Comprehensive ultrasonographic evaluation of normal and fibrotic kidneys in a mouse model with an ultra-high-frequency transducer
Article information
Abstract
Purpose
This study aimed to establish baseline morphological and functional data for normal mouse kidneys via a clinical 33 MHz ultra-high-frequency (UHF) transducer, compare the data with the findings from fibrotic mice, and assess correlations between ultrasonography (US) parameters and fibrosis-related markers.
Methods
This retrospective study aggregated data from three separate experiments (obstructive nephropathy, diabetic nephropathy, and acute-to-chronic kidney injury models). Morphological parameters (kidney size, parenchymal thickness [PT]) and functional (shear-wave speed [SWS], stiffness, resistive index [RI], and microvascular imaging-derived vascular index [VI]) were assessed and compared between normal and fibrotic mouse kidneys. Semi-quantitative histopathologic scores were calculated and molecular markers (epithelial cadherin), Collagen 1A1 [Col1A1], transforming growth factor-β, and α-smooth muscle actin [α-SMA]) were evaluated using western blots. Correlations with US parameters were explored.
Results
Clinical UHF US successfully imaged the kidneys of the experimental mice. A three-layer configuration was prevalent in the normal mouse kidney parenchyma (34/35) but was blurred in most fibrotic mouse kidneys (33/40). US parameters, including size (11.14 vs. 10.70 mm), PT (2.07 vs. 1.24 mm), RI (0.64 vs. 0.77), VI (22.55% vs. 11.47%, only for non-obstructive kidneys), SWS (1.67 vs. 2.06 m/s), and stiffness (8.23 vs. 12.92 kPa), showed significant differences between normal and fibrotic kidneys (P<0.001). These parameters also demonstrated strong discriminative ability in receiver operating characteristic curve analysis (area under the curve, 0.76 to 0.95; P<0.001). PT, VI, and RI were significantly correlated with histological fibrosis markers (ρ=-0.64 to -0.68 for PT and VI, ρ=0.71-0.76 for RI, P<0.001). VI exhibited strong negative correlations with Col1A1 (ρ=-0.76, P=0.006) and α-SMA (ρ=-0.75, P=0.009).
Conclusion
Clinical UHF US effectively distinguished normal and fibrotic mouse kidneys, indicating the potential of US parameters, notably VI, as noninvasive markers for tracking fibrosis initiation and progression in mouse kidney fibrosis models.
Introduction
Kidney fibrosis is a hallmark manifestation of chronic kidney disease (CKD) and is pathologically marked by tubule atrophy, chronic interstitial inflammation, fibrogenesis, glomerulosclerosis, and vascular rarefaction [1]. Animal models, particularly mice, are extensively used in translational research on kidney fibrosis, both in vitro and in clinical studies. Mice are favored due to their short lifespan, ease of breeding, relatively low cost [2], and the simplicity of using transgenic technology [3]. However, their small kidney size renders kidney biopsy impractical, necessitating euthanasia for tissue collection.
Ultrasonography (US) is a representative, noninvasive, real-time imaging modality that is used to evaluate kidney disease without the risks associated with radiation. It is well established that parameters such as kidney size, resistive index (RI), and interval deterioration are indicative of kidney function, disease chronicity, and can predict disease progression in patients with CKD [4-6]. Recent advancements in microvascular flow imaging have enabled detailed assessments of kidney parenchymal flow [7]. Studies have also investigated the use of shear-wave ultrasound elastography (SWE) in assessing kidney fibrosis, both in human biopsy samples and in a rabbit model of obstructive uropathy [8-10]. Furthermore, recent findings have shown that repeated US does not adversely affect the welfare of laboratory mice [11]. Therefore, serial US measurements of these parameters can provide additional insights into disease development and progression without necessitating excessive mouse sacrifice, thereby upholding the ethical standards of animal experiments.
Previous US examinations in mice utilized a dedicated preclinical ultrasound system equipped with an ultra-high-frequency (UHF) transducer (30-40 MHz for mice). Contrast-enhanced US (CEUS) was employed to assess perfusion and vascularity [12]. Recently, a 33-MHz linear transducer has been integrated into a clinical ultrasound system (Aplio i800, Canon Medical Systems). This integration is expected to provide clinicians with easier access to experimental small animal ultrasound, even in the absence of specialized preclinical ultrasound equipment. However, there is currently no reliable documentation of US findings in normal mouse kidneys, nor are there established reference values for the US parameters used in serial examinations of experimental mice. To the authors’ knowledge, no previous experimental mouse studies have employed SWE to assess kidney fibrosis or used microvascular flow imaging as an alternative to CEUS for evaluating kidney perfusion status.
Thus, the aim of this study was twofold: first, to describe the US findings and morphological (size, parenchymal thickness [PT]) and functional (shear-wave speed [SWS], stiffness, RI, and vascular index [VI] from microvascular imaging) parameters of normal mouse kidneys using a clinical ultrasound system equipped with a UHF transducer; second, to compare the US findings and parameters from various kidney fibrosis mouse models with their histological results and molecular markers to establish correlated ultrasound biomarkers for kidney fibrosis.
Materials and Methods
Compliance with Ethical Standards
The animal models were approved by the Institutional Animal Care and Use Committee of the authors’ affiliated hospital (2022-0004, SNU1711115, ILAR-01-21-1010).
Experimental Animal Disease Models
Three different mouse models were used in this study: an obstructive nephropathy (ON) model using unilateral ureter obstruction (UUO), a diabetic nephropathy (DMN) model, and an acute kidney injury to CKD transition (AKI to CKD transition) model.
Seven- to eight-week-old male wild-type mice (C57BL/6; B6), weighing approximately 200-220 g, were obtained and acclimated for seven days prior to the start of the experiments. Anesthesia was administered using xylazine (Rompun; 10 mg/kg; Bayer, Leverkusen, Germany) and Zoletil (30 mg/kg; Virbac, Carros, France) in all cases.
For the ON model, the UUO method was utilized. A left flank incision was made, and the left proximal ureter was exposed. Subsequently, the exposed proximal ureter was ligated at two different points using 5-0 nylon to minimize the risk of incomplete obstruction. One to two weeks after the procedure, the mice were euthanized by exsanguination under the same anesthesia protocol used during the procedure.
For the DMN model, the unilateral nephrectomy–streptozotocin (STZ) method was utilized. One week following unilateral nephrectomy, STZ (Sigma-Aldrich, St. Louis, MO, USA) was administered intraperitoneally to each mouse for 5 consecutive days. Tail vein blood glucose levels were monitored to confirm the onset of diabetes, defined as fasting blood glucose levels exceeding 300 mg/dL. The mice were euthanized 24 weeks after diabetes induction [13].
For the AKI to CKD transition model, the unilateral ischemia-reperfusion injury method was used. The mice were anesthetized using the same protocol as in the previously described ON model. A left flank incision was made to expose the vascular pedicle of the left kidney, which was then clamped for 27 minutes using microaneurysm clamps (Roboz Surgical Instrument Co., Gaithersburg, MD, USA). Six weeks after the procedure, the mice were euthanized by exsanguination under anesthesia.
Forty mice were utilized in the disease models mentioned: 15 for the ON model, three for the DMN model, and 22 for the AKI to CKD transition model. Subsequent analyses involved 15 UUO kidneys and 25 non-UUO kidneys. Additionally, 29 sham-operated mice, which underwent only a flank incision and closure, served as normal controls for each disease model. The total study population included 69 mice. The kidneys of seven normal mice were also evaluated, bringing the total number of kidneys assessed to 76 (Fig. 1).
Schematic diagram of the animal population of this study.
ON, obstructive nephropathy; UUO, unilateral ureter obstruction; DMN, diabetic nephropathy; IP, intraperitoneal injection; AKI, acute kidney injury; CKD, chronic kidney disease; UIRI, unilateral ischemic-reperfusion injury.
Ultrasound Examinations: Grayscale, Doppler, and Microvascular Imaging
Ultrasound examinations were conducted using a clinical US scanner equipped with a UHF linear transducer (i33LX9, center frequency 33 MHz) and a high-resolution linear transducer (i18LX5, center frequency 18 MHz). A board-certified radiologist performed the procedure just before the animal was sacrificed. Initially, the i18LX5 transducer was used to delineate the location, contour, and layer structure of the targeted kidney. Subsequently, detailed grayscale and color Doppler imaging, including RI measurement, was carried out using the i33LX9 transducer. The kidney size was determined by measuring the longest distance from pole to pole. PT was defined as the distance from the cortex-perikidney fat interface (capsule) to the sinus-pyramidal apex interface [14], and was measured at the midportion of the long-axis view, which displayed the greatest kidney length. RI was calculated as the average of two separate measurements taken from different interlobar arteries of the targeted kidney.
Microvascular imaging was conducted using an advanced clutter filter designed to extract low-velocity vessels (Superb Microvascular Imaging, SMI, Canon Medical Systems, Otawara, Japan), accompanied by VI measurement. The settings for SMI included a low wall filter, high frame rates (44-55 frames per second), and a color gain adjusted to minimize color noise while maintaining visibility of the subcapsular vessels (within a range of 38%-40%). VI was calculated as the ratio of the area covered by color-coded pixels to the total pixel area of the kidney, determined by manually drawing the region of interest (ROI) along the kidney contour.
Ultrasound Shear-Wave Elastography
SWE was performed using the i18LX5 with the display setting at 90 kPa. Gel was liberally applied to the skin over the area of interest, and the transducer was placed in direct contact, then slowly lifted vertically. During the examination, the transducer was secured on a direction-adjustable stand with a clamp and gently elevated until the skin was slightly taut due to the gel's viscosity, minimizing the impact of manual compression (Supplementary Fig. 1). The thickness of the ultrasonic gel was consistently maintained at approximately 0.5 cm during imaging to ensure constant pressure from the gel. SWE was carried out in one-shot mode, which involves sending a main pulse across one frame to measure the resultant SWS and tissue elasticity. Two or three linear ROIs, each 1 mm in diameter, were positioned based on the data from the propagation map and the variation (confidence) map. Following the vendor's recommendation, the radiologist placed the ROIs where the contour lines on the propagation map were most parallel and dense, and where variation on the confidence map was lowest. To minimize the effects of anisotropy, the ROIs were positioned as close to the interpolar area as feasible [15]. After four separate main pulse shots, 10 to 12 ROIs from four measurements were obtained in a single measurement session. Each ROI provided the mean SWS (m/s) and the mean stiffness value (kPa), calculated using the equation
Masson trichrome (MT) staining and ultra-high-frequency ultrasonography (US) image of a sham-operated mouse kidney.
A. Normal-looking glomerulus and tubules are seen in MT-stained mouse kidney (×200). B, C. US images of a normal mouse kidney with 33 MHz (B) and 18 MHz (C) linear transducers are presented. The layer distinction is clearer in the image taken with the 33 MHz transducer. B, D, E. Measurements of resistive index (D) and vascular index (E) were also performed using a 33 MHz ultra-high-frequency transducer. F. Shear-wave elastography was performed using an 18 MHz high-frequency transducer.
Masson trichrome (MT) staining and ultra-high-frequency ultrasonography (US) image of a fibrotic mouse kidney 35 days after unilateral ischemic-reperfusion injury.
A. An irregularly shrunken kidney was seen in an autopsy. B. Glomerular sclerosis and tubular atrophy were seen in an MT-stained mouse kidney with blue-colored interstitial fibrosis (×200). C. In a US image using 33MHz linear transducer, the shrunken kidney is well visualized and the size is the same as that of the autopsy finding. D-F. The resistive index was higher (D), vascular index was lower (E), and stiffness was higher (F) compared to normal mouse kidney in Fig. 2.
Histology and Histopathologic Scoring
Paraffin-embedded kidney sections, 4 μm thick, were stained using Masson’s trichrome according to the standard protocol. A board-certified pathologist (J.H.P) conducted semi-quantitative histopathological scoring based on the standardized grading system for chronic changes in native kidney biopsy specimens [17]. This system includes scores for glomerulosclerosis (GS), interstitial fibrosis (IF), tubular atrophy (TA), and atherosclerosis (CV), along with a sum score that represents the total renal chronicity score. The CV score uses a binary scale (0 or 1), whereas the other scores utilize a 4-point scale ranging from 0 to 3.
Western Blotting
Western blotting was conducted following the standard protocol of the authors’ institution [18]. Protein lysates were extracted from homogenized kidney tissues and prepared at equal concentrations for electrophoretic separation using a Bicinchoninic acid protein assay kit. The lysates were denatured and electrophoresed in glycine-SDS buffer (LPS solution), then transferred onto PVDF membranes (Millipore Corporation, Seoul, Korea). The membranes were blocked using 5% skim milk (Becton Dickinson Rowa, France) with 2% BSA buffer, and incubated at 4°C overnight with primary antibodies: epithelial cadherin (Ecad), collagen 1A1 (Col1A1), transforming growth factor-β, α-smooth muscle actin (α-SMA), and glyceraldehyde 3-phosphate dehydrogenase. The following day, the membranes were treated with secondary antibodies at room temperature for 1 hour, and the protein bands were detected using the electrochemiluminescence method (Advansta, San Jose, CA, USA). Band normalization and quantification were carried out using ImageJ software (National Institutes of Health). Since not all kidneys analyzed by western blots underwent US examinations, the bands included in the analysis were marked in blue (normal control) or red (fibrotic) for each disease model group (Supplementary Fig. 2).
Statistical Analysis
US parameters, including kidney size, PT, mean RI, SWS, stiffness, and VI, were compared using a t-test and one-way analysis of variance. The ability of each US parameter to distinguish kidney fibrosis was assessed through receiver operating characteristic (ROC) curve analysis and is presented as the area under the ROC curve (AUC). The correlations between US parameters, molecular markers, and histological scores were determined using Spearman correlation analysis. The feasibility of SWE in experimental mouse kidneys was evaluated in two ways: by comparing the mean of two repeated measurements using the t-test and by assessing repeatability using intraclass correlation coefficients (ICC). A P-value of less than 0.05 was considered statistically significant. Statistical analyses were conducted using MedCalc version 22.006 software (MedCalc Software Ltd., Ostend, Belgium).
Results
Feasibility of SWE in Experimental Mouse Kidneys
Among the 76 kidneys, six were not evaluated by SWE (5 fibrotic kidneys due to technical failure and 1 normal kidney due to limited laboratory rental time). Technical failures were attributed to overly thin kidney parenchyma caused by hydronephrosis (n=1, UUO) or significantly reduced kidney size due to extensive fibrosis (n=4, non-UUO) (Table 1). There was no significant difference between the means of the two repeated measurements (the mean values of the first and second measurements were 1.86±0.32 m/s and 1.88±0.30 m/s, and 10.57±4.10 kPa and 10.77±3.85 kPa, respectively; P>0.05). The ICC values were 0.96 (95% confidence interval [CI], 0.93 to 0.98) for SWS, and 0.96 (95% CI, 0.94 to 0.98) for stiffness, indicating good repeatability. Tables 1 and 2 summarize the acquisition feasibility and suggestive reference values for normal and fibrotic mouse kidneys, respectively.
Feasibility of experimental mouse kidney US with clinical ultrasound transducers
Suggested reference values
Grayscale US
In grayscale US using a 33 MHz UHF transducer, a three-layer configuration was observed in most of the normal mouse kidney parenchyma (Fig. 4), both completely and discontinuously (35 of 36 normal kidneys). The inner medulla, which appeared low-echoic, was often seen discontinuously and was absent in one kidney. When using an 18 MHz linear transducer, the layered structure was poorly differentiated compared to images obtained with the 33 MHz transducer (Fig. 2B, C) in all normal mice. In fibrotic kidneys, regardless of the type of injury, the kidney parenchyma became heterogeneously echoic, and corticomedullary differentiation was obscured in most cases due to a decrease in the echo of the outer layer (33 of 40 fibrotic kidneys) (Fig. 3C).
Correlation of ultra-high-frequency ultrasonography image and histologic structure of normal mouse kidney.
A, B. Axial scan of a normal mouse kidney using a 33 MHz linear transducer (A) and Masson trichrome staining (×40) (B) are shown. In a ultrasonography image, the outer hyperechoic layer (red bracket) corresponds to the cortex (C) and outer medulla (OM), the middle hypoechoic layer (green bracket) corresponds to the inner medulla (IM), and the inner hyperechoic layer (yellow bracket) corresponds to the renal papilla (P).
The mean size and PT of normal mouse kidneys were significantly greater than those of the fibrotic kidneys (P<0.001). The non-UUO kidneys were significantly smaller than both the normal and UUO kidneys (P<0.001) (Fig. 5A). There was no significant difference in PT between the UUO and non-UUO kidneys (Fig. 5B).
Comparison of ultrasonography parameters among normal, unilateral ureter obstruction (UUO), and non-UUO mouse kidney groups (A-F).
PT, parenchymal thickness; RI, resistive index; VI, vascular index; SWS, shear-wave speed. a)P-value between normal kidneys and all fibrotic kidneys, P<0.001. b)P-value between UUO and non-UUO kidneys P<0.001.
Doppler US, Microvascular Imaging, and SWE
RI measurement was successfully conducted in 69 of the 76 kidneys, with measurement failures occurring in one normal kidney and six fibrotic kidneys. The mean RI of the 35 normal kidneys was 0.64±0.05, with a range of 0.54-0.74. This was significantly lower than the RIs of both UUO and non-UUO kidneys (P<0.001) (Fig. 5C). The mean RIs for the 13 UUO kidneys and 21 non-UUO kidneys were 0.76±0.04 (range, 0.70 to 0.87) and 0.76±0.07 (range, 0.67 to 0.92), respectively.
The VI could not be effectively measured in UUO kidneys due to the low color signal and thin parenchyma. In seven non-UUO kidneys, the VI measurements were unattainable because of insufficient effective signals. Additionally, measurements could not be conducted on six normal kidneys due to scheduling constraints. The VI in normal kidneys was significantly higher than in fibrotic (non-UUO) kidneys (P<0.001) (Fig. 5D).
Normal kidneys exhibited significantly lower SWS and stiffness compared to fibrotic kidneys (P<0.001). However, there were no significant differences in SWS and stiffness between UUO and non-UUO kidneys (Fig. 5E, F).
Performance of US Parameters for Detecting Kidney Fibrosis
The ROC curves identified that an SWS of 1.76 m/s and a stiffness of 9.22 kPa were the discrimination points between normal and fibrotic kidneys, demonstrating a sensitivity of 85.7% and a specificity of 91.4% (AUC, 0.92 and 0.92, respectively). An RI of 0.71 and a VI of 17.6% also served as discrimination points, with sensitivities of 88.3% and 95.5%, and specificities of 81.0% and 86.8%, respectively (AUC, 0.95 and 0.93, respectively). A PT of 1.7 mm was identified as the discrimination point in grayscale US, with a sensitivity of 78.4% and a specificity of 91.7% (AUC, 0.89). Due to the inevitable increase in the size of the UUO kidneys as hydronephrosis progressed, they were excluded from the ROC analysis. A size of 9.8 mm was determined as the discrimination point for normal and non-UUO fibrotic kidneys in grayscale US, with a sensitivity of 60.0% and a specificity of 100.0% (AUC, 0.76). The P-values for all AUCs were <0.001.
Correlation of Histologic Fibrosis Scores with US Parameters
UUO kidney size was excluded from the correlation analysis for the reasons previously mentioned. PT and VI demonstrated strong negative correlations with GS, IF, TA and their sum point (SUM) (ρ=-0.64 to -0.68, P<0.001 for PT; ρ=-0.66 to -0.68, P<0.001 for VI). RI exhibited strong positive correlations with histologic fibrosis parameters (ρ=0.71 to 0.76, P<0.001). Elastography markers, including SWS and stiffness, also showed strong correlation with histologic fibrosis parameters (ρ=0.67 to 0.75, P<0.001) (Fig. 6). Since the CV score was a binary category, the averages of each US parameter according to the two CV categories were compared using the t-test. All US parameters, except for kidney size, demonstrated significant differences between the two CV categories (P<0.001). Notably, PT displayed a more pronounced correlation with histological fibrosis score parameters than with kidney size.
Correlations among ultrasonography findings, histological characteristics, and molecular markers.
The top number in each cell is the Spearman correlation coefficient. GS, glomerulosclerosis score; SUM, histological sum score; IF, interstitial fibrosis score; TA, tubular atrophy score; SWS, shear-wave speed; RI, resistive index; α-SMA, α-smooth muscle actin; Col1A1, collagen 1A1; TGF-β, transforming growth factor-β; Ecad, epithelial cadherin; PT, parenchymal thickness; VI, vascular index.
Correlation of Molecular Markers with US Parameters
VI showed strong negative correlations with Col1A1 (ρ=-0.76, P=0.006) and α-SMA (ρ=-0.75, P=0.009). PT had a good correlation with Ecad (ρ=0.59, P=0.012). Other US markers did not show significant correlations with the molecular fibrosis markers (Fig. 6).
Discussion
This study explored the feasibility of using a clinical ultrasound system equipped with a UHF transducer for preclinical imaging of mouse kidneys. This included grayscale, Doppler, microvascular, and SWE imaging modalities. Detailed descriptions of the grayscale US findings for mouse kidneys using a 33 MHz clinical UHF transducer are provided. Additionally, this study investigated the morphological and functional US parameters for both normal and fibrotic kidneys and assessed the efficacy of these parameters in distinguishing fibrosis, using AUC values. The correlation of each US parameter with histological grading and expression of molecular markers was also evaluated.
An important finding was that all mouse kidneys were effectively visualized using the clinical UHF transducer, which benefited from clear layer demarcation and the use of advanced imaging techniques such as microvascular imaging and SWE. The 35 mm transducer footprint and the 15 mm center frequency penetration proved sufficient for evaluating mouse kidneys. The wide bandwidth, advanced beamforming, and the UHF itself may contribute to enhanced spatial resolution and penetration [19]. Notably, the 18 MHz transducer lacked clarity in corticomedullary differentiation during SWE examination, highlighting the superior observation provided by the UHF for this purpose.
Moreover, SWE demonstrated high repeatability (ICC=0.96) compared to previous studies (ICC=0.76-0.85) [20,21]. Since animals are subjected to ultrasound post-anesthesia, confounding factors such as movement are minimized. The use of a clamp to secure the transducer and a height-adjustable table likely contributed to improved repeatability by reducing manual compression.
Regarding the correlation of histologic fibrosis scoring, SWE parameters demonstrated a good correlation with histologic scoring (TA, IF, GS, sum, and grade) and exhibited distinct differences according to the CV category. However, inconsistent findings have been reported regarding the relationship between SWE and kidney fibrosis. Several studies have reported an increase in SWS and stiffness with the progression of fibrosis [8,9,22,23], while others have found a negative correlation between SWE parameters and the degree of fibrosis [10,24]. These conflicting results are likely due to the influence of kidney blood flow [25-27]. In this study, SWE parameters were significantly correlated with the sum score and grade. However, in a subgroup analysis that included only fibrotic kidneys, there was no significant correlation between SWE and histological fibrosis scores (data not shown), likely due to the previously mentioned flow effects. Despite this, SWE parameters demonstrated strong discriminative ability in differentiating between normal and fibrous kidneys, as supported by the ROC analyses and several meta-analyses of human studies [28-30]. Additionally, multiple animal studies [8,31], including studies conducted by the present research group, have shown a notable correlation between kidney fibrosis and SWE parameters, likely due to effective control over the confounding factors mentioned earlier. In summary, SWE holds promise for evaluating the occurrence of kidney fibrosis in patients and for assessing its progression and severity in well-controlled animal models.
Vascular parameters, including RI and VI, demonstrated fair to good correlations with fibrosis scores and grades, effectively distinguishing between normal and fibrotic kidneys. These results indicate that the RI may serve as a predictor for the development and progression of kidney fibrosis.
Few studies have directly visualized the kidney microvasculature using US in both animals and humans, revealing significant reductions in microvascular density in mouse and rat models of kidney ischemia [32,33]. Super-resolution ultrasound, employing a clinical ultrasound scanner and microbubbles, has demonstrated the ability to display the kidney microvasculature in both transplanted and native human kidneys [34,35]. To the authors’ knowledge, no previous studies have explored the diagnostic performance or the correlation between vascular imaging (VI) and kidney fibrosis. SMI and VI align with previous efforts to visualize and quantify the kidney microvasculature. The findings of this study indicate a strong correlation between VI and kidney fibrosis, suggesting that VI may accurately reflect changes in kidney microvascular density associated with the onset and progression of kidney fibrosis. This is further supported by a strong correlation between VI and molecular markers of fibrosis (Col1A1 and α-SMA).
A previous study on US and kidney fibrosis [36] showed that kidney size (length) and PT correlated with histopathological findings, while cortical thickness did not. The kidney cortex is relatively thin and can sometimes be difficult to distinguish from the medulla, making accurate and reproducible measurements challenging [15,36]. Consequently, the authors opted to measure PT instead of cortical thickness. However, in this study, kidney size did not exhibit a significant correlation with fibrosis parameters, whereas PT demonstrated a fair correlation. The mouse flank area was significantly smaller than the footprint of the UHF transducer, leading to shadowing from the rib and adjacent bowel gas that often obscured the bilateral polar areas of the kidney. Therefore, in mouse US, unlike in human US, measurements of kidney size may be less accurate than those of PT. This correlation might be strengthened by using a transducer with better penetrability.
Regarding the correlation with molecular markers, Ecad was found to correlate with PT. Decreased PT is associated with kidney fibrosis. Ecad, a marker of the epithelial-to-mesenchymal transition (EMT), represents one of the various pathological mechanisms underlying kidney fibrosis [37,38]. Therefore, the relationship between Ecad and PT does not imply that these factors directly reflect each other; instead, it suggests that EMT significantly influences the reduction of kidney parenchyma among the different mechanisms of kidney fibrosis. The SWE parameters did not show any correlation with molecular markers, indicating the presence of multiple confounders that affect SWE measurements.
From the standpoint of animal welfare, US parameters can serve as noninvasive indicators of fibrosis induction in an experimental mouse kidney fibrosis model. Specifically, serial VI measurements allow the assessment of fibrosis progression within a single experimental group, eliminating the need to prepare multiple groups of mice for sacrifice at various time points following fibrosis induction.
This study has several limitations. First, there are no established guidelines for SWE in either mouse or human kidneys. This study was conducted with the goal of adhering to previous animal studies [8,31], and special care was taken to avoid excessive compression. To minimize fine movements of the transducer, it was mounted on a direction-adjustable stand. To maintain a constant distance between the transducer and the mouse, a height-adjustable table was used that allowed precise vertical positioning of the mouse. As mentioned earlier, the thickness of the gel was continuously monitored with real-time US, and efforts were made to keep this thickness constant. However, factors such as pulsation and breathing may have impacted the quality of the test. As previously discussed, the measurement of kidney size was compromised by the poor sonic window, and VI could not be effectively obtained in UUO kidneys with thin parenchyma. The correlation between each US or molecular parameter and kidney function could not be evaluated due to the absence of serum creatinine data. Despite the varying schedules for each experiment, all US procedures were performed by the same radiologist. Therefore, interobserver agreement (reproducibility) could not be assessed, although the consistency of the measurements might be assured. This study is limited in that assessing reproducibility remains a significant challenge in studies involving SWE.
In summary, the US evaluation of experimental mouse kidneys using a 33 MHz clinical UHF transducer was both feasible and accessible. The kidney parenchyma in normal mouse kidneys exhibited a three-layer demarcation when imaged with the UHF transducer, a feature typically absent in fibrotic kidneys. All US parameters demonstrated good-to-excellent performance in distinguishing fibrotic kidneys, and the suggestive criterion value for each parameter was provided. Notably, US parameters, particularly VI, could serve as noninvasive indicators for detecting and monitoring fibrosis in a mouse kidney fibrosis model.
Notes
Author Contributions
Conceptualization: Lee MS, Cho JY, Moon MH. Data acquisition: Lee MS, Lee JP, Lee J, Shin N, Jin W. Data analysis or interpretation: Lee MS, Moon MH, Lee JP, Lee J, Cho A. Drafting of the manuscript: Lee MS. Critical revision of the manuscript: Lee MS, Cho JY, Moon MH, Lee JP, Lee J, Shin N, Jin W, Cho A. Approval of the final version of the manuscript: all authors.
No potential conflict of interest relevant to this article was reported.
Acknowledgements
This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (Grant number: 2015R1D1A1A01058519), Focused clinical research grant-in-aid from the Seoul Metropolitan Government Seoul National University Boramae Medical Center (04-2022-0013) and a Korean Fund for Regenerative Medicine (KFRM) grant funded by the Korean government (the Ministry of Science and ICT, the Ministry of Health & Welfare) (23B0104L).
The authors deeply appreciate Professor Jeong Hwan Park in the Department of Pathology of the authors’ affiliated institution for contributing to this article by evaluating histologic fibrosis scoring.
Supplementary Material
Supplementary Fig. 1.
Shear-wave elastography of mouse kidney (https://doi.org/10.14366/usg.24024).
Supplementary Fig. 2.
Western blot bands of the three different kidney fibrosis mouse models used in this study (https://doi.org/10.14366/usg.24024).
References
Article information Continued
Notes
Key point
Ultrasonography (US) parameters that meaningfully reflect kidney fibrosis are still unclear in clinical and translational research. Reference values for various US parameters of normal and fibrotic mouse kidneys are proposed for translational research. US, especially microvascular imaging, can be helpful in the detection and follow-up of kidney fibrosis.