Photo-Cleavable Polycations-Wrapped Upconversion Nanoparticles for Efﬁcient siRNA Delivery and Cancer Therapy

: RNA interference (RNAi) therapy is a promising approach for cancer therapy. However, due to the weak binding afﬁnity between a carrier and small interference RNA (siRNA) and complicated tumor environment, efﬁcient loading and release of siRNA still remain challenging. Here, we design photo-cleavable polycations-wrapped upconversion nanoparticles (PC-UCNPs) for spatially and temporally controllable siRNA delivery. The PC-UCNPs are synthesized by in situ reversible addition − fragmentation chain transfer (RAFT) polymerization of photo-cleaved 5-(2-(dimethylamino)ethoxy)-2-nitrobenzyl acrylat (MENA) monomer and poly(oligo(ethylene oxide) methyl ether acrylate (OEMA) mononer through a chain transfer agent that anchored on the surface of silica-coated upconversion nanoparticles (UCNPs@SiO 2 ). After reacting with CH 3 I, siRNA and hyaluronic acid (HA) are adsorbed on the particle surface to prepare PC-UCNPs/siRNA/HA. The reaction with cell-secreted hyaluronidase (HAase) achieves the intracellular delivery of PC-UCNPs/siRNA/HA, and 980 nm laser irradiation causes siRNA release, which effectively improves the gene silencing efﬁciency in vitro and suppresses tumor growth in vivo; therefore, these processes have a promising potential application in precision medicine.


Introduction
Gene interference therapy using exogenous small interfering RNA (siRNA) to selectively silence gene expression and inhibit protein transcription has become a promising cancer therapeutic approach [1][2][3].However, the key challenge for the further application of RNA interference (RNAi) therapy still remains in the efficient delivery and release of these small, fragile biomolecules.Currently, many types of responsive siRNA delivery systems have been developed with pH [4,5], small molecule [6], light irradiation [7,8] and temperature [9] as stimuli, which enhances the controllability for gene release and the efficiency of gene transfection.
Light irradiation is a promising strategy for the on-demand release of payload [10][11][12].Recently, near-infrared (NIR) light has attracted much attention due to its minimal photodamage on living tissues, low auto-fluorescence background, non-photobleaching and deep penetration [13][14][15].The lanthanide-doped upconversion nanoparticles (UCNPs) can absorb long-wavelength NIR light and generate short-wavelength UV and visible light emissions [16,17].The integration of UCNPs with photoresponsive molecules, including o-nitrobenzy [18], spiropyran [19] and azobenzene [20], have been studied for drug delivery, gene delivery and in vitro/in vivo imaging [18,21].Although NIR-regulated UCNPs delivery systems have been reported for the successful release of siRNA [22,23], surfaceloaded siRNA usually lacks protection, impairing loading efficiency and resulting in siRNA Targets 2023, 1 64 degradation or leakage during the delivery process [24,25].Therefore, a NIR-responsive siRNA delivery system with a satisfactory loading capacity, stability and release efficiency is needed with urgency.
With the positive charge and convenience for functionalization, cationic polymers could enhance siRNA loading efficiency [26][27][28][29].Here, we designed NIR-cleavable polycationsencapsulated UCNPs for efficient siRNA delivery and tumor therapy.UCNPs are coated with SiO 2 and functionalized with chain transfer reaction (CTA) initiator 2-(((propanethio) carbonothioyl)thio)acetic acid.Reversible addition−fragmentation chain transfer (RAFT) polymerization is then carried out on the UCNPs surface with photo-responsive 5-(2-(dimethylamino)ethoxy)-2-nitrobenzyl acrylate (MENA) and poly(oligo(ethylene oxide) methyl ether acrylate (OEMA) as monomers.After reacting with CH 3 I, the as-obtained PC-UCNPs subsequently absorb siRNA and hyaluronic acid (HA) via electrostatic interactions to obtain PC-UCNPs/siRNA/HA (Scheme 1a).The outer layer of negatively charged HA not only extends blood circulation but also achieves siRNA delivery specificity by recognizing the CD44 receptor on the tumor cell membrane.The degradation of HA by hyaluronidase (HAase) exposes inner cationic polymers and promotes intracellular delivery and lysosome escape.Upon subsequent 980 nm laser irradiation, the UV emission of UCNPs cleaves the photosensitive o-nitrobenzyl segment and detaches the polymer coating from the UCNPs surface for efficient siRNA release (Scheme 1b).This strategy demonstrates the effective silencing of target gene expression and suppression of tumor growth; therefore, it should become a universal strategy for efficient NIR-assisted gene therapy.

Preparation of PC-UCNPs/siRNA/HA and In Vitro Verification of HA Degradation and siRNA Release
To load negatively charged siRNA via electrostatic interaction, the tertiary amine groups from UCNPs@SiO 2 -polymer were reacted with CH 3 I to form quaternary ammonium salts.Compared with UCNPs@SiO 2 -polymer, the as-obtained PC-UCNPs had an increased zeta potential of 58 ± 2.2 mV, while its hydrodynamic diameter remained unchanged (Figure S8).A series amount of siRNA was loaded on PC-UCNPs, and the weight ratio of PC-UCNPs/siRNA was optimized as 1/1, as indicated by the significantly retarded siRNA band in agarose gel retardation assay (Figure 2a).HA was further coated with PC-UCNPs/siRNA to obtain PC-UCNPs/siRNA/HA.After siRNA loading, the as-obtained PC-UCNPs/siRNA showed a zeta potential of 26 ± 2.3 mV and a hydrodynamic diameter of 61 ± 1.9 nm (Figure 2b, PC-UCNPs/siRNA).Continuous surface modification of HA decreased zeta potential to −23 ± 2.6 mV due to carboxylate groups of HA and increased the hydrodynamic diameter of particles to 70 ± 2.5 nm (Figure 2b, PC-UCNPs/siRNA/HA).

Preparation of PC-UCNPs/siRNA/HA and In Vitro Verification of HA Degradation and siRNA Release
To load negatively charged siRNA via electrostatic interaction, the tertiary amine groups from UCNPs@SiO2-polymer were reacted with CH3I to form quaternary ammonium salts.Compared with UCNPs@SiO2-polymer, the as-obtained PC-UCNPs had an increased zeta potential of 58 ± 2.2 mV, while its hydrodynamic diameter remained unchanged (Figure S8).A series amount of siRNA was loaded on PC-UCNPs, and the weight ratio of PC-UCNPs/siRNA was optimized as 1/1, as indicated by the significantly retarded siRNA band in agarose gel retardation assay (Figure 2a).HA was further coated with PC-UCNPs/siRNA to obtain PC-UCNPs/siRNA/HA.After siRNA loading, the as-obtained PC-UCNPs/siRNA showed a zeta potential of 26 ± 2.3 mV and a hydrodynamic diameter of 61 ± 1.9 nm (Figure 2b, PC-UCNPs/siRNA).Continuous surface modification of HA decreased zeta potential to −23 ± 2.6 mV due to carboxylate groups of HA and increased the hydrodynamic diameter of particles to 70 ± 2.5 nm (Figure 2b, PC-UCNPs/siRNA/HA).The serum stability of the delivery carrier is important to prolong circulation time and elevate tumor targeting [33].The hydrodynamic diameter and polydispersity index (PDI) of PC-UCNPs/siRNA/HA remained unchanged when incubated with DMEM containing 10% fetal bovine serum (Figure S9), indicating the satisfactory stability of PC-UCNPs/siRNA/HA.
The negatively charged surface of PC-UCNPs/siRNA/HA can effectively extend systemic circulation time [34] and respond to internal tumor microenvironments for surface charge conversion to facilitate intracellular delivery.Due to the surface-adsorbed HA, PC-UCNPs/siRNA/HA was recognized by the CD44 membrane receptor.The HA coating layer was then degraded by the overexpressed HAase enzyme in the tumor cell [35], which exposed the positively charged surface of PC-UCNPs/siRNA to achieve the intracellular delivery and endosomal escape of PC-UCNPs/siRNA/HA.According to the HA degradation period, the zeta potential of PC-UCNPs/siRNA/HA continued its increase towards 25 ± 2.4 mV, indicating the gradual decomposition of The serum stability of the delivery carrier is important to prolong circulation time and elevate tumor targeting [33].The hydrodynamic diameter and polydispersity index (PDI) of PC-UCNPs/siRNA/HA remained unchanged when incubated with DMEM containing 10% fetal bovine serum (Figure S9), indicating the satisfactory stability of PC-UCNPs/siRNA/HA.
The negatively charged surface of PC-UCNPs/siRNA/HA can effectively extend systemic circulation time [34] and respond to internal tumor microenvironments for surface charge conversion to facilitate intracellular delivery.Due to the surface-adsorbed HA, PC-UCNPs/siRNA/HA was recognized by the CD44 membrane receptor.The HA coat-ing layer was then degraded by the overexpressed HAase enzyme in the tumor cell [35], which exposed the positively charged surface of PC-UCNPs/siRNA to achieve the intracellular delivery and endosomal escape of PC-UCNPs/siRNA/HA.According to the HA degradation period, the zeta potential of PC-UCNPs/siRNA/HA continued its increase towards 25 ± 2.4 mV, indicating the gradual decomposition of negatively charged HAs and exposure of the positively charged PC-UCNPs/siRNA surface (Figure 2c, zeta potential).Along with the degradation of HA, the hydrodynamic size of PC-UCNPs/siRNA/HA continued its decrease to 62 ± 2.7 nm (Figure 2c, size), which was similar to that of PC-UCNPs/siRNA (Figure 2b, PC-UCNPs/siRNA, size), indicating the complete degradation of HA from the UCNPs surface.SiRNA release was achieved via external NIR irradiation, and siRNA-loaded PC-UCNPs/siRNA/HA with NIR irradiation showed rapid release, and the release percentage saturated at around 92% at 3 h (Figure 2d).On the contrary, PC-UCNPs/siRNA/HA showed a negligible release of siRNA (2%) in the absence of NIR irradiation (Figure 2d), indicating a satisfactory precise control of siRNA release via NIR light irradiation.

Internalization of PC-UCNPs/siRNA/HA and Intracellular siRNA Release
After the degradation of the HA coating layer, the highly cationic surface of PC-UCNPs/siRNA could facilitate the lysosomal escape of PC-UCNPs/siRNA.Acridine orange (AO) assay was conducted to monitor the lysosomal membrane permeabilization of HepG2 cells.HepG2 cells were incubated with PBS, UCNPs@SiO 2 -polymer, PC-UCNPs/HA, and PC-UCNPs/siRNA/HA, respectively, and stained with AO.Strong green fluorescence was observed from PBS-treated HepG2 cells and UCNPs@SiO 2 -polymertreated HepG2 cells (Figure 3a, PBS, UCNPs@SiO 2 -polymer), while declined green fluorescence and increased red fluorescence was observed from PC-UCNPs/HA and PC-UCNPs/ siRNA/HA-treated HepG2 cells (Figure 3a, PC-UCNPs/HA, PC-UCNPs/siRNA/HA), proving the contribution of cationic surface to lysosomal membrane permeabilization.A lysosomal colocalization experiment was further performed to trace the internalization and lysosomal escape of PC-UCNPs/siRNA/HA by CLSM images.FAM labelled siRNA (siRNA-FAM) was used to prepare PC-UCNPs/siRNA-FAM/HA, which demonstrated good overlap of green fluorescence of FAM and red fluorescence of Lysotracker Red after 2 h incubation, confirming efficient internalization of PC-UCNPs/siRNA-FAM/HA (Figure 3b, 2 h).After 6 h of incubation, the FAM fluorescence had spread to the entire cytoplasm, while the lysosome probe showed weaker fluorescence due to the acidity dependence of Lysotracker red (Figure 3b, 6 h) [36], indicating the efficient lysosomal escape of PC-UCNPs/siRNA-FAM/HA.
To verify the successful release of siRNA from PC-UCNPs to cytoplasm, PC-UCNPs/ siRNA-FAM/HA was incubated with HepG2 cells for 4 h, and UCNPs emission at 454 nm and 475 nm and FAM emission at 529 nm was monitored via CLSM.Separation of FAM fluorescence (green channel) and UCNPs fluorescence (red channel) was observed for PC-UCNPs/siRNA-FAM/HA-treated HepG2 cells after 40 min of NIR light irradiation (Figure 4a, NIR(+)), indicating the efficient release of siRNA from PC-UCNPs.On the contrary, no fluorescence separation was observed for PC-UCNPs/siRNA-FAM/HA-treated HepG2 cells in the absence of NIR irradiation (Figure 4a, NIR(−)).
HA-mediated endocytosis specificity was further confirmed by CLSM images.Strong intracellular FAM fluorescence was observed for PC-UCNPs/siRNA-FAM/HA incubated HepG2 cells (Figure 4b, HepG2), while excess HA pretreated HepG2 cells showed little fluorescence after incubation with PC-UCNPs/siRNA/HA (Figure 4b, HA pretreated HepG2).Moreover, the semiquantitative analysis of intracellular FAM-siRNA fluorescence also showed higher intensity for HepG2 cells than HepG2 cells pretreated with HA (Figure S10).These results indicated good targeting specificity for HA-mediated cancer cell internalization.
by CLSM images.FAM labelled siRNA (siRNA-FAM) was used to prepare PC-UCNPs/siRNA-FAM/HA, which demonstrated good overlap of green fluorescence of FAM and red fluorescence of Lysotracker Red after 2 h incubation, confirming efficient internalization of PC-UCNPs/siRNA-FAM/HA (Figure 3b, 2 h).After 6 h of incubation, the FAM fluorescence had spread to the entire cytoplasm, while the lysosome probe showed weaker fluorescence due to the acidity dependence of Lysotracker red (Figure 3b, 6 h) [36], indicating the efficient lysosomal escape of PC-UCNPs/siRNA-FAM/HA.To verify the successful release of siRNA from PC-UCNPs to cytoplasm, PC-UCNPs/siRNA-FAM/HA was incubated with HepG2 cells for 4 h, and UCNPs emission at 454 nm and 475 nm and FAM emission at 529 nm was monitored via CLSM.Separation of FAM   HA-mediated endocytosis specificity was further confirmed by CLSM images.Strong intracellular FAM fluorescence was observed for PC-UCNPs/siRNA-FAM/HA incubated HepG2 cells (Figure 4b, HepG2), while excess HA pretreated HepG2 cells showed little fluorescence after incubation with PC-UCNPs/siRNA/HA (Figure 4b, HA pretreated HepG2).Moreover, the semiquantitative analysis of intracellular FAM-siRNA fluorescence also showed higher intensity for HepG2 cells than HepG2 cells pretreated with HA (Figure S10).These results indicated good targeting specificity for HA-mediated cancer cell internalization.

In Vivo Therapeutic Efficiency of PC-UCNPs/siRNA/HA
Considering the overexpression of the CD44 receptor on the surface of the membrane, HepG2 cell tumor-bearing nude mice could be optimal tumor models [34].The in vivo antitumor efficiency of PC-UCNPs/siPLK1/HA was evaluated using HepG2 cell tumorbearing nude mice.PC-UCNPs/siPLK1/HA-injected mice after 2 h of NIR irradiation (6 min intervals for every 10 min of light exposure to avoid heating) showed the most effective inhibition of tumor growth compared with the PBS-treated mice group, and the PC-UCNPs/siPLK1/HA-treated mice group without NIR irradiation (Figure 6a), indicating the prominent in vivo antitumor capability of PC-UCNPs/siPLK1/HA.There was no noticeable body weight change in the mice during the treatment period (Figure 6b), indicating satisfactory biocompatibility and therapeutic specificity of PC-UCNPs/siPLK1/HA.All mice were euthanized on day 14, and the tumors were collected, weighed, sliced and stained using H&E and TUNEL to observe apoptosis via a CLSM image.PC-UCNPs/siPLK1/HAtreated mice group with NIR light irradiation also had the smallest tumor weight and size (Figure 6c,d) with the maximum disappearance of tumor nucleus in H&E images and the highest level of cell apoptosis in TUNEL images (Figure 6e).The in vivo applicability of PC-UCNPs/siRNA/HA was only evaluated for HepG2 cell tumor bearing mice and demonstrated satisfactory therapeutic efficiency.By conjugating with tumor targeting molecules, such as folic acid and aptamer AS1411, the as-presented PC-UCNPs/siRNA/HA nanoparticles could be widely suitable for a variety of tumor models.
Targets 2023, 1, FOR PEER REVIEW 9 UCNPs/siPLK1/HA-treated mice group without NIR irradiation (Figure 6a), indicating the prominent in vivo antitumor capability of PC-UCNPs/siPLK1/HA.There was no noticeable body weight change in the mice during the treatment period (Figure 6b), indicating satisfactory biocompatibility and therapeutic specificity of PC-UCNPs/siPLK1/HA.All mice were euthanized on day 14, and the tumors were collected, weighed, sliced and stained using H&E and TUNEL to observe apoptosis via a CLSM image.PC-UCNPs/siPLK1/HA-treated mice group with NIR light irradiation also had the smallest tumor weight and size (Figure 6c, d) with the maximum disappearance of tumor nucleus in H&E images and the highest level of cell apoptosis in TUNEL images (Figure 6e).The in vivo applicability of PC-UCNPs/siRNA/HA was only evaluated for HepG2 cell tumor bearing mice and demonstrated satisfactory therapeutic efficiency.By conjugating with tumor targeting molecules, such as folic acid and aptamer AS1411, the as-presented PC-UCNPs/siRNA/HA nanoparticles could be widely suitable for a variety of tumor models.
Synthesis of 5-(2-(dimethylamino)ethoxy)-2-nitrobenzyl acrylate (MENA): Compound 2 (500 mg, 2.08 mmol) was added to the mixture solution of 7 mL anhydrous THF and 333 µL TEA at 0 • C under N 2 protection.Acryloyl chloride (198 µL, 2.185 mmol) in 500 µL anhydrous THF was then added dropwise.After reaction for 5 h at room temperature, the insoluble salts were removed via filtration.The filtrate was concentrated and further purified using silica gel column chromatography (hexane/ethyl acetate, v/v = 3/1) as the eluent.Additionally, 5-(2-(dimethylamino)ethoxy)-2-nitrobenzyl acrylate (MENA) was obtained as a yellowish solid (380 mg, 73% yield). 1   The UCNPs core NaYF 4 :Yb,Tm was synthesized according to the previous report [39].Briefly, YCl 3 (0.78 mmol), YbCl 3 (0.2 mmol), and TmCl 3 (0.02 mmol) were mixed with 6 mL OA and 15 mL ODE, heated to 150 • C and stirred for 2 h to remove oxygen and water.After cooling down to 30 • C, NH 4 F (4 mmol) and NaOH (2.5 mmol) in 10 mL methanol were added dropwise into the mixture solution within 15 min.The reaction solution was then heated to 45 • C for 30 min, 90 • C for 30 min and 310 • C for 1 h under N 2 protection.The reaction solution was then cooled to room temperature and precipitated with the mixture of ethanol and acetone to obtain UCNPs core and re-dispersed in cyclohexane for further use.
For the synthesis of core-shell structured UCNPs NaYF 4 :Yb,Tm@NaYF 4 , YCl 3 (0.08 mmol) were mixed in ODE (15 mL) and OA (6 mL), heated to 150 • C, stirred for 2 h and subsequently cooled down to 75 • C. The above-obtained UCNPs core NaYF 4 :Yb,Tm was added into the reaction mixture, and the reaction mixture was heated to 90 • C to remove the cyclohexane.After cooling to room temperature, NH 4 F (4 mmol) and NaOH (2.5 mmol) in 10 mL methanol were added and continuously stirred at 45 • C for 30 min, at 90 • C for 30 min and at 290 • C for 45 min under N 2 protection to obtain core-shell structured UCNPs NaYF 4 :Yb,Tm@NaYF 4 .The as-obtained UCNPs NaYF 4 :Yb,Tm@NaYF 4 were precipitated with ethanol and re-dispersed in 10 mL of cyclohexane for future use.The morphology of as-prepared UCNPs was characterized by TEM, and its properties were characterized via fluorescence spectrometer.

Preparation of UCNPs@SiO 2 -CTA
The UCNPs@SiO 2 -NH 2 was synthesized via the reversed-phase microemulsion method according to the previous report with a slight modification [40].Briefly, 500 µL CO-520 surfactant was added into 1 mg/mL of above-obtained UCNPs in 20 mL cyclohexane and sonicated for 10 min.A total of 100 µL NH 3 •H 2 O (wt 30%) was added into the above mixture solution and continuously sonicated for 10 min until it formed a transparent solution.Then, 36 µL TEOS was added into the mixture solution, reacted for 3 h and centrifuged.The as-obtained UCNPs@SiO 2 was redispersed in ethanol, 400 µL of (3-aminopropyl)triethoxysilane and 70 µL of NH 3• H 2 O were added and stirred at 80 • C for 8 h.The as-obtained UCNPs@SiO 2 -NH 2 was centrifuged and redispersed in DMSO for further use.

Preparation of Photo-Cleavable Polycations-Wrapped UCNPs (PC-UCNPs)
The above obtained UCNPs@SiO 2 -CTA (1.2 g), poly(oligo(ethylene glycol)) methyl ether acrylate (OEMA, 4.56 g, 9.6 mmol), MENA (744 mg, 2.4 mmol) and AIBN (12 mg) were added to a glass ampoule with anhydrous 1,4-dioxane (19.2 mL).The ampoule was degassed through three freeze-thaw cycles, sealed under vacuum, maintained in an oil bath preheated to 70 • C and kept for polymerization reaction for 24 h.The polymerization reaction was quenched with liquid nitrogen, and the reaction mixture was centrifuged at 8000 rpm for 5 min and washed with methanol.To endow positive charge to the as-obtained UCNPs@SiO 2 -polymer, it (0.30 g) was dispersed in CH 2 Cl 2 (30 mL), slowly added to CH 3 I (0.1 g, 2 mmol) via a dropping funnel, and kept in the dark for 24 h at room temperature.The as-obtained PC-UCNPs was collected after centrifugation, dispersed in 10 mL PBS and stored at 4 • C for further use.The obtained PC-UCNPs were characterized via TEM, DLS and TG.
The gel permeation chromatographic (GPC) analysis of polycations coating of P(OEMAco-MENA) was performed by dispersing above-obtained PC-UCNPs (200 mg) in hydrofluoric acid (HF, 5 mL), stirring it for 6 h at 50 • C and centrifuging it at 8000 rpm for 5 min.The supernatant was then collected, dialyzed (molecular weight cut off, 3.5 kDa) and freeze-dried.The obtained powder (1 mg) was mixed with THF (1 mL, mass spectrum grade) and analyzed by GPC using THF as eluent and polystyrene as reference polymer.
3.8.Preparation of siRNA-Loaded and HA-Wrapped PC-UCNPs (PC-UCNPs/siRNA/HA) siRNA was loaded into the above-obtained PC-UCNPs via electrostatic interaction.After mixing 2 µg of siRNA in RNase-free DEPC with 2 µL of 6 X siRNA loading buffer, PC-UCNPs were added to PBS with a series of mass ratio (PC-UCNPs/siRNA (w/w) of 0, 0.6/1, 0.8/1, 1/1, 1.2/1), incubated at room temperature for 30 min and applied for the agarose gel retardation assay.Hyaluronic acid (HA) was then wrapped to the as-obtained PC-UCNPs/siRNA via electrostatic interaction to obtain PC-UCNPs/siRNA/HA.HA (50 mM) in PBS buffer was added into the prepared PC-UCNPs/siRNA with a mass ratio of 1/1, incubated for 10 min at room temperature and centrifuged at 8000 rpm for 5 min to obtain PC-UCNPs/siRNA/HA.The surface modification process was then characterized using zeta potential.
3.9.In Vitro Verification of HA Degradation and siRNA Release PC-UCNPs/siRNA/HA was incubated with 0.5 mg/mL of HAase solution (HEPES buffer, pH 6.5) at 37 • C, and the degradation of HA was characterized by DLS and zeta potential over prearranged time intervals.To obtain the release profile of siRNA, the as-treated solutions were exposed under 980 nm of irradiation (2 W/cm −2 ) for different durations, and the as-obtained solutions were centrifuged at predetermined time intervals.The supernatant containing released siRNA was measured via UV-Vis spectroscopy.

Lysosomal Escape of siRNA
To verify the lysosomal escape of siRNA, HepG2 cells were incubated with 5 µg/mL of PC-UCNPs/siRNA-FAM/HA (siRNA 10 nM) for 2 h and 6 h, respectively.The as-treated cells were washed with PBS, incubated with 75 nM LysoTracker Green for 30 min, and observed via CLSM.

MTT Assay
To perform MTT assay, HepG2 cells (4000 cells/well) were seeded in a 96-well plate and incubated overnight in 200 µL of DMEM containing 10% FBS and 1% penicillinstreptomycin, the incubation medium was then replaced with 200 µL of fresh DMEM containing serial concentrations (40, 80, 120, 160 and 200 µg/mL) of PC-UCNPs/siRNA/HA and continuously incubated for 24 h.A total of 20 µL MTT (5 mg/mL in PBS) was subsequently added per well and continuously incubated for 4 h.After removing cell culture medium, 80 µL of DMSO was added to dissolve the crystals precipitate.The optical density was measured at 490 nm with a Bio-Rad microplate reader.Moreover, PBS-treated HepG2 cells were set as control group.The relative cell viability (%) was calculated using (A test /A control ) × 100%.
SiRNA' with scrambled sequence and no biotoxicity was loaded to PC-UCNPs, and the as-obtained PC-UCNPs/siRNA'/HA was used to investigate the biotoxicity of PC-UCNPs/siRNA'/HA (siRNA 5 nM).Therapeutic siPLK1 was loaded to PC-UCNPs, and the as-obtained PC-UCNPs/siPLK1/HA (siPLK1 5 nM) was used to evaluate the therapeutic effect of PC-UCNPs/siPLK1/HA.PC-UCNPs/siPLK1/HA-incubated HepG2 cells were irradiated with the 980 nm laser (2 W/cm 2 ) for 2 h (6 min intervals for every 10 min of light exposure to avoid heating) and continuously incubated for 24 h for MTT assay.At the same time, Hela cells were also used to evaluate the therapeutic effect of PC-UCNPs/siPLK1/HA.

Hemolysis Assay of PC-UCNPs/siRNA/HA
Blood samples from healthy mice were collected in a container filled with heparin and centrifuged at 3000 rpm for 6 min to obtain red blood cells (RBCs).After washing and diluting with saline, RBCs (200 µL) were mixed with PC-UCNPs/siRNA/HA (800 µL) in saline at different concentrations (200, 400, 600, 800 and 1000 µg/mL) and incubated at 37 • C.After incubation for 4 h, the mixture solution was centrifuged, and the supernatant absorbance at 570 nm (A sample ) was measured via UV−vis.Moreover, RBCs were mixed with saline and water as a negative control (A saline ) and a positive control (A water ), respectively.The hemolysis percentage was calculated via (A sample − A saline )/(A water − Asaline ) × 100%.

Gene Silencing Assay
HepG2 cells were seeded into a 6-well plate at a density of 5 × 10 5 cells/well and cultured at 37 • C overnight.The old medium was displaced with fresh medium containing PC-UCNPs/siPLK1/HA at siPLK1 concentrations of 250 nM, incubated for 4 h and exposed under 980 nm laser (2 W/cm 2 ) for 2 h (6 min intervals for every 10 min of light exposure to avoid heating).The as-treated HepG2 cells were continuously cultured for 24 h, and the expression levels of PLK1 mRNA and protein were evaluated using RT-PCR and ELISA kit following the manufacturer's instructions.The control group was set as siRNA' (scrambled sequence)-loaded PC-UCNPs/siRNA'/HA-treated HepG2 cells with NIR irradiation and PC-UCNPs/siPLK1/HA-treated HepG2 cells without NIR irradiation for the comparison of gene-silencing effect.Moreover, Hela cells were also adopted to evaluate the gene-silencing effect of PC-UCNPs/siPLK1/HA.

Cell Apoptosis Assay
HepG2 cells were seeded into a 6-well plate at a density of 2 × 10 5 cells/well and incubated with PC-UCNPs/siPLK1/HA at siPLK1 concentrations of 250 nM at 37 • C for 4 h, irradiated upon with a 980 nm laser for 2 h and continuously incubated for 24 h.Then, the as-treated cells were washed with PBS, stained with a mixture of 5.0 µL Annexin V-FITC and 5.0 µL propidium iodide for 10 min and measured with flow cytometry over FL1 (Annexin V-FITC) and FL3 (PI) channels.Control groups were set as siRNA' (scrambled sequence)-loaded PC-UCNPs/siRNA'/HA-treated HepG2 cells with NIR irradiation and PC-UCNPs/siPLK1/HA-treated HepG2 cells without NIR irradiation.

In Vivo Antitumor Efficiency
Pathogen-free female BALB/c nude mice were purchased from KeyGEN BioTECH (Nanjing, China).All experiment procedures were approved by the Model Animal Research Center of KeyGEN BioTECH.To build the tumor model in mice, HepG2 cells were only selected as models since their overexpressed CD44 receptor on the surface of membrane.1.0× 10 7 HepG2 cells were subcutaneously injected to establish HepG2 tumor xenograft mouse model.After the tumor volumes grew to 50 mm 3 , the tumor-bearing mice were randomly divided into three groups, with five mice in each group, and intratumorally injected, respectively, with 50 µL of PBS (1 group), PC-UCNPs/siPLK1/HA (2 groups) at a dose of 1 mmol siPLK1 per mouse.At 2 h post-injection, one of the PC-UCNPs/siPLK1/HA-injected mouse group was exposed to a 980 nm laser irradiation (2 W/cm 2 ) for 2 h (6 min intervals for every 10 min of light exposure to avoid heating).The injection and irradiation processes were repeated every other two days, a total of six times.The tumor sizes were measured every 2 days with caliper.The tumor volumes were calculated as V = (L × W 2 )/2, where L and W are the length and width of the tumor, respectively.To evaluate the safety of PC-UCNPs/siPLK1/HA in mice, the mice weights were also recorded during the therapeutic process.After treatment for 14 days, all mice were sacrificed, and the obtained tumors were then collected, weighed, photographed and stained using H&E and TUNEL.

Figure 2 .
Figure 2. (a) Agarose gel retardation assay for PC-UCNPs/siRNA at various weight ratios (w/w).(b) Zeta potential and DLS analysis of (1) PC-UCNPs, (2) PC-UCNPs/siRNA, (3) PC-UCNPs/siRNA/HA.(c) Zeta potential and DLS analysis of PC-UCNPs/siRNA/HA in response to HAase at different time points.(d) The release curve of siRNA from PC-UCNPs/siRNA/HA as a function of time in response to NIR irradiation.The error bars indicate means ± SD (n = 4).