Silver Nanoparticles Selectively Treat Neurofibromatosis Type 1-Associated Malignant Peripheral Nerve Sheath Tumors in a Neurofibromin-Dependent Manner

Neurofibromatosis type 1 (NF1) is among the most common neurogenic disorders, characterized by loss of function mutations in the neurofibromin gene (NF1). NF1 patients are extremely susceptible to developing neurofibromas, which can transform into deadly malignant peripheral nerve sheath tumors (MPNSTs). At the center of these tumors are NF1-null Schwann cells. Here, we found that nanomedicine shows promise in the treatment of NF1-associated MPNSTs. We assessed the cytotoxicity of silver nanoparticles (AgNPs) in NF1-null NF1-associated MPNSTs, NF1-wildtype sporadic MPNST, and normal Schwann cells. Our data show that AgNP are selectivity cytotoxic to NF1-associated MPNSTs relative to sporadic MPNST and Schwann cells. Furthermore, we found that sensitivity to AgNPs is correlated with the expression levels of functional neurofibromin. The restoration of functional neurofibromin in NF1-associated MPNSTs reduces AgNP sensitivity, and the knockdown of neurofibromin in Schwann cells increases AgNP sensitivity. This finding is unique to AgNPs, as NF1 restoration does not alter sensitivity to standard of care chemotherapy doxorubicin in NF1-associated MPNSTs. Using an in vitro model system, we then found that AgNP can selectively eradicate NF1-associated MPNSTs in co-culture with Schwann cells at doses tolerable to normal cells. AgNP represents a novel therapy for the treatment of NF1-associated MPNSTs and addresses significant unmet clinical need.


Introduction
Neurofibromatosis type 1 (NF1) is an autosomal dominant disease that affects 1:3000, making it one of the most common neurogenic disorders [1]. NF1 is characterized by loss of function mutations in the neurofibromin gene (NF1), resulting in a host of comorbidities. Nearly 100% of people with NF1 will develop tumors, called neurofibromas, within the nervous system [2]. These neurofibromas can be debilitating and deadly, depending on the type and location. Unfortunately, the current standard of care for unresectable neurofibromas shows limited utility. Peripheral nerve sheath tumors (PNSTs) are a heterogeneous group of neurofibromas that affect predominantly NF1 patients. These PNSTs most often develop after a biallelic loss of functional NF1 in the Schwann cells. PNSTs range from benign PNSTs, including plexiform neurofibromas (pNFs), to highly deadly malignant PNSTs (MPNSTs). MPNSTs are extremely aggressive soft tissue sarcomas that can develop from malignant transformation of pNFs and other atypical neurofibromas [3]. MPNSTs are the most common pediatric sarcoma and the leading cause of death in NF1 patients [4]. gifted from Margaret Wallace (University of Florida). The 293T cells were obtained from the American Type Culture Collection.

Silver Nanoparticles
Spherical silver nanoparticles stabilized by polyvinylpyrrolidone, 25 nm in diameter, were purchased from Nanocomposix, Inc., as dried powders, with a 15% silver content. Silver nanoparticles were dispersed in phosphate buffered saline (PBS), pH 7.4, without calcium or magnesium (Corning) by bath sonication. Silver nanoparticles were dispersed at a final concentration of 5 mg mL −1 silver content by weight, stored at 4 • C in the dark, and used within 14 days of dispersion. Silver nanoparticles were diluted directly in complete cell culture medium for viability and imaging studies. The physicochemical properties of silver nanoparticles (hydrodynamic diameter, colloidal stability, and ζ-potential) were previously characterized [24,25].

Lentivirus Production
The 293T plates were seeded in 100 mm TC dishes and allowed to attach overnight under conditions to reach~70% confluence after 20 h. The following day, 6.4 µg envelope plasmid (psPAX (Addgene Plasmid #12259)), 1.9 µg pMD2.G (Addgene Plasmid #12260) and 5.6 µg lentiviral expression plasmids of interest-pLKO.1-puro Non-Target shRNA Control Plasmid DNA (Sigma SHC016), shNF1 (TRCN0000039713, TRCN0000039714 (Sigma)-dsRed2-RFP (Addgene 109377), or GFP (Addgene 17448) were combined in deionized water to a total volume of 100 µL in a polystyrene culture tube. Polyethylenimine (41.7 µg, linear MW 25000, Polysciences cat# 23966-1) was added to the plasmid mixture and allowed to incubate for 15 min at room temperature. The plasmid:PEI mixture was added to the 293T plates at 70% confluence in 10 mL fresh complete culture medium and incubated for 18 h at 37°C/5% CO 2 . The following day, the medium was aspirated and replaced with 15 mL complete culture growth medium and incubated for 48 h. The medium was removed and filtered using 0.45 µM PES filter (VWR). Lentiviral supernatant was aliquoted for single-use and stored at −80 • C.

Neurofibromin Knockdown
Predesigned and previously validated short hairpin RNA plasmids against neurofibromin from the RNAi Consortium shRNA Library (TRCN0000039713, TRCN0000039714) and non-targeting controls were purchased from Sigma Aldrich. Lentiviral supernatants were prepared using the 293T cell line, as detailed above. STS26T, iHSC1λ, and iHSC-2λ cells (1.6-2.0 × 10 5 cells per well) were seeded in 6-well tissue culture plates (VWR) and allowed to attach overnight. The medium was aspirated, and 0.5 mL viral supernatant (shNT-non-targeting), shNF1-713, shNF1-714, or vehicle was mixed with 0.5 mL complete growth medium and added to each well for 24 h. Cells were allowed to recover for 24 h, then medium containing puromycin (2 µg mL −1 ) was added. After selection was com-plete, pooled clones were propagated until a sufficient number of cells for the experiments was obtained.

Restoration of Functional Neurofibromin Expression
The S462TY (6 × 10 5 cells) cells were seeded in 60 mm TC dishes and allowed to attach for 48 h. GFP expression vector (Addgene 17448) or GFP-tagged neurofibromin 1 expression vector (pLenti-NF1-GFP, kindly provided by Robert F. Hennigan (University of Cincinnati)) or vehicle (water) was added to deionized water (4.67 µg GFP plasmid; 4.67 µg for NF1-GFP-Lo; 9.33 µg for NF1-GFP-Hi) to a total volume of 100 µL. Polyethylenimine (41.7 µg, linear MW 25000) (Polysciences cat# 23966-1) was added to the plasmid mixture, mixed well, and allowed to incubate for 15 min at room temperature. The plasmid:PEI mixture was added to the S462TY cells in 3 mL fresh growth medium and incubated for 18 h at 37°C/5% CO 2 . The following day, the medium was aspirated and replaced with 5 mL fresh growth medium and incubated for 24 h. Cells were puromycin selected (2 µg mL −1 ) until the vehicle transfected control cells were no longer viable. Cells were propagated (2 or fewer passages) until sufficient cells were available for experimental studies.

Realtime Quantitative PCR
Confluent plates of cells (60 mm) were harvested in 1 mL TRIZOL reagent (Invitrogen). RNA was extracted using Direct-zol RNA kits (Zymo Research), according to the manufacturer's protocol. Quality was assessed using a DeNovix DS-11+ spectrophotometer. DNA was transcribed using 0.6-1 µg RNA input with the High Capacity cDNA Reverse Transcription Kit with RNase Inhibitor kit (Applied Biosystems) using a GeneAmp PCR System 9700 (Applied Biosystems), according to the manufacturer's protocol, in a total reaction volume of 20 µL. Realtime quantitative PCR was performed using Taqman reagents specific for neurofibromin 1 (Hs01035108_m1) and reference gene PPIA (Hs04194521_s1) or TBP (Hs00427620_m1) [31], using 1 µL of cDNA mixture in a 384-well format in quadruplicate using Taqman Fast Advanced Master Mix (Applied Biosystems), per the manufacturer's protocol, using a QuantStudio 6 Flex (Applied Biosystems). Relative levels of neurofibromin transcript were calculated using ∆∆Ct methodology.

Fluorescently Labeled Cells
S462TY (6.0 × 10 5 ) or iHSC1λ (7.5 × 10 5 ) cells were seeded in 60 mm TC dishes and allowed to attach overnight. The following day, the medium was aspirated and replaced with 0.5 mL viral supernatant (GFP or RFP, respectively, or vehicle, as detailed above) and 2.5 mL complete growth medium and incubated for 18 h. The following day, the medium was aspirated and replaced with 5 mL fresh growth medium and incubated for 24 h. Cells were then puromycin selected (2 µg mL −1 ) until control-transduced cells were no longer viable. GFP and RFP expression was assessed using fluorescence microscopy; cells termed S462TY-GFP and iHSC-1λ-RFP were propagated under normal growth conditions, and cryogenically preserved.

Other Reagents
Doxorubicin was purchased from TCI. Selumetinib was purchased from Selleck Chemical. Silver nitrate and cobalt chloride were manufactured by Acros.

Imaging Studies
S462TY-GFP (8.0 × 10 4 ) and iHSC1λ-RFP (8.0 × 10 4 ) cells were seeded in the same well in 6-well TC dishes. The following day, cells were treated with AgNP (0-50 µg mL −1 Ag) or vehicle in 3 mL culture medium for 0-72 h. Each day, triplicate wells were washed once in 1 mL Live Cell Imaging Solution (Gibco), and then 1 mL Live Cell Imaging Solution was added. Cells were imaged using a EVOS FL Microscope (Life Technologies) under 10× magnification. Cells were imaged using transmitted light, GFP LED cube, and TxRed LED cube to capture all S462TY-GFP and iHSC1λ-RFP cells, respectively. After imaging, cells were harvested for immunoblotting, as detailed below. Triplicate wells containing 1.6 × 10 5 S462TY-GFP or iHSC1λ-RFP were treated with AgNP and viability assessed by MTT, as described above, to monitor the viability of the remaining cells.

Western Blot
Cell lysates were harvested in MPER protein extraction reagent (Thermo Scientific, Waltham, MA, USA) supplemented with Halt™ Protease Inhibitor Cocktail (Thermo Scientific). Cell lysates were sonicated briefly, and protein content was assessed by BCA protein assay (Thermo Fisher, Waltham, MA, USA) in 96-well plates, per the manufacturer's protocol. Protein (10 µg per lane) was diluted in 4× Laemmli Buffer (Bio-Rad, Hercules, CA, USA) and heated at 95 • C for 8 min. Lysates were loaded into 15-well Mini-Protean TGX Gels (Bio-Rad) and protein ladder (Chameleon Pre-stained Duo Color Ladder, LI-COR, Lincoln, NE, USA), and electrophoresis was performed. Proteins were transferred to nitrocellulose membranes and blocked in Intercept Blocking Buffer (TBS) (LI-COR). Blots were probed with antibodies against GFP (Cell Signaling 2555), RFP (Rockland 600-401-3795), or loading control β-actin (Cell Signaling 4970) at a dilution of 1:1000 in Intercept Blocking Buffer overnight at 4 • C with gentle rocking. Membranes were than washed 3 times for 5 min in TBS, then anti-Rabbit fluorophore-conjugated secondary antibody (IRDye 800 CW Goat Anti-rabbit (926-32211), at a dilution of 1:5000 in Intercept Blocking Buffer, was added. Membranes were rocked at room temperature for 90 min in the dark, then blots washed 3 times for 5 min in TBS, as described above. Membranes were imaged using an Odyssey Clx Imaging System (LI-COR).

Statistical Analysis
IC 50 was calculated by means of Graphpad Prism 9 software using a variable slope with the least squares regression method. Significance between and within groups was determined by one-way ANOVA and Student's t-test, where appropriate. Significance is denoted as follows: *** p < 0.001, ** p < 0.01, * p < 0.05, n/s p > 0.05.

Silver Nanoparticles Are Selectively Cytotoxic to NF1-Associated MPNSTs
MPNSTs most often develop in patients with NF1 and harbor a complete loss of functional neurofibromin. Rarely, MPNSTs can also develop in a sporadic manner and maintain expression of functional neurofibromin. Regardless of NF1 mutational status, all MPNSTs share Schwann cells as the cell of origin [32]; thus, Schwann cells are an appropriate normal cell control. Therefore, to test the possibility of AgNPs as a potential therapeutic for MPNSTs, we evaluated the cytotoxicity of AgNPs in a panel of cell lines representing NF1-associated NF1-deficient MPNSTs (S462TY, ST-8814), sporadic NF1-wildtype MPNST (STS26T), and normal Schwann cells (iHSC-1λ, iHSC-2λ) as a normal cell control (Table 1). We treated our cell panel with AgNPs ranging from 0-500 µg mL −1 (by Ag content) for 72 h. AgNPs were highly cytotoxic to the NF1-associated (NF1-deficient) MPNST cell lines tested and were significantly less cytotoxic to sporadic NF1-wildtype MPNST and control normal Schwann cells ( Figure 1A). NF1-associated NF1-deficient MPNSTs (shown in red) were significantly more sensitive to AgNP than normal Schwann cells (shown in blue). Sporadic MPNST (shown in green) showed a similar sensitivity to normal Schwann cells and were significantly less sensitive compared to NF1-associated MPNSTs ( Figure 1B). Further, there was no intragroup difference in cytotoxicity based on IC 50 , demonstrating evidence of subtype selectivity.
The standard of care for MPNSTs of any grade is surgery, if possible, followed by doxorubicin coupled with radiotherapy. Therefore, we treated our cell panel with standard of care chemotherapy doxorubicin (0-2000 nM) for 72 h and monitored viability. In stark contrast to AgNPs, there was no clear delineation in doxorubicin sensitivity between NF1associated MPNSTs, sporadic MPNST, and normal Schwann cells ( Figure 1C). There were significant differences in doxorubicin sensitivity within groups ( Figure 1D). These data suggest that there could be significant improvement in clinical outcome for NF1-associated MPNST using AgNPs compared to the standard of care doxorubicin chemotherapy. Red is NF1 associated MPNST, green is sporadic MPNST, and blue is normal scwhann cell in entire paper. Red is NF1 associated MPNST, green is sporadic MPNST, and blue is normal scwhan paper.

Intact Silver Nanoparticles Are Required for NF1-Associated MPNST Selectivi
We previously showed that rapid oxidation (activation) of AgNPs in Significance between and within groups was determined by one-way ANOVA and Student's t-test, where appropriate (*** p < 0.001, * p < 0.05, n/s not significant).

Intact Silver Nanoparticles Are Required for NF1-Associated MPNST Selectivity
We previously showed that rapid oxidation (activation) of AgNPs into silver ion (Ag + ) in triple negative breast cancer is likely responsible for the subtype-selective AgNPmediated cytotoxic effect [25]. To test the hypothesis that the oxidation into Ag + in the cell contributes to AgNP selecticity in NF1-associated MPNST, we treated NF1-associated MP-NST cells and normal Schwann cells with equivalent silver using AgNP or AgNO 3 . Silver nitrate was used as a source for Ag + . We found that NF1-associated MPNST ( Figure 2A) and normal Schwann cells ( Figure 2B) were similarly sensitive to Ag + , but showed differential cytotoxicity with equivalent silver mass in the form of AgNP. These data demonstrate the necessity of the nanoparticle formulation of Ag 0 for cancer-selective therapy against NF1-associated MPNST.
2A) and normal Schwann cells ( Figure 2B) were similarly sensitiv differential cytotoxicity with equivalent silver mass in the form demonstrate the necessity of the nanoparticle formulation of Ag 0 for apy against NF1-associated MPNST.

Increased Oxidative Stress Sensitizes Normal Schwann and Sporadic Nanoparticles
Reactive oxygen species can directly degrade (and ionize) Ag 0 totoxic Ag + [28]. Therefore, we sought to investigate the role of oxid tolerant cells. We used cobalt chloride (CoCl2) as a cell tolerable indu Mechanistically, CoCl2 mimics hypoxia through blockade of the Vo tein and the hypoxia-inducible factor, resulting in increased ROS a tathione [33,34] Normal Schwann cells ( Figure 3A) or sporadic N ( Figure 3B) were treated with CoCl2 (0-150 μM) in the presence or 62.5 μg mL −1 Ag) for 48 h and viability assessed. In both cell lines t as single agents had only modest effects on viability. When both age simultaneously, the cytotoxicity was remarkably amplified. This e increasing oxidative stress potentiates AgNP-mediated cytotoxici stress may play a causal role in the mechanism of AgNP mediated

Increased Oxidative Stress Sensitizes Normal Schwann and Sporadic MPNST Cells to Silver Nanoparticles
Reactive oxygen species can directly degrade (and ionize) Ag 0 from AgNPs into cytotoxic Ag + [28]. Therefore, we sought to investigate the role of oxidative stress in AgNP-tolerant cells. We used cobalt chloride (CoCl 2 ) as a cell tolerable inducer of oxidative stress. Mechanistically, CoCl 2 mimics hypoxia through blockade of the Von Hippel-Lindau protein and the hypoxia-inducible factor, resulting in increased ROS and a reduction in glutathione [33,34] Normal Schwann cells ( Figure 3A) or sporadic NF1-wildtype MPNST ( Figure 3B) were treated with CoCl 2 (0-150 µM) in the presence or absence of AgNP (0-62.5 µg mL −1 Ag) for 48 h and viability assessed. In both cell lines tested, CoCl 2 or AgNP as single agents had only modest effects on viability. When both agents were administered simultaneously, the cytotoxicity was remarkably amplified. This evidence suggests that increasing oxidative stress potentiates AgNP-mediated cytotoxicity, and that oxidative stress may play a causal role in the mechanism of AgNP mediated cytotoxicity.

Restoration of Functional Neurofibromin Expression in NF1-Associated MPNST Increases Tolerance to Silver Nanoparticles
Sensitivity to AgNP in MPNSTs and normal Schwann cells correlates with the expression of functional neurofibromin, as we found all neurofibromin-expressing cell models are significantly less sensitive to AgNP therapy relative to cell models that lack functional neurofibromin ( Figure 1A). To test whether functional neurofibromin expression had a direct effect on AgNP sensitivity, we restored neurofibromin expression in a model of NF1-associated MPNST (S462TY). Cells were transfected with two relative doses of an expression vector containing functional neurofibromin tagged with GFP (NF1-GFP), or a control expression vector containing only GFP. After puromycin selection, pooled clones of transfected cells (termed S462TY-GFP, S462TY-NF1-GFP-Lo, and S462TY-NF1-GFP-Hi) were challenged with AgNP for 72 h. Cell models with restored expression of functional neurofibromin were significantly less sensitive to AgNPs, compared to control transfected cells ( Figure 4A). The NF1 gene dose by transfected plasmid amount correlated with the IC 50 of AgNP ( Figure 4B). These data provide evidence that functional neurofibromin expression is inversely related to AgNP-mediated cytotoxicity. ed. 2022, 12, x FOR PEER REVIEW Figure 3. Cobalt chloride-induced hypoxic/oxidative stress se sporadic NF1-wildtype MPNST to AgNP treatment. (A) Normal radic NF1-wildtype MPNSTs (STS26T) were exposed to CoCl2 was assessed by MTT. Data are shown ± SD and are represen ments. Significance between treatments was determined by Stu n/s not significant). sporadic NF1-wildtype MPNSTs (STS26T) were exposed to CoCl 2 and AgNPs for 48 h, and viability was assessed by MTT. Data are shown ± SD and are representative of three independent experiments. Significance between treatments was determined by Student's t-test (*** p < 0.005, * p < 0.05, n/s not significant).

Restoration of Functional Neurofibromin Expression in N Tolerance to Silver Nanoparticles
It is critical to further evaluate whether this loss of therapeutic sensitivity is unique to AgNP, or rather a pan-desensitization mediated through global signaling alterations induced by the restoration of functional neurofibromin. Therefore, we treated our cell models (S462TY-GFP, S462TY-NF1-GFP-Lo, S462TY-NF1-GFP-Hi) with standard of care chemotherapy doxorubicin, as described above. All cell derivatives were equally sensitive to doxorubicin ( Figure 4C). This finding is particularly intriguing, as the alterations in therapeutic sensitivity after neurofibromin restoration is unique to AgNP, and do not occur in all therapies.

Knockdown of Neurofibromin Sensitizes Normal Schwann Cells, but Not Sporadic MPNSTs, to Silver Nanoparticles
Sporadic MPNSTs arise from a distinct mechanism compared to NF1-associated MP-NST and often maintain expression of functional neurofibromin [3]. We next sought to test if we could sensitize NF1-wildtype cells to AgNPs by reducing neurofibromin expression. Thus, sporadic MPNST (STS26T) cells were transduced with neurofibromin-targeting validated short hairpin RNA (shRNA) lentiviral constructs (shNF1-713 and shNF1-714), or non-targeting controls (shNT), sourced from the publicly available RNAi Consortium. Transduced cells were selected with puromycin, and pooled clones were treated with AgNP for 72 h. We found that knockdown of neurofibromin does not alter sensitivity to AgNP in sporadic MPNST (Figure 5A,B). Knockdown was verified by quantitative PCR (Figure 5C). This provides evidence that AgNP would be the most useful in NF1-associated MPNSTs, and likely not effective in sporadic MPNSTs.  In NF1 patients, peripheral nerve sheath tumors develop after the complete loss of functional NF1 in normal Schwann cells. Based on the finding that the restoration of functional NF1 decreases sensitivity to AgNPs, we sought to investigate whether the reduction of neurofibromin expression would increase sensitivity to AgNPs in tumor cells of the Schwann origin cells. Knockdown of neurofibromin expression was performed, as described above, in normal Schwann cell lines (iHSC1λ, iHSC2λ). Knockdown of neurofibromin increased sensitivity to AgNP in both normal Schwann cell lines tested in a gene dose dependent manner ( Figure 6A). Neurofibromin knockdown was validated using quantitative PCR ( Figure 6B). Since loss of functional NF1 in normal Schwann cells is an initiating event in the formation of plexiform neurofibroma, and ultimately, in malignant transformation, we challenged the neurofibromin-reduced and control cells with standard of care MEK inhibitor selumetinib. We found that there were no alterations with sensitivity to selumetinib, regardless of neurofibromin expression ( Figure 6C). Similar to previous data, neurofibromin-mediated alterations in therapeutic sensitivity are unique to AgNP. Our evidence again suggests that AgNP sensitivity correlates with the expression of neurofibromin in a dose-dependent manner in NF1-related MPNSTs and cell models of cancer initiation.

Silver Nanoparticles Selectively Remove NF1-Associated MPNST Cells in Co-Culture with Normal Schwann Cells
The tumor microenvironment contains multitudes of cell types comprising normal cells and cancer cells. Successful treatment of cancer in the clinic ultimately relies on the ability to safely eradicate cancerous tissue while leaving the normal tissues unharmed. Our findings show that there is likely a therapeutic window separating NF1-associated MPNSTs and normal Schwann cells. In order to model whether our AgNPs would be able to safely remove cancerous cells while leaving normal cells unharmed, we employed a co-culture system. In order to track individual cell types, we stably transduced NF1-associated MPNST cells to express green fluorescent protein (termed S462TY-GFP) and normal Schwann cells to express red fluorescent protein (termed iHSC1λ-RFP). An equal number of cells were seeded in the same culture dish and then treated with increasing concentrations of AgNP. Microscopy showed that AgNP could eradicate NF1-associated MPNST S462TY-GFP cells, while allowing normal Schwann cell iHSC1λ-RFP to remain ( Figure 7A). Relative levels of RFP and GFP were also visualized by immunoblot assay (Figure 7B). To ensure that iHSC1λ-RFP were indeed viable at the same concentrations, monocultures of iHSC1λ-RFP and S462TY-GFP were treated with AgNP at the same concentrations as the co-culture, and viability was assessed by MTT ( Figure 7C). This model system demonstrates that a therapeutic window may exist for the safe eradication of MPNST, while leaving normal Schwann cells unharmed. Figure 6. Knockdown of neurofibromin increases sensitivity to AgNPs in normal Schwann with no change in sensitivity to plexiform neurofibroma standard of care selumetinib. N Schwann cell lines iHSC1λ and iHSC2λ were transfected with non-targeting control shNT or (713 or 714) and puromycin selected. (A) iHSC1λ-shNT, -shNF1-713, and -shNF1-714 (l iHSC2λ-shNT, -shNF1-713, and -shNF1-714 (right) were treated with AgNP (0-500 μg mL −1 A 72 h, and viability was assessed by MTT assay. IC50 of AgNP is shown as μg mL −1 Ag. (B) specific for NF1 and the housekeeping gene PPIA was performed in quadruplicate, and relativ transcript levels ± SD were calculated using ΔΔCT methods, as shown. (C) iHSC1λ-shNT shNF1-713 (left) or iHSC2λ-shNT, -shNF1-713 (right) were treated with standard of care selum for 72 h, and viability was assessed by MTT. Data are representative of 3 independent experi each containing at least 4 technical replicates. *** p < 0.001, ** p < 0.01, * p < 0.05, n/s p > 0.05.

Silver Nanoparticles Selectively Remove NF1-Associated MPNST Cells in Co-Culture Normal Schwann Cells
The tumor microenvironment contains multitudes of cell types comprising n cells and cancer cells. Successful treatment of cancer in the clinic ultimately relies o Figure 6. Knockdown of neurofibromin increases sensitivity to AgNPs in normal Schwann cells, with no change in sensitivity to plexiform neurofibroma standard of care selumetinib. Normal Schwann cell lines iHSC1λ and iHSC2λ were transfected with non-targeting control shNT or shNF1 (713 or 714) and puromycin selected. (A) iHSC1λ-shNT, -shNF1-713, and -shNF1-714 (left) or iHSC2λ-shNT, -shNF1-713, and -shNF1-714 (right) were treated with AgNP (0-500 µg mL −1 Ag) for 72 h, and viability was assessed by MTT assay. IC 50 of AgNP is shown as µg mL −1 Ag. (B) qPCR specific for NF1 and the housekeeping gene PPIA was performed in quadruplicate, and relative NF1 transcript levels ± SD were calculated using ∆∆CT methods, as shown. (C) iHSC1λ-shNT and -shNF1-713 (left) or iHSC2λ-shNT, -shNF1-713 (right) were treated with standard of care selumetinib for 72 h, and viability was assessed by MTT. Data are representative of 3 independent experiments, each containing at least 4 technical replicates. *** p < 0.001, ** p < 0.01, * p < 0.05, n/s p > 0.05. model system demonstrates that a therapeutic window may exist for the safe eradicatio of MPNST, while leaving normal Schwann cells unharmed.

Discussion
MPNSTs are the most common pediatric sarcoma and the leading cause of death in NF1 patients [4]. The current standard of care consists of surgery, when possible, chemotherapy (doxorubicin), and radiation therapy. The five-year survival rate of NF1associated MPNST patients is a discouraging 52.5% [35]. Improvements must be made in the clinical management of MPNSTs, as tumors are often unresectable and respond poorly to chemotherapy, and radiation is dangerous and contributes to additional tumor formation. Rationally designed and administered nanomedicine, in this case AgNPs, provides a novel avenue for therapy and addresses significant unmet clinical need. We found that NF1associated MPNSTs are sensitive to AgNPs in a neurofibromin-dependent mechanism at doses that are tolerable to normal Schwann cells.
Silver nanoparticles (AgNPs) are cytotoxic to a variety of cancer cells derived from an assortment of tissues at doses that have a minimal effect on normal cells [13][14][15][16][17][18][19][20][21][22][23][24][25]36]. To the best of the authors' knowledge, this report is the first use of AgNP as a novel therapy for the treatment of MPNSTs. AgNPs show a pleiotropic mechanism of action with a variety of cytotoxic effects. Of note, AgNPs are known to induce protein oxidation [27], DNA damage in the form of double-strand breaks [24,27], misfolded proteins [37], mitochondrial dysfunction [38], redox state imbalance [24,38], endoplasmic reticulum stress [25], and lipid peroxidation [36]. In general, AgNPs induce catastrophic cell injury in specific subsets of cancer cells, resulting in both apoptotic and necrotic death in breast and lung cancers. It is likely that AgNPs induce damages similar to those observed in other cancers, and subsequent apoptosis and necrosis in NF1-associated MPNSTs [25,37]. Across tumors of disparate origins, one major commonality correlated with AgNP sensitivity is the mesenchymal-like status of the cells. Expression of mesenchymal-like markers (ZEB1, CDH1) is correlated with AgNP sensitivity in breast [25,27,36], lung [38,39], and ovarian [24] cancer cells. NF1-associated MPNSTs are also enriched in mesenchymal-like features relative to normal Schwann cells [40], and our data supports a strong correlation between mesenchymal-like status and inherent susceptibility to AgNP therapy.
Selectivity is paramount to the clinical translation of novel therapeutics. We found that NF1-associated MPNSTs are~4.4-fold more sensitive to AgNP relative to normal Schwann cells, showing that there may be a therapeutic window for the clinical use of AgNP. This therapeutic window is unprecedented, compared to standard of care chemotherapy doxorubicin, as we found that there was no clear trend in doxorubicin sensitivity between cancer and non-cancer in the case of MPNSTs and normal Schwann cells (Figure 1). While these studies are limited to in vitro cell models, our data support that AgNP is more selective compared to the current standard of care, and further preclinical studies involving relevant in vivo model systems are warranted.
One major concern with novel therapy is safety during translation from bench to bedside. In preclinical studies using murine models, we found that systemically administered AgNPs (3 times per week, 10 weeks of treatment, 6 mg kg −1 Ag, through intravenous injection) were well tolerated and showed efficacy against triple-negative breast cancer [25]. No major untoward side effects were noted in rodents during the study. Importantly, IC 50 between NF1-associated MPNST cells (S462TY) and breast cancer cells used in vitro and in orthotopic tumors [25] (MDAMB231) were quite similar (11.9 ± 7.8 and 8.3 ± 1.2 µg mL −1 Ag, respectively). This provides further evidence that sufficient doses of AgNP could be systemically administered to safely treat NF1-associated MPNST using in vivo models. AgNPs also show no cytotoxic effect on normal cells derived from liver, kidney, and monocyte lineages at cancer therapeutic doses [27]. A recent case report shows the efficacy of homemade AgNPs against metastatic head and neck carcinoma of the nasal cavity [41]. Selfadministered AgNP treatment was initiated after multiple rounds of standard treatment regimens had failed. Remarkably, no toxicities were observed and a complete radiographic resolution of cancer was achieved. The patient was found to be cancer-free at follow up 18 months later [41]. Our in vitro data, coupled with clinical case reports, provide strong justification for the further evaluation of AgNPs for treatment of NF1-associated MPNSTs using preclinical in vivo model systems.
While our studies show strong evidence for safety, no therapeutic is without risk. To minimize untoward effects, it may be possible to use expression levels of functional neurofibromin as a biomarker for treatment with AgNP to ensure we are only treating patients that would benefit. This would be prudent, based on our findings that AgNP may not be useful in the treatment of sporadic MPNST, as they were significantly less sensitive than NF1-associated MPNST. This stark contrast may have to do with the dramatic differences in the etiology of NF1-associated MPNST compared to sporadic MPNST, as each arises via distinct mechanisms [42]. Firstly, not all sporadic MPNSTs (including the STS26T model used in this study) harbor a loss of functional NF1 and PRC2 components, which are seen in the majority of NF1-associated MPNSTs [43,44]. Further evaluation of these mediators and studies with additional models of sporadic MPNST would be useful. However, it is clear that neurofibromin-regulated pathways which alter sensitivity to AgNP therapy in NF1-associated MPNSTs and the tumor cell of origin Schwann cells exist.
An important finding in this study is that a reduction in neurofibromin expression is sufficient to sensitize normal Schwann cells to AgNPs ( Figure 6A). Loss of neurofibromin 1 is a 'first hit' and a major driving force in the development of atypical neurofibromas, plexiform neurofibromas, and ultimately, MPNSTs in NF1 patients [45]. Our data demonstrates that AgNP may be able to selectively remove pre-cancerous cells from NF1 patients at an extremely early stage. There may be a possible use of AgNP as a preventative therapy in NF1 patients who are prone to develop MPNSTs. This would be a major advancement in the clinical management of NF1 patients and must be investigated further.

Conclusions
In summary, we found that NF1-associated MPNSTs are exquisitely sensitive to AgNPs, compared to normal Schwann cells and sporadic MPNSTs. We then found that AgNPmediated cytotoxicity can be influenced with alterations in basal oxidative stress. Further investigations showed that AgNP-mediated cytotoxicity is dependent on the expression of neurofibromin, as the restoration of functional neurofibromin in NF1-associated MPNST cells decreased the sensitivity to AgNP, and the knockdown of neurofibromin increased the AgNP-mediated cytotoxicity of normal Schwann cells. These data represent a unique approach for a new therapeutic strategy in the clinical management of NF1-associated MPNSTs and show that further preclinical investigations are justified.

Data Availability Statement:
The datasets presented in this study are available upon request from the corresponding author.