| Research BackgroundNecrosis is a common indicator of the occurrence and development of various diseases and is also one of the major risk factors for accelerating the deterioration of diseases. If proper diagnosis and intervention are not achieved in a timely manner, the evolution of necrosis in vital organs may lead to fatal outcomes. Therefore, highly sensitive detection and precise boundary delineation of necrotic lesions are crucial for clinical diagnosis and surgical treatment in order to achieve complete removal of the necrotic tissue as well as to minimize the loss of healthy tissue. Furthermore, these techniques are also extremely valuable for the prognosis of malignant tumors and evaluation of therapeutic effects. Therefore, different imaging strategies and contrast agents or probes have been proposed to detect necrosis.Clinically applied imaging modalities such as ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET) rely on either the perfusion of contrast agents in normal tissues or necrosis-avid imaging probes to indirectly or directly detect necrotic lesions. The indirect approaches have the disadvantages of inaccurate estimation of the necrotic margin and a short observation window. Applying necrosis-avid probes for direct imaging, such as radioisotope-labeled hypericin, can offer better overall performance; however, due to the limitation of conventional imaging modalities, it is challenging to achieve both high sensitivity for small necrotic lesion detection and precise definition of the necrotic boundary. Fluorescence molecular imaging (FMI) and associated intraoperative image-guided surgery have proven to be effective with respect to both sensitivity and boundary definition, demonstrating potential preclinical and clinical applications; however, these novel imaging techniques have not yet been applied for necrosis diagnosis and clinical treatment, mainly because of the lack of a suitable fluorescent probe.The typical method for developing disease-targeted fluorescent probes requires covalent conjugation of a targeting component (for example, a peptide or antibody) and a near-infrared (NIR) fluorophore. Although this strategy works well in preclinical applications, the synthetic conjugates are relatively large molecules and it is thus challenging to obtain immediate clinical translation due to the long time required in obtaining Food and Drug Administration (FDA) approval. Therefore, there is an urgent demand for an ideal fluorescent probe, a small molecule with superb necrosis specificity, with FDA approval for clinical applications. This would potentially enable the use of optical imaging techniques for the clinical diagnosis and treatment of necrosis-associated diseases with high sensitivity and high superficial resolution.Here, we demonstrate that indocyanine green (ICG), an FDA-approved NIR fluorescent dye, has previously undiscovered ability to target necrotic cells because of its interaction with lipoprotein (LP) and phospholipids, which is driven by its inherent chemical structure. We verified this mechanism through a series of in vitro, in vivo, and ex vivo experiments. We also investigated an improved ICG administration strategy to obtain a better signal-to-background ratio (SBR, the ratio of optical efficiency between the necrotic lesion and normal tissue). Furthermore, FMI and real-time image-guided surgery were applied to different animal models of necrosis-associated diseases using an in house-modified fluorescence microscope, which proved the high sensitivity and accurate necrotic boundary delineation ability of this novel imaging technique.We believe that this important discovery and its related optical imaging techniques have great potential for clinical translation in a wide range of necrosis-associated diseases.Materials and MethodsReagents. ICG was purchased from Dandong Medical and Pharmaceutical Co. Ltd., Dandong, China, which has approval for producing ICG for clinical applications by the Chinese FDA. Fetal bovine serum (FBS) was purchased from Life Technologies Corporation, Beijing, China PC was purchased from Sigma-Aldrich Co. LLC, Beijing, China.Mechanism discovery at the cellular level. The same generation of 4T1-fluc cells (obtained from the Chinese Academy of Military Medical Sciences) was used to compare images of the ICG and ICG+LP groups. Cell necrosis was achieved by removing the culture medium and adding distilled water (5 ml) into the culture dishes for 15 min. Fluorescence microscopy confirmed the swelling, rupture, and necrosis of the cells. ICG was dissolved in PBS and FBS, respectively, with the same concentration (1 μg/ml). Because of such a low concentration, all ICG molecules were bound with LP in FBS. Then, the ICG+PBS and ICG+FBS solutions (5 ml each) were directly added into two dishes of viable cells, respectively. Thirty minutes later, the cells were rinsed with PBS 3 times to wash out the free ICG or ICG-LP molecules. For the necrotic cells, the cell debris was separated from the supernatant by centrifugation (Multifuge X1 Centrifuge, Thermo Scientific, Massachusetts,1500 rpm,10 min). Then, the ICG+PBS and ICG+FBS solutions (5 ml each) were added into two centrifuge tubes with cell debris, respectively. Thirty minutes later, the cell debris was separated from the supernatant again and rinsed with PBS 3 times to wash out the free ICG or ICG-LP molecules. All procedures were performed at a constant temperature of 37℃. Finally, the samples of ICG or ICG-LP with viable and necrotic cells were imaged using our in house-designed fluorescence microscope. The experiment was performed in triplicate.Mechanism discovery at the molecular level. ICG was dissolved in PBS and FBS, respectively, with the same concentration (1 μg/ml). The ICG+PBS solution (200 μl) was added into the first column of a 96-well plate as the reference. The ICG+FBS solution (200 μl; containing ICG-LP molecules) was added into the second column. The ICG+FBS solution was mixed with PC, and then 200 μl of the mixture was added into the third column. All procedures were performed at a constant temperature of 37℃. Fluorescent images of the 96-well plate were acquired using the IVIS Spectrum system (default setting for fluorescence imaging of ICG:excitation 780 nm, emission 831 nm, automatic exposure; Caliper Life Sciences, Massachusetts). The experiment was performed in triplicate.Animal experiments. All animal experiments were conducted in compliance with the guidelines of the Animal Care and Ethics Committee of Zhujiang Hospital, Southern Medical University. Animal surgical and imaging procedures were performed under isoflurane gas anesthesia (3% isoflurane-air mixture). The mice and rats were sacrificed via intraperitoneal injection of 4% chloral hydrate (0.15 ml for mice and 1 ml for rats). All efforts were made to minimize suffering. The Balb/c nude mice and Sprague-Dawley (SD) rats were purchased from Vital River Company (Beijing, China). The mice were 6-8 weeks old and 20-25 g. Female mice were used for establishing 4T1 breast cancer xenografts, and male mice were applied for the other mouse models. The SD rats were all male,3-8 weeks old and 150-200 g.Mechanism discovery in mouse models. For the burn-induced hind limb necrosis model, six mice were anesthetized under isoflurane gas, and each received a continuous laser-beam burn (2 W output,808 nm wavelength, and 8 min irradiation) on the backside of its left hind limb using the MDL-H-808-2W system (Changchun New Industries Optoelectronics Technology, Changchun, China). Twenty-four hours later, all mice received IV injections of ICG (0.1 ml,0.5 mg/kg) and were imaged with the IVIS system at 5 min,6 h,12 h,18 h,24 h,5 days, and 9 days post-injection. The region of interest (ROI) of the necrotic area was manually delineated. The ROI or normal area was defined as the symmetrical body area of the right hind limb. Then, the corresponding optical efficiency and SBR were quantitatively measured and calculated for both the necrotic and normal tissues.For the muscle ischemia-reperfusion model, six mice were anesthetized, and the dorsal skin of each left hind limb was surgically opened; a muscle bundle was separated and both of its ends were ligated using a 1-0 silk thread to block blood supply. Twelve hours later, the ligation was surgically removed to achieve reperfusion of the blood24. Another 12 h later, all mice received Ⅳ injection of ICG, following the same protocol for image acquisition and analysis as described above for the burn necrosis mouse model. After the final imaging acquisition on the ninth day, the mice were sacrificed for TTC staining following the standard procedure (31). We then employed our in-house fluorescence microscope to acquire the fluorescent images for ex vivo verification.SBR enhancement experiment. Burn necrosis mouse models (n= 12) were established as described above. The control group (n= 6) received an IV bolus injection of ICG (2.0 mg/kg, which is the maximum dose approved by the FDA23). The experimental group (n= 6) received 4 intermittent injections (0.1 ml each,0.5 mg/kg) at 3-h time intervals. Fluorescent images were acquired from both groups using the IVIS system at 12 h,18 h,24 h, and 30 h post-injection. For the experimental group, the timing started from the last Ⅳ injection of ICG. The Ⅳ bolus injection and the last intermittent injection were performed at the same time for two groups.Fluorescence imaging using the in-house microscopy system. The fluorescence microscope was coupled with a conventional camera and a low-temperature CCD (PIXIS CCD, Princeton Instruments, New Jersey) to acquire both white-light and fluorescent images. The laser provided 775± 25 nm excitation, and the emission was obtained with 845± 25 nm filtering. All of the fluorescent images of organs were acquired with an aperture of F1.4 and exposure of 0.1 s, and all of the fluorescent images of H&E slices were captured with an aperture of F1.4 and exposure of 1.0 s. For the in vivo applications of the intraoperative image-guided surgery, the magnification was reset for each image based on the surgeon’s specific request. Since the fields of view of the two cameras were different, we developed an algorithm to achieve automatic imaging registration of the white-light and fluorescent images.Necrosis boundary definition and small-lesion detection. A 4T1-fluc xenograft mouse model (n= 6) was established by subcutaneously injecting 2 × 106 4T1 cells into the upper torso of each female mouse. One week later, bioluminescent images were acquired by the IVIS system to confirm tumour survival using the intraperitoneal injection of luciferase. Two weeks later, the average tumour lesion size reached 12 ± 2.7 mm, and each mouse received an IV injection of ICG (0.1 ml,0.5 mg/kg). After another 24 h, the mice were sacrificed, and the tumour lesions were resected and split. The white-light and fluorescent images of the tumour specimens were then acquired using our in-house fluorescence microscopy system. The specimens were stained for TTC, and white-light images were taken again using the fluorescence microscope.Five mice were anesthetized, and laser irradiation (2 W output,808 nm wavelength,8 min) was performed on the skin above the left side of the brain to cause brain necrosis. Twelve hours later, the 3 surviving mice received an IV injection of ICG (0.1 ml,0.5 mg/kg). Twenty-four hours after the injection, the mice were sacrificed. The heads were removed, immersed in a tissue-freezing medium (Leica Microsystems Nussloch GmbH, LEICA, Wetzlar, Germany, and frozen at -80℃). Then, slices were prepared using a cryostat microtome (Leica CM1950, LEICA, Wetzlar, Germany) to obtain coronal planes of the necrotic tissue. The sections were imaged and stained for TTC following the same protocol as described above for the 4T1 tumour experiment.Laparotomy was performed on the 3 mice to expose the liver. The blood circulation of a part of the left liver lobe was blocked using a liver occlusion clamp. That part was then injected with 0.1 ml pure ethanol. One minute after the injection, the clamp was removed, and the abdomen was with a two-layer suture. ICG (0.1 ml, 0.5 mg/kg) was Ⅳ-injected 24 h after the operation. Another 24 h later, the entire liver was excised for white-light and fluorescent imaging, and then sliced for H&E staining and imaged using fluorescence microscopy to verify the boundary of the necrotic tissue.Six rats were fed with water only 12 h pre-operatively, and the abdominal fur was removed. Laparotomy was performed on each rat to expose the anterior wall of the stomach. Acetic acid (0.01 ml; Beijing Chemical Works, Beijing, China) was injected in the gastric submucosa to create a gastric ulcer. In order to establish a double-blind experiment, the injection site was randomly chosen in the stomach by a surgeon without notifying the imaging operators. The abdomen was closed with a double suture, and about 72 h later, ICG was Ⅳ-injected (0.3 ml,0.5 mg/kg). Another 24 h later, the rats were sacrificed to obtain their stomachs. Each stomach was surgically opened along the greater gastric curvature, and the serosal and mucosal surface was imaged using the fluorescence microscope. After detection of the gastric ulcer, the tissue samples were stained for H&E, and the lesion boundary definition was verified with microscopy.Preclinical feasibility of the intraoperative image-guided surgery. In vivo escharectomy was performed on 3 burn necrosis mice, and the in-house fluorescence microscopy system was used to provide intraoperative imaging guidance. ICG was injected 24 h before the surgery. During the surgery, all fluorescent images were acquired with the room light off, but the surgical procedures were performed with the room light on. After anesthesia, the first set of fluorescent images was acquired to delineate the necrosis boundary, and then a surgeon removed the necrotic tissue. During this procedure, fluorescent images were acquired sporadically based on the surgeon’s request to evaluate the residual tissues until the necrotic tissue was completely removed.Using a similar procedure, in vivo debridement was performed on 3 bacterial abscess mice with intraoperative imaging guidance using the fluorescence microscopy system. The model was established by subcutaneously injecting 1 × 108 colony forming units of methicillin-resistant Staphylococcus aureus (type:ST-239) into the abdominal wall of each mouse. The strains were obtained from the Institute of Microbiology of Southern Medical University, following the US Committee for Clinical Laboratory Standards criteria. During the surgery, the fluorescent images were acquired sporadically based on the surgeon’s request to achieve complete removal of pus and necrotic tissues.Statistical analyses. Statistical comparisons were made using the Student’s t-test in GraphPad Prism 5 software. SBR values are expressed as mean ± SD. P values less than 0.05 were considered to be statistically significant. Means and SDs were calculated for experiments performed at least 3 times.ResultsThe mechanistic discovery of ICG to target necrosis.The molecular structure of ICG clearly demonstrates amphiphilic properties. Previous reports showed that after intravenous (IV) injection, the majority of ICG was rapidly bound to LP in the blood. The ICG-LP complex is too large (7-20 nm) to penetrate normal blood vessel walls, unless vascular permeability increases due to certain diseases such as a malignant tumor, inflammation, or trauma. However, even if the ICG-LP complex reaches the interstitial space, it is neither actively transported into living cells nor attached to any molecules on the cell membrane surface, except for hepatocytes, which are responsible for quickly excreting ICG into the biliary system and gut without the enterohepatic circulation. Furthermore, the metabolism and pharmacokinetics of ICG-LP in the lesion space of necrotic cells are completely different from those at other sites. Since LP shields the hydrophilic end of ICG, ICG-LP shows enhanced affinity for hydrophobic groups. Moreover, loss of cell membrane integrity due to necrosis exposes the hydrophobic tails of phospholipids, and thus ICG-LP might show distinct targeting ability for the phospholipids from the ruptured lipid bilayer.We here verified this hypothesis using a series of experiments at the cellular and molecular levels. In cellulo white-light and fluorescence images demonstrated that free ICG molecules dissolved in phosphate-buffered saline were bound to both viable and necrotic cells 4T1-fluc, whereas ICG-LP only showed affinity for necrotic cells and was completely washed out from viable cells. This suggested that once ICG was Ⅳ-injected into the in vivo circulation system, the naturally generated ICG-LP would show higher accumulation in necrotic tissues.In vitro quantitative comparison of the fluorescence efficiency between ICG with PBS (baseline), ICG with LP, and ICG with LP and phosphatidylcholine (PC) was performed in 96-well plates (Fig.1C). Significant signal enhancement (105.84% increase, P< 0.001) was observed in the ICG+LP+PC group compared to the ICG +LP group* indicating that the interaction between the ICG-LP complex and PC significantly enhanced the optical signal. PC is the most dominant phospholipid in the cell membrane and is abundant in necrotic tissues. This suggested that once ICG-LP was bound to necrotic cells, significant optical contrast between necrosis and its surroundings would be observed. The fact that optical signal enhancement of ICG-LP was only observed in necrotic cells verified our hypothesis and revealed the mechanism of how ICG-LP targets the ruptured lipid bilayer. Discovery of this new property of ICG indicates its good potential for distinguishing between normal and necrotic tissues for necrosis-associated diseases using FMI.In vivo and ex vivo verificationsFor further confirmation, a series of in vivo and ex vivo animal experiments were conducted. ICG solution (0.1 ml,0.5 mg ICG/kg body weight) was IV-injected into nude mouse models of burn-induced and muscle ischemia-reperfusion-induced hind limb necrosis (each model, n= 6). Fluorescent images were acquired over 24 h at 5 time points. Necrotic lesions showed clear optical contrast with the surrounding tissue at 12 h post-injection in both models. Although the profiles of optical efficiency and the SBR of necrotic lesions differed in the two mouse models, the necrosis avidity of ICG-LP was clearly demonstrated in both cases. The in vivo studies also proved that ICG could achieve a good necrosis targeting effect, even if the causes of necrosis are diverse.Quantitative measurements of optical efficiency and SBR were also performed on the fifth and ninth days, and similar profiles were observed for the two models, indicating that the washing out of ICG from necrotic lesions was much slower than that from normal tissues. This suggested a long observation window up to several days with just a single ICG administration. In addition, since both mouse models demonstrated that the SBR was consistently decreased at 24 h after the injection, we considered 24 h post-injection as the standard in vivo observation time point for subsequent experiments.After the in vivo verification, the ischemic necrosis mice were then sacrificed for TTC staining, and ex vivo fluorescent images were acquired using an in house-modified fluorescence microscope system. The system integrates a conventional microscope, a low-temperature NIR CCD, a color camera, and a laser generator to achieve real-time fluorescence imaging. Since the ex vivo TTC staining does not compromise the ICG distribution, they can corroborate each other for necrosis detection. For all of the specimens, the high optical contrast areas in the fluorescent and merged images were perfectly matched with the white tissue areas (necrotic tissues) in TTC staining. These results confirmed our findings.Improved ICG administration strategy to enhance the SBRThe half-life of ICG in the blood circulation is only 2-4 min, which restricts the delivery of ICG-LP into necrotic tissues. We therefore sought to determine whether the SBR could be increased using intermittent injections instead of a single IV bolus of ICG with the same dose. Using the burn-induced hind limb necrosis mouse model, the control group received a bolus IV injection of 0.1 ml ICG solution (2.0 mg/kg), and the experimental group received 4 intermittent injections of 0.1 ml ICG solution (0.5 mg/kg) at 3-h intervals. FMI was performed at 4 time points from 12 to 30 h post-injection. The optical efficiency of necrotic lesions was significantly larger than that of normal tissues at each time point in both groups (P< 0.001), which was consistent with our previous in vivo verifications. However, at each time point, the SBR of the experimental group was at least 2-fold greater than that of the control (P< 0.001), indicating that intermittent injections improved ICG-LP delivery to necrotic lesions. Therefore, the improved ICG administration strategy could significantly enhance the optical signal intensity of necrotic lesions without requiring an increase in the injection dose.Necrosis detection for various animal disease modelsNecrosis is a common indicator for the development of many serious diseases. The discovery of the necrosis avidity of ICG-LP opened the gate for employing FMI to diagnose necrosis-associated diseases. Therefore, we applied this imaging technique to various disease mouse and rat models, such as spontaneous necrosis of 4T1 breast cancer, brain necrosis, hepatic injury, and gastric ulcer, to evaluate its accuracy for necrosis boundary definition and sensitivity to small-lesion detection.The comparison between fluorescent images and TTC staining of the excised tumor and brain showed that the location and contour of the illuminated areas matched the necrotic lesions perfectly. This implies the potential of applying FMI with ICG for monitoring necrosis-associated oncotherapy and brain diseases.Comparison of hematoxylin and eosin staining and fluorescent images of stomach and liver specimens at the microscopic level showed that the boundary of the necrotic area was precisely delineated by the contrast of the fluorescent signals. In particular, in the liver specimen, the tiny normal tissues embedded in the necrotic areas could be distinguished by the optical images, indicating the high specificity of ICG for necrosis. Moreover, due to the inherent high sensitivity of FMI, tiny foci of necrosis could also be imaged. The smallest necrosis in the ulcer that was successfully detected was 0.6 mm in diameter (average ulcer size:1.0 ± 0.3 mm, n= 6), indicating the potential of this technique for applications requiring high detection sensitivity.The intraoperative image-guided surgeryRecently, real-time intraoperative image-guided surgery with FMI has been applied for sentinel lymph node biopsy, and ovarian and liver cancer resection, which has revealed superior sensitivity and precise boundary definition ability compared to conventional methods. Therefore, we developed an in house-modified fluorescent microscopic system that can be employed for real-time image-guided surgical navigation. Then, we applied the system and ICG administration to preclinical cases for feasibility assessment of the objective removal of necrotic tissues. The goal was to evaluate whether this imaging technique was valuable to conduct clinical translation to assist surgeons.For the escharectomy of a burn (coagulative necrosis) in living mouse models, the fluorescent images provided accurate preoperative localization and objective residue evaluation during the step-by-step resection process. Because of the high sensitivity and accuracy of this technique, complete removal of the necrotic tissue was achieved in all cases (n= 3) with maximum protection of the surrounding normal tissues. Even for a residue that was 1 mm in size, this optical technique could still be used as a guide for the surgeon to achieve accurate escharectomy.For the removal of a bacterial abscess (liquefactive necrosis) in living mouse models, even when the morphology of the necrotic tissue was continuously changing, the real-time fluorescent images could still offer objective intraoperative navigation and residue detection. In particular, when the majority of the abscess was removed, it was very difficult to estimate whether there was residue present with the naked eye, whereas the surgeon could easily obtain valuable information through the guidance of FMI.Conclusion:1. The discovery that ICG could target necrotic tissues and elucidation of the underlying mechanism via interaction between the ICG-LP complex and phospholipids may shift conventional awareness of the medical benefits associated with the use of ICG.2. An improved ICG administration strategy was proposed here, which can double the SBR under the FDA-approved dose. This strategy significantly enhanced ICG-LP delivery to necrotic lesions and thus offered better sensitivity for necrosis detection.3. There are several studies demonstrated the use of ICG for identifying different malignant tumors. We believe that ICG should be accumulated in the necrotic part inside the tumor4. There is no doubt that the application of ICG and intraoperative FMI can provide valuable information of necrosis in addition to that obtained with conventional diagnostic imaging scans and may benefit personalized patient care and contribute to more precise assessments of various necrosis-associated diseases. |
PDF Full Text Request