Biomimetic cell-cell adhesion capillary electrophoresis for studying Gu-4 antagonistic interaction between cell membrane receptor and ligands

Talanta 207 (2020) 120259

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Biomimetic cell-cell adhesion capillary electrophoresis for studying Gu-4 antagonistic interaction between cell membrane receptor and ligands

T

Yiran Zhao1, Linghan Jia1, Chunsu Liang, Cong Li, Meina Li, Yanmeng Liu, ⁎⁎ ⁎ Nurmuhammat Kehrimen, Qing Li , Zhongjun Li, Xiaomei Ling The State Key Laboratory of Natural and Biomimetic Drugs and School of Pharmaceutical Sciences, Peking University, Beijing, 100191, China

ARTICLE INFO

ABSTRACT

Keywords: Biomimetic cell-cell adhesion capillary electrophoresis Interaction between cell membrane receptor and ligands Mac-1 ICAM-1 Screening drugs Lactose derivatives

We report a new method: biomimetic cell-cell adhesion capillary electrophoresis (BCCACE) to screen drugs targeting interactions between cell membrane receptors and ligands under an environment close to physiological conditions, in which the cell membrane receptors/ligands can maintain their natural conformations and bioactivity without being isolated and purified. Firstly, we screened twenty-one lactose derivatives by cellimmobilized capillary electrophoresis and obtained Gu-4 with the best activity (K = 3.58 ± 0.22 × 104) targeting macrophage antigen-1 (Mac-1). Then, BCCACE was performed as follows: HEK 293 cells overexpressed with receptor (intercellular adhesion molecules-1, ICAM-1) were cultured and immobilized on the inner wall of capillaries as stationary phase, which simulated the endothelial cells lining on the inner surface of blood vessels. HEK 293 cells overexpressed with ligand Mac-1 as samples were used to simulate the neutrophils cells in blood vessels. And Gu-4 added into the running buffer solution as the antagonist was used to simulate the drug in blood. The results showed that Gu-4 (40 μM) could selectively inhibit cell-cell adhesion by targeting the interaction between Mac-1 and ICAM-1. Finally, the pharmaceutical efficacy assays of Gu-4 at cellular and animal levels were carried out using the concentration of 40 μM and the dose of 20 mg kg−1 respectively, which showed the anti-cancer metastasis activity of Gu-4 and the validity of the method. This method simulated a complete three-dimensional vascular model, which can easily obtain the suitable blood concentration of drugs. This system simulated the interaction between leukocytes and vascular endothelial cells in the bloodstream antagonized by drugs, and obtained the effective concentration of the antagonist. It can be used as an accuracy and efficient drug screening method and will be expected to become a new method to screen drugs targeting cell-cell adhesion.

1. Introduction Leukocyte migration includes specific adhesive interactions between leukocytes cells and endothelial cells that guide the leukocytes to migrate from the vascular compartment to the extra-vascular tissue. Tethering, rolling, adhesion, and crawling on the apical surface all include heterophilic interactions between one class of molecules on the leukocyte and another class of molecules on the endothelial cells [1,2]. Among the receptors, macrophage antigen-1 (Mac-1) is the most promiscuous one with an α subunit (CD11b) and a β subunit (CD18), and it is an endogenous ligand of intercellular adhesion molecules-1 (ICAM-1) on endothelial cell membrane. Therefore it can form strong interactions with ICAM-1 [3]. Integrins and their ligands are central to the etiology

and pathology of many disease states, functioning in pathological processes such as inflammation, wound healing, angiogenesis, osteoporosis, and tumor metastasis [4,5]. One of lactose derivatives, N-[2(1,3-dilactosyl)-propanyl]-2-amino- pentandia-mide (Gu-4) has been reported to have good binding affinities, excellent chemotactic inhibitory activity, and selectivity towards Mac-1 as a novel antagonist [6]. With the development of researches on the structure and function of cell membrane proteins, more and more cell membrane molecules will become potential targets for the discovery of new drugs [7–9]. For now, most studies on the interactions between hydrophobic receptors and ligands are limited to functional assays and radio ligand-binding assays (RLBA), which have more measurement errors, low accuracy, and low

Corresponding author. Corresponding author. E-mail addresses: [email protected] (Q. Li), [email protected] (X. Ling). 1 Yiran Zhao and Linghan Jia contributed equally to this work and are co-first authors. ⁎

⁎⁎

https://doi.org/10.1016/j.talanta.2019.120259 Received 31 March 2019; Received in revised form 9 August 2019; Accepted 14 August 2019 Available online 16 August 2019 0039-9140/ © 2019 Published by Elsevier B.V.

Talanta 207 (2020) 120259

Y. Zhao, et al.

Fig. 1. Structures of eleven lactose derivatives including O-glycosides (An-2, AcAn-2, TsAn-2, Gu-4, BnGu-4, TsGu-4, NS-7a, NS-19, NS-14b, NS-12a, NS-12b) and ten lactose derivatives including C-glycosides (CL2bF, CL3aF, CL3bF, CL4aF, CL4bF, CL4GF, CL4uF, CLGANF, CLSebF, CLSeF).

sensitivity in quantitative analysis. RLBA is also hazardous and costly [10]. As far as we know, no references have been reported about online studying drug antagonistic interaction between cell membrane receptor and ligand by biomimetic cell-cell adhesion method. Recently, cellular membrane affinity chromatography (CMAC) has been developed to identify the interaction between membrane receptors and ligands [11]. But it is also costly and time-consuming because of a complex preparation procedure, and is operated under unphysiological conditions [12]. According to the attractive features of capillary electrophoresis, we established an online new method: biomimetic cell-cell adhesion capillary electrophoresis (BCCACE) for studying activity of Gu-4 targeting interaction between MAC-1 and ICAM-1 under approximately physiological conditions. Firstly, we set up the immobilized cell capillary electrophoresis (ICCE) [13,14], in which we cultured Mac-1-overexpressed HEK 293 cells on the inner wall of capillaries to form a cell layer as the stationary phase to screen twenty-one lactose derivatives (Fig. 1) targeting to the receptor (Mac-1) and calculated the binding constants, and found Gu-4 with the best targeting activity (K = 3.58 ± 0.22 × 104 M−1) by targeting α subunit (CD11b) of Mac1. And then, ICAM-1-overexpressed HEK 293 cells were cultured and immobilized on the inner wall of capillaries as stationary phase, which simulated the endothelial cells lining on the inner surface of blood vessels. Mac-1-overexpressed HEK 293 cells as samples were used to simulate the neutrophils cells in blood vessels. And Gu-4 added into the running buffer solution as the antagonist was used to simulate the drug in blood. Herein, the antagonism of Gu-4 to the interaction between leukocytes cells and endothelial cells was investigated by BCCACE, and the results showed that Gu-4 (40 μM) could selectively inhibit cell-cell

adhesion by targeting the interaction between Mac-1 and ICAM-1. Furthermore, the results of cell and animal assays showed that Gu-4 had high antitumor metastasis activity and low toxicity. Thereby, the feasibility and availability of BCCACE were verified. This new method has several significant advantages: (1) It can be used to investigate the interaction between cell membrane receptor and ligand, and the antagonism of drugs to cell-cell interactions under approximately physiological environment. (2) The cell membrane molecules are not needed to be isolated and purified, and their native conformations could be kept. (3) The cell-coated capillary columns can resist rinse pressure and the cells did not fall off during performing ICCE and BCCACE procedure, with efficiency in about 30 days. (4) The obtained peak profiles are preferable and consistent with nonlinear chromatography (NLC) theory, and the kinetic parameters (K, ka, kd, and k’) of the interactions can be obtained using NLC technology [15–17]. From the above, BCCACE can effectively simulate the antagonism of drugs to interactions between leukocytes cells and endothelial cells in vivo, and obtain the dynamic parameters and plasma concentrations of drug antagonism under approximately physiological conditions. It can be used as an accuracy and efficient drug screening method and will be expected to become a new method to screen drugs targeting cell-cell adhesion. 2. Materials and methods 2.1. Chemicals and reagents Poly-L-Lysine (PLL, molecular weight 15000–30000) and dimethylsulfoxide (DMSO) were purchased from Sigma Chemical (St. 2

Talanta 207 (2020) 120259

Y. Zhao, et al.

Louis, MO). PLL solution was prepared with sterile water and kept at 4 °C. Eleven lactose derivatives including O-glycosides (An-2, AcAn-2, TsAn-2, Gu-4, BnGu-4, TsGu-4, NS-7a, NS-19, NS-14b, NS-12a, NS-12b) and ten lactose derivatives including C-glycosides (CL2bF, CL3aF, CL3bF, CL4aF, CL4bF, CL4GF, CL4uF, CLGANF, CLSebF, CLSeF) were synthesized in Professor Qing Li's laboratory (Peking University) (Fig. 1). All other chemicals were of analytical grade. 40 mM phosphate buffer (PB, pH 7.40), served as running buffer, was prepared with deionized water derived from a Millipore Milli Q-Plus system (Millipore, Bedford, MA). Samples were diluted with running buffer to the same molar concentration. All CE solutions were filtered through 0.45 μm membranes (Agilent) before use. RPMI 1640 medium and fetal bovine serum (FBS) were purchased from Life Technologies, Inc.

and ICAM-1 were overexpressed on HEK 293 cells’ membranes by electrotransfection, respectively. The cells have been transfected to achieve the desired expression by confocal microscopy images [13,14]. The four kinds of the overexpressing HEK 293 cells were injected into the pretreated capillaries, and then incubated for 12 h at 37 °C in 5% CO2, respectively. To cluster the integrin from cytoplasm towards the cell membrane, the HEK 293 cells overexpressed with CD11b, CD18, and Mac-1 in the capillaries were stimulated with PMA (1 μg/mL) for 2 h. In order to immobilize the cells adhering to the inner wall of capillaries firmly, the capillaries were rinsed with PBS buffer, fixed with 4% paraformaldehyde, and incubated for 20 min. Then the capillaries were rinsed with PBS buffer to remove paraformaldehyde. These cellcoated capillary columns were kept at 4 °C before use.

2.2. Cell culture, electrotransfection and confocal microscopy images

2.4. Instrumentation

Mac-1 is consisted of CD11b (α chain) and CD18 (β chain), and is an endogenous ligand of ICAM-1. To investigate whether Gu-4 inhibits cell-cell adhesion by targeting the interaction between Mac-1 and ICAM-1, CD11b, CD18, Mac-1 (CD11b and CD18), and ICAM-1 were overexpressed on the HEK 293 cells by electrotransfection, respectively. HEK 293 cells were cultured in RPMI 1640 medium supplemented with 10% FBS, penicillin (100 Unit/mL), and streptomycin (100 μg/mL). Each prepared 400 μL HEK 293 cells (4 × 106) were transiently transfected by electroporation with 20 μg pEGFP-CD11b, 20 μg pEYFP-CD18, 10 μg pEGFP-CD11b and 10 μg pEYFP-CD18, or 20 μg pEGFP-ICAM-1at 123 V for 20 ms by an electric pulse generator (Electric Square Porator ECM830, BTX, San Diego, CA), respectively. The HEK 293 cells transfected with pEGFP-CD11b, pEYFP-CD18, pEGFP-Mac-1, and pEGFPICAM-1 were cultured in RPMI 1640 medium with 10% FBS for 24 h, and then were suspended in RPMI 1640 medium with 20% FBS at 1.3 × 107 cells/mL, respectively. It should be noted that, to cluster the integrin from cytoplasm towards the cell membrane, the HEK 293 cells overexpressed with CD11b, CD18, and Mac-1 were stimulated with 1 μg/mL phorbol-12-myristate-13-acetate (PMA) for 2 h [18]. DAPI (4′,6-diamidino-2-phenylindole) was used to stain immobilized cell nucleus: 100 ng/mL DAPI solution was injected into the capillary coated with the overexpressed-receptor cells and stained the cells for 10 min, then DAPI solution was replaced by the mobile phase, and finally the remaining mobile phase was removed. The confocal microscopy images of HEK 293 cells overexpressed with CD11b, CD18, Mac-1, ICAM-1, and the cell-coated capillary columns were completed at room temperature and each field of cells was selected at random using a Leica TCS-NT confocal fluorescence microscope with a 40 × oil immersion lens (Wetzler, Heidelberg). Fluorescence was monitored at 358 nm (excitation wavelength), 461 nm (emission wavelength); 488 nm (excitation wavelength), 518 nm (emission wavelength); and 514 nm (excitation wavelength), 525 nm (emission wavelength). The cell-coated capillary columns were cut into 5 cm pieces. In order to observe the confocal images the polyimide coatings on the surface of the capillaries were removed by scalpel. In the confocal microscopy images, EYFP was a yellow label, while EGFP was a green label and DAPI was a blue label, corresponding to CD11b, CD18, and the nucleus, respectively. The transfection rate was calculated by comparing the number of nuclei with the number of membranes in an imaging field.

CE experiments were performed using Beckman P/ACE MDQ system (Beckman Coulter, Brea, CA) equipped with a photodiode array detector as well as 32 Karat software (version 5.0, Beckman). The bare fused silica capillaries (365 μm O.D.; 200 μm I.D.) were purchased from Yongnian Optical Fibers (Hebei, China) and cell-coated capillaries (365 μm O.D., 200 μm I.D., and a total length of 30.2 cm) were homemade. Fistly, CE experiments of lactose derivatives: the twenty-one lactose derivatives were injected into the bare fused silica capillaries using the pressure injection mode at 0.5 psi for 5 s, respectively, and then the experiment was operated at 3.5 kV, cartridge temperature of 25 °C and DAD wavelength of 214 nm. Secondly, CE experiment of interaction between lactose derivatives and cell-membrane receptor: lactose derivatives were injected into the CD11b, CD18, Mac-1 cellcoated capillary columns, respectively, using the pressure injection mode at 0.5 psi for 5 s, and then the experiment was performed at 1.0 kV, cartridge temperature 25 °C, DAD wavelength of 190 or 214 nm. Third, CE experiment of interaction between Mac-1 and ICAM-1: the capillary column was coated by ICAM-1-overexpressing cells, the samples (Mac-1-overexpressing HEK 293 cells) were injected into the ICAM-1 cell-coated capillary columns using the pressure injection mode at 0.5 psi for 5 s, the Gu-4 was added into the buffer solution, and then the experiment was operated at 1.0 kV, cartridge temperature 25 °C, DAD wavelength of 190 or 214 nm. In addition, 1 pound per square inch (psi) = 6894.76 Pa, the uncoated and cell-coated capillary columns were washed between runs with 40 mM PB (pH 7.40) at 0.5 psi for 3 min, the capillaries coated by cells without over-expressed receptors were used as the control group and each kind of sample was run in duplicate for all the experiments. 2.5. Determination of kinetic parameters In order to determinate the kinetic parameters, several concentrations of Gu-4 (20, 50, 100, 200, and 500 μM) and suitable concentrations of other lactose derivatives were prepared with 40 mM PB (pH 7.40) solution for ICCE analysis. The asymmetries of the observed peaks were analyzed using NLC model derived from impulse input solution as described previously [13]. PeakFit software (version 4.11 for Windows, SPSS, Chicago, IL) was used to calculate the kinetic parameters. The obtained electropherogram profiles will be fitted to the NLC function by adjusting the parameters (a0, a1, a2, and a3; x: reduced retention time) of the nonlinear equation (Eq. (1)).

2.3. Preparation of the cell-coated capillary columns The pretreatment of capillary columns was needed for HPCE experiments. The capillary columns (Yongnian Optical Fibre Corp.) were activated by rinsing with methanol, 0.1 M NaOH solution, and deionized water in sequence. After immerged into medicinal alcohol overnight for sterilization, the capillaries were purged with air in biological safely cabinet (HF safe-1200LC, Shanghai Lishen Scientific Equipment Co.Ltd., Shanghai), charged with 0.1 mM PLL solution and incubated at 37 °C for 2 h to create a positive charged coating. CD11b, CD18, Mac-1,

y=

a0 1 a3

exp

a3 a2

a1

x

1

T

(

I1

(

2 a1 x a2

a1 x , a2 aa2

)

) exp ( 1

x a1 a2

exp

)

( ) a3 a2

(1)

Using the fitted parameters, we can calculate the kd, ka, and K by the equations: kd = 1/a2/t0; ka = Kkd; K = a3/C0. C0 is the concentration of the injected solute multiplied by the width of the injection pulse [15–17]. 3

Talanta 207 (2020) 120259

Y. Zhao, et al.

2.6. Assays of cell adhesion, wound healing and animal anticancer metastasis

stimulate the transfected cells. If PMA was not used, CD11b-GFP remained in the cytoplasm as opposed to moving toward the cell membrane [18]. We observed the state of receptors by confocal microscopy, and found that the receptors clustered on the cell membrane after 2 h of stimulation, and were endocytosed after 4 h of stimulation (Fig. 2- [1], Fig. 2- [2]), so 2 h was used for future experiments. To obtain HEK 293 cells with overexpressed Mac-1, pEGFP-CD11b and pEYFP-CD18 were co-transfected into the cells and verified by fluorescent confocal microscopy at excitation wavelengths of 488 nm and 514 nm, respectively (Fig. 2- [3]). HEK 293 cells overexpressed with pEGFP-ICAM-1 were observed under confocal microscopy using 488 nm excitation (Fig. 2- [5]). These results showed that the expression rate of electrotransfection was > 70% and could meet the requirement of the further experiments.

In order to confirm the bioactivity of Gu-4, the cell adhesion assay was carried out. Human umbilical vein endothelial cells (HUVECs) were seeded in 96-well plates (1 × 105 cells per well) and cultured to confluence. Human acute monocytic leukemia cell line (THP-1) cells were maintained in RPMI 1640 medium supplemented with 15% FBS, and co-incubated with Gu-4 for 30 min at 37 °C. Then, the treated THP1 cells were added onto the monolayer of HUVECs in serum free medium and incubated for 1 h at 37 °C. The non-adherent cells were removed by wash buffer and the adherent cells were quantified by fluorescence measurement with a fluorescence plate reader at 485/ 520 nm. The fluorescence intensity was used to represent the number of adherent cells. Cell wound healing assay was applied to measure the migration activity of cells in order to confirm the bioactivity of Gu-4. Cancer cells (NCI–H520) were seeded in six-well plates and incubated with RPMI 1640 medium containing 10% FBS until they grew to 90% confluence. Then we changed the medium to serum free RPMI 1640 medium and incubated the cells for 24 h to starve them. Wounding was performed using sterile plastic pipette tips. The detached cells were washed with PB solution. The ratio of the remaining wound area was calculated relative to the initial wound area and normalized to control group. In vivo lung cancer metastasis experiment was using murine B16 melanomas in C57BL/6J mice. Murine B16 melanomas cells were cultured in RPMI 1640 medium supplemented with 10% FBS, penicillin (100 Unit/mL), and streptomycin (100 μg/mL) at 37 °C. C57BL/6J mice (male, 8 weeks old, 20.0 ± 2.0 g) were purchased from the Lab Animal Institute of Peking University Health Center. The experimental protocol was approved by the University Ethics Committee for the use of experimental animals and conformed to the Guide for Care and Use of Laboratory Animals. Before initiation of the experimental procedure, the animals were acclimated for at least one week in a standardized temperature (25–28 °C), humidity (50–60%), and light environment (12 h light/12 h dark) with free access to standard food and tap water. The metastasis experiment was carried out by a method described previously [4]. B16 cells were harvested, washed, and resuspended in PBS. The cell suspension (2.5 × 105 cells/200 μL) was injected into mice via tail vein. The mice were randomly divided into five groups (n = 6 for each group). The control group received intraperitoneal injection of saline and the four treatment groups received intraperitoneal injection of cyclophosphamide (10 mg/kg/d), Gu-4 (5, 10, or 20 mg/ kg/d), respectively.

3.2. Preparetion of the cells-coated capillary columns To firmly immobilize CD11b/CD18/Mac-1/ICAM-1 overexpressing cells on the inner wall of capillary columns, respectively, the surface of capillary columns needs to be modified with PLL via electrostatic interaction. The negatively charged molecules on cell surface might be absorbed to positively charged PLL. And the influence of different PLL concentrations (0.05, 0.10, 0.20, 0.50, or 1.00 mg mL−1) on cells adhesion on the inner wall of capillary was investigated. The results indicated that the effect of PLL concentrations on the cell adhesion plateaued when the concentration was larger than 0.1 mg mL−1, so that 0.1 mg mL−1 PLL was applied to this study (Fig. S1). According to the previous research [13], 1.3 × 107 cells mL−1 of transfected cells were injected into the pretreated capillaries, and then incubated in RPMI 1640 medium containing 20% FBS under 5% CO2 at 37 °C for 12 h. And then, the cells on the inner wall of the capillary columns were immobilized with 4% paraformaldehyde. The results showed that the cells displayed strong adhesion on the inner wall of capillary columns (Fig. 2- [4], Fig. 2- [6]). Herein, the overexpressed receptors/ligands could keep natural conformation and activity, and be regarded as “effective targets”. Thus, the binding activity could be achieved as long as the targeting receptors/ligands were overexpressed on the cells [19,20]. 3.3. Investigating the interactions and screening lactose derivatives by BCCACE To better investigate the interactions between Mac-1 and lactose derivates, the buffer solution concentration and the operation voltage need to be optimized in electrophoretic tests under the environments close to physiological conditions. Buffer solution of the 40 mM PB (pH 7.40) should be selected as BGE for the reason that the overlarge buffer solution concentration would result in running current overload. The voltage was set at 3.5 kV and the 20-cm effective-length of capillary column was chosen to provide suitable time for the interaction between ligands and receptors. The baseline and peak shape of the

3. Results and discussion 3.1. Cell culture, electrotransfection and confocal microscopy images The experimental results (Fig. 2) showed that at 36 h after transfection of HEK 293 cells to overexpress CD11b, PMA was required to

Fig. 2. [1] Fluorography of CD11bEGFP on HEK 293 membranes induced by PMA at different time points [2]. Fluorography of CD11b-EGFP on HEK 293 membranes induced by PMA 2h [3]. Fluorography of HEK 293 cells cotransfected with both plasmid DNA of pEGFP-CD11b and pEYFP-CD18 [4]. Confocal microscopy images of HEK 293 cells transfected with Mac-1-FP on the inner wall of a capillary: indicating the cell layer covering the capillary wall. A, B, C, and D are a local amplification of the cell layer, among them, (A) overlay of B, C and D indicates whole cells; (B) blue color indicates cell nuclei; (C) green color indicates cell membranes; (D) yellow color indicates cell membranes [5]. Confocal microscopy images of HEK 293 cells transfected with pEGFP-ICAM-1 in the dish [6]. Confocal microscopy images of HEK 293 cells transfected with ICAM-1-GFP on the inner wall of a capillary. Among them, (A) blue color indicates cell nucleus, (B) green color indicates cell membranes, (C) overlay of A and B indicates whole cells. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) 4

Talanta 207 (2020) 120259

Y. Zhao, et al.

Fig. 3. Electropherograms of (A) lactose, (B) DMSO, and (C) Gu-4 on the different kinds of capillaries (from top to bottom, the capillaries were uncoated, and coated with HEK293 cells, overexpessed-CD18 HEK293 cells, overexpessedMac-1 HEK293 cells, overexpessed-CD11b HEK293 cells, respectively, in A, B and C). (D) Electropherograms of the different concentratios of Gu-4 (20, 50, 100, 200, or 500 μM) on the capillary coated with CD11b-overexpressing HEK 293 cells. (E) The effect curve of Gu-4 concentration on the association constant. Beckman P/ACE MDQ CE system with DAD, 214 nm. Injection: 0.5 psi for 5 s; applied voltage: 3.5 kV; capillary: capillary of 30.2 cm (effective length 20.0 cm), 200 μm i.d.; running buffer: 40 mM PB (pH 7.40).

electropherograms obtained under the optimized conditions were good enough for the qualitative and quantitative analysis. Gu-4 (antagonist of Mac-1), lactose, and DMSO (negative control) were employed to evaluate our method. The electrophoretic behaviors of these compounds were examined respectively in different kinds of capillaries: uncoated; coated with HEK 293 cells; coated with CD11boverexpressing HEK 293 cells, coated with CD18-overexpressing HEK 293 cells; and coated with Mac-1-overexpressing HEK 293 cells. The peak heights of lactose (Fig. 3A) and DMSO (Fig. 3B) in different kinds of capillary columns had no significant differences, the peak widths slightly broadened in the capillaries coated with the cells compared with those in the uncoated ones. The phenomenon might result from the reduction of electroosmotic flow (EOF) in the cell-coated capillaries. Therefore, the electrophoretic peak shape could be applied to estimate the interactions between lactose derivates and stationary phases. When the dissociation rate of a ligand from the stationary phase is less than the adsorption rate, non-Gaussian peaks will appear [13,14]. Additionally, the peak shape of Gu-4 in the capillary column coated with Mac-1-overexpressing HEK 293 cells was a decreased and broadened non-Gaussian peak with significant tailing (Fig. 3C), indicating that strong interactions between the ligand (Gu-4) and the stationary phase (Mac-1) occurred during the ICCE process. We also found that the peak shape of Gu-4 in the capillary column coated with CD11b-overexpressing HEK 293 cells decreased and broadened distinctly while the peak shape of Gu-4 in the capillary column coated with CD18-overexpressing HEK 293 cells didn't change obviously, indicating that the CD11b should be the target receptors rather than CD18 (Fig. 3C). As

Mac-1 is consisted of CD11b (α chain) and CD18 (β chain), and K of Gu4 interacting with Mac-1 was smaller than one of Gu-4 interacting with CD11b (Table 1), indicating that CD11b should contain the binding domain [18]. Using this method, we screened twenty-one lactose derivates by targeting CD11b, of them, fifteen ones (NS-7a, NS-19, NS14b, NS-12a, NS-12b, CL2bF, CL3aF, CL3bF, CL4aF, CL4bF, CL4GF, CL4uF, CLGANF, CLSebF, CLSeF) had similar electrophoretic behaviors as DMSO did, indicating that there were no interactions between them and CD11b; the other five ones [An-2, AcAn-2 TsAn-2, BnGu-4, TsGu-4 Table 1 Kinetic parameters for the interactions of Gu-4, An-2, Bn Gu-4, Ts Gu-4, Ac An2, Ts An-2 with the immobilized CD11b-over-expressing cells and Mac-1-overexpressing cells, respectively.

5

Interaction

K ( × 104M−1)

k′

ka ( × 103M−1s−1)

kd ( × 10−1s−1)

Mac-1/Gu-4 Mac-1/An-2 Mac-1/Bn Gu-4 Mac-1/Ts Gu-4 Mac-1/Ac An-2 Mac-1/Ts An-2 CD11b/Gu-4 CD11b/An-2 CD11b/Bn Gu-4 CD11 b/Ts Gu-4 CD11b/Ac An-2 CD11 b/Ts An-2

3.58 1.40 0.59 1.05 × 10−10 0.14 0.02 4.28 1.99 0.75 0.27 0.21 0.11

1.59 0.46 2.45 4.59 0.67 3.90 1.66 0.52 0.56 2.22 0.70 0.84

2.83 0.55 0.25 6.64 × 10−10 0.10 0.07 2.32 0.76 0.63 0.13 0.12 0.06

0.79 0.39 0.42 6.35 0.70 3.39 0.54 0.38 0.86 0.49 0.54 0.57

Talanta 207 (2020) 120259

Y. Zhao, et al.

possessed similar electrophoretic behaviors, indicating that there were the interactions between them and CD11b, and Gu-4 had the best bioactivity (Table 1). To determine whether the phenomenon was caused by the strong specific interaction or the nonspecific interactions, we designed the parallel control group (capillary columns coated with HEK 293 cells) to compare the electrophoretic behaviors of Gu-4 (Fig. 3C). The decreased and broadened non-Gaussian peaks observed in the parallel control group capillary columns indicated that Gu-4 might have nonspecific interactions with the membrane molecules except for CD11b. Gu-4 peak of capillary coated with CD11b-overexpressing HEK 293 cells decreased and broadened more obviously due to stronger binding activities compared with that of the capillary coated with HEK 293 cells. The differences in peak shape of Gu-4 in different kinds of cell-coated capillaries reappeared in the repeatability tests. It showed that ICCE is able to identify the specific ligands of Mac-1 or CD11b. The intraday and interday reproducibilities of the capillaries coated with the CD11boverexpressing HEK 293 cells (Table S1) were investigated as well, and the RSD values could meet the requirement of the experiment. The activity of cell-coated capillary columns stored at 4 °C on the 0, 10th, and 30th days was also tested, indicating that the cell-coated capillary columns still had a good screening activity after kept for 30 days. One particular technical difficulty is that cells will fall off easily if the capillary is too dry. Therefore, both ends of the capillary need to be inserted in the buffer to maintain a wet state during storage. The experimental results showed that the Gu-4 concentrations (20, 50, 100, 200, or 500 μM) had an obvious influence on the determination of binding constant (Fig. 3D). The smaller the Gu-4 concentration was, the greater the binding constant determined would be (Fig. 3E). When the compound concentration was low, the binding-dissociation process was dominated by strong binding sites (like Mac-1) with a small proportion, which resulted in the higher values of K and ka. On the contrary, when the column was overloaded with high concentration of the compound, the strong binding sites with a small proportion would become saturated whereas the weak binding sites with a large proportion would become a dominating factor, resulting in the lower values of K and ka. The decrease of k′ (k′ = a1) and kd could also be regarded as a result of overloading effects. The different classes of binding sites refer to Mac-1 and other molecules on cell membrane [18]. Therefore, when we determinate the ligand-receptor binding constant, the lower the ligand concentration is, more close to real value the determinated value of binding constant is. To investigate the interaction between Mac-1-overexpressing HEK 293 cells and ICAM-1-overexpressing HEK 293 cells, we optimized the electrophoretic conditions and chose 190 nm and 214 nm as detection wavelengths for the method's sensitivity, and found that the electropherograms were more sensitive and clear at 190 nm (Fig. 4A). We also tested different voltages and found that the electrophoretic peaks were observed extremely unsmooth at 1.5 kV and 1.2 kV, and the reproducibility could not meet the requirement. As a result, we chose 1.0 kV as the separation voltage, at which the sensitivity and reproducibility were both excellent. The density of the injected cell solution (5.0 × 104, 1.0 × 105, 5.0 × 105, 8.0 × 105, 1.0 × 106, 2.5 × 106, 5.0 × 106, 7.5 × 106, or 1.0 × 107 mL−1) were tested as well and the results showed that 5.0 × 106 cells mL−1 was the optimal density (Fig. 4A and B). Gu-4 (antagonist of Mac-1) and lactose (negative control) were used to identify whether the differences in the electropherograms were caused by the strong specific interaction or the nonspecific interactions. For the cell coated capillary electrophoresis, it is easy to run the molecules like protein, compounds, etc. However, when the cells are injected as samples, due to the diameter of HEK 293 cells is about 15−20 μm according to the confocal microscopy images, the capillary column with 200 μm internal diameter is more suitable to invetigate the interaction between coated cells and injected cells. Therefore, we designed series of control experiments to demonstrate the electrophoresis separation is specific and valid. For instance, the

Fig. 4. Electropherograms of different densities of the HEK293 cells in the cellcoated capillaries at (A) 190 nm and (B) 214 nm (From top to bottom, the density of the injected cell solution were 1.0 × 107, 7.5 × 106, 5.0 × 106, 2.5 × 106, 1.0 × 106, 8.0 × 105, 5.0 × 105, 1.0 × 105, and 5.0 × 104 mL−1, respectively, in A and B). (C) Electropherogram of MAC-1-overexpressing HEK293 cells on capillaries coated with ICAM-1-overexpressing HEK 293 cells under the different concentrations of Gu-4 and lactose (From top to bottom, for the top line, the samples was HEK293 cells, and for the other five lines, the samples were overexpessed-Mac-1 HEK293 cells. and tested in the running buffer with the addition of 0, 0.04, 0.4, 4, and 40 μM Gu-4, respectively). Beckman P/ACE MDQ CE system with DAD, 190/214 nm. Injection: 0.5 psi for 5 s; applied voltage: 1.0 kV; capillary: capillary of 30.2 cm (effective length 10.2 cm) 200 μm i.d.; running buffer: 40 mM PB (pH 7.40).

parallel control groups to analyze the electropherogram of HEK 293 cells in different types of capillaries were set up. It was observed that the peak of Mac-1-overexpressing HEK 293 cells decreased and broadened distinctly compared with parallel control group in BCCACE (Fig. 4C), indicating that ICAM-1-overexpressing HEK 293 cells should have interactions with Mac-1-overexpressing HEK 293 cells. In addition, we investigated the bioactivity of Gu-4 as a blocker by adding the different concentrations (0.04, 4.0, or 40 μM) of Gu-4 into the buffer solution. The peak shape became sharper as the concentration of Gu-4 increased, indicating that Gu-4 as an antagonist might inhibit the interaction between Mac-1 and ICAM-1. With the decrease of the Gu-4 concentration, the peak shape could not become sharper efficiently, indicating that Gu-4 might have satisfactory antiadhesion ability at 40 μM. However, when adding 40 μM lactose into the buffer solution, the peak shape could not become sharper, indicating that lactose might have no antiadhesion ability. Therefore, this method can not only simulate in vivo cell-cell interaction system very well in capillary columns, but also identify if the drug in mobile phase has the specific antagonism capability of the cell-cell interaction. Meanwhile, the experimental results could also provide the scientific direction for further pharmaceutical efficacy experiments. BCCACE can be used to investigate cell-cell interaction and small molecule antagonizing cell-cell interaction under the condition close to physiological environment, and also can be easily transferred to a capillary array electrophoresis for high-throughput screening targeted drugs. 6

Talanta 207 (2020) 120259

Y. Zhao, et al.

Fig. 5. (A) Results of cell adhesion assay. (a) HUVEC cells; (b) THP-1 cells; (c) The Leuko Tracker™ labeled THP-1 cells attached to HUVEC cells at 480 nm/520 nm. (B) Results of wound healing assay: statistical significance is indicated as * p < 0.05, **, p < 0.01; all data are shown as mean ± s.e.m. (C) Results from B16–F10 melanoma lung colonization assay: at 16 days after injection of B16 cells, metastatic tumors were counted, and each value represents the mean ± S.D. of triplicate measurements (** represents p < 0.01).

3.4. Assays of cell adhesion, wound healing and animal anticancer metastasis

4. Conclusions Here we established a new method, BCCACE, in which, ICAM-1overexpressing HEK 293 cells were immobilized on the inner wall of capillaries as stationary phase, simulating the endothelial cells lining on the inner surface of blood vessels; Mac-1-overexpressing HEK 293 cells as samples were used to simulate the neutrophils cells in blood vessels; and Gu-4 added into the mobile phase as the antagonist was used to simulate the drug in blood. By this method, we can rapidly screen the potential active antagonists or agonists targeting cell-cell adhesion under an environment close to physiological conditions without the separation and purification process of membrane proteins. The obtained peak profiles are consistent with those fitted by NLC theory and kinetic parameters (K, ka, kd, and k’) were acquired. The cell-coated capillary columns can resist rinse pressure and the cells did not fall off during performing BCCACE procedure, and still have efficiency in about 30 days. Feasibility and availability of BCCACE were verified by cellular and animal assays of Gu-4 anticancer metastasis. Therefore, BCCACE is a new vascular simulation system, which simulates the interaction between leukocytes and drug-antagonistic vascular endothelial cells. Using this system, the effective concentration of antagonists can be easily obtained. It can be used as an accuracy and efficient drug screening method and will be expected to become a new method to screen drugs targeting cell-cell adhesion.

In order to verify the veracity of BCCACE, we performed pharmaceutical efficacy assays in vitro and in vivo. The ability of Gu-4 inhibiting the adhesion between HUVECs and THP-1 cells based on an inflammatory model were carried out. THP-1 cells were labeled with Leuko Tracker so that green fluorescence could be observed under an inverted fluorescence microscope (Fig. 5A). The different concentrations of Gu-4 (0, 0.04, 0.4, 4, or 40 μM) were tested and found that the relative fluorescence unit (RFU) of the labeled cells decreased significantly when adding 40 μM Gu-4, and indicating that Gu-4 as an antagonist could block leukocyte-endothelium adhesion effectively [7], which was consistent with the results of BCCACE. The results of this experiment further declared that the method of BCCACE was feasible and reliable. In order to further investigate the bioactivity of Gu-4, woundhealing assay was applied to measure the migration activity of cancer cells by evaluating wound closure at different concentrations of Gu-4 (0.04, 0.4, 4, or 40 μM) (Fig. 5B). Compared with the control group, Gu4 groups slowed down in vitro cell migration in concentration-dependent manner. And the concentration of 40 μM Gu-4 showed the best inhibitory effect among all the concentrations we have studied. The results not only demonstrated Gu-4 might play a key role during anticancer metastasis process, but also declared the feasibility and reliability of the new established method. The effect of Gu-4 (low-dose, middle-dose, and high-dose) on anticancer metastasis was investigated using a murine B16–F10 melanoma lung colonization assay. Cyclophosphamide (CTX) was chosen as the positive control. Fig. 5C showed the images and average nodule counts of mice with pulmonary metastasis of melanoma in negative control (N·S.), positive control (CTX), Gu-4 (5, 10 or 20 mg kg−1) groups. The results indicated that the tested three concentrations of Gu4 could significantly reduce the number of metastasis nodules on lung surface. Among them, 20 mg kg−1 Gu-4 showed the best anticancer metastatic effect and even as good as CTX did. On the 16th day of the experiment, the death rate of CTX group was 13.3%, while the death rates of low-dose, middle-dose, and high-dose Gu-4 groups were 0, which indicated that Gu-4 could inhibit cancer metastasis very well with low toxicity. The results of cell adhesion and wound healing assay indicated that Gu-4 could be a potential inhibitor of both inflammation and tumor metastasis in vitro. The murine melanoma lung colonization assay further verified the anticancer metastasis activity of Gu-4 in vivo and the validity of the method. According to the results of molecular, cellular, and animal levels, the new method can effectively simulate the antagonism of drugs to antagonise interactions between leukocytes cells and endothelial cells for the first time, and it was screened out that Gu-4 expected to be a novel anticancer metastasis drug with high anti-cancer efficiency and low systemic toxicity targeting.

Conflicts of interest There are no conflicts to declare. Acknowledgements This work was financially supported by the National Natural Science Foundation of China (No. 81673392 and 81373372). We appreciate professor Ying Wang and professor Dalong Ma (Center for Human Disease Genomics, Peking University), professor Suodi Zhai and associate professor Xianhua Zhang (Peking University Third Hospital), for their supports. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.talanta.2019.120259. References [1] S.J. Hopkin, J.W. Lewis, F. Krautter, M. Chimen, H.M. McGettrick, Triggering the resolution of immune mediated inflammatory diseases: can targeting leukocyte migration Be the answer? Front. Pharmacol. (2019), https://doi.org/10.3389/ fphar.2019.00184. [2] S. Nourshargh, F.M. Marelli-Berg, Transmigration through venular walls: a key regulator of leukocyte phenotype and function, Trends Immunol. 26 (2005) 157–165.

7

Talanta 207 (2020) 120259

Y. Zhao, et al. [3] N.P. Podolnikova, A.V. Podolnikov, T.A. Haas, V.K. Lishko, T.P. Ugarova, Ligand recognition specificity of leukocyte integrin αMβ2 (Mac-1, CD11b/CD18) and its functional consequences, Biochemistry 54 (2015) 1408–1420. [4] Byron H. Kwan, Eric F. Zhu, Alice Tzeng, Harun R. Sugito, Ahmed A. Eltahir, Botong Ma, Mary K. Delaney, Patrick A. Murphy, Monique J. Kauke, Alessandro Angelini, Noor Momin, Naveen K. Mehta, Alecia M. Maragh, Richard O. Hynes, Glenn Dranoff, Jennifer R. Cochran, 8 and K. Dane Wittrupcorresponding author, Integrin-targeted cancer immunotherapy elicits protective adaptive immune responses, J. Exp. Med. 214 (2017) 1679–1690. [5] Michael H. Ross, Alison K. Esser, Gregory C. Fox, Anne H. Schmieder, Xiaoxia Yang, Grace Hu, Dipanjan Pan, Xinming Su, Yalin Xu, Deborah V. Novack, Thomas Walsh, Graham A. Colditz, Gabriel H. Lukaszewicz, Elizabeth Cordell, Joshua Novack, James A.J. Fitzpatrick, David L. Waning, Khalid S. Mohammad, Theresa A. Guise, Gregory M. Lanza, Katherine N. Weilbaecher, Bone-induced expression of integrin β3 enables targeted nanotherapy of breast cancer metastases, Cancer Res. 77 (2017) 6299–6312. [6] T. Yan, Q. Li, H. Zhou, Y. Zhao, S. Yu, G. Xu, Z. Yin, Z. Li, Z. Zhao, Gu-4 suppresses affinity and avidity modulation of CD11b and improves the outcome of mice with endotoxemia and sepsis, PLoS One 7 (2012) e30110. [7] R. Aebersold, M. Mann, Mass-spectrometric exploration of proteome structure and function, Nature 537 (2016) 347–355. [8] S. Huang, S.Y. Lim, A. Gupta, N. Bag, T. Wohland, Plasma membrane organization and dynamics is probe and cell line dependent, Biochim. Biophys. Acta Biomembranes. 1859 (2016) 148−1492. [9] R. Aebersold, M. Mann, Mass-spectrometric exploration of proteome structure and function, Nature 537 (2016) 347–355. [10] E.F. Garman, Developments in x-ray crystallographic structure determination of biological macromolecules, Science 343 (2014) 1102–1108. [11] M. Sanghvi, R. Moaddel, C. Frazier, I.W. Wainer, Synthesis and characterization of liquid chromatographic columns containing the immobilized ligand binding domain of the estrogen related receptor alpha and estrogen related receptor gamm, J.

Pharm. Biomed. Anal. 53 (2010) 777–780. [12] R. Moaddel, I.W. Wainer, The preparation and development of cellular membrane affinity chromatography columns, Nat. Protoc. 4 (2009) 197–205. [13] P. Yakufu, H. Qi, M. Li, X. Ling, Y. Wang, CCR4 expressing cells cultured adherently on a capillary wall and formaldehyde fixed as the stationary phase for ligand screening by ACE, Electrophoresis 34 (2013) 531–540. [14] R. Wu, C. Li, X. Sun, S. Zhang, C. Liang, Y. Jiang, X. Ling, Rapid screening of antitumor metastasis drugs targeting integrin macrophage antigen-1 using immobilized cell capillary electrophoresis, Analyst 143 (2018) 4981–4989. [15] J.L. Wade, A.F. Bergold, P.W. Carr, Theoretical description of nonlinear chromatography, with applications to physicochemical measurements in affinity chromatography and implications for preparative-scale separations, Anal. Chem. 59 (1987) 1286–1295. [16] C. Baggiani, P. Baravalle, L. Anfossi, C. Tozzi, Comparison of pyrimethanilimprinted beads and bulk polymer as stationary phase by non-linear chromatography, Anal. Chim. Acta 542 (2005) 125–131. [17] K. Jozwiak, J. Haginaka, R. Moaddel, I.W. Wainer, Displacement and nonlinear chromatographic techniques in the investigation of interaction of noncompetitive inhibitors with an immobilized alpha3beta4 nicotinic acetylcholine receptor liquid chromatographic stationary phase, Anal. Chem. 74 (2002) 4618–4624. [18] H.M. Chen, H.L. Phal, R.J. Scheibe, D.E. Zhang, D.G. Tenen, The Sp1 transcription factor binds the CD11b promoter specifically in myeloid cells in vivo and is essential for myeloid-specific promoter activity, J. Biol. Chem. 268 (1993) 8230–8239. [19] Q. Li, J. Wang, Y.Y. Zheng, L. Yang, Y. Zhang, L. Bian, J. Zheng, Z. Li, X. Zhao, Y. Zhang, Comparison of zonal elution and nonlinear chromatography in determination of the interaction between seven drugs and immobilised β(2)-adrenoceptor, J. Chromatoqr. A 1401 (2015) 75–83. [20] I. Gottschalk, C. Lagerquist, S.S. Zuo, A. Lundqvist, P. Lundahl, Immobilized-biomembrane affinity chromatography for binding studies of membrane proteins, J. Chromatogr. B 768 (2002) 31–40.

8