Controllable CO Release Following NearInfrared Light-Induced Cleavage of Iron Carbonyl Derivatized Prussian Blue Nanoparticles for CO-Assisted Synergistic Treatment Wei-Peng Li,†,‡ Chia-Hao Su,§,‡ Ling-Chuan Tsao,† Chun-Ting Chang,† Ya-Ping Hsu,† and Chen-Sheng Yeh*,† †
Department of Chemistry and Advanced Optoelectronic Technology Center, National Cheng Kung University, Tainan 701, Taiwan Institute for Translational Research in Biomedicine, Kaohsiung Chang Gung Memorial Hospital, Kaohsiung 833, Taiwan
§
S Supporting Information *
ABSTRACT: Carbon monoxide (CO) causes the dysfunction of mitochondria to induce the apoptosis of cancer cells giving a promising choice as an emerging treatment. The currently reported CO-based complexes still suffer from many limitations. Synthesis of CO-release carriers in the manner of on-demand control is highly anticipated. In this study, we present a nearinfrared (NIR) light-responsive CO-delivery nanocarrier, a PEGylated iron carbonyl derivatized Prussian blue (PB) nanoparticle (NP). Taking the structural characteristic containing Fe3+−NC−Fe2+ unit, the −CN− served as the active sites for the coordination of iron carbonyl, while the surface Fe sites chelated with the amine-functionalized polyethylene glycol (NH2−PEG6000−NH2) to yield PEGylated PB NPs carrying CO. The control of light intensity and exposure period is important to release the amount of CO as well as to deliver the hyperthermia effect. The combination therapy including CO and photothermal treatments displayed a synergistic effect against cancer cells. Importantly, the release of CO is inert in the blood circulation without NIR irradiation. The blood oxygen saturation measured by the pulse oximeter and the HCO3, tCO2, and pH values analyzed by the blood assay revealed the steady status from the mice studies, showing no acute CO poisoning. KEYWORDS: carbon monoxide, Prussian blue, iron carbonyl, laser ablation, anthracemia
C
with iron carbonyl led to carrying CO effectively. Third, the iron carbonyl derivatized PB NPs showed no acute CO poisoning in vivo. Chemists developed metal carbonyl-based CO-release molecules (MCCORMs) for the control of CO liberation. Later, the development of nanotechnology has endowed these MCCORMs to combine with nanostructures giving the advantages of high drug payload, more specific drug delivery, and low toxicity. These MCCORMs-constructed nanostructures liberated CO either from the ligand exchange reaction through the electron-donating molecules, such as glutathione,
arbon monoxide (CO) can be endogenously produced through the intracellular heme-oxidation process1 and serves as an important signaling molecule to counteract inflammation, to protect cells, to activate blood vessels, and to participate in intracellular redox reaction.2,3 Unfortunately, CO is also known as the silent killer. CO has greater affinity, 200-fold larger, toward hemoglobin than oxygen. Inhaling a high dose of CO gas causes permanent damage to the human body. In order to employ CO for biological studies, it is, therefore, highly desirable to design the carrier in the manner of an on-demand release of CO. What is more, the gaseous nature has made CO more challenging for controlled delivery. The current study has included three features for CO gas treatment. First, CO gas is applied against malignant tumors through a near-infrared (NIR) light control of release. Second, the approach to conjugate Prussian blue (PB) nanoparticles (NPs) © 2016 American Chemical Society
Received: August 30, 2016 Accepted: November 29, 2016 Published: November 29, 2016 11027
DOI: 10.1021/acsnano.6b05858 ACS Nano 2016, 10, 11027−11036
Article
www.acsnano.org
Article
ACS Nano
Scheme 1. Synthetic Strategy for NIR-Responsive m-PB-CO/PEG NPs Applied for CO and Photothermal Therapy and US Imaginga
a The m-PB NPs were composed of Fe3+−NC−Fe2+ units. The Fe(CO)5 was introduced to coordinate on the surface of m-PB NPs through the ligand exchange process. PEGylated m-PB-CO NPs were prepared via the process of the amine-tailed PEG chelating the active Fe sites. The m-PBCO/PEG NPs can be triggered by NIR laser to initiate CO and photothermal therapy and US imaging applications.
cysteine, in cells,4−6 by the enzyme induction,7,8 or upon the exposure of UV−vis light.8−13 However, CO release from these nanostructures still suffered from different drawbacks including either lack of the precise spatial-temporal control or the phototoxicity accompanied by the limited penetrability. There are only two reports to control CO liberation by NIR light so far.14,15 He et al. conjugated graphene oxide with Mn carbonyl complex receiving 808 nm light, followed by the electron transfer to detach CO for cellular anti-inflammation.14 Ford and co-workers formulated photolabible MnCORM with upconversion NPs releasing CO through up-converted visible emission upon 980 nm irradiation.15 Therefore, NIR-driven CO release is still in its infancy. Previous studies regarding MCCORMs-based nanocarriers have focused on the cellular level for anti-inflammatory activity.4,6,14 Unlike normal cells, cancer cells require a substantial energy supply to sustain rapid growth. Treatment with CO in cancer cells rapidly enhances mitochondria activity that compels cancer cells to consume more oxygen for energy generation. The resulting metabolic processes lead to the mitochondria exhaustion and reactive oxygen species generation, giving the apoptosis of the cancer cells.16 Although the MCCORM-constructed nanostructures provided the advantage in the delivery of CO, the surface modification of nanomaterials with MCCORMs has been a delicate task. Few examples have demonstrated synthetic strategies to from MCCORM-based nanocarriers,11,12,14,17 for example, a click reaction for azide−alkyne cycloaddition to attach Mn carbonyl complexes on NPs,11,12 an amide-bond formation to coordinate Mn carbonyl complex with graphene oxide,14 and a metal−oxygen binding between iron oxide NPs and DOPA for the subsequent coordination to a Ru carbonyl complex.17 Apart from the aforementioned approaches, we present a MCCORM-based nanocarrier using the Fe(CO)5 derivatized PB NPs to release CO through the thermolysis upon NIR light irradiation against a malignant tumor. PB is an
U.S. FDA approved antidote for heavy metal poisoning.18,19 Recently, it has been applied in NIR-driven ablation of cancer because of the characteristic absorption in 600−900 nm corresponding to the charge transfer between Fe2+ and Fe3+ through a cyanide bridge.20−23 Later development has excavated PB with a hollow cavity to carry anticancer drugs for additional chemotherapy.24,25 Notably, PB has the structural composition composed of an Fe3+−NC−Fe2+ unit that has endowed PB as an ideal candidate to process derivatization with iron carbonyl. The exposure of the cyano group −CN− acts as the active sites to replace one CO of Fe(CO)5 for coordination (Scheme 1). The photon-to-heat conversion upon exposure of NIR light is anticipated to cleave the Fe-CO coordinating bond, yielding CO leaving molecules. Although PB NPs have been utilized in the treatment of malignant tumors in nanomedicine, the current study is the example of PB derivatization to provide CO gas-based therapy beyond photothermal ablation and chemotherapy. The present nanoarchitecture and approach resulted in the synergistic effect from the combination of CO and photoinduced hyperthermia therapy. The advantage of the combination treatment is that none of the effect of the individual treatment is lost, but the benefit is that the synergistic effect is raised giving a shorter therapeutic period and a less adverse side effect. Finally, we come to the point of the biosafety: whether or not the current iron carbonyl derivatized PB NPs endanger the oxygen supply in a living body. The measurements of the blood oxygen saturation were thus conducted by the pulse oximeter that monitored the hemoglobin carrying oxygen, i.e., the ratio of oxygenated hemoglobin to deoxygenated hemoglobin. The mice of the SpO2 (peripheral oxygen saturation) values were found to remain steady in the course of 24 h and the mice stayed healthy. The blood assay also displayed stable values in HCO3, tCO2, and pH. 11028
DOI: 10.1021/acsnano.6b05858 ACS Nano 2016, 10, 11027−11036
Article
ACS Nano
RESULTS AND DISCUSSION As shown in Scheme 1, the iron carbonyl (Fe(CO)5) was chosen to coordinate with the exposed cyano groups of the mesoporous Prussian blue nanoparticles (m-PB NPs) through the cyanide-carbonyl ligand exchange reaction, yielding iron carbonyl-modified m-PB NPs (m-PB-CO NPs). Subsequently, the amine-functionalized polyethylene glycol (NH2−PEG6000− NH2) chelated with the surface Fe sites of the m-PB-CO NPs (m-PB-CO/PEG NPs) for the purpose of a better circulation period when receiving an intravenous injection in vivo. Both TEM and SEM images revealed the cubic m-PB NPs with the edge length of 110 nm (Figure 1a,b). The cubic
evidence of the iron carbonyl coordination (Figure S5). Two peaks attributed to the bonding energies of Fe orbitals from 2P2/3 and 2P1/3 were detected for m-PB NPs. Two additional broad bands, respectively, accompanied by 2P2/3 and 2P1/3 appearing around 711 and 724 eV correspond to the iron carbonyl attached on the surface for m-PB-CO and m-PB-CO/ PEG NPs. For O 1s, the peak down shifted from 532 eV (m-PB NPs) to 531 eV (m-PB-CO and m-PB-CO/PEG NPs), which is attributed to the iron carbonyl. This indicates the successful conjugation of iron carbonyl on PB NPs. On the contrary, the binding energy of C(1S) corresponding to 284.6 eV remained the same for m-PB NPs, m-PB-CO NPs, and m-PB-CO/PEG NPs. Last, the hydrodynamic diameter and zeta potential of materials were measured by the dynamic light scattering (DLS) spectrometer (Table S1). The negative charge (−23.3 mV) of m-PB NPs originated from the presence of CN− groups on the surface, and then the surface charge shifted to the positive value of +13.9 mV when conjugated with iron carbonyl. Later, the PEGylation resulted in the surface charge of +7.08 mV because of the NH2−PEG6000−NH2 chelating to Fe sites. PB has a promising application to serve as NIR-responsive material. As-prepared m-PB, m-PB-CO, and m-PB-CO/PEG NPs revealed the nearly identical UV−vis-NIR absorption contour ranging from 600 to 900 nm with close intensity (Figure 2a). Because the photothermal ablation would be conducted using a NIR diode laser with an 808 nm wavelength, the high molar extinction coefficient at 808 nm was estimated as 0.9 × 1011 M−1 cm−1. The photothermal conversion efficiency (ηT) was calculated to be 11% following the method developed by Roper et al.26 Figure 2b−e demonstrated the NIR-responsive performance for NPs. Upon NIR light irradiation, the temperature elevated with the laser intensity and the dosage of the NPs (Figure 2b,c). The reduced hemoglobin (r-Hb) was used as the acceptor to capture CO for the determination of CO release triggered by NIR light. CO binds to the axial position of r-Hb forming the r-Hb-CO accompanied by the appearance at 410 nm absorbance.14 Therefore, the absorbance of r-Hb-CO can be referred as an index for the CO amount. Figure 2d displays the increase of the CO-release quantity as a function of laser intensity and exposure period. A pulsatile release mode was obtained to demonstrate precise control of CO release under light ON− OFF alternation (Figure 2e). The thermolysis of m-PB-CO/PEG NPs was studied by the thermal stability of NPs at 37° and 42°. The CO release gradually increased in the course of 15 min at 42 °C, while CO release was not seen at 37 °C (Figure S6). This evidence suggests that the thermal decomposition of the iron carbonyl coordinated PB NPs to liberate CO could occur when receiving NIR light irradiation. The stability analysis of m-PB-CO/PEG NPs was further investigated for the purpose of the subsequent biomedical studies. The bioenvironment containing enzymes, acidic organelles, phosphate or glutathione species might potentially lead to CO elimination from the PB NPs. The mPB-CO/PEG NPs were then individually dispersed in the phosphate buffered saline (PBS) (pH 7), PBS (pH 5), serum, Dulbecco’s modified Eagle’s medium (DMEM), cysteine, and glutathione solutions and incubated for the observation of 1 week (Figure S7). The absorbance of r-Hb-CO did not show the apparent increase in the course of 7 days from all of the solutions, giving a high stability of the m-PB-CO/PEG NPs. In addition, the structures of m-PB-CO/PEG NPs remained unchanged and with high dispersity (Figure S8). The m-PB-
Figure 1. Electron microscopic analysis. (a) TEM and (b) SEM images (the blue box drawn to reveal the cubic structure) of the mPB NPs. TEM images of (c) m-PB-CO NPs and (d) m-PB-CO/ PEG NPs.
morphology was further confirmed by the face-directed and vertex-directed (tilted by 30°) images (Figure S1a,b). Elemental mapping analysis identified the Fe distribution (Figure S1c). The lattice planes of the single crystalline m-PB NPs were identified by HR-TEM (Figure S1d) and the selected area electron diffraction (Figure S1e). The subsequent iron carbonyl coordination (m-PB-CO NPs) and PEGylation (m-PB-CO/ PEG NPs) have no affect on the NPs morphology (Figure 1c,d). The mesoporous structure was also analyzed by the Brunauer−Emmett−Teller (BET) analyzer, revealing as a type IV mesoporous with surface area of 96 m2/g (Figure S2). X-ray diffraction measurements showed the same pattern with the composition of PB for m-PB, m-PB-CO, and m-PB-CO/PEG NPs (Figure S3). The characteristic groups on the surface were characterized by the Fourier transform infrared (FT-IR) spectrometer (Figure S4). The Fe3+−NC−Fe2+ vibration signal at 2086 cm−1 can be seen from three samples of m-PB, m-PB-CO, and m-PB-CO/PEG NPs. The small intensity of CO vibration at 1646 cm−1 was observed because of the PVP residues on the m-PB. This CO signal enhanced when the iron carbonyl was coordinated to m-PB NPs. For m-PB-CO/ PEG NPs, PEGylation has brought up both C−H (2896 cm−1) and C−O (1180 cm−1) bonds. X-ray photoelectron spectroscopy (XPS) analysis was performed to provide additional 11029
DOI: 10.1021/acsnano.6b05858 ACS Nano 2016, 10, 11027−11036
Article
ACS Nano
Figure 2. UV−vis-NIR spectra and NIR-responsive performance. (a) The NIR absorption behavior for m-PB, m-PB-CO, and m-PB-CO/PEG NPs. (b) 100 ppm m-PB NPs dispersed in H2O irradiated by the different laser intensity using an 808 nm diode laser. (c) Different concentrations of m-PB NPs dispersed in H2O after receiving 0.8 W/cm2 laser irradiation (808 nm). (d) The absorbance of r-Hb-CO taken to reflect the CO-release quantity as a function of laser intensity and exposure period (808 nm) for 10 ppm m-PB-CO/PEG NPs. (e) A pulsatile release performance from 10 ppm m-PB-CO/PEG NPs upon 808 nm laser exposure at 0.8 W/cm2. All data were performed in triplicate.
CO/PEG NPs treated with HeLa cells (human cervical cancer cell lines) were evaluated for cytotoxicity using 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. The dosage of m-PB-CO/PEG NPs up to 300 ppm (in iron concentration) can still remain at a cell viability >90% (Figure S9), showing the biocompatibility for NPs. Prior to performing laser irradiation in vitro, we performed a fluorescence imaging experiment that shows the m-PB-CO/ PEG NP can successfully enter into the cells for the subsequent CO release when receiving laser irradiation. The doxorubicin (DOX) has been loaded into the mesoporous structures of the m-PB-CO/PEG NPs, yielding DOX-loaded m-PB-CO/PEG NPs. The DOX can emit red fluorescence for the observation of NPs location in cells. The HeLa cells were incubated with 100 ppm of DOX-loaded m-PB-CO/PEG for 24 h. The medium was removed and washed by PBS for 3 times to completely remove the residual DOX-loaded m-PB-CO/PEG NPs. Alexa 488 and Hoechst were used to stain the cytoskeleton and cell nucleus, respectively. The Figure S10 clearly indicates that the red fluorescence spots located in the cytoplasmic area. The cells alone without DOX-loaded m-PBCO/PEG NPs were treated as control group. The DOX-loaded m-PB-CO/PEG NPs were presumably through endocytosis into the cells for the subsequent laser exposure. The aforementioned demonstration in Figure 2d indicated the
approximate amount of CO release from the exposure of 0.3 W/cm2 for 15 min (referred as 0.3 W/cm2) and 0.8 W/cm2 for 5 min (referred as 0.8 W/cm2). According to the temperature elevation of 100 ppm of NPs from Figure 2b, using 0.3 W/cm2 only raised the solution temperature to 35 °C, while 0.8 W/cm2 readily heated up to 50 °C. For the subsequent in vitro and in vivo NIR-responsive studies of the synergistic effect (CO plus photoinduced thermal treatments), two conditions of 0.3 W/ cm2 (low thermal temperature) giving CO release only and 0.8 W/cm2 (high thermal temperature) resulting in CO release accompanied by photoablation were thus selected for laser irradiation. Figure 3a shows the cells survival rate upon laser irradiation. Both exposures of 0.3 W/cm2 and 0.8 W/cm2 would not cause any damage for cells only (without treatment of NPs). When the cells were treated with m-PB-PEG NPs (100 ppm), 0.3 W/cm2 had still no effect on the survival rate in the absence of hyperthermia, while 0.8 W/cm2 resulted in 30.2% cells killed because of the thermal effect. If the cells were treated with m-PB-CO/PEG NPs (100 ppm), 0.3 W/cm2 showed 25.6% cell apoptosis from CO toxicity, and 0.8 W/ cm2 caused the viability to drop to 51.4%. Therefore, the combination treatment in m-PB-CO/PEG NPs from 0.8 W/ cm2 was evaluated to possibly exert the synergistic effect. The additive therapy efficiency (Tadditive) can be calculated using the equation:27 11030
DOI: 10.1021/acsnano.6b05858 ACS Nano 2016, 10, 11027−11036
Article
ACS Nano
Figure 3. Cell studies in viability and US imaging upon laser irradiation. (a) The survival rates of HeLa cells were treated under different conditions and evaluated by MTT assay. The groups receiving laser irradiation were performed either at 0.3 W/cm2 for 15 min or at 0.8 W/ cm2 for 5 min. The dosage of the NPs was fixed at 100 ppm in iron concentration. (b) The cell viability dependence of CO amount as a function of laser irradiation period was conducted in HeLa cells treated with m-PB-CO/PEG NPs (100 ppm in iron concentration) at 0.3 W/ cm2. (c) The US images were taken by the microultrasound diagnostic probe (VisualSonics, 40 MHz). The CO generation caused the enhanced echo signal as seen for the bright contrast indicated by a yellow arrow. The laser (808 nm) irradiation for 15 min was conducted using the intensity of 0.3 W/cm2. (d) The US intensity values correspond to the images shown in (c). (e) The US images in tumors followed the pulsatile release behavior from m-PB-CO/PEG NPs (5 mg [Fe]/kg) subjected to 808 nm laser exposure at 0.8 W/cm2. US echo signal in tumors was observed following three repeated laser on/off cycles. (f) The US intensity values correspond to the images shown in (e). The red dotted circles represent US echo signal in tumors. All data were performed in triplicate.
Sonics, 40 MHz, B mode). As seen in Figure 3c, no signal can be found from all of samples without laser irradiation. Once a laser intensity of 0.3 W/cm2 was introduced for 15 min exposure, the noticeably enhanced white contrast attributed to the CO bubbles was detected from m-PB-CO/PEG NPs, while both PBS and m-PB-PEG NPs obtained no echo signal enhancement. The average gray values to quantify imaging revealed that the image intensity of m-PB-CO/PEG NPs was 9.6-fold enhancement triggered by laser (Figure 3d). To further demonstrate laser control of CO release, we have conducted additional in vivo experiments to monitor US echo signals in tumors by laser on/off pulsatile behavior. The m-PB-CO/PEG NPs were intratumorally injected, and then the US images were inspected as shown in Figure 3e,f. Following the same demonstration, as shown in Figure 2e for control of CO release under light ON−OFF alternation, the enhanced US contrast appeared because of CO release when the tumors received 0.8 W/cm2 laser exposure for 5 min. Contrarily, the laser-off action resulted in a drop in US echo signal. However, the tumor image brightened again following the second laser
Tadditive = (1 − fCO × fphotothermal ) × 100%
where f CO and f photothermal are the fractions of the cell viabilities from the individual CO and photothermal treatments. Accordingly, Tadditive was calculated as 48%, which is
