Co-encapsulation of curcumin and doxorubicin in albumin nanoparticles blocks the adaptive treatment tolerance of cancer cells

The adaptive treatment tolerance (ATT) of cancer cells is the main encumbrance to cancer chemotherapy. A potential solution to this problem is to treat cancer cells with multiple drugs using nanoparticles (NPs). In this study, we tested the co-administration of curcumin (Cur) and doxorubicin (Dox) to MCF-7 resistant breast cancer cells to block the ATT and elicit efficient cell killing. Drugs were co-administered to cells both sequentially and simultaneously. Sequential drug co-administration was carried out by pre-treating the cells with albumin nanoparticles (ANPs) loaded with Cur (Cur@ANPs) followed by treatment with Dox-loaded ANPs (Dox@ANPs). Simultaneous drug co-administration was carried out by treating the cells with ANPs loaded with both the drugs (Cur/Dox@ANPs). We found that the simultaneous drug co-administration led to a greater intra-cellular accumulation of Dox and cell killing with respect to the sequential drug co-administration. However, the simultaneous drug co-administration led to a lower intra-cellular accumulation of Cur with respect to the sequential drug co-administration. We showed that this result was due to the aggregation and entrapment of Cur in the lysosomes as soon as it was released from Cur@ANPs, a phenomenon called lysosomotropism. In contrast, the simultaneous release of Dox and Cur from Cur/Dox@ANPs into the lysosomes led to lysosomal pH elevation, which, in turn, avoided Cur aggregation, led to lysosome swelling and drug release in the cytosol, and finally provoked efficient cell killing. Our study shed the light on the molecular processes driving the therapeutic effects of anti-cancer drugs co-administered to cancer cells in different manners.


INTRODUCTION
One of the main challenges in the field of cancer therapy is the poor sensitivity to anti-cancer drugs developed by cancer cells after receiving the therapeutic treatment for a prolonged time, a phenomenon named adaptive treatment tolerance (ATT) (Zhou et al. 2015). ATT could be caused by several processes, including the difficulty of drugs to pass through the plasma membrane and/or escape from the lysosomal compartments of cancer cells or the transport of drugs out of cancer cells due to the activation of multidrug resistance (MDR)-associated proteins (Housman et al. 2014;Logan et al. 2014;Tarragó -Trani and Storrie 2007;Zhitomirsky and Assaraf 2016). One potential approach used to address the ATT is the co-administration of anti-cancer drugs to cancer cells (Wang et al. 2013a). Drugs can be co-administered to cells in two different manners, i.e., sequentially and simultaneously. In the sequential manner of co-administration, cells are pre-treated with a chemo-sensitizer, followed by the administration of a chemo-drug (Neville-Webbe et al. 2005). Pre-treatment of cancer cells with chemosensitizers significantly increases the efficacy of chemodrugs. Pre-treatment with Erlotinib inhibited the epidermal growth factor receptor (EGFR) and increased the therapeutic efficacy of doxorubicin (Dox) in a subset of triple-negative breast cancers (Lee et al. 2012). However, the sequential co-administration of drugs to cancer cells is not always the best option (Neek et al. 2018). Ducreux et al. have shown that the simultaneous co-administration of drugs led to more cytotoxic effects on advanced, resectable colorectal cancer compared to the sequential drug co-administration (Ducreux et al. 2011). One potential explanation for these differences is that anti-cancer drugs often display poor solubility in aqueous environments due to their hydrophobic nature. To address this problem, anti-cancer drugs have been encapsulated into nanoparticles (NPs). NPs can improve the therapeutic efficacy of drugs by increasing their halflife in the blood circulation, facilitating their cell penetration and intra-cellular distribution, and decreasing their side effects on normal tissues and cells (Bronze-Uhle et al. 2017;She et al. 2013;Yan et al. 2016).
Among the NPs used for cancer treatment, albuminbased NPs (ANPs) have been widely used as drug carriers due to their biocompatibility, lack of immunogenicity and toxicity, and great storage stability (Kratz 2008;Li et al. 2015;Lohcharoenkal et al. 2014). ANPs can be prepared in various range of sizes by means of straightforward, inexpensive, and easily controlled preparation procedures (Céspedes et al. 2014). Importantly, ANPs can be loaded with both hydrophilic and hydrophobic cargo (Cheema et al. 2009;Lin et al. 2011) and display a favorable pharmacokinetic profile due to their ability to accumulate in tumors by exploiting the enhanced permeability and retention (EPR) effect (Abbasi et al. 2012) and/or by triggering transcytosis in the endothelial cells by interacting with the albondin/glycoprotein 60 (Azizi et al. 2017). ANPs have been already approved for treatment of different cancers. Paclitaxel-loaded ANPs (Abraxane) have been approved for the treatment of breast cancer and clinical studies on docetaxel-loaded ANPs and are in progress for the treatment of several types of tumors.
Dox has been a widely used anti-cancer anthracycline but is suffering from serious side effects, including a dosedependent and cumulative life-threatening cardiotoxicity and the development of MDR (Carvalho et al. 2014). These shortcomings have been tackled by entrapping Dox into NPs and using chemo-sensitizers (Deepa et al. 2014). In this study, we investigated the co-treatment of MCF-7 resistant breast cancer cells with Dox and the chemosensitizer curcumin (Cur). Cur is a natural compound that inhibits the MDR protein family-member P-glycoprotein (P-gp) and blocks the transport of anti-cancer drugs out of cancer cells (Hsieh et al. 2014;Neerati et al. 2013;Peng et al. 2018;Sahu et al. 2016;Umsumarng et al. 2015;Vinod et al. 2013;Wang et al. 2016;Zhang et al. 2016). Both the drugs were encapsulated into ANPs and co-administered to cells both sequentially and simultaneously. Sequential drug co-administration was carried out by pre-treating the cells with ANPs loaded with Cur (Cur@ANPs) followed by treatment with ANPs loaded with Dox (Dox@ANPs). Simultaneous drug co-administration was carried out by treating the cells with ANPs loaded with both the drugs (Cur/Dox@ANPs). We found that the simultaneous drug co-administration led to a greater intra-cellular accumulation of Dox and cell killing with respect to the sequential drug co-administration. However, the simultaneous drug co-administration led to a lower intra-cellular accumulation of Cur with respect to the sequential drug co-administration. We showed that this result was due to the aggregation and entrapment of Cur in the lysosomes as soon as it was released from Cur@ANPs (lysosomotropism). On the contrary, the simultaneous release of Dox and Cur from Cur/Dox@ANPs into the lysosomes led to lysosomal pH elevation, which, in turn, avoided Cur aggregation, led to lysosome swelling and the release of the drugs in the cytosol, and finally provoked efficient cell killing. Our study showed the importance of tuning the manner of co-administration of chemo-drugs and chemosensitizers in order to maximize the therapeutic outcome. of them (Cur/Dox@ANPs) were prepared by mixing the drug(s) to a solution of BSA. Due to their hydrophobic nature, Cur and Dox will be forced to aggregate in an aqueous environment and interact with the hydrophobic regions of the BSA forming micelle-like structures, which will be stabilized by cross-linking the surface amino groups of BSA with glutaraldehyde (Bronze-Uhle et al. 2017;Gong et al. 2017;Ma et al. 2016). First, we characterized the physical and chemical properties of the NPs by transmission electron microscopy (TEM) and dynamic light scattering (DLS). TEM images showed that all the three types of NPs had a spheroidal shape with sizes ranging from 20 to 30 nm and were individually dispersed in a dehydrated state ( Fig. 1A and Table 1). DLS measurements showed that the NPs had an average hydrodynamic diameter in the 80-100 nm range and a negatively charged surface in a hydrated state ( Fig. 1B and Table 1). It is worth noting that the diameter of the NPs calculated by means of TEM images was smaller than that given by DLS data due to the shrinkage of the NPs during the drying of the TEM grids before imaging. All the NPs were negatively charged with the order Dox@ANPs (-7.53 ± 0.7 mV) [ Cur/ Dox@ANPs (-11.6 ± 1.2 mV) [ Cur@ANPs (-18.2 ± 1.3 mV) ( Table 1). The fact that Dox@ANPs had the highest surface charge among the fabricated NPs was most likely due to the amino groups in Dox.
Next, we tested the encapsulation of Cur and/or Dox in the ANPs. The encapsulation efficacy percentage (EE%) and loading yield percentage (LY%) were measured by means of HPLC. The EE% was greater than 85% for both Cur and Dox in all the NPs tested (Table 1). UV-Vis spectroscopy, differential scanning calorimetry (DSC), and Fourier-transform infrared (FTIR) spectroscopy were used to further test the encapsulation of Cur and/or Dox in the NPs. Free Cur and Dox and their NP counterparts exhibited an absorbance peak at 427 and 495 nm, respectively, whereas both a mixture of free drugs and Cur/ Dox@ANPs exhibited a broad absorbance peak *430 nm with a shoulder at *490 nm (Fig. 1C). These results validated the encapsulation of a single drug or both the drugs in the BSA matrix. Additionally, the DSC spectra of the NPs did not show the melting point peak(s) of Cur and/or Dox, which suggested that the drugs were in an amorphous state upon encapsulation in the NPs (Jenita et al. 2014) (Fig. 1D). Finally, we recorded the FTIR spectra of BSA, Cur, Dox, and the fabricated NPs (Fig. 1E). The FTIR spectra of BSA showed several characteristic peaks including 2925, 1651, and 1531 cm -1 , which are the representing peaks of amid I (mainly -NH stretching vibrations), amid II (mainly C=O stretching vibrations), and amid III (the coupling stretching vibrations of C-N and the coupling bending vibrations of -NH), respectively (Huang et al. 2010). The main FTIR peaks of Cur were found at 3506 cm -1 as a result of -OH stretching vibrations, 1271 cm -1 for aromatic C-O stretching vibrations, and 1023 cm -1 due to C-O-C stretching vibrations (Darandale and Vavia 2013; Mohanty and Sahoo 2010;Wang et al. 2013b;Yallapu et al. 2010). The main characteristics peaks of Dox were found at 3450 cm -1 as a result of -NH 2 stretching vibrations, 1410 cm -1 due to C-C vibrations, and 1073 cm -1 from C=O vibrations (Jayakumar et al. 2012;Unsoy et al. 2014). The presence of the relevant peaks in each type of NP confirmed that the Cur and/or Dox were encapsulated in the BSA matrix.
Finally, we tested the release of the drugs from the NPs in an acidic buffer to mimic the conditions encountered by the NPs upon accumulation in the lysosomes. The conformational changes in the protein structure induced by the acidic conditions led to the release of both Cur and Dox from the NPs with a firstorder kinetics ( Fig. 2A, B). The speed of release of both the drugs increased with the acidity of the environment; however, Cur showed a slower kinetics of release with respect to Dox. This phenomenon was explained by considering that Cur is more hydrophobic than Dox and interacts with the hydrophobic domains of BSA more strongly than Dox Dreis et al. 2007;Li et al. 2008). We also assessed the effect of the shell of glutaraldehyde molecules cross-linking the amino groups on the ANPs' surface on the kinetics of drug release. We reasoned that the kinetics of drug release would decrease with increasing the density of crosslinked surface amino groups. In order to validate this hypothesis, we prepared NPs with higher densities of crosslinked surface amino groups by increasing both the glutaraldehyde: BSA molar ratio and the reaction time. As hypothesized, NPs with higher densities of crosslinked surface amino groups displayed slower kinetics of drug release ( Fig. 2C-E).
Cytotoxicity of albumin nanoparticles loaded with Cur and/or Dox on MCF-7 cancer cells MCF-7 cancer cells were chosen for our experiments because overexpressing the P-gp pump (Lin et al. 2014;Lv et al. 2014). Cells were divided into five groups. Three groups of cells were treated with either Cur@ANPs or Dox@ANPs or Cur/Dox@ANPs loaded with 15, 25, 40, or 60 lmol/L of Cur and/or Dox for 24 h. The fourth and the fifth group of cells was pre-treated with Cur@ANPs loaded with 15, 25, 40, or 60 lmol/L of Cur for 6 or 12 h, respectively, and then the cells were exposed to Dox@ANPs loaded with a concentration of Dox equal to that of Cur in Cur@ANPs for an additional 24 h. The MTT assay showed that the cell viability decreased with the concentration of the drugs independently from the type of treatment. We also found that the cells treated with Cur/Dox@ANPs exhibited a lower viability than the cells treated with either Cur@ANPs or Dox@ANPs for all the drug concentrations tested. Additionally, we found that the cells exposed to Cur/Dox@ANPs showed lower viability than the cells pre-treated with Cur@ANPs for 6 and 12 h, for all the drug concentrations tested (Fig. 3A). We assessed if this phenomenon was due to a greater uptake of the drugs by cells treated with Cur/Dox@ANPs with respect to the cells pre-treated with Cur@ANPs by means of FACS. As expected, we found that the accumulation of Dox in MCF-7 cancer cells was the greatest   E Fourier-transform infrared spectra when they were treated with Cur/Dox@ANPs for all the drug concentrations tested (Fig. 3B). However, cells pretreated with Cur@ANPs took up more Cur than cells treated with Cur/Dox@ANPs (Fig. 3C), which was in contrast with the cell viability data. We hypothesized that the pre-treatment with Cur@ANPs did not lead to efficient cytotoxicity because Cur aggregated and remained entrapped in the lysosomes as soon as it was released from the NPs due to its weak-base nature (lysosomotropism) (Kuzu et al. 2017). In order to validate this hypothesis, we assessed the trafficking profile of Cur in cells treated with Cur/Dox@ANPs for 3 h and in cells pre-treated with Cur@ANPs for 1.5 h and then exposed to Dox@ANPs for an additional 1.5 h by means of confocal laser scanning microscopy (CLSM) (Burger et al. 2015). Cells pre-treated with Cur@ANPs showed a strong punctuate fluorescence arising from Cur co-localized with the lysosomes and a faint fluorescence arising from Dox (Fig. 4B), whereas cells exposed to Cur/Dox@ANPs showed diffused signals    arising from both the drugs, suggesting the rupture of the lysosomes and the release of the drugs in the cytosol (Fig. 4C). Taken together, these results showed that, although the pre-treatment with Cur@ANPs led to a greater accumulation of Cur in the cells, Cur aggregated and remained entrapped in the lysosomes as soon as it was released from the NPs even after cell exposure to Dox@ANPs. The measurement of the lysosomal pH showed that the cells pre-treated with Cur@ANPs had a lysosomal pH lower than that of the cells exposed to Cur/Dox@ANPs (Fig. 5C). Since Dox can increase the lysosomal pH , we reasoned that the ability of Dox to increase the pH of lysosomes and trigger the accumulation of drugs in the cytosol was dependent on the aggregation state of Cur after its release from the NPs. In order to validate this hypothesis, we performed the following experiment. We added either Cur or Dox or a mixture of both the drugs in water, gently spun it down after 0 and 6 h, and measured both the pH and the absorbance spectra of the supernatants. In a separate experiment, Cur was incubated in water for 3 h and either gently spun down and both the pH and the absorbance spectrum of the supernatant were measured, or free Dox was added to the mixture for an additional 3 h. We found that when Cur and Dox were co-incubated in water, the pH slightly raised from 6.5 to 6.8 at 0 h and decreased to 6.5 after 6 h of incubation without forming any visible aggregates (black curves in Fig. 5A, B). The addition of Cur to water led to a decrease in the pH from 6.5 to 5.7 at 0 h (red curve in Fig. 5A). After 3 h of incubation, the pH further decreased to 5.5 and visible aggregates formed and were spun down after gentle centrifugation leaving a clear supernatant (green curve in Fig. 5B). When Dox was added to the Cur solution after 3 h of incubation, the pH increased to 6.1, increasing the solubility of Cur (red curve in Fig. 5B). However, only approximately half of the initial amount of Cur remained in solution in a non-aggregated form at the end of the 6-h experiment as suggested by the intensity of the peak in the Vis-range of wavelengths in the absorbance spectrum of the supernatant collected after gentle centrifugation of the mixture with respect to the intensity of the peak in the spectrum of the supernatant collected after gentle centrifugation of the solution prepared by co-incubating Cur and Dox (red curve vs. black curve in Fig. 5B).
Finally, we measured the expression of P-gp in cells treated with Cur/Dox@ANPs for 24 h and in cells pretreated with Cur@ANPs for 6 h and then exposed to Dox@ANPs for 24 h. We found that the cells treated with Cur/Dox@ANPs exhibited a lower expression of P-gp with respect to cells pre-treated with Cur@ANPs, which suggested that Cur was able to inhibit the expression of P-gp only when it was not entrapped in the lysosomes in an aggregated form (Fig. 5D).

CONCLUSIONS
Taken together our results suggest that the synergism between Cur and Dox co-encapsulated in ANPs enabled to decrease the cell's ATT against Dox. Indeed, the simultaneous release of Cur and Dox in the lysosomes from Cur/Dox@ANPs increased the lysosomal pH and maintained Cur individually dispersed. Concomitantly, the Dox-triggered buffering of the lysosomal acidic pH led to the swelling of the vesicles and the release of the drugs in the cytosol. Once in the cytosol, Cur blocked the transport of Dox out of the cells by inhibiting the expression of P-gp. This ultimately translated into efficient cell killing. On the contrary, due to its weak-base nature, Cur formed aggregates in the lysosomes as soon as it was released from Cur@ANPs. Aggregated Cur decreased the buffering ability of the Dox released from Dox@ANPs added to the cell culture after the pretreatment with Cur@ANPs. This translated into less efficient release of the drugs from the lysosomes and poor inhibition of P-gp expression and, in turn, led to the transport of Dox out of the cancer cells though the P-gp pumps (Scheme 1). In conclusion, this study showed the importance of tuning the manner of co-administration of chemo-drugs and chemo-sensitizers in order to maximize the anti-cancer effects.

LTR
Cur Merged LTR/BF BF Dox A B C Fig. 4 Confocal laser scanning microscopy images of MCF-7 R breast cancer cells untreated (A), pre-treated with Cur@ANPs for 1.5 h and then exposed to Dox@ANPs for an additional 1.5 h (B), and treated with Cur/Dox@ANPs for 3 h (C). Scale bars are 2 lm in all the images. LTR: lysotracker red, BF: bright field Blocking the adaptive treatment tolerance of cancer cells with multiple drugs using nanoparticles

RESEARCH ARTICLE Preparation of Cur@ANPs, Dox@ANPs, and Cur/ Dox@ANPs
First, Dox-HCl was dissolved in DMSO at a final concentration of 15 mg/mL, and then 80 lL of hydrazine (0.1%) was added to 100 lL of Dox solution. The mixture was shaken for 2 min. Next, Cur was dissolved in DMSO at a final concentration of 8.3 mg/mL and the mixture was shaken for 30 s. The Dox solution was added to an aliquot of Cur solution and the mixture was shaken for 30 s. A BSA solution was prepared in double distilled H 2 O (ddH 2 O) and mixed with the Cur, Dox, or Dox/Cur solution to fabricate Cur@ANPs, Dox@ANPs, and Cur/Dox@ANPs, respectively (Table 2). Glutaraldehyde was added to all the mixtures. After 3 h of continuous stirring at room temperature, the solutions were washed with ddH 2 O three times through centrifugal filter devices (100 kDa cut-off, MilliporeSigma, Burlington, MA) and the collected NPs were lyophilized and stored in the dark at 4°C.

Characterization of Cur@ANPs, Dox@ANPs, and Cur/Dox@ANPs
The NPs' size distribution and zeta potential were calculated by a Zetasizer Nano ZS (Malvern Instruments, Worcestershire, UK). The samples were stained by uranyl acetate (1 wt%) and the NPs' morphology was characterized by transmission electron microscopy (TEM, model JEM-200CX, JEOL, Tokyo, Japan). The EE% and LY% of Cur and Dox in Cur@ANPs, Dox@ANPs and Cur/Dox@ANPs were measured by high-pressure liquid chromatography (HPLC, model e2796, Waters, Milford, MA) method using a reverse-phase C-18 Agilent Eclipse Plus (Thermo Fisher Scientific). The mobile phase consisted in 3 mmol/L monopotassium phosphate-acetic acid in MeOH Zhao et al. 2015). The lyophilized NPs were dissolved in MeOH, sonicated for 10 min, and filtered through centrifugal filter devices (30 kDa cut-off, MilliporeSigma). The EE% and LY% were calculated by the following equations: EE% ¼ The amount of loaded DOX or CUR Total amount of DOX or CUR used for ANPs Â 100; The amount of loaded Dox or Cur Total amount of ANPs Â 100: DSC was performed by a diamond differential scanning calorimeter (model Diamond DSC, PerkinElmer Inc., Waltham, MA, USA). Samples were heated from 30 to 230°C at a scan speed of 5°C/min (Jenita et al. 2014). The NPs' chemical structure was assessed by Fourier-transform infrared spectroscopy (FTIR, model IFS 66 s/v, Bruker, Ettlingen, Germany) using KBr discs. The absorption spectra of free drugs and NPs were recorded by a visible spectrophotometer (model Lambda 950, PerkinElmer), whereas their fluorescent spectra were recorded by a fluorescent spectrophotometer (model F-4600, Hitachi, Tokyo, Japan) at Sequential co-administration Simultaneous co-administration Scheme 1 This study investigated the co-administration of the chemo-sensitizer Cur and the chemo-drug Dox to the MCF-7 breast cancer cell line to block the ATT and elicit efficient cell killing. The drugs were entrapped into ANPs and co-administered to cells both sequentially and simultaneously. Sequential drug co-administration was carried out by pre-treating the cells with Cur@ANPs followed by treatment with Dox@ANPs. Simultaneous drug co-administration was carried out by treating the cells with Cur/Dox@ANPs. The simultaneous drug co-administration elicited strong cytotoxicity, whereas the sequential drug co-administration did not. The data described in this study unraveled the molecular mechanisms driving the observed different therapeutic outcomes as summarized in ''Conclusions'' The MTT assay was carried out by seeding the cells at a concentration of 5 9 10 3 cells/well in 96-well plates and growing the cells in RPMI 1640 (FBS 10%) media at 37°C, 5% CO 2 for 24 h. The cells were treated with the NPs both sequentially and simultaneously as described before. The media was replaced with 100 lL of 0.5 mg/mL MTT solution and after 4 h the MTT solution was replaced with 200 lL of DMSO. The absorbance value was recorded at 570 nm with a reference wavelength of 630 nm using an Infinite M200 microplate reader (Tecan Group Ltd., Mä nnedorf, Switzerland).
FACS measurements were carried out by seeding the cells in 12-well plates (three wells for each treatment) at a concentration of 3 9 10 5 cells/well and growing the cells in RPMI 1640 (FBS 10%) media at 37°C, 5% CO 2 for 24 h. The cells were treated with the NPs both sequentially and simultaneously as described before. The cells were trypsinized and washed three times. The fluorescence signals arising from the samples were measured using an Attune Ò acoustic focusing cytometer (Applied Biosystems, Life Technologies, Carlsbad, CA).
The co-localization of drugs within the lysosomes was carried out by seeding the cells in 35-mm glass bottom dishes at a density of 2 9 10 4 and growing the cells in RPMI 1640 (FBS 10%) media at 37°C, 5% CO 2 for 24 h. The cells were treated with the NPs both sequentially and simultaneously as described before. The total drug concentration in the media was 40 lmol/L. Cells were imaged using a confocal laser scanning microscope (LSM710, Carl Zeiss Microscopy) with excitation at 405 nm for Cur and 488 nm for Dox. Lysotracker red DND-99 was used for lysosome coloring.
The measurements of lysosomal pH were carried out by seeding the cells in 12-well plates at a concentration of 2 9 10 5 cells/well and administering the NPs to the cells both sequentially and simultaneously as described before. The total drug concentration in the media was 40 lmol/L. The culture media was poured out of the wells and replaced with culture media containing LysoSensor Green DND-189 at 37°C for 30 min. The cells were trypsinized and washed three times. The fluorescence signals arising from the samples were measured with ex/em wavelengths of 443 nm/505 nm using an Attune Ò acoustic focusing cytometer (Applied Biosystems).
The expression of P-gp was assessed by Western blotting (WB). Briefly, total proteins were extracted from cells, quantified by Bradford method, and run through a 10% polyacrylamide gel (SDS-PAGE). Proteins were transferred to a nitrocellulose membrane for 2 h at 400 mA. The membrane was incubated in 5% w/v non-fat dry milk in TBST for 1 h, at room temperature. The membrane was washed with TBST three times and incubated with primary antibody for P-gp and GAPDH (diluted 1:1000) overnight. The primary antibody was removed, and the membrane was washed with TBST three times and incubated with peroxidaseconjugated goat anti-rabbit IgG (diluted 1:1000) for 2 h. Finally, the secondary antibody was removed, and the membrane was washed with TBST three times, stained with Immobilon Western reagent (Merck Millipore, Burlington, MA), and imaged by a Gel Imaging System (Bio-Rad, Hercules, CA).