| BackgroundDevelopment and progression of cancer are now frequently associated with the changes of cancer metabolism. Since the 1920s, it has been discovered by Otto Warburg that, in contrast to most normal cells, cancer cells would underwent aerobic glycolysis instead of mitochondria tricarboxylic acid cycle regardless of whether oxygen is present. So tumor cells would reorganize to maintain the level of ATP and NADH to sustain their survival, proliferation and metastasis. This Theory has been fully studied and the metabolic switch has been applied in PET-CT clinically.In addition to aerobic glycolysis, another most important metabolic hallmarks of cancer cells is enhanced lipogenesis, this lipogenic conversion starts early and closely related to the development of cancer. Normally, normal cells predominantly absorb exogenous fatty acids and then transfer them into triglycerides for sufficient energy storage. However, cancer cells often undergo deteriorated endogenous fatty acid synthesis (FASN) irrespective of sufficient exogenous lipid supply, tumor cells would rather synthesize up to 95% of saturated and mono-unsaturated fatty acids (FA) depending on different tumor types. De novo FASN is closely related to glucose metabolism and accounts for almost all esterified FAs and more than 93% of triglycerides (TG) synthesis in tumors.However, fatty acid metabolism in tumor cells has not yet been fully explored.Colorectal cancer (CRC) is the third most common cancer in males and second most common in females worldwide. As one of the most common malignant tumors in the world, around one million people would suffer from CRC in each year, and half of them would die of this disease. CRC develops sporadically in an adenoma-carcinoma sequential process, over 95% of the CRC patients would benefit from early diagnosis. For early CRC detection, the colonoscopy is currently regarded as the gold standard for screening gastrointestinal neoplasia. Numerous emerging endoscopic techniques such as chromoendoscopy, magnification endoscopy, endoscopic ultrasonography were developed to improve preclinical diagnosis of neoplasia and acquired higher detection rate. Confocal laser endomicroscopy (CLE) serves as a new screening tool and has led to a new era of endoscopy, allowing for a quick diagnosis and clinical management of colorectal neoplasia. CLE provides real-time in vivo histology at cellular and subcellular resolution with images that are magnified up to 1000× and allow sub-surface imaging of the human mucosa during endoscopy.Fluorescently labeled fatty acid (BODIPY-FA,4,4-difluoro-5-methyl-4-bora-3a,4a-diaza-s-indacene-3-dodecanoic acid), a fatty acid analogue, is activated by an argon ion laser excitation at 488 nm and has generally been applied for experimental research in lipid trafficking. It can also be used in flow cytometry and confocal endomircoscopy.In our preliminary study, analysis of the Clinical Proteomic Tumor Analysis Consortium (CPTAC) data indicated that the expression levels of most rate-limiting enzymes that are involved in a series of metabolic reactions in the glycolytic pathway were increased in colorectal carcinoma tissues compared with matched normal colon tissues. We also found that enzymes that are related to endogenous FA synthesis, including FASN, ACC and ACLY, were expressed at significantly higher levels. In addition, ACT and ACAT showed decreased expression, which suggests that endogenous FAs are poorly utilized in colonic carcinoma. More importantly, we found that the transporters of long-chain fatty acids in CRC, such as FABP1 and Caveolin-1, decreased sharply. Thus, we wondered whether fatty acid absorption differed between normal colons and CRC as a result of this abnormal metabolism.ObjectiveTo detect the phenomenon and related mechanism of fatty acid absorption in colorectal neoplasia, and using confocal endomicroscopy as a carrier to further expand its clinical application.Method1. Q-PCR, Western-Blot and immunohistochemistry(IHC) to detect the different expression of long-chain-fatty-acid transporters.(1) Real-time PCR:total RNA was extracted and its concentration was detected, cDNA was synthesized according to the protocol of Takara PrimeScript RT Master Mix Perfect Real Time Kit, and then using TaKaRa SYBR Premix Ex Taq Kit to subsequent Real-time PCR amplification on an LC480 System.(2) Western-Blot:tissue protein was extracted and applying BCA assay to detect their concentration. Through the process of electrophoresis and electroporator, protein transferred to PVDF membrane. Then blocking the protein by 5%BSA for 1 hour, first antibodies incubated overnight at 4℃ with anti-CD36 (1:200), anti-FABP1 (1:500), anti-caveolin-1 (1:500),and anti-GAPDH (1:2000). Membranes were washed with TBST and were then incubated with a second antibody (1:2000) for 1h, finally membrane-bound second antibodies were detected by enhanced chemiluminescence.(3) Immunohistochemistry (IHC):colon regenerative polyps (10), low-grade intraneoplasia (10), high-grade intraneoplasia (10) and paired tissues of CRC and their matched normal ones (20) were collected. The detailed procedure was as followed:ordinary fix, embedd, deparaffinize and hydrate, antigen retrieval, natural coolind then using UltraVision Quanto Detection System HRP DAB Kit to block, first antibodies for caveolin-1 (1:250),FABP (1:250) and CD36 (1:200) should be incubated overnight at4℃, then DAB coloration. Haemotoxylin staining,backing blue, dehydration and mounting with neutral balsam were the end process.2. Cell culture and fluorescence-activated cell sorter (FACS)-based and confocal laser scanning microscopy (CLSM)-based BODIPY-FA uptake assays(1) All cells were obtained from the American Type Culture Collection (ATCC, Maryland, USA). Different CRC cell lines were cultured in Dulbecco’s Modified Eagle Medium (DMEM,10% fetal calf serum,1% glutamine,1% streptomycin/penicillin) at 37℃ and 5% CO2. When cells reached 60-80% confluence, they were plated in 12-well plates covered with aluminum foil at a density of 500,000 cells/well and incubated overnight for adherence. The growth medium was then removed, and cells were washed twice with phosphate-buffered saline (PBS). BODIPY-FA (D3823, Invitrogen, Carlsbad, California, USA) was dissolved in dimethyl sulfoxide (DMSO) as a stock solution, and the working solution was diluted with lx PBS and 0.1% FA-free bovine serum albumin (BSA). Different cell lines were cultivated in 2 μM BODIPY-FA for 1-15 minutes. Then, the cells were rinsed and harvested immediately for FACS.(2) We also isolated primary colon epithelial cells from cancerous and matched normal tissue samples after obtaining informed consent from the patients. Tissues were immediately placed on culture dish on ice and washed repeatedly. The freshly resected tissue was cut into small pieces with scalpel technique and enzymatically digested with collagenase type IV,10 mg/ml pronase E,2000U/ml DNAase and 0.1mM EGTA in HBSS solution for 30-60min in 37℃ in water bath.The dispersed tissue was mixed well by pipetting and filtered through a 100 μ m cell strainer to remove undigested tissue fragments. Cells then were washed with calcium-and magnesium-free Hank’s Balanced Salt Solution(HBSS) and centrifuged at 1000 r.p.m. for 2-3 min for twice. Cells were purified with Percoll and centrifugated at 600g for 20 min and collect the middle level. Epithelial cells were washed the same as the above. Then, acubate the cells in BODIPY-FA mix in three different time points, the above-mentioned steps were repeated and all experiments were performed in triplicate.(3) FHC and CRC cell lines were cultured in confocal dishes. When cells were fully stretched, the growth media were removed and the cells were incubated with BODIPY-FA mix at a concentration of 2μmol/L for 15 minutes. We then fixed cells with 4% paraformaldehyde for 30 minutes and nuclear was counterstaining with 4’,6-diamidino-2-phenylindole (DAPI,1:20000). Following fixation, cells were washed three times with pre-cold PBS. Images were acquired with a FV1000 CLSM.3. Patients and tissue specimens, imaging and classification(1) The colorectal neoplasia patients were from the Department of Gastroenterology or the Department of General Surgery, Nanfang Hospital, Guangzhou, China. These patients were selected based on the following factors: (1) complete clinical follow-up data were available, and (2) patients who had taken plentiful high-fat food or had been taking lipid medicine were excluded. The Institute Research Medical Ethics Committee of Nanfang Hospital granted approval for the present study (NFEC-2015-083).(2) As our previous experiments have demonstrated that resected mouse colon tissues would retain their biological activity and imaging well until 20 minutes after resection, we performed our agent using freshly excised tissues with the aim to imitate in vivo colonic environment. For ex vivo tissue imaging, human tissue samples of surgical resection of transition zones were immediately incubated with 250 μmol/L BODIPY-FA mix (0.1% BSA,1×PBS) for 5 minutes in 37℃ incubator (shielded from light with aluminum foil), after rinsing off the unbound BODIPY-FA for three times with cold PBS and dried by sterile swabs, tissue samples were immediately put to liquid nitrogen and subjected to subsequent cryostat sectioning. Every sites would prepare 2 pieces of tissue biopsies, one for the comparison of fluorescent signals by fluorescence microscope and nuclear was counterstaining with DAPI, the other was underwent HE pathology to assist diagnosis.(3) Fresh tissues were immediately incubated in the 250umol/L sterilized BODIPY-FA mix (0.1% BSA,1×PBS) for 5 minutes just before the endoscopy was performed. After rinsing off the unbound BODIPY-FA for three times with 0.9% sterilized normal saline, the confocal imaging was performed according to the protocol from the Pentax Corp. Shortly, the confocal window of the CLE laser probe was slightly placed on the surface mucosa and the distal tip of the CLE laser probe scanned the tissue from the superficial to deep layers and mean imaging time was-5 minutes. CLE was performed by two experienced endomicroscopists. The comparison of grayscale between nonneoplastic and cancerous locations was carried under the same lightness control. The targeted tissue samples by biopsies were fixed in 4% buffered formalin and subjected to subsequent H&E staining to facilitate the comparison with the prediction of the confocal images in a double blind pilot trial. The normal process of HE staining was performed according to the sequence of haemotoxylin staining, discrimination, back to blue, eosin staining, gradient dehydration and mounting with neutral balsam.(4) To assess the confocal images, clear images of each site were selected and evaluated according to previous classification by two different endoscopy doctors, with three main histopathologic characteristics taken into consideration. One characteristic was crypt morphology, which helped to classify the images into normal mucosa, hyperplastic locations and neoplasia (intraepithelial neoplasia [IN], adenocarcinoma). The second characteristic was nuclear visualization, which permitted differentiation among low-grade IN, high-grade IN and adenocarcinoma. The last characteristic was the apparent darkness of the neoplasm compared with normal areas of the same patient under the same conditions. The H&E biopsy images were judged by 2 independent, experienced experts and graded according to the modified Vienna classification. Meanwhile, clear images with intravenous fluorescein sodium were also selected, and judgments were made strictly according to previous classification criteria.4. Animal model construction and CLE imagingWe constructed the AOM/DSS-induced and xenograft tumor mouse models for CLE imaging. To generate the chemically induced mouse model,6-8 week-old male Balb/c mice were intraperitoneally injected with 10 mg/kg body weight azoxymethane [AOM]. One week later,2.5% DSS was dissolved in distilled water and was administered for 7 days, followed by 14 days of normal drinking water. This cycle was repeated three times; when the last DSS cycle ended, mice were prepared for in vivo staining. Our run-in phase experiments demonstrated that neoplastic regions always occurred heavy hematochezia with increased thickness and stiffness of the bowel wall macroscopically, which guided the topical application of fluorescence agents. Chosen mice were then fasted for 12h and then reveived intraperitoneal anaesthesia with 0.01g/ml pentobarbital of 6-7ul/g body weight. Exposing the abdomen and colonic tract were isolated and properly ligatured two length of lcm between two sutures, injecting 100ul fluorescent agents into the lumen and staining for 10min, unbound agents were then rinsed with 0.9% sterilized normal saline. The CLE procedure was performed as above and subsequent cryostat sectioning was performed according to above-mentioned protocol.5. Statistical analysis and evaluation and double blind trial for CLE diagnosisMean grayscale and mean fluorescence intensity were analyzed with Image J (NTH, Bethesda, MD, USA). ROIs (regions of interest) within CLE images of 100×100 μm from 10 different regions per image were selected to calculate the mean grayscale and mean fluorescence intensity. Data were analyzed with the statistical software package SPSS v13.0 (SPSS Inc. Chicago, IL, USA). Significance was determined using Student’s t-test, one-way or two-way ANOVA, or nonparametric tests, as appropriate. The Kolmogorov-Smirnov test was performed to compare relative protein expression in the primary data of 95 tumor samples from the Cancer Genome Atlas (TCGA). All p values were generated using 2-sided tests, and P<0.05 was defined as significant. All sections of photos were randomly coded and separately sent to two expert endomicroscopists who were blinded to the macroscopic appearance and H&E diagnosis. All data analysis was performed with SAS 9.4 software. Kappa coefficient were used to evaluate the diagnostic difference and consistency, respectively, between CLE with BODIPY-FA or fluorescein sodium targeting and H&E staining. The Kappa coefficient was also used to evaluate the inter-observer agreement of the CLE diagnosis with BODIPY-FA or fluorescein sodium staining. The Kappa value was ranked as follows:poor (0-0.20), fair (0.21-0.40), moderate (0.41-0.60), good (0.61-0.80), and excellent (0.81-1.0) agreement.Results1. Expression levels of FABP1, Caveolin-1 and CD36 in human colon carcinomaBased on the proteomic analysis, we examined the expression levels of major plasma membrane LCFA transporters, including FABP1, Caveolin-1, CD36, the FATP family (FATP4 mainly in intestine) and FABPpm. Compared with the matched normal tissues, the mRNA expression levels of FABP1, Caveolin-1 and CD36 in human CRC samples were decreased by approximately 4.8-,4.2-and 20.9-fold, respectively, and the protein levels of FABP1, Caveolin-1 and CD36 in the cancerous mucosa were reduced by approximately 1.8-,2.5-and 2.3-fold, respectively. We also performed immunohistochemical staining for FABP1, Caveolin-1 and CD36 in paraffin-embedded human CRC samples and matched normal samples. For FABP1 and Caveolin-1, the normal tissue exhibited more intense staining in the epithelial cell layer than in the adenocarcinoma mucosa. For CD36, the staining intensity progressively decreased from normal colonic mucosa, low grade IN, high grade IN to adenocarcinoma; epithelium staining of the carcinomas was significantly diminished or entirely absent compared with the strongly stained adjacent normal colon tissue. We hypothesized that the down-regulation of FABP1, CD36 and Caveolin-1 in CRC may enhance de novo FASN, which would then impede the import of exogenous fatty acids in colorectal neoplasms.2. Cancerous colon cells uptake BODIPY-FA slower and less than normal epithelial cellsWe then verified the FA metabolic difference in vitro, and FACS was performed to examine the FA intake of three human CRC cell lines (SW480, HCT116, and LOVO) and the normal colonic cell line FHC. Consistent with our hypothesis, BODIPY-FA constantly accumulated in adherent cells within 15 minutes, and FHC cells absorbed more BODIPY-FA than the CRC cell lines (P<0.001). Additionally, the FACS results of three pairs of fresh cancerous and matched normal colon tissues showed the same trend of fluorescence signal strength, and the normal primary colon epithelial cell absorption plateaued within 5 minutes. The ability of CLSM to locate and discriminate BODIPY-FA staining was corroborated, and a stronger cellular signal was seen in FHC compared with the CRC cells (SW480, LOVO). The specific BODIPY-FA signal was cytoplasmic and perinuclear.3. The non-neoplastic and cancerous loctations can be distinguished by BODIPY-FA with fluorescence microscope and CLEWe took analysis of specific fluorescent signals in 50 different sites of transition zones. The fluorescence intensities between non-neoplastic sites and neoplastic locations were then calculated using mean colonocyte fluorescence signal values,45 of 50 (90%) images showed expected fluorescent difference, and the average fluorescent intensity was 0.07±0.004 vs.0.04±0.003, non-neoplastic locations vs. neoplastic locations, respectively (p<0.0001). Upon CLE imaging of human samples, all non-neoplastic locations showed significantly stronger BODIPY-FA signals than the adjacent neoplastic sites, even after we halved the laser brightness on the non-neoplastic locations. We separately analysed the mean grayscale of all distinguishable confocal images (149 non-neoplasia images,103 low-grade IN images,132 high-grade IN images, and 124 adenocarcinoma images) of patients under the same excitation brightness control. The final statistical results verified that BODIPY-FA signals of non-neoplastic sites were almost 1.6-,2-and 2.2-fold higher than those of low-grade IN, high-grade IN and cancerous sites, respectively, but there was no significance between high-grade IN and adenocarcinomas, which indicated that lipogenic conversion may start in early colonic neoplasia.4. Staining characteristics of BODIPY-FA by CLE & Comparison between CLE diagnosis and clinical histological findingsWe also analysed the characteristics of BODIPY-FA staining with CLE. In normal sites, black nuclei on the basal side and bright cytoplasm facilitated visualizing the regular arrangement of crypts surrounded by epithelial cells and goblet cells. With regard to hyperplastic polyps, when focal aggregations of regularly shaped crypts or typical star-shaped luminal crypt openings were visible, the number of goblet cells was normal or slightly reduced. Specifically, BODIPY-FA staining on adenocarcinomas showed greatly modified or even absent glandular structures and identified many atypical cells with increased nuclear size and irregularly shaped nuclei. In low-grade IN, BODIPY-FA showed slightly atypical or elongated glands with a single layer of nuclei on the basal side; additionally, the ratio of nucleus/cytoplasm was almost normal, and the number of goblet cells was slightly decreased. Conversely, in high-grade IN, lateral fusion of destructed crypts, extremely elongated and branch-like crypts, and an increased nucleus-to-cytoplasm ratio were visible. A reduced number or complete absence of goblet cells and loss of cellular junctions were also identified, which were similar to conventional H&E staining. The inter-observer agreement of BODIPY-FA staining (κ=0.72, P＜0.001) was greater than that of fluorescein sodium (κ=0.61,P<0.001). In addition, we observed a non-significant diagnostic difference (chi-square=3.29, p=0.6553) and good diagnostic consistency (K=0.68, P＜0.001) between BODIPY-FA images and H&E diagnosis. However, although the diagnostic difference was slight (chi-square=1.10, p=0.9540), the diagnostic consistency of fluorescein sodium with the histological diagnosis was worse than that of BODIPY-FA (κ=0.43, P＜0.001).5. Colonic neoplasms showed decreased BODIPY-FA uptake with CLE in in vivo mouse modelsBased on the BODIPY-FA signal differences ex vivo, we further performed CLE with topically applied fluorescent agents to explore the absorptive difference between normal epithelium and neoplastic mucosa in vivo. In AOM/DSS-induced mice, we observed that the BODIPY-FA signals underwent a progressive decrease from inflammation to normal to CIN to adenocarcinoma based on acute fluorescent excitation observed with CLE. The fluorescence intensity of the normal mucosa was 2.5- and 1.7-times greater compared with adenocarcinomas and CIN, respectively. In addition, BODIPY-FA staining under CLE revealed more well-defined cellular structures, including disorganized cell arrangement and cellular atypia, in the neoplastic lesions than did other contrast agents. In contrast, topical administration of 2-NBDG with CLE showed no clear outline of the glands, let alone cytoplasmic imaging, and no fluorescent difference was observed between neoplastic and non-neoplastic tissues in the mouse models. For acriflavine staining, the visible nuclei were not specific to normal or malignant tissues.ConclusionsConlusions can be drew from above-mentioned experimental results:1. There is a sharp decrease expression of mRNA and protein level of FABP1, Caveolin-1 and CD36 in human colon carcinomas compared to their matched normal ones, which indicated that lower uptake of long chain fatty acid (LCFA) in colonic tumor might be associated with the decreased expression of plasma LCFA transpoters.2. Cancerous colon cells uptake BODIPY-FA slower and less than normal epithelial cells3. Both staining with fresh humen colon tissues in vitro and in the mouse models in vivo with CLE, the staining intensity progressively decreased from nonneoplastic colonic mucosa, low grade IN, high grade IN to adenocarcinoma; epithelium staining of the carcinomas4. BODIPY-FA is a cytoplasmic staining agent, which contributes to provide histology at cellular and subcellular resolution. In addition, short scanning time, distinct fluorescent differential, and particularly intracellular imaging may therefore make BODIPY-FA an attractive alternative to fluorescein sodium in the future, especially in early detection of colorectal lesions.5. BODIPY-FA staining with CLE showed a good diagnostic consistence with H&E staining and this research will open a new door to tracing pre-neoplastic or neoplastic lesions in scientific research and clinical practice. |
PDF Full Text Request