Ultrasound-triggered Release of Calcein and Doxorubicin from HER2-targeted Liposomes in Breast Cancer Therapy

The functionalization of liposomes with antibodies is a potential strategy to increase the speci�city of liposomes and reduce side-effects of chemotherapeutic agents. The active targeting of Human Epidermal growth factor Receptor 2 (HER2) positive breast cancer cells can be achieved by coating liposomes with an anti-HER2 monoclonal antibody. In this study, we synthesized Calcein and Doxorubicin loaded immunoliposomes functionalized with the monoclonal antibody Trastuzumab. Both liposomes were characterized for size, phospholipid concentration and antibody conjugation. The effect of low-frequency ultrasound (LFUS)-induced drug release was tested using three power densities, 7.46, 9.85 and 17.31 mW/cm 2 ; and the release data were modeled using six different kinetic models. LFUS results established the sonosensitivity of both carrier types, with immunoliposomes being more acoustically sensitive than control liposomes. Results also showed an increase in the release rate as the power density increased from 7.46 to 17.31 mW/cm 2 . Finally, the in vitro cell experiments showed enhanced uptake and cytotoxicity when breast cancer cell lines, SKBr3 and MDAMB-231, were treated with LFUS-triggered HER-liposomes.


Introduction
Breast cancer is the most common type of cancer in females, constituting the second leading cause of death globally 1,2 . The treatment of breast cancer is complex and relies on several factors, including the type of tumor, tumor size, grade, proliferation rate and lymph node status. Existing treatment methods include surgery, chemotherapy, radiotherapy and hormonal treatments 3 . In recent years, the presence of breast cancer tumor markers has been investigated, and several markers were identi ed, including the estrogen receptor (ER), progesterone receptor (PR) and the human epidermal growth factor receptor 2 (HER2) 3,4 . Human epidermal growth factors belong to a family of transmembrane receptor tyrosine kinases (RTKs). HER2 is involved in important stages of growth and cell differentiation and is normally expressed at low levels in the epithelial cells of various organs such as the lungs, bladder, pancreas, breast, and prostate 5,6 . HER2 overexpression can lead to malignant progression, which justi es the unfavorable prognosis of breast, ovarian, gastric, and prostate cancers 3 . HER2 overexpression is present in approximately 25% of all breast cancers and is usually associated with more aggressive disease and endocrine therapy resistance. Studies showed that HER2 is uniformly distributed within the tumor and its expression induces signi cant apoptosis in breast cancer cells. These characteristics make it an appealing choice as a target in targeted therapy 3,5,6 .
One of the signi cant advancements in breast cancer treatment was the development of the humanized monoclonal antibody Trastuzumab, sold under the brand name Herceptin among others, which was FDA approved in 1998 for the treatment of HER2-positive breast cancers [11][12][13] . Trastuzumab can enhance the effects of chemotherapy and reduce the risk of recurrence when used as an adjuvant therapy [14][15][16] . This immunotherapeutic is a humanized IgG(1) kappa monoclonal antibody (145.5 kDa) with a high and speci c a nity towards HER2 receptors 7 . Trastuzumab can prevent HER2 hetero-dimerization and stop tumor development cell signaling via several mechanisms, including reduced PI3K/Akt signaling, enhanced degradation of HER2 receptors, and antibody-dependent cellular cytotoxicity (ADCC). Hence, this drug is considered a HER2 receptor antagonist [8][9][10] .
With regard to novel cancer treatment methods, smart drug delivery systems (SDDSs) have gained prominence in the eld of targeted drug delivery 17,18 . SDDSs are nanostructures capable of reducing drug side effects, increasing blood circulation time, and increasing drug concentrations at target sites, thus ensuring better patient compliance 19,20 . SDDSs release their payloads at target sites in response to internal or external triggers such as temperature, pH, enzymes, light, mechanical waves (e.g., ultrasound) or magnetic elds [21][22][23][24][25] . Several nanostructured delivery systems have been studied for cancer treatment, such as liposomes, solid lipid nanoparticles, metal organic frameworks, micelles, dendrimers, and quantum dots 26 .
Liposomes are one of the most widely used nanocarriers in drug delivery 27 . Liposomes are nanosized to microsized concentric spheres of phospholipid bilayers separated by aqueous compartments. Liposomes assume this bilayer conformation in order to shield the hydrophobic tails of phospholipids from aqueous environments 28,29 . Furthermore, liposomes can be functionalized by conjugating different moieties to their surfaces 30,31 . For instance, the circulation-time of liposomes can be increased and their detection by the reticuloendothelial system (RES) reduced by conjugating stealth-imparting polymers to their surfaces [32][33][34] . The most widely used polymeric substance is poly-ethylene glycol (PEG). Another example of such modi cations involves the conjugation of ligands to the surfaces of liposomes to enhance their selectivity and tumor targetability 35 . These ligands have a binding a nity toward receptors, which tend to be overexpressed on the surface of cancer cells 5,33 . Some commonly used targeting ligands include carbohydrates, proteins, aptamers, antibodies, and peptides 36 .
Liposomes that have monoclonal antibodies or antibody fragments conjugated to their surfaces are called immunoliposomes 37,38 . Each antibody binds to a speci c antigen, making it possible to selectively target that cell, as well as reduce the undesirable interactions with healthy cells 39 . The presence and type of antigen vary according to the malignancy, which affects the e cacy of immunoliposomal therapy.
Highly investigated targets for antibodies include the epidermal growth factor (EGFR), HER2, the vascular endothelial growth factor (VEGFR), and the prostate-speci c membrane antigen (PSMA) 40,41 . As discussed above, once SDDSs reach their target sites, they release their contents when exposed to either an internal or external stimulus 23 . Several external triggers have been investigated; however, the trigger of choice in this work is ultrasound (US). US is a cyclic sinusoidal acoustic wave with frequencies higher than the human hearing range (>20 kHz) 42 . US waves have a high-pressure phase (compression) at the upper peaks and a low-pressure phase (refraction) at the lower peaks 43,44 . In therapeutic applications, the biological effects of US can be either thermal or mechanical 45 . Thermal effects result are caused by energy dissipation, while the mechanical effects occur because of the acoustic wave propagation and pressure variations 46,47 . The mechanical effects of US manifest as acoustic cavitation events, which is the formation of gas bubbles due to changes in pressure and is generally divided into stable and transient cavitation 45 . In stable cavitation, the bubble expands and contracts about an equilibrium radius; however, in transient cavitation, the bubble grows rapidly in size and eventually collapses violently. Cavitation depends on the intensity of the US, and occurs only when a certain threshold (determined using the mechanical index) is reached. The mechanical index (MI) is a parameter that quanti es the probability of transient cavitation happening and is calculated using the following equation 5,6,43 .
Where P neg is the negative pressure in MPa, and f is the frequency in MHz.
In triggered drug delivery, the aim is to nd the optimum US parameters that permit enhanced drug delivery without harming healthy cells. This could be better achieved by disrupting the nanocarriers, enhancing drug distribution in tumor tissue, or increasing the nanocarriers' cellular uptake by enhancing the cell membrane permeability 5,48 . On the other hand, controlled drug delivery systems are systems that release their contents at a predetermined rate, locally or systemically, for a speci ed period of time. In this context, mathematical modeling can provide insight into the mechanisms by which the drug is released.
Common release mechanisms include: release by drug diffusion through the polymer membrane, release through the degradation of the polymer, or release through the chemical disassociation of the drug 5,49 . Release kinetics are usually in uenced by the type of drug, particle size, solubility, and the amount of the drug used. In this research, six kinetic representations were used to model the low-frequency acoustic release of the model drug calcein and the chemotherapeutic drug Doxorubicin (DOX) from immunoliposomes. As will be discussed in the following sections, release was measured by monitoring the changes in uorescence. The percent of drug released from the liposomes was represented by the Cumulative Fraction Release (CFR), as follows: Where I 0 represents the baseline intensity, I t represents the intensity at time, t, and represents the highest uorescence intensity value obtained. In this study we investigated the feasibility of enhancing DOX delivery to breast cancer cells using Herceptin and US.

Preparation of control liposomes
Liposomes were prepared using the thin-lm hydration method. Brie y, liposomes were prepared using cholesterol, DPPC, and DSPE-PEG(2000)-NH 2 at molar ratios of 30:65:5, respectively. The lipids were dissolved in 4 ml chloroform in a round-bottom ask. The chloroform was then evaporated using a rotary evaporator under vacuum at 50 ºC for 15 minutes, until a thin lm was observed on the walls. Next, the lipid lm was hydrated using 2 ml of a 30-mM calcein solution and the pH adjusted to 7.4. To obtain unilamellar vesicles, the solution was sonicated for 2 minutes using a 40-kHz sonicator bath (Elma D-78224, Melrose Park, IL, USA). Whereas for particle size reduction, liposomes were extruded using 200-nm polycarbonate lters (Avanti Polar Lipids, Inc., Alabaster, AL, USA). The puri cation of the formulation was performed using a Sephadex G-100 column (after equilibrating it with borate buffer pH 8.5). Finally, the collected fractions were stored at 4°C 5 .
The same protocol described above was used to prepare liposomes encapsulating ammonium sulfate instead of calcein. The ammonium sulfate method was used to load DOX into the formed liposomes. A 0.11-M solution of ammonium sulfate at a pH of 5.5 was prepared to hydrate the dry lipid lm.
Liposomes were puri ed with a Sephadex G-25 gel ltration column previously equilibrated with HEPES buffer (0.26 M Sucrose, 0.005 M ascorbic acid and 0.016 M HEPES). Next, DOX was added to the liposomal solution providing a DOX to lipid ratio of 1:6 (w/w). The liposome-DOX solution was kept in a water bath at 60℃ for 45 minutes with mild stirring. The resulting solution was then centrifuged through a Sephadex G-25 gel ltration column equilibrated with PBS (pH~7.4) 6 .

Preparation of HER-liposomes
The functionalization of the liposomes with HER was performed using a double substitution reaction. Calcein and DOX liposomes were prepared separately using the procedures detailed in the previous section. As for the functionalization procedure, cyanuric chloride was dissolved in acetone to make a 10 mg/ml solution, of which 9.23 μl was diluted in 0.5 ml deionized water. This solution was then added to 1 ml of the liposomal formulations. The reaction was conducted at a pH of 8.5 and 0 °C and was left to stir for 3 hours, allowing the nucleophilic substitution of the chloride particle on the cyanuric chloride with the proton on the NH 2 group of the liposomes. The second reaction involves the linking of the N-terminus on the amino acids of HER to cyanuric chloride. One milligram of HER was dissolved in 0.5 ml borate buffer (pH 8.5), this solution was then added to the liposomes, and the reaction was kept stirring overnight. Finally, excess HER and any free drugs were puri ed in a Sephacryl S-200 HR column equilibrated with PBS (pH of 7.4). The liposomes were collected and stored at 4 ºC 5,6 .

Particle size and polydispersity evaluation
The particle size and polydispersity index of liposomes and immunoliposomes were measured at 25 °C. The intensity-weighted hydrodynamic radius was determined using DynaPro NanoStar (Wyatt Technology Corp., Santa Barbara, CA, USA).

Estimation of phospholipid content
The phospholipid content of liposomes was determined using the Stewart Assay. The mixing of a phospholipid-containing chloroform solution with ammonium ferrothiocyanate at room temperature yields a colored complex that partitions in the chloroform phase and whose maximal absorbance is 485 nm. The liposome samples were dried under vacuum and then dissolved in chloroform. This solution was then sonicated to break the liposomes to their constituent lipids. The liposomes-chloroform solution was transferred to a centrifuge tube where 2 ml of ammonium ferrothiocyanate was added. The centrifugation step resulted in a biphasic system; the top dark layer was removed and discarded, while the bottom clear chloroform layer was transferred to a quartz cuvette, and its optical density measured using ultraviolet-visible (UV-Vis) spectroscopy at A max =485 nm against chloroform as a blank. An average of the six measurements was taken.

Antibody conjugation e ciency in immunoliposomal formulations
HER conjugation e ciency to liposomes was determined by the Bicinchoninic Acid Assay (BCA). The BCA reagent was prepared by mixing QuantiPro QA buffer, QuantiPro QB, and CuSO 4 in a ratio of 25:25:1, respectively. One milliliter of the reagent was added to 1 ml of PBS and 100 μl of the liposomal solution, followed by incubation at 60 ºC for 1 hour. The optical density of the samples was measured using UV-Vis spectroscopy at 562 nm. Moreover, the molecular weight of Trastuzumab and DPPC, in addition to the results obtained from the BCA and Stewart assay, were used to determine the number of Trastuzumab molecules per vesicle.

Low-Frequency ultrasound release studies
The release of calcein and DOX from liposomes was triggered using a 20-kHz low-frequency ultrasonic probe (model VCX750, Sonics & Materials Inc., Newtown, CT) and monitored by uorescence changes using a QuantaMaster QM 30 Phosphorescence Spectro uorometer (Photon Technology International, Edison NJ, USA). Power densities were measured using a hydrophone ((Bruel and Kjaer 8103, Decatur, GA). Calcein is a uorescent molecule with excitation and emission wavelengths of 495 and 515 nm, respectively. As for DOX the emission wavelength is 595 nm, while its excitation wavelength is around 485 nm. The sample to be tested was prepared by diluting 75 μL of liposomes in 3 mL of PBS in a uorescence cuvette. The initial uorescence intensity I o was measured for 60 seconds before sonication.
Then, US was applied in a pulsed mode, with 20 seconds on and 10 seconds off for calcein liposomes and 20 seconds on, 20 seconds off for DOX liposomes to account for the thermal effects of US. The release was performed using three different power densities, 7.46, 9.85, and 17.31 mW/cm 2 . The pulsed US cycles mentioned earlier were continued until a uorescence plateau was reached, at which point 50 μL of Triton X-100 (Tx100) were added to the sample to lyse liposomes and release all the encapsulated contents, simulating 100% drug release.

MTT assay
SKBr3 (HER2+ cells) and MDA MB-231 (HER2-cells) were cultured in DMEM and RPMI media, respectively, supplemented with 10% Fetal Bovine Serum and 1% penicillin/streptomycin. Cells were grown at 37°C under 5% CO 2 . Aliquots containing 1x10 4 cells per well were seeded into 96-well plates and incubated for 24 hours prior to treatment to ensure proper cellular con uency. After 24 hours, different treatments were added in triplicates to the cells with a concentration of 8 µM per well. The cells were treated with free DOX, DOX-loaded liposomes and DOX-HER liposomes, then incubated for 5 hours. Following incubation, the plates were subjected to continuous LFUS for 20 seconds in a 35-kHz bath. One plate of each cell line was not subjected to US to serve as a control. Next, the plates were incubated for 48 hours, after which the culture medium was replaced with the MTT medium, containing 20 l of the sterile MTT dye (5 mg/ml), and further incubated at 37 ∘ C for 4h. After the incubation, 100 l of DMSO (Sigma, USA) was mixed with the medium and incubated for 15 min. Finally, absorbance values were read using a microplate reader (AccuReader, Nangang, Taipei, Metertech, Taiwan) at 570 nm. For each group, three replicates were analyzed, and a mean value was estimated. The cell viability was calculated by dividing the optical density (OD) value of the experimental group by the mean OD of the control group and multiplying it by 100%.

Flow cytometry analysis
For the ow cytometry assay, both SKBR-3 and MDA-MB-231 cells (2x10 5 cells/ml) were seeded in 6-well plates for 24 h. On the following day, the cells were treated with the control and HER-DOX liposomes and further incubated for 3.5 hours, after which the cells were sonicated with a LFUS bath at a frequency of 35 kHz for 45 sec and a power density of 20 mW/cm 2 . One plate from each cell line was treated with both types of liposomes but was not exposed to US to serve as a reference. Following sonication, the cells were again incubated for 1 h. Trypsin was then added to the plates to detach the cells, which were resuspended in PBS to prepare them for the ow cytometer analysis.

Statistical Analysis
Results were reported as average ± standard deviation (SD). One-way ANOVA tests were used to compare the sizes of the control and immunoliposomes; while two-factor ANOVA tests were employed to analyze LFUS release results and kinetic modeling ndings. Both types of ANOVA tests assumed that both populations have similar variances, and two values are considered statistically different if p<0.05 and if F<F critical (unless otherwise stated).

Liposomes characterization
Dynamic light scattering (DLS) measurements were performed to ensure the formation of liposomes and to measure the size of both types of synthesized liposomes. A polydispersity (Pd) index upper limit of 20% is generally acceptable for DLS measurements. Table 1 summarizes the average sizes of the three batches of liposomes with their respective standard deviations. Based on these ndings, both types of liposomes (both calcein and DOX-loaded) are well within the size range for the EPR effects to take place. A one-way ANOVA analysis was conducted to compare the radii of the NH 2 liposomes and the immunoliposomes. The results showed a higher F-value compared to F critical , as well as a p-value lower than the standard alpha value of 0.05, which indicated that the two types of nanoparticles had statistically different radii.
To con rm the conjugation of Trastuzumab to liposomes, the results from the BCA and Stewart assays were used. As mentioned above, the BCA assay was used to determine the protein concentration in the samples in μg/ml, while the Stewart assay was used to determine the DPPC concentration in mg/l. According to the BCA results, HER-liposomes exhibited a 1.5-fold increase in protein concentration compared to control liposomes (refer Figure 1). Combining the ndings of both assays yielded a weightto-weight (w/w) ratio of proteins to lipids in μg/mg. The results of three batches were averaged to con rm conjugation consistency and calculate the standard deviation. Therefore the average number of lipid molecules constructing a single liposome vesicle would be 80,000 50 . Knowing the concentration of Trastuzumab and the lipids making up the liposomes, as well as their respective molecular weights, it was found that almost 9 Trastuzumab molecules were conjugated per liposome, which is considered well within the optimal range necessary to induce su cient cells cytotoxicity 51  LFUS release was performed on three batches of control liposomes and immunoliposomes (with three replicate measurements for each batch). As mentioned above, US was applied in a pulsed mode, with 20 seconds on and 10 seconds off for calcein liposomes and 20 seconds on, 20 seconds off for DOX liposomes. The plateau after every pulse signi es no release when ultrasound is "off". The release trends for calcein control liposomes are depicted in Figure 2 and Table 2. Figure 2 shows that as the power density increases, the release rate becomes steeper, which is expected due to an increase in cavitation events with increased power density. It can also be seen that the control liposomes released most of their contents (~85 %) within 3 minutes, indicating high sonosensitivity. Additionally, the release trend at 17.31 mW/cm 2 can be seen to decrease slightly after the plateau. This behavior may be attributed to the thermal effects experienced by liposomes. Similar results were obtained for the LFUS release from immunoliposomes. As observed with control liposomes, the release rate of immunoliposomes also increased with increasing power density. Figure 2 and Table 2 illustrate the normalized online release rate of calcein from immunoliposomes, at the three investigated power densities, for three batches of liposomes. The number of pulses needed to release the Trastuzumab liposomal contents decreased with increasing power density, which can be explained by the increase in cavitation events as the power density increased. In addition, as with control liposomes, the release pro le can be seen to decrease slightly after the plateau at the power density of 17.31 mW/cm 2 , which again can be attributed to the thermal effects.
When comparing the LFUS release behavior of the control liposomes and immunoliposomes, it can be noted that both types of liposomes demonstrated what could be considered "ideal" release pro les for controlled drug release systems. However, immunoliposomes were more acoustically sensitive, releasing around 92% of the model drug, compared to approximately 86% released from control liposomes. We attribute this behavior to the presence of the antibody, which appears to slightly destabilize the Power Density (mW/cm 2 ) CFR at plateau membrane making the liposomes more susceptible to acoustic mechanical waves. This enhanced sensitivity of immunoliposomes to US waves is advantageous because therapeutic drug levels could be reached in shorter exposure times.
Similar results were obtained with the release from DOX-loaded liposomes. The release increased with increasing power density for both types of liposomes, and HER-DOX liposomes showed higher sonosensitivity. Figure 3 and Table 3 summarize the ndings for both control and HER-DOX liposomes. Finally, the mechanical index (MI), is a parameter used to indicate the possibility of the occurrence of cavitation. The negative pressure (expressed in units of MPa) in the above equation is dependent upon the acoustic impedance of water, Z, and the intensity of the LFUS, I, (expressed in W/cm 2 ). Negative pressure is denoted by equation (3).
Since the acoustic impedance of soft body tissues is comparable to that of water; hence, when determining P neg , the acoustic impedance of water, i.

In-vitro cytotoxicity analysis
To assess the therapeutic e ciency of DOX encapsulated HER-liposomes, the MTT assay was performed using both HER-positive (HER+) and HER-negative (HER-) breast cancer cells. SKBr3 (HER2+ cells) and MDA MB-231 (HER2-cells) cells were treated with free DOX, DOX-loaded control liposomes and DOXloaded HER-liposomes at a nal DOX concentration of 8 μM for 5 h, followed by continuous LFUS in a 35-kHz sonicating bath. The MTT results revealed the enhanced e cacy of the liposomal formulations upon cell exposure to US. Free DOX at a cytotoxic concentration (8 μM) was used as a positive control. The HER2+ cells viability % treated with HER-DOX liposomes was found to be 39 ± 0.86 %, and further dropped to 27 ± 0.26 % after US treatment (refer Figure 4A). In the HER2-cell line, the performance of targeted liposomes was similar to the control ones; however, more cell death was induced in the cells exposed to US (refer Figure 4B).

Cellular uptake study by ow cytometry analysis
Flow cytometry analysis was performed to compare the DOX uptake by two breast cancer cells SKBr3 (HER2+) and MDA-MB-231 (HER2-). Since HER2 receptors are overexpressed on the surface of SKBr3 cell lines, the intake of DOX when treated with DOX-loaded HER-liposomes was 600% higher compared to control liposomes. This was further improved by the application of US, as a 1000% improvement was observed compared to the control liposomes, refer Figure 5.

Conclusion
Smart drug delivery systems (SDDS), particularly liposomes, are an effective strategy in cancer treatment; by enhancing the pharmacokinetics and bio-distribution of chemotherapeutic agents, these nanocarriers allow the drug to interact with cancer cells, thus reducing unwanted side effects. The selectivity of liposomes can be further enhanced by conjugating targeting moieties to their surface. Once at the tumor site, a stimulus, such as ultrasound, can be used to trigger the release of the drug. This research reports on the synthesis of immunoliposomes, encapsulating the model drug/dye calcein and chemotherapeutic drug doxorubicin to target a HER-positive breast cancer cell line and studies the effects of low-frequency ultrasound, applied at different power densities, on payload release. The monoclonal antibody, Trastuzumab, was selected as a targeting moiety due to its high a nity toward HER2 breast cancer. Control liposomes and immunoliposomes (both Calcein and DOX-loaded) were successfully synthesized using the thin-lm hydration method. The phospholipid content and moiety attachment were con rmed using the Stewart and BCA assays, respectively. Both types of liposomes were tested for their release under LFUS at 20 kHz. The liposomal formulations were nano in size and were small enough to extravasate at the tumor site via the EPR effect.
The quanti cation of phospholipids and antibodies yielded a count of nine Trastuzumab molecules per liposome, which was considered enough to induce the a nity effect needed to target the HER2 overexpressed on breast cancer cells. Moreover, short exposure to LFUS effectively released the model drug calcein and the chemotherapeutic agent DOX from both types of liposomes, thereby improving the therapeutic e cacy signi cantly. Immunoliposomes were found to be more acoustically sensitive than control liposomes. Additionally, in vitro cytotoxicity of DOX-loaded HER-liposomes against HER positive and negative breast cancer cells showed excellent cytotoxic e cacy.
In summary, this research reports important ndings of utilizing ultrasound and immunoliposomes in the treatment of breast cancer. Testing the release at higher frequencies and intensities can provide additional insights into the best way to implement drug delivery in-vivo and in clinical settings.   Normalized release pro les of 3 batches of NH2 and HER-liposomes loaded with calcein loaded liposomes at 7.46 mW/cm2, 9.85 mW/cm2 and 17.31 mW/cm2.

Figure 3
Normalized release pro les of 3 batches of NH2 and HER-liposomes loaded with doxorubicin loaded liposomes at 7.46 mW/cm2, 9.85 mW/cm2 and 17.31 mW/cm2.    Flow cytometry analysis of HER-DOX liposomes (A) SKBr3 and, (B) MDAMB-231 cells were treated with Free Dox, NH2-DOX and HER-DOX (Dox concentration: 8 μM) and incubated for 3.5h followed by LFUS at a frequency of 35 KHz for 45 sec at a power density of 20mW/cm2. Untreated cells served as a negative control for background uorescence. An average of 1 × 104 cells was analyzed from each sample with the DOX uorescence intensity showed on a four-decade log scale. Bars represent the mean uorescence intensity (MFI) values ± SEM of DOX from three independent experiments between treatment groups analyzed for signi cance (* p < 0.05) using a two-tailed Student's t-test. control for background uorescence. An average of 1 × 104 cells was analyzed from each sample with the DOX uorescence intensity showed on a four-decade log scale. Bars represent the mean uorescence intensity (MFI) values ± SEM of DOX from three independent experiments between treatment groups analyzed for signi cance (* p < 0.05) using a two-tailed Student's t-test.