Blocking cholesterol efflux mechanism is a potential target for antilymphoma therapy

Abstract Cholesterol is an essential plasma membrane lipid for the maintenance of cellular homeostasis and cancer cell proliferation. Free cholesterol is harmful to cells; therefore, excessive free cholesterol must be quickly esterified by acetyl‐coenzyme A:cholesterol acetyltransferase (ACAT) and exported by scavenger receptor class B member I (SR‐BI) or ATP‐binding cassette protein A1 from specific cells such as macrophage foam cells, which contain cholesteryl ester‐derived vacuoles. Many vacuoles are present in the cytoplasm of Burkitt lymphoma cells. In this study, we observed that these vacuoles are often seen in high‐grade lymphomas. Cell culture study using lymphoma cell lines found that esterified cholesterol is the main component of these vacuoles and the expression of cholesterol metabolism‐related molecules was significantly upregulated in lymphoma cell lines, with SR‐BI and ACAT inhibitors (BLT‐1 and CI‐976, respectively) impeding lymphoma cell proliferation. Cytoplasmic free cholesterol was increased by ACAT and SR‐BI inhibitors, and the accumulation of free cholesterol induced lymphoma cell apoptosis by inducing endoplasmic reticulum stress. Furthermore, synergistic effects of SR‐BI and ACAT inhibitors were observed in a preclinical study. Treatment with SR‐BI inhibitor suppressed lymphoma progression in a tumor‐bearing mouse model, whereas ACAT inhibitor did not. Therefore, SR‐BI inhibitors are potential new antilymphoma therapeutics that target cholesterol metabolism.


| INTRODUC TI ON
Non-Hodgkin lymphoma, including Burkitt lymphoma, DLBCL, and ATLL, are highly aggressive lymphomas that have difficult clinical courses. 1 The overall survival of patients with Burkitt lymphoma and DLBCL is significantly improved by additional treatment with rituximab, 2,3 although relapsed leukemia/lymphoma is resistant to this treatment, and novel therapeutic strategies are now in clinical trials. Adult T-cell leukemia/lymphoma is associated with retroviral infections, such as human T-cell leukemia virus type I or human T-cell lymphotropic virus type I. Outcomes from aggressive types of ATLL are poor, even when new anti-CCR4 Abs are used as an additional treatment. 4 In addition, most cases of PCNSL are high-grade DLBCL and standard therapeutic protocol is chemotherapy with HD-MTX, combination therapy with rituximab, and WBRT. However, the 5-year overall survival with combined HD-MTX and WBRT is approximately 20%. 5,6 Therefore, the detailed mechanisms underlying the resistance of aggressive lymphomas to antilymphoma therapies should be identified to guide the treatment of lymphoma.
Cholesterol is one of the main components of cell membranes and is a precursor of bile acid and steroidogenic hormones. 7 It is transported around the body in its bound form: LDL, or HDL. Lowdensity lipoprotein is delivered into cells through receptor-mediated endocytosis by the LDLR, while ABCA1 and SR-BI are involved in cholesterol transport mediated by HDL binding. Cholesterol uptake and its metabolic upregulation are important factors in cancer cell growth and proliferation. 8 Low-density lipoprotein receptor and the cholesterol synthesis enzyme HMG-CoA reductase are overexpressed in cancer cells, including leukemia/lymphoma cells. 9 Excess free cholesterol is harmful to cells 10,11 ; therefore, free cholesterol needs to be quickly esterified to CE by ACAT. The efflux of accumulated free cholesterol is a mechanism for regulating cholesterol homeostasis. Many researchers have attempted to treat cancer cells with ACAT or HMG-CoA reductase inhibitors. In brain cancers, an LXR agonist promotes tumor cell death by inducing LDLR downregulation and increasing ABCA1 expression, and the antitumor effect of the LXR agonist is observed in a preclinical model. 12,13 Therefore, targeting cholesterol metabolism could be a promising approach to cancer therapy; however, the effects of these drugs in treating cancers, including leukemia/lymphoma, are currently limited.
Many vacuoles are detected in the cytoplasm of Burkitt lymphoma cells, which are important characteristics in morphological diagnosis. 14 Similarly, intracytoplasmic vacuoles are also observed in some circulating ATLL cells 15 ; however, their components are uncharacterized. Comparable vacuoles composed of accumulated CE are observed in foamy macrophages, which are the primary indicators of atherosclerosis. 16 Therefore, the lymphoma vacuoles might contain large amounts of cholesterol and play an important role in cell proliferation. Targeting cholesterol metabolism could be an effective antilymphoma therapy.
In the present study, we observed vacuoles in high-grade lymphomas and hypothesized that these vacuoles reflected excessive cholesterol uptake with a high accumulation of CE. Molecules related to cholesterol homeostasis were overexpressed in several Tcell and B-cell lymphoma cell lines, including PCNSL. Interestingly, an SR-BI inhibitor induced lymphoma cell apoptosis and prolonged survival in a preclinical model. The SR-BI inhibitor also synergistically increased the cytotoxic effects of the ACAT inhibitor. The present study indicated that blocking cholesterol efflux could be a novel antilymphoma strategy. Intracytoplasmic vacuoles were evaluated in 10 randomly selected areas of high-power field of a microscope by two pathologists who were blinded to information about the patients' backgrounds or their prognosis. Peripheral blood mononuclear cells were acquired from healthy volunteer donors. Written informed consent for sample collection and subsequent analysis was obtained. All protocols using human macrophages were approved by the Kumamoto University Hospital Review Board (No. 486) and carried out in accordance with the approved guidelines.

| Cell staining
May-Grünwald-Giemsa and Sudan black B were purchased from Wako Chemicals, while Filipin III was obtained from Sigma-Aldrich.
Staining was carried out according to the manufacturer's protocol.

| Electron microscopy
The following cells were cultured with or without 40 μg/ml LDL for 24 h, cells were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate buffer for 1 h and postfixed in 1% osmium tetroxide. After the dehydration in a graded series of ethanol solutions with propylene oxide and embedding in Epon 812, ultrathin sections were cut with an ultratome, stained with uranyl acetate and lead citrate, and observed with a Hitachi H-7700 electron microscope (Hitachi).

| Chemicals
The ACAT inhibitor CI-976 was purchased from Santa Cruz Biotechnology.
The SR-BI inhibitor BLT-1 was purchased from Sigma-Aldrich. Lipoproteindeficient serum was purchased from Johnson Matthey. Human LDL and HDL were purchased from Biomedical Technology.

| Cytotoxicity assay
Cell viability was determined with a water-soluble tetrazolium salt (WST-8) colorimetric assay using a CCK-8 (Dojindo). Cells were cultured in a 96-well plate (approximately 10,000-20,000 cells/well) for 1-2 days and then subjected to the indicated treatments. After treatment, 10 µl WST-8 solution was added to each well and incubated for 30-60 min. Absorbance at 450 nm was measured using a microplate reader (SpectraMax i3x system; Molecular Devices). The culture medium containing WST-8 without cells was used to set the background threshold, while culture medium containing WST-8 with cells was used as a control. Cell viability was determined with the following formula: Cell viability (%) = (sample − background) / (control − background) × 100.

| Intracellular lipid analysis
Intracellular total cholesterol, free cholesterol, and triglyceride analyses were carried out using the cholesterol E test, the free cholesterol E test, and the triglyceride E test, respectively (Wako Chemicals), according to the manufacturer's protocol.

| Real-time quantitative PCR
Total RNA was isolated using RNAiso Plus reagent, and reversetranscribed using the PrimeScript RT reagent kit (Takara Bio).
Real-time quantitative PCR was carried out using TaqMan polymerase with SYBR green fluorescence (Takara Bio) and an ABI PRISM 7300 sequence detector (Applied Biosystems). The primer sequences are described in Table S1. and an ImageQuant LAS-4000 mini (Fujifilm).

| Statistics
All data are representative of two or three independent experiments.
Data are expressed as mean ± SD. Differences between the groups were examined for statistical significance using the Mann-Whitney U-test and nonrepeated measures ANOVA with Bonferroni's test.
Overall murine survival was assessed using Kaplan-Meier analysis and compared using the log-rank test. Statistical significance was set at p < 0.05.

| Intracytoplasmic vacuoles observed in lymphoma cells have high malignant potential
We first examined the cytoplasmic vacuoles in the touch imprint cytology samples; we divided the cases into two groups according to the percentage of vacuole-positive lymphoma cells (negative, none or <5%; positive, ≥5%). Follicular lymphoma is low-grade/indolent, whereas ATLL and DLBCL are aggressive subtypes of lymphoma.
Lymphoma cells with intracytoplasmic vacuoles were detected in 57.4%, 69.5%, and 13.3% of DLBCL, ATLL, and follicular lymphoma cases, respectively (Figures 1A and S1A). Cytoplasmic vacuoles were also observed in 6 of 11 PCNS lymphoma patients ( Figure 1A). Those cytoplasmic vacuoles were positively stained with Sudan black, indicating that lipid was contained in the vacuoles ( Figure 1B). Patients with vacuole-positive DLBCL had significantly shorter survival times than those without vacuoles ( Figure 1C). However, the presence of vacuoles was not associated with the GCB/non-GCB subtype, serum sIL-2R, or serum LDH ( Figure S1). These observations suggested that intracytoplasmic vacuoles were often present in high-grade lymphomas. We then investigated whether similar vacuoles could be induced by treatment with lipoproteins such as LDL and HDL under LPDS conditions. We found that treatment with LDL induced lipid accumulation and the appearance of vacuoles in lymphoma cell lines, but HDL did not induce lipid accumulation ( Figure 2D,E). As summarized in Table 1,  Table 1), whereas the content of triglyceride in lymphoma cells was not changed by the uptake of LDL ( Figure S2B). This indicated that intracytoplasmic vacuoles were associated with cholesterol accumulation in lymphoma cells. Low-density lipoprotein receptor is regulated by SREBP2, which is sensitive to cholesterol content in the ER at the transcriptional level. 19 Interestingly, the expression of LDLR and SREBP2 was unchanged in lymphoma cells following LDL stimulation, whereas nonlymphomatous T cells, the EBV-infected B cell line (141-LCL), and B cells showed the downregulation of the expression of both genes ( Figure 3C,D). This indicated that the abrogation of LDLR downregulation was caused by SREBP2 overexpression in lymphoma cells and suggested that LDL metabolism was needed for homeostasis processes, such as cell survival and proliferation in lymphomas.

| Upregulation of cholesterol metabolismrelated molecules in lymphoma cells
Low-density lipoprotein (40-80 µg/ml) enhanced cell proliferation in LDLR-overexpressing lymphoma cell lines using LPDS after 24 h (Figure 3E,F). In contrast, long-term (48 or 72 h) incubation using LDL (40-80 µg/ml) with LPDS decreased lymphoma cell viability, and the cytotoxicity of excessive LDL was abrogated by HDL (80 µg/ml) ( Figure 3F). This suggested that HDL inhibited free cholesterol accumulation by inducing cholesterol efflux through SR-BI or ABCA1. There was increased SR-BI and ACAT1 expression, and decreased ABCA1 expression in lymphoma cell lines ( Figure 3G). Acetyl-coenzyme A:cholesterol acetyltransferase 1, SR-BI, and SREBP2 increased in some lymphoma cell lines, whereas HMGCR did not substantially change ( Figure S4).
This suggested that LDLR, ACAT1, and SR-BI played important roles in cholesterol metabolism in lymphoma cells, and excessive free cholesterol was cleared by esterification and the efflux system ( Figure 3H).

| Acetyl-coenzyme A:cholesterol acetyltransferase and SR-BI inhibitors suppressed lymphoma cell proliferation in vitro
The ACAT inhibitor (CI-976) and SR-BI inhibitor (BLT-1) inhibited lymphoma cell proliferation in a dose-dependent manner ( Figure 4A).
In contrast, these inhibitors had no effect on the viability of activated T cells, 141-LCL cells, or B cells ( Figure 4B). The knockdown of SR-BI and ACAT in lymphoma cells also induced cell death and reduced lymphoma proliferation ( Figure S5). Caspase-3/7 activation in ATN-1 and ED cells was induced by treatment with either CI-976 or BLT-1 ( Figure 4C). The caspase inhibitor z-VAD-fmk protected against cell death induced by treatment with BLT-1 or CI-976 represented by DAUDI, BALL-1, and TK cells ( Figure 4E,F). In addition, CI-976 and BLT-1 synergistically increased the cytotoxic activity of CBDCA ( Figure S7). These findings suggest that cholesterol metabolism regulation using ACAT and SR-BI inhibitors could be a promising therapy for reducing lymphoma proliferation.

| Excessive LDL uptake enhanced the antilymphoma effect of ACAT and SR-BI inhibitors
Low-density lipoprotein uptake enhanced CI-976-and BLT-1induced cell death in lymphoma cells, whereas there was no effect on activated T cell proliferation ( Figure 5A). Treatment with both CI-976 and BLT-1 resulted in a greater accumulation of free cholesterol than that observed in the single-inhibitor treatments ( Figure 5B), and significantly inhibited cell proliferation in the presence of LDL ( Figure 5C). This showed that the enhanced accumulation of LDLderived free cholesterol by ACAT inhibitor blocking free cholesterol esterification, and SR-BI inhibitor blocking cholesterol efflux ( Figure   S8), selectively induced cell death in lymphoma cells. In contrast, HDL significantly attenuated the suppressive effects of CI-976 on lymphoma cell proliferation ( Figure 5D), while having no effect on activated T cell and lymphoma cell proliferation ( Figure 5E). The suppressive effects of BLT-1 on lymphoma proliferation were not influenced by HDL addition ( Figure 5F). These data indicated that free cholesterol efflux through the SR-BI-HDL pathway is involved in lymphoma cell viability.

| Scavenger receptor class B member I inhibitor induces ER stress through cholesterol accumulation in lymphoma cells
Excess cholesterol accumulation induces ER stress in macrophages. 20 The expression of CHOP, a major ER stress response protein, and other ER stress response-related proteins, including Bip, ATF6, p-eIF2α, and p-IRE1α, was induced in lymphoma cells following BLT-1 treatment ( Figure 6A,B). In addition, the MAPK kinases p38 and JNK were activated in lymphoma cells following BLT-1 treatment ( Figure 6C), which correlated with previous results showing that IRE1α induces apoptosis through the activation of p38 and JNK pathways. 21 Furthermore, BLT-1-induced cell death in lymphoma cells was suppressed by incubation with p38 and JNK inhibitors ( Figure S9). These data indicated that BLT-1-induced cholesterol accumulation is involved in apoptosis by promoting ER stress responses.

| Scavenger receptor class B member I inhibitor abrogated lymphoma progression and prolonged survival in vivo
The effects of i.p. administration of CI-976 and BLT-1 after ED cells were injected s.c. into SCID mice showed that BLT-1 significantly suppressed tumor development, whereas CI-976 displayed no antilymphoma effects ( Figure 7A). Cleaved caspase 3-positive cells were increased in s.c. tumor tissues treated with BLT-1 ( Figure 7B). BLT-1 significantly prolonged the survival in the lymphoma-bearing mouse model ( Figure 7C). In addition, BLT-1 significantly prolonged the survival in the PCNS lymphoma-bearing mouse model ( Figure 7D).

Cleaved caspase 3-positive cells were also increased in intracerebral
tumor tissues treated with BLT-1 ( Figure 7E). These results suggest that the SR-BI inhibitor could be a promising candidate agent for antilymphoma therapy.  Figure S8). The effect of SR-BI and ACAT inhibitors in high-grade lymphomas was stronger than that in low-grade lymphoma, indicating that targeting cholesterol metabolism is an effective therapy for high-grade lymphomas (Table S2 and Figure S10). The ACAT inhibitors have anticancer effects in some cancer cell lines [22][23][24] ; however, Cholesterol homeostasis is tightly regulated by a complex protein network that monitors cholesterol import, synthesis, export, metabolism, and esterification. 7  has an antilymphoma effect in combination with cholesterol starvation. 44 These researchers described the potential antilymphoma effect of an SR-BI inhibitor, which is consistent with the results in this study. The SR-BI inhibitor increased the cytotoxic effects of CBDCA on lymphoma cells, indicating that cholesterol efflux through SR-BI was associated with chemoresistance ( Figure S7). These results indicated that targeting cholesterol metabolism could be a promising approach for treating aggressive chemoresistant lymphomas.

| DISCUSS ION
Endoplasmic reticulum stress is a condition in which unfolded proteins possess an incorrect high-order structure and accumulate in the ER. As ER stress interferes with cell function, normal cells have an avoidance system called the UPR. 45,46 The CHOP transcription factor is involved in ER stress-dependent apoptosis if UPR does not work properly or is overstressed. [47][48][49] The accumulation of UPR in mammals is mainly mediated by stress sensor proteins including PERK, IRE1, and ATF6. 48,[50][51][52] The dissociation of a molecular chaperone (Bip) from stress sensor proteins is involved in the activation of UPR signaling by binding to UPR. 53,54 Endoplasmic reticulum stress is suggested to be a therapeutic target for cancer.
Toyocamycin is an IRE1 RNase domain inhibitor that induces apoptosis in myeloma cells. 55 In the present study, SR-BI inhibitor induced apoptosis through the ER stress signaling pathway in lymphoma cells ( Figure 6).
In conclusion, SR-BI and ACAT were overexpressed in lymphoma cells with LDLR expression maintained after LDL stimulation. There was no negative feedback system and cholesterol efflux was acti-

ACK N OWLED G M ENTS
We thank Ms. Ikuko Miyagawa, Ms. Michiko Tokunaga, and Mr.
Takenobu Nakagawa for their technical assistance. This work was supported by Japan Society for the Promotion Science KAKENHI (Grant numbers: 16H05162, 16K09247, 16K15503, and 20H03459). We would like to thank Editage for English language editing.

D I SCLOS U R E
The authors declare no competing interests. The authors A.M. and Y.K. are editorial board members.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets analyzed during the study are available from the corresponding authors on reasonable request.