Open Access
Add to BibSonomy
Abstract
Sepsis is caused by dysregulated host inflammatory response to infection. During sepsis, early identification and monitoring of vascular leakage are pivotal for improved diagnosis, treatment, and prognosis. However, there is a lack of research on noninvasive observation of inflammation-related vascular leakage. Here, we investigate the use of photoacoustic microscopy (PAM) for in vivo visualization of lipopolysaccharide (LPS)-induced ear vascular leakage in mice using Evans blue (EB) as an indicator. A model combining needle pricking on the mouse ear, topical smearing of LPS on the mouse ear, and intravenous tail injection of EB is developed. Topical application of LPS is expected to induce local vascular leakage in skin. Inflammatory response is first validated by ex vivo histology and enzyme-linked immunosorbent assay. Then, local ear vascular leakage is confirmed by ex vivo measurement of swelling, thickening, and EB leakage. Finally, PAM for in vivo identification and evaluation of early vascular leakage using the model is demonstrated. For PAM, common excitation wavelength of 532 nm is used, and an algorithm is developed to extract quantitative metrics for EB leakage. The results show potential of PAM for noninvasive longitudinal monitoring of peripheral skin vascular leakage, which holds promise for clinical sepsis diagnosis and management.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Sepsis is defined as life-threatening organ dysfunction caused by dysregulated host inflammatory response to infection, and around 14,000 people die from complications every day worldwide [1]. Studies have shown that vascular endothelial injury caused by the release of inflammatory mediators is an important mechanism for the pathophysiologic process of sepsis [2–4]. The changes in vascular endothelial structure and function lead to microcirculatory dysfunction and then irreversible organ injury, which increases the cost of clinical treatment and directly impairs the prognosis of patients with sepsis [5–7].
During sepsis, inflammatory pathways are activated, leading to separation of endothelial cell connection, skeletal destruction, and intercellular space enlargement, which impair vascular endothelial barrier function (VEBF) and finally cause vascular leakage [8–11]. The albumin and fluid retained in blood vessels under physiological conditions infiltrate into tissue space, resulting in clinical hypoproteinemia, refractory hypotension, and organ hypoperfusion [11,12]. During inflammation-related vascular leakage, therapeutic fluid and drugs leak into tissues through highly permeable blood vessels, which not only fails to play a therapeutic role but also increases the incidences of irreversible organ injury, prolonging intensive care unit stay, and increasing mortality [13–16]. Early identification of vascular leakage during sepsis is crucial for timely accurate treatment and improved prognosis [17–20].
Currently, traditional methods, including physical examination of edema in the prolapse and effusion in the third space and laboratory tests for plasma albumin levels (e.g., the difference between hematocrit and plasma albumin), are used in clinical evaluation of VEBF [21,22]. However, these methods are indirect, intermittent (i.e., in contrast to continuous monitoring), and subjective, which hinder accurate evaluation of VEBF, especially considering that sepsis patients are complicated. Therefore, the technology to evaluate VEBF in a more direct, continuous, and objective way needs development.
Evans blue (EB) is a kind of azo dye with high albumin affinity. It is blue and non-toxic. It is clinically used to measure blood volume, trace lymph node location, and trace drug injections [23–25]. Under physiological conditions, the EB-albumin (EBA) complex is confined to vessels; EB diffuses to extravascular tissues when the vascular endothelial barrier is impaired and vascular permeability increases [24,26]. Radu et al. reported quantification of the difference in vessel permeability by spectrophotometrically measuring EB captured in different organs [27]. After intravenous injection of EB, EB concentration in extravascular tissues due to vascular leakage was used in the animal models for medical studies related to the blood-brain barrier [28,29], pulmonary vascular permeability [30,31], and intestinal barrier function [32,33]. However, the vascular leakage assessment based on EB concentration in extravascular tissues was performed using biological specimens (i.e., excisions) and cannot be promoted to clinical practice [28–33]. At present, there is limited technology for in vivo monitoring of the changes of vascular leakage through the body surface.
Photoacoustic (PA) microscopy (PAM), an emerging medical imaging technique in recent years, features high penetration and high resolution. It involves the detection of ultrasonic signals generated by short pulsed laser illumination of absorbers [34]. PAM can be used for structural and functional imaging of microcirculation systems by utilizing intrinsic optical absorption of hemoglobin to provide high contrast and high sensitivity [35–37]. Our previous study proved that PAM can quantitatively observe the changes in mouse ear microcirculation caused by lipopolysaccharide (LPS)-induced inflammation in vivo [35]. Besides intrinsic optical absorption, various exogenous chromophores (contrast agents) can be used to label certain tissues for contrast enhancement in PAM [38]. EB has strong absorption in visible and near-infrared light, with peak absorption at wavelength of 620 nm, which is a good contrast agent for PAM. Under physiological conditions, the EBA complex is distributed in blood vessels. Yao et al. found that using EB as a contrast agent can obtain PAM images with more complete capillary networks [39]. It is worth mentioning that PAM for the mouse ear model is particularly useful for biomedical studies [40–42], such as diabetic wound healing and insulin formulations dynamics.
As described previously, EB is a useful assay to test VEBF and is also a good contrast agent for PAM. Therefore, in this study, we propose using PAM for in vivo evaluation of inflammation-related vascular leakage with the following advancement. First, a model for LPS-induced vascular leakage in mice is developed. The model is thoroughly studied by conventional ex vivo methods using quantitative metrics of swelling, thickening, and EB leakage. Besides, vascular leakage of the model mainly occurs on mouse ears, which are easy to access and would facilitate noninvasive clinical evaluation of VEBF. Secondly, for the first time to our knowledge, in vivo investigation of LPS-induced vascular leakage with the use of EB as a tracer is demonstrated. PAM is expected to provide direct, continuous, and objective evaluation of vascular leakage. Our results suggest that PAM can be a promising new tool for quantitative and accurate evaluation of inflammation-related vascular leakage, and PAM has potential to be used in clinical practice for improving sepsis diagnosis and treatment.
2. Methods
2.1 Animal model
Male BALB/c mice aged 6–8 weeks (20–24 g) were purchased from the Institute of Laboratory Animal Science, Chinese Academy of Medical Sciences (Shanghai, China). The mice were housed at constant temperature (24 °C ± 2 °C) and humidity (50–60%) under a 12-h light/dark cycle with free access to standard food and water. Figure 1 illustrates the animal models used in this study. Figure 1(a) shows the sham group and the LPS group, the latter being a LPS-induced ear vascular leakage model. Specifically, the mouse ear hair was removed in advance. After 24 hours, the mouse was anesthetized (RC2 Rodent Anesthesia System) with an isoflurane pill. Then, both the left and right ears were symmetrically pricked (Fig. 1(a)) with a 1 ml insulin syringe (27 G needle) from the backside of the ear. As shown in Fig. 1(a), for each ear, 6 positions around the inner ear and 8 positions around the outer ear were chosen to prick. During pricking, the positions of blood vessels and perforations were carefully avoided. The above procedure was the same for the sham group and the LPS group. Then, for the sham group, both the left and right ears were smeared with normal saline of 20 µl for each ear. As for the LPS group, the left ear was smeared with normal saline of 20 µl, while the right ear was smeared with the same volume of LPS (L2630-100MG; Sigma-Aldrich, Saint Louis, Missouri) with concentration of 125 µg/µL.
Fig. 1. Animal models. (a) Illustration of the sham and LPS groups. (b) The model for the study of histology and ELISA. (c) The model for the study of swelling, thickening, and EB leakage. (d) The model for the study using PAM.
Three studies were conducted, as shown in Figs. 1(b)–1(d). For the study of histology and enzyme-linked immunosorbent assay (ELISA), 12 mice were randomly divided into 2 groups (n = 6 for the sham group and n = 6 for the LPS group). Then, the procedure detailed above (Fig. 1(a)) was applied. After 6 hours, the mouse was sacrificed, and the right ear of each mouse was harvested. The right ear was cut into two parts, one part fixed in 4% paraformaldehyde at 4 °C for histology study and the other part stored at −80 °C for ELISA study. Further steps are described later (Section 2.2). The same procedure in Fig. 1(b) was applied to each mouse.
For the study of swelling, thickening, and EB leakage, the initial procedure (needle pricking and LPS smearing) is the same as the study of histology and ELISA (see Figs. 1(b) and 1(c)). Similarly, 12 mice were used (n = 6 for the sham group and n = 6 for the LPS group). Different from Fig. 1(b), after 4 hours, the mouse was slowly injected with 2% EB of 100 µl (Sigma-Aldrich, Saint Louis, Missouri) by intravenous tail injection using a 1 ml insulin syringe (27 G needle). After 2 hours, the mouse was anesthetized, perfusion with normal saline was applied, and the mouse ear was harvested and cut in a disk with diameter of 8 mm. The mouse was then sacrificed. Similarly, the following steps are described later (Section 2.3), and the same procedure in Fig. 1(c) was conducted for each mouse.
For the study using PAM to allow in vivo evaluation of ear vascular leakage, the initial procedure (needle pricking, LPS smearing, and EB injection) is the same as the study of swelling, thickening, and EB leakage (see Figs. 1(c) and 1(d)). Note that 6 mice (n = 3 for the sham group and n = 3 for the LPS group) were used for the PAM study in Fig. 1(d). The EB injection procedure is the same as Fig. 1(c). 2% EB of 100 µl was injected intravenously into each mouse of both the sham and LPS groups. Different from Fig. 1(c), the mouse ears were imaged by PAM in vivo at 2 hours after EB injection. The PAM system and image processing are described later (Sections 2.4 and 2.5). The same procedure in Fig. 1(d) was also applied to each mouse. Animal-related experimental procedures were approved by the Animal Care and Use Committee of Renji Hospital, School of Medicine, Shanghai Jiao Tong University.
2.2 Histology and ELISA
As mentioned above, for each harvested right ear, one part was for histology study, and the other part was for ELISA study. The part for histology was then embedded in paraffin and sectioned into 5-µm slices. For morphological analysis, the sectioned slices were stained with hematoxylin and eosin (H&E) solution (Beyotime Biotechnology) and analyzed under an optical microscope (IX73, OLYMPUS).
The part for ELISA was warmed from −80 °C to ∼0 °C, and then was triturated with the addition of PSB (G4202-500ML, Servicebio). The triturated parts were then centrifuged at 15000 g at 4 °C for 10 minutes to collect the supernatant, and the protein concentration can be determined. The concentration of IL-6, IL-1β, and TNF-α in mouse right ears was quantified using a sandwich enzyme immunoassay (Mouse IL-6, IL-1β, TNF-α ELISA kit, Liankebio, Hangzhou, China) according to the manufacturer’s instructions, and was normalized to total protein content as applicable.
2.3 Swelling, thickening, and EB leakage
Swelling ratio, thickening ratio, and extravascular EB leakage ratio were used to assess ear vascular leakage. As mentioned previously, the mouse ear was cut in a disk. Each mouse ear disk was then weighted, and its thickness was measured. Then, each ear disk was cut into small pieces, which were then immersed in formamide solution at 50 °C for 72 hours for EB extraction. The solution containing the small pieces of the ear was then centrifuged at 15000 g at room temperature for 10 minutes to collect the supernatant. The EB absorbance of the collected supernatant was measured using a spectrophotometer (wavelength: 630 nm, Tristar2 S LB 942, Berthold). The higher the EB absorbance, the more the EB content. For each mouse, the left ear was used as the baseline for calculation. The swelling ratio is defined as
where ${w_r}$ and ${w_l}$ are the weight of the right ear and left ear, respectively. The thickening ratio is defined as2.4 PAM system
The schematic of the PAM system is shown in Fig. 2. A 532 nm pulsed laser (FDSS532-Q4, CryLaS) with pulse repetition frequency of 1 kHz and pulse duration of <2 ns was used for PA excitation. The laser beam was first split into two beams by a 1:9 beamsplitter (BS025, Thorlabs). The reflected beam (10%) was detected by a photodiode (DET10A2, Thorlabs) and used as trigger signals. The transmitted beam (90%) was attenuated by neutral density filters, and then shaped and expanded by a pinhole and two convex lenses. Then the beam was coupled into a 532 nm single-mode fiber using a fiber coupler (F-915, Newport) and a doublet lens (AC127-030-A-ML, Thorlabs). The beam exiting from the fiber was focused by a plano-convex lens (43-397, Edmund), not shown in Fig. 2, inside a homemade probe to excite PA signals. The probe was also used in our previous work, and the lateral resolution was determined to be 3.2 µm [43]. PA signals were detected by an obliquely placed needle hydrophone (central frequency: 35 MHz). Then, the detected PA signals were enhanced by a preamplifier (ZFL-500LN-BNC+, Mini-Circuits) and a pulser receiver (5073PR, Olympus), recorded by a digitizer (CSE1422, Gage) with sampling rate of 200 MS/s, and stored in a computer for further process. Both the probe and needle hydrophone, which form a scan head, were mounted on a two-dimensional (2D) linear motorized stage (M-404, Physik Instrumente), not shown in Fig. 2, to perform 2D scanning. During the scanning process, the step size was set to 8 µm, and PA signals from a region of 3.2 mm × 3.2 mm (400× 400 pixels) containing several pricked positions were acquired.
Fig. 2. Schematic of PAM system. BS, beamsplitter; PD, photodiode; NDF, neutral density filter; PH, pinhole; L, lens; M, mirror; I, iris; DL, doublet lens; SMF, single-mode fiber; NH, needle hydrophone; AMP, amplifier module; WT, water tank.
2.5 Image processing for quantitative evaluation of vascular leakage
The flowchart of the algorithm for quantitative evaluation of vascular leakage by PAM images is shown in Fig. 3(a). Note that the 2D maximum amplitude projection (MAP) PAM image was used as the original image. First, the Hessian matrix method, which searches for tubular structures in images, was applied to obtain a feature map from the original image [44,45]. Note that the original image was filtered by high-frequency emphasis before applying the Hessian matrix method [44]. From the feature map, the vessels can be better distinguished from the extravascular region. Secondly, the feature map was transformed into the binarized image for segmentation of vascular (the values ‘0’ assigned) and extravascular regions (the values ‘1’ assigned), the latter supposed to be the region containing extravascular EB leakage. Thirdly, element-by-element multiplication of the original image and the binarized image was performed to extract the extravascular image. Fourthly, thresholding was applied to the extravascular image to distinguish EB leakage (above the threshold value) from background noise (below the threshold value), and the EB leakage image was obtained by assigning the values ‘0’ to background noise in the extravascular image. The determination of the threshold value is detailed in Supplementary Section 1. The same threshold value was applied to all the extravascular images (i.e., 12 images). Finally, the EB leakage image was used to calculate quantitative metrics for vascular leakage.
Fig. 3. (a) Flowchart of the algorithm for quantitative evaluation of vascular leakage by PAM images. (b–f) Illustration of image processing by the algorithm using a representative PAM image: Original image (b), contrast-enhanced original image (c), feature map (d), extravascular image (e), and EB leakage image (f). Dynamic range for (b): 0.3 − 1.5 V; dynamic range for (c), (e), and (f): 0 − 1 V; separate colorbars are used for different dynamic range.
To illustrate image processing by the algorithm, a representative PAM image of the right ear (smeared with LPS) of a LPS-group mouse was processed, as shown in Figs. 3(b)–3(f). Figures 3(b)–3(f) show the original image, the contrast-enhanced original image, the feature map, the extravascular image, and the EB leakage image, respectively. Note that the raw data of the original image (Fig. 3(b)) and the contrast-enhanced original image (Fig. 3(c)) are the same. The difference is that Figs. 3(b) and 3(c) are displayed using different dynamic range so that Fig. 3(c) can display the extravascular region more clearly. More details are described as follows. Our experimental results show that the PA signal amplitude in order is “blood vessels” > “EB leakage” > “background noise” (also see Supplementary Section 1). The original image (Fig. 3(b)) is displayed with dynamic range of 0.3 − 1.5 V, and EB leakage is less obvious compared with blood vessels. On the other hand, the contrast-enhanced original image (Fig. 3(c)) is displayed using the same raw data as the original image (Fig. 3(b)) with dynamic range of 0 − 1 V. As a result, in Fig. 3(c), EB leakage, as well as background noise, becomes more obvious. Similarly, the processed images (Figs. 3(e) and 3(f)) displayed with the same dynamic range as Fig. 3(c) can better present EB leakage and background noise.
Three vascular leakage metrics are defined and used. Vascular leakage area is calculated by counting the total number of non-zero pixels of the EB leakage image and is denoted as N. On the other hand, vascular leakage strength is calculated by summing all the pixel values of the EB leakage image. Finally, vascular leakage average strength is calculated by dividing vascular leakage strength by N.
2.6 Statistical analysis
The statistical analysis was performed using GraphPad Prism 7 software (GraphPad, Inc., San Diego, CA, USA). The numerical results were presented as means ± standard deviations for the parameters with normal distribution and medians with interquartile range for the parameters with non-normal distribution. All data were checked for normal distribution with Shapiro-Wilk test. Unpaired t test was performed for the parameters with normal distribution to calculate the p value of comparison between two groups or contralateral ears, while Mann-Whitney test was performed for the parameters with non-normal distribution to calculate the p value of comparison between two groups or contralateral ears. The statistical significance was set at p < 0.05.
3. Results
3.1 Ex vivo evaluation of histology and ELISA
As mentioned in Section 2, the model for the study of histology and ELISA was used (Fig. 1(b)). Figure 4(a) shows a representative photo of a LPS-group mouse at 6 hours after the procedure in Fig. 1(a) (before right ear harvest). Then, the right ear was harvested for histology and ELISA.
Fig. 4. Ex vivo evaluation of histology and ELISA. (a) Representative photo of a LPS-group mouse at 6 hours after the procedure in Fig. 1(a). (b,c) Representative results of histology (H&E staining) from the right ears of a sham-group mouse (b) and a LPS-group mouse (c). Scale bars: 50 µm. (d) The concentration of IL-6, IL-1β, and TNF-α for the sham group and the LPS group. Data are expressed as means ± standard deviations for IL-6 and IL-1β (with normal distribution) and medians with interquartile range for TNF-α (with non-normal distribution). **, p < 0.01 compared with the sham group.
Figures 4(b) and 4(c) show representative results of histology (H&E staining) from the right ears of a sham-group mouse (Fig. 4(b)) and a LPS-group mouse (Fig. 4(c)). Skin, subcutaneous tissue, and cartilage are indicated by i, ii, and iii, respectively, in Fig. 4(b). Similar structures can be observed in Fig. 4(c). Compared with the sham group in Fig. 4(b), the LPS group in Fig. 4(c) shows inflammatory response (neutrophil infiltration indicated by the blue arrow in Fig. 4(c)), hemorrhage (red blood cells indicated by the red arrow in Fig. 4(c)), and edema (thicker tissue thickness indicated by the orange line in Fig. 4(c)). By contrast, no obvious inflammation, hemorrhage, and edema are observed in the sham group, as indicated by the blue arrow, the red arrow, and the orange line, respectively, in Fig. 4(b).
To validate the inflammatory response due to the LPS administration, the concentration of cytokine IL-6, IL-1β, and TNF- α from the harvested right ears of the sham-group mice (n = 6) and the LPS-group mice (n = 6) was measured by ELISA, as shown in Fig. 4(d). It can be seen that the concentration of IL-6, IL-1β, and TNF-α for the LPS group is 2 − 10 times higher than that for the sham group, verifying the LPS-induced inflammation in the right ears of the LPS-group mice.
3.2 Ex vivo evaluation of swelling, thickening, and EB Leakage
The model for the study of swelling, thickening, and EB leakage was used (Fig. 1(c)). Figures 5(a) and 5(b) show representative photos of a sham-group mouse (Fig. 5(a)) and a LPS-group mouse (Fig. 5(b)) right before perfusion and after harvesting ears (see Fig. 1(c)). For all the four harvested ears in Figs. 5(a) and 5(b), EB spread over the ears, particularly around the pricked positions. Besides, for the LPS group, EB’s blue color is more obvious (i.e., blue color is darker or over a larger area.) in the right ear than in the left ear (Fig. 5(b)), while for the sham group, EB’s blue color is similar in the right and left ears (Fig. 5(a)).
Fig. 5. Ex vivo evaluation of LPS-induced ear vascular leakage in mice. (a,b) Representative photos of a sham-group mouse (a) and a LPS-group mouse (b): The mouse head (left) and the ear region (top right) right before perfusion; the harvested ears (bottom right). (c,d) Swelling ratio (c) and thickening ratio (d) for the sham group and the LPS group. (e) The representative cut ear disks from a same sham-group mouse (top left), the representative cut ear disks from a same LPS-group mouse (top right), and the collected supernatant (bottom) of the shame-group mice (n = 6) and the LPS-group mice (n = 6). The supernatant of each ear disk sample is put in two adjacent wells (see the bottom photo), and thus, the EB absorbance is measured twice for each ear disk sample. Then, the average of the EB absorbance from the two measurements is calculated and used, which helps reduce experimental errors. (f) Extravascular EB leakage ratio for the sham group and the LPS group. L, left ear; R, right ear; sham-L, left ears of the sham-group mice; sham-R, right ears of the sham-group mice; LPS-L, left ears of the LPS-group mice; LPS-R, right ears of the LPS-group mice. Data are expressed as means ± standard deviations. *, p < 0.05; **, p < 0.01 compared with the sham group.
As mentioned in Section 2, after measuring the weight and thickness of each cut ear disk (8 mm in diameter; Fig. 5(e) (top)), swelling ratio and thickening ratio of cut ear disks were calculated for the sham group (n = 6) and the LPS group (n = 6) using Eqs. (1) and (2), as shown in Figs. 5(c) and 5(d), respectively. The results show that swelling ratio and thickening ratio of the LPS group are much higher than those of the sham group.
Then, ear vascular leakage was evaluated. As mentioned in Section 2, the supernatant was collected, as shown in Fig. 5(e) (bottom), and the EB absorbance was measured. As shown in Fig. 5(e) (bottom), the right ear in the LPS group (LPS-R) has the darkest blue color, while the rest ears (sham-L, sham-R, and LPS-L) are almost transparent. Then, extravascular EB leakage ratio was calculated for the sham group (n = 6) and the LPS group (n = 6) using Eq. (3), as shown in Fig. 5(f), whose result is consistent with Figs. 5(a), 5(b), and 5(e).
3.3 In vivo evaluation of LPS-induced ear vascular leakage in mice based on PAM
To study LPS-induced ear vascular leakage in mice by PAM in vivo, 6 mice (n = 3 for the sham group and n = 3 for the LPS group) were undergone the procedure in Fig. 1(d). PAM images of both the left and right ears of each mouse were taken. Figures 6(a) and 6(b) show two sets of PAM images, including original images (the top row) and the corresponding EB leakage images (the bottom row) processed by the algorithm mentioned previously, from one representative sham-group mouse and another representative LPS-group mouse, respectively. For original images in Figs. 6(a) and 6(b) (the top row), blob-like patterns (different from tubular structures) are recognized in the right ear of the LPS-group mouse (Fig. 6(b)), which are supposed to be due to EB leakage, but not in the left ear of the LPS-group mouse (Fig. 6(b)) or both ears of the sham-group mouse (Fig. 6(a)). As for EB leakage images in Figs. 6(a) and 6(b) (the bottom row), the right ear of the LPS-group mouse (Fig. 6(b)) shows distinct EB leakage (higher amplitude or over a larger area) compared with the contralateral ear (Fig. 6(b)) and both ears of the sham-group mouse (Fig. 6(a)). Figure 6(c) shows the representative photo of the LPS-group mouse right before in vivo PAM imaging, i.e., at 6 hours after LPS application (Fig. 1(d)). Note that the PAM images in Fig. 6(b) were taken from the mouse in Fig. 6(c). In Fig. 6(c), the blue color of the right ear is more obvious than the contralateral ear, which is consistent with Fig. 5(e) (the left and right ears in the LPS group).
Fig. 6. In vivo evaluation of LPS-induced ear vascular leakage in mice. (a,b) Representative two sets of PAM images from one representative sham-group mouse (a) and another representative LPS-group mouse (b): The original images (the top row; dynamic range: 0.3 − 1.5 V) and the EB leakage images (the bottom row; dynamic range: 0 − 1 V). Separate colorbars are used for different dynamic range. Scale bar: 500 µm. (c) Representative photo of the LPS-group mouse right before in vivo PAM imaging. The dashed boxes indicate the PAM imaged regions, and the acquired PAM images are shown in (b). (d − f) Comparison of the three metrics (vascular leakage area, vascular leakage strength, and vascular leakage average strength) for the sham group and the LPS group. L, left ear; R, right ear. Data are expressed as means ± standard deviations. ns, no significance; *, p < 0.05; **, p < 0.01 compared with the left ear.
For the sham group, signal-to-noise ratio (SNR) of in vivo PAM images (i.e., original images) is 20.4 ± 1.3 dB for the left ears and 19.9 ± 1.6 dB for the right ears of the sham-group mice (n = 3). For the LPS group, SNR of in vivo PAM images is 19.6 ± 0.4 dB for the left ears and 19.2 ± 0.5 dB for the right ears of the LPS-group mice (n = 3). The results show that SNR of in vivo PAM images for the four cases is similar. The calculation of SNR is described in Supplementary Section 2.
Further, the three metrics are compared, as shown in Figs. 6(d)–6(f). We first compare mean values and standard deviations of the two metrics (vascular leakage area and vascular leakage strength) in Figs. 6(d) and 6(e). For the sham group, the mean values of the two metrics are almost the same between the left and right ears. By contrast, for the LPS group, the mean values of the two metrics of the right ears are significantly larger than those of the left ears. The results demonstrate that PAM is capable of in vivo identification and quantification of vascular leakage. Besides, the two metrics are similar between “both ears of the sham-group mice” and “the left ears of the LPS-group mice”, which may indicate that topical smearing of LPS on one ear does not affect the contralateral ear in the LPS group. Interestingly, the metric (vascular leakage average strength) in Fig. 6(f) is similar among the four cases (the left and right ears for the sham group; the left and right ears for the LPS group), which may be explained by similar PA signal amplitude from the leaked EB for all the four cases. The exact reason requires further investigation.
4. Discussion
Vascular leakage is the most common and dangerous manifestation of endothelial injury during sepsis. In vivo identification and monitoring of vascular leakage during sepsis are of great value for clinical diagnosis and treatment. To our knowledge, this is the first time to demonstrate that PAM is capable of in vivo evaluation of LPS-induced vascular leakage using EB as an indicator.
Topical smearing of drugs on the mouse ear was used to induce inflammation and edema [46,47]. Subcutaneous injection of drugs, such as LPS and TNF-α, in the mouse ear skin was also reported to induce inflammation [48]. On the other hand, needle pricking followed by topical smearing of LPS on the mouse ear was used as the animal model for in vivo visualization of skin inflammation [49]. In this study, a model is developed to facilitate “in vivo monitoring of LPS-induced vascular leakage” by combining needle pricking (similar to acupuncture), topical smearing of LPS (to mimic endothelial barrier injury during sepsis), and then EB injection (as an indicator for vascular leakage), as shown in Figs. 1(a) and 1(d). It is worth noting that our model for in vivo evaluation of vascular leakage could also be studied by other optical imaging techniques, e.g., multiphoton microscopy [50], besides PAM demonstrated in this study.
As described previously, the LPS-induced ear inflammation model (Fig. 1(b), without EB injection) was validated by ex vivo evaluation of histology and ELISA (Fig. 4), demonstrating that topical smearing of LPS in our model can mimic the process of pathogen-activated inflammation during sepsis. Besides, the LPS-induced ear vascular leakage model (Fig. 1(c)) was verified by ex vivo evaluation of swelling, thickening, and EB Leakage (Fig. 5). Note that needle pricking may also cause ear swelling and thickening. To evaluate swelling and thickening due to LPS application rather than needle pricking, the contralateral ear was used as the baseline for calculation of swelling ratio and thickening ratio (Eqs. (1) and (2)). Finally, we confirmed LPS-induced vascular leakage by in vivo PAM imaging and image analysis (Figs. 1(d) and 6), which is qualitatively consistent with ex vivo evaluation of EB leakage (Fig. 5).
During sepsis, peripheral skin (e.g., fingers and ears) is often chosen for evaluation of microcirculatory perfusion and endothelial dysfunction [51,52]. Endothelial injury leads to increased vascular permeability and vascular leakage. EB has a high combination rate with plasma albumin and was used for ex vivo evaluation of vascular leakage in the intestine, brain, and lung in animal experiments. In vivo monitoring of vascular leakage may provide guidance during treatment of sepsis in clinical practice. PAM is a noninvasive imaging technology. It has high resolution (a few µm) and moderate penetration in skin (∼1 mm), making PAM very suitable to observe peripheral skin microvasculature down to capillaries. In this study, PAM with common excitation wavelength of 532 nm is used to visualize mouse ear blood vessels as well as EB. In order to differentiate extravascular EB from intravascular EB, efforts have been made to develop an algorithm to automatically extract extravascular EB, an important parameter to quantify vascular leakage. Significant difference in extravascular EB between the mouse ears with and without LPS application is obtained (Fig. 6). The results demonstrate that PAM at 532 nm together with the algorithm is capable to evaluate vascular leakage in vivo. Our work also shows promise of PAM for evaluation and monitoring of peripheral skin vascular leakage during clinical sepsis.
Previously, it has been demonstrated that in vivo PAM can be used to study the diffusion dynamics of EB leaving the bloodstream and the clearance dynamics of the EBA complex using the nude mouse ear model [39], which is without any treatment. Different from obvious EB leakage observed at 2 hours after EB injection in our LPS-induced ear vascular leakage model, obvious EB leakage has not been observed until 12 − 24 hours after EB injection in [39]. Note that similar EB injection is applied (3% of 50 µl in [39] vs. 2% of 100 µl in this study). The results suggest that LPS administration in our model impairs ear VEBF compared with the mouse model without any treatment in [39]. Note that although not demonstrated in this work, EB diffusion dynamics can be investigated using our model by in vivo PAM, which would be of interest for future work. On the other hand, in Ref. [39], we found that EB diffusion presents cloud-like patterns (Fig. 2 in [39]), while EBA clearance presents blob-like patterns (Fig. 3 in [39]). In this study, we also noticed blob-like patterns in the right ear in Fig. 6(b), which may result from the EBA diffusion according to [39].
The normality test with Shapiro-Wilk test shows that all the data in Figs. 4–6 were normally distributed, except for the concentration of TNF-α in the evaluation of ELISA in Fig. 4(d). Therefore, unpaired t test was used to calculate the p values in Figs. 4–6 except that Mann-Whitney test was used to calculate the p value of the concentration of TNF-α in Fig. 4(d).
In Fig. 2, the photodiode arranged before the pinhole is used to obtain trigger signals. Since the laser energy fluctuation was measured to be within ±5.3%, normalization of PA signals to laser energy to compensate for the laser energy fluctuation was not applied in this study. To accommodate the laser energy fluctuation, another beamsplitter and photodiode can be placed after the pinhole to obtain the laser energy for normalization of PA signals. This is because the photodiode placed before the pinhole may have the issue of nonlinear correlation between the measured laser energy and PA signals.
In this study, vascular leakage is caused by endothelial dysfunction due to LPS application. Compared with our previous work to evaluate structural changes of microvasculature [35], we demonstrate that PAM can evaluate functional changes in vascular endothelium in this work. As a potential tool for clinical diagnosis and treatment of sepsis, PAM can be applied for more explorations besides the experiments demonstrated in this work. First, in this study, mouse ears applied with certain LPS concentration are imaged by PAM at a fixed time (Figs. 1(d)). Mouse ears applied with different LPS concentration and monitored by PAM at different time points can be conducted to show more comprehensive evaluation of quantitative changes of EB leakage with time. Secondly, microvascular oxygen saturation, another key index for tissue oxygen delivery and organ dysfunction during sepsis, can also be monitored by PAM. These investigations would expedite the progress of PAM for clinical sepsis applications.
5. Conclusions
In summary, we demonstrated in vivo evaluation of LPS-induced vascular leakage by PAM. A model combining needle pricking, topical smearing of LPS, and then EB injection was developed, which allows local induction of vascular leakage in skin. The model with needle pricking and topical smearing of LPS (without EB injection) was first evaluated by ex vivo histology and ELISA, validating inflammatory response due to LPS application. Then, the model (with EB injection) was studied by ex vivo measurement of swelling, thickening, and EB leakage, verifying local vascular leakage due to topical smearing of LPS. Finally, PAM at 532 nm was employed to image mouse ears of the model in vivo, and the algorithm was developed to extract EB leakage images and the three quantitative metrics to evaluate vascular leakage. Our work shows promise for using PAM as a potential tool to continuously monitor peripheral skin vascular leakage during clinical sepsis for early intervention and treatment plan optimization.
Funding
National Natural Science Foundation of China (61775134, 81770067); Natural Science Foundation of Shanghai (22ZR1428900); Shanghai Jiao Tong University (YG2019ZDB04).
Disclosures
The 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.
Supplemental document
See Supplement 1 for supporting content.
References
1. L. Evans, A. Rhodes, W. Alhazzani, et al., “Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021,” Intensive Care Med. 47(11), 1181–1247 (2021). [CrossRef]
2. I. Cinel and S. M. Opal, “Molecular biology of inflammation and sepsis: a primer,” Crit. Care Med. 37(1), 291–304 (2009). [CrossRef]
3. C. Ince, “The microcirculation is the motor of sepsis,” Crit. Care 9(Suppl 4), S13 (2005). [CrossRef]
4. J. Joffre, J Hellman, C. Ince, and H. Ait-Oufella, “Endothelial responses in sepsis,” Am. J. Respir. Crit. Care Med. 202(3), 361–370 (2020). [CrossRef]
5. V. S. Kanoore Edul, C. Ince, and A. Dubin, “What is microcirculatory shock?” Curr. Opin. Crit. Care 21(3), 245–252 (2015). [CrossRef]
6. C. Fleischmann, A. Scherag, N. K. Adhikari, C. S. Hartog, T. Tsaganos, P. Schlattmann, D. C. Angus, and K. Reinhart, “International forum of acute care trialists. “Assessment of global incidence and mortality of hospital-treated sepsis. Current estimates and limitations,” Am. J. Respir. Crit. Care Med. 193(3), 259–272 (2016). [CrossRef]
7. S. M. Opal and T. van der Poll, “Endothelial barrier dysfunction in septic shock,” J. Intern. Med. 277(3), 277–293 (2015). [CrossRef]
8. C. Ince, P. R. Mayeux, T. Nguyen, H. Gomez, J. A. Kellum, G. A. Ospina-Tascón, G. Hernandez, P. Murray, D. De Backer, and ADQI XIV Workgroup, “The endothelium in sepsis,” Shock 45(3), 259–270 (2016). [CrossRef]
9. L. Martin, P. Koczera, E. Zechendorf, and T. Schuerholz, “The endothelial glycocalyx: New diagnostic and therapeutic approaches in sepsis,” Biomed. Res. Int. 2016, 1–8 (2016). [CrossRef]
10. A. Krüttgen and S. Rose-John, “Interleukin-6 in sepsis and capillary leakage syndrome,” J. Interferon Cytokine Res. 32(2), 60–65 (2012). [CrossRef]
11. D. L. Traber, “Microvascular leakage in sepsis,” Crit. Care Med. 28(3), 882–883 (2000). [CrossRef]
12. D. Zheng, J. Zhang, Z. Zhang, L. Kuang, Y. Zhu, Y. Wu, M. Xue, H. Zhao, C. Duan, L. Liu, and T. Li, “Endothelial microvesicles induce pulmonary vascular leakage and lung injury during sepsis,” Front. Cell Dev. Biol. 8, 643 (2020). [CrossRef]
13. J. Bakker, E. Kattan, D. Annane, et al., “Current practice and evolving concepts in septic shock resuscitation,” Intensive Care Med. 48(2), 148–163 (2022). [CrossRef]
14. E. Siddall, M. Khatri, and J. Radhakrishnan, “Capillary leak syndrome: etiologies, pathophysiology, and management,” Kidney Int. 92(1), 37–46 (2017). [CrossRef]
15. M. L. N. G. Malbrain, T. Langer, D. Annane, L. Gattinoni, P. Elbers, R. G. Hahn, I. De Laet, A. Minini, A. Wong, C. Ince, D. Muckart, M. Mythen, P. Caironi, and N. Van Regenmortel, “Intravenous fluid therapy in the perioperative and critical care setting: Executive summary of the International Fluid Academy (IFA),” Ann. Intensive Care 10(1), 64 (2020). [CrossRef]
16. P. Kumpers, F. Gueler, S. David, P. V. Slyke, D. J. Dumont, J. K. Park, C. L. Bockmeyer, S. M. Parikh, H. Pavenstadt, H. Haller, and N. Shushakova, “The synthetic tie2 agonist peptide vasculotide protects against vascular leakage and reduces mortality in murine abdominal sepsis,” Crit. Care 15(5), R261 (2011). [CrossRef]
17. J. L. Vincent, C. Ince, and P. Pickkers, “Endothelial dysfunction: a therapeutic target in bacterial sepsis?” Expert Opin. Ther. Targets 25(9), 733–748 (2021). [CrossRef]
18. M. M. Chen and X. C. Ma, “[Sepsis and endothelial injury: the important role of glycocalyx],” Zhonghua Nei Ke Za Zhi 59(6), 418–420 (2020). [CrossRef]
19. J. Bakker and C. Ince, “Monitoring coherence between the macro and microcirculation in septic shock,” Curr. Opin. Crit. Care 26(3), 267–272 (2020). [CrossRef]
20. I. Ocak, A. Kara, and C. Ince, “Monitoring microcirculation,” Best Pract. Res. Clin. Anaesthesiol. 30(4), 407–418 (2016). [CrossRef]
21. C. A. Brito, F. Albuquerque Mde, and N. Lucena-Silva, “Evidência de alterações de permeabilidade vascular na dengue: quando a dosagem de albumina sérica define o quadro? [Plasma leakage detection in severe dengue: when serum albumin quantification plays a role?],” Rev. Soc. Bras. Med. Trop. 40(2), 220–223 (2007). [CrossRef]
22. D. Dai, K. Tang, W. Xu, M. Li, Y. Su, Y. Wang, and R. Hu, “[The difference between hematocrit and plasma albumin in the course of systemic capillary leak syndrome: a systematic review],” Zhonghua Wei Zhong Bing Ji Jiu Yi Xue 30(10), 920–924 (2018). [CrossRef]
23. M. I. Harrell, B. M. Iritani, and A. Ruddell, “Lymph node mapping in the mouse,” J. Immunol. Methods 332(1-2), 170–174 (2008). [CrossRef]
24. L. Yao, X. Xue, P. Yu, Y. Ni, and F. Chen, “Evans blue dye: A revisit of its applications in biomedicine,” Contrast Media Mol. Imaging 2018, 7628037 (2018). [CrossRef]
25. E. B. Ehlerding, X. Lan, and W. Cai, ““Albumin hitchhiking” with an Evans blue analog for cancer theranostics,” Theranostics 8(3), 812–814 (2018). [CrossRef]
26. M. J. Wick, J. W. Harral, Z. L. Loomis, and E. C. Dempsey, “An optimized Evans blue protocol to assess vascular leak in the mouse,” J. Vis. Exp. 139, 57037 (2018). [CrossRef]
27. M. Radu and J. Chernoff, “An in vivo assay to test blood vessel permeability,” J. Vis. Exp. 73, e50062 (2013). [CrossRef]
28. B. Ahishali and M. Kaya, “Evaluation of blood-brain barrier integrity using vascular permeability markers: Evans blue, sodium fluorescein, albumin-alexa fluor conjugates, and horseradish peroxidase,” Methods Mol. Biol. 2367, 87–103 (2021). [CrossRef]
29. M. C. Wu, J. L. Hsu, and T. W. Lai, “Evans blue dye as an indicator of albumin permeability across a brain endothelial cell monolayer in vitro,” Neuroreport 32(11), 957–964 (2021). [CrossRef]
30. J. Moitra, S. Sammani, and J. G. Garcia, “Re-evaluation of Evans Blue dye as a marker of albumin clearance in murine models of acute lung injury,” Transl. Res. 150(4), 253–265 (2007). [CrossRef]
31. P. Smith, L. A. Jeffers, and M. Koval, “Measurement of lung vessel and epithelial permeability in vivo with Evans blue,” Methods Mol. Biol. 2367, 137–148 (2021). [CrossRef]
32. M. Shakeri, J. J. Cottrell, S. Wilkinson, W. Zhao, H. H. Le, R. McQuade, J. B. Furness, and F. R. Dunshea, “Dietary betaine improves intestinal barrier function and ameliorates the impact of heat stress in multiple vital organs as measured by Evans blue dye in broiler chickens,” Animals (Basel) 10(1), 38 (2020). [CrossRef]
33. T. V. Schäfer, J. J. Ivnitsky, and V. L. Rejniuk, “Permeability of intestinal and blood-tissue barriers in rats for Evans blue dye under conditions of acute intoxication with cyclophosphamide,” Bull Exp. Biol. Med. 168(1), 38–40 (2019). [CrossRef]
34. H. F. Zhang, K. Maslov, G. Stoica, and L. V. Wang, “Functional photoacoustic microscopy for high-resolution and noninvasive in vivo imaging,” Nat. Biotechnol. 24(7), 848–851 (2006). [CrossRef]
35. Z. Guo, Z. Li, Y. Deng, and S. L. Chen, “Photoacoustic microscopy for evaluating a lipopolysaccharide-induced inflammation model in mice,” J. Biophotonics 12(3), e201800251 (2019). [CrossRef]
36. Y. Ding, M. Zhang, J. Lang, J. Leng, Q. Ren, J. Yang, and C. Li, “In vivo study of endometriosis in mice by photoacoustic microscopy,” J. Biophotonics 8(1-2), 94–101 (2015). [CrossRef]
37. V. M. Sciortino, A. Tran, N. Sun, R. Cao, T. Sun, Y. Y. Sun, P. Yan, F. Zhong, Y. Zhou, C. Y. Kuan, J. M. Lee, and S. Hu, “Longitudinal cortex-wide monitoring of cerebral hemodynamics and oxygen metabolism in awake mice using multi-parametric photoacoustic microscopy,” J. Cereb. Blood Flow Metab. 41(12), 3187–3199 (2021). [CrossRef]
38. J. Yao and L. V. Wang, “Photoacoustic tomography: fundamentals, advances and prospects,” Contrast Media Mol. Imaging 6(5), 332–345 (2011). [CrossRef]
39. J. Yao, K. Maslov, S. Hu, and L. V. Wang, “Evans blue dye-enhanced capillary-resolution photoacoustic microscopy in vivo,” J. Biomed. Opt. 14(5), 054049 (2009). [CrossRef]
40. K. Maslov, H. F. Zhang, S. Hu, and L. V. Wang, “Optical-resolution photoacoustic microscopy for in vivo imaging of single capillaries,” Opt. Lett. 33(9), 929–931 (2008). [CrossRef]
41. N. Sun, B. Ning, K. M. Hansson, A. C. Bruce, S. A. Seaman, C. Zhang, M. Rikard, C. A. DeRosa, C. L. Fraser, M. Wågberg, R. Fritsche-Danielson, J. Wikström, K. R. Chien, A. Lundahl, M. Hölttä, L. G. Carlsson, S. M. Peirce, and S. Hu, “Modified VEGF-A mRNA induces sustained multifaceted microvascular response and accelerates diabetic wound healing,” Sci. Rep. 8(1), 17509 (2018). [CrossRef]
42. A. Khadria, C. D. Paavola, K. Maslov, F. A. Valenzuela, A. E. Sperry, A. L. Cox, R. Cao, J. Shi, P. L. Brown-Augsburger, E. Lozano, R. L. Blankenship, R. Majumdar, S. A. Bradley, J. M. Beals, S. S. Oladipupo, and L. V. Wang, “Photoacoustic imaging reveals mechanisms of rapid-acting insulin formulations dynamics at the injection site,” Mol. Metab. 62, 101522 (2022). [CrossRef]
43. W. Yang, W. Wang, L. Jing, and S. L. Chen, “Label-free photoacoustic microscopy: a potential tool for the live imaging of blood disorders in zebrafish,” Biomed. Opt. Express 12(6), 3643–3657 (2021). [CrossRef]
44. Z. Yang, J. Chen, J. Yao, R. Lin, J. Meng, C. Liu, J. Yang, X. Li, L. Wang, and L. Song, “Multi-parametric quantitative microvascular imaging with optical-resolution photoacoustic microscopy in vivo,” Opt. Express 22(2), 1500–1511 (2014). [CrossRef]
45. A. F. Frangi, W. J. Niessen, K. L. Vincken, and M. A. Viergever, “Multiscale vessel enhancement filtering,” Med. Image Comput. Comput. Assist. Interv. 1496, 130–137 (1998).
46. A. Bánvölgyi, G. Pozsgai, S. D. Brain, Z. S. Helyes, J. Szolcsányi, M. Ghosh, B. Melegh, and E. Pintér, “Mustard oil induces a transient receptor potential vanilloid 1 receptor-independent neurogenic inflammation and a non-neurogenic cellular inflammatory component in mice,” Neuroscience 125(2), 449–459 (2004). [CrossRef]
47. H. Inoue, T. Asaka, N. Nagata, and Y. Koshihara, “Mechanism of mustard oil-induced skin inflammation in mice,” Eur. J. Pharmacol. 333(2-3), 231–240 (1997). [CrossRef]
48. S. Arokiasamy, R. King, H. Boulaghrasse, R. N. Poston, S. Nourshargh, W. Wang, and M. B. Voisin, “Heparanase-dependent remodeling of initial lymphatic glycocalyx regulates tissue-fluid drainage during acute inflammation in vivo,” Front. Immunol. 10, 2316 (2019). [CrossRef]
49. B. Kim, S. H. Lee, C. J. Yoon, Y. S. Gho, G. O. Ahn, and K. H. Kim, “In vivo visualization of skin inflammation by optical coherence tomography and two-photon microscopy,” Biomed. Opt. Express 6(7), 2512–2521 (2015). [CrossRef]
50. J. L. Li, C. C. Goh, J. L. Keeble, J. S. Qin, B. Roediger, R. Jain, Y. Wang, W. K. Chew, W. Weninger, and L. G. Ng, “Intravital multiphoton imaging of immune responses in the mouse ear skin,” Nat. Protoc. 7(2), 221–234 (2012). [CrossRef]
51. W. Mongkolpun, P. Bakos, J. L. Vincent, and J. Creteur, “Monitoring skin blood flow to rapidly identify alterations in tissue perfusion during fluid removal using continuous veno-venous hemofiltration in patients with circulatory shock,” Ann. Intensive Care 11(1), 59 (2021). [CrossRef]
52. Y. Marcusohn, A. Ar, and J. J. Dirckx, “Perfusion and diffusion limitations in middle ear gas exchange: the exchange of CO2 as a test case,” Hear. Res. 265(1-2), 11–14 (2010). [CrossRef]
