The gastric mucosa specimens were divided into four groups according to the pathology results (1): Chronic Superficial Gastritis (CSG) group, characterized by the gastric mucosa having lymphocytes and plasma cell infiltration without changes in atrophy or intestinal metaplasia (2); Chronic Atrophic Gastritis (CAG) + Intestinal Metaplasia (IM) group, characterized by atrophy of the intrinsic glands in the gastric mucosa and by gastric mucosal epithelium that has transformed into the mucosal epithelium of the small or large intestine containing goblet cells (3); Dysplasia (Dys) group, also is preneoplastic lesion, characterized by gastric mucosal epithelium and glands that have deviated from normal differentiation and were atypical in shape and function, and can be divided in two three subgroups depending on the degree of atypia (mild, moderate, and severe); and (4) Gastric Cancer (GC) group, characterized by AGC lesions with cancer cells invasion beyond the submucosa to the muscularis propria or beyond. Gastric mucosa specimens were collected from CSG, CAG+IM, Dys, and GC patients through biopsy during their endoscopy procedures at the Wuxi People’s Hospital between October and December 2019. Three samples were collected for each group, and total 12 samples were kept in RNAlater (AM7020, Thermo Fisher Scientific) for the mRNA high throughput sequencing (HTSeq v0.11.2) analysis and reverse transcription‑quantitative polymerase chain reaction (RT‑qPCR) verification.

A total of 15 fresh specimens (mild=5, moderate=5, severe dysplasia=5) obtained from the ESD procedures at the Wuxi People’s Hospital between January and December 2020 were used for the in vitro probe studies, and then the specimens were fixed with formalin for the HE and IHC staining.

A total of 89 tumor and matched adjacent tumor tissues (62 from male patients and 27 from female patients) were collected from GC patients during their surgical procedures at the Wuxi People’s Hospital between January 2013 and December 2014 (the clinical data are shown in Table 1 ). The tissue chips were made by Shanghai ZuoCheng Bio Co., Ltd (Shanghai, China). All patients had not received chemotherapy or radiotherapy before their surgeries. The UICC TNM method was used for pathological classification (14), and all of the tissue chips were used for IHC staining.

The human GC lines AGS, MKN-45, and normal gastric epithelial cell lines GES-1 were purchased from the cell bank of the Chinese Academy of Sciences (Shanghai, China). Cells were cultured in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 g/mL streptomycin separately, the cells were culture in a 37°C humidified incubator with 5% CO2.

Src inhibitor was purchased from MCE MedChenExpress (cat no.179248-59-0), it was dissolved with dimethyl sulfoxide (DMSO) and got the finial concentration of 10μM. The 2.5×10⁵/ml AGS or MKN-45 cells were seeded in 6-well plates, 2ml/well, after 24h, the 2μl solution was added into cells/well, the control group was added the DMSO only, after 24 h, the medium was changed and the induced cells were infected with lentivirus for the following studies.

CEACAM6 overexpression lentivirus (Lenti-CEACAM6) and CEACAM6 RNAi lentivirus (CEACAM6 RNAi) with fluorescein expression purchased from Genechem company (Shanghai Genechem Co., Ltd), the AGS, MKN-45, and GES-1 cells were seeded in 6-well plates to reach about 30% confluence. HiTransG A (60 µl) and 7×10⁸ TU/ml CEACAM6 lentivirus (Lenti-CEACAM6, CEACAM6 RNAi) or negative control lentivirus (Lenti-NC) (4 µl) were added to 1.5 ml of the complete medium. After 12 h of incubation, the medium was replaced with the complete medium for 72 h for the in vitro and in vivo experiments described below.

The hand-held detectors of wide field multispectral microscope and a high-resolution microendoscope were supplied by the Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences (Suzhou, China).

To investigate the GC CEACAM6 mRNA expression in published works, we performed a search in TCGA (http://ualcan.path.uab.edu) using the combined keywords “CEACAM6 AND stomach adenocarcinoma.” The CEACAM6 expression in different clinic stages of GC was also investigated.

Total RNA was extracted from the gastric mucosa specimens of patients using Trizol (Invitrogen, Grand Island, NY, USA), according to the manufacturer’s instructions. The QIAGEN QIAseq FastSelect RNA Removal Kit (QIAGEN, Germany) was used to remove rRNA and achieve RNA fragmentation (average fragment length of approximately 200 nt). The purified RNAs were subjected to first strand and second strand cDNA synthesis following by adaptor ligation and enrichment with a low-cycle according to instructions of NEBNext^® Ultra™ RNA Library Prep Kit for Illumina (NEB, USA). The purified products were evaluated using the Agilent 2200 TapeStation and Qubit^®2.0(Life Technologies, USA) and then diluted to 10 pM for cluster generation in situ on the pair-end flow cell followed by sequencing (2×150 bp) HTSeq.

The clean reads were obtained after removal of reads containing adapter, ploy-N and at low quality from raw data. HTSeq was subsequently employed to convert aligned short reads into read counts for each gene model. Differential expression was assessed by DEGseq using read counts as input. The Benjamini–Hochberg multiple test correction method was enabled. Differentially expressed genes were chosen according to the criteria of fold change > 2 and adjusted q-value < 0.05. All the differentially expressed genes were used for volcano map analysis, the expression of CEACAM6 mRNA was verified by RT‑qPCR using the same total RNA for the sequencing above.

RT‑qPCR primer set for CEACAM6 (MQPS0000872‑1‑100) and β-actin (snRNa, MQPS0000002‑1‑100) purchased from Guangzhou RiboBio Co. Ltd. A total of 0.5 µg total RNA was used to synthesize cDNA in a 20 µl reaction volume at 42°C for 45 min using a PrimeScriptsis kit (cat. no. RR047a; Takara Bio, Inc.). RT‑qPCR was performed in a 20 µl reaction volume using iQ™ SYBR Green Supermix (cat. no. 1708882aP; Bio‑Rad Laboratories, Inc.) The thermocycling conditions were as follows: 95°C for 3 min, followed by 40 cycles of 95°C for 20 sec, 60°C for 30 sec, and 70°C for 30 sec. The β-actin was used as the internal control for CEACAM6, which was quantified using the 2^‑ΔΔCt method. All RT‑qPCR experiments were performed in triplicate.

The tissues were fixed and dehydrated, followed by paraffin embedding and cutting. The 5 µm-thick sections were rehydrated and subjected to antigen retrieval buffer, followed by incubation with primary antibodies against CEACAM6 (1:40, cat. no., ab275022, Abcam, USA), p-Akt (1:100, cat. no. 4060, CST, USA), p-PI3K (1:200, cat. no. 17366, Abcam, USA), p-Src (1:50, cat.no. 2105S, CST, USA), and MMP9 (1:100, cat.no. ab58803, Abcam, USA) for 1 h at room temperature. After washing thrice with tris-buffered saline (TBS), the sections were incubated with Envision-labeled polymer-horseradish peroxidase (HRP)-conjugated rabbit antibody (1:2,000; cat. no. 7074; Cell Signaling Technology, Inc.) for 1 h at room temperature, and were then visualized using diaminobenzidine.

All immunohistochemical sections were scanned by 3D HISTECH (pannoramic MIDI) and 10 randomly selected fields were checked under pannoramic viewer, the percentage of positive cells was analyzed by the densito quant software. The staining intensity was: strongly positive showed in dark brown was 3, the moderately positive showed in brown was 2, the weakly positive showed in light yellow was 1 and the negative showed in blue was 0. The positive cell percentage was calculated, and finally the histochemistry score (H-Score) was calculated. H-Score = ∑ (PI ×I) = (percentage of cells of weak intensity ×1) + (percentage of cells of moderate intensity ×2) + percentage of cells of strong intensity ×3), where PI is the percentage of positive cells in all cells in the section, I is the staining intensity.

All histopathological sections were read, diagnosed, and recorded by two senior pathologists.

For the cell invasion assay, 2.5×10⁴ infected AGS and MKN-45 cells obtained as mentioned above were seeded in 24-well plates with inserts. The two-chamber system was equipped with a cell-permeable membrane coated with Matrigel (R&D Systems, USA). After culturing for 48 h, cells that invaded through the matrigel and reached the bottom of the insert were fixed with formalin, and stained with crystal violet, the cells in the bottom chamber were considered invaded and were counted using Image-Pro Plus software.

AGS and MKN-45 cells infected with Lenti-CEACAM6 or Lenti-NC were seeded in 96-well plates with 10⁴ cells/well, with three replicates in each group. After 48 h in complete medium, cell proliferation was detected using Lights’ EdU Apollo567 kit (cat. no.C10310-1; Guangzhou RiboBio Co. Ltd) according to the manufacturer’s instructions. The cells were observed under fluorescence microscope, with the total live cells in blue and the proliferated cells in red. The cells number were estimated using Image-Pro Plus software and the proliferation rate= (number of proliferated cells/total number of cells) ×100%.

Cells were harvested and then lysed with radioimmunoprecipitation assay buffer (Pierce; Thermo Fisher Scientific, Inc.). The protein concentration was determined using the bicinchoninic acid assay, and proteins were separated using 10% SDS‑PAGE and were transferred onto polyvinylidene difluoride membranes. The membranes were blocked with 5% skimmed milk in TBS‑Tween‑20 (TBST), followed by overnight incubation at 4°C with rabbit anti-human CEACAM6 (1:500, cat. no. ab275022, Abcam, USA), Akt (1:1000, cat. no. 9272, CST, USA), p-Akt (1:2000, cat. no. 4060, CST, USA), PI3K (1:1000, cat. no. 4255, CST, USA), p-PI3K (1:500, cat. no. 17366, Abcam, USA), MMP9 (1:500, cat. no. ab58803, Abcam, USA), and p-Src (1:1000, cat.no.2105S, CST, USA), Src (1:1000, cat.no. 2109T, CST, USA), β-Tubulin (1:1000, cat. no.2146S, CST, USA) and β-actin (1:1000, cat. no.4970S, CST, USA) antibodies at 4°C. The membranes were washed with TBST thrice, followed by incubation with HRP-conjugated goat anti-rabbit IgG (cat. no.7074, Sigma-Aldrich) for 1 h at room temperature. The protein bands on the membrane were visualized using the SuperSignal West Pico Chemiluminescent Substrate (Pierce) and were quantified using Image-Pro Plus software.

The CEACAM6 monoclonal antibody (CEACAM6 mAb; NP_002474.4; Met 1-Gly 320) was purchased from Sino Biological (cat. no.10823-R408). CEACAM6 mAb was obtained from a rabbit immunized, the gene of antibody was characterized by phage display technology, and then, the targeted gene was constructed into eukaryotic expression vector, the vector was transfected into the 293 cells, lastly, the antibody was expressed and purified.

0.1 mg Alexa Fluo 488 was dissolved into 0.1 ml PBS(pH=8.4)buffer solution in the dark room, then 1 ml anti-CEACAM 6 mAb (1 mg/ml, cat. no.10823-R408, Sino Biological) was added into the above solution, mixed together and placed at room temperature for 2 h, then the mixture was centrifuged thrice at 1200 rpm using a desalination column (Pierce Zebra, 1 ml) and 0.01 M PBS was used as the mobile phase. The filtrate was collected and the anti-CEACAM6-mAb-Alexa Fluo488 probe was obtained.

0.5 mg IR Dye 800CW NHS was dissolved into 0.5 ml PBS (pH=8.4) buffer solution in the dark room, then 5 ml anti-CEACAM 6 mAb (1 mg/ml, cat. no.10823-R408, Sino Biological) was added into the above solution, mixed together and placed at room temperature for 2 h, then the reaction, the mixture was centrifuged thrice at 1200 rpm using a desalination column (Pierce Zebra, 5 ml) and 0.01M PBS was used as the mobile phase. The filtrate was collected and the CEACAM6-mAb-IRDye 800CW probe was obtained.

AGS, MKN-45, and GES-1 cells were cultured in dishes for 24 h for the confocal experiment. Then, the cells were fixed with 4% paraformaldehyde at 4 for 20 min, blocked with goat serum for 1 h at room temperature, and incubated with CEACAM6- mAb-Alexa Fluo488 probe synthesized as described above (1:20) at 4°C overnight. Finally, the nuclei were stained with DAPI and the cells were observed under confocal microscope. The average fluorescence intensity was calculated using the Image-Pro Plus software (Media Cybernetics, Rockville, MD, USA).

In order to test if the conjugate (IRDye800CW NHS) affect the activity of the antibody, the ELISA (Enzyme Linked Immunosorbent Assay) method was applied. The recombinant CEACAM6 protein (CD66c, cat. no. 10823-H08H, Sino Biological) was dissolved with PBS, then put the 50ul CEACAM6 dissolving solution (100ug/ml) into the 4950ul coating buffer and got the coating solution (1ug/ml). Put the coating solution into the 96-well ELISA plates (100ul/well) and incubated at 4°C for 20 h, then the plates were washed for three times and blocked with 1%BSA-PBST blocking buffer at 37°C for 2 h, after washing, the CEACAM6 mAb or CEACAM6 mAb-IRDye800CW with different dilution times was added into the 96-wells (100ul/well), and incubated at 37°C, after 2 h, the plates were washed and the anti-rabbit secondary antibody was added into the wells (100ul/well), after incubating at room temperature for 1 h, the plates were washed, the color was developed with tetramethylbenzidine (TMB) ELISA substrate (100ul/well) (abcam,ab171523) and terminated with 1M H₂SO₄ (50ul/well), finally the OD value was read at 450nm.

when calculating half maximal inhibitory concentration (IC50) value, subtracted the OD value when the concentration of antibody is 0 from each OD value, then the value obtained by subtraction was input into graphpad prism 8.0.2 software for the curve fitting, and the IC50 value was got according to the fitting curve, the lower the IC50, the higher the activity of the antibody.

The kinetic measurement was performed at 25°C. HBS-EP buffer (10X) was used to dilute CEACAM6 mAb or CEACAM6-mAb-IRDye800CW antibody, 20μg/ml was the highest concentration, 2-fold dilution of 20μg/ml till got 0.3125μg/ml. The gradient diluted antibodies were injected to Fc1 or Fc2, the flow rate was 30µl/min, the association time was 180s and dissociation time was 600s. 10mM glycine HCl (pH 1.5, cytiva, cat no.BR100354) was selected to regenerate the combined CEACAM6 mAb or CEACAM6-mAb-IRDye800CW antibody, and the regeneration time was 30s. Different concentrations of the antibody were injected repeatedly to bind, dissociate and regenerate. Finally, the association (Ka) and dissociation (Kd) rate constants were evaluated using Biacore T200 evaluation software (GE Healthcare Bio-sciences, Sweden), and the average dissociation constant (KD= Kd/Ka) was got. The 1:1 binding model was used for evaluating the binding of CEACAM6 mAb or CEACAM6-mAb-IRDye800CW to CEACAM6 (CD66).

Female nude mice (6–8 weeks old) were provided by the Changzhou Kaiwen Laboratory Animal Co., Ltd. (Jiangsu, China). The mice were maintained under specific pathogen-free conditions and were housed in ventilated cages with free access to food and water. The mice were allowed to acclimate for one week. They were anesthetized with 2% isoflurane for euthanasia and the cervical dislocation method was used to confirm death. The animal procedures were carried out in accordance with the guidelines of the Nanjing Medical University Laboratory Animal Center.

MKN-45 cells transfected with Lenti-CEACAM6, CEACAM6-RNAi or Lenti-NC were cultured in 10% FBS-containing DMEM and were then harvested. Approximately 4×10⁷ cells were resuspended in 100 µl of saline and were subcutaneously inoculated into the mice. 30 days after inoculation, when the tumors have grown to about 200 mm³, the tumor volume (0.5×length×width²) was recorded every 5 days, and the mice were used for the in vivo imaging study described below, the tumor and liver tissues were harvested to HE and IHC staining.

The CEACAM6mAb-IRDye800CW was injected through the tail vein using the following doses: 2.5 mg/kg, 5 mg/kg, and 10 mg/kg. The animals were observed at 2 h, 1 d, 2 d, 4 d, and 7 d after the injection, and were anesthetized with sodium pentobarbital by intraperitoneal injection at a dose of 40 mg/kg. After several minutes, the animals were placed in the in vivo imaging system (IVIS) (PerkinElmer) for near infrared imaging. The animals were euthanized after the last examination, with the cervical dislocation method used to confirm death, then the tumor as well as other organs were removed for near infrared imaging. The data were collected and analyzed using living image software (Xenogen) according to the manufacturer’s instructions.

The CEACAM6 mAb-IRDye800CW probe was sprayed on the surface of fresh tissues obtained from the ESD procedure, staining for 30 min at room temperature. Then, the tissues were washed thrice with running water and the fluorescent probe was completely removed from the surface. Fluorescence was checked using the fluorescent endoscopy detector and the surface microstructure was observed using detector of the fluorescence microscopic endoscopy detector.

All experiments were performed in triplicate, and all data were expressed as mean ± standard error. GraphPad Prism 5 software (GraphPad Software, Inc., GPW5-384305-RAG-5235) was used to perform statistical analyses. Differences between two groups were analyzed using the unpaired t-test, and the one-way analysis of variance (ANOVA) followed by Bonferroni’s post hoc test was used to analyze the differences between more than two groups. The paired datasets (adjacent tumor and tumor) were analyzed using the Wilcoxon matched-pairs signed rank test. A p value <0.05 was considered statistically significant. The correlation between fluorescence intensity and CEACAM6 staining H-Score was studied by calculation of Pearson correlation coefficients. Kaplan-Meier survival curve analysis was used to identify the difference between patients with high and low CEACAM6 expression, and the Cox risk proportional regression model was used to determine the GC risk factors.

Data from TCGA showed that CEACAM6 mRNA is up regulated in cancer tissues compared to their levels in normal tissues in stomach adenocarcinoma ( Figure 1A ). Furthermore, the CEACAM6 expression in cancer tissues with clinical grades 2, 3, and 4 are much higher than in the normal tissues ( Figure 1B ). However, there are currently few studies about CEACAM6 expression during GC progression. Thus, in order to investigate the expression and possible role of CEACAM6 in gastric mucosa carcinogenesis, the mRNA sequencing was performed on gastric mucosa specimens with CSG, CAG+IM, Dys, and Cancer, and the genes of CEA family member were verified by RT-qPCR method (only the data of CEACAM6 mRNA was showed). Compared with CSG, there were 882 up regulated and 633 down regulated mRNAs in the CAG+IM group, and among these mRNAs, the CEACAM18 was verified up regulated and CEACAM21 was down regulated ( Figure 1C ). Moreover, there were 1581 up regulated and 876 down regulated mRNAs in the Dys group, and among these, CEACAM1, 3, 4, 6, 18, 20, and 22 were verified up regulated ( Figure 1D ). In the Cancer group, 4429 mRNAs were up regulated and 2937 mRNAs were down regulated, and among these, CEACAM 3, 4, 5, 6 and 20 were verified up regulated and CEACAM19 was down regulated ( Figure 1E ). From the data above, it was showed CEACAM6 was up regulated (P<0.05) in both Dys and Cancer groups. The mRNA sequencing data has been submitted to the GEO repository and the series record is GSE163416.

According to the data provided above, CEACAM6 mRNAs were over expressed in both the Dys and Cancer groups. To detect the CEACAM6 protein in the GC, we sought to detect the expression patterns of CEACAM6 in GC tissue chips using IHC staining. The results showed that the positive rate of CEACAM6 was 66.29% (59/89) and 4.49% (4/89) in the tumor tissues and the adjacent tumor tissues, respectively, and the staining H-Score in tumor was 39.56 ± 6.01 ( Figures 2A, B ). From the clinical data shown in Table 2 , patients with TNM classifications of stage T3-4 have higher CEACAM6 expression than those with stages T1-2 (P=0.026), and the patients with pathological classifications of grades 3-4 have higher CEACAM6 expression than those with grade 2 (P=0.007). Furthermore, the data showed that distant metastasis and N classification were closely related with the prognosis of patients and were independent risk factors for GC prognosis.

During the end point of 60 months of follow-up, a total of 51 patients died whereas the other 38 remained alive. The survival analysis showed that the patients with higher positive CEACAM6 expression have shorter survival time than those with lower CEACAM6 expression ( Figure 2C ). To investigate the CEACAM6 expression in dysplastic gastric mucosa, tissues of patients with mild, moderate or severe dysplasia obtained from the ESD procedures were stained with anti-CEACAM6 mAb using IHC staining. The results show that compared to the adjacent gland, CEACAM6 was highly expressed in the gland with dysplastic changes ( Figures 2D–F ). *** P <0.001.

In order to study the roles of CEACAM6 in GC cells, the AGS and MKN-45 cells were infected with Lenti-CEACAM6 or Lenti-NC. The infection efficiency was checked using fluorescence microscopy and western-blot ( Figure 3A ), and the cell invasion and proliferation were detected 48 h after infection. The transwell assay results showed that CEACAM6 over expression promoted GC cell invasion ( Figure 3B ). Moreover, the EdU assay showed that CEACAM6 can promote the proliferation of GC cells ( Figure 3C ). *P<0.05, **P<0.01, *** P <0.001.

Studies found Src/PI3K/Akt signaling pathway activation is responsible for the GC tumorigenesis (15, 16), furthermore, the western-blot was applied to study the roles of CEACAM6 in proteins of Src/PI3K signaling pathway, with the results showing that CEACEM6 over expression can up regulate the expression of p-Src, p-PI3K, p-Akt and MMP9 proteins in the AGS and MKN-45 cells, however, when the Src was inhibited, the down regulation of p-PI3K, p-Akt and MMP9 cannot be recovered by CEACAM6 ( Figure 3D ). It was speculated that CEACAM6 can accelerate the GC tumorigenesis and progression by Src/PI3K/Akt signaling pathway ( Figure 3E ).

To further investigate the role of CEACAM6 in GC cell tumorigenesis in vivo, the subcutaneous tumors from CEACAM6 over expressed MKN-45 cells in a xenograft mouse model were studied. We found that CEACAM6 (Lenti-CEACAM6) over expression led to a significant increase in the tumor volume in a time-dependent manner, as compared to the Lenti-NC group ( Figures 4A, B ). From the data obtained above, it was found that CEACAM6 can increase the expression of p-Src, p-PI3K, p-Akt, and MMP9 proteins. Then, we checked the expression of these proteins in the subcutaneous tumor tissues and found that high CEACAM6 expression can also promote the expression of p-Src, p-PI3K, p-Akt, and MMP9 proteins in vivo ( Figure 4C ). The results further confirmed that the activation of the Src/Akt/PI3K/MMP9 signaling pathway may be a mechanism by which CEACAM6 promotes GC cell metastasis in vivo

In order to check the affinity of the fluorescent-labeled anti-CEACAM6 probe with GC cells, the CEACAM6-mAb-Alexa Fluo488 probe was synthesized, the AGS, MKN-45, and GES-1 cells were cultured in confocal specific dishes, and the cells were incubated with the fluorescent-labeled anti-CEACAM6 probe (CEACAM6- mAb-Alexa Fluo488). After incubation, the cells were observed under confocal microscopy. It was found that the probe was assembled on the cell membrane, as expressed by the CEACAM6 (green fluorescence) ( Figure 5A ). Moreover, the fluorescence intensity was highest in the MKN-45 cells ( Figure 5B ). ***P<0.001. The chemical molecular structure model diagram of CEACAM6-mAb-Alexa Fluo488 was showed in Figure 5C . It was suggested that the probe had the strongest affinity to MKN-45 cells.

The near infrared-labeled anti-CEACAM6 probe (CEACAM6 -mAb-IRDye800CW) was synthesized to study the tracking role of the anti-CEACAM6 probe in GC in vivo, the chemical molecular structure model diagram of CEACAM6 -mAb-IRDye800CW was showed in Figure 5D . The activity of the probe was evaluated using the ELISA method. The results showed that the CEACAM6 antibody labeled with IRDye800CW has the similar IC50 value (0.3501 ng/μl) with that of the CEACAM6 antibody without labeling (the IC50 value was 0.5705 ng/μl). Therefore, it was indicated that the activity of the CEACAM6-mAb-IRDye800CW was high and was not influenced by conjugates ( Figure 5E ).

In order to further evaluate the affinity of CEACAM6-mAb-IRDye800CW to CEACAM6, the surface plasmon resonance (SPR) analysis was performed. For CEACAM6-mAb, the association (Ka) rate constant was 1.422×10⁵ M^-1s^-1, the dissociation (Kd) rate constant was 2.2876×10^-5s^-1, and the average dissociation constant (KD) was 2.022×10^-10 M ( Figure 5F ). For CEACAM6-mAb-IRDye800CW, the Ka rate constant was 4.996×10⁴ M^-1s^-1, the Kd rate constant was 2.989×10^-5 s^-1, and the KD was 5.983×10^-10 M ( Figure 5G ), from the data above, it was showed the CEACAM6-mAb-IRDye800CW had the similar Kd rate constant to CEACAM6 mAb, the Ka rate constant of CEACAM6-mAb-IRDye800CW was slightly lower than CEACAM6 mAb, then it was found the affinity of labeled antibody was slightly lower than unlabeled antibody accord to the KD value.

The anti-CEACAM6 demonstrated strongest affinity to the MKN-45 cells, hence the MKN-45 cells infected with Lenti-CEACAM6 or CEACAM6-RNAi were used for the xenograft mouse GC model construction and in vivo study. The CEACAM6-mAb-IRDye800CW probe was injected into the mouse model through the caudal vein and the whole-body imaging was acquired using animal computed tomography. When the CEACAM6-mAb-IRDye800CW with different dose was injected in the caudal vein of the mice, the signal intensity was observed from 2 h up to 7 days. In the Lenti-CEACAM6 group, the signal intensity was enhanced with the dose increasing but was attenuated with passing time. Therefore, the fluorescence signals were strongest at 2 h with the dose of 10 mg/kg. In the CEACAM6 RNAi group, the signal intensity showed the similar change pattern from 2 h to day 7, but it was more weak when compared with Lenti-CEACAM6 group, the signal was very weak at the day 4 and there was almost non signal was detected at day 7 ( Figure 6A ). The tumor, heart, liver, spleen, lung, kidney, stomach, and muscles from Lenti-CEACAM6 were removed from the mice at day 7. The highest signal intensity was observed in the tumor with the dose of 10 mg/kg CEACAM6-mAb-IRDye800CW probe injection. The signal was also observed in the liver ( Figure 6B ), it was speculated that the tumor has metastasized to the liver. The result was supported by the HE staining that the surface of the liver became spotty, the arrangement of hepatic lobules was disordered, there were metastases in the tissue (white box) and a large number of cancer cells infiltrated (white arrows) ( Figure 6C ).

From the data described above, it was shown that the anti-CEACAM6 probe has high affinity with GC cells and can track the tumor as well as the metastases in vivo in the mice model. Therefore, we performed the analysis of gastric specimens with dysplasia after ESD operations. The hand-held fluorescent detector was used to check the fluorescent signals. The CEACAM6-mAb-IRDye800CW probe could label the neoplastic lesion exactly ( Figure 7A ), and the mucosa with the strongest fluorescence intensity had the highest CEACAM6 expression when compared with adjacent mucosa ( Figure 7B ). The fluorescence intensity was increasing with the pathological changes progression of the mucosa, in another words, the mucosa with the cancer had the strongest fluorescence intensity ( Figure 7C ). Furthermore, it was showed that the fluorescence intensity closely related with the CEACAM6 staining H-Score (R² = 0.6944) ( Figure 7D ). Meanwhile, the surface micro structure of the gastric mucosa was detected using the hand-held detector of the fluorescent microendoscopy. It was investigated the normal mucosa had no signal, the mucosa with mild dysplasia had weak signals that looked similar to regular circles, the mucosa with moderate to severe dysplasia had stronger signals than that with mild dysplasia, with the signals having an appearance similar to irregular circles. The mucosa with the adenocarcinoma had the strongest signal, and the signals characterized as messy ( Figure 7E ).