Purinergic receptor P2RY14 cAMP signaling regulates EGFR-driven Schwann cell precursor self-renewal and nerve tumor initiation in neurofibromatosis

Neurofibromatosis type 1 (NF1) is a genetic disorder characterized by nerve tumors called neurofibromas, in which Schwann cells (SCs) lack NF1 and show deregulated RAS signaling. NF1 is also implicated in regulation of cAMP. Gene expression profiling and protein expression identified P2RY14 in SCs and SC precursors (SCPs) implicating P2RY14 as a candidate upstream regulator of cAMP in EGF-dependent SCP. We found that SCP self-renewal was reduced by genetic or pharmacological inhibition of P2RY14. In NF1 deficient SCs and malignant peripheral nerve sheath tumor (MPNST) cells, P2RY14 inhibition decreased EGFR-driven phospho-Akt and increased cAMP signaling. In a neurofibroma mouse model, genetic deletion of P2RY14 increased mouse survival, delayed neurofibroma initiation and rescued cAMP signaling. Conversely, elevation of cAMP diminished SCP number in vitro and diminished SC proliferation in neurofibroma bearing mice in vivo. These studies identify the purinergic receptor P2RY14 as a critical G-protein-coupled receptor (GPCR) in NF1 mutant SCPs and SCs and suggest roles for EGFR-GPCR crosstalk in facilitating SCP self-renewal and neurofibroma initiation via cAMP and EGFR-driven phospho-Akt.


Lead Contact 24
Further information and requests for resources and reagents should be directed to and 25 will be fulfilled by the Lead Contact, Nancy Ratner, PhD (nancy.ratner@cchmc.org). Neurofibromatosis type 1 (NF1) is an autosomal dominant disease that affects up 71 to 1:2,000 individuals worldwide (Kallionpää et al., 2018). To date, there is no cure for 72 NF1, which is characterized by multiple, variable, clinical manifestations (Friedman, 1998;73 Tabata et al., 2020). At least half of the children with NF1 develop plexiform 74 neurofibromas (PNs), which are tumors within peripheral nerves. PN may be present at 75 birth and show most rapid growth during the first decade of a child's life (Nguyen, et al., 76 2012). PNs can occur in any cranial or peripheral nerve and have the potential to 77 transform into lethal malignant peripheral nerve sheath tumors (MPNST) (Prudner,Ball,78 Rathore, & Hirbe, 2020). Neurofibroma infiltration of normal nerves in NF1 patients results 79 in a complicated risk profile because it can cause nerve damage and compress nearby 80 vital organs (Kim, et al., 2017). Therefore, understanding how neurofibromas form and 81 how to treat them is under intense investigation. 82 Peripheral nerve glial cells, Schwann cells (SCs), are the only cell type in 83 neurofibromas that show bi-allelic loss-of-function mutations in the NF1 tumor suppressor 84 gene (Serra et al., 1997(Serra et al., , 2001. Neurofibroma SCs also show aberrant properties ex vivo, 85 consistent with it being the primary pathogenic cell type in neurofibromas (Sheela et al., 86 1990; Kim et al., 1995). In mice, neural crest cells develop into Schwann cell precursors 87  increases in GTP-bound RAS (Kim et al., 1995;Sherman et al., 2000), and RAS-GTP 115 stimulates the mitogen-activated protein kinase (MAPK) pathway and other downstream 116 pathways. Loss of NF1 also causes reduced levels of cyclic AMP (cAMP) in Nf1 mutant 117 mouse, fly and zebrafish (Hegedus, et al., 2007;Tong, et al., 2002;Wolman et al., 2014;118 Anastasaki and Gutmann, 2014). Whether cAMP deregulation occurs downstream of 119 increased RAS-GTP is unclear. Neurofibromin shares homology with the yeast proteins 120 Ira1 and Ira2, which are inhibitory regulators of the RAS-cAMP adenylyl cyclase pathway, 121 but no evidence shows a direct role for NF1 or IRA proteins in direct regulation of cAMP 122 in mammals (Ballester, et al., 1990;Martin, et al., 1990;Xu, et al., 1990). It is unclear 123 how, or if, regulation of cAMP is relevant to neurofibroma initiation or growth but reducing 124 cAMP drove formation of brain tumors in cells lacking Nf1 (Warrington et al., 2010). 125 We sought to identify molecules that might affect neurofibroma development and 126 regulate EGFR signaling expression in the SC lineage. We identified P2RY14 as a G- We dissociated cells from human plexiform neurofibromas resected for therapeutic 143 purposes from 3 neurofibroma patients. Cells were sorted into p75 + /EGFR -SCs and 144 p75 + /EGFR + SCP-like tumor initiating cells. We performed gene expression analysis on 145 these cells and found that P2RY14 mRNA is elevated in p75 + /EGFR + SCP-like tumor 146 initiating cells ( Figure 1A). Western blot also showed expression in human SCs and 147 neurofibroma SCPs, with 1.9-fold increase of P2RY14 protein in SCP-like cells ( Figure  148 1B). To test if P2RY14 + SCP-like cells derived from human neurofibromas have altered 149 ability to self-renew, we performed fluorescence activated cells sorting (FACS) and sorted 150 SCP-like cells into p75 + /EGFR + /P2RY14and p75 + /EGFR + /P2RY14 + cells and plated 151 them at low density to generate unattached spheres in vitro ( Figure 1C, D, E). Unsorted 152 neurofibroma cells rarely form SCP-like spheres. FACS analysis (of cells from 3 additional 153 neurofibroma tumors) showed that on average p75 + /EGFR + /P2RY14cells formed 154 spheres at a frequency of 23.4%, while 64.8% of p75 + /EGFR + /P2RY14 + cells formed 155 spheres. The p75 + /EGFR + /P2RY14 + cells maintained their significantly enhanced ability 156 to form spheres in vitro for three passages (Figures 1F, 1G). Thus, P2RY14 is 157 overexpressed in human neurofibroma SCs and SCPs in vitro, and marks SCP with the 158 potential to self-renew in vitro. 159 Mouse Nf1 mutant SCPs use P2RY14 signaling through Gi to regulate self-renewal. 160 We verified P2RY14 expression in cultured SCP spheres from wild-type (WT) and Nf1-/-161 mouse embryos. P2RY14 protein was slightly elevated in Nf1 -/-SCPs (Figure 2A)  To confirm these results, we silenced P2RY14 gene expression using short-hairpin 172 RNAs (shRNAs) targeting P2RY14. WT and Nf1 -/mouse SCPs were treated with non-173 target control (shNT) or shRNA P2RY14 (shP2RY14). shP2RY14 treated cells showed 174 reduced P2RY14 mRNA ( Figure 2C) and P2RY14 protein ( Figure 2D). Sphere formation 175 was significantly decreased in Nf1 -/-SCPs but not wild-type treated with P2RY14 shRNA, 176 and this phenotype also persisted for 3 passages ( Figure 2E). Photomicrographs are 177 shown for one shRNA (sh09) (Supplemental figure 1B); but the experiment was 178 repeated with 2 additional P2RY14 shRNAs in 3 biological replicates each, with similar 179 results (Supplemental figure 3A-3D). To delineate signaling pathways affected by 180 P2RY14 shRNA in SCPs, we analyzed cell lysates in western blots. We found that 181 shP2RY14 caused a slight decrease in pERK, a read-out of RAS/MAPK signaling, which 182 we know to be elevated in Nf1-/-cells ( Figure 2F). We analyzed cAMP-dependent protein 183 kinase (PKA) substrate phosphorylation (using anti p-PKA substrate antibody) as an 184 indirect read out of cAMP levels in cells. p-PKA substrate phosphorylation increased in 185 WT and Nf1-/-SCPs shP2RY14 treated cells. Interestingly, knockdown of P2RY14 in Nf1 -186 /-SCPs increased levels of pPKA, similar to those in WT shNT levels ( Figure 2G). 187 To understand if P2RY14-mediated changes in cAMP signaling affect SCP self-188 renewal, we treated WT and Nf1 -/-SCPs with the specific phosphodiesterase-4 (PDE4) 189 inhibitor, rolipram. Rolipram blocks degradation of cAMP by PDE4, increasing intracellular Given that EGFR and P2RY14 similarly modulate SCP self-renewal, we 225 hypothesized that P2RY14 inhibition might affect EGFR downstream signaling. To test 226 this idea, we treated 1λ and 1λ∆NF1 with epidermal growth factor (EGF), with and without 227 the P2RY14 inhibitor (PPTN). We monitored EGFR-phosphorylation at the Grb2 228 activation site to measure EGFR activation, and examined downstream signaling with 229 pERK1/2 and phosphorylation of Akt at the threonine 308 (T308) activation site (Vincent, 230 et al., 2011). In 1λ∆NF1 cells, P2RY14 inhibition reduced phosphorylation of EGFR at the 231 Grb2 activation site and shortened the duration of p-AKT signaling ( Figure 3D; 232 Supplemental figure 4). Similar results were seen in the NF1 deficient MPNST cell line, 233 ST88-14 ( Figure 3E-3F). To explain these results, we hypothesized that P2RY14 might 234 form a complex with EGFR. To test this idea, we transfected ST88-14 cells with P2RY14-235 GFP and/or EGFR-myc tagged proteins. Then, we immunoprecipitated EGFR using an 236 anti-myc antibody and detected P2RY14 in a complex with anti-GFP EGFR ( Figure 3G).  To study an additional SC phenotype in these mice, we examined the effects of 294 P2RY14 loss on nerve disruption phenotype using electron microscopy. At 4-months-old 295  In conclusion, purinergic receptor P2RY14 cAMP signaling regulates EGFR-driven 400 Schwann cell precursor self-renewal and nerve tumor initiation in neurofibromatosis. 401 Several groups have suggested that combined inhibition of EGFR and GPCRs could be 402 a promising strategy in cancer therapy (Köse, 2017). P2RY14 is a candidate for this type 403 of therapeutic intervention as a whole-body knockout of this receptor showed no effect on 404

Lead Contact 553
Further information and requests for resources and reagents should be directed to and 554 will be fulfilled by the Lead Contact, Nancy Ratner, PhD (nancy.ratner@cchmc.org). 555

Materials Availability 556
This study did not generate new unique reagents. 557

Data and Code Availability Statement 558
The data sets and original figures generated during this study will be available at Synapse 559 Project (https://www.synapse.org/). 560

Human neurofibroma sample collection 573
Fresh plexiform neurofibromas (n=3) were obtained after medically mandated surgeries. 574 All samples were obtained with patient consent under IRB approval. 575

FACS Analysis 576
Fresh surgical plexiform neurofibroma (PNF) specimens were enzymatically dissociated 577 as described (Williams et al, 2008). For cell sorting, we incubated cell suspensions with 578 anti-P2RY14 receptor antibody (Rabbit, polyclonal, Alomone labs, # APR-018) on ice for 579 30 minutes, washed with PBS twice. We then incubated cells with goat-anti-rabbit-APC 580 (Selleck, Cat# S1430). To passage, we centrifuged sphere cultures, treated with 0.05% 617 Trypsin for 3 min., dissociated and plated at 2 × 10 4 cells/ml in 50% conditioned and 50% 618 fresh medium. We counted secondary spheres after 14 days. For every cell line three 619 biological replicates with three technical replicates were done. Of the three biological 620 replicates the best one was reported as a representative (n=3). Spheres were counted 621 with an inverted phase contrast microscope after 6 days of plating.  Cells were pre-incubated with MEK inhibitor for two hours prior to stimulation. Stimulation 653 was carried out for 2 minutes and cAMP measurement was done according to 654 manufacturer's protocol (Direct cAMP ELISA kit by Enzo Life Sciences, Cat# ADI-900-655 066) using the acetylated format. An aliquot prior to cAMP measurements was set aside 656 for protein quantification using the Bio-Rad protein assay kit. For every cell line two 657 biological replicates with three technical replicates were done. Of the two biological 658 replicates the best one was reported as a representative (n=3 biological replicates, 3 659 technical replicates). 660

Immunostaining 661
For frozen sections, OCT was removed by incubation with 1XPBS. We permeabilized 662 cells in ice cold MeOH for 10 min., followed by incubation in normal donkey serum 663

Mouse dissection and quantification of neurofibroma number and size 681
To quantify neurofibroma number and size, we perfused mice and used a Leica dissecting 682 microscope to dissect the spinal cord with attached DRG and nerve roots at the ages of 683 4-months and 7-month, as previously described (Wu, et al., 2016). A neurofibroma was 684 defined as a mass surrounding the DRG or nerve roots, with a diameter greater than 685 1mm, measured perpendicular to DRG/nerve roots. Neurofibroma diameter for each 686 mouse were measured with Image J. 687

RT-PCR 688
We isolated total RNA from WT and Nf1-/-spheres treated with shP2RY14 using the 689 RNeasy Plus Micro-Kit (QIAGEN, Cat# 74034) and made cDNA using the High Capacity 690 Reverse Transcription Kit (Thermo Fisher, Cat# 4368813). We conducted rt-PCR as 691  comparisons used Student's t-tests. When single agents were tested at different 704 concentration in a single cell type, we used a one-way ANOVA with a Dunnett's multiple 705 comparisons test. When multiple genotypes were analyzed in a single experiment, we 706 used a Two-way ANOVA with multiple comparisons, without matching, and correction 707 with the Holm-Sidak test. Mann-Whitney test was used for comparisons between 708 genotypes for tissue widths and neurofibroma incidence (GraphPad Prism V9). All data 709 unless otherwise stated is represented as average ± SD, and was analyzed in GraphPad 710 Prism 7.