Ex vivo and in vivo capture and deactivation of circulating tumor cells by dual-antibody-coated nanomaterials Jingjing Xie, Yu Gao, Rongli Zhao, Patrick J. Sinko, Songen Gu, Jichuang Wang, Yuanfang Li, Yusheng Lu, Suhong Yu, Lie Wang, Shuming Chen, Jingwei Shao, Lee Jia PII: DOI: Reference:
S0168-3659(15)00275-8 doi: 10.1016/j.jconrel.2015.04.036 COREL 7657
To appear in:
Journal of Controlled Release
Received date: Revised date: Accepted date:
21 January 2015 8 April 2015 27 April 2015
Please cite this article as: Jingjing Xie, Yu Gao, Rongli Zhao, Patrick J. Sinko, Songen Gu, Jichuang Wang, Yuanfang Li, Yusheng Lu, Suhong Yu, Lie Wang, Shuming Chen, Jingwei Shao, Lee Jia, Ex vivo and in vivo capture and deactivation of circulating tumor cells by dual-antibody-coated nanomaterials, Journal of Controlled Release (2015), doi: 10.1016/j.jconrel.2015.04.036
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Ex vivo and in vivo capture and deactivation of circulating
T
tumor cells by dual-antibody-coated nanomaterials
IP
Jingjing Xie1, Yu Gao1, Rongli Zhao1, Patrick J. Sinko2, Songen Gu1, Jichuang Wang1,
SC R
Yuanfang Li1, Yusheng Lu1, Suhong Yu1, Lie Wang3, Shuming Chen3, Jingwei Shao1, Lee
MA
NU
Jia1, *
1
D
Cancer Metastasis Alert and Prevention Center, and Biopharmaceutical Photocatalysis of
TE
State Key Laboratory of Photocatalysis on Energy and Environment, College of Chemistry,
2
CE P
Fuzhou University, Fuzhou 350002, China.
08854, USA.
AC
Rutgers, The State University of New Jersey, 160 Frelinghuysen Road, Piscataway, NJ,
3
Department of Oncology, East Hospital, Fuzhou, 350025, China.
*Corresponding authors: Lee Jia ([email protected] or [email protected]), 523 Industry Road, Science Building, 3FL., Fuzhou University, Fuzhou, Fujian, China, 350002. Phone and Fax: +86-0591-8357-6912.
1
ACCEPTED MANUSCRIPT
ABSTRACT
MA
NU
SC R
IP
T
Circulating tumor cells (CTCs) have been detected by us and others in cancer patient blood. However, little is known about how to specifically capture and deactivate CTCs in vivo, which may lead to successful metastasis prevention in asymptomatic cancer survivors after surgery. We hypothesize that the dual antibody conjugates may have the advantage of capturing CTCs specifically over their single antibody counterparts. Here we show that the surface-functionalized dendrimers can be sequentially coated with two antibodies directed to surface biomarkers (EpCAM and Slex) of human colorectal CTCs. The dual antibody-coated dendrimers exhibit a significantly enhanced specificity in capturing CTCs in the presence of interfering blood cells, and in both eight-patient bloods and nude mice administered with the labeled CTCs in comparison to their single antibody-coated counterparts. The dual antibody-coated conjugates down-regulate the captured CTCs. This study provides the first conceptual evidence that two antibodies can be biocompatibly conjugated to a nanomaterial to capture and down-regulate CTCs in vivo with the enhanced specificity.
AC
CE P
TE
D
KEYWORDS: Cancer metastasis; Circulating tumor cells; Dual antibody conjugates; Capture efficiency; Capture specificity
2
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
NU
SC R
IP
T
1. Introduction The root cause of cancer metastasis can be traced down to the presence of circulating tumor cells (CTCs) or tumor-specific DNA [1] in the blood [2]. As the estimated number of cancer survivors will reach 18 million in the USA alone by 2022 [3], specifically restraining the residual CTCs in cancer survivors becomes increasingly important [4, 5]. CTCs in cancer survivors often show a low rate of proliferation when cancer survivors are in remission and/or asymptomatic [6]. Thus current post-metastatic chemotherapy strategies that only target highly proliferating cancer and non-cancer cell lines can't be effective in targeting and killing CTCs [7]. As a matter of fact, anticancer chemotherapy may enhance metastasis formation [8], and the current nanotechnology approaches that conjugate nanomaterials with a typical antibody and a representative anticancer drug have not resulted in a successful cancer treatment. Indeed, most tumor-targeting nanotechnology reported to date is developed to utilize the unique property of the enhanced permeability and retention of the post-metastatic tumor tissues for treatment [9, 10]. Other cancer nanotechnology developed is based on the CTCs surface biomarkers for early in vitro cancer metastasis diagnosis, including utilization of a single surface biomarker of CTCs such as the anti-epithelial cell adhesion molecule (aEpCAM) to coat silicon-nanopillars [11, 12], nanoscale poly(amidoamine) (PAMAM) dendrimers [13], or functionalized graphene oxide nanosheets [14] for specific in vitro capturing CTCs. Recent reports revealed the co-existence of various surface biomarkers such as HER-2/EGFR/HPSE/Notch1 [15], EpCAM/CD44/CD47/MET [16] and EpCAM/HER-2/EGFR [17] on the single types of breast CTCs. Various combinations of the corresponding antibodies have thus been developed as an antibody cocktail-coated microchip [18], a microfluidic herringbone-chip [17], or an epoxy-functionalized glass slide with three channels immobilizing an individual PEG-dendrimer-antibody on each channel [19] for in vitro CTCs diagnosis. However, little is known about utilization of bionanotechnology for specifically capturing CTCs and restraining their activity in the bloodstream of the asymptomatic cancer survivors in order to prevent the future cancer metastasis. The stage-2 and -3 colorectal cancer patients often tragically confront 70% and 80% metastatic rate within 3-5 years after surgical removal of the primary cancer [20]. Human colorectal carcinoma HT29 cells often express two CTCs biomarkers, the EpCAM [21] and the saliva acidifying louis oligosaccharides X (Sialyl Lewis X, Slex) [22, 23]. However, EpCAM down-regulation was found during the epithelial-to-mesenchymal transition of CTCs. Thus, only utilization of antibody against EpCAM may not capture the EpCAM-negative CTCs. In addition, the abundance of a single surface biomarker may vary chronometrically with the cell cycle phases, resulting in variable binding affinity between the biomarker and its corresponding ligand [24]. Moreover, one CTCs surface biomarker (e.g., CD44) may also express on other normal cell lines leading to a false biological recognition [16]. Based on the above analysis, we hypothesized that conjugating the functionalized PAMAM dendrimers with two clinically-used surface biomarker antibodies, instead of a single antibody, may significantly enhance the sensitivity and specificity of the antibody-coated conjugates in identifying and capturing the CTCs, and consequently result in down-regulation of the activity of the captured CTCs. We therefore developed the surface-functionalized PAMAM dendrimers as a scaffold to accommodate both antiEpCAM antibody (aEpCAM) and antiSlex antibody (aSlex) in harmony to capture and restrain colorectal CTCs in vivo. 3
ACCEPTED MANUSCRIPT
SC R
IP
T
Here, we show that the dual antibody-coated conjugate can bind and capture CTCs more selectively and specifically than the single antibody-coated conjugates in the presence of interfering cells (i.e., HL-60 cells used as a leukocyte model) and red blood cells (RBCs). Furthermore, the dual antibody-coated conjugate can capture CTCs in the colorectal cancer patient blood as well as in animal blood when the cancer cells and the antibody-coated conjugates were injected sequentially to the nude mice. HT29 cells were selected as human carcinoma CTC model because they are often found in patients’ blood [25]. This study provides the strong evidence that the dual antibody-coated conjugate possesses high specificity in capturing and restraining CTCs in blood.
AC
CE P
TE
D
MA
NU
2. Methods 2.1 Synthesis of G6 PAMAM dendrimer-antibody conjugates After surface modification (see Supplementary information), the completely-carboxylated (CC G6) dendrimers were ready for conjugating antibodies [aEpCAM (or abbreviated as E hereafter), aSlex (or abbreviated as S hereafter)] or fluorescence-labeled antibodies [phycoerythrin (PE) linked aEpCAM (aEpCAM-PE), fluorescein isothiocyanate (FITC) linked aSlex (aSlex-FITC)]. The conjugation reaction was conducted by dissolving CC G6 (0.83 μg, 11.86 pmol) in 3 ml phosphate-buffered saline (PBS). Its carboxylic ends reacted with 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) (113.66 ng, 592.90 pmol, 50 molar excess) and N-hydroxysuccinimide (NHS) (68.24 ng, 592.90 pmol, 50 molar excess) at 25 0C for 1 h. The activated dendrimers were divided into three vials, each reacted with aEpCAM (19.75 pmol; 5 molar excess) for G6-5E conjugate, aSlex (11.85 pmol; 3 molar excess) for G6-3S conjugate, and the combination of aEpCAM (19.75 pmol) and aSlex (11.85 pmol) for the dual antibody conjugation (G6-5E-3S). The fluorescence-labeled single or dual antibody conjugates (G6-5E-PE, G6-5S-FITC, G6-3S-FITC, PE-3E-G6-3S-FITC and PE-5E-G6-3S-FITC) and other conjugates (G6-3E, G6-5S and G6-5E-5S) were synthesized following the similar procedures with or without fluorescence-labeled antibodies. All the reactions were conducted under vigorous stirring overnight in dark. Purification of the conjugates were obtained after dialysis (10,000 MWCO) against DDI water overnight followed by lyophilization. 2.2 Characterization of the conjugates Physiochemical characterization of all the dendrimer derivatives and conjugates were conducted by using 1H NMR (AVANCE III) in D2O for changes in H atom, FTIR (Nicolet 360) for changes in active groups, Zeta potential/ dynamic light scattering analyzer (Zetaplus/ 90plus, Brookharen) for changes in surface charge and particle size (diameter, nm) of the synthesized nanomaterials, respectively. Size of the conjugate was measured immediately after 30-min ultrasonic treatment. Morphology of the dendrimers and their conjugates were determined by both scanning electron microscope (SEM, Nova NanoSEM 230) and atomic force microscope (AFM, Agilent Technologies, 40N/ m tapping probe, AC mode) measurements. The number of FITC molecules conjugated to each partially carboxylated G6 PAMAM (PC G6) was back-calculated based on the standard curve of free FITC concentration-fluorescence intensity using a fluorescence spectrophotometer (Hitachi, Japan). The amount of antibody in each conjugate was determined by its characteristic UV absorption with the aid of a fluorescence inverted microscope to assure the conjugation of fluorescence-labeled antibodies to dendrimer surface. 4
ACCEPTED MANUSCRIPT
MA
NU
SC R
IP
T
2.3 Microscopic analysis of the captured CTCs Cells were cultured as described in Supplementary information. Adherent cells (105 cells ml-1) were evenly seeded on 35 mm dishes with glass coverslips on the bottom and incubated with fluorescence-labeled antibodies (E-PE and S-FITC), or the conjugates (G6-E-PE, G6-S-FITC and PE-E-G6-S-FITC) at the same concentration (20 μg ml-1, 1 h, 37 0 C) in dark after 30-min pretreatment with PBS containing 1% bovine serum albumin (BSA). Cells were stained with PBS containing the nuclei stain dihydrochloride (DAPI, 10 μg ml-1, blue) for 15 min, and then rinsed and covered with serum-free medium for analysis by using an Olympus FluoView 1000 laser scanning confocal microscope. The non-adherent HT29 cells (106 cells) were first stained with Hoechst 33258 (blue) for 15 min, and then mixed with fluorescence-labeled antibodies and conjugates for 1 h at 37 0C in dark after blocking the non-specific binding with 1% BSA for 30 min. Fluorescence images were taken following removal of the unbound conjugates. The captured cells were quantified by using the Becton Dickinson (BD) multiparametric fluorescence-activated cell sorting (FACS) Aria III analyzer.
CE P
TE
D
2.4 Flow cytometric analysis of captured HT29 in the presence of interfering HL-60 or RBCs A series of artificial CTC blood samples were prepared by mixing HT29 with a large population of either HL-60 cells, or RBCs as mentioned above. Cell mixture was treated with individual fluorescence-labeled conjugates (20 μg ml-1) as mentioned above, the capture efficiency of dual antibody conjugates was finally determined based on the % FITC+PE+ HT29 cells and the number of FITC+PE+ HT29 cells by the flow cytometry. The flow cytometric procedures were explicitly described in Supplementary information.
AC
2.5 Analyses of cell viability, cell cycle and apoptosis after capture The analyses were conducted as we described previously [26, 27]. Briefly, the cell viability of the captured HT29 was determined by the MTT {[3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrasodium bromide] tetrazolium salt)} assay after the cells were exposed to the purified conjugates (0, 5, 10, 20 μg ml-1) for 48 h (Tecan Infinite M200 pro). In parallel, the effects of the conjugates on the cell population distribution in every phase (G0/G1, S, and G2/M) and the cell apoptotic status were analyzed. The morphologic change of the suspensory cells was evaluated by using a fluorescence inverted microscope (Zeiss Axio Observer A1). 2.6 CTCs captured in vivo and in patient blood Human colon cancer HT29 cells were pre-labeled with red fluorescence protein (RFP) by using a lentiviral vector with an RFP transgene. The selected RFP-labeled HT29 cells with the stable fluorescence yield (>90%) were cultured and counted for the following capture assay. The PBS containing RFP-labeled HT29 cells (5×104) were intravenously injected into the nude mice. Ten-min later, the fluorescence-labeled conjugates G6-5E-PE, G6-5S-FITC, and PE-5E-G6-3S-FITC (80 μg per mouse; the dual antibody conjugate has a low molar concentration) were administered to the mice separately. One hour later, one-ml blood was individually withdrawn and treated with the RBCs lysis buffer. The isolated cells via centrifugation were then stained with Hoechst 33258 (25 μg ml-1) for imaging under a 5
ACCEPTED MANUSCRIPT
MA
NU
SC R
IP
T
fluorescence inverted microscope (Zeiss Axio Observer A1). Blood samples were drawn from eight colorectal cancer patients when they were under operation to remove the primary tumors. The study was reviewed and approved by the hospital institutional review board (IRB). Blood was collected into vacutainer tubes containing the anticoagulant EDTA. The single and dual antibody conjugates (20 μg ml-1) labeled with fluorescence (PE and/or FITC) were individually spiked into the patient blood (1 ml) pretreated with 1% BSA for 30 min. Blood without the conjugates was used as the control. After 1 h of incubation at 37 0C, the unbound conjugates were removed together with the lysed RBCs, and the captured cells and white blood cells (WBCs) were obtained via centrifugation. The obtained cells were blocked with 1% BSA at room temperature for another 30 min. Finally, APC-conjugated anti-CD45 was used to exclude the WBCs. After normalization of the fluorescence intensity with the corresponding isotype controls, CTCs captured by the single and dual antibody conjugates were quantified by the flow cytometry. After labeling the cell nuclei with DAPI dye (10 μg ml-1) in dark for 15 min, the captured CTCs were distinguished from the WBCs by the fluorescence images under a laser confocal microscope (Leica TCS SPE).
AC
CE P
TE
D
3. Results 3.1 Preparation of antibody-coated functionalized PAMAM dendrimers The surface modification and functionalization of G6 PAMAM dendrimers in the present study were mainly performed by starting carboxylation of the purified G6 PAMAM at their primary amine end groups with succinic anhydride (SA) (Fig. 1a). G6 PAMAM dendrimers (MW 62400 g mol-1, theoretically) were chosen not only because of their biocompatibility and biodegradability, but also owing to their number of surface functional groups (256) and large surface area to accommodate multiple aEpCAM and aSlex per dendrimer to enable more specific recognizing and binding HT29 cells by the biomimetic conjugates. The carboxylate-functionalized PAMAM surfaces were first activated using an 1:1 mixture of EDC and NHS for 1 h followed by the covalent conjugation with the aEpCAM and aSlex [19] (Fig. 1a). The conjugates were purified by dialysis in DDI water overnight and then lyophilized. To visualize the binding and capturing effects of the conjugates on the targeted HT29 cells, aEpCAM and aSlex were linked to PE (orange fluorescence) and FITC (green fluorescence), respectively, to exhibit high levels of fluorescence signal. The green fluorescence intensities from aSlex-FITC substantially increased, in a nonlinear fashion, with an increase in the amount of aSlex conjugated. The orange fluorescence from PE and the green fluorescence from FITC have little to no spectral overlap after normalization of the fluorescence intensities. The even distribution of the merged fluorescence indicated that the uniform coating of both aEpCAM-PE and aSlex-FITC was achieved. The aEpCAM and aSlex reacted with the completely-carboxylated dendrimers, and the resulted materials showed no absorption at 220 nm, the wavelength that can be used as an indicator of the presence of either aEpCAM or aSlex in preparation for the conjugates [13]. The preliminary conjugation experiments suggested that conjugation of the aEpCAM to the G6 PAMAM was facile, especially when the dual antibody conjugation was performed at molar ratios of CC G6: aSlex: aEpCAM in 1:3:3, or 1:3:5 (the optimal condition for the conjugation). The synthesized and purified conjugates are designated hereafter as G6-3E, G6-5E, G6-3S, G6-5S, G6-5E-3S and G6-5E-5S, respectively, based on the reaction molar ratios of CC G6 6
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
NU
SC R
IP
T
(G6) to aEpCAM (E) or aSlex (S) molecules. Fluorescence imaging was conducted using a laser scanning confocal microscope to monitor the conjugation process.
Fig. 1 Conceptual illustration and physicochemical characterization of re-engineered G6 PAMAM dendrimers coated with two antibodies targeting typical HT29 surface biomarkers (EpCAM and Slex). a, Chemical surface modification of G6 PAMAM and sequential link of aEpCAM and aSlex to the modified G6 PAMAM; Synthesis procedures of G6 PAMAM dendrimer derivatives and the related conjugates. 1, unmodified G6; 2, completely carboxylated G6 (CC G6); 3, partially carboxylated G6 (PC G6); 4, G6 single or dual antibody conjugates (G6-aEpCAM, e.g., G6-3E, G6-5E; G6-aSlex, e.g., G6-3S, G6-5S; and G6-aEpCAM-aSlex, e.g., G6-5E-3S, G6-5E-5S); 5, PC G6 conjugated with fluorescein; 7
ACCEPTED MANUSCRIPT
T
6, G6-antibody conjugates labeled with FITC; 7, CC G6 modified with fluorescence-labeled antibody (G6-5E-PE, G6-5S-FITC, PE-3E-G6-3S-FITC and PE-5E-G6-3S-FITC). b, FTIR spectra of CC G6 and its corresponding conjugates. c, SEM image of the dual antibody-coated conjugate G6-5E-5S. d, A typical AFM scan of the single antibody-coated conjugate G6-5E. e, Merged images of fluorescence-labeled G6-antibody conjugates in PBS (pH 7.4) solution obtained by a laser confocal microscope.
AC
CE P
TE
D
MA
NU
SC R
IP
3.2 Characterization of antibody-coated functionalized PAMAM dendrimers To measure the numbers of aEpCAM and aSlex conjugated to one dendrimer, a series of known concentrations of aEpCAM and aSlex were spiked into the running buffer to establish a standard curve for UV/Vis measurements of sample concentrations of the conjugates. Based on the linear curves of free aEpCAM and aSlex at 220 nm, the numbers of the conjugated aEpCAM and aSlex molecules per G6 PAMAM were calculated to be about 2 or 6 moles of aEpCAM or aSlex molecules conjugated to one G6 PAMAM (Supplementary Fig. 1) according to the initial amount of the aEpCAM or aSlex added. The numbers of E and S molecules in dual antibody conjugate could also be inferred and calculated based on the molar reaction ratios of the functionalized dendrimers (CC G6) to E and/or S by using the fixed amount of CC G6, E and/or S for the conjugation reaction. According to the UV analysis, the molecular weights of G6-5E, G6-3S and G6-5E-3S conjugates were estimated as 377 kDa, 924 kDa and 1240 kDa, respectively, based on the numbers of E and/or S molecules in each single or dual antibody conjugate. The molecular weight of G6 PAMAM dendrimers was known as 62.4 kDa, and that of CC G6 dendrimers was calculated to be 69.8 kDa. So when the same mass concentration unit of these conjugates (μg ml-1) was converted to the molar concentration (nM), the molar concentrations of G6-5E-3S were about 16-18-fold less than CC G6 dendrimers and 2-4-fold less than other conjugates. Surface-modified PAMAM derivatives were confirmed by 1H NMR and size/ zeta potential analyses (Supplementary Fig. 2). FTIR spectra showed increasing intensity of amide (-CONH-) IR signals when the surface functionalized PAMAM was linked to the increasing numbers of aEpCAM and aSlex (Fig. 1b). Various conjugated PAMAM dendrimers were separately eluded by high performance size exclusion chromatography. The corresponding chromatography showed distinct retention times for each individual derivative and conjugate. The size and morphology of the antibody-coated PAMAM conjugates were characterized by the SEM. Fig. 1c showed that the size of the dual antibody conjugate G6-5E-5S was less than 100 nm and in a pie shape. The tapping mode of AFM showed the characteristic morphology of G6 PAMAM conjugated with either aEpCAM or aSlex. The mean diameter measured by the AFM of CC G6, aSlex, G6-5E, G6-5S and G6-5E-5S was 6, 25, 35, 40 and 100 nm, respectively (Fig. 1d and Supplementary Table 1). The thickness of the conjugates in the assembly was 1-2 nm. Dynamic light scattering (DLS) measurement showed that the hydrodynamic diameter of the conjugates significantly increased when G6 PAMAM was coated with aEpCAM or aSlex, separately, or in combination (Supplementary Table 1). The increase in diameter of the G6 PAMAM as a result of the surface coating, and the uniform size and morphology of the conjugates (Fig. 1c, d) clearly indicated the formation of the conjugates. Zeta-potential measurement indicated the increased negative charges of G6 PAMAM after surface modification with –COOH group and functionalization with aSlex and aEpCAM, separately or in combination (Supplementary Table 1). The numbers of FITC molecule in 8
ACCEPTED MANUSCRIPT
MA
NU
SC R
IP
T
G6-COOH-FITC derivatives, aEpCAM in G6-aEpCAM and aSlex in G6-aSlex conjugates were estimated about 40, 2 and 6 molecules based on the corresponding intensity of fluorescence and UV absorption measured (Supplementary Fig. 1). The fluorescent intensity was enhanced with the increasing number of the conjugated antibodies. When aEpCAM and aSlex were separately conjugated to the same G6 PAMAM scaffold, the merged fluorescence exhibited the typical green-yellow color, indicating the successful conjugation (Fig. 1e; right two panels). The above data well characterized and confirmed the successful synthesis of the single and dual antibody-coated conjugates. Biostability of the conjugates was evaluated by measuring the changes in UV absorption of the conjugates at different pHs, temperatures and agitation rates. Under static condition and at pHs 5.6, 7.4 and 9.13 and temperatures 4 and 37 0C, it was found that those conjugates were most stable at pH 7.4. Both acidic and basic conditions expedited the degradation of the conjugates at room temperature. Individual conjugates were stable at 4 0C for as long as the tested 11 days without significant degradation. At 37 0C, some degradation was observed after 24 h incubation. Shaking the conjugates at velocity of 80, 120 and 210 rpm for as long as 30 min at 25 0C did not cause significant degradation of the conjugates.
AC
CE P
TE
D
3.3 Sensitivity and efficiency of capturing adherent and non-adherent CTCs To make sure that the HT29 cells express EpCAM and Slex, we quantitatively analyzed the expression levels of EpCAM and Slex on the HT29 cells under the same conditions (Supplementary Fig. 3). The result showed the overexpression of EpCAM and Slex on the surface of HT29 cells. Moreover, the result explained why the PC G6 and CC G6 without aEpCAM and/ or aSlex conjugated showed no noticeable binding to the HT29 cells. Whereas when the adherent HT29 cells were incubated with aEpCAM-PE and G6-5E-PE (orange), or aSlex-FITC and G6-3S-FITC (green), or PE-5E-G6-3S-FITC (green-yellow) for 1 h at 37 0C, the cells bound to these conjugates exhibited different degrees and colors of fluorescence that could be quantified by using the laser scanning confocal microscope (Fig. 2a). The mean fluorescence images have identical exposure times and normalization. The numbers of captured cells increased when the amounts of aEpCAM or aSlex coated on G6 increased, or when the concentrations of either single or dual antibody conjugates increased from 10 µg ml-1 to 20 µg ml-1. The metastatic ability of the adherent CTCs [15] may differ from that of characteristic non-adherent CTCs [28], where, the latter remain viable in the bloodstream after loss of attachment to endothelial membrane. We therefore assessed the capture specificity of the conjugates to the suspensory HT29 cells by using both fluorescence inverted microscope analysis and flow cytometry (Fig. 2b-d). The cells were stained with Hoechst 33258 (blue) to show each individual cell following incubation of the cells with 1% BSA in PBS for 30 min and mixed with fluorescence-labeled conjugates at 37 0C for 1 h. G6-5E-PE, G6-3S-FITC and PE-5E-G6-3S-FITC showed typical merged images upon binding to the stained cells (Fig. 2b). To determine the optimal incubation time, the non-adherent HT29 cells were cultured in the presence of the conjugates (20 µg ml-1) for 0.5-4 h. The conjugates bound the cells within 1 h, indicating the efficient binding period. Capture efficiency of the synthesized conjugates targeting HT29 cells was determined by using the FACS Aria III (Becton Dickinson) with laser excitation at λex 488 nm. Immunoglobulins labeled with the same fluorochromes were used as isotype controls to exclude the autofluorescence and 9
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
NU
SC R
IP
T
non-specific binding. Data acquisition was based on the threshold of 10,000 cells to meet the criteria of the light scatter. Dot plots were adjusted to show the number of fluorescence-labeled conjugates bound to HT29 cells after normalization for both FITC and PE dyes. A gating strategy was first focused on quantitatively distinguishing the conjugate-captured FITC+PE+ HT29 cells from the free cells. The capture efficiency (% positive cells) was defined as the number of the conjugate-captured FITC+PE+ HT29 cells divided by the number of the cells shown in the P1 gate. The quantitative flow cytometric analysis demonstrated that the dual conjugates PE-5E-G6-3S-FITC and PE-3E-G6-3S-FITC (both 20 µg ml-1) captured more suspensory cells in a cytometric tube than its single conjugate counterparts G6-5E-PE and G6-3S-FITC, or G6-3E-PE and G6-3S-FITC (both 20 µg ml-1) added together to the tube (Fig. 2c, d). Moreover, PE-5E-G6-3S-FITC captured more HT29 cells than its dual conjugate counterpart PE-3E-G6-3S-FITC (containing less aEpCAM), suggesting the importance of aEpCAM for capturing HT29 cells [12, 29-32] (Fig. 2c), which was consistent with what we found that the HT29 cells expressed EpCAM higher than Slex (Supplementary Fig. 3). The average number of the captured FITC+PE+ HT29 cells by PE-5E-G6-3S-FITC was greater than that by the individual single antibody conjugates G6-5E-PE and G6-3S-FITC added together (Fig. 2d). The above results suggest that the dual antibody-coated conjugates possess capture capability greater than their single conjugate counterparts, and the capture efficiency is at least EpCAM-dependent.
10
AC
CE P
TE
D
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
Fig. 2 Capability of aEpCAM- and aSlex-coated conjugates to recognize and capture adherent or non-adherent HT29 cells. a, Recognizing and binding of fluorescence-labeled antibody aEpCAM-PE (E-PE), aSlex-FITC (S-FITC), G6-antibody conjugates (G6-5E-PE, G6-3S-FITC, PE-5E-G6-3S-FITC) to adherent HT29 cells examined by using confocal microscope at FITC λex 488 nm, λem 500-535 nm, PE λex 550 nm, λem 570-610 nm, and DAPI λex 405 nm, λem 425-475 nm; b, to non-adherent HT29 cells examined by using fluorescent microscope at λex 470/40 nm, λem 525/50 nm and λex 365 nm, λem 445/50 nm. The confocal images have identical exposure conditions (1 h, 37 0C, 5% CO2). All the images were the merged pictures taken after DAPI, PE and/or FITC stain. The merged image colors distinguished the binding of HT29 cells to the single and dual antibody conjugates from the binding of HT29 to the antibody alone. c, Comparison of capture efficiency quantified by the FACS analysis between dual antibody-coated 11
ACCEPTED MANUSCRIPT
IP
T
conjugates (PE-E-G6-S-FITC) and their single antibody counterparts added together (G6-3E-PE+ G6-3S-FITC; or G6-5E-PE+ G6-3S-FITC) at the same concentration 20 µg ml-1 for capturing HT29. Note, the added concentration of the combined single antibody conjugates was therefore 40 µg ml-1; capture efficiency (%) was calculated based on the control. d, Dot plot analysis by gating plots of FITC+PE+ HT29 cells; the number in Q2 plot represents the positive cells for each sorting in the presence of the conjugates. S stands for aSlex, and E, aEpCAM, with the molecular number prefixed.
AC
CE P
TE
D
MA
NU
SC R
3.4 Capture specificity in the presence of interfering HL-60 or RBCs To evaluate the ability of the conjugates in recognizing and capturing CTCs, HT29 was used as a CTC model. The capture specificity, on the basis of our previous definition [33], was defined as the ability to assess unequivocally the CTCs captured in the presence of components (RBCs, WBCs) that may interrupt the quantitative measurement, while maintaining their own characteristics. The specificity of the conjugates binding to the surface of HT29 cells was examined by mixing a series of HT29 with either HL-60 cells (as a leukocyte model that lacks HT29 biomarker EpCAM [34-36]), or RBCs at a mixture ratio of 1:103 or 1:105 (HT29: HL-60; or HT29: RBCs) with fixed total 106 cells per incubation tube (Fig. 3a), or by mixing 1,000 HT29 cells with 106 or 108 RBCs per incubation tube (Fig. 3b, c). HL-60 cells usually have no interaction with aEpCAM [36]. The ratio mimicked the clinical situation in which, roughly 1 CTC presented in 10 3-106 leukocytes [37], 14-5,000 CTCs per milliliter blood [6], about 175 CTCs in 3 ml of blood [17], or 100 CTCs in 108 RBCs [38, 39]. We can regularly sort about 10 colorectal CTCs from 1 ml of patient blood. Under optimal gate conditions, the synthesized conjugates that recognized and captured the suspended HT29 could be visualized by fluorescence intensity of the flow cytometric images. In general, the conjugates recognized and bound HT29 in a cooperative manner in the presence of either HL-60 or RBCs, whereas, the isotype IgG control did not bind or capture the suspended HT29 cells. The dual antibody conjugate PE-5E-G6-3S-FITC captured 13.3% and 19.7% more HT29 cells than its single antibody conjugates G6-5E-PE and G6-5S-FITC in a solution where 1,000 HT29 cells were spiked into 1×10 6 RBCs (Fig. 3c). The numbers or percentage of bound and captured HT29 by those conjugates decreased when the mixed cell ratio of HT29 to HL-60 or RBCs changed from 1:103 to 1:105 (HL-60), or from 106 to 108 (RBCs) (Fig. 3a-c) because of the interfering effect. For example, the dual antibody conjugate captured less HT29 cells (8-10% less) when the total numbers of RBCs were increased to 108 cells (Fig. 3c). Similarly, when more RBCs were added to a fixed number of HT29 cells (103), dot plots showed a reduction in the number of the captured HT29 (Fig. 3b). These results collectively demonstrated that the dual antibody conjugates showed the significant advantages over single ones in terms of the capture specificity and capture capability.
12
AC
CE P
TE
D
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
Fig. 3 Specificity and capacity of single and dual antibody-coated G6 conjugates in capturing HT29 cancer cells in the presence of interfering HL-60 cells (a leukocyte 13
ACCEPTED MANUSCRIPT
NU
SC R
IP
T
model) and RBCs. a, Flow cytometric analysis of the total numbers of HT29 cells captured by the PE-FITC- fluorescence-labeled single or dual antibody-conjugates. HT29 cells were mixed with either HL-60 cells (left) or RBCs (right) at the ratio of 1:103 (HT29: HL-60, or RBCs; open bar), or 1:105 (filled bar) with the total 106 cells combined. b, Dot plots showed enhanced capture capacity when PE-3E-G6-3S-FITC was replaced with PE-5E-G6-3S-FITC conjugate. The HT29 cells (1,000) co-existed with 106 (left) or 108 (right) RBCs. The number shows the captured FITC+PE+ HT29 cells and indicates the importance of aEpCAM for the capture. c, Comparison of capture efficiency between dual antibody conjugates (PE-E-G6-S-FITC) and single ones in capturing HT29 cells (1,000) spiked to 106 (open bar) or 108 (filled bar) RBCs. Note, HT29 cells incubated with two immunoglobins labeled with PE and FITC were used as the isotype controls to normalize the fluorescence intensities. S stands for aSlex, and E, aEpCAM, with the molecular number prefixed. The significance analysis was performed on the basis of the PE-5E-G6-3S-FITC conjugate or control. Error bars denote the standard deviation of Student t test.
AC
CE P
TE
D
MA
3.5 Down-regulation of the captured CTCs and related mechanisms To explore whether the conjugates had an effect on cell viability, the conjugates were individually incubated with various cell lines for 48 h at concentrations ranging from 1.25-20 μg ml-1. The cell lines investigated included the human umbilical vein endothelial cells ECV-304, melanoma cells A-375, cervical cancer cells Hela, hepatocellular carcinoma cells HepG2 [26, 27] and HT29. These cell lines usually express lower levels of EpCAM and/or Slex than HT29 cells [21-23, 30, 40]. EpCAM and Slex expression levels of the five cell lines were also confirmed by using flow cytometry. In comparison to the control, the conjugates without the fluorescence-labeling produced a concentration-dependent reduction in cell viability when the conjugate concentrations increased from 5, 10 to 20 μg ml-1 (Table 1, 2). The conjugate G6-5E inhibited HT29 cells more significantly than other tested cell lines that usually express low EpCAM and/or Slex activities (Table 1). Further analysis demonstrated that both G6-5E and G6-5E-3S exhibited a significant concentration-dependent inhibition on HT29 cell viability. In contrast, CC G6 and G6-3S at molar concentration of five-fold excess over their conjugates did not show the significant effect (Table 2). The result further demonstrated the capture capability of the conjugates, and suggested that the interruption of EpCAM activity of HT29 cells by the aEpCAM-coated conjugates may be attributed to their inhibition on HT29. Table 1. Deactivation of the captured HT29 cells by G6-5E single antibody conjugate. The conjugate showed the effect on cell viability in a concentration dependent manner. However, it has low effects on other cell lines that express less EpCAM and/or Slex. Cell lines HT29 A-375 ECV-304 Hela HepG2
Cell viability (%) 10 μg ml-1 50.21±7.97## 75.89±7.64 77.00±8.17 82.60±5.77 82.57±5.81
-1
5 μg ml 55.20±6.37## 82.41±22.00 84.80±1.36 88.30±1.32 89.10±6.62 14
20 μg ml-1 33.80±6.71## 71.61±11.78 62.23±7.90## 70.81±7.04 78.43±1.94
ACCEPTED MANUSCRIPT #
P
S0168-3659(15)00275-8 doi: 10.1016/j.jconrel.2015.04.036 COREL 7657
To appear in:
Journal of Controlled Release
Received date: Revised date: Accepted date:
21 January 2015 8 April 2015 27 April 2015
Please cite this article as: Jingjing Xie, Yu Gao, Rongli Zhao, Patrick J. Sinko, Songen Gu, Jichuang Wang, Yuanfang Li, Yusheng Lu, Suhong Yu, Lie Wang, Shuming Chen, Jingwei Shao, Lee Jia, Ex vivo and in vivo capture and deactivation of circulating tumor cells by dual-antibody-coated nanomaterials, Journal of Controlled Release (2015), doi: 10.1016/j.jconrel.2015.04.036
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Ex vivo and in vivo capture and deactivation of circulating
T
tumor cells by dual-antibody-coated nanomaterials
IP
Jingjing Xie1, Yu Gao1, Rongli Zhao1, Patrick J. Sinko2, Songen Gu1, Jichuang Wang1,
SC R
Yuanfang Li1, Yusheng Lu1, Suhong Yu1, Lie Wang3, Shuming Chen3, Jingwei Shao1, Lee
MA
NU
Jia1, *
1
D
Cancer Metastasis Alert and Prevention Center, and Biopharmaceutical Photocatalysis of
TE
State Key Laboratory of Photocatalysis on Energy and Environment, College of Chemistry,
2
CE P
Fuzhou University, Fuzhou 350002, China.
08854, USA.
AC
Rutgers, The State University of New Jersey, 160 Frelinghuysen Road, Piscataway, NJ,
3
Department of Oncology, East Hospital, Fuzhou, 350025, China.
*Corresponding authors: Lee Jia ([email protected] or [email protected]), 523 Industry Road, Science Building, 3FL., Fuzhou University, Fuzhou, Fujian, China, 350002. Phone and Fax: +86-0591-8357-6912.
1
ACCEPTED MANUSCRIPT
ABSTRACT
MA
NU
SC R
IP
T
Circulating tumor cells (CTCs) have been detected by us and others in cancer patient blood. However, little is known about how to specifically capture and deactivate CTCs in vivo, which may lead to successful metastasis prevention in asymptomatic cancer survivors after surgery. We hypothesize that the dual antibody conjugates may have the advantage of capturing CTCs specifically over their single antibody counterparts. Here we show that the surface-functionalized dendrimers can be sequentially coated with two antibodies directed to surface biomarkers (EpCAM and Slex) of human colorectal CTCs. The dual antibody-coated dendrimers exhibit a significantly enhanced specificity in capturing CTCs in the presence of interfering blood cells, and in both eight-patient bloods and nude mice administered with the labeled CTCs in comparison to their single antibody-coated counterparts. The dual antibody-coated conjugates down-regulate the captured CTCs. This study provides the first conceptual evidence that two antibodies can be biocompatibly conjugated to a nanomaterial to capture and down-regulate CTCs in vivo with the enhanced specificity.
AC
CE P
TE
D
KEYWORDS: Cancer metastasis; Circulating tumor cells; Dual antibody conjugates; Capture efficiency; Capture specificity
2
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
NU
SC R
IP
T
1. Introduction The root cause of cancer metastasis can be traced down to the presence of circulating tumor cells (CTCs) or tumor-specific DNA [1] in the blood [2]. As the estimated number of cancer survivors will reach 18 million in the USA alone by 2022 [3], specifically restraining the residual CTCs in cancer survivors becomes increasingly important [4, 5]. CTCs in cancer survivors often show a low rate of proliferation when cancer survivors are in remission and/or asymptomatic [6]. Thus current post-metastatic chemotherapy strategies that only target highly proliferating cancer and non-cancer cell lines can't be effective in targeting and killing CTCs [7]. As a matter of fact, anticancer chemotherapy may enhance metastasis formation [8], and the current nanotechnology approaches that conjugate nanomaterials with a typical antibody and a representative anticancer drug have not resulted in a successful cancer treatment. Indeed, most tumor-targeting nanotechnology reported to date is developed to utilize the unique property of the enhanced permeability and retention of the post-metastatic tumor tissues for treatment [9, 10]. Other cancer nanotechnology developed is based on the CTCs surface biomarkers for early in vitro cancer metastasis diagnosis, including utilization of a single surface biomarker of CTCs such as the anti-epithelial cell adhesion molecule (aEpCAM) to coat silicon-nanopillars [11, 12], nanoscale poly(amidoamine) (PAMAM) dendrimers [13], or functionalized graphene oxide nanosheets [14] for specific in vitro capturing CTCs. Recent reports revealed the co-existence of various surface biomarkers such as HER-2/EGFR/HPSE/Notch1 [15], EpCAM/CD44/CD47/MET [16] and EpCAM/HER-2/EGFR [17] on the single types of breast CTCs. Various combinations of the corresponding antibodies have thus been developed as an antibody cocktail-coated microchip [18], a microfluidic herringbone-chip [17], or an epoxy-functionalized glass slide with three channels immobilizing an individual PEG-dendrimer-antibody on each channel [19] for in vitro CTCs diagnosis. However, little is known about utilization of bionanotechnology for specifically capturing CTCs and restraining their activity in the bloodstream of the asymptomatic cancer survivors in order to prevent the future cancer metastasis. The stage-2 and -3 colorectal cancer patients often tragically confront 70% and 80% metastatic rate within 3-5 years after surgical removal of the primary cancer [20]. Human colorectal carcinoma HT29 cells often express two CTCs biomarkers, the EpCAM [21] and the saliva acidifying louis oligosaccharides X (Sialyl Lewis X, Slex) [22, 23]. However, EpCAM down-regulation was found during the epithelial-to-mesenchymal transition of CTCs. Thus, only utilization of antibody against EpCAM may not capture the EpCAM-negative CTCs. In addition, the abundance of a single surface biomarker may vary chronometrically with the cell cycle phases, resulting in variable binding affinity between the biomarker and its corresponding ligand [24]. Moreover, one CTCs surface biomarker (e.g., CD44) may also express on other normal cell lines leading to a false biological recognition [16]. Based on the above analysis, we hypothesized that conjugating the functionalized PAMAM dendrimers with two clinically-used surface biomarker antibodies, instead of a single antibody, may significantly enhance the sensitivity and specificity of the antibody-coated conjugates in identifying and capturing the CTCs, and consequently result in down-regulation of the activity of the captured CTCs. We therefore developed the surface-functionalized PAMAM dendrimers as a scaffold to accommodate both antiEpCAM antibody (aEpCAM) and antiSlex antibody (aSlex) in harmony to capture and restrain colorectal CTCs in vivo. 3
ACCEPTED MANUSCRIPT
SC R
IP
T
Here, we show that the dual antibody-coated conjugate can bind and capture CTCs more selectively and specifically than the single antibody-coated conjugates in the presence of interfering cells (i.e., HL-60 cells used as a leukocyte model) and red blood cells (RBCs). Furthermore, the dual antibody-coated conjugate can capture CTCs in the colorectal cancer patient blood as well as in animal blood when the cancer cells and the antibody-coated conjugates were injected sequentially to the nude mice. HT29 cells were selected as human carcinoma CTC model because they are often found in patients’ blood [25]. This study provides the strong evidence that the dual antibody-coated conjugate possesses high specificity in capturing and restraining CTCs in blood.
AC
CE P
TE
D
MA
NU
2. Methods 2.1 Synthesis of G6 PAMAM dendrimer-antibody conjugates After surface modification (see Supplementary information), the completely-carboxylated (CC G6) dendrimers were ready for conjugating antibodies [aEpCAM (or abbreviated as E hereafter), aSlex (or abbreviated as S hereafter)] or fluorescence-labeled antibodies [phycoerythrin (PE) linked aEpCAM (aEpCAM-PE), fluorescein isothiocyanate (FITC) linked aSlex (aSlex-FITC)]. The conjugation reaction was conducted by dissolving CC G6 (0.83 μg, 11.86 pmol) in 3 ml phosphate-buffered saline (PBS). Its carboxylic ends reacted with 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) (113.66 ng, 592.90 pmol, 50 molar excess) and N-hydroxysuccinimide (NHS) (68.24 ng, 592.90 pmol, 50 molar excess) at 25 0C for 1 h. The activated dendrimers were divided into three vials, each reacted with aEpCAM (19.75 pmol; 5 molar excess) for G6-5E conjugate, aSlex (11.85 pmol; 3 molar excess) for G6-3S conjugate, and the combination of aEpCAM (19.75 pmol) and aSlex (11.85 pmol) for the dual antibody conjugation (G6-5E-3S). The fluorescence-labeled single or dual antibody conjugates (G6-5E-PE, G6-5S-FITC, G6-3S-FITC, PE-3E-G6-3S-FITC and PE-5E-G6-3S-FITC) and other conjugates (G6-3E, G6-5S and G6-5E-5S) were synthesized following the similar procedures with or without fluorescence-labeled antibodies. All the reactions were conducted under vigorous stirring overnight in dark. Purification of the conjugates were obtained after dialysis (10,000 MWCO) against DDI water overnight followed by lyophilization. 2.2 Characterization of the conjugates Physiochemical characterization of all the dendrimer derivatives and conjugates were conducted by using 1H NMR (AVANCE III) in D2O for changes in H atom, FTIR (Nicolet 360) for changes in active groups, Zeta potential/ dynamic light scattering analyzer (Zetaplus/ 90plus, Brookharen) for changes in surface charge and particle size (diameter, nm) of the synthesized nanomaterials, respectively. Size of the conjugate was measured immediately after 30-min ultrasonic treatment. Morphology of the dendrimers and their conjugates were determined by both scanning electron microscope (SEM, Nova NanoSEM 230) and atomic force microscope (AFM, Agilent Technologies, 40N/ m tapping probe, AC mode) measurements. The number of FITC molecules conjugated to each partially carboxylated G6 PAMAM (PC G6) was back-calculated based on the standard curve of free FITC concentration-fluorescence intensity using a fluorescence spectrophotometer (Hitachi, Japan). The amount of antibody in each conjugate was determined by its characteristic UV absorption with the aid of a fluorescence inverted microscope to assure the conjugation of fluorescence-labeled antibodies to dendrimer surface. 4
ACCEPTED MANUSCRIPT
MA
NU
SC R
IP
T
2.3 Microscopic analysis of the captured CTCs Cells were cultured as described in Supplementary information. Adherent cells (105 cells ml-1) were evenly seeded on 35 mm dishes with glass coverslips on the bottom and incubated with fluorescence-labeled antibodies (E-PE and S-FITC), or the conjugates (G6-E-PE, G6-S-FITC and PE-E-G6-S-FITC) at the same concentration (20 μg ml-1, 1 h, 37 0 C) in dark after 30-min pretreatment with PBS containing 1% bovine serum albumin (BSA). Cells were stained with PBS containing the nuclei stain dihydrochloride (DAPI, 10 μg ml-1, blue) for 15 min, and then rinsed and covered with serum-free medium for analysis by using an Olympus FluoView 1000 laser scanning confocal microscope. The non-adherent HT29 cells (106 cells) were first stained with Hoechst 33258 (blue) for 15 min, and then mixed with fluorescence-labeled antibodies and conjugates for 1 h at 37 0C in dark after blocking the non-specific binding with 1% BSA for 30 min. Fluorescence images were taken following removal of the unbound conjugates. The captured cells were quantified by using the Becton Dickinson (BD) multiparametric fluorescence-activated cell sorting (FACS) Aria III analyzer.
CE P
TE
D
2.4 Flow cytometric analysis of captured HT29 in the presence of interfering HL-60 or RBCs A series of artificial CTC blood samples were prepared by mixing HT29 with a large population of either HL-60 cells, or RBCs as mentioned above. Cell mixture was treated with individual fluorescence-labeled conjugates (20 μg ml-1) as mentioned above, the capture efficiency of dual antibody conjugates was finally determined based on the % FITC+PE+ HT29 cells and the number of FITC+PE+ HT29 cells by the flow cytometry. The flow cytometric procedures were explicitly described in Supplementary information.
AC
2.5 Analyses of cell viability, cell cycle and apoptosis after capture The analyses were conducted as we described previously [26, 27]. Briefly, the cell viability of the captured HT29 was determined by the MTT {[3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrasodium bromide] tetrazolium salt)} assay after the cells were exposed to the purified conjugates (0, 5, 10, 20 μg ml-1) for 48 h (Tecan Infinite M200 pro). In parallel, the effects of the conjugates on the cell population distribution in every phase (G0/G1, S, and G2/M) and the cell apoptotic status were analyzed. The morphologic change of the suspensory cells was evaluated by using a fluorescence inverted microscope (Zeiss Axio Observer A1). 2.6 CTCs captured in vivo and in patient blood Human colon cancer HT29 cells were pre-labeled with red fluorescence protein (RFP) by using a lentiviral vector with an RFP transgene. The selected RFP-labeled HT29 cells with the stable fluorescence yield (>90%) were cultured and counted for the following capture assay. The PBS containing RFP-labeled HT29 cells (5×104) were intravenously injected into the nude mice. Ten-min later, the fluorescence-labeled conjugates G6-5E-PE, G6-5S-FITC, and PE-5E-G6-3S-FITC (80 μg per mouse; the dual antibody conjugate has a low molar concentration) were administered to the mice separately. One hour later, one-ml blood was individually withdrawn and treated with the RBCs lysis buffer. The isolated cells via centrifugation were then stained with Hoechst 33258 (25 μg ml-1) for imaging under a 5
ACCEPTED MANUSCRIPT
MA
NU
SC R
IP
T
fluorescence inverted microscope (Zeiss Axio Observer A1). Blood samples were drawn from eight colorectal cancer patients when they were under operation to remove the primary tumors. The study was reviewed and approved by the hospital institutional review board (IRB). Blood was collected into vacutainer tubes containing the anticoagulant EDTA. The single and dual antibody conjugates (20 μg ml-1) labeled with fluorescence (PE and/or FITC) were individually spiked into the patient blood (1 ml) pretreated with 1% BSA for 30 min. Blood without the conjugates was used as the control. After 1 h of incubation at 37 0C, the unbound conjugates were removed together with the lysed RBCs, and the captured cells and white blood cells (WBCs) were obtained via centrifugation. The obtained cells were blocked with 1% BSA at room temperature for another 30 min. Finally, APC-conjugated anti-CD45 was used to exclude the WBCs. After normalization of the fluorescence intensity with the corresponding isotype controls, CTCs captured by the single and dual antibody conjugates were quantified by the flow cytometry. After labeling the cell nuclei with DAPI dye (10 μg ml-1) in dark for 15 min, the captured CTCs were distinguished from the WBCs by the fluorescence images under a laser confocal microscope (Leica TCS SPE).
AC
CE P
TE
D
3. Results 3.1 Preparation of antibody-coated functionalized PAMAM dendrimers The surface modification and functionalization of G6 PAMAM dendrimers in the present study were mainly performed by starting carboxylation of the purified G6 PAMAM at their primary amine end groups with succinic anhydride (SA) (Fig. 1a). G6 PAMAM dendrimers (MW 62400 g mol-1, theoretically) were chosen not only because of their biocompatibility and biodegradability, but also owing to their number of surface functional groups (256) and large surface area to accommodate multiple aEpCAM and aSlex per dendrimer to enable more specific recognizing and binding HT29 cells by the biomimetic conjugates. The carboxylate-functionalized PAMAM surfaces were first activated using an 1:1 mixture of EDC and NHS for 1 h followed by the covalent conjugation with the aEpCAM and aSlex [19] (Fig. 1a). The conjugates were purified by dialysis in DDI water overnight and then lyophilized. To visualize the binding and capturing effects of the conjugates on the targeted HT29 cells, aEpCAM and aSlex were linked to PE (orange fluorescence) and FITC (green fluorescence), respectively, to exhibit high levels of fluorescence signal. The green fluorescence intensities from aSlex-FITC substantially increased, in a nonlinear fashion, with an increase in the amount of aSlex conjugated. The orange fluorescence from PE and the green fluorescence from FITC have little to no spectral overlap after normalization of the fluorescence intensities. The even distribution of the merged fluorescence indicated that the uniform coating of both aEpCAM-PE and aSlex-FITC was achieved. The aEpCAM and aSlex reacted with the completely-carboxylated dendrimers, and the resulted materials showed no absorption at 220 nm, the wavelength that can be used as an indicator of the presence of either aEpCAM or aSlex in preparation for the conjugates [13]. The preliminary conjugation experiments suggested that conjugation of the aEpCAM to the G6 PAMAM was facile, especially when the dual antibody conjugation was performed at molar ratios of CC G6: aSlex: aEpCAM in 1:3:3, or 1:3:5 (the optimal condition for the conjugation). The synthesized and purified conjugates are designated hereafter as G6-3E, G6-5E, G6-3S, G6-5S, G6-5E-3S and G6-5E-5S, respectively, based on the reaction molar ratios of CC G6 6
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
NU
SC R
IP
T
(G6) to aEpCAM (E) or aSlex (S) molecules. Fluorescence imaging was conducted using a laser scanning confocal microscope to monitor the conjugation process.
Fig. 1 Conceptual illustration and physicochemical characterization of re-engineered G6 PAMAM dendrimers coated with two antibodies targeting typical HT29 surface biomarkers (EpCAM and Slex). a, Chemical surface modification of G6 PAMAM and sequential link of aEpCAM and aSlex to the modified G6 PAMAM; Synthesis procedures of G6 PAMAM dendrimer derivatives and the related conjugates. 1, unmodified G6; 2, completely carboxylated G6 (CC G6); 3, partially carboxylated G6 (PC G6); 4, G6 single or dual antibody conjugates (G6-aEpCAM, e.g., G6-3E, G6-5E; G6-aSlex, e.g., G6-3S, G6-5S; and G6-aEpCAM-aSlex, e.g., G6-5E-3S, G6-5E-5S); 5, PC G6 conjugated with fluorescein; 7
ACCEPTED MANUSCRIPT
T
6, G6-antibody conjugates labeled with FITC; 7, CC G6 modified with fluorescence-labeled antibody (G6-5E-PE, G6-5S-FITC, PE-3E-G6-3S-FITC and PE-5E-G6-3S-FITC). b, FTIR spectra of CC G6 and its corresponding conjugates. c, SEM image of the dual antibody-coated conjugate G6-5E-5S. d, A typical AFM scan of the single antibody-coated conjugate G6-5E. e, Merged images of fluorescence-labeled G6-antibody conjugates in PBS (pH 7.4) solution obtained by a laser confocal microscope.
AC
CE P
TE
D
MA
NU
SC R
IP
3.2 Characterization of antibody-coated functionalized PAMAM dendrimers To measure the numbers of aEpCAM and aSlex conjugated to one dendrimer, a series of known concentrations of aEpCAM and aSlex were spiked into the running buffer to establish a standard curve for UV/Vis measurements of sample concentrations of the conjugates. Based on the linear curves of free aEpCAM and aSlex at 220 nm, the numbers of the conjugated aEpCAM and aSlex molecules per G6 PAMAM were calculated to be about 2 or 6 moles of aEpCAM or aSlex molecules conjugated to one G6 PAMAM (Supplementary Fig. 1) according to the initial amount of the aEpCAM or aSlex added. The numbers of E and S molecules in dual antibody conjugate could also be inferred and calculated based on the molar reaction ratios of the functionalized dendrimers (CC G6) to E and/or S by using the fixed amount of CC G6, E and/or S for the conjugation reaction. According to the UV analysis, the molecular weights of G6-5E, G6-3S and G6-5E-3S conjugates were estimated as 377 kDa, 924 kDa and 1240 kDa, respectively, based on the numbers of E and/or S molecules in each single or dual antibody conjugate. The molecular weight of G6 PAMAM dendrimers was known as 62.4 kDa, and that of CC G6 dendrimers was calculated to be 69.8 kDa. So when the same mass concentration unit of these conjugates (μg ml-1) was converted to the molar concentration (nM), the molar concentrations of G6-5E-3S were about 16-18-fold less than CC G6 dendrimers and 2-4-fold less than other conjugates. Surface-modified PAMAM derivatives were confirmed by 1H NMR and size/ zeta potential analyses (Supplementary Fig. 2). FTIR spectra showed increasing intensity of amide (-CONH-) IR signals when the surface functionalized PAMAM was linked to the increasing numbers of aEpCAM and aSlex (Fig. 1b). Various conjugated PAMAM dendrimers were separately eluded by high performance size exclusion chromatography. The corresponding chromatography showed distinct retention times for each individual derivative and conjugate. The size and morphology of the antibody-coated PAMAM conjugates were characterized by the SEM. Fig. 1c showed that the size of the dual antibody conjugate G6-5E-5S was less than 100 nm and in a pie shape. The tapping mode of AFM showed the characteristic morphology of G6 PAMAM conjugated with either aEpCAM or aSlex. The mean diameter measured by the AFM of CC G6, aSlex, G6-5E, G6-5S and G6-5E-5S was 6, 25, 35, 40 and 100 nm, respectively (Fig. 1d and Supplementary Table 1). The thickness of the conjugates in the assembly was 1-2 nm. Dynamic light scattering (DLS) measurement showed that the hydrodynamic diameter of the conjugates significantly increased when G6 PAMAM was coated with aEpCAM or aSlex, separately, or in combination (Supplementary Table 1). The increase in diameter of the G6 PAMAM as a result of the surface coating, and the uniform size and morphology of the conjugates (Fig. 1c, d) clearly indicated the formation of the conjugates. Zeta-potential measurement indicated the increased negative charges of G6 PAMAM after surface modification with –COOH group and functionalization with aSlex and aEpCAM, separately or in combination (Supplementary Table 1). The numbers of FITC molecule in 8
ACCEPTED MANUSCRIPT
MA
NU
SC R
IP
T
G6-COOH-FITC derivatives, aEpCAM in G6-aEpCAM and aSlex in G6-aSlex conjugates were estimated about 40, 2 and 6 molecules based on the corresponding intensity of fluorescence and UV absorption measured (Supplementary Fig. 1). The fluorescent intensity was enhanced with the increasing number of the conjugated antibodies. When aEpCAM and aSlex were separately conjugated to the same G6 PAMAM scaffold, the merged fluorescence exhibited the typical green-yellow color, indicating the successful conjugation (Fig. 1e; right two panels). The above data well characterized and confirmed the successful synthesis of the single and dual antibody-coated conjugates. Biostability of the conjugates was evaluated by measuring the changes in UV absorption of the conjugates at different pHs, temperatures and agitation rates. Under static condition and at pHs 5.6, 7.4 and 9.13 and temperatures 4 and 37 0C, it was found that those conjugates were most stable at pH 7.4. Both acidic and basic conditions expedited the degradation of the conjugates at room temperature. Individual conjugates were stable at 4 0C for as long as the tested 11 days without significant degradation. At 37 0C, some degradation was observed after 24 h incubation. Shaking the conjugates at velocity of 80, 120 and 210 rpm for as long as 30 min at 25 0C did not cause significant degradation of the conjugates.
AC
CE P
TE
D
3.3 Sensitivity and efficiency of capturing adherent and non-adherent CTCs To make sure that the HT29 cells express EpCAM and Slex, we quantitatively analyzed the expression levels of EpCAM and Slex on the HT29 cells under the same conditions (Supplementary Fig. 3). The result showed the overexpression of EpCAM and Slex on the surface of HT29 cells. Moreover, the result explained why the PC G6 and CC G6 without aEpCAM and/ or aSlex conjugated showed no noticeable binding to the HT29 cells. Whereas when the adherent HT29 cells were incubated with aEpCAM-PE and G6-5E-PE (orange), or aSlex-FITC and G6-3S-FITC (green), or PE-5E-G6-3S-FITC (green-yellow) for 1 h at 37 0C, the cells bound to these conjugates exhibited different degrees and colors of fluorescence that could be quantified by using the laser scanning confocal microscope (Fig. 2a). The mean fluorescence images have identical exposure times and normalization. The numbers of captured cells increased when the amounts of aEpCAM or aSlex coated on G6 increased, or when the concentrations of either single or dual antibody conjugates increased from 10 µg ml-1 to 20 µg ml-1. The metastatic ability of the adherent CTCs [15] may differ from that of characteristic non-adherent CTCs [28], where, the latter remain viable in the bloodstream after loss of attachment to endothelial membrane. We therefore assessed the capture specificity of the conjugates to the suspensory HT29 cells by using both fluorescence inverted microscope analysis and flow cytometry (Fig. 2b-d). The cells were stained with Hoechst 33258 (blue) to show each individual cell following incubation of the cells with 1% BSA in PBS for 30 min and mixed with fluorescence-labeled conjugates at 37 0C for 1 h. G6-5E-PE, G6-3S-FITC and PE-5E-G6-3S-FITC showed typical merged images upon binding to the stained cells (Fig. 2b). To determine the optimal incubation time, the non-adherent HT29 cells were cultured in the presence of the conjugates (20 µg ml-1) for 0.5-4 h. The conjugates bound the cells within 1 h, indicating the efficient binding period. Capture efficiency of the synthesized conjugates targeting HT29 cells was determined by using the FACS Aria III (Becton Dickinson) with laser excitation at λex 488 nm. Immunoglobulins labeled with the same fluorochromes were used as isotype controls to exclude the autofluorescence and 9
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
NU
SC R
IP
T
non-specific binding. Data acquisition was based on the threshold of 10,000 cells to meet the criteria of the light scatter. Dot plots were adjusted to show the number of fluorescence-labeled conjugates bound to HT29 cells after normalization for both FITC and PE dyes. A gating strategy was first focused on quantitatively distinguishing the conjugate-captured FITC+PE+ HT29 cells from the free cells. The capture efficiency (% positive cells) was defined as the number of the conjugate-captured FITC+PE+ HT29 cells divided by the number of the cells shown in the P1 gate. The quantitative flow cytometric analysis demonstrated that the dual conjugates PE-5E-G6-3S-FITC and PE-3E-G6-3S-FITC (both 20 µg ml-1) captured more suspensory cells in a cytometric tube than its single conjugate counterparts G6-5E-PE and G6-3S-FITC, or G6-3E-PE and G6-3S-FITC (both 20 µg ml-1) added together to the tube (Fig. 2c, d). Moreover, PE-5E-G6-3S-FITC captured more HT29 cells than its dual conjugate counterpart PE-3E-G6-3S-FITC (containing less aEpCAM), suggesting the importance of aEpCAM for capturing HT29 cells [12, 29-32] (Fig. 2c), which was consistent with what we found that the HT29 cells expressed EpCAM higher than Slex (Supplementary Fig. 3). The average number of the captured FITC+PE+ HT29 cells by PE-5E-G6-3S-FITC was greater than that by the individual single antibody conjugates G6-5E-PE and G6-3S-FITC added together (Fig. 2d). The above results suggest that the dual antibody-coated conjugates possess capture capability greater than their single conjugate counterparts, and the capture efficiency is at least EpCAM-dependent.
10
AC
CE P
TE
D
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
Fig. 2 Capability of aEpCAM- and aSlex-coated conjugates to recognize and capture adherent or non-adherent HT29 cells. a, Recognizing and binding of fluorescence-labeled antibody aEpCAM-PE (E-PE), aSlex-FITC (S-FITC), G6-antibody conjugates (G6-5E-PE, G6-3S-FITC, PE-5E-G6-3S-FITC) to adherent HT29 cells examined by using confocal microscope at FITC λex 488 nm, λem 500-535 nm, PE λex 550 nm, λem 570-610 nm, and DAPI λex 405 nm, λem 425-475 nm; b, to non-adherent HT29 cells examined by using fluorescent microscope at λex 470/40 nm, λem 525/50 nm and λex 365 nm, λem 445/50 nm. The confocal images have identical exposure conditions (1 h, 37 0C, 5% CO2). All the images were the merged pictures taken after DAPI, PE and/or FITC stain. The merged image colors distinguished the binding of HT29 cells to the single and dual antibody conjugates from the binding of HT29 to the antibody alone. c, Comparison of capture efficiency quantified by the FACS analysis between dual antibody-coated 11
ACCEPTED MANUSCRIPT
IP
T
conjugates (PE-E-G6-S-FITC) and their single antibody counterparts added together (G6-3E-PE+ G6-3S-FITC; or G6-5E-PE+ G6-3S-FITC) at the same concentration 20 µg ml-1 for capturing HT29. Note, the added concentration of the combined single antibody conjugates was therefore 40 µg ml-1; capture efficiency (%) was calculated based on the control. d, Dot plot analysis by gating plots of FITC+PE+ HT29 cells; the number in Q2 plot represents the positive cells for each sorting in the presence of the conjugates. S stands for aSlex, and E, aEpCAM, with the molecular number prefixed.
AC
CE P
TE
D
MA
NU
SC R
3.4 Capture specificity in the presence of interfering HL-60 or RBCs To evaluate the ability of the conjugates in recognizing and capturing CTCs, HT29 was used as a CTC model. The capture specificity, on the basis of our previous definition [33], was defined as the ability to assess unequivocally the CTCs captured in the presence of components (RBCs, WBCs) that may interrupt the quantitative measurement, while maintaining their own characteristics. The specificity of the conjugates binding to the surface of HT29 cells was examined by mixing a series of HT29 with either HL-60 cells (as a leukocyte model that lacks HT29 biomarker EpCAM [34-36]), or RBCs at a mixture ratio of 1:103 or 1:105 (HT29: HL-60; or HT29: RBCs) with fixed total 106 cells per incubation tube (Fig. 3a), or by mixing 1,000 HT29 cells with 106 or 108 RBCs per incubation tube (Fig. 3b, c). HL-60 cells usually have no interaction with aEpCAM [36]. The ratio mimicked the clinical situation in which, roughly 1 CTC presented in 10 3-106 leukocytes [37], 14-5,000 CTCs per milliliter blood [6], about 175 CTCs in 3 ml of blood [17], or 100 CTCs in 108 RBCs [38, 39]. We can regularly sort about 10 colorectal CTCs from 1 ml of patient blood. Under optimal gate conditions, the synthesized conjugates that recognized and captured the suspended HT29 could be visualized by fluorescence intensity of the flow cytometric images. In general, the conjugates recognized and bound HT29 in a cooperative manner in the presence of either HL-60 or RBCs, whereas, the isotype IgG control did not bind or capture the suspended HT29 cells. The dual antibody conjugate PE-5E-G6-3S-FITC captured 13.3% and 19.7% more HT29 cells than its single antibody conjugates G6-5E-PE and G6-5S-FITC in a solution where 1,000 HT29 cells were spiked into 1×10 6 RBCs (Fig. 3c). The numbers or percentage of bound and captured HT29 by those conjugates decreased when the mixed cell ratio of HT29 to HL-60 or RBCs changed from 1:103 to 1:105 (HL-60), or from 106 to 108 (RBCs) (Fig. 3a-c) because of the interfering effect. For example, the dual antibody conjugate captured less HT29 cells (8-10% less) when the total numbers of RBCs were increased to 108 cells (Fig. 3c). Similarly, when more RBCs were added to a fixed number of HT29 cells (103), dot plots showed a reduction in the number of the captured HT29 (Fig. 3b). These results collectively demonstrated that the dual antibody conjugates showed the significant advantages over single ones in terms of the capture specificity and capture capability.
12
AC
CE P
TE
D
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
Fig. 3 Specificity and capacity of single and dual antibody-coated G6 conjugates in capturing HT29 cancer cells in the presence of interfering HL-60 cells (a leukocyte 13
ACCEPTED MANUSCRIPT
NU
SC R
IP
T
model) and RBCs. a, Flow cytometric analysis of the total numbers of HT29 cells captured by the PE-FITC- fluorescence-labeled single or dual antibody-conjugates. HT29 cells were mixed with either HL-60 cells (left) or RBCs (right) at the ratio of 1:103 (HT29: HL-60, or RBCs; open bar), or 1:105 (filled bar) with the total 106 cells combined. b, Dot plots showed enhanced capture capacity when PE-3E-G6-3S-FITC was replaced with PE-5E-G6-3S-FITC conjugate. The HT29 cells (1,000) co-existed with 106 (left) or 108 (right) RBCs. The number shows the captured FITC+PE+ HT29 cells and indicates the importance of aEpCAM for the capture. c, Comparison of capture efficiency between dual antibody conjugates (PE-E-G6-S-FITC) and single ones in capturing HT29 cells (1,000) spiked to 106 (open bar) or 108 (filled bar) RBCs. Note, HT29 cells incubated with two immunoglobins labeled with PE and FITC were used as the isotype controls to normalize the fluorescence intensities. S stands for aSlex, and E, aEpCAM, with the molecular number prefixed. The significance analysis was performed on the basis of the PE-5E-G6-3S-FITC conjugate or control. Error bars denote the standard deviation of Student t test.
AC
CE P
TE
D
MA
3.5 Down-regulation of the captured CTCs and related mechanisms To explore whether the conjugates had an effect on cell viability, the conjugates were individually incubated with various cell lines for 48 h at concentrations ranging from 1.25-20 μg ml-1. The cell lines investigated included the human umbilical vein endothelial cells ECV-304, melanoma cells A-375, cervical cancer cells Hela, hepatocellular carcinoma cells HepG2 [26, 27] and HT29. These cell lines usually express lower levels of EpCAM and/or Slex than HT29 cells [21-23, 30, 40]. EpCAM and Slex expression levels of the five cell lines were also confirmed by using flow cytometry. In comparison to the control, the conjugates without the fluorescence-labeling produced a concentration-dependent reduction in cell viability when the conjugate concentrations increased from 5, 10 to 20 μg ml-1 (Table 1, 2). The conjugate G6-5E inhibited HT29 cells more significantly than other tested cell lines that usually express low EpCAM and/or Slex activities (Table 1). Further analysis demonstrated that both G6-5E and G6-5E-3S exhibited a significant concentration-dependent inhibition on HT29 cell viability. In contrast, CC G6 and G6-3S at molar concentration of five-fold excess over their conjugates did not show the significant effect (Table 2). The result further demonstrated the capture capability of the conjugates, and suggested that the interruption of EpCAM activity of HT29 cells by the aEpCAM-coated conjugates may be attributed to their inhibition on HT29. Table 1. Deactivation of the captured HT29 cells by G6-5E single antibody conjugate. The conjugate showed the effect on cell viability in a concentration dependent manner. However, it has low effects on other cell lines that express less EpCAM and/or Slex. Cell lines HT29 A-375 ECV-304 Hela HepG2
Cell viability (%) 10 μg ml-1 50.21±7.97## 75.89±7.64 77.00±8.17 82.60±5.77 82.57±5.81
-1
5 μg ml 55.20±6.37## 82.41±22.00 84.80±1.36 88.30±1.32 89.10±6.62 14
20 μg ml-1 33.80±6.71## 71.61±11.78 62.23±7.90## 70.81±7.04 78.43±1.94
ACCEPTED MANUSCRIPT #
P
