Mechanisms of Sensitivity and Resistance of Primary Effusion Lymphoma to Dimethyl Fumarate (DMF)

PEL is a rare B cell lymphoma associated with KSHV that mainly arises in immune-deficient individuals. The search for new drugs to treat this cancer is still ongoing given its aggressiveness and the poor response to chemotherapies. In this study, we found that DMF, a drug known for its anti-inflammatory properties which is registered for the treatment of psoriasis and relapsing–remitting MS, could be a promising therapeutic strategy against PEL. Indeed, although some mechanisms of resistance were induced, DMF activated NRF2, reduced ROS and inhibited the phosphorylation of STAT3 and the release of the pro-inflammatory and immune suppressive cytokines IL-6 and IL-10, which are known to sustain PEL survival. Interestingly, we observed that DMF displayed a stronger cytotoxic effect against fresh PEL cells in comparison to PEL cell lines, due to the activation of ERK1/2 and autophagy in the latter cells. This finding further encourages the possibility of using DMF for the treatment of PEL.


Introduction
Dimethyl fumarate (DMF; trade name: Tecfidera), a drug registered for the treatment of relapsing-remitting multiple sclerosis (MS) and psoriasis [1], has been shown to also display anti-cancer activity [2,3]. Although the underlying mechanisms have not been completely elucidated, most DMF activities seem to be related to its capacity to deplete cellular GSH (via GSH succination) and activate nuclear factor erythroid 2 (NF-E2)-related factor 2 (NRF2) through the succination and inhibition of Kelch-like ECH-associated protein 1 (KEAP1), removing the inhibitory activity on NRF2 [4]. Indeed, the activation of NRF2 and the inhibition of nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) [5,6] underly the capacity of DMF to reduce the inflammatory process. Through NRF2 activation, DMF modulates the expression of a variety of anti-oxidant enzymes, including heme oxygenase 1 (HO-1) and NADPH dehydrogenase (quinone 1) (NQO1), thus dysregulating the intracellular radical oxygen species (ROS) level, whose fine balance is required for cancer cell survival [7]. Accordingly, we have previously shown that Apigenin (a natural compound abundantly contained in vegetables and fruits) inhibited signal transducer and activator of transcription 3 (STAT3) [8] (a pathway whose activation plays a key role in the survival of Primary Effusion Lymphoma (PEL) cells [9,10]) by reducing ROS. However, Apigenin, as well as other flavonoids, has been reported to mediate its cytotoxic effect against other cancer types by increasing ROS rather than reducing them [11], which nonetheless confirms the dependency of cancer cell survival on the appropriate amount of these molecules, which are involved in multiple biological processes [12].
PEL is a rare, aggressive B cell lymphoma carrying Kaposi's Sarcoma Herpesvirus (KSHV) and, in most of the cases, also Epstein Barr virus (EBV) infection. The search for more effective 2 of 12 therapies for this cancer is needed, given its poor response to conventional chemotherapies such as cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) [13,14]. Drugs targeting STAT3 may represent a promising therapeutic option against this cancer [9,15,16], even if several other oncogenic pathways contribute to driving lymphoma cell survival, including PI3K/AKT/mTOR, NFkB, ERK1/2 and mevalonate [17][18][19][20][21]. Interestingly, STAT3 has also been reported to be among the pathways inhibited by DMF, e.g., in hepatocellular carcinoma [22]. AKT/mTOR/ERK signaling has also been shown to be inhibited by DMF in melanoma cells when used in combination with Vemurafenib [23]. Based on this collection of evidence and the anti-inflammatory activity of DMF, in this study, we investigated the possibility of treating PEL cell lines and fresh PEL cells with this drug. Indeed, besides the above-mentioned oncogenic pathways, PEL survival strongly relies on the autocrine effect of pro-inflammatory cytokines such as IL-6 [24], whose release is strictly interconnected with the activation of oncogenic pathways such as STAT3 [25,26]. Moreover, the impact of DMF on autophagy was evaluated in PEL cell lines and fresh PEL cells, given that the autophagic process, interconnected with several oncogenic pathways, may strongly influence the response to anti-cancer treatments [27].

Results
2.1. DMF Impairs PEL Cell Survival by Activating NRF2, Reducing Intracellular ROS and STAT3 Activation, and Counteracting the Release of IL-6 and IL-10 Cytokines BC3 and BCBL-1 PEL cells were treated with different concentrations (20 and 50 µM) of DMF for 24 h; cell viability was then assessed by trypan blue assay. As shown in Figure 1A,B, DMF induced a dose-dependent and a time-dependent reduction of cell survival in both cell lines. DMF is known to activate NRF2 [28]; accordingly, in this study, we found that NQO1, one of the NRF2 targets, [29], was upregulated in a dose-dependent manner by this drug ( Figure 1C). We then evaluated the expression level of p62/SQSTM1, a protein which is another NRF2 target able to establish a positive feedback loop with this transcription factor [30]. We found that p62/SQSTM1 expression increased following DMF treatment either at the protein ( Figure 1D) or mRNA level ( Figure 1E). In correlation with the activation of the antioxidant response, ROS were reduced by DMF, as demonstrated by FACS analysis using DCF-DA staining ( Figure 1F). Of note, DMF also altered the mitochondrial dynamics, which is strictly correlated with the function of these organelles, as it downregulated Mitofusin (MFN) 2 ( Figure 1G), a protein which regulates mitochondrial fusion. The reduction of ROS was previously found to be a mechanism leading to a reduction of PEL cell survival following treatment with Apigenin as well as the antioxidant N-acetylcysteine (NAC), which inhibited STAT3 by reducing ROS [8]. A proper level of ROS contributes to the activation of oncogenic pathways and thus may help tumor cell proliferation [7]. In accordance with these previous findings [22], in this study, we observed that DMF reduced STAT3 tyrosine 705 phosphorylation in both PEL cell lines ( Figure 1H), suggesting that this could be a molecular mechanism underlying its cytotoxic effect.
Given that STAT3 activation is strongly interconnected with the release of the proinflammatory and immunosuppressive cytokines such as IL-6 and IL-10, we then performed a Luminex assay to investigate the impact of DMF on their production. As shown in Figure 1I, DMF reduced both IL-6 and IL-10 cytokines, which are required for PEL cell growth, by acting in an autocrine fashion [24]. This suggests that the interruption of the feedback loop between STAT3 and these pro-survival cytokines by DMF could play an important role in the effect exerted against PEL cells. Interestingly, we found that counteracting the activation of NRF2 by using Brusatol, the cytotoxic effect of DMF decreased in both PEL cell lines ( Figure 1J), suggesting that its cytotoxicity involved NRF2 activation.

p-4EBP1 Activation Counteracts the Cytotoxic Effect of DMF at a Low Dose
Given that STAT3 was de-phosphorylated by DMF at both 20 and 50 µM doses, particularly in BC3 cells, we then evaluated whether the activation of pro-survival pathway(s) could counteract its cytotoxic effect when used at the lower dose of 20 µM. We investigated the phosphorylation of 4EBP1 (an mTOR target), as the activation of this pathway sustains PEL cell survival [19] and can be influenced by DMF treatment [23]. As shown in Figure 2A, at 20 µM dose, p-4EBP1 was hyperactivated by DMF in both BC3 and BCBL1 cells. To demonstrate that the activation of p-4EBP1 could represent a mechanism of resistance to DMF treatment in this setting, we used this drug at 20 µM with NVP-BEZ235, a dual mTOR inhibitor. Such a combination strongly impaired p-4EBP1 activation ( Figure 2B) and potentiated the cytotoxic effect of DMF ( Figure 2C) as well as PARP cleavage ( Figure 2D), suggesting that mTOR activation represented a mechanism of resistance of PEL cells to the treatment with a low dose of DMF. µM DMF for 24 h. (J) Percentage of cell viability of BC3 and BCBL1 cells, co-treated with 20 µM DMF and 5 nM Brusatol (BRU) for 24 h, evaluated by Trypan blue exclusion assay. In the figure, mean ± standard deviation (SD) of three independent experiments is shown. * p < 0.05.

p-4EBP1 Activation Counteracts the Cytotoxic Effect of DMF at a Low Dose
Given that STAT3 was de-phosphorylated by DMF at both 20 and 50 µM doses, particularly in BC3 cells, we then evaluated whether the activation of pro-survival pathway(s) could counteract its cytotoxic effect when used at the lower dose of 20 µM. We investigated the phosphorylation of 4EBP1 (an mTOR target), as the activation of this pathway sustains PEL cell survival [19] and can be influenced by DMF treatment [23]. As shown in Figure 2A, at 20 µM dose, p-4EBP1 was hyperactivated by DMF in both BC3 and BCBL1 cells. To demonstrate that the activation of p-4EBP1 could represent a mechanism of resistance to DMF treatment in this setting, we used this drug at 20 µM with NVP-BEZ235, a dual mTOR inhibitor. Such a combination strongly impaired p-4EBP1 activation ( Figure  2B) and potentiated the cytotoxic effect of DMF ( Figure 2C) as well as PARP cleavage ( Figure 2D), suggesting that mTOR activation represented a mechanism of resistance of PEL cells to the treatment with a low dose of DMF.  was assessed by western blotting in BC3 and BCBL1 cells, treated or not (CT) with 20 and 50 µM DMF for 24 h. β-actin was used as loading control, and one representative experiment out of three is shown. The histograms represent the mean plus SD of the densitometric analysis of the cl PARP/β-actin ratio of three different experiments. In the figure, mean ± standard deviation (SD) of three independent experiments is shown. * p < 0.05.

DMF Induces a Lower Cytotoxic Effect against BC3 Cell Line Compared to Fresh PEL Cells Due to ERK1/2 Activation
We then compared the cytotoxic effect exerted by DMF on PEL cell lines with that induced on fresh PEL cells grown in SCID mice. To this end, BC3 cells were engrafted i.p. and, after 35 days, the ascites was withdrawn and cells were isolated, as this lymphoma mainly grows in liquid effusions. After having checked that the cells contained in ascites were PEL cells, based on positivity for LANA1 (a latent KHSV antigen always expressed in these lymphoma cells ( Figure 3A), we treated them with DMF at 20 and 50 µM. We found that fresh PEL cells were more sensitive to DMF than the BC3 cell line when the drug was used at the 50 µM dose ( Figure 3B). Indeed, at the lower dose of 20 µM, DMF also induced the activation of p-4EBP1 in these cells ( Figure 3C). Searching for the molecular mechanism(s) underlying the different cytotoxicity of DMF used at 50 µM against BC3 and fresh PEL cells, we observed an increase of ERK1/2 phosphorylation in the BC3 cell line, an effect that was not induced by DMF in fresh PEL cells ( Figure 3D). To investigate whether ERK1/2 activation could be responsible for the lower cytotoxic effect of DMF against the BC3 cell line, we pre-treated these cells with PD98059 ERK1/2 inhibitor before exposing them to DMF and found that such a combination was more cytotoxic than DMF or PD98059 alone ( Figure 3E). The efficiency of ERK1/2 inhibition by PD98059 was also demonstrated by western blot analysis, and the higher cytotoxicity of DMF in combination with PD98059 was confirmed by the increase in PARP cleavage ( Figure 3F).

The Induction of a Pro-Survival Autophagy by ERK1/2 Activation Counteracts the Cytotoxic Effect of DMF in BC3 Cell Line
We then evaluated the impact of DMF on autophagy, in the BC3 cell line and in fresh PEL cells, to investigate whether a different activation of this process could contribute to the different cytotoxic effect of DMF against these two cell types. To this end, considering that p62/SQSTM1 was transcriptionally upregulated by DMF through NRF2 activation ( Figure 1C,D), we decided that p62/SQSTM1 could not be used as an autophagic marker; thus, we assessed the expression level of LC3II, the lipidated form of LC3 that is formed during autophagy activation. LC3I/II was evaluated in the presence or in the absence of chloroquine, as this drug prevents LC3II degradation, which also occurs during the autophagic process. We found that LC3II increased in the presence of chloroquine in the BC3 cell line, while such increase was not observed in fresh PEL cells, suggesting that autophagy was only activated in the BC3 cell line ( Figure 4A). We then observed that chloroquine increased the cytotoxic effect of DMF in the BC3 cell line but not in fresh PEL cells ( Figure 4B), indicating that autophagy played a pro-survival role in the DMF-treated BC3 cell line. We finally investigated whether the induction of autophagy could correlate with ERK1/2 activation in the DMF-treated BC3 cell line. To this end, we pre-treated these cells with PD98059 ERK1/2 inhibitor and found that it counteracted autophagy induction by DMF ( Figure 4C). Overall, these findings suggest that DMF, by inducing ERK1/2 activation, triggered autophagy in the BC3 cell line, and this effect reduced its cytotoxic effect.
cell line, an effect that was not induced by DMF in fresh PEL cells ( Figure 3D). To investigate whether ERK1/2 activation could be responsible for the lower cytotoxic effect of DMF against the BC3 cell line, we pre-treated these cells with PD98059 ERK1/2 inhibitor before exposing them to DMF and found that such a combination was more cytotoxic than DMF or PD98059 alone ( Figure 3E). The efficiency of ERK1/2 inhibition by PD98059 was also demonstrated by western blot analysis, and the higher cytotoxicity of DMF in combination with PD98059 was confirmed by the increase in PARP cleavage ( Figure 3F).

The Induction of a Pro-Survival Autophagy by ERK1/2 Activation Counteracts the Cytotoxic Effect of DMF in BC3 Cell Line
We then evaluated the impact of DMF on autophagy, in the BC3 cell line and in fresh PEL cells, to investigate whether a different activation of this process could contribute to the different cytotoxic effect of DMF against these two cell types. To this end, considering that p62/SQSTM1 was transcriptionally upregulated by DMF through NRF2 activation activation in the DMF-treated BC3 cell line. To this end, we pre-treated these cells with PD98059 ERK1/2 inhibitor and found that it counteracted autophagy induction by DMF ( Figure 4C). Overall, these findings suggest that DMF, by inducing ERK1/2 activation, triggered autophagy in the BC3 cell line, and this effect reduced its cytotoxic effect.

Discussion
This study shows for the first time that DMF could represent a promising therapeutic option against PEL, particularly against fresh PEL cells, in which such treatment did not trigger ERK1/2 and autophagy activation, as a defense mechanism that counteracted its cytotoxic effect. Besides ERK1/2, which was hyper-phosphorylated by DMF at the dose of 50 µM in the BC3 cell line, the mTOR/p-4EBP1 axis was activated by DMF in both PEL cell lines and fresh PEL cells when it was used at the lower dose of 20 µM. This pathway also reduced DMF cytotoxicity, as suggested by the stronger cell death induced by the combination of DMF with NVP-BEZ235. Drug resistance is a major obstacle to cancer therapies; therefore, it is of pivotal importance to unveil the molecular mechanisms and the processes that sustain cancer cell survival in the course of anti-cancer treatments, such as p-4EBP1 and ERK1/2 in the case of PEL cells treated by DMF at low and high doses, respectively. However, despite this resistance, DMF still induced a cytotoxic effect against PEL and PD98059 EK1/2 (PD) inhibitor. β-actin was used as loading control, and one representative experiment out of three is shown. The histograms represent the mean plus SD of the densitometric analysis of the LC3II/β-actin ratio of three different experiments. In the figure, mean ± standard deviation (SD) of three independent experiments is shown. * p < 0.05.

Discussion
This study shows for the first time that DMF could represent a promising therapeutic option against PEL, particularly against fresh PEL cells, in which such treatment did not trigger ERK1/2 and autophagy activation, as a defense mechanism that counteracted its cytotoxic effect. Besides ERK1/2, which was hyper-phosphorylated by DMF at the dose of 50 µM in the BC3 cell line, the mTOR/p-4EBP1 axis was activated by DMF in both PEL cell lines and fresh PEL cells when it was used at the lower dose of 20 µM. This pathway also reduced DMF cytotoxicity, as suggested by the stronger cell death induced by the combination of DMF with NVP-BEZ235. Drug resistance is a major obstacle to cancer therapies; therefore, it is of pivotal importance to unveil the molecular mechanisms and the processes that sustain cancer cell survival in the course of anti-cancer treatments, such as p-4EBP1 and ERK1/2 in the case of PEL cells treated by DMF at low and high doses, respectively. However, despite this resistance, DMF still induced a cytotoxic effect against PEL cells, and this occurred in correlation with NRF2 activation, ROS reduction, and de-phosphorylation of STAT3. These lymphoma cells are indeed strongly addicted to this pathway [15], which is constitutively activated by the release of pro-inflammatory cytokines that act in an autocrine fashion and, due to the expression of proteins encoded by the oncovirus KSHV [31], are carried in a latent state in PEL cells. Of note, STAT3 inhibition by DMF correlated with the reduction of IL-6 and IL-10, in accordance with previous findings [32], and these pro-inflammatory and immunosuppressive cytokines are known to be strongly involved in PEL cell survival [24]. Of note, it has been reported that one of the most important therapeutic effects of DMF relies on its anti-inflammatory properties, which is why this drug has been registered for the treatment of inflammatory-based diseases such as psoriasis and multiple sclerosis [33,34]. Regarding cancer, previous studies have demonstrated the anti-cancer effects of DMF against several solid cancers [35,36] and hematologic cancers, including cutaneous T-cell lymphoma [37], diffuse large B cell lymphoma (DLBL), and HTLV-1-associated adult T cell leukemia [38]. At the molecular level, NF-kB was found to be inhibited by DMF and, in the latter two cancer types, the de-phosphorylation of STAT3 was also induced. The other important effect of DMF that we observed in this study was that it promoted autophagy in the BC3 cell line, while this process was not activated in fresh PEL cells. Autophagy represented a pro-survival mechanism induced in correlation with the activation of the ERK1/2 pathway [39]; indeed, through autophagy, cancer cells may often be resistant to chemo and radiotherapies. Therefore, it is worthwhile to evaluate the impact of treatments on this process and to explore whether its inhibition can be used as strategy to potentiate the outcome of treatments [27]. Interestingly, autophagy was not induced by DMF in fresh PEL cells; indeed, we observed a higher cytotoxic effect against these cells. Cells were treated with DMF for 24 h, Chloroquine was added for the last 18 h, and all inhibitors were added to the cells 30 min before DMF.

Trypan Blue Exclusion Assay
Cells were harvested after treatments and then centrifuged and counted by Trypan blue exclusion assay (Trypan blue, Sigma-Aldrich, St. Louis, MO, USA, cat. n. T8154) by using a Neubauer hemocytometer in a phase contrast microscope. Each experiment was performed in triplicate.

Antibodies
The following antibodies were diluted in PBS/T containing BSA 3% and used to detect the expression of specific proteins on western blot membranes: Goat anti-rabbit IgG-horseradish peroxidase HRP (Santa Cruz Biotechnologies, Dallas, TX, USA, sc-2004) and goat anti-mouse IgG-horseradish peroxidase HRP (Santa Cruz Biotechnologies, Dallas, TX, USA, sc-2005) were diluted 1:10,000 in PBS/T containing BSA 3% and used as secondary antibodies.

Densitometric Analysis
Bands evidenced by western blot were subjected to densitometric analysis by using ImageJ software, which was downloaded from the NIH web site (available online: http://imagej.nih.gov (accessed on 5 May 2022)).

Quantitative Real-Time PCR
Total RNA levels from PEL cell lines treated with DMF (50 µM) for 24 h and from those untreated (CT) were extracted by using TRIzol™ Reagent (Thermo Fisher, Waltham, MA, USA, cat. n.15596026) according to the manufacturer's instructions. After in vitro reverse transcription, quantitative Real-Time PCR (qRT-PCR) was performed and TaqMan gene expression assay specific to p62 mRNA was used as described elsewhere [40]. All values were normalized to β-actin (endogenous gene controls).

Statistical Analysis
The results are represented by the mean ± standard deviation (S.D.) of at least three independent experiments, and a two-tailed Student's t-test was used to demonstrate statistical significance. Differences were considered statistically significant when the pvalue was at least <0.05.

Conclusions
In conclusion, this study presents for the first time that DMF could be a promising treatment against PEL, particularly because it was even more effective against fresh lymphoma cells. We also evidenced the molecular pathways underlying its cytotoxic effect and those that counteracted it, which could be targeted to potentiate the outcome of DMF treatment against PEL, an aggressive lymphoma that still remains an orphan drug cancer.