Advertisement
Journal of Investigative Dermatology Home

Combined Cyclin-Dependent Kinase Inhibition Overcomes MAPK/Extracellular Signal–Regulated Kinase Kinase Inhibitor Resistance in Plexiform Neurofibroma of Neurofibromatosis Type I

  • Author Footnotes
    3 These authors contributed equally to this work.
    Wei Wang
    Footnotes
    3 These authors contributed equally to this work.
    Affiliations
    Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
    Search for articles by this author
  • Author Footnotes
    3 These authors contributed equally to this work.
    Xi-Wei Cui
    Footnotes
    3 These authors contributed equally to this work.
    Affiliations
    Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
    Search for articles by this author
  • Author Footnotes
    3 These authors contributed equally to this work.
    Yi-Hui Gu
    Footnotes
    3 These authors contributed equally to this work.
    Affiliations
    Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
    Search for articles by this author
  • Cheng-Jiang Wei
    Affiliations
    Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
    Search for articles by this author
  • Yue-Hua Li
    Affiliations
    Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
    Search for articles by this author
  • Jie-Yi Ren
    Affiliations
    Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
    Search for articles by this author
  • Man-Hon Chung
    Affiliations
    Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
    Search for articles by this author
  • Re-han-gu-li Aimaier
    Affiliations
    Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
    Search for articles by this author
  • Hai-Bing Zhang
    Affiliations
    CAS Key Laboratory of Nutrition, Metabolism and Food Safety, Shanghai Institute of Nutrition and Health, Chinese Academy of Sciences, Shanghai, China
    Search for articles by this author
  • Author Footnotes
    4 These authors contributed equally to this work.
    Qing-Feng Li
    Footnotes
    4 These authors contributed equally to this work.
    Affiliations
    Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
    Search for articles by this author
  • Author Footnotes
    4 These authors contributed equally to this work.
    Zhi-Chao Wang
    Correspondence
    Correspondence: Zhi-Chao Wang, Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, 639 Zhizaoju Road, Shanghai 200011, China.
    Footnotes
    4 These authors contributed equally to this work.
    Affiliations
    Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
    Search for articles by this author
  • Author Footnotes
    3 These authors contributed equally to this work.
    4 These authors contributed equally to this work.
Open AccessPublished:September 14, 2021DOI:https://doi.org/10.1016/j.jid.2021.07.164
      MAPK/extracellular signal–regulated kinase kinase (MEK) 1/2 inhibitors (MEKis) have recently achieved surprising success in treating unresectable plexiform neurofibromas (PNFs). However, few studies have investigated the mechanisms of MEKi resistance in patients with PNF. We determined the efficacy of six different MEKis for treating PNFs, explored drug resistance mechanisms, and identified potential combination therapies to overcome resistance. By screening drug efficacy among six MEKis in human NF1-deficient PNF cell lines, TAK-733 was found to reduce PNF cell viability the most. We then cultured the TAK-733‒resistant cells and explored the potential targets for further treatment. Both high-throughput drug screening and RNA sequencing analyses of MEKi-resistant PNF cells identified cyclin-dependent kinase inhibitors as potential agents for PNFs. Dinaciclib, a cyclin-dependent kinase inhibitor, showed synergistic effects on MEKi-resistant cells. Coadministration of dinaciclib and TAK-733 significantly reduced cell viability and inhibited sphere formation and colony formation. Dinaciclib did not affect MEK signaling but decreased the expression of several prosurvival proteins, including survivin and cyclin-dependent kinase 1, to induce apoptosis and inhibit mitosis. TAK-733/dinaciclib combination therapy induced tumor reduction in PNF patient‒derived xenografts mouse models. Therefore, the combination of MEKi and cyclin-dependent kinase inhibitor may be promising for treating inoperable PNFs, especially when drug resistance exists. Our findings provide evidence for future clinical trials with MEKi-resistant patients with PNF.

      Abbreviations:

      CDK (cyclin-dependent kinase), CDKi (CDK inhibitor), ERK (extracellular signal–regulated kinase), MEK (MAPK/extracellular signal–regulated kinase kinase), MEKi (MAPK/extracellular signal–regulated kinase kinase 1/2 inhibitor), NF1 (neurofibromatosis type 1), p-ERK (phosphorylated extracellular signal–regulated kinase), p-MEK (phosphorylated MAPK/extracellular signal–regulated kinase kinase), PNF (plexiform neurofibroma), PNF-PDX (plexiform neurofibroma patient‒derived xenograft), RNA-seq (RNA sequencing)

      Introduction

      Neurofibromatosis type 1 (NF1) is caused by mutations in the NF1 gene encoding neurofibromin and affects approximately 1 in 3,000–3,500 live births worldwide (
      • Le L.Q.
      • Parada L.F.
      Tumor microenvironment and neurofibromatosis type I: connecting the GAPs.
      ). Overall, 20–50% of patients with NF1 present with plexiform neurofibromas (PNFs) (
      • Gross A.M.
      • Singh G.
      • Akshintala S.
      • Baldwin A.
      • Dombi E.
      • Ukwuani S.
      • et al.
      Association of plexiform neurofibroma volume changes and development of clinical morbidities in neurofibromatosis 1.
      ). Patients with PNFs suffer from pain, disfigurement, and dysfunction of the affected cranial or large peripheral nerves with disease progression (
      • Dombi E.
      • Solomon J.
      • Gillespie A.J.
      • Fox E.
      • Balis F.M.
      • Patronas N.
      • et al.
      NF1 plexiform neurofibroma growth rate by volumetric MRI: relationship to age and body weight.
      ;
      • Gross A.M.
      • Singh G.
      • Akshintala S.
      • Baldwin A.
      • Dombi E.
      • Ukwuani S.
      • et al.
      Association of plexiform neurofibroma volume changes and development of clinical morbidities in neurofibromatosis 1.
      ;
      • Prada C.E.
      • Rangwala F.A.
      • Martin L.J.
      • Lovell A.M.
      • Saal H.M.
      • Schorry E.K.
      • et al.
      Pediatric plexiform neurofibromas: impact on morbidity and mortality in neurofibromatosis type 1.
      ).
      MAPK/extracellular signal–regulated kinase (ERK) kinase (MEK) 1/2 inhibitors (MEKis) have demonstrated promising results in inoperable pediatric patients with PNF with a significant reduction in PNF volume and functional improvements. A phase I trial of selumetinib as a treatment for 24 inoperable pediatric patients with PNF showed a positive response in >70% of patients, with >20% decrease in tumor volume (
      • Dombi E.
      • Baldwin A.
      • Marcus L.J.
      • Fisher M.J.
      • Weiss B.
      • Kim A.
      • et al.
      Activity of selumetinib in neurofibromatosis type 1-related plexiform neurofibromas.
      ). This is consistent with the result of the phase 2 trial (
      • Gross A.M.
      • Wolters P.L.
      • Dombi E.
      • Baldwin A.
      • Whitcomb P.
      • Fisher M.J.
      • et al.
      Selumetinib in children with inoperable plexiform neurofibromas [published correction appears in N Engl J Med 2020;383:1290].
      ). Several other MEKis, including trametinib, cobimetinib, and TAK-733, also exhibit antitumor effects by interrupting growth signals for tumor cell proliferation and are currently in clinical trials (
      • Adjei A.A.
      • LoRusso P.
      • Ribas A.
      • Sosman J.A.
      • Pavlick A.
      • Dy G.K.
      • et al.
      A phase I dose-escalation study of TAK-733, an investigational oral MEK inhibitor, in patients with advanced solid tumors.
      ;
      • Perreault S.
      • Larouche V.
      • Tabori U.
      • Hawkin C.
      • Lippé S.
      • Ellezam B.
      • et al.
      A phase 2 study of trametinib for patients with pediatric glioma or plexiform neurofibroma with refractory tumor and activation of the MAPK/ERK pathway: TRAM-01.
      ;
      • Schadendorf D.
      • Fisher D.E.
      • Garbe C.
      • Gershenwald J.E.
      • Grob J.J.
      • Halpern A.
      • et al.
      ).
      Drug resistance can be observed in benign tumors such as pituitary adenomas (
      • Stark R.
      • Grzelak M.
      • Hadfield J.
      RNA sequencing: the teenage years.
      ). Because they grow slowly and persistently, acquired resistance may occur after long-term use. Acquired MEKi resistance was observed in phase 2 trial of selumetinib (
      • Gross A.M.
      • Wolters P.L.
      • Dombi E.
      • Baldwin A.
      • Whitcomb P.
      • Fisher M.J.
      • et al.
      Selumetinib in children with inoperable plexiform neurofibromas [published correction appears in N Engl J Med 2020;383:1290].
      ). Resistance to MEKi can also be found in various types of cancers, including malignant peripheral nerve sheath tumor, the malignant transformation of PNFs, and other tumors such as melanoma or colorectal cancer (
      • Martinelli E.
      • Morgillo F.
      • Troiani T.
      • Ciardiello F.
      Cancer resistance to therapies against the EGFR-RAS-RAF pathway: the role of MEK.
      ;
      • Samatar A.A.
      • Poulikakos P.I.
      Targeting RAS-ERK signalling in cancer: promises and challenges.
      ;
      • Wang J.
      • Pollard K.
      • Calizo A.
      • Pratilas C.A.
      Activation of receptor tyrosine kinases mediates acquired resistance to MEK inhibition in malignant peripheral nerve sheath tumors.
      ). Investigating the underlying mechanisms of drug resistance is vital for designing combination therapies. There are currently few studies investigating the resistance mechanisms of MEKi in PNFs.
      In this study, we first evaluated phosphorylated ERK (p-ERK) and phosphorylated MEK (p-MEK) expression in PNFs and its association with clinicopathological parameters. The efficiency of MEKi was tested in PNF models both in vitro and in vivo. We also employed a high-throughput drug screening system to identify potential combination agents, especially for patients in whom drug resistance may present. To further validate our findings in clinical specimens, we compared the feasibility and efficacy of combination therapies with those of MEKi monotherapy using a PNF patient‒derived xenograft (PNF-PDX) mouse model. This study provides evidence for a therapeutic option in the emergence of MEKi resistance in patients with PNF and informs the design of future clinical trials with patients with PNF.

      Results

       p-MEK/p-ERK expression heterogeneity was observed in PNF tissues and cell lines

      To evaluate cancer molecular heterogeneity, we employed a PNF tissue microarray containing two tissue cores per tumor. Figure 1a displays the different expression intensities of different markers, including p-MEK, p-ERK, MEK, and ERK, in the PNF tissue microarray. The expression of these markers was assessed as a score ranging from 0 to 6 on the basis of the intensity and area of positive staining with different proportions. A total of 94 of 141 (66.66%) of PNF tumor areas showed a relatively strong (staining scores ranging from 3 to 6) p-MEK expression (Figure 1b). More information about the relationships between the expression of p-MEK/p-ERK and the clinical characteristics, including sex, tumor sites, and size, is presented in Supplementary Tables S1 and S2. The expression of p-ERK was moderately correlated with that of p-MEK (r = 0.4601, P < 0.001) (Figure 1c), so as the expression of ERK and MEK (Figure 1d). The expression levels of p-MEK and p-ERK in tumor tissue derived from different individuals were compared to assess tumoral heterogeneity (Figure 1e). Moreover, the p-MEK and p-ERK protein levels in different clinical PNF tumor tissues (Figure 1f) and cell lines (Figure 1g) were also heterogeneous, which may provide an explanation for the differences in treatment efficacy.
      Figure thumbnail gr1
      Figure 1PNF tumor tissues and cell lines presented tumor heterogeneity. (a, b, e) IHC results showed the heterogeneous p-MEK/p-ERK and MEK/ERK expression in intensity in the PNF tissue microarray. (c, d) Correlations between MEK/ERK and p-MEK/p-ERK expression were presented. Tumor heterogeneity was further confirmed by western blotting of (f) PNF biopsy tissue and (g) cell lines. Bar = 100 μm. ERK, extracellular signal–regulated kinase; IHC, immunohistochemistry; MEK, MAPK/extracellular signal–regulated kinase kinase; p-ERK, phosphorylated extracellular signal–regulated kinase; p-MEK, phosphorylated MAPK/extracellular signal–regulated kinase kinase; PNF, plexiform neurofibroma.

       PNF cell lines but not normal Schwann cells were responsive to MEKi

      To evaluate how PNF cells respond to different MEKi in vitro, we screened the effectiveness of selumetinib, trametinib, PD0325901, TAK-733, cobimetinib, and refametinib (all purchased from MedChemExpress, Monmouth Junction, NJ) in three NF1-deficient PNF cell lines (ipNF05.5, ipNF9511.bc, and ipNF95.6) and one Schwann cell line (Ipn02.3 2λ) (all purchased from ATCC, Manassas, VA) as a control after 72 hours of incubation. The MEKis TAK-733, trametinib, and cobimetinib were determined to be more effective for inhibiting PNF cell viability than the other three inhibitors with <5 μM half-maximal inhibitory concentration values (Figure 2a and b). In addition, they exhibited a more sensitive trend to NF1-deficient PNFs, with a lower inhibitory concentration level, than the normal Schwann cell control. We then chose trametinib and TAK-733 for further evaluation in vitro. Treatment with high concentrations induced morphological changes in the cells, resulting in a poached egg shape (Figure 2c). In addition, apoptotic rate was significantly increased in both 10 μM and 1 μM concentrations of TAK-733 (P < 0.05) (Figure 2d and f). Cell cycle assays displayed the blockade of S phase (Figure 2e and g). TUNEL assays additionally proved that MEKis induced apoptosis in PNF cells, and this inhibition was stronger when the concentration of MEKis increased (Figure 2h and i). Inhibition of colony formation and growth (as evidenced by fewer colonies in the trametinib-treated groups than in the control group) were also observed in MEKi-treated cells (Figure 2j and k), which was in line with the decreased ratio of living cells in Cell Counting Kit-8 assay (Dojindo Laboratories, Kumamoto, Japan).
      Figure thumbnail gr2
      Figure 2PNF cell lines were responsive to MEK1/2 inhibitors. The six MEK1/2 inhibitors reduced the viability of NF1-deficient PNF cancer cell lines and Schwann cells after treatment for 72 h. (a, b) The mean viability is plotted with IC50 values. (c) ipNF9511.bc cells incubated in DMEM or 1 μmol/l TAK-733 for 48 h and 72 h were observed and photographed. (d, f) TAK-733 increased cell apoptosis, (e, g) the proportion of cells in the G0/G1 phase of the cell cycle. (h, i) TUNEL assays and (j, k) colony formation assays were performed for the evaluation of cell apoptosis and colony formation ability. ∗P < 0.05 and ∗∗P < 0.01 referred to control condition. Bar = 100 μm. h, hour; IC50, half-maximal inhibitory concentration; MEK, MAPK/extracellular signal–regulated kinase kinase; n.s., not significant; PI, propidium iodide; PNF, plexiform neurofibroma; TRA, trametinib.

       MEKi-related proliferation inhibition and apoptosis induction occurred in a time- and dose-dependent manner

      The therapeutic effects of TAK-733 and trametinib were further evaluated in vitro, and the expression levels of MEK/ERK signaling pathway proteins in the PNF cell line ipNF9511.bc are analyzed in Supplementary Figure S1. The results showed that TAK-733 was more effective in reducing the level of p-ERK than trametinib because its expression could not be detected in ipNF9511.bc cells treated with 0.003 μM TAK-733 after both 4- and 24-hour incubations. In contrast, the reduction in p-MEK levels induced by trametinib remained for less than 24 hours when the concentration was lower than 0.01 μM. The p-MEK reduction observed with a higher dose of trametinib treatment (0.03–0.1 μM) was equivalent to that seen with 0.003 μM TAK-733, and this result indicates that TAK-733 induces strong inhibition of p-MEK expression than trametinib. Similarly, TAK-733 also induced potent decreases in the levels of p-ERK, a member of the MEK/ERK signaling pathway. Moreover, a significant decrease in the expression of cyclin D1 was observed with a lower concentration of TAK-733, and this pattern was not seen with trametinib. This result was in line with the results from the cell cycle assays regarding the S phase blockade induced in TAK-733‒treated PNF cells (Figure 2e and g).

       TAK-733‒resistant cells exhibited increased survival abilities compared with parental cell lines

      To address the mechanism of drug resistance after MEKi monotherapy and to further evaluate the potential synergistic effects in reducing tumor growth, we established drug-resistant PNF cell lines from parental ipNF05.5 and ipNF9511.bc cells by chronic treatment with low-concentration TAK-733. Drug resistance to TAK-733 was obtained, and both ipNF05.5-resistant and ipNF9511.bc-resistant cells exhibited significantly higher viability than parental cells according to Cell Counting Kit-8 assay (Figure 3a). The results of cell scratch migration showed a 10% increment in tumor cell migration ratio after the induction of resistance (Figure 3b). In the long-term colony formation assay, we observed that resistant PNF cells can grow and form more colonies under DMSO and low concentrations (0.1, 0.01 μM) of TAK-733 (Figure 3c). Similarly, as shown in the flow cytometry analysis, the ratios of apoptotic cells were significantly higher than those of control cells (Figure 3d and g). More resistance-induced PNF cells were observed in the S phase of the cell cycle after TAK-733 treatment (Figure 3e and h). As for the sphere-forming assay, the size of spheroids that formed by resistant PNF cells displayed no significant change under 1 μM TAK-733. However, the size of spheroids that formed by nonresistant PNF cells was significantly reduced by treatment with 1 μM TAK-733, relative to the size of resistant spheroids (Figure 3f and i). The mechanism of acquired drug resistance and possible treatment options were then further explored.
      Figure thumbnail gr3
      Figure 3TAK-733‒resistant cells exhibit increased malignant abilities compared with parental cell lines. The (a) viability, (b) migration, (c) colony formation, (d, g) apoptosis, (e, h) cell cycle, and (f, i) sphere-forming ability of resistant ipNF05.5 (ipNF05.5R) and ipNF9511.bc (ipNF9511.bcR) cells were analyzed and compared with those of parental ipNF05.5 (ipNF05.5C) and ipNF9511.bc (ipNF9511.bcC) cells, respectively. ∗P < 0.05 and ∗∗P < 0.01 referred to control condition. Bar = 100 μm. C, control; h, hour; PI, propidium iodide; R, resistant.

       Quantitative high-throughput drug screening identified potential therapeutic agents for PNFs

      Long-term monotherapy can induce drug resistance. To select the possible therapeutic targets for PNF, a panel of 5,558 compounds assembled from MedChemExpress Chemicals, Selleck Chemicals (Shanghai, China), CSNpharm Chemicals (Shanghai, China), and other chemical companies was used to screen in both TAK-733‒resistant (Figure 4a) and ‒nonresistant (Supplementary Figure S2) ipNF9511.bc cells. This panel included various inhibitors of well-explored oncogenic targets, such as cyclin-dependent kinases (CDKs), histone deacetylases, and sodium ion‒potassium ion ATPases. The pharmacological outcomes in the various tested drugs illustrated their drug sensitivity profiles, and the details are displayed in the Supplementary Materials. We further performed RNA sequencing (RNA-seq) on both ipNF05.5 and ipNF9511.bc cells before and after induction of TAK-733 resistance, and the overall differences in gene transcription levels were visualized with heatmaps to explore the mechanisms underlying PNF cell resistance to TAK-733 (Figure 4b). We also performed Kyoto Encyclopedia of Genes and Genomes and Gene Ontology enrichment analysis of each parental and resistant cell line pair, and there were substantial overlap pathways across comparisons (Figure 4c and d). The analysis identified pronounced differences in cell cycle signaling between parental and resistant cell lines. That is, cell cycle gene sets were significantly upregulated in both resistant cell lines, and specifically, CDK1 gene activation was observed in both resistant cell lines (Figure 4e and g). The CDK inhibitor (CDKi) dinaciclib (the red triangle in Figure 4a), was thus selected as the compound with the most potential for combination therapy for PNF and was thus further analyzed.
      Figure thumbnail gr4
      Figure 4Potential mechanisms of drug resistance identified by drug screening and RNA-seq. (a) Quantitative high-throughput drug screening identified potential therapeutic agents for PNFs. Each agent in the screening panel is shown as a dot in the figure (the red triangle represents the CDKi dinaciclib). The name and chemical formula of dinaciclib are displayed. The horizontal dotted line indicates 70% inhibition of cell viability. RNA-seq of both ipNF05.5 and ipNF9511.bc cells before and after induction of resistance. (b) Heat map showing the differences in RNA levels in TAK-733‒resistant ipNF05.5 and ipNF9511.bc cells. (c, d) KEGG and (e, f) GO Enrichment Map analysis are displayed. (g) Significantly changed genes related to CDK signaling are listed. Akt, protein kinase B; BP, biological process; C, control; CC, cellular component; CDK, cyclin-dependent kinase; CDKi, cyclin-dependent kinase inhibitor; FC, fold change; GO, Gene Ontology; GPCR, G-protein coupled receptor; KEGG, Kyoto Encyclopedia of Genes and Genomes; PI3K, phosphatidylinositol 3 kinase; PNF, plexiform neurofibroma; R, resistant; RNA-seq, RNA sequencing; tRNA, transfer RNA; STAT, signal transducer and activator of transcription.

       The CDKi dinaciclib was identified as a promising agent in combination with TAK-733 on the basis of in vitro evaluation in PNF cell cultures

      The western blot results confirmed the pivotal role of CDK1 in treating PNFs with TAK-733 resistance, with resistant tumor cells having higher CDK1 expression than parental cells (Figure 5a). The expression of survivin was also higher in these cells than in parental cells, suggesting that the resistant cells developed inhibited apoptosis after the induction of resistance (Figure 5a). This result is consistent with the previous drug screening results and RNA-seq analysis because CDK1 was one of the identified targets of dinaciclib. Dinaciclib may become a promising therapeutic agent for PNFs, especially when resistance occurs. In vitro studies proved that for resistant ipNF9511.bc cells, the combination therapy with dinaciclib resulted in a significant reduction in cell viability (Figure 5b) and spheroid formation (Figure 5c and d), an increase in cell apoptosis (Figure 5e and f), and G2/M phase blockade (Figure 5g and h). For resistant ipNF05.5 cells, combination therapy can significantly inhibit cell viability but not spheroid formation and apoptosis. It is noteworthy that higher concentrations of dinaciclib may cause substantial decreases in the viability of PNF cells (Figure 5b).
      Figure thumbnail gr5
      Figure 5The CDKi dinaciclib was identified as a promising agent in combination with TAK-733 on the basis of in vitro evaluation in PNF cell cultures. TAK-733 monotherapy (1 μM; TAK), dinaciclib monotherapy (10 nM; DINA) and a TAK-733 and dinaciclib combination (COMBO/COMBO 1, 1 μM TAK-733 + 10 nM dinaciclib and COMBO 2, 1 μM TAK-733 + 20 nM dinaciclib) were administered to resistant ipNF05.5 and ipNF9511.bc cell lines. Treated cell lines were analyzed to evaluate (a) viability, (b, c) apoptosis, (d, g) sphere formation, and (e, f) the cell cycle. (h) Western blots determined the expression levels of ERK/p-ERK, survivin, and CDK1, and GAPDH served as the internal control. ∗P < 0.05 and ∗∗P < 0.01 referred to control condition. Bar = 100 μm. C, control; CDKi, cyclin-dependent kinase inhibitor; ERK, extracellular signal–regulated kinase; n.s., not significant; p-ERK, phosphorylated extracellular signal–regulated kinase; PNF, plexiform neurofibroma; R, resistant.

       The CDKi dinaciclib in combination with TAK-733 demonstrated efficacy in treating PNF-PDX mice

      To evaluate in vivo efficacy, we generated a patient‒derived xenograft model by grafting freshly extracted PNF tissue into the hypervascular armpit areas of NOD scid gamma mice; the procedure of the model establishment is displayed in Supplementary Figure S3. On confirmation of successful grafting, mice were first given trametinib (1 mg/kg), TAK-733 (30 mg/kg), or control medium daily by gavage for 4 weeks. After administration, TAK-733 significantly reduced the tumor weight by 50–60% compared with the control treatment. Trametinib also showed a trend for decreasing tumor weight, although it was not statistically significant (Figure 6a and b). Then, we evaluated the effectiveness of combination therapy with the CDKi dinaciclib (the combination group). After intraperitoneal injection of dinaciclib (20 mg/kg) three times a week over a 4-week treatment cycle, the tumors in the combination group were significantly reduced compared with those in the TAK-733 monotherapy group (Figure 6c and d). After administration, mice in the combination group had a significant reduction in their weight after the second-week treatment, compared with those in the TAK-733 monotherapy group (Figure 6e). We also evaluated the expression of specific markers in tumor sections to determine the pathological characteristics of PNFs before and after combination therapy. Both kinds of MEKis induced a robust decrease in p-ERK and Ki-67, suggesting that they were potent inhibitors of the growth of PNF cells (Figure 6f). The expression of p-ERK, Ki-67, c-Myc, and cyclin D1 decreased dramatically (Figure 6g). After identifying the different expression intensity of CDK1 (Figure 6h), we analyzed the expression and correlation between CDK1 and p-ERK with a tissue microarray (Figure 6i and j), and the result showed a moderate correlation (r = 0.4921, P < 0.001). H&E staining of main organs was performed and captured after the intervention of different chemical agents at the end of therapy. No obvious toxicity was observed to the main organs after a drug treatment of 1 month (Supplementary Figure S4). Information about the relationships between the expression of CDK1 and clinical characteristics is presented in Supplementary Table S3.
      Figure thumbnail gr6
      Figure 6The CDKi dinaciclib in combination with TAK-733 demonstrated efficacy in treating PNF-PDX mice. (a–d) Tumor bulk images and the weight of each group. Bar = 1 cm. (f, g) Representative images of IHC staining for S100-β, p-ERK, Ki-67, c-Myc, and cyclin D1, and H&E staining of grafts after treatment for 20 days. (e) Significant weight loss was seen with the combination therapy, and (h) H&E staining was performed to evaluate any major organ toxicity. (i, j) The expression of CDK1 and (k) its correlation with p-MEK are displayed. ∗P < 0.05 and ∗∗P < 0.01 referred to CON condition. Bar = 100 μm. CDKi, cyclin-dependent kinase inhibitor; CON, control; PDX, patient-derived xenograft; p-ERK, phosphorylated extracellular signal–regulated kinase; p-MEK, phosphorylated MAPK/extracellular signal–regulated kinase kinase; PNF, plexiform neurofibroma; TRA, trametinib.

      Discussion

      Although PNFs comprise various cell types, such as Schwann cells, fibroblasts, and mast cells, Nf1-deficient Schwann cell precursors have been reported to drive the formation of PNFs in mouse models (
      • Ratner N.
      • Miller S.J.
      A RASopathy gene commonly mutated in cancer: the neurofibromatosis type 1 tumour suppressor.
      ;
      • Yu Y.
      • Choi K.
      • Wu J.
      • Andreassen P.R.
      • Dexheimer P.J.
      • Keddache M.
      • et al.
      NF1 patient missense variants predict a role for ATM in modifying neurofibroma initiation.
      ). For this reason, medical agents used to treat PNFs must target Schwann cells. We compared the expression of several phenotypic biomarkers and confirmed the similarities and representativeness between the PNF cell line ipNF9511.bc and the PNF primary cells (Supplementary Figure S5). Currently, the most common models used to study PNF treatment are genetically engineered mouse models with Nf1 ablation driven by the Cre/lox system, and these models feature multiple histologically identical neurofibromas (
      • Dombi E.
      • Baldwin A.
      • Marcus L.J.
      • Fisher M.J.
      • Weiss B.
      • Kim A.
      • et al.
      Activity of selumetinib in neurofibromatosis type 1-related plexiform neurofibromas.
      ;
      • Jessen W.J.
      • Miller S.J.
      • Jousma E.
      • Wu J.
      • Rizvi T.A.
      • Brundage M.E.
      • et al.
      MEK inhibition exhibits efficacy in human and mouse neurofibromatosis tumors.
      ). Other models, including porcine models with Nf1 exon 41 or 42 deletion at the animal level and NF1−/− pluripotent stem cells at the cell level have the capacity to mimic many features in NF1-related processes (
      • Anastasaki C.
      • Wegscheid M.L.
      • Hartigan K.
      • Papke J.B.
      • Kopp N.D.
      • Chen J.
      • et al.
      Human iPSC-derived neurons and cerebral organoids establish differential effects of germline NF1 gene mutations.
      ;
      • Carrió M.
      • Mazuelas H.
      • Richaud-Patin Y.
      • Gel B.
      • Terribas E.
      • Rosas I.
      • et al.
      Reprogramming captures the genetic and tumorigenic properties of neurofibromatosis type 1 plexiform neurofibromas [published correction appears in Stem Cell Reports 2019;12:639–41].
      ;
      • Isakson S.H.
      • Rizzardi A.E.
      • Coutts A.W.
      • Carlson D.F.
      • Kirstein M.N.
      • Fisher J.
      • et al.
      Genetically engineered minipigs model the major clinical features of human neurofibromatosis type 1.
      ;
      • White K.A.
      • Swier V.J.
      • Cain J.T.
      • Kohlmeyer J.L.
      • Meyerholz D.K.
      • Tanas M.R.
      • et al.
      A porcine model of neurofibromatosis type 1 that mimics the human disease.
      ). On the basis of clinical considerations, we developed PNF-PDX NOD scid gamma mouse models, which are an easy and repeatable tool for in vivo benign tumor studies, compared with the relatively long and complicated methods for generating conditional models. PNF-PDX mouse models also retain most of the original PNF characteristics of the clinical patients from whom the tumor tissue was derived, and in our mice, the expression of S100-β and other protein markers remained the similar expression for as long as 4 weeks (Supplementary Figure S6).
      The mechanisms underlying the agents used in the various clinical trials related to PNFs are shown in Supplementary Figure S7. We also presented our findings in this figure. MEKis have been proven to be effective and tolerable in both preclinical assessments in NF1 mouse models and children with NF1 with inoperative, diffuse PNFs. However, there are potential drawbacks with continuous usage, such as gastrointestinal toxicities, elevation of creatinine kinase, and skin toxicities (
      • Klesse L.J.
      • Jordan J.T.
      • Radtke H.B.
      • Rosser T.
      • Schorry E.
      • Ullrich N.
      • et al.
      The use of MEK inhibitors in neurofibromatosis type 1-associated tumors and management of toxicities.
      ). Differences in response to these drugs are another challenge affecting clinical outcomes. On the basis of the data from this study, 66.67% of PNF tumor areas showed medium to high p-MEK expression, and this may correlate with the 70% positive response rate to the MEKi selumetinib (
      • Dombi E.
      • Baldwin A.
      • Marcus L.J.
      • Fisher M.J.
      • Weiss B.
      • Kim A.
      • et al.
      Activity of selumetinib in neurofibromatosis type 1-related plexiform neurofibromas.
      ). In addition, the variable expression levels of p-MEK/p-ERK in PNF tissues derived from different patients with NF1 and PNF cell lines in our results emphasize the heterogeneity present in these tumors, and such heterogeneity is another reason for the differences in MEKi efficacy. Moreover, drug resistance remains a problem. It is noteworthy that in the phase 2 trial of selumetinib, four patients were observed to have disease progression after confirmed partial response (patient identification: 1019010, 1019023, 2019002, and 2019003), indicating acquired resistance in these patients. The limited efficacy of monotherapies is due to the adaption and evolution of tumor cells, and small subpopulations of tumor cells can persist in the presence of drugs and eventually develop further mutations that allow them to grow and become the dominant populations (
      • Lim Z.F.
      • Ma P.C.
      Emerging insights of tumor heterogeneity and drug resistance mechanisms in lung cancer targeted therapy.
      ).
      Our research discussed the increased possibilities of resistance to MEKi in patients with NF1 with PNFs receiving long-term treatment. In both in vitro and in vivo studies, TAK-733 induced a stronger reduction in tumor growth and in markers related to cancer proliferation, such as Ki-67 and c-Myc, than trametinib. In this research, CDKi was proven to be a promising therapeutic agent for PNFs by both high-throughput drug screening and in vivo and in vitro experiments. Cancers display an uncontrolled cell cycle. CDKis have critical roles in regulating cell proliferation and transcription in tumor cells (
      • Asghar U.
      • Witkiewicz A.K.
      • Turner N.C.
      • Knudsen E.S.
      The history and future of targeting cyclin-dependent kinases in cancer therapy.
      ). Not long before this research, the CDK4/6 inhibitor palbociclib was determined to inhibit the growth of both malignant peripheral nerve sheath tumor cells and orthotopic xenografts, providing perspectives on CDKi in neurofibroma treatment (
      • Kohlmeyer J.L.
      • Kaemmer C.A.
      • Pulliam C.
      • Maharjan C.K.
      • Samayoa A.M.
      • Major H.J.
      • et al.
      RABL6A is an essential driver of MPNSTs that negatively regulates the RB1 pathway and sensitizes tumor cells to CDK4/6 inhibitors.
      ). In addition, the combination of CDKi and MEKi was also reported by
      • Kohlmeyer J.L.
      • Kaemmer C.A.
      • Pulliam C.
      • Maharjan C.K.
      • Samayoa A.M.
      • Major H.J.
      • et al.
      RABL6A is an essential driver of MPNSTs that negatively regulates the RB1 pathway and sensitizes tumor cells to CDK4/6 inhibitors.
      , who proposed that this combination could significantly promote cellular senescence and suppress the proliferation of lung cancer cells with Ras mutations (
      • Ruscetti M.
      • Leibold J.
      • Bott M.J.
      • Fennell M.
      • Kulick A.
      • Salgado N.R.
      • et al.
      NK cell-mediated cytotoxicity contributes to tumor control by a cytostatic drug combination.
      ).
      We examined the transcriptional profiles of PNF cells by RNA-seq and confirmed that tumor resistance to MEKi monotherapy may arise owing to CDK1 hyperactivation. CDK1, as a key determinant of mitotic progression, is needed for cell proliferation, as proven by mouse-knockout models (
      • Santamaría D.
      • Barrière C.
      • Cerqueira A.
      • Hunt S.
      • Tardy C.
      • Newton K.
      • et al.
      Cdk1 is sufficient to drive the mammalian cell cycle.
      ). The combination of both MEKi and CDKi significantly suppressed tumor growth in resistant cells and in vivo PNF-PDX models. This combination therapy showed strong inhibition of p-ERK expression, as evidenced by comparing the protein level before and after the induction of resistance cells and mouse models. The Ki-67 expression in the combination group was also observed lower than in the monotherapy group.
      It is worth noting that other MEKis showed limited effectiveness in reducing cell viability. This could be because traditional two-dimensional culture methods lacked the interactions between tumor and supporting infrastructure that presented in vivo, and instead, three-dimensional cell models provided a better pathomimetic recapitulation of PNF biology for more precise medical studies (
      • Kraniak J.M.
      • Chalasani A.
      • Wallace M.R.
      • Mattingly R.R.
      Development of 3D culture models of plexiform neurofibroma and initial application for phenotypic characterization and drug screening.
      ). In addition, significant weight loss in mice and viability inhibition in cell culture were found in the combination group compared with those in the TAK-733 monotherapy group. A reduced dinaciclib dose may be better tolerated in mice and may be a more appropriate dose for improved preclinical/clinical treatment, reducing adverse drug reactions and obtaining the best outcomes of combination therapy.
      Although contrasting results can be observed between preliminary drug studies and studies in model animals, which may be linked to tumor heterogeneity; the results mentioned earlier suggest that MEKi have good therapeutic effects in PNFs. We showed that drug resistance to the MEKi TAK-733 is likely caused by genetic differences leading to evasion of cell cycle arrest. In this study, we developed comprehensive preclinical models for testing MEKi monotherapy and MEKi/CDKi combination therapy: PNF tumor cell lines and PNF-PDX mouse models. By considering the mechanisms of resistance to MEKi in individual tumors, we identified this promising combination therapy to maximize tumor control. Our results provide a scientific basis for future clinical trials and the development of precision medicine to treat PNFs in patients with NF1.

      Materials and Methods

       Cell lines

      The human NF1-deficient PNF cell lines, including ipNF05.5, 9511.bc, 95.6, Schwann cell lines Ipn02.3 2λ, and NF1-deficient malignant peripheral nerve sheath tumor cell line S462, were purchased from the ATCC.

       Human PNFs samples and xenograft study

      Generally, patient‒derived xenograft models were established by transplanting a fresh PNFs tumor biopsy (cut into 3 × 3 × 3 mm tumor mass) from a patient with NF1 into the armpit of immunocompromised humanized NOD scid gamma mice aged 4 weeks under general anesthesia. Mice were divided into different treatment groups randomly (n = 4 per group) and were assigned to receive mono/combined drug therapies or vehicle by oral gavage or intraperitoneal injection. One week after operation, humanized NOD scid gamma mice were treated daily with trametinib (1 mg/kg), TAK-733 (30 mg/kg), or the drug diluted vehicle (10% DMSO + 40% PEG300 + 5% Tween-80 + 45% saline) by oral gavage. As for the combination therapy group, dinaciclib (20 mg/kg) was injected intraperitoneally three times a week. After 4 weeks of different treatment, mice were killed. Tumor xenografts were isolated and their weights were measured and photographed. Tissues were fixed overnight in 4% paraformaldehyde for further studies. All studies related to human or mice were designed in accordance with the Declaration of Helsinki and were approved by the Human Research Ethics Committee and Animal Care and Use Committee of Shanghai Ninth People’s Hospital Shanghai Jiao Tong University School of Medicine (Shanghai, China). Written informed consent for the use of the tissues for research was obtained from patients at the time of procurement of tumor specimens.

       High-throughput screening

      A library of 5,558 compounds was assembled from MedChemExpress Chemicals, Selleck Chemicals, CSNpharm Chemicals, and other chemical companies. Drug screening was performed as previously reported (
      • Lin G.L.
      • Wilson K.M.
      • Ceribelli M.
      • Stanton B.Z.
      • Woo P.J.
      • Kreimer S.
      • et al.
      Therapeutic strategies for diffuse midline glioma from high-throughput combination drug screening.
      ) and was used to culture the PNF cell line, ipNF9511.bc, at a density of 2,000 cells/well into 17 × 384-well plates and diluted at a final concentration of 200 μM. After 3 days of incubation, Celltiter Glo (Promega, Madison, WI) was added to examine the effect of inhibitors to evaluate cell viability. Agent selections for further combination strategies were determined by analyzing individual compound efficacy.

       RNA isolation and RNA-seq

      RNA isolation was performed using the Qiagen RNeasy Plus Mini kit (Qiagen, Hilden, Germany). RNA-seq libraries of poly (A) RNA from 500 ng total RNA were obtained from PNFs cell samples and were generated using the cBot Cluster Generation System of TruSeq PE Cluster Kit v3-cBot-HS for Illumina multiplexed sequencing (Illumina, San Diego, CA). Sequencing was performed from Illumina on a Novaseq 6000 with a paired-end module (Illumina). All the clustering and sequencing procedures were performed by Novogene Experimental Department. RNA-seq raw data of our samples were deposited to Gene Expression Omnibus (identification: GSE173577).

       Statistical analysis

      All the data are presented as mean ± SEM. All the graphs were plotted and analyzed with GraphPad Prism 8 Software (GraphPad Software, San Diego, CA) using Student’s t-test. P < 0.05 was considered statistically significant.
      More details were shown in Supplementary Materials and Methods.

       Data availability statement

      Related data that support the findings of this manuscript are available from the corresponding author on reasonable request.

      ORCIDs

      Conflict of Interest

      The authors state no conflict of interest.

      Acknowledgments

      We thank Qing-Feng Li for help in providing surgical tissues and clinical data and in supervising this manuscript. This work was supported by grants from Youth Doctor Collaborative Innovation Team Project (QC201803) of Shanghai Ninth People’s Hospital, Shanghai Jiaotong University School of Medicine; the Shanghai Youth Top-Notch Talent Program (201809004); Chenguang Program supported by Shanghai Education Development Foundation and Shanghai Municipal Education Commission ( 19CG18 ); Science and Technology Commission of Shanghai Municipality ( 19JC1413 ); Shanghai Rising Star Program ( 20QA1405600 ); Innovative research team of high-level local universities in Shanghai ( SSMU-ZDCX20180700 ); and Shanghai Municipal Key Clinical Specialty ( shslczdzk00901 ). For additional correspondence queries, please contact Zhi-Chao Wang ( [email protected] ).

      Author Contributions

      Formal Analysis: WW, YHG, MHC, RHGLA; Investigation: WW, YHG, CJW, YHL, XWC, JYR; Methodology: WW, YHG, QFL, ZCW; Writing - Original Draft Preparation: WW, YHG, HBZ, QFL, ZCW; Writing - Review and Editing: WW, YHG, HBZ, QFL, ZCW

      Supplementary Data

      Supplementary Materials and Methods

       Reagents

      All cells were cultured in DMEM (Gibco, Waltham, MA), supplemented with 10% fetal bovine serum (Gibco), glutamine, and ×1 penicillin/streptomycin (Gibco) at 37 °C in a double-distilled water humidified atmosphere with 5% carbon dioxide. To establish a TAK-733‒resistant plexiform neurofibromas (PNFs) cell subline, ipNF05.5 and 9511.bc cells were incubated in a one-sixth half-maximal inhibitory concentration of TAK-733 for 2 weeks at first. After cells stably growing, the drug concentration was elevated gradually over 6 months to a final dose of five times the previous half-maximal inhibitory concentration and withdrawn for two passages before functional studies. Selumetinib, trametinib, TAK-733, cobimetinib, refametinib, and PD0325901 were all purchased from MedChemExpress Chemicals (Monmouth Junction, NJ). Antibodies against CDK1 (ab133327), survivin (ab76424), and S100-β (ab52642) were purchased from Abcam (Cambridge, United Kingdom). Antibodies against GAPDH (#5174), MAPK/extracellular signal–regulated kinase (ERK) kinase (#4694), phosphorylated MAPK/ERK kinase (#9154), ERK (#4696), phosphorylated ERK (#4370), protein kinase B (#4691), phosphorylated protein kinase B (#4060), caspase-3 (#9662), cleaved caspase-3 (#9661), cyclin D1 (#55506), Ki-67 (#9449), and c-Myc (#18583) were purchased from Cell Signaling Technology (Danvers, MA).

       Immunohistochemistry, histological analysis, and immunocytochemistry

      A total of 141 tumor samples diagnosed as PNFs were collected and constructed as tissue microarray with two cores and 1 mm in diameter. Sections were cut into 4-μm paraffin-embedded sections and stained by either H&E or immunohistochemistry. For immunohistochemistry, after deparaffinization and dehydration, the pretreated slides were incubated in ×1 EDTA buffer (Sangon Biotech, Shanghai, China) and treated in a microwave oven at 95 °C for 25 minutes. The staining procedures strictly followed the instructions provided by UltraSensitive SP IHC Kit (KIT-9730) purchased from Maxim Biomedical (Fuzhou, China). For each tissue sample, the percentage of phosphorylated MAPK/ERK kinase, phosphorylated ERK, and CDK1-positive tumor cells was estimated depending on the staining intensity and on distribution separately by two pathologists, and the score was recorded semiquantitatively as 0, 1+, 2+ or 3+. For statistical analysis, the staining results were added ranging from 0 to 6 accordingly. For immunocytochemistry, cells were cultured on sterile glass coverslips, and the assay was performed as previously described (
      • Wang W.
      • Yan X.
      • Lin Y.
      • Ge H.
      • Tan Q.
      Wnt7a promotes wound healing by regulation of angiogenesis and inflammation: Issues on diabetes and obesity.
      ).

       Western blotting

      Protein lysis was extracted from PNFs tumors or cell lines by RIPA buffer (Beyotime Biotechnology, Shanghai, China) supplemented with complete protease/phosphatase inhibitor cocktails (MedChemExpress). After being separated by 10% SDS-PAGE and transferred to a polyvinylidene difluoride membrane, samples were blocked with 5% nonfat milk for 1 hour at room temperature and incubated with different primary antibodies, diluted according to the manufacturer’s instructions, at 4 °C overnight. Followed by incubation with horseradish peroxidase‒conjugated secondary antibodies for 1 hour, the immunoreactive bands were visualized by ImageQuant LAS 4000 Imager after using enhanced chemiluminescence‒immunoprecipitation buffer.

       Cell viability assay

      A total of 2,500 cells were seeded in each well of 96-well plates in the presence or absence of various concentrations of MAPK/ERK kinase 1/2 inhibitor. Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan) was added (1:10) to each well of different treatment after 72 hours of treatment and further incubated for 1.5 hours. Fluorescence was measured using a multiwell plate reader.

       Generation of tumor spheroids and colony formation assay

      For sphere assay, cells were seeded at a density of 500 cells in each well in 100 μl complete DMEM media in round bottom ultralow attachment 96-well plates (Corning, Corning, NY). One day after, different concentrations of drugs were added and incubated for an additional 5 days. As for colony assay, 800 cells were seeded in the presence or absence of inhibitors in each well of six-well plate. Growth medium was changed every 3 days. Colonies were formed within 10 days, stained by crystal violet (Beyotime Biotechnology), and imaged for calculating colony amount when cell counts were >50.

       Apoptosis assay and cell cycle assay

      A total of 1 × 106 cells were seeded in each well of six-well plates in the presence or absence of different concentrations of inhibitors. Apoptosis ratio was analyzed using the Annexin V and PI Kit (BD Biosciences, Franklin Lakes, NJ), and cell cycles were stained with RNA stain (Thermo Fisher Scientific, Waltham, MA). Results were tested by flow cytometry after seeding for 24 hours.

       TUNEL assay

      Four percent paraformaldehyde-fixed PNF cells on glass slides were washed with PBS and permeabilized by 0.5% Triton X-100 (Beyotime Biotechnology) for 10 minutes, followed by incubation with 100 μl of TUNEL reaction mixture according to the instructions from Yeasen Biotechnology (Shanghai, China). Nuclei were then stained with DAPI (Abcam) and observed under a fluorescence microscope.
      Figure thumbnail fx1
      Supplementary Figure S1Expression of MEK/ERK signaling pathway proteins was inhibited after 4 h and 24 h of TAK-733 and trametinib treatment in ipNF9511.bc cells. Akt, protein kinase B; Cl, cleaved; ERK, extracellular signal–regulated kinase; h, hour; MEK, MAPK/extracellular signal–regulated kinase kinase; p-Akt, phosphorylayed protein kinase B; p-ERK, phosphorylayed extracellular signal–regulated kinase; p-MEK, phosphorylated MAPK/extracellular signal–regulated kinase kinase; TRA, trametinib.
      Figure thumbnail fx2
      Supplementary Figure S2The result of a 5,558 compounds panel in quantitative high-throughput drug screening of TAK-733‒nonresistant ipNF9511.bc cells. The red triangle represented dinaciclib. The horizontal dotted line indicates 70% inhibition of cell viability. Akt, protein kinase B; GPCR, G-protein couped receptor; PI3K, phosphoinositide 3-kinase; STAT, signal transducer and activator of transcription
      Figure thumbnail fx3
      Supplementary Figure S3The main steps about PDX model establishment. NF1, neurofibromatosis type 1; NSG, NOD scid gamma; PDX, patient-derived xenograft; PNF, plexiform neurofibroma.
      Figure thumbnail fx4
      Supplementary Figure S4H&E staining of the main organs of PNF-PDX NSG mice model after 4 weeks of drug therapy. Bar = 100 μm. CON, control; NSG, NOD scid gamma; PDX, patient-derived xenograft; PNF, plexiform neurofibroma.
      Figure thumbnail fx5
      Supplementary Figure S5The expression of main phenotypic biomarkers to compare the differences among the PNF cell line ipNF9511.bc, MPNST cell line S462, and the PNF primary cells. Bar = 100 μm. MPNST, malignant peripheral nerve sheath tumor; PNF, plexiform neurofibroma.
      Figure thumbnail fx6
      Supplementary Figure S6Variable phenotypes of xenograft tumors on their histological characteristics by immunohistochemistry before implantation. The phenotypes at 2 wk and 4 wk after implantation are displayed. Bar = 100 μm. d, day; PDX, patient-derived xenograft; wk, week.
      Figure thumbnail fx7
      Supplementary Figure S7Clinical trials/mechanisms related to PNFs and the potential application of dinaciclib are presented. Akt, protein kinase B; ERK, extracellular signal–regulated kinase kinase; GDP, guanosine diphosphate; GTP, guanosine triphosphate; MEK, MAPK/extracellular signal–regulated kinase kinase; PI3K, phosphatidylinositol 3 kinase; PNF, plexiform neurofibroma.
      Supplementary Table S1Clinical Parameters with p-MEK Expression
      p-MEKHigh (4–6)Low (1–3)NegativeP-Value
      Sex0.869
       Male32447
       Female20316
      Age (y)0.207
       0–1525426
       15–4022336
       >40051
      Course of disease (y)0.016
      The expression of p-MEK is significantly associated with the course of disease (P = 0.016).
       0–318140
       3–104122
       >1012113
      Area0.643
       Head and neck41613
       Arms572
       Trunk76
      With CALMs or cNFs0.174
       Yes36538
       No16225
      Tumor size (cm)0.310
       0–39123
       3–107101
       >1016101
      Family history0.807
       Yes792
       No456611
      Abbreviations: CALM, cafe-au-lait macules; cNF, cutaneous neurofibroma; p-MEK, phosphorylated MAPK/extracellular signal–regulated kinase kinase.
      1 The expression of p-MEK is significantly associated with the course of disease (P = 0.016).
      Supplementary Table S2Clinical Parameters with p-ERK Expression
      p-ERKHigh (4–6)Low (1–3)NegativeP-Value
      Sex0.583
       Male34473
       Female27273
      Age (y)0.649
       0–1532393
       15–4027322
       >40231
      Course of disease (y)0.715
       0–312201
       3–1028332
       >1021213
      Area0.767
       Head and neck48594
       Arms861
       Trunk591
      With CALMs or cNFs0.839
       Yes44504
       No17242
      Tumor size (cm)0.304
       0–37113
       3–107101
       >1016101
      Family history0.532
       Yes7110
       No54636
      Abbreviations: CALM, cafe-au-lait macules; cNF, cutaneous neurofibroma; p-ERK, phosphorylated extracellular signal–regulated kinase.
      Supplementary Table S3Clinical Parameters with CDK1 Expression
      CDK1High (4–6)Low (1–3)NegativeP-Value
      Sex0.131
       Male32231
       Female38451
      Age (y)0.877
       0–1529390
       15–4038280
       >40322
      Course of disease (y)0.699
       0–315130
       3–1029330
       >1025191
      Area0.538
       Head and neck55540
       Arms951
       Trunk680
      With CALMs or cNFs0.703
       Yes49490
       No20202
      Tumor size (cm)0.292
       0–38131
       3–101170
       >1014130
      Family history0.817
       Yes990
       No60581
      Abbreviations: CALM, cafe-au-lait macules; cNF, cutaneous neurofibroma.

      References

        • Adjei A.A.
        • LoRusso P.
        • Ribas A.
        • Sosman J.A.
        • Pavlick A.
        • Dy G.K.
        • et al.
        A phase I dose-escalation study of TAK-733, an investigational oral MEK inhibitor, in patients with advanced solid tumors.
        Invest New Drugs. 2017; 35: 47-58
        • Anastasaki C.
        • Wegscheid M.L.
        • Hartigan K.
        • Papke J.B.
        • Kopp N.D.
        • Chen J.
        • et al.
        Human iPSC-derived neurons and cerebral organoids establish differential effects of germline NF1 gene mutations.
        Stem Cell Reports. 2020; 14: 541-550
        • Asghar U.
        • Witkiewicz A.K.
        • Turner N.C.
        • Knudsen E.S.
        The history and future of targeting cyclin-dependent kinases in cancer therapy.
        Nat Rev Drug Discov. 2015; 14: 130-146
        • Carrió M.
        • Mazuelas H.
        • Richaud-Patin Y.
        • Gel B.
        • Terribas E.
        • Rosas I.
        • et al.
        Reprogramming captures the genetic and tumorigenic properties of neurofibromatosis type 1 plexiform neurofibromas [published correction appears in Stem Cell Reports 2019;12:639–41].
        Stem Cell Reports. 2019; 12: 411-426
        • Dombi E.
        • Baldwin A.
        • Marcus L.J.
        • Fisher M.J.
        • Weiss B.
        • Kim A.
        • et al.
        Activity of selumetinib in neurofibromatosis type 1-related plexiform neurofibromas.
        N Engl J Med. 2016; 375: 2550-2560
        • Dombi E.
        • Solomon J.
        • Gillespie A.J.
        • Fox E.
        • Balis F.M.
        • Patronas N.
        • et al.
        NF1 plexiform neurofibroma growth rate by volumetric MRI: relationship to age and body weight.
        Neurology. 2007; 68: 643-647
        • Gross A.M.
        • Singh G.
        • Akshintala S.
        • Baldwin A.
        • Dombi E.
        • Ukwuani S.
        • et al.
        Association of plexiform neurofibroma volume changes and development of clinical morbidities in neurofibromatosis 1.
        Neuro Oncol. 2018; 20: 1643-1651
        • Gross A.M.
        • Wolters P.L.
        • Dombi E.
        • Baldwin A.
        • Whitcomb P.
        • Fisher M.J.
        • et al.
        Selumetinib in children with inoperable plexiform neurofibromas [published correction appears in N Engl J Med 2020;383:1290].
        N Engl J Med. 2020; 382: 1430-1442
        • Isakson S.H.
        • Rizzardi A.E.
        • Coutts A.W.
        • Carlson D.F.
        • Kirstein M.N.
        • Fisher J.
        • et al.
        Genetically engineered minipigs model the major clinical features of human neurofibromatosis type 1.
        Commun Biol. 2018; 1: 158
        • Jessen W.J.
        • Miller S.J.
        • Jousma E.
        • Wu J.
        • Rizvi T.A.
        • Brundage M.E.
        • et al.
        MEK inhibition exhibits efficacy in human and mouse neurofibromatosis tumors.
        J Clin Invest. 2013; 123: 340-347
        • Klesse L.J.
        • Jordan J.T.
        • Radtke H.B.
        • Rosser T.
        • Schorry E.
        • Ullrich N.
        • et al.
        The use of MEK inhibitors in neurofibromatosis type 1-associated tumors and management of toxicities.
        Oncologist. 2020; 25: e1109-e1116
        • Kohlmeyer J.L.
        • Kaemmer C.A.
        • Pulliam C.
        • Maharjan C.K.
        • Samayoa A.M.
        • Major H.J.
        • et al.
        RABL6A is an essential driver of MPNSTs that negatively regulates the RB1 pathway and sensitizes tumor cells to CDK4/6 inhibitors.
        Clin Cancer Res. 2020; 26: 2997-3011
        • Kraniak J.M.
        • Chalasani A.
        • Wallace M.R.
        • Mattingly R.R.
        Development of 3D culture models of plexiform neurofibroma and initial application for phenotypic characterization and drug screening.
        Exp Neurol. 2018; 299: 289-298
        • Le L.Q.
        • Parada L.F.
        Tumor microenvironment and neurofibromatosis type I: connecting the GAPs.
        Oncogene. 2007; 26: 4609-4616
        • Lim Z.F.
        • Ma P.C.
        Emerging insights of tumor heterogeneity and drug resistance mechanisms in lung cancer targeted therapy.
        J Hematol Oncol. 2019; 12: 134
        • Lin G.L.
        • Wilson K.M.
        • Ceribelli M.
        • Stanton B.Z.
        • Woo P.J.
        • Kreimer S.
        • et al.
        Therapeutic strategies for diffuse midline glioma from high-throughput combination drug screening.
        Sci Transl Med. 2019; 11eaaw0064
        • Martinelli E.
        • Morgillo F.
        • Troiani T.
        • Ciardiello F.
        Cancer resistance to therapies against the EGFR-RAS-RAF pathway: the role of MEK.
        Cancer Treat Rev. 2017; 53: 61-69
        • Perreault S.
        • Larouche V.
        • Tabori U.
        • Hawkin C.
        • Lippé S.
        • Ellezam B.
        • et al.
        A phase 2 study of trametinib for patients with pediatric glioma or plexiform neurofibroma with refractory tumor and activation of the MAPK/ERK pathway: TRAM-01.
        BMC Cancer. 2019; 19: 1250
        • Prada C.E.
        • Rangwala F.A.
        • Martin L.J.
        • Lovell A.M.
        • Saal H.M.
        • Schorry E.K.
        • et al.
        Pediatric plexiform neurofibromas: impact on morbidity and mortality in neurofibromatosis type 1.
        J Pediatr. 2012; 160: 461-467
        • Ratner N.
        • Miller S.J.
        A RASopathy gene commonly mutated in cancer: the neurofibromatosis type 1 tumour suppressor.
        Nat Rev Cancer. 2015; 15: 290-301
        • Ruscetti M.
        • Leibold J.
        • Bott M.J.
        • Fennell M.
        • Kulick A.
        • Salgado N.R.
        • et al.
        NK cell-mediated cytotoxicity contributes to tumor control by a cytostatic drug combination.
        Science. 2018; 362: 1416-1422
        • Samatar A.A.
        • Poulikakos P.I.
        Targeting RAS-ERK signalling in cancer: promises and challenges.
        Nat Rev Drug Discov. 2014; 13: 928-942
        • Santamaría D.
        • Barrière C.
        • Cerqueira A.
        • Hunt S.
        • Tardy C.
        • Newton K.
        • et al.
        Cdk1 is sufficient to drive the mammalian cell cycle.
        Nature. 2007; 448: 811-815
        • Schadendorf D.
        • Fisher D.E.
        • Garbe C.
        • Gershenwald J.E.
        • Grob J.J.
        • Halpern A.
        • et al.
        Melanoma. Nat Rev Dis Primers. 2015; 1: 15003
        • Stark R.
        • Grzelak M.
        • Hadfield J.
        RNA sequencing: the teenage years.
        Nat Rev Genet. 2019; 20: 631-656
        • Wang J.
        • Pollard K.
        • Calizo A.
        • Pratilas C.A.
        Activation of receptor tyrosine kinases mediates acquired resistance to MEK inhibition in malignant peripheral nerve sheath tumors.
        Cancer Res. 2021; 81: 747-762
        • White K.A.
        • Swier V.J.
        • Cain J.T.
        • Kohlmeyer J.L.
        • Meyerholz D.K.
        • Tanas M.R.
        • et al.
        A porcine model of neurofibromatosis type 1 that mimics the human disease.
        JCI Insight. 2018; 3e120402
        • Yu Y.
        • Choi K.
        • Wu J.
        • Andreassen P.R.
        • Dexheimer P.J.
        • Keddache M.
        • et al.
        NF1 patient missense variants predict a role for ATM in modifying neurofibroma initiation.
        Acta Neuropathol. 2020; 139: 157-174