Effects of Recovery Time during Magnetic Nanofluid Hyperthermia on the Induction Behavior and Efficiency of Heat Shock Proteins 72

In this study, we investigated the effects of recovery time during magnetic nanofluid hyperthermia (MNFH) on the cell death rate and the heat shock proteins 72 (HSP72) induction behavior in retinal ganglion cells (RGCs-5) to provide a possible solution for highly efficient ocular neuroprotection. The recovery time and the heat duration time during MNFH were systematically controlled by changing the duty cycle of alternating current (AC) magnetic field during MNFH. It was clearly observed that the cell death rate and the HSP72 induction rate had a strong dependence on the recovery time and the optimizated recovery time resulted in maximizing the induction efficiency of HSP72. Controlling the recovery time during MNFH affects not only the cell death rate but also HSP72 induction rate. The cell death rate after MNFH was dramatically decreased by increasing the recovery time during MNFH. However, it was also found that the HSP72 induction rate was slightly decreased by increasing the recovery time. These results indicate that applying the appropriate or optimized recovery time during MNFH can improve the induction efficiency of HSP72 by minimizing the cell death caused by cytotoxic effects of heat.

cells have been shown to be biologically related with the exposure time (heat duration time) and intensity of the heat. Therefore, it is expected that the cell death rate can be minimized by controlling the intensity of the heat and its duration time in cells during hyperthermia, thereby improving the induction efficiency of HSPs.
In this study, we systematically controlled the ratio of the recovery time (τ R ) to the heat duration time (τ H ) of AC magnetic field (or AC heating stress) during magnetic hyperthermia to explore the effects of recovery time on the induction efficiency behavior of HSP72 in RGCs-5 for ocular neuroprotection in future glaucoma clinics. Magnetic nanofluid hyperthermia (MNFH) with tailored Mn 0.5 Zn 0.5 Fe 2 O 4 (T-Mn 0.5 Zn 0.5 Fe 2 O 4 ) superparamagnetic nanoparticles (SPNPs), which have a good electrical field absorption and a large power loss (200-500 W/m 3 ) at a low frequency (<100 KHz) 21 , was employed as a hyperthermia modality to effectively control the τ R and τ H of the AC magnetic field during the HSPs induction process in RGCs-5 (Fig. 1a). In order to control the ratio of τ R to τ H of the AC magnetic field, the duty cycle (D) value of the applied AC magnetic field was systematically varied from 0.25 (25%) to 1 (100%) during MNFH (Fig. 1b). The induction behavior of HSP72 including the induction rate, the induction efficiency, and the cell death characteristics during HSPs induction process were studied by changing the τ R during MNFH. The optimal D value (or the ratio of the τ R to the τ H ) to provide highly efficient induction of HSP72 in RGCs-5 was empirically investigated.

Results and Discussion
T-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles were prepared using a modified one-pot thermal decomposition method, which includes metal precursors and reductant in the presence of surfactants and solvent. As can be clearly seen in Fig. 2a, the synthesized T-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles had a spherical (round) shape with a mean particle (core) size of 6.5 nm ± 0.78 nm. In addition, as shown in Fig. 2b, high-resolution transmission electron microscopy (HR-TEM) analysis confirmed that the synthesized T-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles have a highly oriented typical cubic spinel structure with (400) and (220) preferred lattice planes. Crystal structure was determined using an X-ray diffractometer (XRD), and doping level of Mn 2+ or Zn 2+ ions was qunatititively analyzed by using an energy dispersive X-ray spectroscopy (EDS) and an inductively coupled plasma atomic emission spectroscopy (ICP-AES) (Spplementary Information Fig. S1). Conventional Mn 0.5 Zn 0.5 Fe 2 O 4 (C-Mn 0.5 Zn 0.5 Fe 2 O 4 ) nanoparticles were also explored for comparison. C-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles obtained using a conventional experimental method showed similar particle size, size distribution and particle shape to those of T-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles (Spplementary Information Fig. S2). Figure 2c shows the direct current (DC) minor hysteresis loops of as-synthesized T-Mn 0.5 Zn 0.5 Fe 2 O 4 and C-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles (solid state) measured at a sweeping field of H appl = ± 140 Oe. The two nanoparticles did not exhibit any DC minor hysteresis indicating that the synthesized T-Mn 0.5 Zn 0.5 Fe 2 O 4 and C-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles have typical superparamagnetic characteristics. However, as can be seen in Fig. 2d, a larger AC hysteresis loss was observed from the T-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles. This result indicates that T-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles are magnetically softer than that of C-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles under AC magnetic field. Furthermore, it implies that T-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles would generate a higher AC magnetically-induced heating temperature since the larger AC hysteresis loss area of those indicating that it has a higher AC softness, which is directly related to the larger "Néel relaxation loss power" 22 . Figure Figure 2f shows the measured DC hysteresis loop of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG (40 μg/100 μL D.I. water) nanofluids. The observed high mono-dispersity of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanofluids can be confirmed from the DC magnetic hysteresis loop, because agglomerated nanoparticles in nanofluids can generally cause undesirable changes in dipole-dipole interactions or degradation of magnetic properties directly resulting in asymmetric hysteresis behaviors 22,23 . The measured hysteresis was symmetric and severe degradation was not observed. No severe magnetic degradation or no change of magnetic intrinsic properties compared to those of solid state indicating that the PEG coated T-Mn 0.5 Zn 0.5 Fe 2 O 4 SPNPs are well dispersed in the nanofluid and form chemically stable colloidal suspension. Furthermore, this well controlled coating status of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG SPNPs is expected to improve the efficiency of the cellular uptake and the cell viability associated with cell cytotoxicity [24][25][26] .
Prior to controlling D value of the AC magnetic field during MNFH with RGCs-5 to study the effects of recovery time on the rate of cell death and the corresponding efficiency of the induction of HSP72, the T AC,mag and intrinsic loss power (ILP) of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG SPNPs were measured in three different media (ethanol, D.I. water, and RGCs-5 cells) to explore AC self-heating characteristics of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG SPNPs under different microenvironments. For this measurement, the micro-centrifuge tube containing the T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG SPNPs with ethanol, D.I. water, and RGCs-5, was placed at the center of AC magnetic coil ( Fig. 3a and b). The concentration of the T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanofluid was a 5 mg/mL and the applied f appl , and H appl were fixed at a 110 kHz, and a 140 Oe, respectively. After the power is turned on, the T AC,mag of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanofluids was rapidly increased and saturated at a 43 °C (in ethanol) and a 41.3 °C °C (in D.I. water) within 1000 sec. with different AC heating up rate due to the different specific heat capacities and viscosities of the two nanofluids (Fig. 3c). In addition, after the AC magnetic field was removed, T AC,mag was decreased sharply for both two nanofluids. To measure the T AC,mag of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG SPNPs in RGCs-5, RGCs-5 were incubated with a 5 mg/10 mL (DMEM) of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanofluid for 24 hours at 37 °C. After the incubation process, the RGCs-5 uptake with T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG SPNPs were centrifuged to form a cell pellet and a 100 μL of DMEM was added to the RGCs-5 pellet. The T AC,mag of RGCs-5 pellet was measured at a fixed AC magnetic field of f appl = 140 kHz and H appl = 170 Oe. The T AC,mag of the RGCs-5 pellet treated with T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG SPNPs was increased up to a 40 °C within a 122 sec, and then it was stably saturated at a typical HSPs induction temperature of 40.5 °C ± 0.5 °C by adjusting the intensity of H appl (Fig. 3d).
The ILP values of the T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG SPNPs obtained in ethanol, D.I. water, and RGCs-5 were a 3.5 nHm 2 kg −1 , a 3.7 nHm 2 kg −1 , and a 2.4 nHm 2 kg −1 , respectively (Fig. 3e). The differences in obtained ILP values of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG SPNPs in different microenvironments (ethanol, D.I. water, and RGCs-5 cells) directly indicates that the contribution of "Néel" and "Brownian" relaxation heating power to the total heating power is dependent on the nanofluidic ambient environment. A higher viscosity resulted in a lower heating power due to the reduction of "Brownian relaxation loss power" caused by the increase of Brownian relaxation time. This indirectly implies that the "Néel relaxation loss power" would be dominant in characterizing the AC heat generation of superparamagnetic nanoparticles in intercellular or in-vivo hyperthermia 27,28 . In addition, the higher ILP of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanofluid allows us to apply it for a MNFH agent to systematically and readily control the D value (or τ R and τ H ) to adjust the induction characteristics during MNFH so that the highly efficient induction of HSP72 in RGCs-5 can be expected to be achieved.
After After MNFH was carried out at the different conditions of D values of AC magnetic field in RGCs-5, the Western blot analysis was carried out to identify the induction of HSP72 and subsequently, the dependence of HSP72 induction. In addition, the cell death rate depending on the D values of AC magnetic field was investigated by using optical microscope images of the stained RGCs-5. The inset image in Fig. 4e shows the Western blot result of the RGCs-5 after MNFH. As can be seen in the Figure, a strong immunoreactivity for HSP72 was detected. Figure 4e-h show the images for the stained HSP72 and nuclei of the RGCs-5 treated with MNFH using T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanofluid (experimental group, left) and the RGCs treated with only AC magnetic field but no T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG (control group, right). The right images (Fig. 4i-l) apparently show that the HSP72 were not induced in the control group. However, the images on the left side clearly demonstrate that the HSP72 were successfully induced by MNFH. The HSP72 induction rate (defined by "visible number of HSP72") was monotonically increased by increasing the D value of applied AC magnetic field, but in contrast, it was also observed from the microscopic images that the cell survival rate (defined by "visible number of cells or nuclei") was gradually decreased by increasing the D value of applied AC magnetic field (or AC heating stress). These results clearly illustrate that the change of D value during MNFH directly influence on the extent of the induction of HSP72 as well as the cell death behavior in RGCs-5 during MNFH induced heat shock protein process. In order to numerically analyze the effects of controlling D value of the applied magnetic field on the induction efficiency of HSP72, the relative numerical calculation methods were employed for quantitative analysis. First, the HSP72 induction rate (denoted by "R H " = (visible number of HSP72/visible number of nuclei) × 100 (%)) and the cell (nucleus) death rate (denoted by "R C " = 1 -(visible number of nuclei after MNFH / visible number of nuclei in the control group) × 100 (%)) were calculated using the optical microscopic images of stained RGCs-5 by considering for both experimental and control groups, and accordingly, the induction efficiency, η, of HSP72 was obtained from the analyzed R H and R C , as shown in equation (1).
H C Table 1 shows the analyzed results of R H and R C in RGCs-5 after MNFH in which the D value of the applied AC magnetic field was changed from 0.25 (25%) to 1 (100%). According to the analyzed results, the D value of 25% (duty factor: 0.25) exhibited the lowest R H of 10.3% and the R C of 22.1%, while the D value of 100% (duty factor: 1) had the highest R H of 26.9% and the R C of 65.4%. In addition, Table 1 shows that the both R C and R H were decreased by increasing the τ R . (In other words, the R C and R H were proportional to the heat duration time, τ H ). This result simply indicates that the induction behavior (induction rate and cell death rate) of HSP72 in RGCs-5 has a dependence on the duty cycle of applied AC magnetic field. However, it is interestingly noted from Table 1 that the R C (cell death rate) was more significantly reduced than that of R H by increasing the recovery time (τ R ). Accordingly, we speculate that the increase of the τ R during MNFH can contribute to improve the cell survival rate by allowing the cells to have the time to recover from the applied AC magnetic field. Thus, in order to achieve a high η of HSP72, the ratio of the τ R to the τ H during MNFH in RGCs-5 was tried to be optimized. The η of HSP72 determined by the ratio of the τ R to the τ H (or duty factor) was calculated using equation 1. As shown in Table 1, the optimal ratio of the τ R to the τ H for the highest η of HSP72 induction was found to be a 1: 1 (duty factor D = 0.5 (50%)). This result demonstrates that applying the same period of τ R as the τ H to the AC magnetic field during MNFH can maximize the η of HSP72 induction by reducing the cell death rate, R C . It is well known that cells respond to various kinds of stresses in selective ways that help them to either survive the stressful stimuli or signal cell death pathway that can eventually eliminate injured cells. Whether cells mount a protective or detrimental stress response is determined by the cell type and especially, by the nature and the duration of the stress 29,30 . To our knowledge, our study is the first report that has manipulated the duration of applied AC magnetic field (AC magnetic heating stress) by introducing recovery time, during MNFH to induce HSP72 in RGCs-5. Our research on the effects of recovery time during MNFH induced HSP72 in RGCs-5 clearly demonstrates that an appropriate recovery time is required to minimize the death of cells while maximizing the induction rate of HSP72. When the same period of τ R and τ H is applied to the AC magnetic field, it was clear that cells undergo heat absorption time and heat dissipation time recurrently at regular intervals. This could provide a large population of cells not only with adequate heat duration time (τ H ), but also with enough recovery time (τ R ) from the stress before entering into cell death pathways, reducing the undesirable side effects of MNFH and eventually leading to the highest η of HSP72. In light of this, this research strongly supports the clinical feasibility of MNFH that induces the localized HSPs for the efficient ocular neuroprotection modality in glaucoma clinics.

Conclusions
We investigted the effects of τ R during MNFH on the induction rate of HSP72 and cell death behavior of RGCs-5 to achieve the most efficient η of HSP72 induction for clinical applications. It was experimentally demonstrated that the controlling τ R during MNFH affects not only the cell death rate but also HSP72 induction rate. The cell death rate after MNFH was dramatically decreased by increasing the τ R during MNFH. However, it was also found that the HSP72 induction rate was slightly decreased by increasing the τ R . These results indicate that applying the appropriate or optimized τ R during MNFH can improve the η of HSP72 by minimizing the cell death caused by the cytotoxic effect of the heat. Furthermore, this study is expected to provide a useful solution for the safe and effective induction of HSP72 for the highly efficient ocular neuroprotection for glaucoma. Duty cycle value (D) Cell death rate (Mean ± SD) HSP72 induction rate (Mean ± SD) HSP72 induction efficiency (%) 0. 25 22. SPNPs. In order to coat the nanoparticles with PEG, methoxy-PEG-silane 500 Da, and triethylamine, which were purchased from Gelest Inc., and Sigma-Aldrich, respectively, were used. The conventional Mn 0. 5  were mixed in a three neck flask and magnetically stirred at room temperature. The ramping up rate, and the heat treatment time in the nucleation process (process temperature: ~200 °C) were 11 °C/min, and 60 min, respectively. Then, the reaction solutions were heated again up to 296 °C for 30 min (3.2 °C/min, the second ramping rate) and the heat treatment time in the growth process (process temperature: ~296 °C) was 46.5 min. After the heat treatment process, the mixed solution was cooled down to room temperature and ethanol (40 mL) was added to the mixed solution to rinse the synthesized nanoparticles. The rinsed nanoparticles were collected by centrifugation and then dried at room temperature. The synthesized T-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles were coated with PEG to form a nanofluid. For coating the PEG layer, the nanoparticles were first dispersed in toluene (7.5 mL), and then PEG (0.75 mL) and triethylamine (catalyst, 3.75 mL) were added to the toluene containing the nanoparticles. The mixed solution was shaken well for 24 hours at room temperature, and then the PEG coated nanoparticles were collected by centrifugation and dispersed in D.I. water. AC magnetically-induced heating temperature, T AC,mag , and intrinsic loss power, ILP, of T-Mn 0.5 Zn 0.5 Fe 2 o 4 @PEG nanoparticles. The T AC,mag of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanofluids was measured in ethanol, D.I. water, and RGCs-5 using an AC magnetic field generation system consisting of AC coils, capacitors, DC power supplies, and wave generators (anytech. Korea). The applied frequency (f appl ), and H appl were 110 kHz ~ 140 kHz, and 140 Oe ~170 Oe, respectively. The ILP of the T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanoparticles was determined based on the measured T AC,mag and using equation (2)  The SLP was determined based on the measured T AC,mag and using equation (3)  Cell culture, in-vitro cytotoxicity assay, and cellular uptake. The rat RGCs-5 were cultured in Dulbecco Modified Eagle Medium (DMEM, Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum (FBS, Introgen), 100 μg/ml penicillin/streptomycin (Invitrogen), and 2 mM glutamine (Invitrogen) in a humidified incubator with 5% CO 2 at 37 °C. The in-vitro cytotoxicity assay of RGCs with T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanoparticles was conducted at the concentration of 0, 1, 3, 5, 10, 30, 100, 250, and 500 μg/mL using a Cell Counting Kit-8 (CCK-8) assay (Fig. S3). The RGCs-5 cells were exposed to the nanoparticles for one day. For cellular uptake of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanoparticles by RGCs-5, the RGCs-5 was differentiated with 1 μM staurosporine for 1 hour and incubated again with a 500 μg/mL of T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanoparticles in a 10 mL of DMEM for 24 hours. The RGCs-5 containing T-Mn 0.5 Zn 0.5 Fe 2 O 4 @PEG nanoparticles were washed with phosphate buffered saline (PBS) and trypsinized with trypsin-EDTA (Life technologies, USA) followed by addition of 100 μL of DMEM, and then collected in a microcentrifuge tube and centrifuged to form a cell pellet. The volume of the