EpCAM aptamer activated 5-FU-loaded PLGA nanoparticles in CRC treatment; in vitro and in vivo study

Abstract In this study, epithelial cell adhesion molecule (EpCAM) aptamer-activated nanoparticles (Ap-NPs) were synthesised to enhance treatment efficiency in colorectal cancer (CRC). PLGA [poly(d, l-lactide-co-glycolide)] copolymer was fabricated by conjugation of COOH-PEG-NH2 to PLGA-COOH through an EDC/NHS-mediated chemistry. Afterwards, 5-fluorouracil-loaded (FU) nanoparticles were prepared using the water/oil/water double emulsion solvent evaporation method. The in vitro cytotoxicity of formulations was evaluated using the MTT assay in HCT-116, CT-26 and HEK-293 cell lines. For in vivo study, tumour-bearing BALB/c mice were established by subcutaneous injection of CT-26 cell line. The results indicated that fabricated AP-FU-NPs had 101 nm size with a spherical surface, relatively homogeneously and, satisfactory encapsulation efficiency (83.93%). In vitro experiments revealed that Ap-FU-NPs had a superior in vitro cytotoxicity than both FU-NPs and free 5-FU in CT-26 and HCT-116 cells but, were significantly low toxic against HEK-293 cells relative to free 5-FU. Furthermore, in vivo results showed no significant haemolytic effect, hepatic and renal injury, or weight loss. After treatment of various animal groups with formulations, notable tumour growth delay was observed following the order: Ap-FU-NPs < FU-NPs < 5-FU < PBS. The results suggest that AP-FU-NPs could be an effective and promising carrier for 5-FU delivery to the EpCAM overexpressing CRC cells. GRAPHICAL ABSTRACT The present study was carried out in the three steps summarised: (1) synthesis and characterisation. At first, PEG was conjugated to PLGA polymers by EDC/NHS chemistry method, then the resulting polymers were used to prepare PLGA-PEG NPs via the W/O/W technique. Finally, EpCAM Aptamers were attached to fabricated NPs. (2) In the second step cytotoxicity of formulated NPs was evaluated in vitro experiments in HEK-293, CT-26 and HCT-116 cell lines. (3) In the third step in vitro results further was investigated in CT-26 tumour-bearing BALB/c mice. ‘Created with BioRender.com’.


Introduction
Colorectal cancer (CRC) is the second and third (women and men, respectively) most frequent cancer leading to approximately 3,100,000 deaths worldwide [1,2]. Conventional treatments such as radiotherapy and/or chemotherapy are generally chosen in the early stage (non-metastatic CRC) whereas systemic chemotherapy is employed for patients with metastatic stage [3][4][5][6]. To ameliorate surgical outcomes and reduce tumour recurrence chances, surgical and adjuvant or neoadjuvant therapy are often used. Adjuvant therapy is defined as cytotoxic chemotherapies used alone or in combination with targeted therapies. One of the frequently used adjuvant chemotherapies in CRC treatment is 5-fluorouracil (5-FU), a synthetic fluorinated pyrimidine analog that inhibits thymidylate synthase leading to purine synthesis inhibition resulting in reduced DNA replication and repair, ultimately causing tumour cell growth inhibition [7][8][9]. The 5-FU is the first-line drug therapy in CRC treatment but, some shortages such as short plasma half-life (thus need for continuous infusion), patient resistance to 5-FU in some cases, water-soluble nature (slightly pass across cell membrane) and finally adverse effects (such as bone marrow myelosuppression, hair-loss, nausea and diarrhea) can seriously hamper the drug anti-tumour efficacy. Such challenges in clinical practice motivate researchers to design different strategies, which enhance 5-FU half-life and reduce adverse effects on normal cells. Hence 5-FU is a water-soluble drug and cannot pass through diffusion from cell membranes; therefore, specific carriers are needed to help cells uptake and improve the half-life of such drugs [10] so, entrapment in nanoparticles (NPs) could be the benefit choice in this regard. Many NPs have been used to develop the 5-FU formulation for improved anti-tumour efficacy such as gold nanoparticles [11], Ag 2 S quantum dots [12], chitosan nanoparticles [13], magnetic nanoparticles [14], hydroxyapatite-gelatin polymer composites [15], single-wall carbon nanotubes (SWCNT) [16], bacterial ghosts [17] and liposome nanoparticles [18]. Although several types of NPs have been previously employed to enhance the half-life of 5-FU, of these NPs PLGA [poly(d,l-lactide-co-glycolide)] presents many desirable properties such as biodegradability, biocompatibility, adjustability of size and shape, no inflammatory responses and possessing functionalised end groups that make it an attractive candidate for drug delivery systems [19,20]. The major disadvantage of the PLGA NPs to use in drug delivery is its hydrophobic surface leading to limit its cellular membrane permeability and rapid elimination from systemic circulation. This limitation can be overcome with surface modification such as PEGylating. Polyethylene glycol (PEG) has hydrophilic properties that has been widely used to increase the half-life of NPs in blood and prevent opsonisation by reticuloendothelial system [21][22][23].
Another strategy to improve drug delivery efficacy and reduce the side effect of chemotherapy is active targeting via conjugating tumour-specific ligands (such as antibodies or their fragments, peptides, aptamers, carbohydrates and vitamins such as folate, etc.) to the surface of NPs. El-Hammadi et al. showed that conjugation of the folic acid (FOL) to the PLGA-PEG nanoparticles leading to reduce IC 50 value of FOL-FU-PLGA nanoparticles approximately fourfold less than that of the FU-PLGA nanoparticles [24]. Similar result was reported about sulfadiazine (SUL) decorated 5-FU-PLGA [25]. Liu et al. fabricated cyclic arginine-glycine-aspartic peptide (cRGD)-modified 5-FU-loaded PLGA nanoparticles that have great potential drug delivery than non-targeted PLGA nanoparticles [26]. Recently, anticancer drug-loading NPs have been functionalised via monoclonal antibody (mAb) for enhancing the efficacy and reducing unwanted side effects of cytotoxic chemotherapies agents in cancer treatment. In vitro results of 5-FU-loaded nanoliposomes demonstrated that the 5-FU has higher cytotoxic activity when encapsulated in the anti-FZD10 mAb-decorated nanoliposomes as compared to the free 5-FU [27]. It was reported that PEGylated liposomes that was decorated with ITGB6 mAbs showed 1.5-fold higher cellular apoptosis rate than non-decorated liposomes in colon carcinoma [28]. In another report, Anti-EGFR mAb activated 5-FU-loaded PLGA nanoparticles showed superior drug delivery potency to EGFR positive CRC cells [29].
Aptamers are short single-stranded oligonucleotides that could be a proper candidate for active targeting because of the following advantages over antibody: (1) have a very broad spectrum of targets, (2) little or no immunogenicity, (3) do not increase nanoparticle size, and (4) relatively easy to make and to store, etc. [30]. The epithelial cell adhesion molecule (EpCAM) is an over-expressed protein in various solid tumours whereas low-expressed in a variety of healthy cells. Song et al. introduced an EpCAM aptamer (SYL3C) that could specifically and effectively bind to the surface of EpCAM-positive cells [31,32].
To the best of my knowledge it is not reported fabrication of 5-FU-SYL3C-PLGA in the CRC treatment, therefore, we hypothesised that engineering and fabricating an aptamer-functionalised PLGA drug delivery system might be able to enhance the stability and cellular uptake of 5-FU from the serum, decrease the toxicity to healthy cells or tissues and reduce various side effects. We then assessed our fabricated NPs drug delivery potency against CRC. The PEG polymer was conjugated to PLGA via EDC/NHS-mediated chemistry. The 5-FU-loaded NPs was made by water/oil/water (W/O/W) technique. After confirmation of formation of spherical NPs and their stability, HCT-116, CT-26 and HEK-293 cell lines were subjected to formulated NPs and their efficacy was evaluated. Furthermore, the effect of formulated NPs on tumour was investigated in CT-26 tumour-bearing BALB/c mice. All result together was confirmed that formulated NPs were able to delay and decrease tumour growth with lower serological and tissue damage side effects. We have concluded that our fabricated NPs can be used as suitable carrier for 5-FU delivery in CRC and further evaluation needs to be survey their efficacy in clinical trial.
The overall strategy of this study is summarised in Graphical Abstract.

Synthesis and characterisation of PLGA-b-PEG-COOH copolymer
Carboxylate-functionalised copolymer PLGA-b-PEG was accomplished by the conjugation of COOH-PEG-NH 2 to PLGA-COOH through an EDC/NHS-mediated chemistry. Briefly, PLGA (100 mg) was reacted with EDC (0.0275 mmol) and NHS (0.0275 mmol) in 5 mL of DCM and poured into the 50-mL round-bottom flask and attached to a reflux system to prepare PLGA-NHS, then gently stirred at room temperature (RT, 25 ± 2 °C) for 4 h. The activated polymer (PLGA-NHS) was precipitated with an anti-solvent solution (ice-cold methanol/diethyl ether 50:50, vol/vol) and subsequently washed three times with cold freezing solution (containing 80% diethyl ether and 20% methanol) to remove residual NHS and EDC. After drying, activated PLGA-NHS and HCl.NH2-PEG-COOH (3:1 PLGA:PEG molar ratio) were mixed and dissolved in 5 mL of chloroform in the presence of 0.2 mmol DIPEA (N,N-diisopropylethylamine). The resulting mixture was attached to the reflux system and gently stirred at RT for 24 h. The final obtained PLGA-PEG block copolymer was precipitated using 50 mL of anti-solvent solution and subsequently washed three times with cold freezing anti-solution. The final PLGA-PEG block copolymer was dried under the vacuum for 24 h and stored at -20 °C until use. The conjugated copolymer was confirmed by FTIR and 1 H-NMR spectra [33,34].

Formulation of 5-FU-loaded nanoparticles
Blank or 5-FU-loaded nanoparticles were prepared using the water-in-oil-in-water (W/O/W) double emulsion solvent evaporation method by slightly modifications to load the hydrophilic drug [35]. To prepare 5-FU-loaded nanoparticles, 30 mg of PLGA-b-PEG-COOH (was synthesised in previous step) and 70 mg of PLGA were dissolved in 5 mL of ethyl acetate containing Tween 20 (5 mg) as an organic phase surfactant (OPS), then 5-FU (10 mg/ mL) in 2 mL of PVA (0.5% vol/vol) as an aqueous phase surfactant (APS) was added to the polymer solution and sonicated at 30% amplitude for 6 min in an ice bath using a probe sonicator (BANDELIN SONOPULS, Germany) to perform first emulsion W/O. The second emulsion (W/O/W) was formed by adding the first one (W/O) dropwise into 18 mL of an aqueous solution containing 0.5% PVA and sonicated at 40% amplitude for 6 min in an ice bath. The formed double emulsion was continuously stirred overnight at RT to evaporate the organic solvent. The resulting 5-FU loaded NPs were centrifuged by Amicon® Ultra 15 mL (10 kDa cut-off; Ultrafree, MC Millipore, MA, USA) at 5000 rpm for 5 min to remove free drug and excess PVA. The supernatant and filtered liquid were collected for later experiments. All previous steps were performed without 5-FU for the synthesis of blank nanoparticles. To confirm the in vitro cellular uptake of nanoparticles, fluorescein isothiocyanate (FITC) was encapsulated into nanoparticles by a similar formulation protocol in dark.

Conjugation of EpCAM Ap to FU-loaded-PLGA-PEG nanoparticles
Aptamer conjugation to the surface of FU-NPs was carried out using EDC/NHS chemistry [36]. The resulting PEG-PLGA-FU NPs (10 mg/mL) in Section 2.2.2 were dispersed in nuclease-free water with excess EDC (200 mM) and NHS (100 mM) for 30 min at RT with gentle stirring. For removing the residual EDC and NHS, the obtained solution was repeatedly washed three times with Amicon Ultra 15 mL (Cut-off: 10 kDa) and nuclease-free water. Then, aptamers were denatured at 80 °C for 10 min and subsequently allowed to be renatured during cooling on the ice leading to the formation of the 3D binding conformation. Finally, the resulting activated NPs were incubated with 1 mL of 3′-NH2-Ap (300 μg/mL) for 4 h under gentle stirring. Then, the synthesised Ap-NPs were washed three times with nuclease-free water by ultrafiltration to take away residual aptamers and kept at 4 °C until use.

Confirmation of Ap-conjugation on the surface of nanoparticles
The Ap conjugation on the surface of NPs was confirmed by agarose gel electrophoresis following a previously reported protocol [37]. Samples were loaded on the gel as follows: naked NPs (as the negative control), free Ap (as the positive control), unwashed Ap-NPs, washed Ap-NPs, and DNA marker that were subjected to 2% (wt/vol) agarose gel electrophoresis in 1 M Tris-acetate-EDTA buffer. Samples were mixed with loading buffer and loaded onto agarose gel at 90 mA for 40 min and visualised under UV illumination. For further examination of Ap and NPs conjugation, the absorbance of naked NPs, Ap-NPs, and free Ap was measured at UV absorption of DNA at 260 nm (n = 3).

Aptamer conjugation efficiency
An ultrafiltration-based approach was used to evaluate the aptamer conjugation efficiency of the synthesised NPs. Briefly as mentioned in the Section 2.2.3, the aptamers were incubated with NPs, and the mixture of Ap/NPs was centrifuged with the Amicon Ultrafilter (MWCO 10 kDa) at 4,000 rpm for 30 min at RT to eliminate the unbounded aptamers. The unbounded aptamers were measured by comparing their absorption at 260 nm with the Ap calibration curve. Finally, the aptamer conjugation efficiency (ACE %) of the nanoparticles was indirectly calculated using Eq. (1)

ACE %=
Total amount of Ap added-unbounded Ap Total amount of Ap added ×1 100 (1)

Physicochemical characterisation of nanoparticles
Mean particle size distribution (diameter, nm SD), polydispersity index (PDI), and zeta potential (in PBS, pH 7.4) of blank NPs, FU-NPs, and Ap-FU-NPs were determined utilising dynamic light scattering (DLS) (Malvern Instruments, Zetasizer ver. 7.11) at RT. Furthermore, the particle shape and surface topography of the formulated NPs were characterised by FE-SEM (ZEISS SIGMA VP Field, Germany). For this purpose, the drug-loaded NPs were suspended in milli-Q water then were dropped on a thin 1 cm 2 glass slide and allowed to dry at RT to take the measurements.

Determination of encapsulation efficiency and drug loading
The encapsulation efficiency (EE%) of 5-FU into fabricated NPs was indirectly evaluated by determining the percentage of 5-FU entrapped in the NPs compared with the total 5-FU used for the formulation. For this purpose, NPs suspension was placed in the Amicon Ultra 15 mL (10 kDa Cut Off ) and centrifuged (Osterode am Harz, Germany) for 10 min at 14,000 × g. Next, the eluates were combined with the initial eluate containing non-entrapped 5-FU, and the resulting mixture was analysed using the spectrophotometer at 266 nm. Equations (2) and (3) were used to calculate encapsulation efficiency (EE%) and drug content (DC%) respectively (each test was done three times).

In vitro release of 5-FU from developed formulations
The releasing profile of 5-FU from fabricated NPs was studied using the dialysis diffusion method. For this purpose, 2 mL of each formulated NPs were transferred into a cellulose dialysis bag (MWCO, 12 kDa) and placed in two beaker containing 50 mL of recipient phosphate-buffered saline (PBS) at pH 7.4 (simulating physiologic condition) and pH 5.5 (simulating tumour acidic microenvironment) in 0.05 M respectively. The media was gently agitated at RT and 100 rpm to maintain homogeneity. At distinct time intervals (0.25, 0.5, 1, 2, 4, 6,12,24,48,72,96,120,144,168,192, and 216 h), 2 mL of released media was removed and replaced via the same volume of fresh. The amount of 5-FU in the released media was quantified by the UV-visible spectrophotometer (5-FU λ max = 266 nm).

Physical stability studies and drug leakage
To evaluate physical stability, the prepared NPs were dispersed in distilled water and kept for six months in microtubes (1.5 mL) at 4 ± 0.5 °C and protected from light. At regular time intervals (0, 1, 3, and 6 months), aliquots were withdrawn and assayed periodically for several key physicochemical properties including zeta potential, PDI, and the hydrodynamic diameter of NPs. The leakage of 5-FU from the formulated NPs was also evaluated by assessing the differences of EE% primary prepared NPs at the same regular time intervals.

Haemolysis assay
To investigate the blood compatibility profile of the synthesised NPs, the spectrophotometric determination method was selected. Briefly, about 10 mL of human blood sample (EDTA-tube) was collected from healthy volunteers and centrifuged for 10 min at 2500 rpm. Then, red blood cells (RBCs) were isolated from the plasma and washed three times with PBS (pH 7.4). A proper amount of washed RBCs (1% vol/vol) was mixed in 11 mL of PBS to make stock dispersion. In a sterile 6-well microtiter, 900 μL of formulated NPs were distributed then, 100 μL RBC suspension was added to all samples, then incubated at room temperature (RT) for 1 h. Triton X-100 (5%) was used as the positive control (100% haemolysis) and PBS as the negative control (0% haemolysis) compared with all samples. After incubation, samples were centrifuged at 2500 rpm for 10 min to eliminate the non-lysed RBCs. The absorbance of haemoglobin that was released in the supernatant at 540 nm refers to the percentage of the haemolytic activity [38,39]. Finally, the percentage of haemolytic activity was determined according to Eq. (4) HemolyticHaemolytic Activity (%)= AbsS-AbsN AbsP-AbsN 100 × (4) Abbreviation in Eq. (4) refers to AbsS, AbsN, and AbsP as the average absorbance of the sample, negative control, and positive control, respectively.

Specificity cellular binding of aptamers
To investigate the specifically targeting of the EpCAM on the cell surface by SYL3C aptamers, flow cytometry and fluorescent microscopy analyses were chosen. Murine CT-26, human HCT-116 and HEK-293 cell lines were selected for the analysis. Briefly, the cells were cultured in 5% CO 2 , 95% air humidified and 37 ± 5 °C conditions. After approximately 60% confluency, the cells were detached using trypsin (0.025% vol/vol) and seeded in a 6-well plate. After 24-h incubation, the cells were trypsinized and incubated on ice for 30 min. Then Cy-3-labeled SYL3C aptamer (200 μM) was added on cells in binding buffer (0.2 μM sterile filtered 0.1 M HEPES [pH 7.4], 1.4 M NaCl and 25 mM CaCl2 solution) in dark place. After 1-h incubation, the cells were washed (three times) with 1 mL of PBS and re-suspended in 0.5 mL of binding buffer containing 0.1% sodium azide. For assessment of the intensity of fluorescence as an index for the binding of Cy-3-labeled SYL3C aptamer to the EpCAM protein of cell surface, the flow cytometry technique was used. To further confirm that the aptamers could specifically bind to the EpCAM on the cells surface, the cells were seeded onto a 6-well plate and incubated for 24 h. Afterward Cy-3-labeled SYL3C aptamers were added to cells (200 μM) in a dark place. Fluorescence microscopy (Olympus Corp., Tokyo, Japan) was used to directly image the localisation of Cy-3-labeled Syl3c aptamer on the surface of cell lines. To further predict the produced NPs' stability, the Cy-3-labeled Ap-NPs were dispersed in distilled water and kept for six months in microtubes at 4 ± 0.5 °C and protected from light.

In vitro cellular uptake of Ap-conjugated PLGA nanoparticles
Fluorescence microscopy was used to directly visualise the in vitro cellular uptake of NPs and Ap-NPs. In summary, FITC was encapsulated with of NPs and Ap-NPs according to the Section 2.2.2 description. The cells were seeded in a 6-well plate and after 24 h, the culture medium was removed and replaced with the FBS-free medium containing free-FITC, FITC-NPs and Ap-FITC-NPs. After 4-h incubation, the cells were washed (three times) with PBS and then analysed by fluorescence microscopy.

In vitro cytotoxicity assay
The in vitro cytotoxicity of 5-FU was evaluated using the MTT assay. In summary, the cells were grown in 96-well microplates at a density of 4.0 × 10 3 cells/well/150 μL for HEK-293 and CT-26, and 6.0 × 10 3 cells/well/150 μL for HCT-116 cells. After 24-h of incubation (in 5% CO 2 at 37 °C), 50 μL of prepared serial dilution of 5-FU was added to per triplicate well and incubated for 24, 48 and 72 h. After the mentioned time intervals, 10 mL of fresh MTT reagent (5 mg/mL) was added and the plates were incubated for 4 h. After 4-h incubation, the MTT-containing media were removed and 200 μL of DMSO was added to each well and incubated at 37 °C for 15 min. The absorbance was detected at 570 nm using an ELx 800 microplate reader (BioTek). The same concentrations of media without 5-FU were used for control cells. After determination IC 50 of 5-FU, in vitro cytotoxicity of formulated NPs was analysed using the MTT assay as mentioned.

Animal experiments
The BALB/c mice (4-6 weeks old, weighing 17-21 g, female) were purchased from the Pasteur Institute of Iran (Tehran, Iran). The cells were cultured in a DMEM medium with 10% FBS, 50 IU/mL of penicillin and 50 mg/mL of streptomycin in 5% CO 2 at 37 °C. The tumour-bearing mice were established by subcutaneous injection of 1.0 × 10 6 viable suspensions of CT-26 cells into the left flank. Twelve days after inoculation, the mice that showed a distinct tumour 5-6 mm in diameter, were randomly divided into four groups (n = 5). Group 1 (as control) received PBS, whereas each of the animals in groups 2, 3, and 4 received 5-FU, FU-NPs and Ap-FU-NPs respectively. Animals thrice weekly received 25 mg 5-FU/kg per intraperitoneal injection for two weeks. Changes in tumour volume were monitored by using formula: where a is the largest diameter of tumour, b is the smallest diameter of tumour, and V is the total tumour volume (mm 3 ). The body weight change of each animal was monitored one day after each injection to evaluate the toxicity of different drug formulations. Two weeks after the final treatment with the various formulations, the mice were euthanized and the solid tumours were removed for comparison of treatment effects. Liver and kidney from each animal were also extracted and stored in a 4% formalin solution, stained using haematoxylin and eosin (H&E) for histopathological examination. Serum was collected for serological assessment [40]. The biochemical parameters including aspartate transaminase (AST) and alanine transaminase (ALT) were analysed using a colorimetric method. For analysis, 1.5 mL of heart blood was collected and transferred into blood collection tubes. Then the serum was isolated by centrifugation at 3000 × g for 15 min at RT. The AST and ALT activity were measured using a biochemical analyser (Hitachi 902 Automatic Analyzer; Hitachi, Japan) and data were recorded as U/L [41].

Statistical analysis
To compare the significance of differences, one-way ANOVA followed by Tukey's test was performed with p ≤ 0.05 as the significance level using SPSS 22.0. All experiments were repeated at least three times and analysed using GraphPad Prism 8.4.3 (IBM Corporation, NY, USA). The obtained results were expressed as the mean values ± standard deviation (SD).

Synthesis and characterisation of PLGA-b-PEG-COOH co-polymer
In the FTIR spectrum of pure PLGA polymer, a strong band located at 1759.28 cm −1 was attributed to the stretching of carbonyl groups (C = O). The peak at 2999 cm −1 belonged to the stretching of the -CH 3 group of LA, and the peak at 2956 cm −1 was due to the stretching of the -CH 2 group of GA. In the spectrum of pure PEG, the absorption peaks at 2885 cm −1 belonged to the stretching of the methylene group (-CH) in PEG. The peaks observed from 1182 to 1094 cm −1 are due to C-O and C-O-C of the stretching vibration of PEG respectively [42,43]. The peaks observed at 1650 cm −1 and 1530 cm −1 attributed to the amide I and amide II bands, respectively. Formation of the -CH stretching peak at 2854 cm −1 and the amide bonds at 1530-1650 cm −1 confirmed the conjugation of PEG in PLGA copolymers (Figure 1(a)) [44].
In Figure 1(b), the basic chemical structure of PLGA-PEG showed the characteristic chemical peak values for the -CH 3 , -CH 2 and -CH groups. The peaks at 1-2 ppm (-CH 3 ) and 5.2 ppm (-CH) were the references for the methyl groups of LA monomer of pure PLGA moiety. The multiple peaks at 4.8 ppm served as the references for GA monomer (-CH 2 ) of pure PLGA moiety. A peak at around 2.5 ppm corresponds to residual DMSO as solvent. In addition, a peak at 3.65 ppm due to methylene groups as the characteristic peak of PEG confirms the successful formation of PLGA-PEG copolymer [44][45][46].

Characterisation of the fabricated nanoparticles
The zeta potentials, PDI, mean hydrodynamic diameter and size distribution of the formulated NPs were measured via dynamic light scattering at the ambient temperature given in Table 1. Also, Figure 1 and the corresponding zeta potentials were -7.84 mV, -9.06 mV and -13.9 mV in deionised water, respectively. The surface morphology and size of NPs were evaluated using an FE-SEM micrograph (Figure 1(d)). The results of DLS studies are in agreement with the FE-SEM micrograph which depicted the nanometric size of NPs with a spherical surface and relatively homogeneously distribution around at 100-200 nm range. The spherical shape of NPs showed surface energy minimisation and no aggregation occurring during their synthesis. Physicochemical measurements revealed no significant change in the size, PDI, zeta potential and surface charge of PLGA-PEG and FU-PLGA-PEG NPs. Also compared with PLGA-PEG and FU-PLGA-PEG NPs, there was only a small difference in the size of AP-FU-PLGA-PEG nanoparticles. The mean size of all formulated NPs in this study was less than 200 nm thus, would be suitable for internalisation into targeted cells. It was suggested that cells preferentially uptake NPs with a diameter of less than 200 nm [47]. Very small NPs in size could be rapidly eliminated from the blood by renal clearance (<5 nm) or quick liver uptake (10-20 nm) on the other hand, large NPs could be removed in the spleen (>200 nm) or cleared by the reticular endothelial system [48]. Some studies suggest that NPs with small size could avoid clearing by phagocytes in plasma, which is an advantage in drug delivery [49]. In our experiment, the zeta potential of all formulated NPs was negatively charged that attribute to anionic possibilities due to the presence of the PLGA-PEG chain in NPs. As reported in the literature, negatively charged NPs could be significantly accumulated in cells than positively charged NPs [50] and escape from elimination by the reticular endothelial system leading to avoiding NPs aggregation and enhancing the stability of the NPs [51]. Other additional advantages of negatively charged NPs compared to positively charged are biocompatibility, lower induction of inflammation, lower stimulation of the T-cell proliferation and cytokine production [52,53], and less damage to the erythrocyte membranes [54].

Encapsulation efficiency and drug loading
The EE (%) and DL (%) of 5-FU to the PLGA NPs are summarised in Table 1. The optimised formulated NPs showed a higher percentage of drug encapsulation efficacy and no significant differences between FU-PLGA-PEG (84.21 ± 3.29%) and AP-FU-PLGA-PEG (83.93 ± 3.67%) NPs. These data suggest that the activation of nanoparticles with aptamers does not significantly affect the encapsulation efficiency or drug loading capacity of the NPs. The EE (%) value in this research is in agreement with average entrapment values in a variety of literature [55,56]. Drug content values are a determining factor in release kinetics. As shown in Table 1, the corresponding percentages of drug content are (7.17 ± 4.76 and 6.83 ± 3.06%) for FU-PLGA-PEG and AP-FU-PLGA-PEG, respectively. We obtained a drug content value of roughly 7.0%, which is relatively lower than some reports. By decreasing particle size, the surface area of NPs was increased leading to an entrapment volume decrease [29]. Although drug content values were relatively low, we preferred small particle size (<200 nm) due to the cytotoxicity of nanoparticles depending on their size [57]. As mentioned previously, larger particles are more likely to be filtrated via the renal and immune system. On the other hand, the relatively low loading capacity of NPs in this research might be due to the hydrophilic nature of 5-FU which may accelerate the fast diffusion of the 5-FU to the aqueous phase during preparation [58].

Physical stability studies and drug leakage
The physical stability of long-term stored NPs is a pivotal parameter in drug delivery until their clinical application [59]. Acquired data from the long-term storage stability of prepared NPs are presented in Table 1. The data suggest that no significant differences were observed in physiochemical properties such as zeta potential, PDI, hydrodynamic diameter and EE% of NPs. These findings show enough physical stability of NPs during six months of storage at the stability conditions of 4 ± 0.5 °C and protection from light.

In vitro release of 5-FU from developed formulations
To investigate correlations between in vitro and in vivo release of formulated NPs, drug release experiments were performed by using two different release media including PBS (pH 7.4, simulating blood environment, Figure 2) and an acidic medium (pH 5.5, simulating tumour microenvironment, Figure 2). Using the 5-FU drug release data in two pH media, all 5-FU loaded NPs showed a biphasic release pattern including an initial burst, followed by a slower and prolonged sustained release. In acidic medium (pH 5.5) as well as physiologic medium (pH 7.4) 5-FU-loaded NPs showed a biphasic release pattern including an initial burst, followed by a slower and prolonged sustained release. The initial burst release profile might happen because of the drug locating on or near the surface of the NPs therefore, available for immediate release. Slower release profiles showed a plateau for a certain period, due to only diffusion of entrapped drugs from the core of the NPs matrix and prolonged sustained release profiles due to diffusion of the drug through the NPs during its degradation. As shown in Figure 3, all the NPs showed a similar drug release profile, approximately 20% and 25% of 5-FU was released from the NPs within 1 h at pH 7.4 and 5.5 values respectively (burst release). Within 48 h approximately 52% and 56% of the drug were released at pH 7.4 and 5.5 values respectively (slower release) finally, about 78%-89% of the drug was released over the 9-day experiment. This obtained data and previous literature reports indicate that 5-FU encapsulated PLGA NPs show a sustained release pattern in each two pH values at different time intervals [60,61]. In addition, 5-FU-loaded NPs exhibited a higher release rate at pH 5.5 than pH 7.4 which might be due to the improved solubility of protonated NPs in water thereby, accelerating NPs hydrolysis by acidic pH. It was reported that most of the vesicles may be in the second row, 1 day represents measurement data after formulation and 6 months represent the measurement data after 6 months of storage at the stability conditions of 4 ± 0.5 °C.
hydrolysed in the endolysosomal environment (i.e. pH 5.5 at 37 °C) thus, could result in enhanced release rate of drugs [62]. Recent investigations about the metabolism of cancer cells hint that cancer cells choose aerobic glycolysis as the main mode of glucose metabolism. Cancer cells, therefore, increase glucose uptake and secretion of more lactic acids to the microenvironment leading to acidification of the microenvironment [63].

Confirmation of Ap-conjugation and Ap-conjugation efficiency
Agarose gel electrophoresis was applied to determine the conjugation of aptamers to the NPs surface. As demonstrated in Figure 3(a), no band was observed for naked-NPs (line 4) and a bright band was shown in line 5 representing the free aptamer movement on the gel. Since NPs size are large, they are not able to move on the gel therefore, the band observed at the loading site of washed Ap-NPs (line 2) suggested successful conjugation of aptamer on the surface of NPs. A low light band was observed in line 3 that showed unbounded aptamer in unwashed NPs sample. After washing by ultrafiltration, no residual aptamer was observed at the loading site. As expected, the detected absorbance of Ap at 260 nm strongly suggested an effective aptamer conjugation (Figure 3(b)). According to Eq. (1) results, the conjugation efficiency of the aptamer was 77.63%, which indicates most of the aptamers were decorated on the surface of NPs. In the recent studies, conjugation efficiency of AS1411 aptamer to the PLGA NPs was reported over 90% [64], also conjugation efficiency of MUC1 aptamer to the mesoporous silica NPs was reported as 18.5% [65] indicating our successful aptamer conjugation.

In vitro cellular targeting and cellular uptake of fluorescent NPs
The flow cytometry results showed that the fluorescent intensities of HCT-116 and CT-26 cells were higher than that of HEK-293 cells. Furthermore, the HCT-116 cell line demonstrated a higher binding to the EpCAM aptamer than the CT-26 cell line. The results indicated that the Ap-PLGA NPs were specifically recognised as EpCAM-positive tumour cells (Supplementary Figure S2). Consistent with previous reports, the SYL3C aptamer can be specifically recognised and attach to EpCAM protein with high affinity while no obvious aptamer interaction has appeared in the HEK-293 cells [66][67][68]. To gain more insight into the targeting ability of Ap-FU-NPs, the interaction between Ap-FU-NPs and HEK-293, CT-26 and HCT-116 cells was directly visualised by fluorescence microscopy. As shown in Figure 4(a) after incubating Cy3-Apt-NPs with CT-26 and HCT-116 cells for 1 h at RT, a clear red fluorescence signal was observed around at cell membrane, because of the high expression of EpCAM on the cell surface and ligand-receptor interaction [69,70]. To further ensure the red fluorescence response was indeed due to the interaction of Ap-FU-NPs and EpCAM on the cell surface, 30 min before the addition of Ap-FU-NPs, excess quantities of free SYL3C aptamer, as a competitive ligand, was incubated with CT-26 and HCT-116 cells to completely block the active sites. As expected, almost no fluorescence signal was observed with these blocked cells (competitive ligand data not shown). Moreover, the HEK-293 cell line with low expression of EpCAM protein, as a negative control [71,72], was employed to further confirm the efficacy and specificity of Ap-FU-NPs for targeted delivery. Almost no fluorescence signal was observed on the membranes of HEK-293 cells indicating no internalisation of Ap-FU-NPs into HEK-293 cells due to very low expression of EpCAM aptamer target on the surfaces of these cells  agarose gel electrophoresis (2% wt/vol); since nPs size are large, they are not able to move on the gel therefore, the band observed at the loading site (line 2) confirms successful conjugation of aptamer on the surface of nPs: B: uV spectrophotometer; comparing the total added aptamers to unbounded aptamers indicate that most of the aptamers were successfully decorated on the surface of nPs. Data represent mean ± SD (n = 3; *p < 0.05; **p < 0.01; ***p < 0.001, ns: no significant).
which was in consistent with the results obtained by several studies [67]. Furthermore, all of these experiments were carried out with FITC-loaded NPs, to explore the cellular uptake in vitro. As shown in Figure 4(b) fluorescence intensity of CT-26 and HCT116 cells incubated with Ap-FITC-NPs was noticeably stronger than that of naked FITC-NPs. Also, naked FITC-NPs demonstrated greater cellular uptake than the free FITC (Supplementary Figure S1). Noticing Figure 4(a,b) revealed that fluorescent intensity is low in the CT-26 cells relative to HCT-116 cell lines. This fact might be due to low EpCAM expression on the CT-26 cells [73]. However, in HEK-293 cells incubated with Ap-FITC-NPs fluorescence intensity was similar to that of FITC-NPs. These observations indicated that attachment of aptamer to their special receptors on the cell surface considerably enhanced cellular uptake of aptamer-targeted NPs; therefore, could specifically increase drug delivery to EpCAM overexpressing living cells and could be used as an effective targeting agent.

Evaluation of in vitro cytotoxicity
After ensuring the ability of fabricated NPs to target cells and release 5-FU into them, the anti-proliferation activity of free 5-FU, blank NPs, free EpCAM, FU-NPs and Ap-FU-NPs was evaluated using an MTT assay. As revealed in Figure 5 Table 2, by increasing the treatment time interval, IC 50 value of CT-26 and HCT-116 cell lines were decreased, whereas the reduction in IC 50 value of HEK-293 cells was lower than cancerous cells. Also, Ap-FU-NPs and FU-NPs were significantly low toxic against HEK-293 cells relative to free 5-FU. Since the HEK-293 cell lines were unable to actively uptake Ap-FU-NPs, the observed low cytotoxicity of Ap-FU-NPs to HEK-293 cells was more likely due to the uptake of released 5-FU after the dissociation from NPs [74]. Furthermore, we can conclude from Figure 5(a-c) and Table  2 that the cellular viability exhibits a dose-dependent and time-dependent effect, especially for Ap-FU-NPs. These results suggest that the presence of Ap could considerably elevate the cytotoxicity of Ap-FU-NPs compared to both FU-NPs and free 5-FU therefore, aptamers may have an outstanding targeting advantage on the surface of NPs. Hence, free 5-FU has a short half-life time, therefore, during administration the 5-FU efficacy is lost very quickly. By encapsulation of 5-FU into NPs such as PLGA and targeting, NPs can first attach to the cell's surface and then achieve a sustained and delayed drug release [75]. Such a targeting drug delivery system showed the lowest IC 50 , which could significantly reduce the amount of dosage from the system.

Side effect of formulation NPs
Toxicity is a critical parameter in drug delivery; thus, in vivo safety evaluations such as haematoxylin-eosin (H&E) staining, body weight, haemolytic percent and blood biochemistry analysis were  also performed to ensure the safe application of formulated NPs in vivo.

Haemolysis assay
It is necessary to know NPs behaviour with blood and haemolysis assay is a frequently used indicator for in vivo toxicity studies. The study results indicated that free 5-FU leads up to 28.45% haemolytic toxicity in comparison to the negative control ( Figure 5(d)). FU-NPs and Ap-FU-NPs showed haemolytic toxicity of less than 5.0% which is generally regarded as a critical limit for haemolytic toxicity [76,77]. Therefore, formulated NPs found to be harmless to RBC membrane integrity. This may be attributed to the encapsulation of drug molecules in the NPs and subsequent prolonged release.

Serology assay
The liver is the largest detoxification organ and is frequently exposed to damage via different harmful factors such as pathogens or chemicals. By liver damage, hepatic function markers ALT and AST levels in the blood are elevated and can be sensitive indicators for such hepatic injury [78]. The mice treated with free 5-FU had elevated ALT and AST levels compared to the PBS treatment control group, confirming early signs of damage in the liver and kidney (Supplementary Figure S3). It was reported that administration of 5-FU alone dramatically elevated serum levels of ALT and AST but the combined treatment or nanoencapsulation of 5-FU significantly decreased liver tissue injury, and levels of ALT and AST [79]. In contrast, the mice treated with formulated NPs resulted in similar levels of ALT and AST compared to the control group, indicating that 5-FU loaded-NPs did not cause any serious damage to the mice's liver and kidneys. Therefore, Ap-FU-PLGA and FU-PLGA NPs can be safe therapeutic carriers for tumour treatment. Similar results were shown in previous studies [76,80,81].

Histological study
In the initial dose administration time (12 days), no significant changes in the tumour volume were observed among the groups but, after only 2 doses (day 16) notable tumour growth delay were observed following the order: Ap-FU-NPs < FU-NPs < 5-FU < PBS (Figure 6(a,b)). At the end time of the experiment, the tumour volume of the groups followed the same order: Ap-FU-NPs < FU-NPs < 5-FU < PBS. The body weight of animals was continuously recorded during the experiment to follow adverse effects. The results of Figure 6(d) indicate that in all groups at day 12-18 interval, body weight was increased slightly without significant difference among the groups. At the end of the study, significant body weight loss was observed in the free 5-FU receiving group. Interestingly, the volume of the Ap-FU-NPs administrating group was significantly decreased as compared to the respective initial dose receiving level (day 12; 45.84 mm 3 , p < 0.05). Tumour weight in all of the groups shows a similar change pattern with respective tumour volume in each group (Figure 6(c)). The result of both tumour growth inhibition and the body weight suggest that Ap conjugated PLGA NPs could be an effective and promising carrier for 5-FU delivery to the EpCAM overexpressing colorectal cancer . results of the a-C section indicated that aP-fu-nPs show a superior in vitro cytotoxicity than both fu-nPs and free 5-fu. also, aP-fu-nPs have low toxicity to normal cells compared to cancerous cell lines. also, free aptamers and blank nPs showed no significant cytotoxicity. D: haemolytic toxicity assay of formulated nPs; results of the haemolytic assay showed no significant haemolytic toxicity in red blood cells treated by fu-nPs and aP-fu-nPs compared to free 5-fu. Data represent mean ± SD (n = 3; *p < 0.05; **p < 0.01; ***p < 0.001, ns: no significant).
cells. Although CT-26 cells express a low level of EpCAM on their surface relative to HCT-116 cells, in vitro and in vivo results of this study showed effective tumour growth delay by exposure to FU-NPs and Ap-FU-NPs. The obtained results together in this research suggest that uptake of Ap-FU-NPs was more efficient than FU-NPs therefore, Ap-FU-NPs present high uptake of fewer drug doses with lower side effects and more anti-tumour efficacy. The histological examination of H&E stained slides of the major organs such as liver and kidney showed that the formulated NPs receiver mice had fewer histopathological lesions compared to the mice that received the free 5-FU. As shown in Figure 7, the histopathological result from the liver sections of mice treated with free 5-FU indicated cell nuclei being hyperchromatic, deformed and heterogeneous. In contrast, no obvious hepatic tissue side effect was observed in animals treated with formulated NPs. The kidney histopathological results also showed significantly less severe alterations except in the 5-FU treatment group, in which all the structures of the glomerulus and renal tubule were normal. In mice treated with 5-FU, the kidney had mild tubular necrosis, glomerular cell proliferation, thickened basement membrane and fusion of glomerulus and renal capsule. Since 5-FU itself is greatly toxic, nanotechnology can target and control drug release therefore remarkably improve the therapeutic effect of drugs and finally minimise drug side effects. In this regard, a study was reported that exosome-entrapped 5-FU has an effective anti-tumour effect with fewer systemic toxic and vital organ damage compared to free 5-FU administrated BALB/c mice [82].
The results of the another study showed that exploiting the ligands-functionalised micelle NPs leading to greater targeting efficiency of HCC, higher tumour growth inhibition, minimum liver and kidney histopathological injury and insignificant body loss compared to non-targeted group (Combined Ultrasound Treatment with Transferrin-Coupled Nanoparticles Improves Active Targeting of 4T1Mammary Carcinoma Cells).

Conclusion
In summary, we successfully developed an engineered NPs delivery system to simultaneously target and deliver chemotherapeutic drug (5-FU) to the CT-26 and HCT-116 colorectal cancer cells leading to lower adverse side effects with higher tumour suppression efficacy than the free drug and non-targeted drug-NPs. It has been demonstrated that the fabricated nanoparticles had desirable characteristics including negative surface charge higher stability and uniformity and spherical shape. Also, in vitro release data showed fabricated NPs deliver a higher quantity of 5-FU in acidic pH (simulating tumour microenvironment) and could prolong the half-life of the drug. Furthermore, H&E analysis did not show notable damage to the liver, or kidney in the NPs receiving groups also, did not observe haemolytic effect, ALT and AST change level or body weight loss during antitumor therapy at the tested doses suggesting that in vivo results of the present study confirm the in vitro results. The obtained results exhibited that the Ap-FU-NPs had a higher anticancer potency and effective cellular targeting specificity towards colon cancer cells, and lower side effects against normal cells. Therefore, the Ap-FU-NPs may be suitably used as vehicles for targeting colorectal cancer cells and deliver 5-FU into them.

Funding
This study was financially supported by the