is limited, so k io The correlation analysis between and AQP4 expression is performed only among rats with orthotopic gliomas. As expected, the mean k across the tumors io A linear correlation was observed between the AQP4 expression level and the AQP4 expression level (R = 0.92, p < 0.01, n = 7, Fig. 13c). Note that the k io 13.0 seconds respectively -1 and 1.3 seconds -1 These results suggest that AQP4 expression in gliomas may be an imaging biomarker for the expression of AQP4. io This further demonstrates the robustness of

[0117] As shown in Figure 13, in a rat orthotopic model of C6 glioma, kio A linear correlation between AQP4 expression and a,b was observed, with low k io (a) and high k io (b) kA of AQP4 from two animals io Examples of profiles and typical IHC results. From top to bottom, a contrast-enhanced T1-weighted image (tumor location indicated by a white dotted circle), k-weighted images overlaid on the T1-weighted image, io A typical IHC result of AQP4 at the map and the position of the white arrow. MRI scale: 2 mm, IHC scale: 25 μm. c The mean k in the whole tumor in seven rats with orthotopic gliomas. io Proportion and AQP4 + A linear correlation is observed between the proportions. The solid line reflects the linear regression analysis and the two dashed lines represent the 95% confidence intervals. n=7.

[0118] Specific pharmacological inhibition of AQP4 reduces kappa glioma cell death in a rat glioma model. io Decreased To further verify that water-exchange DCE-MRI is a highly sensitive imaging method related to AQP4 expression in vivo, TGN020 is further used to specifically inhibit AQP4 function in a rat subcutaneous glioma model. As shown in Figure 14, TGN020 (3 mg / kg, dissolved in saline) is injected via the tail vein 15 minutes before water-exchange DCE-MRI scanning in each animal (n=9). The injection time of TGN020 is selected so that TGN020 achieves and maintains a stable collection level during water-exchange DCE-MRI acquisition. As a control group, water-exchange DCE-MRI measurement is performed in the same animals using the same amount of saline instead of TGN020 compared to the day before TGN020 treatment. The results show that the glioma tumors treated with TGN020 showed a significant improvement in k io It shows a significant decrease in the value of k io As shown in the map (Fig. 14b), the average k io is 6.3±0.6 seconds -1 (Saline injection group) 0.5 ± 0.7 seconds -1(TGN020-injected group) (43%, n=9, p=0.0246, Fig. 14c). Note that TGN020 also significantly reduced the mean K trans 0.060±0.015 minutes -1 (Saline) to 0.045±0.024 min -1 (TGN020), most tumor voxels still passed the error analysis and had a reliable k io This means that we can generate measurements, io This does not significantly affect the estimates (see example in Figure 14b). In the control group, each animal (n=4) was treated with saline on days 1 and 2, but there was no significant k io No change is shown (k in FIG. 12). The dose of TGN020 used here is relatively low due to safety concerns, but since TGN020 only partially inhibits AQP4, k io is not completely inhibited by TGN020. However, the current qualitative results suggest that AQP4 expression levels may be a biomarker for kappa-like tumors in gliomas. io The sensitivity and specificity of the assay can be fully demonstrated.

[0119] FIG. 14 shows the effect of specific pharmacological inhibition of AQP4 in a rat subcutaneous glioma model. io a is a timeline and schematic diagram of the animal experiment, which shows that the AQP4 specific inhibitor (TGN020) exerts an anti-kappa io a) Experimental process showing the effect of TGN020 on the cerebral cortex on the cerebral cortex. For the TGN020 group, the same animals are injected with saline on day 1 and with TGN020 on day 2. For the control group, the animals are injected with saline on both days. b) The k-weighted TSE images registered with T2-weighted TSE images. io A map is used to represent the effect of AQP4 inhibition on gliomas (C6) in the TGN020 group. Scale: 2 mm. c is the mean k across the tumor after AQP4 inhibition with TGN020. iodecreases by 43% (from 6.3 ± 0.6 s-1 to 3.5 ± 0.7 s-1). Two-tailed paired t-test, *p<0.05, where p=0.0246. Data points cover the corresponding boxes, bar heights and error bar widths represent the mean and standard deviation, respectively, n=9.

[0120] Human glioma io By guided stereotactic biopsy, io Further verifying the linear correlation between and AQP4 expression Water-exchange DCE-MRI was performed to visualize the k-images of human gliomas by bolus injection of a clinically approved contrast agent (Gd-DTPA, 0.1 mmol / kg body weight). io The present invention discloses intratumoral heterogeneity of AQP4. Instead of AQP4 spatial distribution, the present invention uses k io The aim of this study is to use a frameless neuronavigation technique. It is known that intraoperative brain transduction can affect the sampling accuracy of frameless neuronavigation techniques. This may not affect the spatial accuracy of radiopathological correlation. The present invention avoids the effect of brain metastases on the time from preoperative MRI to tumor sampling by frameless stereotactic biopsy techniques. Therefore, k io Quantitative analysis of images and biopsy specimens is expected to provide reliable assessments that allow accurate interpretation of radiological-histopathological correlation results at the voxel level.

[0121] This is an observational study, approved by the Institutional Review Board (IRB) of Shandong Provincial Hospital Affiliated to Shandong First Medical University, and written informed consent is obtained from each subject. Between May 2019 and August 2021, 21 patients with suspected glioma (both histological and molecular diagnosis confirmed by biopsy) were recruited according to the inclusion criteria of the IRB, using a Leksell® Model G stereotactic frame system (Elekta AB, Stockholm, Sweden). Before biopsy, water-exchange DCE-MRI was performed to determine the k io A map is obtained (c in Fig. 15). Next, k ioThe values ​​are integrated into the structural images used to guide stereotactic biopsy, and each patient has a different k io Collect multiple ROIs with values ​​(Figure 15a). After excluding two patients with poor image registration, the final analysis uses 45 biopsies from 19 patients (6 WHO I-II, 13 WHO III-IV, 10 of which have recurrent gliomas).

[0122] Typical K in the same section of the patient trans and k io The parameter diagrams are shown in Fig. 15b and c, where K trans and k io shows different spatial distributions, representing different pathophysiological conditions. Fig. 15d shows the results of AQP4-IHC of a typical biopsy sample, and k io =10.0 seconds -1 The sample of k io <0.5 seconds -1 The results of in situ ultrastructural detection using IHC combined with transmission electron microscopy (TEM)-based AQP4 antibody immunoelectron microscopy (IEM) also showed that k io The results show a high density of AQP4-positive gold nano-labeled spots in the biopsy samples with high k, supporting the conclusions of this paper at the subcellular level (Figure 15e). The AQP4+ fraction in each biopsy was further quantified, and then summed across all 45 biopsy samples, revealing k io Cellular percentage and AQP4 + A significant linear correlation was found between the cell ratio (R=0.97, p<0.0001) (FIG. 15 f). The linear regression equation is as follows:

number

[0123] This is because AQP4-related k io (14.9 seconds -1 ) is non-AQP4-related k io (i.e., k io Baseline, 0.2 seconds -1), and suggests that kio is primarily regulated by the AQP4-regulated pathway in human gliomas. io This linear relationship between IL-1 and AQP4 expression persists across multiple biopsy samples taken from the same patient, as shown in FIG. 15f, where data from each patient are marked with different symbols.

[0124] As shown in Figure 15, the k obtained from water-exchange DCE-MRI io The map discloses intratumoral AQP4 distribution in human gliomas, a, k io k in human gliomas using guided stereotactic biopsy io -AQP4 relationship. 19 glioma patients were randomly assigned to receive a randomized controlled trial. io Multiple biopsies with different signal intensities are acquired. Shown here is a schematic map of two points acquired using the same needle (overlaid on frame 70 of the water-exchange DCE-MRI). io (b, c) K of the same patient shown in (a) and a photograph of the stereotactic biopsy system. trans Maps b and k io Map c is registered to a water-exchange DCE-MRI image (frame 70) (a, b, c, staff length: 20 mm). (d, e) are the high (left) and low (right) k io Typical examples of AQP4 IHC (d, 100 μm) and IEM (e, 0.1 μm) from stereotactic biopsy. In IHC, AQP4 is marked in brown and cell nuclei in blue. In IEM, AQP4 is marked by gold particles (black arrows). Red arrows indicate glioma microfibrils. In both results, k io Tissues with higher AQP4 expression have higher AQP4 expression, and vice versa. io Cellular percentage and AQP4 + A linear correlation between the cellular proportions was observed in 45 stereotactic biopsy sites from 19 glioma patients. Data for each patient are marked with a different symbol. The solid line reflects the linear regression analysis and the two dashed lines represent the 95% confidence intervals.

[0125] Figure 16 also shows a special case of a patient with recurrent glioblastoma who underwent multi-needle radiofrequency ablation due to the large tumor area. In this particular case, the physician safely collects 10 biopsy samples along the planned trajectory. The present invention also shows that k in the 10 biopsy samples io We observed a good linear correlation between k and AQP4 (R = 0.96), which indicates that k in measuring AQP4 in a single case application io We further prove the accuracy of . Note that when averaging all biopsy data for each patient, the average k io It can be seen that a good correlation is maintained between the mean AQP4 and the mean AQP4 (R = 0.98, d in Figure 16).

[0126] Figure 16 shows the average statistical data of a special case and samples of a glioma patient, a patient with recurrent glioblastoma who underwent several biopsy surgeries due to the large tumor area. In this particular case, 10 biopsy samples are safely collected along the planned trajectory of the biopsy needle. a is the mean value of k io a) Examples of maps and locations where biopsy samples were taken (white arrows). b) Examples of AQP4 IHC results for the three biopsy sites shown in a. Scale: 25 μm. c) Ten stereotactic biopsy sites from this patient, with k io Cell fractionation and AQP4 + A linear correlation between the cell proportions is observed. io and AQP4 + The linear correlation between the cell percentages does not change across 19 data points for 19 patients. Here, for patients in whom multiple biopsies were obtained, the average result from all biopsies for that patient is used as the final representative biopsy result for that patient. The solid line reflects the linear regression analysis, and the two dashed lines represent the 95% confidence interval.

[0127] In summary, the above experiments demonstrated that the glioma-related markers of glioma are highly potent in vivo in two rat models of glioma, and in human clinical cases. io We show that AQP4 is a potent and sensitive imaging biomarker in human gliomas.

[0128] low k io reflects glioma treatment resistance Tumor recurrence is common after radiochemotherapy for glioma. io It is shown that C6 (i.e., low AQP4) cells may represent a therapy-resistant cell subtype in glioma. A closer look at the image texture of C6 cell nuclei after TMZ treatment reveals the following: Some cells show nuclear damage after TMZ treatment, while others show essentially intact and uniform nuclear structure, indicating that this type of cell subtype is resistant to the nuclear damaging anticancer drug TMZ (a in FIG. 17). The AQP4 expression of these cells resistant to TMZ chemotherapy was significantly lower than that of cells sensitive to TMZ treatment (c in FIG. 17). Meanwhile, during TMZ treatment, an increase in the number of quiescent slow-proliferating cells (SCC) was also simultaneously observed, indicating that these cells are more resistant to TMZ than fast-proliferating cells (FCC) (b in FIG. 17). Here, SCCs were marked by the cell fluorescence tracking method described above (see Methods) and positive cells (OG) were identified in which SCCs were marked with CellTrace reagents (Oregon Green (OG) for the C6 cell line, CellTrace Violet (CTV) for primary human glioma cells). + and CTV +) is used because SCCs do not divide or proliferate, and retain more OG and CTV fluorescent dyes, so they appear as bright images under the microscope. On the 3rd and 7th days of TMZ treatment, the SCC percentage increases from 6.0 (±0.4)% before TMZ treatment to 16.0 (±0.1)% (p<0.0001) and 28.9 (±0.3)% (d in Fig. 17, f in Fig. 9), respectively. To further prove the close relationship between AQP4 expression and cell cycle, on the 7th day of TMZ chemotherapy, EdU (5-ethynyl-2'-deoxyuridine) is used to rapidly label newly synthesized DNA while double staining for AQP4 to mark proliferative C6 cells. The corresponding flow cytometry test results show that AQP4 expression in SCC (low EdU) cells is lower than that in FCC cells (e in Fig. 17), and vice versa. In the cell morphology image of C6 cells on day 7 after TMZ treatment, several new long tail structures or “pseudopodia” (red arrows) (Figure 17b) were found, and these features are often considered to be treatment-resistant stem cell-like properties. Therefore, low k io (i.e., AQP4) tends to be associated with more SCC subtypes, slow proliferation rates and a treatment-resistant phenotype (Fig. 17 e, f).

[0129] More importantly, the therapeutic resistance status of clinical glioma biopsy samples is further characterized using the biomarker ZEB1 (zinc finger enhancer binding protein 1, a transcription factor that controls DNA damage) (the immunohistochemical and quantitative methods are the same as AQP4), and ZEB1 is used as a marker of therapeutic resistance of glioma. We show that low kappa-like protein (KF) from recurrent gliomas is a promising candidate for treatment of gliomas with low glioma resistance. io Biopsy samples were collected from the same subjects using high-k io We found that the IL-16 expression was higher in the IL-16 mice than in the biopsy samples (Fig. 17f), and that the low-k io This conclusion is further supported by flow cytometry results (Figure 17g) of primary cells from biopsy samples (see Methods), which showed that low k ioCells from the sample were io A higher proportion of cells from the sample exhibit an SCC (or CTV) phenotype.

[0130] As shown in Figure 17, low-k io (AQP4) reflects the phenotype of therapy-resistant glioma. a, Confocal microscopy images of AQP4 (red) and DAPI (blue) in C6 cells on day 3 after TMZ treatment, some of these cells have damaged cell nuclei (long-tailed arrows), while others show TMZ resistance and have more intact cell nuclear structures (short-tailed dashed arrows). From left to right, the scale is 25 μm and 5 μm, respectively. b, Results of cell tracking marked with OG, where slow-cycling cells (SCCs) show high signals (long-tailed arrows) in the fluorescent images. c, Quantitative results of AQP4 expression in therapy-sensitive and therapy-resistant cells of C6 on day 3 of TMZ treatment. Data are shown as mean ± standard deviation. *p<0.05, where p=0.0465, n=(4, 8). d, Bar graph shows the increase in the percentage of SCC (OG+) after TMZ treatment, ****p<0.0001, n=(7, 3, 3). e, Characterization of AQP4 and EdU expression results at day 7 of TMZ treatment using flow cytometry. The two-dimensional scatter plot shows two different cell phenotypes marked as FCC and SCC. Slowly proliferating glioma cells (SCC, i.e., low EdU) show lower AQP4 expression (red) than fast-proliferating cells (FCC). f, ZEB1 results obtained from stereotactic biopsy samples from recurrent glioma patients, showing low k io Sample (Right io <0.5 seconds -1 ) is a high-k io Sample (k io >0.5 seconds -1 ) shows a higher ZEB1 expression level than that of the control. Scale: 50 μm. As can be seen from the histogram statistics, k io For biopsy sites with <0.5 sec-1 (n=5), k ioMore abundant ZEB1 expression (ZEB1+ percentage) is shown compared to >0.5s-1 (n=5). Data for cell nuclei (blue) and ZEB1 (brown) are shown as mean ± SE. *pM0.05, where p<0.0260. In c, d, f, data points (blue point plots) cover the corresponding boxes (i.e., two-tailed unpaired t-test). g, CTV flow sorting results of primary cells from a clinical glioma biopsy sample obtained by stereotaxy, low k io Primary cells obtained from the sample (right) were io The samples show a much higher percentage of SCC (CTV+) than primary cells obtained from the samples (left); in e and g, the black lines represent the flow cytometry gating selection.

[0131] TMZ inhibits proliferation by inducing DNA double-strand breaks in glioma cells, killing FCC subtypes while sparing SCC subtypes. TMZ treatment of C6 cell line discloses similar phenomena, such as decreased proliferation and increased SCC rate. It is noted that cell subtypes with low AQP4 expression levels and long tail or pseudopodia morphology on day 7 of TMZ are often considered to be treatment-resistant stem cell-like cells (GSCs) in GBM. Similarly, U87MG cell line also shows decreased proliferation and downregulation of AQP4 during quiescence and death phases upon nutrient and oxygen deprivation. In these two microenvironments, the k obtained from water-exchange DCE-MRI. io and the dynamic regulation of AQP4 are closely related. A proportional increase in the number of SCCs with low AQP4 expression levels is observed upon TMZ treatment (C6) and in the quiescent reduced proliferation stage (U87MG). This phenomenon is reasonable because SCCs are low-proliferative cell subtypes and AQP4 expression levels are related to cell proliferation. More importantly, low AQP4 expression may slow down the transport rate across the cell membrane and protect SCCs from TMZ and other therapies. Indeed, it was observed that the C6 cell line with low AQP4 expression showed no or little damage to the cell nucleus under TMZ treatment, indicating its TMZ-resistant status. In addition, low kAg expression from human gliomas was observed to be significantly increased in the number of SCCs with low AQP4 expression levels. io(i.e., low AQP4) biopsy samples also show high expression of the treatment resistance biomarker ZEB1 and a high rate of SCC. Collectively, these results suggest that kappa B-cell markers may be used as a potential method to image SCC and predict treatment response. io Prove that.

[0132] The spatial and temporal heterogeneity of AQP4 expression level profiles in gliomas has great potential to facilitate precision therapy and predict therapeutic response. AQP4 expression is correlated with the degree of therapeutic resistance, and therefore k io The heterogeneity of AQP4 may indicate differences in treatment resistance across the glioma region. This AQP4 expression profile has great potential in the assessment of the pathological status of recurrent gliomas, further influencing therapeutic strategies and leading to many alternative methods for the treatment of recurrent gliomas. For example, the heterogeneity of AQP4 may indicate differences in treatment resistance across the glioma region. io If the tumor shows high AQP4 expression levels, secondary radiation therapy and chemotherapy are recommended, since high levels of AQP4 expression indicate that the tumor cells are sensitive to damage caused by radiation therapy and chemotherapy. io is measured, surgical intervention is recommended as these tumor cells are resistant to treatment with radiation or chemotherapy.

[0133] The present invention uses the water-exchange DCE-MRI efflux rate constant k of intracellular molecules across the cell membrane as a highly sensitive MRI biomarker of AQP4 expression level. ioBy specifically and quantitatively measuring AQP4, we can accurately measure AQP4 heterogeneity within and between glioma tumors. In a previous study, ADC on DWI also showed sensitivity to AQP4 silencing by RNA, but subsequent results questioned this result. As we all know, water exchange can contribute to ADC, but the use of ADC lacks the specificity of detecting transmembrane water exchange in vivo, and further questions arise because there are many possible biophysical mechanisms that cause changes in ADC, such as cell swelling or shrinkage, changes in cell density and shape, and the shape of the gap. Specifically, ADC has long been shown to mainly reflect cell density in human gliomas, further hindering the use of ADC as a specific biomarker of water exchange in gliomas. In clinical practice, water-exchange DCE-MRI can be used for glioma diagnosis in parallel with conventional MRI scanning without additional economic or time costs by adding a water-exchange DCE-MRI sequence during contrast agent (CA) injection. It should be noted that conventional MRI can also be used to identify glioma tumor areas from normal or other tissues, and then water-exchange DCE-MRI can be used to further analyze intratumoral AQP4. Since individual differences may still be observed in rats (Figure 12) and human gliomas, the water-exchange DCE-MRI scanning protocol used in the present invention can be further optimized to detect AQP4 in the tumor tissues. io The accuracy of estimation and AQP4 detection can be improved (FIG. 15f). Areas that need improvement include, but are not limited to, optimization of MRI sequences, CA injection, and SS model analysis. Aquaporin 4 (AQP4) plays a key role in determining the fate of gliomas, including tumor migration, proliferation, and therapeutic resistance. Although tissue biopsies can quantitatively characterize AQP4 expression levels in vivo, they cannot provide information about the heterogeneous distribution of AQP4 throughout the tumor. The present invention provides a non-invasive magnetic resonance biomarker (MRI) for detecting and mapping AQP4 in gliomas in vivo as a biomarker sensitive to AQP4 expression levels. ioAQP4 is visualized by MRI by quantitatively measuring AQP4-regulated water exchange across the plasma membrane. AQP4 is the major pathway regulating water exchange across the plasma membrane, k io We demonstrate for the first time that water-exchange DCE-MRI is a highly sensitive biomarker of AQP4 expression in gliomas. We then demonstrate that water-exchange DCE-MRI accurately detects the dynamic changes in AQP4 expression and function during different stages of glioma growth, temozolomide (TMZ) treatment, gene knockout, and AQP4 inhibition by TGN020, and captures the spatial heterogeneity of AQP4 expression within rat glioma models and human gliomas. io The cells showed greater resistance to therapy, suggesting the potential value of AQP4 profiling in assessing therapy resistance in gliomas.More importantly, this method can easily diagnose and evaluate gliomas radiologically on whole-tumor MRI, and this technique can greatly improve the accurate evaluation and treatment of human gliomas.

[0134] The above specific embodiments have described in detail the technical solutions and advantageous effects of the present invention, and it should be understood that the above contents are merely the most preferred embodiments of the present invention, and do not limit the present invention. Any modifications, additions, equivalent replacements, etc. made within the scope of the principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. 1. A method for measuring the flux rate of water molecules across a cell membrane, the method comprising: (1) setting dynamic contrast-enhanced magnetic resonance imaging parameters to measure a noise level during a dynamic contrast-enhanced magnetic resonance scan; (2) optimizing a flip angle in the dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) acquisition parameters by Monte Carlo simulation to set a new flip angle for dynamic contrast-enhanced magnetic resonance imaging; (3) scanning quantitative T1 magnetic resonance imaging; (4) scanning dynamic contrast-enhanced magnetic resonance imaging and injecting a contrast agent; (5) Full Shutter Speed ​​Model SSM ｆｕｌｌ A pixel analysis is performed for each pixel point in the tumor region using ｉｏ and obtaining a steady-state flux coefficient of water molecules from the cell to the interstitium (also called a steady-state flux coefficient of water molecules from the cell to the interstitium).

2. In step (2), the measured vascular contrast agent concentration C ｐ Then, by simulating the situation of real tissue and scanning parameters, DCE-MRI data is synthesized under different flip angles, and the time series signal of the synthesized DCE-MRI data is obtained. is a full shutter speed model SSM ｆｕｌｌ Based on the noise level estimated by the DCE-MRI experiment, Add white noise of the same noise level to the full shutter speed model SSM ｆｕｌｌ Noise was added using A nonlinear least-squares fitting is performed on the k ｉｏ Finally, k ｉｏ The flip angle for which the fitting result is closest to the simulation given value and the variance is the smallest is the optimal flip angle; In step (2), additional quantitative measurements of the spatial distribution of the actual flip angles are performed in order to optimize the flip angles under ultra-high magnetic fields; 2. The method of claim 1, wherein in step (3) quantitative T1 magnetic resonance imaging uses a short repeated time series of multiple flip angles.

3. In step (5), an automatic shutter speed analysis method is used to obtain the vascular leakage rate Ktrans of the contrast agent in the tumor region, and then the SSM ｆｕｌｌ Model fitting was performed for Ktrans>0.01 min. －１ The steady-state water flux coefficient k from the cell to the interstitium at each pixel point was calculated only in the tumor region. ｉｏ 2. The method of claim 1, further comprising obtaining a velocity of water molecules passing through a cell membrane.

4. Application of the flux rate of water molecules across cell membranes as a magnetic resonance imaging marker of gliomas to assess AQP4 expression levels.

5. The flux rate of water molecules across the cell membrane, k ｉｏ The linear relationship between AQP4 expression level and AQP4 cell positivity rate = (k ｉｏ -A) / B, In the formula, A is 0.1 to 0.2 seconds. －１ , B is 13.07 to 15.04 seconds －１ 5. The application of the flux rate of water molecules across a cell membrane according to claim 4, characterized in that:

6. 1. A method for measuring the flux rate of water molecules across a cell membrane for non-disease diagnostic purposes as a magnetic resonance imaging marker for glioma to assess AQP4 expression levels, said method comprising: (1) The steady-state flow rate coefficient k of tissue water molecules from cells to the interstitium ｉｏ Quantitatively measuring the (2) stereotactically taking a biopsy tissue using a biopsy planning system and quantifying the AQP4 expression level in the tissue; (3) k ｉｏ and establishing a linear relationship between the two based on the expression levels of AQP4. (4) The steady-state flow rate coefficient k of water molecules from the cells to the interstitium in the tissue to be measured ｉｏ and obtaining the AQP4 expression level of the tissue to be measured based on the linear relationship in step (3).

7. The measurement method includes: (1) setting dynamic contrast-enhanced magnetic resonance imaging parameters to measure a noise level during a dynamic contrast-enhanced magnetic resonance scan; (2) optimizing a flip angle in the dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) acquisition parameters by Monte Carlo simulation to set a new flip angle for dynamic contrast-enhanced magnetic resonance imaging; (3) scanning quantitative T1 magnetic resonance imaging; (4) scanning dynamic contrast-enhanced magnetic resonance imaging and injecting a contrast agent; (5) Full Shutter Speed ​​Model SSM ｆｕｌｌ A pixel analysis is performed for each pixel point in the tumor region using the steady-state flow rate coefficient k of water molecules from the cell to the interstitium at each pixel point. ｉｏ and obtaining (6) stereotactically taking a biopsy tissue using a biopsy planning system based on dynamic contrast-enhanced magnetic resonance imaging and quantifying the AQP4 expression level in the tissue; (7) k ｉｏ A linear regression analysis was performed on the AQP4 expression level and k ｉｏ obtaining a linear equation for (8) Repeat steps (3) to (5) and calculate k according to the linear equation in step (7). ｉｏ A method for measuring the flux rate of water molecules across a cell membrane for non-disease diagnostic purposes as a magnetic resonance imaging marker for glioma to evaluate AQP4 expression level as described in claim 6, characterized in that it includes a step of converting the image into an AQP4 expression level image to achieve intratumoral AQP4 expression imaging.

8. The method for measuring the flux rate of water molecules across cell membranes for non-disease diagnostic purposes as a magnetic resonance imaging marker for glioma to evaluate AQP4 expression level as described in claim 7, characterized in that in step (6), an AQP4 immunohistochemical photograph of the biopsy tissue is obtained to quantify the AQP4 expression level in the tissue.

9. In step (7), the AQP4 expression level and k ｉｏ The linear relationship between AQP4 cell positivity rate = (k ｉｏ -A) / B, In the formula, A is 0.1 to 0.2 seconds. －１ , B is 13.07 to 15.04 seconds －１ A method for measuring the flux rate of water molecules across a cell membrane for non-disease diagnostic purposes as a magnetic resonance imaging marker for glioma to evaluate the AQP4 expression level according to claim 7, characterized in that

10. Application of the flux rate of water molecules across cell membranes as a magnetic resonance imaging marker of gliomas in the preparation of a product for predicting the sensitivity of gliomas to radiotherapy and chemotherapy.

11. Application of the flux rate of water molecules across cell membranes according to claim 10, characterized in that the drug used in radiotherapy and chemotherapy is temozolomide.

12. 1. A system for measuring a glioma magnetic resonance imaging marker, the system comprising: The steady-state flow rate coefficient k of tissue water molecules from cells to the interstitium ｉｏ an image extraction and processing module for quantitatively measuring The k obtained by the image processing module ｉｏ and a post-processing module for establishing a linear relationship between the AQP4 expression level of the biopsied tissue. An image extraction module and an image processing module are performed on the tissue to be measured. ｉｏ and a prediction module for predicting the AQP4 expression level of the tissue to be measured based on the linear relationship of the post-processing module.

13. The measurement system comprises: a pre-processing module for setting dynamic contrast-enhanced magnetic resonance imaging parameters, measuring a noise level during dynamic contrast-enhanced magnetic resonance scanning, and optimizing a flip angle in the dynamic contrast-enhanced magnetic resonance imaging DCE-MRI acquisition parameters through Monte Carlo simulation to reset the flip angle of the dynamic contrast-enhanced magnetic resonance imaging; an image extraction module for scanning quantitative T1 magnetic resonance imaging, scanning dynamic contrast-enhanced magnetic resonance imaging, and injecting a contrast agent; Full shutter speed model SSM ｆｕｌｌ A pixel analysis is performed for each pixel point in the tumor region using the steady-state flow rate coefficient k of water molecules from the cell to the interstitium at each pixel point. ｉｏ an image processing module for acquiring the The k obtained by the image processing module ｉｏ and a post-processing module for establishing a linear relationship between the AQP4 expression level of the biopsied tissue. An image extraction module and an image processing module are performed on the tissue to be measured. ｉｏ and a prediction module for predicting the AQP4 expression level of the tissue to be measured based on the linear relationship of the post-processing module.

Images