Characterization of the constructed DNA nanoprobe
In the macroscopic world, nanoprobes are assemblies of components designed to achieve a specific function. Each component of the assembly performs a simple action, while the entire assembly performs a more complex, useful function that is characteristic of that particular device or machine. AS1411 could form stable G-quadruplex structures in the presence of K+ and specifically bind to nucleolin on the tumor cells membrane [10, 11]. In the previous study, we screened a chiral Ru (II) complex (laevoisomer) with alkyne, which exhibited strong affinity to AS1411 G-quadruplex DNA in groove binding mode through π-π stacking, and it is observed that two intramolecular hydrogen bond formation between N atom in midazole ring of RuPEP with two H atoms of G15 and T16 residues by molecular docking (Figure 1A). Here, the construction of this nucleolin-targeting nanoprobe is achieved through liquid-liquid phase separation of aptamer AS1411 and anchoring a phosphorous chiral ruthenium(II) complex RuPEP [29]. Moreover, it is observed that the nanoprobe displayed stronger fluorescence than equimolar RuPEP, which may be attributed to switch-on assay of RuPEP interacting with AS1411 to enhance the fluorescence emission of nanoprobe (Figure 1B) [30]. After initial assembly of nanoprobe, a high-order structural re-arrangement was observed by TEM (Figure 1C). Observably, the monodispersed nanoparticles with average diameter of 200 nm were seen for composites extracted from solution, the proposed nanoparticle structures that spontaneously assembled were also supported by AFM observations that showed well-dispersed, uniform size distribution with a mean diameter of 200 nm (Figure 1D).
We couple elemental mapping using energy-dispersive X-ray spectroscopy in a transmission electron microscope (TEM-EDS) with colocation analysis to study the elemental distribution and the degree of homogeneity in the nanoprobe. The elemental composition analysis employing EDS showed the presence of a strong signal from the P atoms (16.41%) contributed to AS1411 molecules, and a obvious signal from the Ru atoms (4.83%) affiliated to RuPEP. Moreover, other obvious peaks for other elements C (33.24%), N (18.32%), and O (27.20%) that from AS1411 and RuPEP observed. Above EDX analysis suggesting the assembly of AS1411 and RuPEP constructed nanoprobe successfully. Furthermore, the elemental maps clearly demonstrate that the C, P and Ru elements are not distributed homogeneously (Figure S3A). P and Ru are more majorly centralized at the particle core while C are preferentially found closer to the particle surface. We observe similar properties throughout the sample, with strong spatial correlation between P and Ru, and the enrichment of two atoms in the particle centre (Figure 1F). In addition, the nanoprobe exhibited mean lengths ranging from 200–500 nm that was confirmed by data from a Malvern laser particle analyzer, showing multiple nanoprobe bound structures (≈ 200 nm, Figure 1G) [31]. After binding to RuPEP, the zeta potential of AS1411 increased negatively by almost double the initial value, indicating that the NPs became more colloidally stable than free AS1411 particles with amorphous structure (Figure S3B) [32]. To further confirm the stability of the nanostructures in solution, the changes of nanometer size was monitored by a Malvern laser particle analyzer over time. We found that the size of nanoprobe increased slightly with the increasing time from 2 h to 72 h, and stabilized to about 200-500 nm after 72 h (Figure S3C). We regarded this as evidence that the nanoprobe exhibits some degree of stability over time in aqueous solutions.
In addition to observe the aggregation of AS1411 in the presence of RuPEP in the aqueous solution by laser diffraction we used analytical ultracentrifugation (AUC) to evaluate molecular weight accretion as a function of time. The results are shown in Figure 1G(in supporting information). The nanoprobe accretion complex displayed three possible states having the molecular weights of 47.3, 75.8, and 131 kDa, in which the corresponding sedimentation coefficients (SC) were about 3.5 S (polymer), 4.5 S (polymer), and 9.5 S (polymer), respectively Figure 1G(in supporting information). These structures are larger than those of free AS1411, which exhibited two possible states with molecular weights of 27.9 kDa (SC is about 3.2 S) and 43.0 kDa (SC is about 4.2 S) [33]. After prepared nanoprobe for 72 h, increasing sedimentation coefficients and molecular weights were observed for AS1411 after the addition of RuPEP. These data indicate that in the presence of equimolar concentrations of RuPEP, more rigid, high-order nanoparticles structures self-assemble from random coil AS1411 [34].
Cellular uptake and Elevated Localization of nanoprobe in Nuclei of Tumor Cells
Then, we used NCL high-expression breast cancer MDA-MB-231 cells to study the tumor-targeting recognition ability of the nanoprobe. After incubation in MDA-MB-231 breast cancer cells, nanoprobe is completely absorbed by the cells and emits strong red phosphorescence from the cell nuclei (Figure 2B). We observed that the bright red phosphorescence (nanoprobe) co-localized at the same site and completely overlaid the blue fluorescence band. In the magnified images, two-color fluorescence bands were confined to the cell nuclei. The overlap ratio of the three color bands originating from nanoprobe and DAPI was very close to 100%. In addition, the red fluorescence in 3D tomoscan imaging from depth sectioned images filled the entire nucleus and matched the staining pattern observed for nanoprobe and DAPI. These results indicated that nanoprobe were efficiently absorbed and retained by tumor cells and localized in the nuclei of these cells.
To ascertain the cellular uptake mechanism of nanoprobe for nuclear translocation from extracellular environment to nucleus (Figure 2C and 2D) [35], we cultured MDA-MB-231 cells with nanoprobe (5 μM) at either 37°C or 4°C for 6 h. The majority of nanoprobe localized in the nucleus when incubated at 37°C, whereas the nanoprobe remained in the cell cytoplasm when incubated at 4°C. Based on above results, we hypothesize that nanoprobe enters the cell nucleus through an energy-dependent pathway deriving from an active transport mechanism that drives NCL to the nucleus by intra-cytoplasmic trans-localization. These processes are slowed at 4°C. Usually, endocytosis describes an energy-dependent process for a general entry mechanism for various extracellular materials. In this process clathrin-coated pits are the primary plasma membrane specialization vehicle involved in the uptake of a wide variety of molecules [36, 37]. To clearly confirm the specific endocytotic pathway involved in cellular internalization of nanoprobe, we pretreated MDA-MB-231 cells with chlorpromazine (clathrin-dependent inhibitor) for 1 h before incubation with nanoprobe. We then observed that the fluorescence signals from nanoprobe mainly localized at the cell surface membrane, while little fluorescence was distributed in the cytoplasm (Figure 2D). These data suggested that nanoprobe are processed by living cancer cells through an endocytotic pathway [38]. Initially, 2-deoxy-D-glucose and oligomycin, which is a common inhibitor combination acting as an ionophore that reduces the ability of ATP synthesis to function optimally, were employed to determine the mechanism underlying essential nuclear accumulation [39]. Interestingly, cells treated with 2-deoxy-D-glucose and oligomycin exhibited significant inhibition of staining by nanoprobe in the nucleus (Figure 2D). Again, this data supports the view that the uptake of nanoprobe into the nucleus is mainly caused by an energy-dependent active transport pathway.
Bio-TEM was performed to shed more light on the cellular uptake of nanoprobe in breast cancer cells. After incubation with nanoprobe, MDA-MB-231 cells were harvested and sectioned for bio-TEM analysis [40]. As shown in Figure 2E, nanoprobe were trapped inside vesicles that were observed in the cytoplasm and nucleus. It is observed that nanoprobe may induce the MDA-MB-231 cells to produce several vesicles to carry them entered the cytoplasm and moved near nuclear envelope (yellow arrow in Figure 2E, step 1 and step 2). Numerous nanoprobe complexes with different sizes and shapes were found in these vesicles, which are illustrated (yellow arrows). The vesicles containing nanoprobe particles gradually approached the nucleus and their contact with the nuclear membrane appears to have triggered its disruption (step 2), after which the particles enter the nucleus through ATP-dependent endocytosis (step 3). Images of the nanoprobe complex escaping from vesicles are shown by pointing red arrows in cell the nucleus (step 4). The escape from vesicles is an important function for completion of their multi-faceted operation. We assume that the nuclear distribution of nanoprobe particles is related to ATP-dependent NCL transport processes that rely on the AS1411 component of nanoprobe to recognize and bind to NCL.
Imaging tumor cells by the constructed nanoprobes through targeting recognize nucleolin
NCL is a major nucleolar protein that is able to shuttle between the cell surface, the cytoplasm, and the nucleus- a property that makes NCL an attractive target for the selective delivery of anti-tumor drugs without affecting normal cells [41]. A number of studies indicated that NCL is over-expressed in human breast cancer cells and largely distributed on the surface of the cell membrane [10, 11]. However, in normal epithelial cells, NCL is mainly confined within the cell nucleus and deficient in cell membrane [10, 11]. As shown in Figure 3B, in both cases, it is clearly located in the nucleoli that perfectly matched DAPI staining cell nucleus and non-staining nucleoli. It is confirmed that NCL is mainly distributed in cell nucleolus and only a little in cell membrane, as well as is riched in tumor cells. Literature reported NCL is over-expressed (from three to six fold increase) in human breast cancer cell lines compared with normal cells [42]. And then, further study shown that the expression of NCL in MDA-MB-231 is obviously higher than in MCF-10A cells (Figure 3C and 3D). Moreover, according to the analysis of NCL expression on transcription level with patients’ survival in breast cancer using GEPIA2 and Kaplan-Meier plotter based ATGC database [18]. The data also shown that NCL is higher expression in BRCA (breast cancer) patients tissue than normal tissue (Figure S6), the patients with high expression of NCL also exhibited poorer survival outcomes.
We were curious to examine the distribution of nanoprobe in the cytoplasm of cancer cells to see if they should distribute intracellularly in the absence of metabolic transport. For MCF-10A cells, nanoprobe fails to enter the cells (Figure 3E), NCL targets in normal epithelial cells exhibit weak and diffuse uptake. But for MDA-MB-231 cells in the presence of nanoprobe (red phosphorescence, Figure 3E) was overlaid by DAPI stain in the cell nucleus, and the number of NCL loci shows a significant increase than MCF-10A cells (Figure 3F). These results suggested that nanoprobe may selectively recognize and activate transport of cell surface NCL receptors to cell nuclei and thereby facilitate imaging of breast cancer cells.
To further evaluate the selectivity of the nanoprobe for breast cancer cells, we developed a cell culture model in which MDA-MB-231 and MCF-10A cells were co-cultured on microscope slides. Considering that NCL is over-expressed in MDA-MB-231 human breast cancer cells and is deficient in MCF-10A normal immortalized human epidermal cells, it should be expected that uptake of the nanoprobe should preferentially localize in the breast cancer cell line. To differentiate between the two cell lines in co-culture, we used MCF-10A cells with green fluorescence through GFP-labeled actin. All cells in co-culture system were labeled with blue fluorescence by Hoechst 33258. We incubated nanoprobe with the co-cultures at 5 μM for 6 h. Figure 3G shows strong red phosphorescence in the nucleus of MDA-MB-231 cells, while only a feeble red phosphorescence was observed in MCF-10A cells. These results clearly demonstrate that nanoprobe specifically target and identify tumor cells in mixed cultures.
In vivo Imaging tumor cells
After confirming that the probe selectively binds and translocates as expected to breast cancer cells in culture, we investigated its performance in vivo in MDA-MB-231 tumor-bearing BALB/c mice. Specific tumor-targeting images were obtained from nanoprobe interrogated at different injection time points (Figure 4A). Mice that were imaged in NIR before injection of the probes showed virtually no signal. NIR phosphorescence became visible immediately after intravenous (iv) injection in the tail vein due to the rapid distribution of the probes [43]. Tumor areas were well defined in the mice within the first 6 h as the nanoprobe rapidly recognizes and binds its NCL targets in tumor tissues through the enhanced permeability and retention effect (EPR effect) (Figure 4A). At 12 h, the illuminated tumor area had increased due to the retained signal from the tumor site, augmented by interference in the fluorescence background from normal tissue. With increasing time, fluorescence from normal tissues, originating from clearance pathways and non-specific uptake, caused the tumor area to become less well defined (Figure 4B). By contrast, strong fluorescence from the non-targeting probe RuPEP was observed in the entire mouse within 6 h, that indicates that free RuPEP rapidly distributes in the entire body and increases with time (Figure S7). The above results suggest that the constructed nanoprobe selectively and rapidly define tumor tissues after systemic administration with 6 h. Eventually, nanoprobe distributes throughout the entire body, but nevertheless predominately accumulates in tumors. Ex vivo images of organs and tumors taken at autopsy from experimental animals showed that the probe retention in tumors taken at 24 and 48 h were comparable (Figure 4C). These data confirm the long tumor retention time of nanoprobe. The ex vivo image after dissection shows that the quantitative distributions of nanoprobe and their non-targeting RuPEP component were determined by fluorescence intensity in the different organs. Signals for the two probes arising in brain tissues were markedly higher at 24 h than at 48 h, while the signals in the kidney were significantly lower at 24 h than at 48 h (Figure 4C). This indicates that the two probes transport across the Blood-Brain-Barrier and are cleared from the body through kidney filtration. Low metabolism and slow kidney clearance produced high tumor accumulation with greater tumor uptake and stronger fluorescence of the nanoprobe at 24 h than 48 h [44]. The consistency of these data showing the accumulation of the probes in the tumor and kidney from the in vivo measurements and after organ extraction clearly indicates that noninvasive real-time in vivo imaging for localizing specific tumors is feasible in spite of renal clearance of these probes (Figure 4D). This work demonstrates that nanoprobe can specifically and rapidly image tumor tissues in vivo, but the potential application of nanoprobe for in vivo detection is limited severely by aggregated distribution in different organs.
In vivo preliminary safety evaluation.
The unforeseen side-effect of metal-materials for application in biomedicine is always a major concern. For safety's sake, we evaluated systematic toxicity of nanoprobe in healthy kunming mouse after tail intravenous injection for nanoprobe at a dosage of 50 mg/kg per days for three day. Then, primary tissues (containing heart, liver, spleen, lung, kidney and brain) were histopathologically observed under light microscope by H&E staining (Figure 4E). Compared with the control group, no death and serious body weight loss were found in all test groups during the study period. Major tissues including brain, heart, liver, spleen, lung and kidney have no obvious histopathological abnormalities or lesions in the two groups [45]. These results indicated that multiple dosing of nanoprobe had minimal impact in these tissues, showing that there was no significant side-effect caused by this nanoprobe. But in order to improve the potential application of this nanoprobe in clinic, the long-term toxic effects should be further investigate in the future study.
Human breast cancer section imaging
To further investigate the potential application of nanoprobe as a diagnostic agent for breast cancer in clinical tissue specimens, we used five fresh biopsy specimens of patients with invasive ductal carcinoma of the breast to evaluate the availability of nanoprobe for targeting NCL to image tumor tissue. Histology in the resection specimen revealed that obvious neoplastic lesion was composed of large polygonal cells arranged in infiltrating solid and micropapillary formations, with abundant eosinophilic, vacuolated, and foamy cytoplasm. In situ areas of the lesion contain cells arranged in an alveolar pattern with a hobnail appearance (Figure 5A) [46]. Also, it is apparent that a segmentation is produced for most of the nuclei in the image, with few contours corresponding to non-epithelial nuclei objects. However, there are clear differences between tumorous and paracancerous tissues. It is observed that the tumor cells are disorganized with incompact structure and deep-dyed bigger nucleolus than normal cells. As mentioned before, the preferential tumor accumulation is considered to be caused by NCL-mediated active transport of nanoprobe. Although the contribution of nanoprobe to tumor accumulation is clear in co-culture system of MDA-MB-231 and MCF-10A cells lines, the potential effectiveness of nanoprobe for targeting tumor imaging is unclear in human breast cancer biopsy specimens.
In histological analysis for the ex vivo tumor samples shown in Figure 5B, visible blue fluorescence is observed at DAPI mode in the pathological section. Moreover, there is a distinct demarcation between cancerous area with high-expressed nucleolin (green fluorescent spot) and paracancerous area with low-expressed nucleolin. Then, in enlarged image (Figure 5C), nucleolin merged greatly in cell nucleus with red fluorescence in turmorous area, but no fluorescence signal of nanoprobe and nucleolin in paracancerous area. Importantly, it is found that the red fluorescent of nanoprobe merged perfectly with green fluorescent nucleolin in cancerous tissue, but no apparent red fluorescence in paracancerous tissue (Figure 5D). Above results suggesting that the nanoprobe could effectively and differentially highlight cancerous tissue in biopsy specimens of patients with invasive ductal carcinoma of the breast.
And then, we also evaluated the expression of NCL in tumor and neighbor normal breast tissues by Western blotting. It is found that most of the tumor tissues exhibit significantly up-regulated NCL levels when compared against neighbor normal tissues (Figure 5E). Combined with the results of the clinical diagnosis report, the higher expression of NCL showed higher grade malignancy (Figure 5F), indicating the expression level of NCL is a feasible defining features in human invasive ductal carcinoma of different malignancy grade and can be used in predicting tumor malignancy [47]. In that way, the nonoprobe might be available for distinguishing the malignancy grade of invasive ductal carcinoma in clinic.
Potential clinical application in tumor grade diagnosis grade
Then, we used more samples, which are definitely diagnosed and divided into stages of invasive ductal carcinoma, to evaluate the availability of nanoprobe act as a convenient and rapid probe to define the tumor grade through testing the luminescent intensity in biopsy tissue section (Figure 6A). HE staining showed that the arrangement of normal tissue cells was tight with light red staining, but it revealed the well-defined tumor without obvious invasion to adjacent normal tissue for grade I samples (Figure 6B). However, with the tumor development to grade II and III, it invades a tissue area as a large number of interlocked tumors and the boundaries between malignant tissue and healthy tissue are blurred and, eventually dissolved.
As mentioned earlier, the expression levels of NCL in invasive ductal carcinoma are increased in tumor initiation and progression, that it's a critical factor to distinguish tumor grade. It is found that the nanoprobe displayed prominent differences in imaging capability for different grades of invasive ductal carcinoma, the higher degree of malignancy, the higher phosphorescent intensity (Figure 6C). In these clinical specimens, the nanoprobe emitted extremely weak red signal in normal tissues with intensity at the trace of mark line about 0~60 a.u., but it is observed quite strong red phosphorescence in grade I-III tissues with intensity at the trace of mark line range of 110-260 a.u. (Figure 6D). To further make clear the effectiveness and credibility of the nanoprobe distinguish tumor grade through phosphorescent intensity range, it still need to be improved by expanding the specimens' quantity and extending the number of repeats. Through statistical analysis of five specimens of every group for three repeats, it is found that the average intensity in equal area of nanoprobe in normal tissues is range of 7~21, in grade I tissues is range of 15~68, in grade II tissuess is range of 54~134 and in grade III tissues is range of 88~152 (Figure 6E). Compared with common used tumor marker-based methods in clinilical, which is expensive, or the operation is complex, high maintenance costs, is not conducive to the promotion. This study indicated that the synthetic nanoprobe has the potential to act as a convenient and cost-effective diagnostic agent for diagnosis of breast cancer.