Open Access
Add to BibSonomy
Abstract
Hepatocellular carcinoma is one of the most lethal cancers worldwide, causing almost 700,000 deaths annually. It mainly arises from cirrhosis, which, in turn, results from chronic injury to liver cells and corresponding fibrotic changes. Although it is known that chronic liver injury increases the elasticity of liver tissue, the role of increased elasticity of the microenvironment as a possible hepatocarcinogen is yet to be investigated. One reason for this is the paucity of imaging techniques capable of mapping the micro-scale elasticity variation in liver and correlating that with cancerous mechanisms on the cellular scale. The clinical techniques of ultrasound elastography and magnetic resonance elastography typically do not provide micro-scale resolution, while atomic force microscopy can only assess the elasticity of a limited number of cells. We propose quantitative micro-elastography (QME) for mapping the micro-scale elasticity of liver tissue into images known as micro-elastograms, and therefore, as a technique capable of correlating the micro-environment elasticity of tissue with cellular scale cancerous mechanisms in liver. We performed QME on 13 freshly excised healthy and diseased mouse livers and present micro-elastograms, together with co-registered histology, in four representative cases. Our results indicate a significant increase in the mean (×6.3) and standard deviation (×6.0) of elasticity caused by chronic liver injury and demonstrate that the onset and progression of pathological features such as fibrosis, hepatocyte damage, and immune cell infiltration correlate with localized variations in micro-elastograms.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Liver cancer is the fourth largest cause of cancer-related death globally with a 5-year survival rate of 18% [1]. Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer [2]. In 90% of cases, HCC develops in patients with chronic liver disease, whereby liver cells undergo continual cycles of necrosis and regeneration [1,3]. Although the liver has a remarkable regenerative capacity, chronic liver injury may lead to the progression from initial fibrotic wound healing responses to the development of hepatic fibrosis [4,5], which typically results in increased liver elasticity, i.e., higher stiffness [6]. Importantly, liver elasticity routinely measured in clinics on the millimeter to centimeter scale using ultrasound elastography and magnetic resonance elastography is a noninvasive and indirect measure of liver fibrosis that is positively correlated with the risk of HCC development [7–10]. Several studies have suggested important roles for fibrosis in driving HCC, through mechanisms such as stiffness-induced signaling pathways, activation of hepatic stellate cell-derived extracellular matrix and immunosuppression [11,12]. Other studies demonstrated the correlation between the distribution and differentiation of liver cancer cells and the elasticity of the microenvironment [13,14]. However, such studies are either based on in vitro experiments on cells [14] or point elasticity measurements on tissue in mouse models of liver disease [13]. An elastography technique that can map the depth-resolved elasticity of liver micro-structures over millimeter-scale fields-of-view (FOV), covering several liver lobules, hepatocyte structures, bile ducts, and sinusoids, has potential to further improve our understanding of the role of microenvironment elasticity in driving cellular mechanisms resulting in HCC development. Although, ultrasound elastography and magnetic resonance elastography are routinely used clinically to diagnose chronic liver disease [5,15–17], they typically provide a spatial resolution in the millimeter range and, therefore, are of limited use in studying variations in the elasticity of liver micro-structures caused by chronic liver injury. Furthermore, whilst atomic force microscopy (AFM) provides subcellular resolution [13], it is typically limited to mapping the surface mechanical properties of a limited number of cells, making it unsuitable to map the elasticity of the liver micro-structures over an adequate FOV.
Optical elastography describes a collection of optical methods to map the elasticity of biological tissue [18] that have the potential to provide a more complete understanding of liver mechanics and their impact on disease progression. Optical coherence elastography (OCE) is one of the most prominent optical elastography techniques [19,20] that utilizes optical coherence tomography (OCT) as the underlying imaging technique [21] to detect depth-resolved tissue deformation in response to mechanical loading. The deformation is then used to estimate the mechanical properties of the tissue using a mechanical model [19]. Quantitative micro-elastography (QME) is a variant of OCE [22–24] that uses compression as the loading mechanism. Compared to other OCE techniques, QME provides relatively rapid three-dimensional (3–D) acquisition (several seconds to several minutes), high resolution (tens of micrometers), and high elasticity sensitivity (sub-kPa range) [25], making it well-suited to assessing the elasticity of freshly excised liver tissue.
In this study, we propose QME [22] for 3–D assessment of the elasticity of liver micro-structure in mouse models of chronic liver disease and demonstrate correlation between the progression of pathological features and heterogeneity in the elasticity of the microenvironment of liver tissue, validated against co-registered histology. We scanned 13 freshly excised liver tissues, including five healthy wild type mice, four mice with chronic liver injury resulting from a choline-deficient and ethionine-supplemented (CDE) diet, and four mice with chronic liver injury induced through thioacetamide (TAA) administration [26]. The diseased liver tissues were scanned at three, five, and 12 weeks to obtain a range of pathologies from early to advanced liver injuries. To validate micro-elastograms, we stained liver sections with both haematoxylin and eosin (H&E), to identify cells and reveal changes in micro-structure, and Sirius Red, to highlight regions of fibrosis. We co-registered OCT images and micro-elastograms with the corresponding histology in each case. The QME results demonstrate that healthy liver tissue is relatively soft (mean = 6.5 kPa), with relatively low elasticity heterogeneity (standard deviation = 8.3 kPa) caused mainly by vasculature, bile duct networks, and lobular micro-structures. Our results also indicate a relatively mild increase in liver elasticity (mean = 10.4 kPa) and mechanical heterogeneity (standard deviation = 17.8 kPa) at early stages of liver injury, which correlate with the onset of fibrosis, immune cell infiltration, and moderate hepatocyte damage. However, in advanced stages of liver injury, we observe a significant increase in the elasticity (mean = 40.9 kPa) and mechanical heterogeneity (standard deviation = 49.5 kPa) of the tissue, mainly due to severe hepatocyte damage, bridging fibrosis, and cirrhosis. Furthermore, strong correlation between the heterogeneity in micro-elastograms and pathological features such as fibrosis, immune cell infiltration, and hepatocyte damage is shown in the case of severely affected liver tissue. Our results demonstrate that QME can reveal the mechanical heterogeneity of liver tissue in mouse models of chronic liver injury from the micro- to the milli-scale. This indicates that QME has the potential to investigate the role of microenvironment elasticity in the onset and development of advanced liver disease and HCC by correlating microenvironment elasticity with cellular scale cancerous mechanisms in liver.
2. Materials and methods
2.1 Quantitative micro-elastography
The QME system used in this study, illustrated in Fig. 1, has been described in detail previously [22], and is briefly described here. It utilizes a fiber-based spectral-domain OCT system (TEL320, Thorlabs Inc., USA). The light source is a superluminescent diode with a mean wavelength of 1300 nm and a full width at half maximum (FWHM) spectral bandwidth of 170 nm, providing an axial resolution of 4.8 µm (FWHM) in air. A two-axis galvanometer scanning system was used to scan the beam through a scan lens (LSM03, Thorlabs) with a numerical aperture (NA) of 0.063, providing an OCT lateral resolution of 7.2 µm (FWHM). OCT imaging was performed in a common-path configuration with the reference reflection provided by the interface of the imaging window and a compliant layer, placed on top of the sample. The compliant layer is fabricated using silicone and is used to determine the stress applied at the surface of the sample [27].
Fig. 1. Schematic diagram of the QME system. A liver sample encapsulated in gelatin methacryloyl (GelMA) and sliced using a vibratome is placed in the sample arm, with a compliant layer placed on top of the sample. The sample/layer combination is compressed between a rigid plate attached to an axial translation stage and an imaging window attached to an annular piezoelectric actuator. The axial translation stage generates the preload on the sample/layer combination while the annular piezoelectric actuator generates the micro-scale loading. SLD, superluminescent diode; Galvo, galvanometer.
To perform QME, we compressed the sample/layer combination between the imaging window attached to an annular piezoelectric actuator (Piezomechanik GmbH, Germany) and a rigid plate attached to a motorized axial translation stage. A bulk strain (preload) of 20% was applied to the sample/layer combination to ensure uniform contact between the sample, compliant layer, and both the imaging window and the rigid plate. The PZT actuator was then driven with a 10 Hz square wave, synchronized with the acquisition of B-scans. We acquired two B-scans at each y-location such that alternate B-scans were acquired at different compression levels (referred to as loaded and unloaded levels, respectively). The phase difference between the two B-scans, $\Delta \phi ({x,z} )$, was calculated by multiplying the complex conjugate of the unloaded OCT B-scan, $a_U^\ast ({x,z} )$, by the complex loaded OCT B-scan, ${a_L}({x,z} )$, as $a_U^\ast ({x,z} ).{a_L}({x,z} )= B({x,z} )\textrm{exp}[{i.\Delta \phi ({x,z} )} ]$ (using phase-sensitive detection). The phase difference was then used to calculate the local axial displacement, ${u_z}({x,z} )$, [28]:
where ${\lambda _0}$ is the central wavelength of the OCT light source, 1300 nm, and n is the refractive index of the tissue, assumed to be 1.4. In OCE, the depth resolved local strain in the sample, ${\varepsilon _z}$, is defined as the change of local axial displacement, $\Delta {u_z}$, divided by the depth $\Delta \textrm{z}$, i.e., ${\varepsilon _z} = \Delta {u_z}/\Delta \textrm{z}$. Here, a more complex model for local axial strain was used and calculated using weighted least squares linear regression over a fitting length of 40 pixels, which is equivalent to ∼100 µm [29]. By performing this process at each location in the y-plane, 3-D axial strain maps were generated. Stress at the sample surface was estimated through knowledge of the stress-strain response of the pre-calibrated, transparent, compliant layer placed between the sample and the imaging window and it was assumed that stress was uniform with depth in the sample [22,27]. Local stress at the sample surface, ${\sigma _z}$, was then divided by local axial strain at each location in the sample, ${\varepsilon _z}$, to estimate elasticity as the tangent modulus at the point of preload, which is equal to Young’s modulus, E, under the assumption of linear elasticity [22].2.2 Mouse models
Two mouse models of chronic liver injury were included in this study. The first model is based on a CDE diet, which is widely used to study chronic inflammation, fibrosis, liver progenitor cell (LPC) proliferation, and HCC [26,30]. In this model, hepatotoxicity is induced by the deficiency of choline in combination with the hepatocarcinogen ethionine. The second model is based on TAA administration which is delivered orally in drinking water (300 mg/L). TAA is metabolized and produces hepatotoxic compounds which subsequently induce fibrosis, cirrhosis, and cholangiocarcinoma [26,31]. We included both CDE and TAA models to represent liver pathology that were induced by a fatty diet and chemicals, respectively. Additionally, in the CDE model, damage is induced in the periportal area of the liver lobe, whereas in the TAA model, damage first appears centrally. Overall, 13 male C57BL/6J mice (Animal Resources Centre, Murdoch, Australia) were scanned in this study. These mice were categorized into three groups of: (1) healthy: five control mice fed on a normal diet; (2) early stage: four mice fed on the CDE diet for three weeks and two mice treated with TAA for 5 weeks; and (3) advanced stage: two mice treated with TAA for 12 weeks. All animal experiments were performed in accordance with the Australian code for the care and use of animal for scientific purposes with local animal ethics committee approval from Curtin University.
2.3 Tissue encapsulation in hydrogel
One or two lobes were cut from each freshly excised liver and were kept in Wisconsin solution on ice for transport to the imaging laboratory and subsequent imaging, which occurred within six hours of excision. To overcome the irregular surface profile of the liver lobes and to provide a consistent and smooth flat contact with the mechanical loading mechanism (which is a requirement to generate even preload across the FOV in QME [22]), we embedded the liver samples in gelatin methacryloyl (GelMA), a protein-based hydrogel used frequently in biomedical applications [32]. We placed each lobe in a 3–D printed mold with an inner diameter of 18 mm and a height of 5 mm. We filled the molds with liquid GelMA solution containing 0.1% w/v light-sensitive photoinitiator (IrgaCure, Sigma-Aldrich). In addition, to increase the optical scattering to ensure adequate OCT intensity in the GelMA, we added 3% v/v of suspended polystyrene microspheres with a mean diameter of 0.19 µm (Polysciences, Inc., Warrington, FL). We covered both sides of the mold containing the GelMA solution with coverslips pre-lubricated with dimethyldichlorosilane and placed the sample on an ultra-violet (UV) transilluminator with a central wavelength of 365 nm to polymerize and cure the GelMA for approximately three minutes. Pre-lubrication reduces the adherence of GelMA to the coverslips after curing. Furthermore, to remove the liver capsule and to provide access to deeper tissue, we removed the top 0.5 mm of the GelMA blocks encapsulating the liver lobes using a vibratome (Leica Microsystems, Germany). This had the added advantage of providing a sample with a flatter and more even surface for QME imaging.
2.4 Histology
Following QME imaging, the samples were fixed for histology using Carnoy’s fixative (60% ethanol, 30% chloroform, and 10% glacial acetic acid). The samples were then embedded in paraffin wax blocks, cut into 4 µm sections using a microtome and mounted onto positively-charged microscope slides. Sections were cut at depths of 50 µm, 100 µm, and 150 µm from the imaged surface. The tissue sections were then stained with both H&E and Sirius Red. These different stains were chosen as they visualize different constituents of liver tissue. Haematoxylin stains nuclei blue through binding to DNA and RNA and eosin stains cellular components pink through binding to basic cytoplasmic proteins, providing visualization of tissue micro-structures. Sirius red, in contrast, stains collagen in a range from yellow to red, providing an indication of the level of fibrosis in the tissue. To stain the tissue sections, the paraffin wax was first removed from the slides. Then, tissue sections were rehydrated through graded ethanol washes, and immersed in the stain solutions. Slides were then dehydrated, dried, and mounted under coverslips using DePeX Mounting Medium. Subsequently, stained histology slides were scanned with a 3DHISTECH high resolution scanner (3DHISTECH, Hungary).
3. Results
In this section, we present representative OCT images and micro-elastograms, co-registered with histology, from four mouse liver lobes excised from healthy, early stage CDE (three weeks), early stage TAA (five weeks), and advanced stage TAA (12 weeks) mouse models. The OCT images are presented as the SNR of OCT intensity and the micro-elastograms are presented in kiloPascals (kPa), both on logarithmic scales. The OCT images and micro-elastograms are presented in the en face plane from the depth that provided the closest match with the corresponding histology. The selected depths of the OCT images and micro-elastograms are in the range of ∼100-200 µm from the tissue surface. The FOV of the OCT images and micro-elastograms is 5 × 5 mm2 and the FOV of the histology images is chosen to best match with the OCT images and micro-elastograms. The FOV in the presented histology images in this study is ∼0.6× smaller than the FOV of the OCT images and micro-elastograms (∼40% shrinkage), as a result of tissue shrinkage during the fixation and paraffin embedding processes [33]. In addition, magnified 1 × 1 mm2 and 0.4 × 0.4 mm2 insets of the OCT images and micro-elastograms and corresponding histology are included to show the correlation between specific features in the OCT images and micro-elastograms and the micro-scale histology. Furthermore, statistical analysis of the tissue elasticity during the course of chronic liver injury in mice is presented in Section 3.5, which includes histograms, mean and standard deviation values, and a null-hypothesis significance test on the healthy, early stage, and advanced stage group.
3.1 Healthy model
Figure 2 shows results from a representative healthy liver lobe. Figures 2(a)–2(c) show the Sirius Red histology images. Figures 2(b) and 2(c) are the magnified regions indicated by square insets in Figs. 2(a) and 2(b), respectively. In these images, a high concentration of collagen, indicated by the red regions of the Sirius Red histology (Figs. 2(b) and 2(c)) can be seen in the lining of the vasculature and duct network, while there is a very low level of collagen elsewhere in the lobular area of the liver tissue. Enclosed broken lines in Fig. 2(c) highlight the high collagen region in the lining of a vein. The H&E histology images are shown in Figs. 2(d)–2(f). Figures 2(e) and 2(f) are the magnified regions indicated by square insets in Figs. 2(d) and 2(e), respectively. In these images, an even distribution of healthy hepatocytes (well-separated large cells) can be observed in the lobular region, while smaller cells (probably sinusoidal endothelial cells) can be observed in the lining of the vasculature network. Enclosed broken lines in Fig. 2(f) highlight the location of the smaller cells in the lining of the vein, and an enclosed solid line indicates a region of healthy hepatocytes. The top left region of the histology images in Figs. 2(a) and 2(d) are lost due to an artifact of the histology process, as the sample was not perfectly flat during the fixation, paraffin embedding, and slicing steps.
Fig. 2. Healthy liver tissue: (a)–(c) Sirius Red histology, (d)–(f) H&E histology, (g)–(i) en face OCT images, and (j)–(l) en face micro-elastograms. The second and third columns are magnified regions indicated by square insets in the first and second columns, respectively. The fields-of-view of the OCT images and micro-elastograms in the first, second, and third columns are 5 × 5 mm2, 1 × 1 mm2, and 0.4 × 0.4 mm2, respectively, and the fields-of-view of the histology images are adjusted to co-register them with the OCT images and micro-elastograms in each column. Arrows in the first column indicate the anatomical features which are used to co-register the histology with the OCT images and micro-elastograms. The enclosed broken lines in the third column indicate the lining of a vein region, and the enclosed solid line in (f) indicates an area containing healthy hepatocytes.
The OCT images are shown in Figs. 2(g)–2(i). Figures 2(h) and 2(i) are the magnified regions indicated by square insets in Figs. 2(g) and 2(h), respectively. The OCT images show the micro-structure of the liver tissue with slightly higher OCT intensity where there is a higher concentration of collagen. For example, the enclosed broken lines in Fig. 2(i) correlate with a higher concentration of collagen in the lining of the vein. The micro-elastograms are shown in Figs. 2(j)–2(l). Figures 2(k) and 2(l) are the magnified regions indicated by square insets in Figs. 2(j) and 2(k), respectively. The micro-elastograms present relatively low elasticity in this sample (∼3.2-32 kPa). A relatively low heterogeneity in the elasticity is also observed, which is mainly caused by the vasculature and duct network. An enclosed broken line in Fig. 2(j) indicates a lobular region with a homogeneous elasticity at ∼5 kPa which does not include large vasculature and duct networks. Enclosed broken lines in Fig. 2(l) highlight slightly higher elasticity regions (∼32 kPa) corresponding to the higher concentration of collagen in the lining of the vein and a soft region (∼3.2 kPa) between them caused by the vein lumen.
3.2 Early stage CDE model
Figure 3 presents results from a representative early stage CDE liver tissue. The Sirius Red histology images are shown in Figs. 3(a)–3(c). Figures 3(b) and 3(c) are the magnified regions indicated by square insets in Figs. 3(a) and 3(b), respectively. In these images, a moderate increase in fibrosis is observed, caused by both collagen deposition in lobules and the collagen concentration in the lining of vasculature and ducts. Enclosed broken lines in Fig. 3(c) highlight the onset of fibrosis extending out of a PV region. Fibrosis is a common response to cell damage in mice on a CDE diet and the fibrosis pattern observed in Figs. 3(b) and 3(c) has been referred to as chicken wire pattern fibrosis [26]. Corresponding H&E histology images are presented in Figs. 3(d)–3(f). Figures 3(e) and 3(f) are the magnified regions indicated by square insets in Figs. 3(d) and 3(e), respectively. Although a relatively even distribution of healthy hepatocytes can still be seen in H&E histology, it is observed that clusters of immune cells infiltrate hepatocytes with a higher concentration near PV regions. Immune cell infiltration is present when small cells can be observed between hepatocytes and away from the lining of vasculature (where the endothelial cells reside). The immune cells are circular and have small nuclei and scarce cytoplasm, much smaller than hepatocytes, and often present as clusters. An enclosed broken line in Fig. 3(f) indicates immune cell infiltration in the vicinity of hepatocytes close to a PV region, which corresponds with the onset of fibrosis indicated in the Sirius Red histology in Fig. 3(c). Such immune cell infiltration has previously been reported for mice on a CDE diet [26].
Fig. 3. CDE liver tissue: (a)–(c) Sirius Red histology, (d)–(f) H&E histology, (g)–(i) en face OCT, and (j)–(l) en face micro-elastograms. The second and third columns are magnified regions indicated by square insets in the first and second columns, respectively. The fields-of-view of the OCT images and micro-elastograms in the first, second, and third columns are 5 × 5 mm2, 1 × 1 mm2, and 0.4 × 0.4 mm2, respectively, and the fields-of-view of the histology images are adjusted to co-register them with the OCT images and micro-elastograms in each column. Arrows in the first column indicate the anatomical features which are used to co-register the histology with the OCT images and micro-elastograms. The enclosed broken lines in (c), (f), and (l) show co-registered onset of fibrosis, immune cell infiltration, and local elevated elasticity, respectively, in the vicinity of a portal vein (PV) region.
The OCT images are presented in Figs. 3(g)–3(i). Figures 3(h) and 3(i) are the magnified regions indicated by square insets in Figs. 3(g) and 3(h), respectively. The OCT intensity is slightly higher where there is a higher concentration of collagen or fibrosis, however, micro-scale liver pathological features caused by the CDE diet are not easily observed in the OCT images. The corresponding micro-elastograms are presented in Figs. 3(j)–3(l). Figures 3(k) and 3(l) are the magnified regions indicated by square insets in Figs. 3(j) and 3(k), respectively. The micro-elastograms show a relatively higher elasticity (∼10-100 kPa) for the CDE liver sample, in comparison to the softer healthy tissue in Figs. 2(j)–2(l). Furthermore, there is an increase in the heterogeneity of the elasticity of the CDE liver tissue compared to the healthy tissue in Fig. 2 which may be caused by the increase in fibrosis and immune cell infiltration. An enclosed broken line in Fig. 3(l) indicates a region of high elasticity (∼100 kPa) in the vicinity of the PV region, correlated with the onset of fibrosis and immune cell infiltration which is observed in Figs. 3(c) and 3(f), respectively.
3.3 Early stage TAA model
Figure 4 shows the results from a representative early stage TAA liver. The Sirius Red histology images are presented in Figs. 4(a)–4(c). Figures 4(b) and 4(c) are the magnified regions indicated by square insets in Figs. 4(a) and 4(b), respectively. In these figures, a clear increase in fibrosis, indicated by the appearance of red regions, is observed compared to the healthy and CDE samples in Figs. 2(a)–2(c) and 3(a)–3(c). It is observed that bands of fibrosis are present between central vein (CV) regions. An enclosed broken line in Fig. 4(c) indicates a fibrotic band forming in the vicinity of a CV region and extending out of the CV area. It has previously been reported that mouse liver treated with TAA forms bands of fibrosis between CV regions [26]. The corresponding H&E histology images are presented in Figs. 4(d)–4(f). Figures 4(e) and 4(f) are the magnified regions indicated by square insets in Figs. 4(d) and 4(e), respectively. Immune cell infiltration can be observed in these images as clusters of small cells between hepatocytes. For example, an enclosed broken line in Fig. 4(f) indicates an area affected by immune cell infiltration corresponding to the fibrotic region in Sirius Red histology in Fig. 4(c). Furthermore, it is observed in H&E histology that some of the hepatocytes have a darker pink color, which suggests a higher accumulation of protein caused by the inability of hepatocytes to synthesize proteins. An enclosed solid line in Fig. 4(f) indicates a region with partial healthy hepatocytes and a region with damaged hepatocytes indicated with a blue arrow.
Fig. 4. Early TAA liver tissue: (a)–(c) Sirius Red histology, (d)–(f) H&E histology, (g)–(i) en face OCT, and (j)–(l) en face micro-elastograms. The second and third columns are magnified regions indicated by square insets in the first and second columns, respectively. The fields-of-view of the OCT images and micro-elastograms in the first, second, and third columns are 5 × 5 mm2, 1 × 1 mm2, and 0.4 × 0.4 mm2, respectively, and the fields-of-view of the histology images are adjusted to co-register them with the OCT images and micro-elastograms in each column. Arrows in the first column indicate the anatomical features which are used to co-register the histology with the OCT images and micro-elastograms. The enclosed broken lines in (c), (f), (i), and (l) show co-registered fibrotic band, immune cell infiltration, higher OCT intensity, and relatively low elasticity, respectively, in the vicinity of a central vein (CV) region. The enclosed solid lines in (f) and (l) indicate co-registered partially damaged hepatocyte and relatively high elasticity regions, respectively.
The OCT images are presented in Figs. 4(g)–4(i). Figures 4(h) and 4(i) are the magnified regions indicated by square insets in Figs. 4(g) and 4(h), respectively. Higher OCT intensity is observed in the regions containing fibrotic bands. An enclosed broken line in Fig. 4(i) indicates higher OCT intensity corresponding with the fibrotic band in Sirius Red histology in Fig. 4(c). The micro-elastograms are presented Figs. 4(j)–4(l). Figures 4(k) and 4(l) are the magnified regions indicated by square insets in Figs. 4(j) and 4(k), respectively. The micro-elastograms show a higher elasticity (∼10-100 kPa) for the early TAA tissue, compared to the elasticity of the healthy liver in Figs. 2(j)–2(l). In addition, a moderate increase in the heterogeneity of the elasticity correlated with the micro-structure of the liver and fibrotic bands can be observed in these images. An enclosed broken line in Fig. 4(l) indicates a relatively low elasticity region (∼10-32 kPa) in the vicinity of a CV region which corresponds to the fibrotic band in Fig. 4(c), and an enclosed solid line that indicates a higher elasticity region (∼32-100 kPa) which corresponds to the regions of partially damaged hepatocytes in Fig. 4(f). Interestingly, the fibrotic band correlates with a region of relatively low elasticity, which is somewhat unexpected given that fibrosis typically presents with high elasticity in ultrasound elastography and magnetic resonance elastography [16,17]. A possible reason for this is the higher concentration of micro-vasculature resulting from fibrosis-associated angiogenesis, indicated by green arrows in the H&E histology image in Fig. 4(f), which generate a porous region in the fibrotic band [34].
3.4 Advanced stage TAA model
Figure 5 shows the results from a representative advanced stage TAA liver lobe. The Sirius Red histology images are presented in Figs. 5(a)–5(c). Figures 5(b) and 5(c) are the magnified regions indicated by square insets in Figs. 5(a) and 5(b), respectively. A large increase in fibrosis is observed in these images, in comparison to the results presented in Figs. 2(a)–2(c), Figs. 3(a)–3(c), and Figs. 4(a)–4(c), in the form of fibrotic bands typically between CV regions, which is known as bridging fibrosis [26]. Enclosed solid and broken lines in Fig. 5(b) indicate bridging fibrosis and a localized cirrhotic region, respectively. Cirrhotic regions present when fibrotic bands thicken and tightly occupy the lobular region. An enclosed broken line in Fig. 5(c) also highlights a magnified cirrhotic region. The corresponding H&E histology images are presented in Figs. 5(d)–5(f). Figures 5(e) and 5(f) are the magnified regions indicated by square insets in Figs. 5(d) and 5(e), respectively. An enclosed broken line in Fig. 5(f) indicates heterogeneous cell distribution with a high concentration of immune cells infiltrating damaged hepatocytes in the cirrhotic region, while an enclosed solid line presents a region of damaged hepatocytes with low concentration of immune cells and high accumulation of proteins (indicated by a darker color in H&E histology) in the lobular region. Furthermore, blue arrows in this figure indicate micro-vascularization due to fibrosis-associated angiogenesis in the cirrhotic region [34].
Fig. 5. Advanced TAA liver tissue: (a)–(c) Sirius Red histology, (d)–(f) H&E histology, (g)–(i) en face OCT, and (j)–(l) en face micro-elastograms. The second and third columns are magnified regions indicated by square insets in the first and second columns, respectively. The fields-of-view of the OCT images and micro-elastograms in the first, second, and third columns are 5 × 5 mm2, 1 × 1 mm2, and 0.4 × 0.4 mm2, respectively, and the fields-of-view of the histology images are adjusted to co-register them with the OCT images micro-elastograms in each column. Arrows in the first column indicate the anatomical features which are used to co-register the histology with the OCT images and micro-elastograms. The enclosed solid and broken lines in (b) indicate bridging fibrosis and cirrhosis, respectively. The enclosed broken lines in (c), (f), (i), and (l) show co-registered cirrhosis, immune cell infiltration, higher OCT intensity, and relatively low elasticity, respectively, in the vicinity of a central vein (CV) region. The enclosed solid lines in (f) and (l) indicate co-registered totally damaged hepatocyte and relatively high elasticity regions, respectively.
The OCT images are presented in Figs. 5(g)–5(i). Figures 5(h) and 5(i) are the magnified regions indicated by square insets in Figs. 5(g) and 5(h), respectively. Close correlation between the OCT images and the histology images is observed with higher OCT intensity corresponding to fibrotic regions. An enclosed broken line in Fig. 5(i) indicates a region with relatively high OCT intensity which corresponds to a cirrhotic area in the Sirius Red histology image (Fig. 5(c)). The micro-elastograms are presented in Figs. 5(j)–5(l). Figures 5(k) and 5(l) are the magnified regions indicated by square insets in Figs. 5(j) and 5(k), respectively. The micro-elastograms show elevated elasticity (∼32-320 kPa), a ∼10-fold increase in elasticity compared to healthy liver tissue. In addition, the micro-elastograms demonstrate strong heterogeneity in elasticity which corresponds very well with the micro-structure of liver tissue observed in the histology images, such as regions of advanced fibrosis, hepatocyte damage, and immune cell infiltration. An enclosed broken line in Fig. 5(l) highlights a region of relatively low elasticity (∼32 kPa) which corresponds to the cirrhotic region in Fig. 5(c), while an enclosed solid line in this image shows a region of relatively high elasticity (∼320 kPa) which corresponds with damaged hepatocytes in Fig. 5(f) using an enclosed solid line. Interestingly, the region containing damaged hepatocytes looks much stiffer than the cirrhotic region. Again, a possible reason for this is the higher concentration of micro-vasculature resulting from fibrosis-associated angiogenesis (which makes the tissue porous) in the cirrhotic region [34], while the region containing damaged hepatocytes is less porous and shows high accumulation of proteins.
3.5 Statistical analysis of elasticity variation in chronic liver injury
We analyzed the elasticity of the liver samples across all the animals included in the study. Based on the pathological features observed in histology, we grouped the liver samples into healthy (N = 5: fed on normal diet, with no pathological features), early stage (N = 6: four fed on CDE for three weeks and two treated with TAA for five weeks, with moderate pathological features), and advanced stage (N = 2: treated with TAA for 12 weeks, with severe pathological features). First, we combined the elasticity values of the samples in each group, to generate a dataset for each of the three groups. The elasticity values were selected from a cuboid region of the liver tissue in each scan, cropping out the edges of the scan with low quality of elasticity and the encapsulating GelMA regions. The cuboids were at least 6 × 6 mm2 in xy and consisted of 50 pixels in z (∼120 µm) close to the top part of the tissues. As freshly excised liver tissue considerably attenuates OCT intensity after a few hundred micrometers, this cuboid thickness was selected to have a consistently high OCT SNR and elasticity quality across the samples. Furthermore, as the top 0.5 mm of the liver lobes were cut after GelMA encapsulation, the cuboids were situated sufficiently deep into the liver tissue. As the cuboid regions with lateral size of several millimeters and thickness of 120 µm include a large number of liver microstructures with typical sizes of >100 µm, the combination of cuboids within each group is a reliable representation of tissue variation within each group of mouse models. Furthermore, the histology images of tissue on 4 µm thick sections at depths of 50, 100, and 150 µm show even distribution of pathological features over the size of cuboids that further show their suitability as reliable representation of liver samples. Figure 6(a) shows the normalized histograms of the elasticity of the healthy, early stage, and advanced stage datasets, on a logarithmic scale. The mean and heterogeneity of elasticity, expressed as the mean ± standard deviation, for the healthy, early stage, and advanced stage datasets are 6.5 ± 8.2 kPa, 10.4 ± 17.8 kPa, and 40.9 ± 49.5 kPa, respectively, also annotated in Fig. 6(a). The histograms in Fig. 6(a) indicate a significant increase in the mean and heterogeneity of the elasticity by the progress in chronic liver injury. Although the mean and heterogeneity of elasticity of healthy dataset increase moderately by 1.6× and 2.2×, respectively, in case of the early stage dataset, those values increase significantly by 6.3× and 6×, respectively, in case of the advanced stage dataset. This correlates with the severity of the pathological features observed in the healthy, early stage CDE, early stage TAA, and advanced stage TAA histology images, presented in Figs. 2–5, respectively. Furthermore, the median elasticity of the healthy, early stage, and advanced stage datasets are 5.1 kPa, 5.6 kPa, and 22.4 kPa, respectively, which are smaller than the mean values for these datasets. Smaller median values indicate that datasets are positively skewed toward larger elasticity, as it can also be observed in the histograms in Fig. 6(a). Positively skewed elasticity indicates an abnormal distribution with large variation toward positive values and is a possible indicator of a larger standard deviation of elasticity compared to the mean elasticity.
Fig. 6. The statistical analysis of the variation in the elasticity due to the progression of chronic liver injury. (a) Histograms of the elasticity of the healthy, early stage, and advanced stage samples pooled into three datasets, and annotated mean ± standard deviation for each dataset. (b) Mean and standard deviation of the elasticity of the individual liver samples within the healthy, early stage, and advanced stage groups, including the null-hypothesis test for healthy/early stage groups (non-significant, NS) and healthy/advanced stage groups (statistically significant, * p < 0.05).
Furthermore, to analyze the elasticity variation across individual liver samples, we calculated the median elasticity value for each sample within the groups and considered them as individual values. Figure 6(b) presents the mean and heterogeneity of the elasticity of the individual samples (as mean ± standard deviation) in the healthy (5.4 ± 1.1 kPa), early stage (6.8 ± 4.2 kPa), and advanced stage (28.3 ± 17.5 kPa) groups, respectively. We also carried out a null-hypothesis significance test between the healthy and early stage groups and also between the healthy and advanced stage groups, to determine the statistical significance of the elasticity variation across the groups. To do this, we calculated the p-values using the two sample t-test, and considered p < 0.05 as statistically significant. The p-value for the healthy/early stage test was 0.47 (non-significant), while its value for the healthy/advanced stage test was 0.018 (statistically significant). The results of the null-hypothesis test indicate that the variation in the elasticity of the individual liver samples between the healthy and advanced stage groups is large enough to be considered as statistically significant, while this variation is non-significant between the individual samples in the healthy and early stage groups.
4. Discussion
In this study, QME was utilized to assess the mechanical properties of mouse models of healthy and chronically diseased liver tissue and elasticity was correlated with the micro-structures of liver. In addition, statistical analysis of the measured elasticity was performed and it was demonstrated that the mean and standard deviation of elasticity significantly increased in the advanced stage of chronic liver injury. This was documented for the onset and progression of liver pathological features commonly identified by histology. To visualize the cell structure and collagen distribution, we performed H&E and Sirius Red staining on fixed and paraffin embedded samples after QME experiments. We co-registered the OCT images and micro-elastograms with Sirius Red and H&E histology shown for representative samples from each group. Liver pathological features are typically categorized into four main groups: (1) fibrosis and its functional contributions, (2) cell-damage where hepatocytes are damaged, rounded, ballooned (swelled and enlarged), (3) immune-response where macrophages invade and reside in the tissue, and (4) regeneration response, where LPCs proliferate and differentiate into cholangiocytes and hepatocytes. In this study, we considered pathological features such as: (1) collagen deposition representing fibrosis, (2) change in the morphology and staining intensity of hepatocytes representing hepatocyte damage, and (3) appearance of small circular cells with scarce cytoplasm among hepatocytes representing immune cell infiltration. More extensive investigation of liver pathological features by comprehensive cell-typing requires further immunohistochemical staining, immunofluorescence microscopy, and other biochemical and molecular approaches, which is planned for future studies.
Biological tissues with established fibrosis typically have higher elasticity than normal tissue due to increased extracellular matrix, particularly because of collagen deposition [35]. For instance, multiple studies have reported higher elasticity for fibrotic lung and liver tissue compared to normal tissue [36–38]. Other studies suggest that elasticity and the level of fibrosis are not linearly correlated [35,38]. The results of our study demonstrate a ∼6-fold increase in the elasticity of the fibrotic liver tissue, indicated by an enclosed broken line in Fig. 5(l) at ∼32 kPa, relative to the elasticity of the homogeneous healthy liver tissue, indicated by an enclosed broken line in Fig. 2(j) at ∼5 kPa. However, the elasticity of a damaged hepatocyte region at ∼320 kPa, indicated by an enclosed solid line in Fig. 5(l), is ∼10 times higher than the elasticity of the fibrotic tissue (∼32 kPa) and ∼60 times higher than the elasticity of the homogeneous healthy tissue (∼5 kPa). One possible reason for the lower elasticity of the fibrotic tissue relative to the tissue containing damaged hepatocytes, observed in Fig. 5(l), is that fibrosis progression in mice is strongly linked to enhanced angiogenesis [34,39]. The blue arrows in Fig. 5(f) indicate higher concentration of micro-vasculature in the fibrotic region. Consequently, higher density of micro-vasculature results in the fibrotic tissue being porous with a higher concentration of liquid. This causes the fibrotic region to undergo higher deformation compared to the surrounding less porous region. Higher deformation is associated with higher strain (lower elasticity) in QME [40]. Another possible reason for the relatively lower elasticity in the fibrotic region is the higher concentration of immune cells in this region, indicated by a higher concentration of small cells in Fig. 5(f), which may affect the mechanical properties of the fibrotic tissue [37]. However, further investigation is required to determine the effect of pores and cell-type on the measured elasticity of the fibrotic region.
Our findings demonstrate a significant increase in the mean and standard deviation of the elasticity of the liver tissues affected by bridging fibrosis and cirrhosis. As severe fibrosis and cirrhosis increase the risk of HCC, it is expected that elevated and heterogeneous microenvironment elasticity promotes carcinogenic processes in liver [41]. Importantly, in mechanotransduction, cell function is governed by the elasticity of its microenvironment [42], and it is suggested that heterogeneous mechanical properties of tissue can promote oncogenesis [43]. To investigate the role of microenvironment elasticity in cancer development, in a recent study, mechanical heterogeneity in liver tissue at different stages of HCC development was reported using AFM [13]. However, this study presented the heterogeneity by averaging repeated surface measurements at multiple sites rather than a spatially resolved elastogram at different depths [13]. We propose that QME can be used to map spatially resolved liver elasticity on the micro-scale that can be directly co-registered with liver pathological features over the course of chronic liver disease development. In addition, it is suggested that LPCs may play a role in HCC development. Such studies investigate the correlation between the distribution of LPCs with liver pathological features such as fibrosis [44]. It has also been proposed, using traction force microscopy, that elasticity can control the differentiation of LPCs [14]. However, such studies were performed in vitro on LPCs rather than being implemented in vivo. We suggest that QME can be a useful in vivo tool in the form of micro-endoscopes [45] to investigate the role of elasticity in proliferation and differentiation of liver cancerous cells such as LPCs. Furthermore, LPCs are considered as potential candidates for cell therapy and tissue engineering in alternative approaches to whole organ transplantation in chronic liver disease treatment [46,47]. QME can also be employed in these areas as a tool to assess the impact of mechanical properties on their biology.
Ultrasound elastography is routinely used in clinics for non-invasive diagnoses of chronic liver disease. Using ultrasound elastography, an elasticity range of 5.5–75.4 kPa has been reported for human patients with cirrhosis with a median of 31.1 kPa [48]. Also, an elasticity range of 2.3–5.9 kPa has been reported in people without any known liver disease (normal liver) with a median of 4.1 kPa [49]. Elasticity measured by ultrasound elastography has a relatively low resolution and cannot be correlated with liver micro-structures and is often presented as an average value in such studies. In this work, the median elasticity values of the pooled datasets in the healthy and advanced stage groups (presented in Section 3.5) are 5.1 kPa and 22.4 kPa, respectively. This indicates that the elasticity values measured by QME are relatively comparable with the elasticity values measured by ultrasound elastography presented in such studies [48,49]. Ultimately, although ultrasound elastography is a powerful tool that is better suited than QME to non-invasive chronic liver disease diagnosis, QME provides micro-scale elasticity resolution, which is required to correlate the elasticity of the microenvironment of liver tissue with pathological features and can therefore provide insight on the physiological basis for contrast in clinical elastography.
5. Conclusion
We utilized QME for biomechanical assessment of mouse models of healthy and chronically diseased mouse liver. We performed Sirius Red and H&E histology on the samples after QME and co-registered elasticity with liver micro-structures and pathological features such as fibrosis, hepatocyte damage, and immune cell infiltration on the micro-scale. The QME experiments show that elevated elasticity and heterogeneity due to chronic liver injury correlate both visually and statistically with pathological features in the liver. The findings of this study demonstrate that QME has potential for assessment of liver biomechanics on the scale of tens of micrometer to several millimeters.
Funding
Cancer Council Western Australia; Department of Health, Australian Government; Australian Research Council.
Acknowledgements
The authors thank Yu Suk Choi from School of Human Sciences, The University of Western Australia, for providing GelMA, and acknowledge the efforts of Mary Lee from School of Human Sciences, University of Western Australia, in providing advice and preparation of histological sections.
Disclosures
BFK: OncoRes Medical (F, I). The other authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. A. J. Craig, J. V. Felden, T. Garcia-Lezana, S. Sarcognato, and A. Villanueva, “Tumour evolution in hepatocellular carcinoma,” Nat. Rev. Gastroenterol. Hepatol. 17(3), 139–152 (2020). [CrossRef]
2. J. M. Llovet, R. K. Kelley, V. Augusto, A. G. Singal, P. Eli, R. Sasan, L. Riccardo, K. Kazuhiko, J. Zucman-Rossi, and R. S. Finn, “Hepatocellular carcinoma (Primer),” Nat. Rev. Dis. Primers 7(1), 6–28 (2021). [CrossRef]
3. E. Kim and P. Viatour, “Hepatocellular carcinoma: old friends and new tricks,” Exp. Mol. Med. 52(12), 1898–1907 (2020). [CrossRef]
4. R. Loomba and L. A. Adams, “Advances in non-invasive assessment of hepatic fibrosis,” Gut 69(7), 1343–1352 (2020). [CrossRef]
5. L. J. M. Heyens, D. Busschots, G. H. Koek, G. Robaeys, and S. Francque, “Liver fibrosis in non-alcoholic fatty liver disease: from liver biopsy to non-invasive biomarkers in diagnosis and treatment,” Front. Med. 8(615978), 1–20 (2021). [CrossRef]
6. L. Sandrin, B. Fourquet, J.-M. Hasquenoph, S. Yon, C. Fournier, F. Mal, C. Christidis, M. Ziol, B. Poulet, F. Kazemi, M. Beaugrand, and R. Palau, “Transient elastography: a new noninvasive method for assessment of hepatic fibrosis,” Ultrasound Med. Biol. 29(12), 1705–1713 (2003). [CrossRef]
7. B. Suh, S. Park, D. W. Shin, J. M. Yun, H.-K. Yang, S. J. Yu, C.-I. Shin, J.-S. Kim, E. Ahn, H. Lee, J. H. Park, and B. Cho, “High liver fibrosis index FIB-4 is highly predictive of hepatocellular carcinoma in chronic hepatitis B carriers,” Hepatology 61(4), 1261–1268 (2015). [CrossRef]
8. M. N. Kim, S. U. Kim, B. K. Kim, J. Y. Park, D. Y. Kim, S. H. Ahn, K. J. Song, Y. N. Park, and K.-H. Han, “Increased risk of hepatocellular carcinoma in chronic hepatitis B patients with transient elastography–defined subclinical cirrhosis,” Hepatology 61(6), 1851–1859 (2015). [CrossRef]
9. H.-M. Wang, C.-H. Hung, S.-N. Lu, C.-H. Chen, C.-M. Lee, T.-H. Hu, and J.-H. Wang, “Liver stiffness measurement as an alternative to fibrotic stage in risk assessment of hepatocellular carcinoma incidence for chronic hepatitis C patients,” Liver Int. 33(5), 756–761 (2013). [CrossRef]
10. T. Akima, M. Tamano, and H. Hiraishi, “Liver stiffness measured by transient elastography is a predictor of hepatocellular carcinoma development in viral hepatitis,” Hepatol. Res. 41(10), 965–970 (2011). [CrossRef]
11. A. Filliol and R. F. Schwabe, “Contributions of fibroblasts, extracellular matrix, stiffness, and mechanosensing to hepatocarcinogenesis,” Semin. Liver Dis. 39(03), 315–333 (2019). [CrossRef]
12. Y. Liu, S. Zhuo, Y. Zhou, L. Ma, Z. Sun, X. Wu, X. W. Wang, B. Gao, and Y. Yang, “Yap-Sox9 signaling determines hepatocyte plasticity and lineage-specific hepatocarcinogenesis,” J. Hepatol. 76(3), 652–664 (2022). [CrossRef]
13. Y. Sun, H. Li, Q. Chen, Q. Luo, and G. Song, “The distribution of liver cancer stem cells correlates with the mechanical heterogeneity of liver cancer tissue,” Histochem. Cell Biol. 156(1), 47–58 (2021). [CrossRef]
14. A. P. Kourouklis, K. B. Kaylan, and G. H. Underhill, “Substrate stiffness and matrix composition coordinately control the differentiation of liver progenitor cells,” Biomaterials 99, 82–94 (2016). [CrossRef]
15. S. K. Venkatesh, M. Yin, J. F. Glockner, N. Takahashi, P. A. Araoz, J. A. Talwalkar, and R. L. Ehman, “Magnetic resonance elastography of liver tumors-preliminary results,” AJR, Am. J. Roentgenol. 190(6), 1534–1540 (2008). [CrossRef]
16. A. Tang, G. Cloutier, N. M. Szeverenyi, and C. B. Sirlin, “Ultrasound elastography and MR elastography for assessing liver fibrosis: part 1, principles and techniques,” AJR, Am. J. Roentgenol. 205(1), 22–32 (2015). [CrossRef]
17. A. Tang, G. Cloutier, N. M. Szeverenyi, and C. B. Sirlin, “Ultrasound elastography and MR elastography for assessing liver fibrosis: part 2, diagnostic performance, confounders, and future directions,” AJR, Am. J. Roentgenol. 205(1), 33–40 (2015). [CrossRef]
18. B. F. Kennedy, P. Wijesinghe, and D. D. Sampson, “The emergence of optical elastography in biomedicine,” Nat. Photonics 11(4), 215–221 (2017). [CrossRef]
19. B. F. Kennedy, K. M. Kennedy, and D. D. Sampson, “A review of optical coherence elastography: fundamentals, techniques and prospects,” IEEE J. Sel. Top. Quantum Electron. 20(2), 272–288 (2014). [CrossRef]
20. B. F. Kennedy and J. C. Bamber, Introduction to Optical Coherence Elastography (AIP Publishing, 2021).
21. D. Huang, E. A. Swanson, C. P. Lin, J. S. Schuman, W. G. Stinson, W. Chang, M. R. Hee, T. Flotte, K. Gregory, C. A. Puliafito, and J. G. Fujimoto, “Optical coherence tomography,” Science 254(5035), 1178–1181 (1991). [CrossRef]
22. K. M. Kennedy, L. Chin, R. A. McLaughlin, B. Latham, C. M. Saunders, D. D. Sampson, and B. F. Kennedy, “Quantitative micro-elastography: Imaging of tissue elasticity using compression optical coherence elastography,” Sci. Rep. 5(1), 15538 (2015). [CrossRef]
23. V. Y. Zaitsev, A. L. Matveyev, L. A. Matveev, A. A. Sovetsky, M. S. Hepburn, A. Mowla, and B. F. Kennedy, “Strain and elasticity imaging in compression optical coherence elastography: the two-decade perspective and recent advances,” J. Biophotonics 14(2), e202000257 (2021). [CrossRef]
24. J. Li, K. Y. Foo, M. S. Hepburn, A. Mowla, L. Chin, and B. F. Kennedy, Compression Optical Coherence Elastography (AIP Publishing, 2021).
25. J. Li, M. S. Hepburn, L. Chin, A. Mowla, and B. F. Kennedy, “Analysis of sensitivity in quantitative micro-elastography,” Biomed. Opt. Express 12(3), 1725–1745 (2021). [CrossRef]
26. J. Köhn-Gaone, B. J. Dwyer, C. A. Grzelak, G. Miller, N. A. Shackel, G. A. Ramm, G. W. McCaughan, C. L. Elsegood, J. K. Olynyk, and J. E. E. Tirnitz-Parker, “Divergent inflammatory, fibrogenic, and liver progenitor cell dynamics in two common mouse models of chronic liver injury,” Am. J. Pathol. 186(7), 1762–1774 (2016). [CrossRef]
27. K. M. Kennedy, S. Es’haghian, L. Chin, R. A. McLaughlin, D. D. Sampson, and B. F. Kennedy, “Optical palpation: optical coherence tomography-based tactile imaging using a compliant sensor,” Opt. Lett. 39(10), 3014–3017 (2014). [CrossRef]
28. R. K. Wang, S. Kirkpatrick, and M. Hinds, “Phase-sensitive optical coherence elastography for mapping tissue microstrains in real time,” Appl. Phys. Lett. 90(16), 164105 (2007). [CrossRef]
29. B. F. Kennedy, S. H. Koh, R. A. McLaughlin, K. M. Kennedy, P. R. T. Munro, and D. D. Sampson, “Strain estimation in phase-sensitive optical coherence elastography,” Biomed. Opt. Express 3(8), 1865–1879 (2012). [CrossRef]
30. B. Akhurst, E. J. Croager, C. A. Farley-Roche, J. K. Ong, M. L. Dumble, B. Knight, and G. C. Yeoh, “A modified choline-deficient, ethionine-supplemented diet protocol effectively induces oval cells in mouse liver,” Hepatology 34(3), 519–522 (2001). [CrossRef]
31. T. Y. Low, C. K. Leow, M. Salto-Tellez, and M. C. M. Chung, “A proteomic analysis of thioacetamide-induced hepatotoxicity and cirrhosis in rat livers,” Proteomics 4(12), 3960–3974 (2004). [CrossRef]
32. M. Zhu, Y. Wang, G. Ferracci, J. Zheng, N.-J. Cho, and B. H. Lee, “Gelatin methacryloyl and its hydrogels with an exceptional degree of controllability and batch-to-batch consistency,” Sci. Rep. 9(1), 1–13 (2019). [CrossRef]
33. J. Rieger, S. Twardziok, H. Huenigen, R. M. Hirschberg, and J. Plendl, “Porcine intestinal mast cells. Evaluation of different fixatives for histochemical staining techniques considering tissue shrinkage,” Eur. J. Histochem. 57(3), 21–142 (2013). [CrossRef]
34. H. Sahin, E. Borkham-Kamphorst, C. Kuppe, M. M. Zaldivar, C. Grouls, M. Al-samman, A. Nellen, P. Schmitz, D. Heinrichs, M.-L. Berres, D. Doleschel, D. Scholten, R. Weiskirchen, M. J. Moeller, F. Kiessling, C. Trautwein, and H. E. Wasmuth, “Chemokine Cxcl9 attenuates liver fibrosis-associated angiogenesis in mice,” Hepatology 55(5), 1610–1619 (2012). [CrossRef]
35. R. G. Wells, “Tissue mechanics and fibrosis,” Biochim. Biophys. Acta, Mol. Basis Dis. 1832(7), 884–890 (2013). [CrossRef]
36. F. Liu, J. D. Mih, B. S. Shea, A. T. Kho, A. S. Sharif, A. M. Tager, and D. J. Tschumperlin, “Feedback amplification of fibrosis through matrix stiffening and COX-2 suppression,” J. Cell Biol. 190(4), 693–706 (2010). [CrossRef]
37. A. J. Booth, R. Hadley, A. M. Cornett, A. A. Dreffs, S. A. Matthes, J. L. Tsui, K. Weiss, J. C. Horowitz, V. F. Fiore, T. H. Barker, B. B. Moore, F. J. Martinez, L. E. Niklason, and E. S. White, “Acellular normal and fibrotic human lung matrices as a culture system for in vitro investigation,” Am. J. Respir. Crit. Care Med. 186(9), 866–876 (2012). [CrossRef]
38. P. C. Georges, J.-J. Hui, Z. Gombos, M. E. McCormick, A. Y. Wang, M. Uemura, R. Mick, P. A. Janmey, E. E. Furth, and R. G. Wells, “Increased stiffness of the rat liver precedes matrix deposition: implications for fibrosis,” Am. J. Physiol. Gastrointest. Liver Physiol. 293(6), G1147–G1154 (2007). [CrossRef]
39. C. Bocca, E. Novo, A. Miglietta, and M. Parola, “Angiogenesis and fibrogenesis in chronic liver diseases,” Cell. Mol. Gastroenterol. Hepatol. 1(5), 477–488 (2015). [CrossRef]
40. P. Wijesinghe, L. Chin, and B. F. Kennedy, “Strain tensor imaging in compression optical coherence elastography,” IEEE J. Sel. Top. Quantum Electron. 25(1), 1–12 (2019). [CrossRef]
41. S. Huang and D. E. Ingber, “Cell tension, matrix mechanics, and cancer development,” Cancer Cell 8(3), 175–176 (2005). [CrossRef]
42. D. E. Jaalouk and J. Lammerding, “Mechanotransduction gone awry,” Nat. Rev. Mol. Cell Biol. 10(1), 63–73 (2009). [CrossRef]
43. K. Pogoda, M. Cieśluk, P. Deptuła, G. Tokajuk, E. Piktel, G. Król, J. Reszeć, and R. Bucki, “Inhomogeneity of stiffness and density of the extracellular matrix within the leukoplakia of human oral mucosa as potential physicochemical factors leading to carcinogenesis,” Transl. Oncol. 14(7), 101105 (2021). [CrossRef]
44. J. Köhn-Gaone, J. Gogoi-Tiwari, G. A. Ramm, J. K. Olynyk, and J. E. E. Tirnitz-Parker, “The role of liver progenitor cells during liver regeneration, fibrogenesis, and carcinogenesis,” Am. J. Physiol. Gastrointest. Liver Physiol. 310(3), G143–G154 (2016). [CrossRef]
45. Q. Fang, A. Curatolo, P. Wijesinghe, Y. L. Yeow, J. Hamzah, P. B. Noble, K. Karnowski, D. D. Sampson, R. Ganss, J. K. Kim, W. M. Lee, and B. F. Kennedy, “Ultrahigh-resolution optical coherence elastography through a micro-endoscope: towards in vivo imaging of cellular-scale mechanics,” Biomed. Opt. Express 8(11), 5127–5138 (2017). [CrossRef]
46. T. Agarwal, B. Subramanian, and T. K. Maiti, “Liver tissue engineering: challenges and opportunities,” ACS Biomater. Sci. Eng. 5(9), 4167–4182 (2019). [CrossRef]
47. T. Kido, Y. Koui, K. Suzuki, A. Kobayashi, Y. Miura, E. Y. Chern, M. Tanaka, and A. Miyajima, “CPM is a useful cell surface marker to isolate expandable bi-potential liver progenitor cells derived from human iPS cells,” Stem Cell Rep. 5(4), 508–515 (2015). [CrossRef]
48. J. Foucher, E. Chanteloup, J. Vergniol, L. Castera, B. L. Bail, X. Adhoute, J. Bertet, P. Couzigou, and V. D. Ledinghen, “Diagnosis of cirrhosis by transient elastography (FibroScan): a prospective study,” Gut 55(3), 403–408 (2006). [CrossRef]
49. J. Fung, C.-K. Lee, M. Chan, W.-K. Seto, D. K.-H. Wong, C.-L. Lai, and M.-F. Yuen, “Defining normal liver stiffness range in a normal healthy Chinese population without liver disease,” PLoS One 8(12), e85067 (2013). [CrossRef]
