Derivatization of inhibitor of apoptosis protein (IAP) ligands yields improved inducers of estrogen receptor α degradation

Aberrant expression of proteins often underlies many diseases, including cancer. A recently developed approach in drug development is small molecule-mediated, selective degradation of dysregulated proteins. We have devised a protein-knockdown system that utilizes chimeric molecules termed specific and nongenetic IAP-dependent protein erasers (SNIPERs) to induce ubiquitylation and proteasomal degradation of various target proteins. SNIPER(ER)-87 consists of an inhibitor of apoptosis protein (IAP) ligand LCL161 derivative that is conjugated to the estrogen receptor α (ERα) ligand 4-hydroxytamoxifen by a PEG linker, and we have previously reported that this SNIPER efficiently degrades the ERα protein. Here, we report that derivatization of the IAP ligand module yields SNIPER(ER)s with superior protein-knockdown activity. These improved SNIPER(ER)s exhibited higher binding affinities to IAPs and induced more potent degradation of ERα than does SNIPER(ER)-87. Further, they induced simultaneous degradation of cellular inhibitor of apoptosis protein 1 (cIAP1) and delayed degradation of X-linked IAP (XIAP). Notably, these reengineered SNIPER(ER)s efficiently induced apoptosis in MCF-7 human breast cancer cells that require IAPs for continued cellular survival. We found that one of these molecules, SNIPER(ER)-110, inhibits the growth of MCF-7 tumor xenografts in mice more potently than the previously characterized SNIPER(ER)-87. Mechanistic analysis revealed that our novel SNIPER(ER)s preferentially recruit XIAP, rather than cIAP1, to degrade ERα. Our results suggest that derivatized IAP ligands could facilitate further development of SNIPERs with potent protein-knockdown and cytocidal activities against cancer cells requiring IAPs for survival.


degrade ER␣. Our results suggest that derivatized IAP ligands could facilitate further development of SNIPERs with potent protein-knockdown and cytocidal activities against cancer cells requiring IAPs for survival.
Aberrant or increased expression of pathogenic proteins often results in the initiation of many diseases, including a variety of cancers. However, a limited number of pathogenic proteins are currently targeted by pharmacological inhibitors, primarily because many proteins do not have suitable binding pockets where a pharmacophore may interact to regulate protein activity (1,2). An alternative approach has been to down-regulate expression of the pathogenic proteins; this is usually achieved in vitro through use of genetic methods involving antisense oligonucleotides, dsRNAs, and CRISPR-Cas9 technology. However, clinical application of these technologies remains challenging because delivery of oligonucleotides to the target tissues is not easily accomplished (3,4).
As a novel strategy to down-regulate pathogenic proteins in a nongenetic manner, we and others have devised a proteinknockdown system that uses small molecules with sufficient membrane permeability to induce selective degradation of target proteins. These small-molecule compounds, designated as proteolysis-targeting chimeras (PROTACs) 7 and specific and non-genetic IAP-dependent protein erasers (SNIPERs), are chimeric molecules that contain two different ligands connected by a linker; one ligand is specific for an E3 ubiquitin ligase, and the other is specific for a target protein (5)(6)(7). The PROTACs and SNIPERs are designed to cross-link the E3 ubiquitin ligase and the target protein to induce polyubiquitylation and proteasomal degradation of the target protein within cells. To recruit the von Hippel-Lindau (VHL) E3 ligase complex and the cereblon (CRBN) E3 ligase complex, a VHL inhibitor (based on the HIF-1␣ peptide) and a phthalimide moiety have been respectively integrated into PROTAC constructs (8 -11). In a similar manner, an IAP antagonist has been incorporated into SNIPERs to recruit either cellular inhibitor of apoptosis protein 1 (cIAP1) or X-linked inhibitor of apoptosis protein (XIAP) E3 ligase (12)(13)(14)(15)(16)(17)(18). To date, a range of PROTAC and SNIPER compounds have been developed, allowing degradation of a variety of proteins, such as estrogen receptor ␣ (ER␣), oncogenic kinase BCR-ABL, and epigenetic regulator bromodomaincontaining protein 4 (BRD4) (19 -30). Some PROTACs and SNIPERs also have demonstrated the ability to degrade target proteins in vivo, suggesting that this technology is feasible for use in novel drug development. These include VHLbased and CRBN-based PROTACs directed against BRD4, which are capable of inducing BRD4 protein degradation in mouse xenograft models (10,31). By incorporating an LCL161 derivative, a high-affinity IAP ligand, we developed SNIPER(ER)-87, which induces in vivo degradation of ER␣ and growth inhibition of an ER␣-positive human breast tumor in a xenograft model (13).
IAPs are a family of antiapoptotic proteins containing one or three baculoviral IAP repeat (BIR) domains (32)(33)(34). Some family members, such as cIAP1, cIAP2, and XIAP, directly interact with and regulate caspases via the BIR domain, thus inhibiting apoptosis (35)(36)(37)(38). These IAPs are attractive targets for tumor therapy because of their frequent overexpression in multiple human malignancies and their implications in tumor progression, treatment failure, and poor prognosis (39 -45). Based on the IAP-binding tetrapeptides of second mitochondria-derived activator of caspases/direct IAP-binding protein with low pI (SMAC/DIABLO), many potent and cell-permeable peptidomimetic IAP antagonists (also known as SMAC mimetics) have been developed; some of these are under evaluation in clinical phase studies as antitumor drugs (32,46,47). These IAP antagonists interact with BIR domains in IAP proteins to directly inhibit XIAP or to induce autoubiquitylation and proteasomal degradation of cIAP1 and cIAP2 (48 -51). Because SNIPERs utilize IAP antagonists as IAP-ligand modules, SNIPERs are able to down-regulate IAPs beyond the initial target proteins (12)(13)(14)(15)(16)(17)(18); this is likely to be advantageous when attempting to kill cancer cells that require IAPs for survival.
In this paper, we demonstrate that, by derivatizing the IAPligand module, we have developed novel SNIPER(ER)s whose protein knockdown and antitumor activities are more potent than those of SNIPER(ER)-87. These SNIPER(ER)s have higher affinity for IAPs and exhibit more consistent abilities to degrade ER␣ and IAPs. In addition, we discuss the significance of IAP down-regulation in pharmacological efforts to induce cancer cells to undergo apoptosis.

Degradation of cIAP1 and XIAP by SNIPER(ER)s
As observed in studies of many IAP antagonists, SNIPERs in our study rapidly induced proteasomal degradation of cIAP1, along with degradation of their target proteins (12)(13)(14)(15)(16)(17)(18). SNIPER(ER)-105, -110, -113, -119, and -126 each reduced the cIAP1 level more potently than SNIPER(ER)-87 (Figs. 2a and 3b), consistent with their increased binding affinities for cIAP1 (Table 1). Further, these SNIPER(ER)s reduced XIAP expression in MCF-7 cells after 48 h in a more potent fashion than SNIPER(ER)-87 (Fig. 2a). The reduction of XIAP by SNIPER(ER)s was prominent in T47D cells after 48 h but was weak after 4 h of exposure (Fig. 3b). The difference in the degradation pattern between cIAP1 and XIAP suggests that these IAPs are degraded through a different mechanism, as discussed below.

XIAP is required for the ER␣ degradation by SNIPER(ER)s
We previously reported that XIAP is preferentially recruited to ER␣ upon treatment with SNIPER(ER)-87 (13). To examine which IAP is recruited by the novel SNIPER(ER)s (SNIPER(ER)-105, -110, and -126) to induce ER␣ degradation, MCF-7 and T47D cells were treated with or without the SNIPER(ER)s in the presence of MG132, a proteasome inhibi-

Development of potent SNIPER derivatives against ER␣
tor; cell lysates were analyzed by co-immunoprecipitation. Fig.  4a shows that, when cells were treated with SNIPER(ER)s, the anti-ER␣ precipitates contained both XIAP and cIAP1. We did not examine cIAP2, because it is not expressed in MCF-7 and T47D cells under normal cell culture and cIAP1-down-regulated conditions (Fig. S1, a-c). Upon comparison of IAP protein levels in total cellular lysates, we found that SNIPER(ER)s more effectively recruited XIAP to ER␣, compared with cIAP1. Consistent with this, depletion of XIAP by siRNA substantially suppressed the ER␣ degradation induced in MCF-7 and T47D cells by our SNIPER(ER) molecules, whereas depletion of cIAP1 marginally suppressed the degradation (Fig. 4b). These results indicate that SNIPER(ER)-105, -110, and -126 preferentially recruit XIAP to degrade ER␣ in target cells, similar to the mechanism initiated by SNIPER(ER)-87.

SNIPER(ER)s induce apoptosis in tumor cells requiring IAPs for survival
Because ER␣ plays an important role in the proliferation of most primary breast tumors (56,57)  ). Consistent with this, a 100 nM concentration of these SNIPER(ER)s activated caspase-7 and caspase-6 (detected by reduction of their precursor forms), which resulted in the cleavage of two caspase substrates, PARP and lamin B1, in MCF-7 cells (Fig. 5e). These results indicate that SNIPER(ER)-105, -110, and -126 each induced apoptosis in MCF-7 cells more effectively than SNIPER(ER)-87; apoptosis was not observed in T47D cells.
To understand why these new SNIPER(ER)s effectively induce apoptosis in MCF-7 cells, but not in T47D cells, and because SNIPER(ER)-105, -110, and -126 each induced the degradation of cIAP1 and XIAP more potently than SNIPER (ER)-87 (Figs. 2a and 3b), we investigated the effects of IAP depletion in MCF-7 and T47D cells. In MCF-7 cells, individual silencing of cIAP1 and XIAP by siRNA slightly reduced the number of cells; combined silencing of cIAP1 and XIAP reduced the number of cells by ϳ50%, compared with untreated cells (Fig. 6a), and was accompanied by caspase activation, as demonstrated by cleavage of lamin B1 and PARP (Fig.  6b). In contrast, similar silencing of IAPs in T47D cells neither reduced cell number nor activated caspases. When ER␣ was

Development of potent SNIPER derivatives against ER␣
silenced in MCF-7 and T47D cells, the numbers of cells in both cell lines were effectively reduced without a requirement for caspase activation. However, triple silencing of ER␣, XIAP, and cIAP1 effi-ciently reduced MCF-7 cell number, compared with T47D cells; further, caspase activation was observed only in MCF-7 cells. Additionally, an IAP antagonist, LCL161, sensitized tumor necro-

Development of potent SNIPER derivatives against ER␣
sis factor ␣ (TNF␣)-dependent apoptosis in MCF-7 cells, but not T47D cells (Fig. S4). These results suggest that MCF-7 cells, but not T47D cells, require IAPs for survival; thus, IAPs could be involved in the selective induction of apoptosis in MCF-7 cells that have been exposed to SNIPER(ER)-105, -110, and -126.

Development of potent SNIPER derivatives against ER␣
and carefully selected SNIPER(ER)-110 and -126 as test compounds for in vivo comparison with SNIPER(ER)-87. When female BALB/c mice were intraperitoneally injected with SNIPER(ER)-87, -110, and -126 (30 mg/kg body weight), ER␣ protein levels in ovarian tissue were effectively reduced by all SNIPERs (Fig. 7a). To evaluate SNIPER(ER)-knockdown activities in a tumor model, we next applied MCF-7 breast tumor xenografts to nude mice. The ER␣ protein levels in orthotopic

Development of potent SNIPER derivatives against ER␣
tumors were reduced by 40 -50% in SNIPER(ER)-treated mice, compared with those in vehicle-treated mice (Fig. 7, b and c); SNIPER(ER)-110 significantly showed a significant increase in protein-knockdown activity at 48 h, compared with SNIPER(ER)-87. Treatment with SNIPER(ER)-87, -110, and -126 each reduced cIAP1 and XIAP protein levels in tumor xenografts (Fig. 7, b and c), consistent with the results of our in vitro analyses (Fig. 2a). Notably, SNIPER(ER)-110 again exhibited a significant increase in the ability to reduce cIAP1 and XIAP at 24 and 48 h, compared with SNIPER(ER)-87.
To demonstrate the therapeutic significance of these findings, we also evaluated the in vivo antitumor activity of SNIPER(ER)-110, compared with SNIPER(ER)-87, in an MCF-7 tumor xenograft model. When tumor-bearing mice received daily intraperitoneal injections of SNIPER(ER)-87 or SNIPER(ER)-110 (30 mg/kg body weight) for 16 days, SNIPER(ER)-110 attenuated MCF-7 tumor progression more effectively than SNIPER(ER)-87, as shown in measurements of tumor volume and tumor weight (Fig. 8, a-c). Notably, no obvious toxicities, including body weight changes, were observed during treatment (Fig. 8d). Thus, compared with SNIPER(ER)-87, SNIPER(ER)-110 showed superior antitumor activity against ER␣-positive breast tumors that require IAPs for cellular survival.

Discussion
Chemical protein knockdown technologies, utilizing chimeric molecules, such as PROTACs and SNIPERs, have been recently developed to induce degradation of target proteins in cells (5-7). Our SNIPER compounds degrade cIAP1 and XIAP along with their initial target proteins, which could be advantageous in attempts to kill cancer cells because of the frequent overexpression of IAPs to suppress apoptosis in cancer cells. In this study, we developed novel SNIPER(ER)s by incorporating various IAP ligands with higher binding affinities for IAP proteins. The resulting SNIPER(ER)s (SNIPER(ER)-105, -110, and -126) each induce degradation of ER␣, cIAP1, and XIAP more potently than SNIPER(ER)-87, resulting in the induction of apoptosis in MCF-7 breast tumor cells. In an in vivo tumor xenograft model, the novel SNIPER(ER)s exhibited activity to reduce ER␣ and IAPs that was respectively comparable with, and better than, that of SNIPER(ER)-87; further, the novel SNIPER(ER)-110 showed superior antitumor activity against MCF-7 tumor xenograft, compared with SNIPER(ER)-87. These results suggest that incorporating IAP ligands with higher binding affinities may be a promising strategy to develop SNIPERs with better protein knockdown and antitumor activities.
The novel SNIPER(ER)s also exhibited better proteinknockdown activity in other ER␣-positive breast tumor cells, compared with SNIPER(ER)-87 (Fig. 3). However, the novel SNIPER(ER)s did not exhibit improved inhibition of growth or induction of apoptosis in T47D cells (Fig. 5), which contrasts with their activities in MCF-7 cells. We speculate that the differences in sensitivity between the two cell lines derive from differing need for IAPs, because depletion of IAPs induces apoptosis in MCF-7 cells, but not in T47D cells (Fig. 6). The role of IAPs in cancer cell survival has been suggested in many papers (32,(42)(43)(44)(45)58), and IAP antagonists have shown antitumor activity against many tumor cells, including breast tumor (46,47,59). Therefore, we hypothesize that the ability of the SNIPERs to degrade IAPs could be advantageous in treatment of cancer cells, especially those dependent on the activity of IAPs. Recently, it has been reported that the different molecular signature other than IAP frequency also specifies a sensitivity to IAP antagonists in cancer cells; the mechanism depends on autocrine release of TNF␣ (60 -64), constitutive ubiquitylation of receptor-interacting protein kinase 1 (RIP1) (61), and expression of cellular FLICE-inhibitory protein (c-FLIP) (60) and leucine-rich repeats and immunoglobulin-like domains protein 1 (LRIG1) (65). Thus, these signatures could be used to predict sensitivity to SNIPERs as well.
Although SNIPER(ER)s induce degradation of ER␣, cIAP1, and XIAP, the degradation mechanism appears different for each protein. Degradation of ER␣ appears to require formation of a ternary complex that is composed of ER␣, SNIPER, and XIAP; we suspect this because a treatment with the individual ligand or with an structural SNIPER(ER) analog that cannot bind to IAPs does not induce degradation of ER␣ (13) (Fig. S5). The pharmacological hook effect that is observed during SNIPER(ER)-induced ER␣ degradation further supports the hypothesis that ternary complex formation is required (13). In contrast, the degradation of cIAP1 is triggered by binding of the IAP antagonist module to the BIR3 domain of cIAP1, thus inducing autoubiquitylation. In this case, the ternary complex is not required, and no hook effect is observed (13). Notably, the degradation mechanism of XIAP remains ambiguous. Degradation of XIAP was induced by SNIPER(ER)s but not by IAP antagonists nor by the mixture of IAP antagonist and 4-hydroxytamoxifen (Fig. S5), suggesting that the ternary complex is required for XIAP degradation. Consistent with this hypothesis, XIAP degradation by SNIPER(ER)-87 was not observed in the ER␣-negative MDA-MB-231 breast cancer cells (Fig. S5). Thus, XIAP degradation seems to require the same ternary complex formation that is required for ER␣ degradation. However, XIAP degradation requires more time and is inefficient, compared with degradation of the ER␣ target protein (Figs. 2a  and 3). It is likely that part of the population of XIAP in the ternary complex is subjected to ubiquitylation and proteasomal degradation. Because XIAP degradation was more prominently observed in T47D cells than in MCF-7 and ZR-75-1 cells, we

Development of potent SNIPER derivatives against ER␣
measured the expression of ER␣, cIAP1, and XIAP in these cells (Fig. S6). We found that the ratio of XIAP to ER␣ is one-third and one-thirteenth less in T47D cells than in MCF-7 cells and in XR-75-1 cells, respectively, suggesting that the lower expression of XIAP may increase the significance of its degradation in T47D cells. Thus, although the degradation mechanisms differ, SNIPER(ER)s induce degradation of ER␣, cIAP1, and XIAP, culminating in an anti-tumor effect that is directed against breast cancer cells.
It has been reported that ER␣ translocates from the cytosol to the nucleus upon binding with tamoxifen (66,67). This is also observed after treatment with SNIPER(ER)s, because the relative amount of ER␣ increased in the nucleus, compared with the cytosol, upon treatment with SNIPER(ER)s; this was more clearly observed in the presence of MG132 (Fig. S7, a and b). Interestingly, the amount of XIAP increased in the nucleus upon treatment with SNIPER(ER)s; this was also more clearly observed in the presence of MG132. These results suggest that when a SNIPER(ER) penetrates a cell, it induces formation of the ternary complex composed of ER␣, SNIPER(ER), and XIAP in the cytosol; ER␣ then translocates to the nucleus together with the bound XIAP. Thus, in the presence of SNIPER(ER), ER␣ interacts with XIAP sufficiently to tether XIAP in the nucleus, consistent with the preferential recruitment of XIAP, but not cIAP1, to ER␣.
In summary, we have developed potent SNIPER(ER)s by incorporation of high-affinity IAP ligands. These SNIPER(ER)s induce degradation of ER␣ and IAPs and exhibit antitumor activity against breast tumors that require IAPs for survival. These IAP ligand modules could be utilized to develop novel SNIPERs with activity against various oncogenic proteins and enhanced antitumor capabilities.

Chemistry
Chemical synthesis and physicochemical data for SNIPER compounds are provided in the supporting information.

Cell culture
Human breast carcinoma MCF-7, T47D, and ZR-75-1 cells were maintained in RPMI 1640 medium containing 10% fetal bovine serum (FBS) and 100 g/ml kanamycin. Human breast carcinoma MDA-MB-231 cells were maintained in Dulbecco's modified Eagle's medium containing 10% FBS and 100 g/ml kanamycin. All of the cell lines were purchased from ATCC (Manassas, VA), and their cell authentications were confirmed by morphology, karyotyping, and PCR-based approaches in ATCC. Cells were treated with various concentrations of SNIPER compounds for the indicated times.

Western blotting
Cells were lysed with SDS lysis buffer (0.1 M Tris-HCl, pH 8.0, 10% glycerol, 1% SDS) and immediately boiled for 10 min to obtain clear lysates. Nuclear and cytoplasmic extracts were prepared with NE-PER nuclear and cytoplasmic extraction reagents (Thermo Fisher Scientific). Protein concentration was measured by the BCA method (Pierce); lysates containing equal amounts of proteins were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore, Darmstadt, Germany) for Western blot analysis using the

Measurement of binding affinities of ER␣ and IAPs
The bindings between test compounds and cIAP1, cIAP2, or XIAP were determined by a time-resolved FRET assay using His-IAP proteins (XIAP, cIAP1, or cIAP2), FITC-Smac, Tb-SA, and biotinylated anti-His antibody, as described previously (14).
The binding between test compounds and the human ER␣ protein was determined using the PolarScreen TM Estrogen Receptor-␣ Competitor Assay Green (Thermo Fisher Scientific, A15882) containing recombinant ER␣ full-length protein, Fluormone ES2 Green2, and ES2 screening buffer. Purified ER␣ and Fluormone ES2 were diluted with assay buffer to final concentrations of 25 and 4.5 nM, respectively, and 4 l of each dilution was added to each well of a 384-well black low-volume assay plate (Greiner, 784076). Then 2 l of the ES2 screening buffer, containing test compounds or DMSO, was added to the well. The plate was subjected to centrifugation and incubated at room temperature for 1 h; then the intensity of the fluorescence polarization signal was measured by a plate reader (Envision, PerkinElmer Life Sciences). Wells containing ER␣ and Fluormone ES2 were used as a positive control, and wells containing only Fluormone ES2 were used as a negative control. IC 50 values and 95% confidence interval were calculated by XLfit (ID Business Solutions, fit model 204) from the data expressed as percentage of control inhibition.

Cell viability assay
Cell viability was evaluated by crystal violet staining. Cells were treated with graded concentrations of the SNIPER compounds for 72 h and then stained with 0.1% crystal violet (Wako, Osaka, Japan) in 1% ethanol for 15 min at room temperature. Cells were rinsed thoroughly with distilled water and then lysed in 1% SDS. The absorbance of the cell lysate at 600 nm was measured using an EnVision multilabel plate reader (PerkinElmer Life Sciences).

In vivo protein knockdown
All procedures of the animal experiments were reviewed and approved by the Institutional Animal Experiment Review Committee of the National Institute of Health Sciences (NIHS), and those experiments were conducted in accordance with the guidelines approved by the Care and Use of Animals published by the Institutional Animal Ethical Committee. Mice were housed in pathogen-free animal facilities at the NIHS with 12-h light/dark cycles and were fed rodent chow and water ad libitum. For Fig. 7a, female 6-week-old BALB/c mice (Clea Japan, Tokyo, Japan) were randomized and divided into eight treatment groups (n ϭ 3) and then treated with vehicle or 30 mg/kg SNIPER(ER)s for 24 h or 48 h. For Fig. 7 (b and c), each suspension of 1 ϫ 10 7 MCF-7 cells was mixed with an equal volume of Matrigel (Corning Inc.) and injected (100 l total) into the left and right mammary fat pads of 6-week-old female BALB/c nude mice (Clea Japan). After cell inoculation, ␤-estradiol solution was subcutaneously injected into the neck twice at intervals of 6 days. Fourteen days after the last ␤-estradiol injection, tumor-bearing mice were randomized and divided into two or three treatment groups as follows: for Fig. 7b

Development of potent SNIPER derivatives against ER␣
pounds were administered via intraperitoneal injection. After the indicated treatment periods, the mice were sacrificed, and tissues were excised. Total lysates from the ovaries and tumors were analyzed by Western blotting.

In vivo tumor growth inhibition
Each suspension of 1 ϫ 10 7 MCF-7 cells was mixed with an equal volume of Matrigel (BD Biosciences) and inoculated (100 l total) into the left and right mammary fat pads of 6-week-old female BALB/c nude mice (Clea Japan) that had received an ␤-estradiol pellet (6 g/day) (Innovative Research of America, Sarasota, FL) under the neck skin. After 4 days, mice bearing tumors of ϳ100 mm 3 in size were randomized and divided into two groups (n ϭ 7). One group served as a control for the dosing vehicle; the other groups were administered SNIPER(ER)-87 or SNIPER(ER)-110 (30 mg/kg, intraperitoneally, every 24 h), respectively. Tumor volumes were measured every 2 days using a caliper and calculated according to the standard formula: (length ϫ width 2 )/2. At 16 days, mice were sacrificed, and the tumors were weighed and excised.

Cell cycle analysis
After treatment, cells were gently trypsinized and washed with serum-containing medium. Cells were collected by centrifugation and then washed with PBS and fixed in 70% ice-cold ethanol for 1 h on ice. The cells were then washed, treated with 1 mg/ml RNase A for 1 h at 37°C, and stained in PI solution (50 g/ml in 0.1% sodium citrate, 0.1% NP-40). The stained cells were analyzed in a FACScan flow cytometer (BD Biosciences).

Measurement of apoptosis by flow cytometer
Apoptosis was analyzed with an annexin V-FITC apoptosis detection kit (BioVision). After treatment, cells were gently trypsinized and washed with serum-containing medium. Cells were collected by centrifugation and then washed with PBS and resuspended in binding buffer. The cells were stained with annexin V-FITC and PI at room temperature for 5 min in the dark, according to the manufacturer's instructions, and analyzed in a FACScan flow cytometer (BD Biosciences).

Statistical analysis
Student's t test was used to evaluate differences among the experimental groups. Values of p Ͻ 0.05 were considered significant.