Mutation of PTPN11 (Encoding SHP-2) Promotes MEK Activation and Malignant Progression in Neurofibromin-Deficient Cells in a Manner Sensitive to BRAP Mutation

Simple Summary Germline mutations of NF1 cause neurofibromatosis type 1 (NF1), which is characterized by multiple benign peripheral nerve sheath tumors known as neurofibromas. In some individuals with NF1, plexiform neurofibromas can give rise to malignant peripheral nerve sheath tumors. Here, we applied genomic DNA sequencing to NF1-derived tumors and identified additional genetic alterations in PTPN11 (encoding Src homology region 2 domain-containing phosphatase-2 (SHP)-2) and BRAP associated with NF1 tumor malignancy. We found that the forced expression of the mutant form of SHP-2 activated the protein kinase MEK and increased tumorigenic activity in NF1 cells, and that these effects were attenuated by the forced expression of the mutant form of BRCA1-associated protein (BRAP). This suppressive action of mutant BRAP was not apparent in NF1-intact cells. Our data indicate that the combination of NF1 mutation and PTPN11 mutation drives the malignancy of NF1 cells and that SHP-2 inhibition by BRAP is a potential therapeutic strategy for NF1-associated malignant tumors. Abstract Germline mutations of NF1 cause neurofibromatosis type 1 (NF1) through the activation of the RAS signaling pathway, and some NF1 patients develop malignant peripheral nerve sheath tumors (MPNSTs). Here, we established subclones of the human NF1-MPNST cell line sNF96.2 that manifest increased tumorigenic activity and increased phosphorylation of the protein kinases MEK and Akt relative to the parental cells. Genomic DNA sequencing identified 14 additional heterozygous mutations within the coding regions of 13 cancer- and other disease-related genes in these subclones. One of these genes, PTPN11, encodes SHP-2, and the forced expression of the identified G503V mutant of SHP-2 increased both tumorigenic activity and MEK phosphorylation in parental sNF96.2 cells, suggesting that the combination of PTPN11 and NF1 mutations induces the pathological activation of the RAS pathway. These effects of SHP-2 (G503V) were inhibited by the coexpression of the G370A mutant of BRAP, which was also detected in the highly malignant subclones, and this inhibition was accompanied by the calpain-dependent cleavage of SHP-2 (G503V). The cleavage of SHP-2 (G503V) and suppression of MEK phosphorylation mediated by BRAP (G370A) were not detected in NF1-intact (HeLa) cells. Tumor promotion by SHP-2 (G503V) and its suppression by BRAP (G370A) may serve as a basis for the development of new treatment strategies for NF1.


Introduction
Malignant peripheral nerve sheath tumors (MPNSTs) are soft-tissue sarcomas that surround peripheral nerves and occur either sporadically or more often in association with neurofibromatosis type 1 (NF1, von Recklinghausen disease), an inherited disorder caused by germline mutations of the NF1 gene [1,2]. NF1 is inherited in an autosomal-dominant manner and affects 1 in~3000 live births, with 30% to 50% of NF1 mutations arising de novo [3][4][5]. Individuals with NF1 manifest a variety of clinical symptoms, including peripheral and central nervous system tumors, pigmentary changes such as multiple café au lait spots, bone defects, cardiovascular abnormalities, and learning disabilities [6]. The development of single or multiple benign peripheral nerve sheath tumors known as neurofibromas is a common symptom of NF1. These lesions can occur at virtually any site in the body and are generally divided into nodular plexiform, diffuse plexiform, subcutaneous, and cutaneous subtypes [7,8]. Whereas the life expectancy of individuals with NF1 is almost normal, the malignant transformation of benign tumors is associated with a reduction in overall survival. Most MPNSTs arise from plexiform neurofibromas, and the overall prognosis of patients with such tumors is poor, with a 5-year survival rate of~34% to 52% having been reported [9]. Another recent cohort study showed that, among 1607 NF1 patients, 243 individuals (15.1%) developed MPNSTs, and that the 5-year disease-specific survival rate of these individuals was lower than that of patients with MPNSTs not related to NF1 (31.6% versus 43.4-71.9%) [10]. A better understanding of the mechanisms underlying the malignant transformation of neurofibromas would be expected to contribute to the development of therapeutic strategies that improve the prognosis of NF1 patients.
NF1 mutations have been identified both in NF1 patients and in individuals with sporadic brain, breast, or lung cancer [11]. NF1 is a tumor suppressor gene and encodes a 2818-amino acid protein termed neurofibromin that functions as a GTPase-activating protein for the small GTPase RAS (RAS-GAP) and therefore negatively regulates RAS signaling by promoting the conversion of the GTP-bound (active) form of RAS to the GDP-bound (inactive) form. The loss or mutation of neurofibromin thus results in the activation of RAS and the engagement of its downstream effectors [12,13].
The RAS-RAF-MEK-ERK (RAS-ERK) signaling pathway is affected in many human cancers as a result of the abnormal activation of upstream receptor tyrosine kinases or gain-of-function mutations predominantly of the RAS or RAF genes [14,15]. This pathway is thus considered a potential therapeutic target for such cancers [16]. We hypothesize that NF1 mutation alone is not sufficient to induce the continuous activation of the RAS-ERK pathway and that the pathological activation of this pathway promotes the development of NF1-associated malignant tumors.
To identify additional genetic alterations that promote the malignancy of NF1-associated tumors, we established highly malignant NF1-mutated cells and analyzed their genomic DNA using next-generation sequencing. We detected 14 heterozygous variants within the coding regions of 13 cancer-and other disease-related genes, including single-base substitutions in the PTPN11 (c.1508G > T, p.Gly503Val) and BRAP (c.1109G > C, p.Gly370Ala) genes. We examined the effects of these additional mutations on malignant progression in NF1-mutated cells.

Cell Transplantation in Mice
Balb/c nu/nu and NOD/SCID immunodeficient mice were obtained from Charles River. All mouse experiments were approved by the ethics committee of Keio University School of Medicine (approval number 13037), and the animals were treated in compliance with the regulations for animal experiments of Keio University School of Medicine. Mice were anesthetized via the intraperitoneal injection of a combination of medetomidine (0.3 mg/kg), midazolam (4.0 mg/kg), and butorphanol (5.0 mg/kg). For subcutaneous transplantation, cells (5 × 10 6 in 100 µL of phosphate-buffered saline) were injected into the lower flank of 4-week-old female Balb/c nu/nu mice. For kidney subcapsular transplantation, cells (1 × 10 6 in 10 µL of Matrigel (Corning)) were injected below the kidney capsule of 4-week-old female Balb/c nu/nu mice.

Establishment of sNF96.2-GFP Tumor-Derived Cells
sNF96.2-GFP cells (3 × 10 6 ) were mixed with Matrigel and injected below the kidney capsule of immunodeficient (Balb/c nu/nu or NOD/SCID) mice. After 5 to 6 months, the tumor cells were isolated from the developed tumors and cell subclones were established. A-1 and C-1 cells were obtained from Balb/c nu/nu mice at 154 and 176 days, respectively, after cell injection. B-1 and D-1 cells were obtained from NOD/SCID mice at 163 and 191 days, respectively. These cell subclones were maintained in DMEM supplemented with 10% FBS.

STR Profile Analysis
The short tandem repeat (STR) profiles of sNF96.2-GFP and A-1 cells were determined using polymerase chain reaction (PCR) analysis as performed by the Human Cell Line Authentication Service (Promega, Madison, WI, USA).

Genomic DNA Analysis
Genomic DNA of sNF96.2-GFP, A-1, B-1, C-1, and D-1 cells was analyzed using wholeexome sequencing. The sequence reads were aligned to the reference genome assemblies (hg19) with the use of BWA [17]. MuTect version 1.14 [18] was applied for comparison of the exome data derived from the parental sNF96.2-GFP cells with those derived from A-1, B-1, C-1, and D-1 cells. The default parameters were adopted with the exceptions that max_alt_alleles_in_normal_count and minimum_mutation_cell_fraction were set to 0 and 0.1, respectively.

RNA Interference
NF1 and negative control (GAPD) small interfering RNAs (siRNAs) were obtained from Dharmacon and were introduced into cells via transfection with the use of the Lipofectamine RNAiMAX reagent (Invitrogen).

Trypan Blue Staining and Cell Viability Assay
For the evaluation of live cell number, cells (2 × 10 4 per well) were seeded in sixwell plates, cultured for the indicated times, and then stained with trypan blue. For the evaluation of cell viability, cells (1 × 10 3 per well) were seeded in 96-well plates, incubated for 24 h before exposure to trametinib or SHP099 for 72 h, and then assayed for viability with the use of a Cell TiterGlo assay (Promega).

Statistical Analysis
Data are presented as means ± SD and were analyzed with the two-tailed unpaired Student's t test. A p value of <0.05 was considered statistically significant.

Establishment of Tumor-Derived Cells after sNF96.2-GFP Cell Injection into Nude Mice
As we previously reported, sNF96.2 cells generally show low tumorigenic activity in mice. We found that injecting sNF96.2 cells into the subcutaneous tissues or into the sciatic nerves of immunodeficient mice did not result in tumor formation [19]. We injected a GFP-expressing human NF1-MPNST cell line, sNF96.2-GFP, which harbors a frameshift mutation (c.3683delC, p.Asn1229MetfsTer11) in the NF1 gene, under the kidney capsule of an immunodeficient mouse. The developed tumor at 154 days after cell injection manifested as a high-grade, atypical, spindle-shaped cell sarcoma with necrotic areas (Figure 1(Aa-Ac)) and was diagnosed as spindle cell sarcoma, consistent with an MPNST, on the basis of the histological findings. Many mitotic cells were observed in the tumor, with a Ki-67 labeling rate of >60%, indicating that it consisted of highly proliferative tumor cells (Figure 1(Ae)). The tumor cells had spread outside of the kidney and infiltrated skeletal muscle (Figure 1(Ad)). We performed immunostaining for S-100 ( Figure 1(Af)) as well as for Sox10, glial fibrillary acidic protein (GFAP), and CD57 (data not shown), all of which are commonly examined for the diagnosis of human MPNST. About half of all human MPNST cases are negative for these proteins, and the developed tumor was also negative.
We then established cells from the developed tumor and designated them as A-1 cells. To eliminate the possibility of cell contamination, we performed STR analysis. The STR profile of the A-1 cells was identical to that of the parental sNF96.2-GFP cells, and we confirmed that these latter cells were the same as those registered with ATCC (CRL-2884 sNF96.2) by comparison with the JCRB Cell Bank database ( Figure S1). We found that the proliferative activity of A-1 cells in vitro was significantly higher than that of the parental sNF96.2-GFP cells ( Figure 1B). Immunoblot analysis and immunofluorescence analysis also revealed that the phosphorylation (activation) levels of Akt and MEK were much higher in A-1 cells than in sNF96.2-GFP cells ( Figure 1C,D). In addition, A-1 cells were more sensitive to the MEK inhibitor trametinib than sNF96.2-GFP cells were ( Figure 1E).

Identification of Genetic Alterations Related to NF1 Tumor Malignancy
We also found that the tumorigenic activity of A-1 cells was much higher than tha of the parental sNF96.2-GFP cells when the cells were injected subcutaneously into nud mice ( Figure 2A). The tumors formed after the subcutaneous injection of A-1 cell recapitulated the histological features of those formed by sNF96.2-GFP cells in the kidney ( Figure 2B). To identify additional genetic alterations that promote the malignancy o NF1-associated tumors, we analyzed the genomic DNA of sNF96.2-GFP parental cells and four subclones of tumor-derived cells ( Figure S2). Three additional cell subclones (B-1, C

Identification of Genetic Alterations Related to NF1 Tumor Malignancy
We also found that the tumorigenic activity of A-1 cells was much higher than that of the parental sNF96.2-GFP cells when the cells were injected subcutaneously into nude mice ( Figure 2A). The tumors formed after the subcutaneous injection of A-1 cells recapitulated the histological features of those formed by sNF96.2-GFP cells in the kidney ( Figure 2B). To identify additional genetic alterations that promote the malignancy of NF1-associated tumors, we analyzed the genomic DNA of sNF96.2-GFP parental cells and four subclones of tumor-derived cells ( Figure S2). Three additional cell subclones (B-1, C-1, and D-1) were thus established from tumors formed after the injection of sNF96.2-GFP cells below the renal capsule of immunodeficient mice. Next-generation sequencing of the genomic DNA was performed in order to compare the whole-exome sequences of the parental cells and four subclones (A-1, B-1, C-1, and D-1). The NF1 germline mutation was identified as a biallelic base pair deletion in exon 21 (3683delC) that results in a frameshift and the generation of a premature stop codon before the RAS-GAP domain of the encoded protein. In addition, the tumor-derived cells showed the somatic loss of heterozygosity on chromosome 17. The MuTect program revealed that, whereas each of the four subclones harbored unique genetic alterations, all four subclones shared mutations in 42 genes that were not present in the parental cells ( Figure S3). Among these 42 genes, 13 genes are related to cancer and other diseases (Table 1). Of note, three subclones established from secondary tumors formed after A-1 cell injection ( Figure S2) did not harbor any common alterations other than these 13 gene variants. We therefore speculated that these 13 gene variants might be associated with the malignant progression of NF1-associated tumors. Although the NF1 gene encodes a RAS-GAP, NF1 mutation appears to not be sufficient to induce the persistent activation of the RAS-ERK pathway. Given that the phosphorylation levels of Akt and MEK were increased in the tumor-derived A-1 cells, we focused on the relation of additional mutations that might affect this pathway to tumor malignancy. Table 1. Identification of 14 heterozygous mutations of 13 cancer-and other disease-related genes commonly present in tumor-derived subclones (A-1, B-1, C-1, and D-1). Chrom., chromosome.

Gene Name
Chrom. Among these 13 genes, we have a special interest in the BRAP and PTPN11 mutations because these are the regulators that control the activity of the RAS-ERK pathway. We identified missense mutations of BRAP (c.1109G > C, p.Gly370Ala) and PTPN11 (c.1508G > T, p.Gly503Val) in the tumor-derived subclones but not in the parental sNF96.2-GFP cells ( Figures 2C,D and S4A,B). Quantitative RT-PCR analysis showed that the abundance of BRAP and PTPN11 mRNAs did not differ substantially between sNF96.2-GFP and A-1 cells ( Figure S4C,D).

Role of Missense Mutations of BRAP and PTPN11 in Tumorigenic Activity
To determine the role of these mutations of BRAP and PTPN11 in the malignant progression of NF1-associated tumors, we established three sNF96.2 cell lines that stably overexpress BRAP (G370A), SHP-2 (G503V), or both BRAP (G370A) and SHP-2 (G503V). We injected these cells as well as the parental sNF96.2-GFP cells either below the kidney capsule or subcutaneously in nude mice. The tumors formed below the kidney capsule by the SHP-2 (G503V)-expressing cells were much larger than those formed by the parental cells ( Figure 3A,B). Only parental and SHP-2 (G503V)-expressing cells formed

Role of Missense Mutations of BRAP and PTPN11 in Tumorigenic Activity
To determine the role of these mutations of BRAP and PTPN11 in the malignant progression of NF1-associated tumors, we established three sNF96.2 cell lines that stably overexpress BRAP (G370A), SHP-2 (G503V), or both BRAP (G370A) and SHP-2 (G503V). We injected these cells as well as the parental sNF96.2-GFP cells either below the kidney capsule or subcutaneously in nude mice. The tumors formed below the kidney capsule by the SHP-2 (G503V)-expressing cells were much larger than those formed by the parental  Figure 3A,B). Only parental and SHP-2 (G503V)-expressing cells formed subcutaneous tumors, with formation rates of 16.7% (1/6) and 100% (6/6), respectively ( Figure 3C). The subcutaneous tumors formed by SHP-2 (G503V)-expressing cells manifested a higher density of tumor cells ( Figure 3D) and were larger ( Figure 3E) compared with those formed by the parental cells, with the mean tumor volume at 10 weeks after cell injection being 75.89 versus 1.69 mm 3 . The higher tumorigenic activity of SHP-2 (G503V)-expressing cells relative to parental sNF96.2-GFP cells after both kidney subcapsular and subcutaneous injection thus suggested that additional mutation of PTPN11 is able to enhance the malignant characteristics of NF1-associated tumors. In contrast, cells expressing BRAP (G370A) or both BRAP (G370A) and SHP-2 (G503V) did not form tumors in subcutaneous tissue ( Figure 3C,E), and the tumors formed below the kidney capsule by these cells were of a similar size to those formed by parental cells (Figure 3A,B), suggesting that BRAP (G370A) suppresses SHP-2 (G503V)-mediated tumorigenic activity.
Cancers 2022, 14, x 9 of 17 subcutaneous tumors, with formation rates of 16.7% (1/6) and 100% (6/6), respectively ( Figure 3C). The subcutaneous tumors formed by SHP-2 (G503V)-expressing cells manifested a higher density of tumor cells ( Figure 3D) and were larger ( Figure 3E) compared with those formed by the parental cells, with the mean tumor volume at 10 weeks after cell injection being 75.89 versus 1.69 mm 3 . The higher tumorigenic activity of SHP-2 (G503V)-expressing cells relative to parental sNF96.2-GFP cells after both kidney subcapsular and subcutaneous injection thus suggested that additional mutation of PTPN11 is able to enhance the malignant characteristics of NF1-associated tumors. In contrast, cells expressing BRAP (G370A) or both BRAP (G370A) and SHP-2 (G503V) did not form tumors in subcutaneous tissue ( Figure 3C,E), and the tumors formed below the kidney capsule by these cells were of a similar size to those formed by parental cells (Figure 3A,B), suggesting that BRAP (G370A) suppresses SHP-2 (G503V)-mediated tumorigenic activity.

Effects of Missense Mutations in BRAP and PTPN11 on MEK Phosphorylation
Immunoblot analysis revealed that, as in A-1 cells, the phosphorylation level of MEK was increased in SHP-2 (G503V)-expressing cells as well as in cells derived from a tumor formed by these cells (SHP-2 (G503V)-3 cells), whereas MEK phosphorylation in BRAP (G370A)-or SHP-2 (G503V)/BRAP (G370A)-expressing cells did not differ from that in parental cells ( Figure 4A). The phosphorylation level of Akt was also increased in the SHP-2 (G503V) tumor-derived cells, as it was in A-1 cells ( Figure 4A). Consistent with these results, SHP-2 (G503V)-expressing cells showed a greater sensitivity to trametinib than SHP-2 (G503V)/BRAP (G370A)-expressing cells did ( Figure 4B). We next established sNF96.2 cells that overexpress wild-type (WT) SHP-2 and compared them with those expressing SHP-2 (G503V). Immunoblot analysis revealed that, whereas MEK phosphorylation was increased in SHP-2 (G503V)-expressing cells, it was unaffected in the SHP-2 (WT)-expressing cells ( Figure 4C). We also performed a cytotoxicity assay with the SHP-2 inhibitor SHP099 [23] and found that the sensitivity of SHP-2 (WT)-expressing cells to this agent was similar to that of parental cells, whereas SHP-2 (G503V)-expressing cells as well as tumor-derived A-1 cells were less sensitive to SHP099 ( Figure 4D).
Cancers 2022, 14, x 10 of 17 phosphorylation was increased in SHP-2 (G503V)-expressing cells, it was unaffected in the SHP-2 (WT)-expressing cells ( Figure 4C). We also performed a cytotoxicity assay with the SHP-2 inhibitor SHP099 [23] and found that the sensitivity of SHP-2 (WT)-expressing cells to this agent was similar to that of parental cells, whereas SHP-2 (G503V)-expressing cells as well as tumor-derived A-1 cells were less sensitive to SHP099 ( Figure 4D).

Mechanism of MEK Inhibition by BRAP Missense Mutation
The identification of the mechanism by which BRAP (G370A) suppresses SHP-2 (G503V)-mediated tumorigenic activity may inform the development of new therapeutic approaches to the inhibition of the RAS-ERK pathway. Given that a lower-molecularweight protein reactive with antibodies to SHP-2 was detected when BRAP (G370A) was coexpressed with SHP-2 (G503V) (Figure 4A), we considered the possibility that SHP-2 (G503V) might undergo proteolytic cleavage in a manner dependent on BRAP (G370A). SHP-2 has previously been shown to be cleaved in a calpain-dependent manner [24], and we found that the calpain inhibitor calpeptin attenuated the generation of the immunoreactive lower-molecular-weight protein in a concentration-dependent manner in cells expressing both SHP-2 (G503V) and BRAP (G370A) ( Figure 5A).

Mechanism of MEK Inhibition by BRAP Missense Mutation
The identification of the mechanism by which BRAP (G370A) suppresses SHP-2 (G503V)-mediated tumorigenic activity may inform the development of new therapeutic approaches to the inhibition of the RAS-ERK pathway. Given that a lower-molecularweight protein reactive with antibodies to SHP-2 was detected when BRAP (G370A) was coexpressed with SHP-2 (G503V) (Figure 4A), we considered the possibility that SHP-2 (G503V) might undergo proteolytic cleavage in a manner dependent on BRAP (G370A). SHP-2 has previously been shown to be cleaved in a calpain-dependent manner [24], and we found that the calpain inhibitor calpeptin attenuated the generation of the immunoreactive lower-molecular-weight protein in a concentration-dependent manner in cells expressing both SHP-2 (G503V) and BRAP (G370A) ( Figure 5A).
We next established sNF96.2 cells that stably express the WT form of BRAP in order to clarify the relevance of the G370A variant. Immunoblot analysis revealed that both the cleavage of SHP-2 and inhibition of MEK phosphorylation apparent in SHP-2 (G503V)expressing cells that also express BRAP (G370A) were much less pronounced in those expressing BRAP (WT) ( Figure 5B). These data suggested that the mutant form of BRAP mediates SHP-2 (G503V) cleavage and thereby downregulates MEK phosphorylation.   We next established sNF96.2 cells that stably express the WT form of BRAP in order to clarify the relevance of the G370A variant. Immunoblot analysis revealed that both the cleavage of SHP-2 and inhibition of MEK phosphorylation apparent in SHP-2 (G503V)expressing cells that also express BRAP (G370A) were much less pronounced in those expressing BRAP (WT) ( Figure 5B). These data suggested that the mutant form of BRAP mediates SHP-2 (G503V) cleavage and thereby downregulates MEK phosphorylation.
Finally, we examined the effects of the knockdown of neurofibromin by siRNA transfection in HeLa cells stably expressing SHP-2 (G503V) or both SHP-2 (G503V) and BRAP (G370A) ( Figure 6B). In the SHP-2 (G503V)-expressing cells, the knockdown of neurofibromin increased the phosphorylation of MEK and ERK, but not that of Akt, indicating that MEK-ERK signaling was preferentially activated. The activation of MEK by SHP-2 (G503V) was thus also confirmed in the neurofibromin-depleted HeLa cells. The knockdown of neurofibromin tended to increase the phosphorylation of MEK and ERK in HeLa cells expressing both BRAP (G370A) and SHP-2 (G503V). The comparison of neurofibromin-depleted SHP-2 (G503V)-expressing cells and neurofibromin-depleted SHP-2 (G503V)/BRAP (G370A)-expressing cells revealed that the phosphorylation of Akt was increased and that of MEK and ERK was slightly decreased in the latter. The cleavage of SHP-2 was not observed in HeLa cells, however. Finally, we examined the effects of the knockdown of neurofibromin by siRNA transfection in HeLa cells stably expressing SHP-2 (G503V) or both SHP-2 (G503V) and BRAP (G370A) ( Figure 6B). In the SHP-2 (G503V)-expressing cells, the knockdown of neurofibromin increased the phosphorylation of MEK and ERK, but not that of Akt, indicating that MEK-ERK signaling was preferentially activated. The activation of MEK by SHP-2 (G503V) was thus also confirmed in the neurofibromin-depleted HeLa cells. The knockdown of neurofibromin tended to increase the phosphorylation of MEK and ERK in HeLa cells expressing both BRAP (G370A) and SHP-2 (G503V). The comparison of neurofibromin-depleted SHP-2 (G503V)-expressing cells and neurofibromin-depleted SHP-2 (G503V)/BRAP (G370A)-expressing cells revealed that the phosphorylation of Akt was increased and that of MEK and ERK was slightly decreased in the latter. The cleavage of SHP-2 was not observed in HeLa cells, however.

Discussion
Genomic DNA sequencing allowed us to identify additional genetic alterations including mutations in the PTPN11 (encoding SHP-2) and BRAP genes that were associated with enhanced NF1-related tumor malignancy. BRAP negatively regulates ERK activation by inactivating the KSR1 scaffold protein and limiting the formation of the RAF-MEK complex [25][26][27]. Upon the activation of RAS, BRAP binds to RAS-GTP and undergoes autoubiquitylation, resulting in the stabilization of the KSR1-RAF-MEK complex and the activation of MEK-ERK signaling [26]. SHP-2 is a member of the PTP family and regulates SRC family kinase activity and RAS-ERK activation [28]. We found that the forced expression of the mutant form of PTPN11 (c.1508G > T, p.Gly503Val) promoted MEK activation and tumorigenic activity in neurofibromin-deficient cells, whereas these effects were attenuated by the coexpression of the mutant form of BRAP (c.1109G > C, p.Gly370Ala). Moreover, the suppressive effect of the BRAP (G370A) protein on MEK phosphorylation was not observed in neurofibromin-intact cells. Our data indicate that PTPN11 mutation drives malignancy in NF1-mutated cells and that the inhibition of SHP-2 by BRAP may provide the basis for a new therapeutic strategy for malignant NF1-associated tumors.
Evidence from transgenic mouse models has implicated dysfunction of the proteins p53, Rb, and p16 in the malignant progression of NF1-associated tumors [29]. In addition, somatic mutations of SUZ12 and EED which encode core components of the Polycomb repressive complex 2 (PRC2) have been frequently identified in human NF1-associated MPNSTs, indicating that mutations of these genes play a role in the establishment of these tumors [30][31][32]. Our whole-exome sequence analysis revealed that the parental sNF96.2-GFP cells examined in this study harbor a homozygous missense mutation in EED (p.S241R). Zebrafish models have also shown that the overexpression of PDGFRA cooperates with the loss of NF1 and the p53 gene to promote MPNST formation [33]. Furthermore, the activation of the PI3K (phosphatidylinositol 3-kinase)-Akt-mTOR (mammalian target of rapamycin) signaling pathway caused by the deletion of the PTEN gene has been shown to induce the malignant transformation of NF1-associated tumors [34]. We found that the phosphorylation of Akt was increased in tumor-derived sNF96.2 cells (A-1 cells), suggesting that Akt was activated in all tumor cells or that tumor cells with activated Akt were selected in vivo. We also found that the phosphorylation level of MEK was increased in tumorderived cells as well as in sNF96.2 cells stably expressing SHP-2 (G503V). Both the RAS-ERK and PI3K-Akt-mTOR signaling pathways thus appear to play key roles in the malignant transformation of neurofibroma into MPNST.
MEK-ERK signaling contributes to the regulation of cellular functions such as proliferation, differentiation, and apoptosis as well as to that of normal tissue and tumor formation [35]. Molecular and pharmacological studies have shown that MEK and ERK are required for the transforming activities of RAS and various other oncogenes, and preclinical data support an important role for the RAS-ERK pathway in cancer biology and its potential as a therapeutic target in human cancers [36]. The inhibition of MEK-ERK signaling was found to restore colon tumor cells to a nontransformed state in vitro and to inhibit tumor growth in vivo [37], suggesting that increased activity of such signaling is related to tumor development. We found that ERK was constitutively phosphorylated and that the expression of SHP-2 (G503V) increased the level of MEK phosphorylation in NF1-mutated cells, indicating that the activity of the MEK-ERK pathway was increased further by the pathogenic mutation of PTPN11 in these cells. Such enhanced activation of MEK-ERK signaling may promote the transformation of benign neurofibroma cells into malignant tumor cells.
The PTPN11 missense mutation identified in the present study induced pathological MEK activation and promoted tumor malignancy. Constitutional gain-of-function mutations of PTPN11 cause Noonan syndrome in humans as a result of the activation of RAS signaling [38]. PTPN11 mutations have been detected in human cancers including leukemias and solid tumors, with the G503V missense mutation being associated with childhood leukemia [39] and adult acute myelogenous leukemia [40]. Although the 13 genes that we identified were not on the list of genes identified using a forward genetic screen in Schwann cells utilizing the Sleeping Beauty (SB) transposon-based somatic mutagenesis system in mice to determine the driver genes of MPNST [41], PTPN11 missense mutation and NF1 homozygous loss were recently detected in clinical specimens of MPNSTs [42], suggesting that the combination of PTPN11 and NF1 mutations may give rise to the persistent activation of both RAS-ERK and PI3K-Akt-mTOR pathways and to malignant transformation. A small-molecule allosteric inhibitor of SHP-2 (RMC-4550) was found to inhibit the activity of oncogenic RAS-ERK signaling and cancer growth in models of human cancer harboring mutated BRAF, NF1 loss, or the G12C mutation of KRAS [43], suggesting that the inhibition of SHP-2 is a promising molecular therapeutic strategy for NF1-associated tumors. The combination of a MEK inhibitor plus the SHP-2 inhibitor SHP099 was recently shown to be effective in models of NF1-MPNST and plexiform neurofibroma [44]. However, our data show that SHP-2 (G503V)-expressing sNF96.2 cells were less sensitive to SHP099 than cells overexpressing SHP-2 (WT) were.
We identified 14 heterozygous mutations within the coding regions of 13 genes-AMER1, BCAT1, BRAP, C3, CNTNAP2, COL6A6, NOS3, PCNT, PTPN11, SCNN1B, SLC5A11, SYNE1, and WIPF1-that were commonly present in highly malignant NF1-mutated cells. We speculate that the cell subclones with mutations taking advantages for proliferation and survival in vivo were selected and grew in mice. Given that the phosphorylation of Akt and MEK was increased in the malignant A-1 cells, among these 13 genes, we focused on those (BRAP and PTPN11) related to the RAS-ERK pathway. Our results suggest that the BRAP mutation might have developed as a negative feedback response to the PTPN11 mutation. Whereas the overexpression of BRAP (G370A) strongly suppressed MEK activity mediated by the overexpression of SHP-2 (G503V), in A-1 cells, BRAP (G370A) may adjust MEK activity so as to stabilize cell survival and proliferation, suggesting that the expression levels may be related to the suppressive effect of mutant BRAP. The BRAP mutation may thus serve to modulate cell survival and proliferation signals in order to optimize the growth of malignant NF1-mutated cells. Further characterization of the other gene alterations found to be associated with malignant transformation in our study may provide the basis for a better understanding of NF1-MPNST.
The cleavage of SHP-2 (G503V) and suppression of MEK phosphorylation mediated by BRAP (G370A) were apparent in NF1-mutated (sNF96.2) cells but not in NF1-intact (HeLa) cells, suggesting that the phenotypes induced by BRAP and PTPN11 missense mutations are related to NF1 loss. Whereas the marked activation of MEK by SHP-2 (G503V) was detected in neurofibromin-depleted HeLa cells, the cleavage of SHP-2 was not, suggesting that the attenuation of MEK phosphorylation and cleavage of SHP-2 (G503V) dependent on BRAP (G370A) may only occur in cells in which NF1 is chronically inactivated. Different downstream signals may be activated depending on whether NF1 expression is acutely or chronically deficient. Indeed, sNF96.2 cells showed a high level of ERK phosphorylation, whereas neurofibromin depletion increased ERK phosphorylation without affecting Akt phosphorylation in SHP-2 (G503V)-expressing HeLa cells. Furthermore, the knockdown of neurofibromin tended to increase the phosphorylation of MEK and ERK in SHP-2 (G503V)/BRAP (G370A)-expressing HeLa cells. Our findings indicate that tumor progression induced by SHP-2 (G503V) and its suppression by BRAP (G370A) may provide the basis for the development of new strategies for the treatment of NF1-MPNST as well as for the prevention of the malignant transformation of neurofibroma in individuals with NF1.

Conclusions
Here, we showed that MEK phosphorylation was increased in a highly malignant subclone of NF1-MPNST cells. Genomic DNA sequencing identified missense mutations of PTPN11 (encoding SHP-2) and BRAP in these cells. We also found that the forced expression of the SHP-2 mutant promoted MEK activation and malignancy in the parental NF1-MPNST cells, and that these effects of the SHP-2 mutant were attenuated by the forced expression of the BRAP mutant. The combination of PTPN11 and NF1 mutations may thus give rise to malignant transformation in NF1 cells, and SHP-2 inhibition by BRAP is a potential therapeutic strategy for NF1-MPNST.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cancers14102377/s1. Figure S1: Confirmation of the identity of A-1 cells with parental sNF96.2-GFP cells via STR analysis. Figure S2: Experimental scheme for genome DNA analysis of tumor-derived cells. sNF96.2-GFP cells were injected below the renal capsule of immunodeficient mice, and tumor-derived cells were established from each mouse. Figure S3: Identification of 42 genes commonly harboring mutations in tumor-derived cells (A-1, B-1, C-1, and D-1 cells). Figure S4: Alignment of the nucleotide sequences of BRAP (A) and PTPN11 (B) for sNF96.2-GFP (P-G) and A-1 cells as performed with ClustalW (base differences are boxed in red) as well as quantitative RT-PCR analysis of BRAP (C) and PTPN11 (D) mRNA abundance in these cells (data are means ± SD from three independent experiments). Figure   Institutional Review Board Statement: The animal study protocol was approved by the ethics committee of Keio University School of Medicine (approval number 13037).

Informed Consent Statement: Not applicable.
Data Availability Statement: Representative data are provided in this published article and its Supplementary Material Files. Other data sets generated or analyzed during the current study are available from the corresponding author upon reasonable request.