Magnetic thermoresponsive nanocomposite for targeted PARP-1 in colorectal adenocarcinoma, an approach for tumor dual therapy

Background: The current study presents a bimodal therapeutic platform for cancer treatment. Bimodal implies that the presented drug loaded core-shell structure is capable of elevating the tumor tissue temperature (hyperthermia) through the superparamagnetic iron oxide core and simultaneously release a Poly (ADP-ribose) polymerase-1(PARP-1)-modifying agent from the thermoresponsive shell. The capability of Superparamagnetic iron oxide nanoparticles (SPIONs) as successful hyperthermia agents is well established. Likewise, poly-n-isopropylacrylamide (PNIPAAm) is a widely used thermoresponsive polymer. Together, they constitute the magnetic thermoresponsive nanocomposite (MTN). To the authors knowledge, the combination of magnetic nanocomposites with PARP-1 modifying agents has not been investigated. Therefore, in this work, 5-aminoisoquinoline (5-AIQ) is loaded on the thermoresponsive polymer to constitute MTN.5-AIQ. Results: Structural characterization of the formed composite is studied via various experimental tools. The lower critical solution temperature (LCST) is determined by differential scanning calorimetry (DSC) method. The results confirm the formation of magnetic thermoresponstive nanocomposite (MTN) with excellent potential for hyperthermia. A high drug loading efficiency (85.72%) is obtained with convenient temperature dependent drug release kinetics. Biocompatibility and cytotoxic efficacy are tested on an in vivo and in vitro colorectal-adenocarcinoma models, respectively. MTN.5-AIQ exhibits normal hepatic and as well as lower toxic effect on normal tissue. In addition, the composite Caco-2 cells Conclusions: Based on the obtained results, the proposed therapeutic platform can be considered as a novel, promising candidate as dual therapy for colorectal adenocarcinoma exhibiting a PARP-1 overexpression.


Conclusions:
Based on the obtained results, the proposed therapeutic platform can be considered as a novel, promising candidate as dual therapy for colorectal adenocarcinoma exhibiting a PARP-1 overexpression.

Background
Enormous interest has been focused on developing smart delivery vehicles capable of targeting and controlling the release of chemotherapeutic agents. Controlled drug release is very attractive as it overcomes various drawbacks of conventional chemotherapy. Limitations of using chemotherapy with subsequent adverse outcomes on healthy tissues can be attributed to incompatible pharmacokinetics as well as narrow therapeutic indices (Senapati, Mahanta, Kumar, & Maiti, 2018), (Cho, Wang, Nie, Chen, & Shin, 2008). Recently, the utilization of smart polymers showed great potential for controlling drug release when subjected to external stimuli such as temperature, PH, electric/magnetic field, light, etc. (Kost & Langer, 2012), (Anderson, Burdick, & Langer, 2004), (Xiong et al., 2011), (Priya James, John, Alex, & Anoop, 2014). Among the various thermoresponsive polymers, poly-n-isopropylacrylamide (PNIPAAm) is the most extensively used candidate. This paper reports the design of a core-shell drug loaded platform, in which SPIONs constitutes the core that acts as the heat source, and PNIPAAm acts as the thermoresponsive shell, together they constitute the core shell magnetic thermoresponsive nanocomposite (MTN). 5aminoisoquinoline (5-AIQ) acts as a PARP inhibitor loaded on MTN. This drug loaded magnetic thermoresponsive nanocomposite (MTN.5-AIQ) can be intravenously injected and targeted to the tumor site by a magnetic field gradient. An alternating magnetic field (AMF) then triggers heat production by the SPION cores. which concequently triggers drug release. The elevated temperature along with the chemotherapeutic agent are awaited to play synergistic roles in tumor treatment.
The thermoresponsive Poly(N-isopropylacrylamide) (PNIPAAm) is frequently used for drug delivery owing to its low critical solution temperature (LCST ≈ 32 C); close to the human body temperature(Priya James et al., 2014), (Ward & Georgiou, 2011). This polymer exhibits coilglobule phase transition when the temperature changes around the LCST of the polymer. At temperatures below its LCST, PNIPAAm becomes hydrophilic as it swells and extends in virtue of the intermolecular hydrogen bonds formed between polymer chains and water molecules. In contrast, above LCST, the polymer chain collapses, releases much of the water content and eventually shrinks. These reversible shrinking and water release of PNIPAAm are exploited for drug loading and release (Sun et al., 2004), (Schmaljohann, 2006).
SPIONs act as tiny antennae that generate heat upon exposure to alternating magnetic field (AMF) (Salimi, M., Sarkar, S., Saber, R. et al. (2018)). They exhibit superparamagnetic behavior at room temperature; they can be magnetized by applying magnetic field and the effect is abolished once the applied field is ceased (Gould, 2006). Heat dissipation from these superparamagnetic nanoparticles is caused by two relaxation mechanisms; either by the rotation of the moments within the particle (Néel relaxation) or by the physical rotation of the particle itself (Brownian or viscous loss) (Abenojar, Wickramasinghe, Bas-Concepcion, & Samia, 2016) upon exposure to AMF. Néel (τN) and Brownian (τB) magnetic relaxation times of a particle are given by Equation 1 and Equation 2 respectively; where τN is the Néel the relaxation time, τB is the Brownian relaxation time, Vm is the volume of the magnetic core, K is the anisotropy constant, kB is the Boltzmann constant, T is the absolute temperature, ° is an attempt frequency for changes in the dipole direction, η is the viscosity of the carrier fluid and Vhyd is the hydrodynamic volume of the particle.
PARP-1 belongs to the superfamily of more than seventeen enzymes that catalyzes the transfer of ADP-ribose units from its substrate NAD + to several protein acceptors as single-strand break repair and base excision repair factors, which contribute to DNA repair (Watson, Whish, & Threadgill, 1998), (Ba & Garg, 2011). In general, PARP inhibitors act through competitive blocking of the NAD + binding domain of the enzyme (Virág & Szabó, 2002); the effect which inhibits the repair of damaged DNA and facilitates apoptosis-dependent death of tumor cells (Threadgill, 2015). 5aminoisoquinoline (5-AIQ) is an active, water soluble inhibitor of Poly (ADP-ribose) polymerase-1 (PARP-1) and adjunctly promotes radiotherapy and chemotherapy of various cancer types (Vinod, Chandra, & Sharma, 2010), (Cuzzocrea et al., 2002). The elevated mortality associated with metastatic colon cancer is attributed to further development of resistant microenvironments towards current drugs. Latest chemotherapeutic protocols for colon cancer include DNAmodifying pharmaceuticals, like oxaliplatin or irinotecan, combined with 5-fluorouracil (Raftery & Goldberg, 2010). Such treatment regimens are initially successful in numerous patients, but ultimately, the majority of them become resistant. A significant interpretation of this resistance in colon cancer is the frequent DNA repair mechanism-linked to high expression of PARPs enzymes (Nosho et al., 2006), (Sulzyc-Bielicka et al., 2012). Previous studies investigated the use of commercially available PARP-1 inhibitors like Olaparib alone or in combination with the chemotherapeutic agents was accepted as a promising candidate for treatment of colon cancer (Davidson, Wang, Aloyz, & Panasci, 2013), (Augustine, Maitra, Zhang, Nayak, & Goel, 2019).

Structure and morphology
The X-ray diffraction (XRD) pattern of the iron oxide core is shown in Figure 1A. The obtained pattern shows a face centered cubic (FCC) structure, the peaks positioned at 2θ= 29. 942 , 35.386 , 43.123 , 57.019 ̊ and 62.581̊ correspond to reflections from the planes; (220), (311), (400), (511) and (440) respectively (Wang, Wei, & Qu, 2013). These positions and their relative intensities are consistent with ICDD 89 0688 card . This confirms that the synthesized particles are single phase magnetite or maghemite. The calculated average crystallite for the obtained SPIONs size is 12.7 nm and the average lattice parameter is 8.39Å.
The Fourier transform infrared spectroscopy (FTIR) spectrum of the magnetite (Fe3O4) is shown in Figure 1B (lower plot in black). The Fe3O4 sample exhibits one intense peak at 550 cm -1 ; due to stretching vibration mode associated with the metal-oxygen absorption band (Fe-O bonds) in the Fe3O4 sample (Lesiak et al., 2019).There are less intense peaks at 1632 and 865 cm -1 resulting from the bending vibration of the O-H in plane and out of plane bonds of water respectively (Lopez, González, Bonilla, Zambrano, & Gómez, 2010) in addition to a broad peak centered at 3377 cm -1 due to the O-H bond stretching vibrations of water (Coates, 2004). FTIR spectrum of SPION-PNIPAM sample is also shown in Figure 1B (upper plot in red). The spectrum illustrates the characteristic peaks of Fe3O4 and PNIPAAm. The peak at 550 cm -1 represents Fe-O bond with no observed shift because the SPION-PNIPAAm is prepared after the magnetite is synthesized (post synthesis polymerization). The characteristic peaks at 1545cm -1 and 1650 cm -1 represent the amide II (N-H bending) and amide І (C=O stretching) vibrational mode found in NIPAAm respectively, the peaks at 1365 and 1395 cm -1 arise from bending vibration of isopropyl group -CH(CH3)2 found in NIPAAm, the peak at 3300 cm -1 corresponds to secondary N-H amide symmetric stretching vibration (Chou, Lai, Shih, Tsai, & Lue, 2013), (Omer, Haider, & Park, 2011), the peaks at 2972 cm -1 and 1460 cm -1 correspond to (C-H) stretching and bending vibration respectively (Coates, 2004), the peak at 1245 cm -1 corresponds to (C-N) bending vibration found in NIPAAm (Narain, 2010).
High transmission electron microscope (HRTEM) analysis shows the size and morphology of MTN. TEM micrograph of MTN in Fig. 1C reveals the spherical shape of MTN with core shell structure. The MNPs appears as dark spots coated with PNIPAM that constitutes the grayish layer as clarified by the arrows. The observed aggregation of the particles is a consequence of imaging under high vacuum conditions. The average diameter of magnetite was estimated with automated size distribution analysis software (image J). The size distribution curve of magnetite obtained from TEM micrographs shown in Fig. 1D. The magnetite diameter was determined to be 11 nm, which is consistent with the size obtained from XRD data, suggesting each particle is a single crystal. Moreover, the statistical analysis of the particles reveals the narrow size distribution of the nanoparticles (5-20 nm). Selected area electron diffraction pattern is shown in Fig. 1E. Peak indexing was performed using Equation 3; where d is the interplanar spacing, Rr is the ring radius, Lf is the camera focal length and λ is the wavelength of electron, and it is given by Equation 4.
Energy dispersive X-ray (EDX) spectrum of MTN is shown in Fig. 2A. The polymer content of magnetic thermoresponsive nanocomposite (MTN) was determined using TGA analysis. TGA thermogram estimates the weight loss occurring in the sample upon increasing the temperature from room temperature to 850 C as shown in Fig. 2B. The thermogram shows that weight loss occurred in two steps; one below 200˚C due to loss of water, the other occurred in the range from 200 to 600 C. The latter loss was attributed to the degradation of the side chain functional group and back bone structure of the PNIPAM (Purushotham & Ramanujan, 2010). The results show that the amount of PNIPAAm in MTN is 5.08%.

Thermal response
The LCST for NIPAAM is studied via the turbidity test.
The transmittance versus temperature is shown in Fig. 3A (Ward & Georgiou, 2011) , (Schmaljohann, 2006). It is worth noticing that, the turbidity test could not be conducted for MTN sample in which magnetite is the major constituent (~91.6%). For this purpose, DSC was utilized and gave an endothermic peak at 48 C as shown in Fig. 3B, designating it as the LCST. A slight increase in LCST of the polymer upon encapsulating magnetite has been reported previously (Dionigi et al., 2014), (Pich, Bhattacharya, Lu, Boyko, & Adler, 2004), (Rubio-Retama et al., 2007) and was explained on the basis of the steric hindrance of the magnetite core to the collapse of the polymeric chain at the phase transition temperature. Herein, the shift of LCST on MTN is higher than expected. This observation may be attributed to the high magnetite concentration as well as utilization of aluminum oxide (Al2O3) as the reference in the DSC measurement rather than the magnetite. Another influencing factor may be the high heating rate compared to the rate of collapse of the polymer chains.

Magnetic and specific absorption rate (SAR) measurements of magnetic core
The magnetization curve of magnetite (Fe3O4) is shown in Fig. 3C. The curve reveals the superparamagnetic behavior of Fe3O4 at room temperature with inconspicuous value of remnant magnetization. The saturation magnetization of is 35.412 emu/g which is less than the saturation magnetization of bulk magnetite (92 emu/gm and 74-80 emu/gm) for maghemite (Shokrollahi, 2017). The result is consistent with the fact that the saturation magnetization decreases with size.
As the particle size decreases, the surface to volume ratio increases and in turn, the magnetically dead layer fraction increases because of the canted and disordered spins forming a magnetically dead layer on the surface of the nanoparticles (Khairy, 2013). Likewise, the coercivity decreases with decreasing size below the critical size, which is consistent with previous theoretical estimations (Li et al., 2017), (Huber, 2005) The low values of coercive field (36.3 emu/g) and remanent magnetization (0.93694 G) unveil that the sample is superparamgnetic nature of the sample at room temperature. Slightly lower values of saturation magnetization (32.74emu/g) is obtained for MTN due to the non-magnetic polymer layer. It is worth noticing that the decrease in magnetization is consistent with the contribution of PNIPAM in the MTN structure as obtained from the TGA analysis. Coercive field and remanent magnetization are 34.522 G and 0.83349 emu/g denoting a superparamagnetic behavior as well.
Assessment of specific absorption rate of SPIONs was performed by exposing the sample to alternating magnetic field (AMF) of strength equal to 9.4 kA/m and frequency of 198 kHz obtained by the induction heater. Fig. 3D shows the temperature rise with time curve for two concentrations of magnetite colloid (3 and 7mg/ml) exposed to AMF. The field intensity and the operating frequency are within the acceptable range; concurring the patients-safety regulations for respective clinical application (Spirou, Basini, Lascialfari, Sangregorio, & Innocenti, 2018

Drug loading and release kinetics
Magnetic thermoresponsive nanocomposite (MTN) sample was loaded with 5-aminoisoquinoline (5-AIQ) by resuspension of MTN in four different concentrations of the drug as listed in Table  S1. MTN.5-AIQ samples were separated by a strong magnet then the residual concentration of 5-AIQ in the supernatants were quantified using High performance liquid chromatography (HPLC) analysis to measure the amount of residual drug. The HPLC chromatogram of 5-AIQ is shown in Figure S1A. Table 1 illustrates the value of drug loading efficiency of the samples and Fig. 4A illustrates the results. It was noticed that the drug loading efficiency increased as the concentration of feed drug increase and sample MTN.5-AIQ.4 has the highest drug loading capacity. Based on these results, the drug release kinetics were studied for sample, MTN.5-AIQ.4.
After incubation of MTN.5AIQ.4 sample at five temperatures -35, 37, ,39, 41, and 43 Cfor 30 minutes, the solution was separated by strong magnet and the supernatant was analyzed by UVvisible spectrophotometer. As shown in Fig. 4Bthe drug release was tested at different temperatures and the maximum release of the 5-AIQ was at 39.5 ̊ C. In addition, Fig. S1Bdisplays the drug release concentration at various temperatures analyzed by HPLC, which supports the findings obtained using UV-spectrophotometer and confirms that the maximum amount of drug is released at hyperthermia within the acceptable therapeutic range.

Kidney and liver functions
Serum levels of creatinine, uric acid, glutamic pyruvic transaminase (GPT) and glutamic oxaloacetic transaminase (GOT) were evaluated for the control, cisplatin-treated and MTN.5-AIQtreated groups. Cisplatin impaired the renal function as is apparent from the significantly higher serum creatinine level (p<0.05) compared with the control as shown in Fig. 5A. the renal dysfunction induced by cisplatin was previously depicted in literature (Wu et al., 2020). The serum creatinine and uric acid levels of the MTN.5AIQ-treated group, however, are not significantly different from those of the control group as illustrated in Fig. 5A&B; proving the biocompatibility of MTN.5-AIQ. The significant (P<0.05) elevation of serum level of GPT for the cisplatin group compared to control refers to the hepatotoxicity of cisplatin. On the other hand, the administration of MTN.5-AIQ did not produce any significant changes in liver function compared to the control as shown in Fig. 5C&D. This result proves the biocompatibility of MTN.5-AIQ.

Histopathology examination
Upon cisplatin administration, a marked congestion and atrophy of glomeruli are observed in mice kidneys of cisplatin-treated group compare to normal (Fig. 6. A-C). Furthermore, liver histopathological abnormalities are detected, marked by swelling and vascular degeneration in hepatocytes ( Fig. 6. I&J) comparable with normal control. These observations confirming the nephrotoxicity and hepatotoxicity of cisplatin (Wu et al., 2020). In MTN.5-AIQ-treated group, normal renal and collecting tubules are shown in Fig. 6 (E-F), only slight congestion can be detected in the interstitial blood vesicles. As for hepatotoxicity examination, the livers of MTN.5-AIQ -treated mice shows normal histopathological pictures with slight congestion in portal veins with normal bile ductuli (Fig. 6. K). Such results supporting the biocompatibility of MTN.5-AIQ encouraged by undetectable toxicity of the examined organ applied for histological study.

Cytotoxic effect of MTN.5-AIQ on colorectal adenocarcinoma
The cytotoxicity of MTN and MTN.5-AIQ was studied on Caco-2 cells as shown in Fig. 7 and

Synthesis of magnetite (Fe 3 O 4 ) nanoparticles
Colloidal magnetite nanoparticles were prepared by the co-precipitation method. Briefly

Characterization
XRD pattern for magnetite nanoparticles was obtained by using an X-ray diffractometer (Philips X,pert MPD) with CuKα radiation λ= 1.5418 Å. The crystallite size (T) of the sample was calculated by Scherrer (Equation 5 ) (Sharma, Bisen, Shukla, & Sharma, 2012); where β is the width of the peak at half maximum intensity for a specific plane with miller indices h, k, and l,whereas, C is the shape factor (taken as 0.9), λ is the wavelength of the incident X-ray and θ is the half angle between the incident and diffracted beams (2θ) in radians. The lattice parameter was calculated by using Equation 6; where dhkl is the interplanar spacing, and a is the lattice parameter.

Biocompatibility and toxicological parameters
Male Swiss albino mice weighing ~ 25-31 g were kept at animal facility of Faculty of Science, Ain Shams University. The animals were housed in an air-conditioned facility with a 12-h light/dark cycle, allowed free access to food and water. Mice were humanely treated in accordance with the ARRIVE guidelines for animal care. All experimental procedures were approved by the Ain Shams University Research Ethics Committee.
Fifteen mice were randomly divided into three groups. The first group was i.p. injected with saline and served as sham-operated control, the second one administered a single dose of cisplatin (15mg/kg) as a positive control (cisplatin-treated group). The last group administered a single intraperitoneal injection of (MTN.5-AIQ) (5mg/Kg). All mice were then euthanized and blood samples were collected from the heart and sera form different groups were separated by centrifugation (1500/10 min) and kept at -20 ˚C until the time of analysis. The kidney and liver tissue samples were collected and fixed in 10% buffered neutral formalin for further histopathological studies.

Kidney and liver functions
Renal functions were evaluated by measuring serum creatinine and uric acid according to manufacturer ' s instructions using commercial calorimetric kits (Bio-diagnostic Co., Cairo, Egypt).

Histological examinations
Kidney and liver samples obtained from different experimental groups were fixed in 10% buffered neutral formalin. The samples were then routinely dehydrated in graded series of ethanol, cleared in xylol and mounted in molten paraplast at 58-62ºC. Paraffin sections of about 4±5 μm were obtained, stained with H&E stain (Page, 1983), and examined under light microscope (LICA, German, provided with HD camera).
Cells were passaged at 80-90% confluency after trypsinization with pre-warmed trypsin-EDTA solution. The cytotoxicity of 5-AIQ, MTN and MTN.5-AIQ.4 samples was investigated on Caco-2 cell lines by cell viability MTT assay. The experiment was conducted for determination of the IC50 (the concentration of the drug which causes 50% cell death). The selected samples were twofold diluted in culture media. Cells were treated with 100 µl of each sample and incubated for further 24 hours. Then, 20µl of MTT solution were added to each well and incubated for 1-5 hours.
Finally, the produced formazan was dissolved in 200 µl DMSO and the mean absorbance of three replicates was measured at 570 nm.

Statistical analysis
All values are presented as means ± SE. Statistical analysis of experimental data was performed using a one-way analysis of variance (ANOVA) followed by Donnett's multiple comparison test for comparing means from different treatment groups.