Lipid nanoparticle siRNA cocktails for the treatment of mantle cell lymphoma

Abstract Mantle cell lymphoma is an aggressive and incurable subtype of non‐Hodgkin B cell lymphoma. Patients typically present with advanced disease, and most patients succumb within a decade of diagnosis. There is a clear and urgent need for novel therapeutic approaches that will affect mantle cell lymphoma through a unique mechanism compared to current therapies. This study examined the use of RNA interference (RNAi) therapy to attack mantle cell lymphoma at the mRNA level, silencing genes associated with cancer cell proliferation. We identified a lipid nanoparticle formulated with the lipidoid 306O13 that delivered siRNA to JeKo‐1 and MAVER‐1 mantle cell lymphoma cell lines. Three therapeutic gene targets were examined for their effect on lymphoma growth. These included Cyclin D1, which is a cell cycle regulator, as well as Bcl‐2 and Mcl‐1, which prevent apoptosis. Gene knockdown with siRNA doses as low at 10 nM increased lymphoma cell apoptosis without carrier‐mediated toxicity. Silencing of Cyclin D1 induced apoptosis despite a twofold “compensation” upregulation of Cyclin D2. Upon simultaneous silencing of all three genes, nearly 75% of JeKo‐1 cells were apoptosing 3 days post‐transfection. Furthermore, cells proliferated at only 15% of their pretreatment rate. These data suggest that lipid nanoparticles‐formulated, multiplexed siRNA “cocktails” may serve as a beneficial addition to the treatment regimens for mantle cell lymphoma and other aggressive cancers.

protein levels, RNA interference (RNAi) therapy may offer a viable means to kill cancer at the mRNA level. When used in combination with current treatments, RNAi may help to induce remission using lower doses of chemotherapy, which would better preserve treatment options upon relapse. Furthermore, an RNAi approach is apropos because mantle cell lymphoma cells overexpress several genes that encourage cell proliferation. 13,14 For example, Cyclin D1 (CCND1), a protein that facilitates cell cycle progression, is overexpressed in more than 90% of mantle cell lymphoma patients due to a t(11;14) (q13;q32) translocation of Cyclin D1 and immunoglobulin heavy chain genes (IgH). 5 Additionally, the anti-apoptotic proteins Bcl-2 and Mcl-1 are commonly overexpressed in mantle cell lymphoma and may contribute to chemotherapy resistance. [15][16][17] Although protein inhibitors of Bcl-2 and Mcl-1 have shown promise, [18][19][20][21][22] downregulation at the mRNA level with a viable delivery system has been overlooked.
Unfortunately, B-cells are notoriously difficult to transfect, and only a limited number of studies have reported on short interfering RNA (siRNA) delivery to lymphoma cells. [23][24][25][26][27] For example, we have previously described the ability of siRNA-loaded lipid nanoparticles (LNPs) to silence Mcl-1 expression and increase the apoptosis rates of mantle cell lymphoma. 28 Because our LNPs are potent transfection agents, [29][30][31][32][33] we herein investigate their use for multiplexed gene silencing in mantle cell lymphoma cells. Our data show that the simultaneous silencing of three key genes involved in mantle cell lymphoma growth induces apoptosis in the majority of lymphoma cells. As such, siRNA "cocktails" may hold promise as a mechanistically distinct addition to current mantle cell lymphoma treatment regimens.

| R E SU LTS
In this study, we sought to identify an RNAi treatment with potential to enhance multi-pronged mantle cell lymphoma therapy. Mantle cell lymphoma patients have a poor prognosis and patients that do not participate in clinical trials have a life expectancy of only 3 years after initial diagnosis. 2,4 The genetic mutations present in many mantle cell lymphoma patients make this malignancy an ideal candidate for RNA interference therapy. 25,34 Although lymphocytes can be particularly challenging to transfect compared to monocytes and adherent cells, we have previously established that lipidoid-containing LNPs facilitate gene silencing in mantle cell lymphoma cells. 28 In the present study, we first evaluated the ability of three lipidoids-303O 13 , 304O 13 , and 306O 13 (Figure 1a)-to durably silence mRNA expression in JeKo-1 immortalized mantle cell lymphoma cells. Our goal was to identify the most potent lipidoid, as it would be the best candidate for multiplexed gene silencing. For all three lipidoids, formulated nanoparticles were 70-100 nm in diameter with PDI <0.15 and siRNA entrapments greater than 75%.

| LNPs potently silenced GAPDH in human mantle cell lymphoma cells in vitro
LNPs containing the lipidoids 303O 13 , 304O 13 , and 306O 13 were examined for their ability to silence the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in vitro. Following a 24 hr incubation period with cells, each LNP mediated doseresponsive GAPDH knockdown ( Figure 1b). The lipidoid 306O 13 was the most potent of the three, with an EC 50 of less than 10 nM siGAPDH. No significant reduction in gene expression was observed for cells treated with control 306O 13 LNPs, which contained 100 nM siGFP. Additionally, no reduction in cell viability was observed in cells treated with 306O 13 LNPs (Supporting Information Figure 1). 306O 13 LNPs also silenced GAPDH (albeit more modestly) in MAVER-1 cells, a more aggressive human mantle cell lymphoma line, in a dose-responsive manner (Supporting Information Figure 2). We anticipate that this ability to mediate silencing in lymphoma cells with varying degrees of gene mutation will be important in developing a broadly applicable therapeutic. Because of their higher potency compared to 303O 13 13 , and 306O 13 -were formulated into siRNA-loaded LNPs. (b) Treatment with each LNP resulted in dose-dependent silencing of GAPDH in JeKo-1 mantle cell lymphoma cells 24 hr post-transfection. PBS and LNPs (siCntrl) at the maximum dose of 100 nM served as negative controls. (c) Following a single dose of 100 nM siGAPDH, 306O 13 LNPs mediated near-complete GAPDH knockdown for at least 3 days with maximal silencing achieved by 36 hr (blue circles). In each panel, error bars represent standard deviation (n 5 3). Statistically significant differences compared to cells given PBS were determined using two-tailed Welch's t tests. *, **, and **** indicate p .05, 0.01, and .0001, respectively 2.2 | LNPs silenced genes commonly overexpressed in mantle cell lymphoma After establishing that three-tailed 306O 13 is the most potent LNP for gene silencing in mantle cell lymphoma cells, we turned our focus to three genes that are upregulated in mantle cell lymphoma patients.
Cyclin D1 facilitates cell cycle progression from the G1 to the S phase through complex formation with Cyclin D kinases 4 and 6, which phosphorylate retinoblastoma 1. 35 Cyclin D1 overexpression occurs in more than 90% of mantle cell lymphoma patients as a consequence of a t (11;14) (q13;q32) translocation of Cyclin D1 and the IgH heavy chain. 36 This overexpression deregulates the cell cycle, leading to rapid progression from the G1 to the S phase and subsequent proliferation. It has been shown that Cyclin D1 downregulation leads to cell cycle arrest, apoptosis, and an increased sensitivity to chemotherapy. [23][24][25] Bcl-2 and Mcl-1 are antiapoptotic proteins that are also commonly overexpressed in mantle cell lymphoma. These proteins inhibit apoptosis by binding and sequestering two pro-apoptotic proteins, BAX and BAK. BAX and BAK's presence on the mitochondrial membrane is needed for the release of cytochrome c to the cytoplasm allowing for the activation of caspases during apoptosis. 16,[37][38][39][40] Downregulation of Bcl-2 has been shown to increase apoptosis in mantle cell lymphoma cells treated with immunotoxins in vitro. 41 Other studies have demon- Cyclin D1 is one of three Cyclin D homologues, all of which regulate the G1 to S phase transition in the cell cycle. 44 It has been shown previously that knockdown of Cyclin D1 can cause a compensatory increase in Cyclin D2 expression. 23  . Statistically significant differences compared to untreated cells were determined using two-tailed Welch's t tests. *, **, ***, and **** indicate p .05, 0.01, .001, and .0001, respectively experiment, the two treated samples each received a total siRNA dose of 100 nM. That is, the first group received a 50 nM siCCND1 dose 1 50 nM siControl dose, while the second group received a 50 nM siCCND1 dose 1 50 nM siCCND2 dose. It was not clear, however, that silencing Cyclin D2 would be necessary for effective therapy. This is because Cyclin D2 mRNA expression was more than 8,000-fold lower than Cyclin D1 levels in untreated samples (data not shown).
To determine whether Cyclin D2 siRNA was a necessary inclusion in a therapeutic siRNA cocktail, we measured JeKo-1 apoptosis rates following Cyclin D1 and Cyclin D2 protein silencing. We first measured the duration of Cyclin D1 mRNA silencing following a single 50 nM dose of siCCND1 to determine appropriate timing for the apoptosis measurement ( Figure 4a). Cyclin D1 mRNA levels were silenced by 75% for 3 days before returning to baseline by day 6. We, therefore, shown). Therefore, we decided to target only Cyclin D1 instead of a Cyclin D1-D2 combination in future experiments.

| Multiplexed gene silencing improved apoptosis rates
In previous work, we showed that silencing Mcl-1 in JeKo-1 and MAVER-1 cells using siRNA-loaded LNPs caused cells to undergo apoptosis. 28 We thought it may be possible to improve upon these results by targeting multiple genes in one treatment, given the potency of our LNPs. Therefore, we attempted to simultaneously knockdown Mcl-1, Cyclin D1, and another antiapoptotic protein, Bcl-2 using a cocktail of the three siRNAs. Figure 5  Given these positive results, we examined whether or not the formulation procedure for the LNP cocktail affected apoptosis rates. One formulation was made by pre-mixing the three siRNAs and then  Figure 4), suggesting that the cell entry of LNPs is not an effect-limiting step in vitro. We recommend the first, pre-mixed siRNA formulation strategy, as it is simpler.

| D I SCUSSION
Mantle cell lymphoma is one of the most deadly subtypes of B-cell non-Hodgkin lymphoma. 1 Although new treatments (e.g., rituximab) have improved outcomes over the last 20 years, survival in the general patient population remains very low. 2,4 As clinical therapies have progressed from chemotherapy to small molecules drugs to immunotherapy, RNA interference therapy remains an untapped clinical option with the potential to improve treatment outcomes through a unique therapeutic mechanism.
In this study, we formulated LNPs containing synthetic, ionizable, lipid-like materials, termed "lipidoids," to deliver siRNA to mantle cell lymphoma cells. Because lipidoid-containing LNPs had previously been shown to effectively silence genes in a variety of cell types, we sought to examine their ability to transfect notoriously difficult B-cells.
Although several groups have used nanotherapeutics to knockdown Cyclin D1 with siRNA or shRNA in mantle cell lymphoma, [23][24][25]47 other genes have been largely overlooked.
We found that lipid nanoparticles formulated from the lipidoid 306O 13 mediate potent and durable silencing in human mantle cell lymphoma cells (Figure 1). We then used these lipidoid nanoparticles to examine the effect of silencing a trio of genes-Cyclin D1, Bcl-2, and Mcl-1-on mantle cell lymphoma apoptosis and proliferation rates. The overexpression of Cyclin D1, which occurs in more than 90% of mantle cell lymphoma patients, 5,36 is correlated with higher rates of tumor cell growth and decreased survival. 48 We also examined two anti-apoptotic proteins in the Bcl-2 family: Bcl-2, which is commonly overexpressed in  (Figures 3 and 4). Klier and colleagues found that Cyclin D1 exists at greater than 1,000-fold higher levels than Cyclin D2 in JeKo-1 cells. 53 Our qPCR analysis also determined that total Cyclin D1 mRNA existed at a much higher level compared to Cyclin D2 mRNA, regardless of treatment (data not shown). This likely explains why the upregulation of Cyclin D2 following the knockdown of Cyclin D1 does not preclude therapeutic effect (i.e., increased apoptosis). Because we found that silencing Cyclin D2 did not increase apoptosis rates (Figure 4d), we chose not to include it as a target in siRNA cocktails.
Multiplexed gene silencing has not previously been attempted for the treatment of mantle cell lymphoma. Simultaneous silencing of multiple gene targets requires a delivery system that is sufficiently potent to induce therapeutic levels of protein knockdown without causing toxicity. Lipid delivery systems, particularly those with a permanent positive charge, can be associated with local toxicity, such as cell irritation and cell lysis, and systemic toxicity causing inflammatory cytokines to be released. 54 In this study, we demonstrate that a triple siRNA cocktail delivered with nanoparticles made from the ionizable lipidoid 306O 13 caused increased apoptosis and decreased proliferation in mantle cell lymphoma cells. Based on apoptosis rates ( Figure 5), it appears that silencing at least one protein from each pathway (i.e., Cyclin D1 plus either Mcl-1 or Bcl-2) results in higher apoptosis rates than when Cyclin D1 is excluded. Figure 6, however, shows greater reductions in proliferation

| C ONC LUSI ON
These results highlight the ability of LNPs to deliver a triple siRNA cocktail targeting multiple pathways for therapeutic effect in mantle cell lymphoma. LNPs delivering multiplexed siRNA were more potent than LNPs carrying siRNA targeting a single gene at equivalent siRNA doses. RNAi therapy, which offers a unique mechanism compared to current treatments, has potential to enhance currently available treatment options by increasing their potency while reducing resistance and treatment-related toxicity.

| Lipid nanoparticle formulation
Lipid nanoparticles were formulated as previously described. 28 Briefly, a lipid solution containing lipidoid, DSPC, cholesterol, and C14 PEG2000 (50:10:38.5:1.5 molar ratio) in ethanol with 5% sodium citrate buffer (pH 3-4) was added to an equal volume of siRNA diluted in sodium citrate buffer. The mixture was briefly vortexed and further diluted 1:1 in PBS (pH 7.4). 306O 13 LNPs were dialyzed for 4 hr in PBS. siRNA entrapment and particle size were determined using a Quant-iT Ribogreen RNA Reagent assay (Thermofisher Scientific) and dynamic light scattering, respectively. Particle sizes are reported as number mean.

| Cell culture
Immortalized human mantle cell lymphoma cell lines JeKo-1 and MAVER-1 were cultured at 378C in 5% CO 2 in RPMI 1640 with 20% and 10% FBS, respectively. About 100 U/ml Penicillin Streptomycin was added to the cell culture media.

| Cell viability experiment
JeKo-1 cells were seeded at 250,000 cells/ml in 96 well plates and treated with LNPs as described above. After 24 hr, cell viability was quantified using a MTT Cell Proliferation Assay Kit (ATCC) and Synergy H1 Hybrid Reader (BioTek Instruments, Winooski, Vermont) following the manufacturer's protocol.

| Apoptosis experiment
Cells were seeded at 250,000 cells/ml. At several time points between 1 and 8 days following transfection with siRNA-loaded LNPs, the fraction of cells undergoing apoptosis was determined using an Alexa Fluor 488 Annexin V/Dead Cell Apoptosis Kit and BD Accuri C6 flow cytometer (BD Biosciences, San Jose, CA) following the manufacturer's protocol. FlowJo software was used for flow cytometry data analysis.

| Cell counting experiment
JeKo-1 cells were seeded at 250,000 cells/ml in 24 well plates. Cells were treated with 200 nM doses of total siRNA. The cell count per 20 ll aliquot was determined using a BD Accuri C6 flow cytometer with a HyperCyt Autosampler (Intellicyt, Albuquerque, NM) for high throughput processing. FlowJo software was used for flow cytometry data analysis.

| Statistical analysis
All statistical analysis was performed using GraphPad Prism (La Jolla, CA) software. Error bars represent standard deviation (n 5 3).