Tailoring the morphology of AIEgen fluorescent nanoparticles for optimal cellular uptake and imaging efficacy

We have demonstrated a new approach for regulating the morphology and emission of AIE-active organic nanoparticles by assembling different amphiphilic copolymers.

Understanding the interactions of nanoparticles with living cells is a key point to engineering ideal nanoparticles for bioimaging. 41,57,58 Recent literature suggests that alterations in the parameters of nanoparticles, such as the surface functionalization, size, geometry, and charge, can signicantly affect the pathway of endocytosis and the intracellular fate of the nanoparticles. 40,[59][60][61][62] Among these parameters, the geometrical shape of the nanoparticles has not received enough attention to date, partly because of the great challenge of designing nonspherical nanoparticles. 60,63,64 Recently, there have been several observations about the design of optimal nanoparticles for imaging and therapy. For example, Zhou et al. demonstrated that rod-like micelles exhibited accelerated cellular internalization compared to sphere-shaped micelles. 65 He et al. revealed that an appropriate increase of the aspect ratio would facilitate the cellular uptake of mesoporous silica nanoparticles. 66 Our previous work has also showed similar results. 67 In general, rodshaped particles appear to be more favorably engulfed compared to their spherical counterparts. 68 In this work, we have developed a universal approach to produce a group of AIEgen uorescent nanoparticles with different shapes, and investigated their cellular uptake (Scheme 1). The differently shaped particles were readily internalized in HeLa cells, and the rod-like micelles had faster internalization rates than their spherical counterparts, leading to a better imaging effect in vitro and in vivo.

Results and discussion
Herein, the rod-like micelles were obtained when we used poly(ethylene glycol)-block-poly(L-lactic acid) (PEG 5k -PLA 10k ) to encapsulate 9,10-distyrylanthracene (DSA), but the spherical micelles were made using poly(ethylene glycol)-block-poly-(caprolactone) (PEG 5k -PCL 10k ). The rod-like and spherical nanoparticles were named as DPP NRs and DPP NSs, respectively. For comparison, we also prepared DSA nanoparticles (spherical, DSA NSs) through a nanoprecipitation method. We systematically compared the properties of the three AIE nanoparticles in detail. DSA was synthesized according to our previously reported procedures. 50 The chemical structures of PEG 5k -PLA 10k and PEG 5k -PCL 10k were conrmed using 1 H NMR (Fig. S1 †). DPP NRs were prepared by adding a mixed THF solution of DSA and PEG 5k -PLA 10k dropwise into water with vigorous stirring for 1 h, followed by dialysis to remove the residual THF. The DPP NSs were prepared by following the same protocol, but using PEG 5k -PCL 10k instead of PEG 5k -PLA 10k . The DSA NSs were made in aqueous solution in the absence of polymer. The size distribution and morphologies of the DPP NRs/DPP NSs were characterized by dynamic light scattering (DLS), transmission electron microscopy (TEM) and confocal laser scanning microscopy (CLSM). As shown in Fig. 1, the DPP NRs were about 35.2 nm in diameter and about 137.1 nm in length, with a PDI value of 0.216. The DPP NSs and DSA NSs possessed an average size of 85.3 nm and 235.4 nm, and a PDI of 0.152 and 0.234, respectively. TEM images revealed the smooth rod-like morphology of the DPP NRs, while the DPP NSs and DSA NSs were spherical, but the spherical shape of the DSA NSs was not homogeneous. The DLS and TEM results of the DPP NRs and DPP NSs were different from those of the micelles of PEG 5k -PLA 10k and PEG 5k -PCL 10k , respectively (Fig. S2 †). In addition, the critical micelle concentration of PEG 5k -PCL 10k was lower than that of PEG 5k -PLA 10k (Fig. S3 †). Before observing the morphologies of the nanoparticles by CLSM, the AIE effect of DSA, the DPP NRs and the DPP NSs was conrmed, as shown in Fig. S4 and S5. † As shown in Fig. 1c, the DPP NRs were in the form of well-dened and monodisperse nanorods with green uorescence, while the DPP NSs (Fig. 1f) and DSA NSs (Fig. 1i) exhibited spherical morphologies with yellow uorescence. These results indicate that AIEgen nanoparticles (AIE NPs) with different morphologies emit different uorescence.
We compared their optical properties by UV-Vis absorption and photoluminescence spectra. The concentration of DSA in all samples was the same, which was adjusted according to the UV-vis standard curves (Fig. S6 †). As shown in Fig. 2, the absorption peak of the DPP NRs was slightly blue-shied compared with that of the DPP NSs, and their photographs under room light were almost the same (inset Fig. 2a). For their PL spectra (Fig. 2b), the maximum emission of the DPP NRs appears at 500 nm, while those of the DPP NSs and DSA NSs were at 540 nm, which is consistent with the noticeable color changes shown in Fig. 1 and the inset in Fig. 2b. Moreover, the three AIE nanoparticles possess large Stokes shis of about 100 nm, which greatly minimizes self-absorption and thus improves the signal-to-noise ratio for imaging. Although the Scheme 1 An illustration showing the preparation of spherical and rod-like AIEgen nanoparticles from DSA, and the comparison of their cellular uptake and imaging in vitro and in vivo. DPP NRs showed different uorescent emission to that exhibited by the DPP NSs and DSA NSs, their excitation spectra were almost the same (Fig. 2c). The quantum yield of the DPP NRs, DPP NSs and DSA NSs in water was 58.67%, 67.71% and 60.39%, respectively, which are much higher than that of DSA in THF (28.41%) (Table S1 †). The uorescence lifetimes of the DPP NRs, DPP NSs and DSA NSs were 2.02, 1.26 and 1.66 ns, respectively, which are shorter than that of free DSA (2.77 ns) ( Fig. 2d and S7 †). All these data are collected in Table S1. † These results indicate that the AIE NPs are totally different formulations with AIE molecules.
In order to reveal the assembly mechanism, we used Fourier transform infrared (FT-IR) spectroscopy and powder X-ray diffraction (PXRD) to further study the aggregates formed between the copolymers and DSA. As shown in Fig. 3a, the spectrum of the DPP NRs was red shied compared with that of PEG 5k -PLA 10k , indicating that the interactions between PEG 5k -PLA 10k and DSA were strong, presumably as a result of the synergy of noncovalent supramolecular interactions including p-p stacking and hydrophobic interactions. In contrast, the spectrum of the DPP NSs was almost the same as that of PEG 5k -PCL 10k , revealing that the interactions between PEG 5k -PLA 10k and DSA were weak (Fig. 3b). Furthermore, the PXRD of the freeze-dried DPP NRs showed well-resolved peaks, which is different to the situation with PEG 5k -PLA 10k . Meanwhile, the PXRD of the DPP NSs is similar to that of PEG 5k -PCL 10k (Fig. 3c  and d). These results demonstrated that the DSA molecules in the DPP NRs possess higher crystallinity than those of the DPP NSs, which leads to the different optical properties exhibited between the DPP NRs and DPP NSs. Moreover, to study whether this strategy can be a general approach to regulate the morphology of AIEgen-encapsulated organic nanoparticles, we used these two copolymers to form assemblies with three other AIEgens. As shown in Fig. S8, † only AIE3@PEG-PLA showed a rod shape, indicating that the structure of the AIEgens also played an important role in this study. In addition, we also studied the effect that the copolymer concentration has on the morphology of the micelles. As shown in Fig. S9, † the shape of the DPP NRs changed obviously with an increase in the concentration of copolymer, while still keeping a general rodlike shape. As a control, the DPP NSs changed slightly. These results suggested that this assembling strategy is special for DSA and its derivatives.
Excellent stability is essential for retaining the shape and function of the nanoparticles in blood circulation. Here, we evaluated the stabilities of the AIE NPs by monitoring the size distribution, absorbance and uorescence spectra in various conditions. As displayed in Fig. S10, † the DPP NRs and DPP NSs stored in Dulbecco's modied Eagle's medium (DMEM) with 10% fetal calf serum (FBS) and 1% penicillin/streptomycin exhibited unchanged sizes and size distributions aer 5 days. In contrast, the size of the DSA NSs changed obviously, and their PDI value increased in 2 days (Fig. S11 †). Moreover, the appearance of all the nanoparticles was still transparent aer ve days, without obvious aggregates or precipitates (Fig. S12 †). Furthermore, we studied the effect of human serum albumin (HSA) on the morphology of the micelles during storage. As shown in Fig. S13, † the size distribution of the DPP NRs and DPP NSs only increased slightly, which was mainly due to the absorption of the protein on the surface of the micelles. The above results indicated that the DPP NRs and DPP NSs could keep a stable nanostructure under physiological conditions. Furthermore, we collected the absorbance and uorescence spectra of the AIE NPs in aqueous solution over 7 days. As depicted in Fig. 4a-h, the absorbances of the DPP NRs and DPP  NSs all decreased slowly and retained more than 60% of the original value within one week, while that of the DSA NSs decreased signicantly and was reduced to 42.5% of the original value. Changes of the uorescence intensity gave similar results. These spectral results illustrate that both the DPP NRs and DPP NSs possess good physical and optical stability, which is favorable for biomedical applications. The photostability of the three AIE NPs was investigated by monitoring the uorescence intensity upon continuous laser irradiation. As shown in Fig. 4i, aer continuous laser irradiation at 488 nm for 30 min, the uorescence intensity of the three AIE NPs decreased only slightly and maintained about 90% of their initial value. Furthermore, we monitored the green uorescence signals from human cervical carcinoma (HeLa) cells pretreated with the three AIE NPs under laser irradiation for 30 min (Fig. S14 †). There was no obvious bleaching of the uorescence aer 30 min of laser irradiation. For direct comparison, we also studied the photostability of BODIPY dyes (BDP), which are believed to possess robust photostabilities, under the same conditions. 67 As shown in Fig. 4i, the uorescence intensity of the BDP maintained only about 20% of the initial value aer continuous laser irradiation. Moreover, the uorescence intensity of BDP in HeLa cells rapidly diminished and became negligible due to severe photobleaching (Fig. S14 †). The above results suggest that the DPP NRs and DPP NSs possess excellent physical and optical stability.
Biocompatibility is imperative for the use of uorescent nanoparticles as bioimaging agents. We rstly studied the biocompatibility of DSA, PEG 5k -PLA 10k and PEG 5k -PCL 10k toward HeLa cells using an MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay. As shown in Fig. S15a-c, † DSA, PEG 5k -PLA 10k and PEG 5k -PCL 10k all have low cytotoxicity toward HeLa cells at different concentrations aer incubation for 24 h. Similarly, low cytotoxicity of the DPP NRs, DPP NSs and DSA NSs against cells was observed, and more than 90% of those cells were alive at different incubation concentrations. To further demonstrate visually the biocompatibility of the DPP NRs, DPP NSs and DSA NSs, we stained the cells with calcein-AM and propidium iodide to identify live (green) and dead/late apoptotic (red) cells, respectively. As exhibited in Fig. S15d-f, † no red uorescence was observed for all of the samples, suggesting that the three AIE NPs have low cytotoxicity toward HeLa cells, which agrees well with the MTT experiments. Fig. S16 † shows the morphology of the HeLa cells aer incubation with different concentrations of the DPP NRs, DPP NSs and DSA NSs for 24 h; the cells maintain their normal morphology. These results conrmed the good biocompatibility of the DPP NRs, DPP NSs and DSA NSs.
Cellular uptake is necessary for nanomaterials to exert their functions, especially for live cell imaging. HeLa cells were used to investigate the cellular uptake of AIE NPs by CLSM. Aer incubating with DPP NRs at various concentrations for 2 h at 37 C, cellular nuclei were dyed using 4,6-diamidino-2phenylindole (DAPI). As presented in Fig. S17, † the homogeneous green uorescence was located in the cytoplasm, suggesting that the DPP NRs can pass across the cell membrane into the cytoplasm. Moreover, the DPP NRs exhibit internalization by living cells in a concentration-dependent manner. The DPP NSs and DSA NSs showed similar results ( Fig. S18 and S19 †). Furthermore, the sub-cellular location of the internalized nanoparticles was carried out using lyso-tracker red. As shown in Fig. S20, † the AIE nanoparticles were mainly located within the endosome, and the co-localization of the DSA nanoparticles (green) with the endosome (red) produced an orange uorescence in the merged images. All of these results conrmed that the DSA nanoparticles could be internalized effectively by cancer cells.
To evaluate the effects of the nanoparticle morphology on the cellular uptake efficiency, the HeLa cells were cultured with the DPP NRs, DPP NSs, and DSA NSs. As shown in Fig. 5, S21 and S22, † the intracellular uorescent intensity increased gradually with the incubation time from 1 to 4 h, demonstrating that these AIE NPs had a sustained cellular uptake in HeLa cells. In addition, the DPP NRs exhibited the strongest green uorescence, followed by the DSA NSs and DPP NSs, indicating that the DPP NRs were more easily internalized by cells compared with the DPP NSs and DSA NSs. Meanwhile, ow cytometry was employed to quantify the cellular uptake of the three AIE NPs. As shown in Fig. 6, the DPP NRs had relatively higher uptake efficiencies in comparison to the DPP NSs and DSA NSs. These results are in agreement with the CLSM results. In addition, we detected the cellular uptake efficiency of the NPs using UV-vis spectra (Fig. S23 †). The absorbance of DSA extracted from the HeLa cells aer different times increased from 1 to 4 h. The possible reason for the higher cellular uptake of the DPP NRs is that the rod-like nanoparticles have multivalent contact points with the cell membranes, resulting in stronger adhesions and an enhanced uptake relative to the spheres. To study the inuence of the stability of the nanoparticles on the cellular uptake, we studied the cellular uptake of the DSA NSs in two conditions: freshly made, and aer storing for 48 hours. As shown in Fig. S24, † the freshly made DSA NSs could pass into cells, but aer 48 hours of storage time the DSA NSs were less able to enter cancer cells. Thus, we used the DPP NRs and DPP NSs for further studies in long-term imaging.
To determine whether the AIE NPs with different morphologies are internalized via different endocytic pathways, we explored the uptake of polymer micelles by HeLa cells. Three types of inhibitor, namely sucrose, genistein and amiloride, were chosen to inhibit clathrin-mediated endocytosis, caveolae-mediated endocytosis, and macropinocytosis, respectively. Low temperature (4 C) treatment was used to determine whether the endocytosis process was energy-dependent. To decrease the inuence of the inhibitors on the cancer cells, experimental conditions were optimized according to previous publications. 65 The ow cytometry results are shown in Fig. 7. The cells treated under low temperature all showed drastic decreases in the uptake of the AIE NPs, conrming an energy-dependent endocytosis process. The internalization at 37 C showed that the AIE NPs with different morphologies show diverse uptake proles. The uptake of the DPP NRs by the HeLa cells primarily occurred via the clathrin-mediated endocytic and micropinocytic pathways. It is more probable that macropinocytosis and/or phagocytosis are the mechanisms of uptake of the DSA NSs, due to their relatively large size. These observations suggest that the mechanisms of endocytosis are dependent on the shapes of the nanoparticles. A spherical nanoparticle has only one face that can interact with the cell surface, while rod-like nanoparticles have multiple faces with large variations in size in each dimension. Therefore, we concluded that the cellular uptake of nanoparticles with various shapes appears to be mediated by multiple pathways.
To investigate and compare the long-term cellular tracking capability of the DPP NRs and DPP NSs, we captured the uorescence images aer different incubation periods (Fig. 8). The HeLa cells were rst incubated with the DPP NRs for 6 h at 37 C (labelled as day 0). The treated cells were then subcultured for designated time intervals. For each cell passage, the old culture medium was extracted and the HeLa cells were washed with PBS twice to remove the DPP NRs present in the culture medium. At the initial stage (day 0), strong and bright green uorescence from the DPP NRs can be clearly observed in Fig. 8. With the increase of incubation time (from day 3 to day 15), the green uorescence gradually decreases because of cell proliferation.   Interestingly, aer 15 days of subculture, the green uorescence from the DPP NRs was still clearly observed in the HeLa cells, which indicates that the DPP NRs can act as a uorescent probe for long-term cellular imaging. By contrast, the DPP NSs (Fig. S25 †) showed very weak uorescence aer 15 days of incubation. The DPP NRs showed stronger intracellular uorescence than the DPP NSs at every time point (Fig. S26 †). More importantly, this long-term imaging strategy is based on cellular proliferation and only needed a one-time addition of organic nanoprobes, rather than a continuous exogenous addition of imaging agents. All of these results indicate the superior cell tracing ability of the DPP NRs.
In order to further study the imaging capacity, the DPP NRs were intratumorally injected into tumor-bearing BALB/c mice, and then an in vivo optical imaging system was used to monitor the uorescence over 21 days. As displayed in Fig. 9a and S27, † uorescence from the site of the DPP NR injection could be readily detected aer 21 days, while that of the DPP NSs showed very weak uorescence aer 15 days (Fig. S28 and S29 †). The DPP NRs exhibited mild uorescence at the injection site one day aer injection. Over time, the uorescence intensity gradually increased and showed the strongest uorescence at day 6 ( Fig. 9b). Then, the uorescence intensity gradually decreased, but still could be readily detected aer 21 days, suggesting that the DPP NRs still remained in the tumor, showing better imaging performance than the DPP NSs. Moreover, to further conrm that the uorescence is from the DSA nanoparticles, we extracted the DSA from the tumor by tissue extraction with THF. As shown in Fig. S30, † the absorbance of DSA could be detected in the tumor extraction solution, and the absorbance intensity of the DPP NR group was stronger than that of the DPP NSs. The body weight of the mice in the two imaging groups gradually increased over time (Fig. 9c and S28c †), suggesting that the DPP NRs and DPP NSs had no distinct systemic toxicity. In addition, we also study the imaging capacity of the DSA nanoparticles by intravenous injection. As shown in Fig. S31, † uorescence at the tumor site of the DPP NR group could be readily detected aer 24 h. Over time, the uorescence intensity gradually decreased, but could still be readily detected aer 168 h, suggesting that the DPP NRs could accumulate in a tumor and show better imaging performance than the DPP NSs. Next, we studied the biodistribution of the DPP NRs and DPP NSs by detecting the uorescence in the tumor and the major organs excised from the mice. As shown in Fig. S32 and S33, † at 48 h post-injection a strong uorescence intensity could be clearly observed in the tumor, while weak uorescence was observed in the organs, and the uorescence of the DPP NR group was higher than that of the DPP NSs. These results indicated that the DPP NRs could accumulate and be retained around the tumor. The biosafety of the nanoparticles was evaluated by hematology analysis. As shown in Fig. S34, † the DPP NRs and DPP NSs all exhibited a negligible inuence on the AST, ALT, and blood urea nitrogen (BUN) or creatinine (CREA) indexes, compared to those of the saline group. All of this information consistently demonstrated the good biocompatibility of the DSA nanoparticles. The above results all conrmed that the DPP NRs are potentially promising for in vivo imaging.

Conclusions
In summary, stable AIEgen nanoparticles with different shapes were prepared, by the assembling of copolymers and AIE molecules, and used for noninvasive long-term imaging. The formulated nanoparticles exhibit superior physical and photo stability under physiological conditions. In vitro experiments have veried that these tailor-made AIE-active organic nanoparticles are biocompatible and are internalized through various pathways of cellular uptake. The long-term imaging ability was validated by in vitro and in vivo experiments. More importantly, the rod-like nanoparticles were signicantly more internalized than the spherical particles, resulting in a better imaging effect. Our ndings may provide useful information for the development of new strategies for the design of efficient AIEgen nanoparticles for bioimaging.