ADT-030, a novel PDE10 inhibitor, demonstrates potent antitumor activity in pancreatic ductal adenocarcinoma

33 Phosphodiesterase 10 (PDE10) was previously reported to be overexpressed in various cancers 34 and essential for cancer cell prolif eration and survival. Here, we studied a novel PDE10 inhibitor , 35 ADT-030, and found it to potently and selectively inhibit KRAS mutant PDAC cell prolif eration 36 and clonogenicity by inducing G2/M arrest and apoptosis. ADT-030 also inhibited motility of 37 PDAC cells in vitro . These eff ects were mediated by increased cAMP /cGMP levels and activation 38 of PKA/PK G. The growth inhibitory activity of ADT-030 was associated with reduced β- catenin 39 and RAS signaling. Notably , ADT-030 also inhibited the growth of KRAS G12D and KRAS G12 C mutant 40 PDAC cells resistant to allele-specific KRAS inhibitors. Oral administration of ADT-030 41 significantly suppressed tumor growth, reduced lung and liver metastasis, and increased 42 survival without systemic toxicity in syngeneic and patient-derived xenograft (PDX) PD AC 43 models. ADT-030 also increased chemotherapy response in orthotopic PDAC models. Immune 44 phenotyping and single-cell RNA sequencing revealed remodeling of the tumor 45 microenvironment by ADT-030 with a more favorable i mmune suppressive profile to activate 46 anti-tumor immunity . These results show that ADT-030 is a promising drug development 47 candidate for the treatment of KRAS-mutant PDAC capable of simultaneously targeting k ey 48 oncogenic signaling pathways, r esulting in tumor-intrinsic and immunomodulatory eff ects. 49 Key words: KRAS-mutant PDAC, targeted therapy , immune checkpoint inhibition, T and NK cell 50 activation, PKA/PK G signaling, myeloid polarization, PDE10, β-catenin. 51 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint 52 53

54 Pancreatic ductal adenocarcinoma (PDAC) is the fourth most common cause of cancer-55 related mortality in the U.S., with a five-year survival of under 13% 1 . Sur vival of patients with 56 metastatic PDAC r emains poor even with the current chemotherapeutic regimens such as 57 FOLFIRINO X or gemcitabine and nab-paclitax el 2 . The aggressive nature of PDAC is mainly 58 attributed to activation of multiple compensatory signaling path ways driv ing proliferation an d 59 survival, along with a hypoxic microenviroment driven by dense desmoplastic stroma and 60 decreased vascular perfusion 3- 7 . Although allele-specific KRAS inhibitors have demonstrated 61 promising activity in early-phase clinical trials in patients with PDAC, the development of 62 resistance remains a major challenge and highlights the need to identify new therapeutic 63 targets and agents with broader activity 8-1 0 . A better understanding of the complexity of 64 oncogenic signaling, the importance of stroma, and the role of immune evasion in PDAC 65 progression is critical for the development of more effective target-directed drugs for the 66 treatment of PDAC 11 . 67 Mutations in the KRAS gene have been reported in about 90% of PDAC patients, with the 68 majority at the 12 th codon (KRAS G12 D , KRAS G12V , and KRAS G12 C ) 12 . These mutations result in the 69 constitutive activation of downstream pathways such as RAS/RAF /MEK and PI3K/ AKT /mT OR 70 signaling to promote the proliferation, survival, and metabolic reprogramming of PDAC 13 . 71 Although the mutation frequency of CTNNB1 is relatively low in PDAC , β-catenin signaling is 72 aberrantly activated from WNT overexpression, which, along with KRAS mutations, contributes 73 to the aggressive behavior of PDAC 6 . KRAS has been reported to form a complex with β- cateni n 74 to modulate the phosphorylation of the transcription factor TCF4, leading to crosstalk between 75 these two oncogenic signaling pathways 14 , 15 . In addition, both β-catenin and RAS signaling have 76 been reported to be activated with gemcitabine treatment, suggesting that these pathways play 77 a major role in therapy resistance in PDAC 16 . Given the interactions between RAS and β- cateni n 78 in PDAC, a strategy that targets both pathways with single inhibitor could off er more robust 79 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint therapeutic approach. Emerging evidence also suggests that simultaneous inhibition of multiple 80 oncogenic pathways not only increases the potential for efficacy of target-directed anticancer 81 drugs but also reduces the potential for resistance by overcoming compensatory signaling 82 mechanisms 17,1 8 . KRAS mu tated tum ors can u til ize β-caten in si gnaling to maintain a stem-lik e, 83 immune-depleted, niche within the tumor immune microenvironment (TiME) resulting in 84 relapse following chemotherapy . T arg eting both pathways with one inhibitor could also sensitize 85 cancer cells to undergo apoptosis while modulating the TiME to favor immune activation 86 leading to inhibition of tumor growth 19 . 87 Phosphodiesterase (PDE) isoenzymes hydrolyze and inactivate the second messengers, 88 cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP) 20 . 89 Although understudied, PDEs have been reported to play a role in the initiation and progression 90 of PDAC and other cancers 21 , 22 . Notably , the cAMP /PKA and cGMP /PK G signaling ax es have 91 been reported to suppress MAPK signaling downstream from KRAS 23 , 2 4 . In addition, PDE 92 isoenzymes have been shown to regulate β-catenin signaling, which can also influence RAS 93 signaling 25- 27 . Several PDE isozyme families, most notably , PDE4, PDE5, and PDE10 have been 94 investigated as anti-cancer targets inhibition of which can impact cancer cell prolif eration, 95 survival, and immune responses 22, 28 . Notably , isozyme-specific inhibitors of the dual 96 cAMP /cGMP degrading PDE10 isozyme and gene silencing approaches have been reported to 97 selectively inhibit the prolif eration and induce apoptosis of cells from colon, lung, and ovarian 98 cancers through activating cGMP /PKG signaling and disrupting RAS signaling and WNT /β-99 catenin-mediated transcription 25, 27 ,2 9 ,30 . These findings established the basis for our hypothesis 100 that ADT-030, a novel PDE10 inhibitor with properties distinct from known PDE10 inhibitors 101 developed for CNS disorders, can block both RAS and β-catenin signaling and result in tumor 102 inhibition and modulation of the TiME in PDAC. 103 104 105 106 107 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint 108 109 110

111 ADT-030 synthesis 112 The synthesis of ADT-030 [(S,Z)-2-(5-methoxy-2-methyl-1-(3,4,5-trimethoxybenzylidene)-1H-113 inden-3-yl)-N-(1-methylpyrrolidin-3-yl)acetamide] is based on a procedure originally described 114 in US patent 20200223815 using 3-(4-methoxyphenyl)-2-methylacrylic acid as the starting 115 material. 116 Human scRNA-seq datamining 117 The Single Cell RNA seq Pancr eatic Cancer A tlas R Data Serialization (RDS) file 118 ( https://zenodo.org /records/14199536 ) was downloaded 31 . This dataset has normalized and 119 scaled scRNA-seq (10x genomics sequencing) data from 12 studies containing 229 patients 120 across the groups. The ductal cells were identified from the main dataset and the expression 121 levels of PDE10 was queried across the tissue samples (donor , adjacent normal, primary tumor , 122 and meta static lesi on) as described in the previously published paper 31 . 123 Cellular target engagement assay 124 HEK293 cells expressing PDE10 fused to the MICRO-T AG reporter were subjected to a 125 temperature-series denaturation assay to determine the aggregation midpoint under cellular 126 conditions as previously described 32 . Cells were heated for 10 min across a defined 127 temperature range, followed by non-denaturing lysis and fluorescence complementation 128 quantification. Fitting the resulting thermal curve yielded a T agg₅₀ of 44°C for PDE10. This 129 defined T agg₅₀ provided the fix ed challenge temperature for subsequent experiments to 130 determine if ADT-030 binds PDE10 in intact cells. 131 Cell culture 132 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint Panc-1 (A TCC# CRL-1469; RRID:CVCL_0Q68), AsPC-1 (A TCC# CRL-1682; RRID:CVCL_0152), Panc 133 02.03 (A T CC# CRL-2553; RRID:CVCL_1633), Panc 10.05 (A T CC# CRL-2547; RRID:CVCL_1639), MIA 134 PaCa-2 (A T CC# CRL-1420; RRID:CVCL_0428), BxPC-3 (A TCC# CRL-1687; RRID:CVCL_0186), and 135 KLE (A TCC# CRL-1622; RRID:CV CL_1329) cell lines were obtained from American T ype Culture 136 Collection (A TCC, Manassas, V A, USA) and maintained as recommended. Mouse PDAC cell line 137 2838c3 (Kerafast# EUP013-FP; RRID:CVCL_YM18) was purchased from Kerafast (Boston, MA, 138 USA). MKN1 (Accegen# ABC-T C0685; RRID:CVCL_1415) cell line was purchased from Accegen 139 Inc (F airfield, NJ, USA). Dr . Gregory Lesinski, Emory University , USA, gifted the KPC cell line. Dr . 140 Denis C Guttridge, Medical University of South Carolina, USA, gifted the KPCML1 cell line. All 141 cells were grown in appropriate medium as recommended by the A TCC and Kerafast with either 142 Dulbecco’ s Modified Eagle Medium (DMEM; A TCC# 30-2002) or Roswell Park Memorial Institute 143 (RPMI)-1600 Medium (RPMI; A T CC# 30-2001) supplemented with 10% f etal bovine serum (FBS; 144 A TCC# 30-2020) and 1% penicillin/streptomycin (A TCC# 30-2300) under 5% CO 2 . Additionally , 145 MIA PaCa-2 cells received 2.5% horse serum (Thermo# 26050088). 146 Phosphodiesterase assay 147 The enzymatic activity of recombinant PDE10 was measured using the Immobilized Metal 148 Affinity Particle (IMAPTM) fluorescence polarization (FP) progressive binding system (Molecular 149 Devices; San Jose, CA; USA) as previously described to determine the inhibitory eff ect of ADT-150 030 33 . FP was measured using a Synergy H4 Hybrid plate reader (BioT ek; Santa Clara, CA; USA). 151 Recombinant PDE10 was purchased from BPS Biosciences (San Diego, CA; USA). 152 Proliferation assa y 153 Human and murine PDAC cells with KRAS mutations (KRAS G12D and KRAS G12 C ) and wild-type cells 154 (BxPC-3) were plated at a density of 5×10 3 cells/well in 96-well plates. After 20 hrs, cells were 155 treated with ADT-030 or PF-2545920 (a known PDE10 inhibitor) 34 in a dose-dependent manner . 156 After 72 hrs, the medium was removed, 10-μL methylthiazole tetrazolium (MTT ; 5 mg /mL in 157 PBS; Sigma-Aldrich# 475989) was added, and cells were incubated for another 2 hrs at 37 °C. 158 The resulting formazan crystals were solubilized in 100 μL DMSO (Sigma-Aldrich# D2438), and 159 absorbance was measured at 570 nm with a ref erence wavelength of 630 nm. 160 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint Clonogenic as say 161 Mouse derived PDAC cell line 2838c3 (KRAS G12D ) and human derived cell line MIA PaCa-2 162 (KRAS G12C ) were plated in 6-well culture plates (2 × 10 3 cells/well). The cells were then treated 163 with either DMSO , ADT-030 at 0.5, 1, 2, and 5 µM, or PF-2545920 at 5, 10, 25, 50, and 100 µM 164 every 3 days. After 10 days, the cells were stained with a 0.005% Coomassie Brilliant Blue R-250 165 solution, and plates were imaged using an Epson Perf ection V850 Pro Photo Scanner (USA). The 166 resulting colonies were counted using ImageJ (RRID:SCR_003070). 167 Motility assa y 168 T o measure eff ects on cell motility , 2838c3 and MIA PaCa-2 cells were grown in 6-well plates 169 until they reached confluence. The cells were treated with DMSO or ADT-030 (0.5, 1, 2, and 5 170 µM). A scratch was created using a sterile 10-µL pipette tip, and cell migration was monitored 171 daily using light microscopy . Quantification of cell movement was performed using ImageJ 172 software. 173 Apoptosis as say 174 The binding of annexin V to cells was measured using the PE-Annexin V Apoptosis Detection Kit 175 I (BD Biosciences# 559763), according to the manufacturer ’ s protocol. Briefly , 2838c3 and MIA 176 PaCa-2 cells were treated with either DMSO or ADT-030 (at 2 and 5 µM) for 72 hrs. After 177 treatment, cells were collected, washed twice with cold 1x PBS (A T CC# 30-2200), and suspended 178 in 1x Binding Buff er . The cells were then stained with 300 µL PE Annexin V F ITC and 5-µL of 179 propidium iodide (PI) and incubated for 15 min in the dark. Flow cytometry analysis was 180 performed using a BD LSR Fortessa Flow Cytometer and data were analyzed using FlowJo 181 (RRID:SCR_008520). 182 Cell cycle assay 183 2838c3 and MIA PaCa-2 cells were treated with either DMSO or ADT-030 (2 and 5µM) for 24 hrs 184 and the cells were trypsinized and centrifuged at 1,000x g for 3 min at 4 °C. Cells were then 185 washed with PBS and fix ed using 70% ethanol at 4 °C overnight. The following day , the cells 186 were incubated with 1 mL of RNAse solution for 30 min in the dark and stained with PI for 30 187 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint min. The cells were then analyzed for cell cycle arrest using a BD LSR Fortessa Flow Cytometer . 188 The experiment was repeated thrice independently , and the results were analyzed using FlowJo. 189 Immunoblotting 199 Whole-cell protein extracts were prepared using RIP A L ysis Buff er (Pierce Chemical, Rockford, IL, 200 USA) containing protease inhibitor cocktail (Roche, Basel, Switzerland) and phosphatase 201 inhibitor cocktail (Sigma-Aldrich, St. Louis, MO , USA). L ysed samples were centrifuged at 12,000 202 rpm for 40 min, and clarified supernatants were stored at −80 °C. Protein concentrations were 203 determined using the Pierce Bicinchoninic Acid (BCA) protein assay kit. Equal amounts of 204 protein samples were electrophoresed on 4-20% sodium dodecyl sulfate (SDS)- polyacrylamide 205 gels (BIO-RAD , #4568096) and transferred onto PVDF membranes (Invitrogen, #IB34001). The 206 membr anes were then incubated with antibodies diluted in 2% Bovine Serum Albumin (BSA, 207 Fisher Scientific, #BP1600) for 2 hrs at room temperature. Primary antibodies were pERK, ERK, 208 pAKT , AKT , pmT OR, mT OR, pP70s6 kinase, p70s6 kinase, pCREB, CREB, Bcl-2, VEGF A, PDE3B, 209 PDE4C, PDE4D , LC3A/B, cleaved P ARP , cleaved caspase 3, non phospho β-catenin, β-catenin, 210 pV ASP , V ASP , and anti-β-actin. Incubation with HRP-link ed secondary antibodies (CST , 211 #7074/7076;) at a dilution of 1:3000 in a 2% BSA solution was carried out for 1 hr at room 212 temperature. The signal was then detected on a LI-COR Odyssey DLx Imager using the ECL 213 chemiluminescence detection system (Thermo Fisher Scientific, #34577). 214 Measurement of intracellular cAMP and cGMP levels 215 2838c3 and MIA PaCa-2 cells were treated with ADT-030 at varying doses, harvested, and the 216 intracellular cAMP and cGMP levels were measured. Enzyme immunoassay kits were used to 217 detect cAMP (Cat# 581001, Cayman) and cGMP (Cat# 581021, Ca yman) by following 218 manufacturer ’ s instructions. The results were expressed as picomoles/µg of total protein. 219 Immunohistochemistry (IHC) and immunofluor escence (IF) 220 Par affin-embedded tumor tissue slides from 2838c3 and KPC orthotopic studies were used for 221 detecting the expression patterns of extracellular matrix (ECM) remodeling, apoptosis and 222 autophagy mark ers through IHC and IF . H&E-stained sections from lungs, liver , and primary 223 tumor were evaluated by a board-certified veterinary anatomic pathologist (JBF) to quantify the 224 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint number of metastatic lesions in a blinded fashion. For IHC, tumor slides were deparaffinized 225 with xylene for 20 min and rehydrated using 100% and 90% ethanol for 20 min each. The slides 226 were then washed twice with deionized water for 5 min. Antigen retrieval was performed in 10 227 mM citrate bu ffer by microwaving for 10 min followed by two w ashes with deionized water for 5 228 min each. The slides were then quenched in BLO XALL blocking solution (Cat# PK-8200, V ector 229 Labs) for 15 min to block endogenous peroxidase activity and the slides were block ed with 2.5% 230 normal horse serum for 30 min. Primary antibodies were diluted in 2.5% normal horse serum 231 and added to the slides and incubated overnight at 4°C in a humidified chamber . The slides 232 were then washed twice with 1% serum in PBS-T for 10 min each. For IF , the secondary 233 antibodies were diluted in 1% serum in PBS-T and incubated for 2 hrs at room temperature. The 234 slides were washed twice with 1% serum in PBS-T , and nuclear labelling was performed with 235 DAPI containing anti-fade mounting medium. A coverslip was placed and sealed with nail polish. 236 For IHC, the slides were incubated with prediluted biotinylated horse anti-mouse/rabbit IgG 237 secondary antibody for 30 min and washed in PBS-T for 15 min. Then slides were incubated with 238 VECT AST AIN elite ABC reagent for 30 mins and washed in PBS-T for 15 mins. DAB staining was 239 performed until the intensities were reached and then the counterstaining was performed with 240 hematoxylin (cat# 51275, Sigma). The slides were then washed with deionized water and 241 dehydration was performed in 90% and 100% ethanol for 1 min each followed by 1 min in 242 xylene incubation. The slides were then mounted using the mounting medium (Cat# 1442, 243 ePredia). Stained slides were imaged using the Echo Revolution automated microscope (ECHO , 244 USA) at 20× magnification, and quantified using ImageJ (RRID:SCR_003070) with the same 245 threshold for each stain. The results were expressed as percent staining per visual field. 246 Active RAS detection assay 247 RAS activation (RAS-GTP) levels were measured using the active RAS activation assay kit (Cell 248 Signaling Technology, Cat# 8821). L ysates were prepared from cell lines or tumor tissues from 249 various in vivo experiments by performing the steps provided by the manufacturer ’ s protocol. 250 T umors were lysed using the provided lysis buff er supplemented with protease and 251 phosphatase inhibitors. A total of 1 mg/mL lysate in 1x lysis buff er was employed for the 252 experiment. Equal amounts of protein were then incubated with the GST-Raf1-RBD protein, and 253 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint the reaction mixture was loaded onto a RAS affinity resin to capture activated RAS. Following 254 extensive washing to remove unbound proteins, bound protein was eluted in sample buffer and 255 subjected to immunoblotting using a mouse RAS mAb (1:200 dilution) with gentle agitation 256 overnight at 4 °C. The membrane w as then probed with anti-mouse IgG, HRP-link ed antibody 257 (Cell Signaling T echnology , Cat# 7076, RRID:AB_330924; 1:2000), and HRP-conj ugated ant i-258 biotin antibody (Cell Signaling T echnology Cat# 7075, RRID:AB_10696897; 1:1000) to detect 259 biotinylated protein mark ers. Activated RAS levels were measured using chemiluminescent 260 reagents and quantified using the ImageJ system. 261 Pharmacokinetics, tissue distribution and histopathological examination of AD T-030 in mice 262 Pathogen-free 8-week-old f emale C57BL/6J mice (Envigo#044; RRID:IMSR_ENV :HSD-044) were 263 housed in the Biologic Research Labo rato ry at t he Univers ity of South Alabama (U of SA), 264 College of Medicine. Following acclimatization, mice were treated with ADT-030 at a dose of 265 100 mg /kg once daily for 14 days by oral gavage. Bl ood was collected at 0.5, 1, 2, 4, 8, and 24 266 hrs (n=4 per time point) following the last treatment into K 2 EDT A-microtainer tubes (BD 267 Biosciences; Franklin Lak es, NJ; USA) to obtain plasma. Major organs (lungs, kidneys, spleen, 268 heart, liver , brain, colon, and ovaries) were collected at 8 hrs (n=4). ADT-030 levels in plasma 269 and organs were determined by LC-MS/MS. The study followed established guidelines and 270 adhered to the approved protocol of the U of SA Institutional Animal Care and Use Committee 271 (IACUC). 272 In another study , following acclimatization at the University of Alabama at Birmingham (U AB) 273 animal facility , 5-6-week-old male C57BL/6J mice (The Jackson Labo rato ry #000664; 274 RRID:IMSR_JAX:000664) were randomly assigned to two groups (n=5) and received ADT-030 275 (150 mg /kg) by oral gavage for 2 weeks. At the end of the treatment, blood was drawn for 276 serum biochemical analysis, and mice were necropsied, organs were collected and fixed for 277 histopathological analysis, and bone marrow smears were prepared for cytology . A board-278 certified veterinary anatomic pathologist (JBF) performed blinded assessment of organ viscera 279 (heart, lung, kidney , liver , duodenum, pancreas, colon, spleen, thymus, testes, and brain) 280 following standard procedures. The study followed established guidelines and adhered to the 281 approved protocol of the UAB IACUC. 282 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint Open field locomotor activity 283 Mice had ad libitum access to food and water throughout the experiments. Behavioral 284 experiments were performed during the light cycle (between 8 a.m. and 6 p.m.). Before 285 evaluation, mice were habituated for at least 30 min in the testing room. Mice were placed in 286 an open-field arena (44×44×30 cm) in a dimly lit room (7 lx) and allowed to freely explore for 10 287 min as previously described 35 . Locomotor activity was identified as the cumulative distance 288 traveled du ring the entire 10 min. Statistical analyses were performed using GraphPad Prism 289 (V ersion 10.4.1) using Student ’ s t-test for 2 groups and one-way ANOV A for comparing more 290 than 2 treatment groups. Experimenters were blinded to treatment for all comparisons. 291 Orthotopic gr afting of PD AC cells in mice 292 In vivo studies followed established guidelines and adhered to the approved protocol of the 293 UAB IACUC. 4-5-weeks-old male C57BL/6J mice (The Jackson Labo rato ry #000664; 294 RRID:IMSR_JAX:000664) were subjected to isoflurane anesthesia, followed by an intra-295 abdominal incision to access the spleen and pancreas. A matrigel suspension (40 μL), containing 296 KPC- f -luc (1 × 10 5 ), 2838c3-f- luc (1 × 10 5 ), or KPCML1- f -luc (1 × 10 5 ) cells was injected into the 297 pancreas. The skin and abdominal wall were then closed by suturing. Successful engraftment of 298 the tumor cells was confirmed by PerkinElmer IVIS Lumina III In Vivo Imaging System 299 (RRID:SCR_025239) one week later , and mice with tumors were randomized into four groups 300 (n=5 per group) for KPC and 2838c3, and two groups for KPCML1 for monotherapy studies, and 301 five groups for KPCML1 for chemotherapy combination study . Mice in PK C and 2838c3 studies 302 were given oral dosages as follows: the first group received a vehicle, and the other three 303 groups received ADT-030 at varying oral doses (50, 100, and 150 mg /kg). In the KPCM1 study , 304 mice received vehicle or ADT-030 at 150 mg /kg. In the combination study with KPCML1 cells 305 implanted, mice received vehicle, ADT-030 (150 mg/kg), or PF-2545920 (10 mg /kg, IP) daily for 306 4 weeks, gemcitabine (50 mg /kg, IP) plus nab-paclitax el (30 mg /kg, IP), weekly twice (GPT x), and 307 ADT-030 plus GPT x. T umor tracking and response to therapy were monitored using D-lucif erin 308 injection and were conducted bi-weekly throughout the studies. The total luminescence from 309 tumor-bearing regions was quantified using the Living Image in vivo imaging software. Body 310 weights of the animals were measured twice a week. A t termination, all animals were subjected 311 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint to imaging of the whole body , followed by euthanasia, at which time tumors were collected, 312 weighed, and used for subsequent experiments. 313 Single-cell tumor processing 314 FFPE blocks of the tumor tissues from KPC orthotopic experiments treated with vehicle or ADT-315 030 (150 mg /kg) were used for sc-RNA sequencing, as per the standard protocol used at 316 Admera Health (South Plainfield, NJ, USA). After sequencing, the data were analyzed by 317 demultiplexing and aligned to the mouse ref erence genome (GRCm39) for gene expression 318 quantification, and processed with Cell Ranger 9.0.1. The count matrices were then analyzed in 319 Seurat (v .5.3.0) R package v .4.5.1. Cells with more than 200 genes and less than 5% 320 mitochondrial content were k ept for downstream analysis. After SC T ransform analysis, PCA and 321 UMAP were used for dimensionality adjustment and clustering. Diff erentially expressed genes 322 (DEGs) were identified with FindallMark ers, and clusters were labeled using known mark ers. 323 T umor gr owth inhibition studies using PD AC PD X tumors 324 Mouse experiments were conducted using PDAC PD X models with KRAS G12D and KRAS G12 C 325 mutations as described previously 36 . 4-5-week-old male NSG mice (The Jackson Laboratory# 326 005557; RRID:IMSR_JAX:005557) were utilized for the experiments. In brief , F1 generation 327 tumors were cut into 2-mm × 2-mm fragments and subcutaneously implanted through a small 328 incision made in the right flanks of NSG mice while they were anesthetized. T umor size and 329 body weight were monitored biweekly . T umor volume was calculated using the formula: length 330 × width 2 × 0.5. Once tumors reached approximately 80–100 mm 3 , mice (n=5 per group) were 331 randomly assigned into two groups. The first group received vehicle, and the second group 332 received ADT-030 (150 mg /kg). A survival study was conducted for 70 days after 28 days of 333 treatment. The UAB IACUC approved the experimental protocol for these mouse studies. 334 Flow cytometry 335 T umors derived from 2838c3 -f- luc and KPC-f- luc PD AC cells implanted into the pancreas of 336 C57BL/6J mice were digested using a solution containing 0.1 mg /mL DNase 1 and 1 mg /mL 337 collagenase IV (W orthington Biochemical, Lak ewood, NJ) in Hank's Balanced Salt Solution (HBSS) 338 at 37 °C with shaking for 45 min. Following digestion, the samples were rinsed, and enzymes 339 were quenched with RPMI-1640-supplemented with 10% FBS, and a 70 μm strainer was used to 340 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint filter them to produce single-cell suspensions. The separated cells were labeled for 60 min at 4 341 °C using primary antibodies conjugated to fluorophores and a live/dead dye ( Supplementary 342 T able 1 ). Cells were then rinsed and suspended in a F ACS buff er (PBS + 2% FBS). After labeling 343 the cell surface, the cells were fix ed at room temperature in 4% paraformaldehyde or FoxP3 344 transcription buff er set (eBioscience# 00-5523-00) for 45 min, then washed with 1x Perm/W ash 345 (BD , 554723) followed by resuspension in F ACS buff er for data acquisition. Flow cytometry was 346 performed. Data acquisition was conducted using a Symphony A5 flow cytometer , and analysis 347 was performed using FlowJo. 348 Statistical analysis 349 Statistical analyses and data visualization were done using GraphPad Prism (RRID:SCR_005375). 350 The data are represented as means accompanied by either standard deviation (SD) or standard 351 error of the mean (SEM). A repeated measures analysis of variance (ANOV A) or ANOV A with 352 Bonf erroni correction was conducted to evaluate and apply multiple corrections for assessing 353 stat ist ica l signifi c ance between groups. A stati sti cal s ignifi c ance threshold was set at p < 0.05. 354 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint

355 PDE10 over expr ession in PD AC 356 T o investigate the role of PDE10 in PDAC development, we used 10x Genomic sequencing and 357 measured PDE10 expression from 229 patients across 12 diff erent study groups. Employing sc-358 RNA seq data, we found that PDE10 mRNA is significantly enriched in primary tumors and 359 metastatic lesio ns c ompared to tissues from normal donors and adjacent uninvolved tissues 360 (Figure 1A). We then tested the PDAC cell lines used in our experiments and found that all 361 expressed PDE10 protein ( Figure 1B ). 362 ADT-030 inhibition and binding of PDE10 363 ADT-030 is an indene chemically related to the nonsteroidal anti-inflammatory drug, sulindac, 364 designed to block cyclooxygenase (COX) inhibitory activity while targeting PDE10 ( Figure 1C ). 365 The potency of ADT-030 to inhibit the enzymatic activity of recombinant PDE10 was 366 determined by measuring cGMP and cAMP hydrolysis using a fluorescence polarization assay. 367 ADT-030 inhibited cAMP and cGMP hydrolysis with IC 50 values of 0.95 and 1.15 µM, respectively 368 ( Figure 1D ). Molecular modeling studies using the PDE10 (2OUN) structure were performed by 369 induced-fit molecular dynamics and simulation interaction analysis to identify a potential 370 binding site on PDE10 for ADT-030. An optimal GLIDE docking score of -10.3 was calculated with 371 ADT-030 bound in the PDE10 catalytic domain with the tri-methoxy benzylic moiety oriented 372 toward the deep hydrophobic region of the pocket, while the more polar substituents were 373 oriented towards the pocket entrance ( Supplementary Figure 1A ). The amide carbonyl and 374 neighboring heteroatoms are predicted to form hydrogen bonds with His525 through bridging 375 water molecules ( Supplementary Figure 1B ). Hydrophobic and aromatic contacts with residues 376 Tyr524, Leu635, Phe639, Ile692, Tyr693, Phe696, Met714, Phe729, Val733, and Ala734 help to 377 anchor the aromatic system within the binding site. These results suggest a favorable binding 378 for ADT-030 in the PDE10 catalytic domain, which is supported by the docking score and a 379 network of direct and water-mediated interactions, and consistent with a competitive 380 mechanism of enzyme inhibition as previously reported for an analog, ADT-061 30 . 381 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint We next evaluated the antiproliferative activity of ADT-030 against a series of PDAC cell lines 382 harboring various KRAS mutations, as well as wild-type RAS, by performing cell viability 383 measurements using the MTT assay and determining potency (IC 50 ) values. As shown in Figure 384 1E, ADT-030 inhibited the proliferation of all KRAS G12D and KRAS G12C mutant PDAC cell lines 385 tested with IC 50 values in the low micromolar range (1.8-4.5 µM). Notably, the KRAS wild-type 386 PDAC cell line, BxPC-3, was found to be essentially insensitive to ADT-030, suggesting that ADT-387 030 selectively inhibits the proliferation of KRAS mutant PDAC cells ( Figure 1E ). 388 Experiments were also conducted to confirm that ADT-030 binds PDE10 in intact cells. In brief , 389 HEK-293 cells expressing PDE10 Micro-T ag were treated with ADT-030. PDE10 thermal stability 390 was measured by Micro-T ag enzyme compleme ntat ion as described in the Materials and 391

section ( Supplementary Figure 1C-F ). The results revealed an EC 50 value of 0.9 µM for 392 ADT-030 to bind PDE10, which paralleled the potency ranges of ADT-030 to inhibit the 393 enzymatic activity of PDE10 and the prolif eration of PDAC cells. 394 ADT-030 inhibits the clonog enicity and mi gration of KRAS mutant PDAC cells 395 A mouse PD AC cell line, 2838c3 (KRAS G12D mutant), and a human PDAC cell line, MIA 396 PaCa-2 (KRAS G12C mutant), were selected to further study the anti-cancer activity of ADT-030. 397 The long-term inhibitory eff ect of ADT-030 on cancer cell survival was evaluated in both PDAC 398 cell lines by colony formation assays. ADT-030 treatment significantly reduced the number and 399 size of colonies in both PDAC cell lines across a concentration range comparable to the potency 400 (IC 50 ) values to inhibit prolif eration ( Figure 1F-G ). In addition, 2838c3 and MIA PaCa-2 cells 401 showed significant impairment in motility after treatment with ADT-030 at non-cytotoxic 402 concentrations ( Figure 1H - I). T ogether , these results indicate that ADT-030 inhibits the 403 prolif eration, colony formation, and motility of PDAC cell lines harboring KRAS G12D and KRAS G12 C 404 mutations within the same concentration range as that is required to inhibit recombinant PDE10 405 and bind PDE10 in cells. 406 ADT-030 induces apoptosis and G2/M cell cycle arr est in PD AC cells 407 T o determine the eff ect of ADT-030 on apoptosis and cell cycle progression, the 2838c3 and MIA 408 PaCa-2 PDAC cell lines were treated with ADT-030 for 24 and 72 hrs, respectively . Flow 409 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint cytometry analysis of apoptosis as measured by Annexin V /PI staining showed that ADT-030 410 increased both early and late apoptotic cells at concentrations of 2 and 5 µM. In vehicle-treated 411 2838c3 cells, apoptotic cells comprised 1.6% of the population, whereas treatment with ADT-412 030 increased the percentage to 3.5% (2 µM) and ~8% (5 µM) ( Supplementary Figures 2A and 413 2C). For MIA PaCa-2 cells, vehicle-treated cells had 2.3% apoptotic cells within the population, 414 whereas ADT-030 treatment increased the number to 14.5% (2 µM) and 32% (5 µM) 415 ( Supplementary Figur es 2B and 2 D ). Analysis of cell cycle distribution revealed that ADT-030 416 treatment increased the percentage of cells arrested in the G2/M phase in both PDAC cell lines 417 ( Supplementary Figures 2E-H ). 418 ADT-030 inhibits PDE10 and activates PKA/PK G signaling 419 Since PDE10 inhibition by ADT-030 is expected to increase intracellular levels of cAMP and 420 cGMP , we measured both levels in 2838c3 and MIA PaCa-2 PD AC cell lines following treatment 421 with ADT-030 by ELISA. Consistent with a PDE10 inhibitor , ADT-030 significantly increased the 422 levels of cAMP and cGMP in a concentration-dependent manner in both PDAC cell lines. 423 Notably , the eff ect was apparent at concentrations that paralleled the concentration range 424 eff ective for inhibiting recombinant PDE10 and prolif eration of both PDAC cell lines, as well as 425 for inducing cell cycle arrest and apoptosis ( Figure 2A-D ). T o determine if increased cyclic 426 nucleotide levels by ADT-030 activated downstream protein kinases PKA and PK G, in 2838c3 427 and MIA PaCa-2 cells, we measur ed the phosphorylation of V ASP (vasodilator-stimulated 428 phosphoprotein), a known subst rate for PKA and PK G 37 . Consistent with a PDE10 inhibitor , ADT-429 030 increased V ASP phosphorylation in both PD AC cell lines ( Figure 2E-F) , demonstrating that 430 ADT-030 induced elevation of intracellular cAMP and cGMP levels results in the activation of 431 PKA and PK G. Finally , it should be noted that ADT-030 treatment did not aff ect the expression o f 432 PDE10 ( Figur e 2E ) or other cGMP and cAMP-degrading PDE isozymes, PDE3 or PDE4, 433 respectively ( Supplementary Figur e 3A ). W e also determined if ADT-030 can activate canonical 434 downstream signaling from PKA/PK G activation commonly reported in normal cells. T reatment 435 of 2836c3 and MIA PaCa-2 cells with ADT-030 did not have any eff ect on level of activated 436 phospho-CREB, VEGF-A, and Bcl-2 ( Supplementary Figur e 3B-C ). This suggests that the 437 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint activation of PKA and/or PK G by AD T-030 in PD AC cells may involve downstream targets unique 438 to cancer cells. 439 ADT-030 attenuates RAS and β-catenin signaling to promot e apoptosis 440 Previous reports suggest that known PDE10 inhibitors and activation of PKG can phosphorylate 441 the oncogenic pool of β-catenin in cell lines from various cancers 2 5, 27, 29, 30 . W e therefore 442 performed W estern blot analysis to measure levels of the unphosphorylated (stable) form of β-443 catenin, representing the oncogenic pool of β- caten in requ ired for TCF/LEF tr anscriptional 444 activity in PDAC cells treated with ADT-030. AD T-030 treatment significantly reduced levels of 445 the unphosphorylated form of β-catenin in 2838c3 and MIA PaCa-2 cell lines ( Figure 2G ). 446 Previous research also suggested that PDE10 inhibitors and activation of PKG could suppress 447 MAPK and AKT signaling in lung and ovarian cancer cells 25 , 27 . Hence, we determined if ADT-030 448 has a similar eff ect in PDAC cells by measuring phosphorylated levels of ERK (pERK) and mT OR 449 (pmT OR) within the MAPK and AKT signaling nodes, respectively . ADT-030 decreased pERK 450 levels at its activating phosphorylation sites, Thr 20 2 and T yr 20 4 , as well as pmT OR at its activation 451 site (Ser 24 48 ) ( Figure 2G ). These results suggest that the PDE10 inhibitory activity of ADT-030 can 452 simultaneously suppress RAS and β-catenin signaling in PDAC cells. 453 T o further study the eff ects of ADT-030 on RAS signaling, RAS-GTP pulldown assays were 454 performed to measure activated RAS levels following the treatment of PDAC cells with ADT-030 455 for 24 hrs at concentrations of 2 and 5 µM. ADT-030 treatment did not reduce activated RAS 456 levels in KRAS wild-type BxPC-3 and P anc 02 cells, as well as in KRAS amplified KLE (endometrial 457 adenocarcinoma) and MKN1 (gastric adenocarcinoma) cancer cell lines ( Figure 2H - I and 458 Supplementary Figure 4A-D ). Conversely , ADT-030 reduced activated RAS levels in 2838c3 cells 459 expressing KRAS G12D and MIA PaCa-2 cells expressing KRAS G12C mutations at concentrations 460 eff ective for inhibiting prolif eration and PDE10 ( Figure 2J and Supplementary Figure 4E-F ). 461 These results suggest that ADT-030 c an inhibit activated RAS levels in KRAS-mutant PDAC cells 462 but not in KRAS wild-type or KRAS amplified cells. This may explain the selective growth 463 inhibitory activity of ADT-030 observed between RAS-mutated and RAS-wild-type PDAC cells. 464 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint W e also analyzed the eff ect of ADT-030 treatment on autophagy in 2838c3 and MIA PaCa-2 465 PDAC cell lines, given previous studies reporting that RAS signaling and autophagy are 466 interconnected and that RAS can modulate autophagy to promote tumorigenicity 38 . Autophagy 467 was assessed by W estern blotting using LC3A/B as a mark er . ADT-030 treatment increased the 468 expression of LC3A/B, indicating its capacity to disrupt autophagic flux ( Figure 2G ). The cells 469 were also treated with ADT-030 alone or in combination with hydroxychloroquine (HCQ), a 470 known autophagy inhibitor . ADT-030 in combination with HCQ did not increase the levels of 471 LC3A/B, suggesting that ADT-030, lik e HCQ , inhibits autophagic flux ( Supplementary Figure 5A ), 472 and are consistent with a previously reported analog of ADT-030 39 . Furthermore, ADT-030 473 treatment reduced the expression of p70s6 kinase in 2838c3 and MIA P aCa-2 cells, which is 474 associated with reduced mig rato ry ca pacity (Figure 2G ) 40 . 475 Hematologic, clinical chemistry , histopathologic and behavioral assessment of ADT-030 476 treated mice 477 Our next objective was to evaluate the tolerance of mice to ADT-030 treatment. T en mice were 478 randomly assigned to vehicle (n=5) or ADT-030 (150 mg /kg, n=5) treatment by oral gavage for 479 two weeks. Complete blood counts (WBC, RBC, HGB, HCT , MCV , MCH, MCHC, RDW , PL T , MPV , 480 neutrophils, lymphocytes, monocytes, eosinophils, and basophils) and serum biochemistry 481 (albumin, AL T , ALP , amylase, total bilirubin, BUN, phosphorus, creatinine, glucose, electrolytes 482 (calcium, sodium and potassium), total protein, and globulin were measured following two 483 weeks of treatment (Supplementary Figure 6A-B) . W e observed no significant differences 484 between the vehicle and ADT-030-tr eated mice, except for a slight reduction of total bilirubin 485 levels in the treatment group. In addition, gross examination of multiple organs, including lungs, 486 liver , kidney , pancreas, heart, duodenum, colon, spleen, thymus, brain, and testis, showed no 487 histopathological abnormalities in ADT-030-treated mice compared with vehicle-treated mice 488 (Figure 3A-B) . 489 Sedation is a well-known side effect of conventional PDE10 inhibitors that were designed to 490 cross the blood-brain barrier and developed for the treatment of CNS disorders (schizophrenia 491 and Huntington’ s disease). W e therefore performed an open-field locomotor test to determine 492 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint if ADT-030 causes sedation. These behavioral experiments revealed no significant diff erences 493 between vehicle and ADT-030-treated mice, suggesting that ADT-030 does not cause sedation, a 494 common side eff ect from previously developed PDE10 inhibitors (Supplementary Figure 7A) . 495 Pharmacokinetics and tissue distribution of ADT-030 496 A PK study was conducted in f emale C57BL/6J mice after oral gavage administration of 100 497 mg /kg ADT-030 once daily for 14 days. ADT-030 generated plasma levels that exceeded those 498 required to inhibit PDE10 and PD AC cell growth in vitro ( Supplementary Figur e 7B ). Plasma 499 levels of ADT-030 reached a Cmax of 7 µM by 1 hr post-treatment and remained unchanged for 500 an additional hour before decreasing by 4 hrs post-treatment to the level (5 µM). High levels of 501 ADT-030 were also detected in various organs (lungs, kidneys, spleen, heart, liver , ovaries, and 502 colon) 8 hrs after administration, but low levels were measured in brain ( Supplementary Figure 503 7C). The low concentration of ADT-030 measured in the brain following oral administration 504 lik ely account for the absence of sedation, a known side eff ect of conventional PDE10 inhibitors 505 developed for the treatment of CNS disorders that were designed to cross the blood-brain 506 barrier to achieve high concentrations in the brain 41 . 507 ADT-030 suppresses tumor growth in orthotopic PD AC model and r ep rograms the TiME 508 A mouse model of PDAC involving orthotopically implanted 2338c3- f-luc PDAC cells in the 509 pancreas was initially used to assess the in vivo antitumor activity of ADT-030. Mice that 510 established palpable tumors one week following implantation were randomized into four 511 groups and treated by oral g avage administration with vehicle or ADT-030 at dosages of 50, 100, 512 and 150 mg /kg once daily 5x/week for 23 days. T umor progression was monitored using 513 bioluminescence imaging ( Figure 3C ). All dosages of ADT-030 were eff ective, with the highest 514 dose of ADT-030 tested showing tumor regression in all mice in the group. Quantitation of 515 bioluminescence confirmed a significant reduction in tumor mass in ADT-030-treated mice 516 compared to vehicle treatment ( Figure 3D ). Both tumor images and tumor weight 517 measurements confirmed tumor shrinkage in ADT-030-treated groups in a dose-dependent 518 manner compared to vehicle treatment ( Figure 3E-F and Supplementary Figure 8A-B ). AD T-030 519 treatment did not cause apparent systemic toxicity , as evidenced by no effect on body weight 520 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint gain during treatment, suggesting the potential for greater efficacy at higher dosages or a more 521 frequent dosing schedule ( Figure 3G ). 522 T o study the immunomodulatory eff ects of ADT-030 relevant to PDAC, we performed 523 multiparametric flow cytometry on orthotopic 2838c3 tumors from vehicle and ADT-030-524 treated mice. The analysis revealed that ADT-030 induced profound shifts in immune cell 525 composition in the TiME, favoring a more immunostimulatory phenotype. ADT-030 treatment 526 enhanced the immune cell infiltration within the TiME, resulting in a significant increase in 527 overall populations of CD45 + leuk ocytes compared with vehicle treatment ( Supplementary 528 Figure 9A ). Further characterization of the T-cell compartment revealed an increase in CD3 + T 529 cells ( Supplementary Figure 9B ), observed with both CD4 + ( Supplementary Figur e 9C ) and CD8 + 530 T cells ( Supplementary Figur e 9D ). T reatment with ADT-030 induced higher levels of several 531 immune checkpoint mark ers, including CTLA-4 ( Supplementary Figure 9E ), PD-1 532 ( Supplementary Figur e 9F ), LAG-3 ( Supplementary Figur e 9G ), and TIGIT ( Supplementary 533 Figure 9H ) in the total T cell populations. In another experiment involving the KPC -f- luc model, 534 the immune checkpoint mark ers were diff erentially regulated with a decrease in CD8 + T c e l l s , 535 and an increase in CD4 + T cells. These results indicate that there was a concurrent adaptive 536 immune regulatory response, lik ely representative of an acute but low , probably exhausted, T 537 cell phenotype within the TiME. Apart from modulating the T cell compartment, ADT-030 also 538 elevated NK cell infiltration within the TiME. Flow cytometry results revealed a pronounced 539 increase in the NK1.1 + cell population in the tumors of the ADT-030-treated mice compared to 540 vehicle-treated mice, indicating enhanced activation of the innate immune system 541 ( Supplementary Figur e 9I ). In addition to augmenting NK cell numbers, increased expression of 542 immune checkpoint receptors was measured on NK cells by ADT-030 treatment. These 543 observations were paralleled within the CD3 + T cell population, where ADT-030 enhanced 544 infiltration of CD4 + and CD8 + T cells while upregulating CTLA-4, PD-1, LAG-3, and TIGIT 545 ( Supplementary Figure 9E-H ). These data demonstrate that ADT-030 has the capacity to broadly 546 remodel the immune landscape, attracting both adaptive and innate eff ector cells into the 547 tumor while simultaneously engaging checkpoint regulatory pathways. 548 ADT-030 alters RAS and β-catenin signaling in tumor s 549 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint T o determine the activity of ADT-030 treatment to inhibit oncogenic signaling in vivo , 2838c3 550 tumors were har vested from mice treated with vehicle or ADT-030 and analyzed for k ey 551 signaling and apoptosis mark ers. W estern blot analysis revealed a mark ed decrease in pERK 552 ( Figure 4A-B ) and pAKT ( Figure 4A and 4 C ) levels with no effect on total ERK or AKT levels in 553 tumors from ADT-030-treated mice, indicating the concurrent inhibition of the MAPK and 554 PI3K/ AKT pathw ays, respectively , reflective of upstream RAS inhibition. In addition, ADT-030 555 treatment reduced levels of the non-phosphorylated form of β-catenin, indicative of the stable 556 pool of β-catenin driving transcription of proteins involved in oncogenesis, for example, from 557 aberrant activation of WNT signaling ( Figure 4A and 4 D ). Along with these alterations in k ey 558 signaling node proteins, ADT-030 treatment also increased the expression of LC3A/B, consistent 559 with in vitro experiments, indicating that ADT-030 inhibits autophagic flux ( Figure 4A and 4 E ). 560 Increased levels of cleaved P ARP ( Figure 4A and 4 F ) and cleaved caspase 3 ( Figure 4A and 4 G ) 561 were also observed in the ADT-030-treated group, again consistent with in vitro experiments 562 showing apoptosis induction by ADT-030 treatment. Also consistent with in vitro experiments, 563 ADT-030 treatment reduced levels of activated (GTP-bound) RAS as measured by RAS-RBD 564 pulldown assays ( Figure 4H and Supplementary Figur e 9J ). IHC showed significantly reduced 565 expression of the prolif eration mark er , Ki-67, in tumors from AD T-030-treated mice, 566 corroborating the antiprolif erative activity of ADT-030 as observed in vitro ( Figure 4I and 4 M ). IF 567 microscopy evaluation was used to analyze the treatment impact on autophagy and 568 mesenchymal-to-epithelial transition (MET). Mice treated with ADT-030 at 150 mg /kg showed 569 an increased level of LC3A/B in tumors, indicative of a disruption of autophagic flux ( Figure 4J 570 and 4 N ). ADT-030 also reduced vimentin expression ( Figure 4K and 4 O ) and increased 571 expression of E-cadherin ( Figure 4L and 4 P ), which are associated with MET , and indicative of 572 transforming cancer cells to a more normal epithelial phenotype with lower invasive capability . 573 ADT-030 enhances antitumor immune r esponses and inhibits metastasis in an orthotopic 574 PD AC model 575 T o confirm the anti-tumor eff ects of ADT-030 observed in the orthotopic 2838c3 tumors, we 576 evaluated ADT-030 in the KPC orthotopic mouse model of PDAC to study specific immune cell 577 subsets and functional T cell responses. T o accomplish this, we implanted KPC -f- luc cells in 578 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint C57BL/6J mice followed by treatment with ADT-030 (50, 100, and 150 mg/kg) or vehicle by oral 579 gavage. T u mor growth was monitored by bioluminescence imaging on Day 0 and on Day 23 580 prior to euthanasia ( Figure 5A ). The KPC model recapitulated the major findings from the 581 2838c3 model. Normalized bioluminescence intensities showed statistically significant 582 diff erences between the vehicle and treatment groups on Day 23 ( Figure 5B ). Both tumor size 583 ( Figure 5C) and tumor weights ( Figure 5D ) at the end of the experiment showed substantial 584 shrinkage in a dose-dependent manner in the treated groups compared to the vehicle-treated 585 group. W e then performed IHC analysis for prolif eration using the Ki-67 antibody in tumor 586 sections from the KPC orthotopic model. T umors treated with ADT-030 showed a mark ed 587 decrease in the number of Ki-67 positive cells compared to the vehicle group, confirming the 588 anti-prolif erative activity of ADT-030 in vivo ( Supplementary Figures 10A and 10E ). T o further 589 substantiate the above findings, we performed immunofluorescence microscopy with the tumor 590 tissues using autophagy (LC3A/B) and the MET mark ers, E-cadherin and vimentin. These 591 analyses showed a significant increase in LC3A/B, suggestive of disrupted autophagy flux 592 ( Supplementary Figures 10B and 10F ) and elevated levels of E-cadherin ( Supplementary 593 Figures 10C and 1 0 G ), along with a decrease in vimentin expression, reflective of MET 594 ( Supplementary Figures 10D and 10H ). 595 ADT-030 promotes anti-tumor immunity in a mouse model of PD AC 596 W e then performed multiparametric flow cytometry on the excised KPC- f -luc tumors to assess 597 the immune changes underlying ADT-030 treatment. A significant elevation in overall immune 598 infiltration was observed, following ADT-030 treatment, as determined by increased frequency 599 of CD45 + cells ( Supplementary Figur e 11A ). Among the infiltrating pool of immune cells, there 600 was an increase in αβ + T c e l l s ( Supplementary Figure 11B ), γδ + T c e l l s ( Supplementary Figur e 601 11C ), TNK cells ( Supplementary Figure 11D ), and conventional NK cells ( Supplementary Figur e 602 11E ), suggesting the activation of a broad-nature innate and adaptive immune response. The 603 reduced expression of immune checkpoint molecules PD-1 ( Supplementary Figure 11F ) and 604 CTLA-4 ( Supplementary Figure 11G ) on NK1.1 + cells, indicates NK cell exhaustion. In contrast, a 605 mark ed increase in the frequencies of CD4 + T cells ( Supplementary Figure 11H ), as well as 606 higher expression levels of PD-1, TIGIT , CTLA4 + , PD-1 + CTLA4 + and F ASr subsets ( Supplementary 607 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint Figure 11I ), were observed in the CD4 + T cell compartment of ADT-030-treated tumors. The 608 number of effec tor CD4 + T cells increased in the ADT-030-treated mice, signifying the 609 involvement of helper T cell activation and functionality diff erences ( Supplementary Figure 610 11J ). Similar significant increases were present in overall CD8 + T cell numbers after ADT-030 611 treatment ( Supplementary Figur e 11K ). In comparison, we measured decreased expression of 612 immune checkpoint mark ers, including PD-1, CTLA-4, PD-1 + CTLA-4 + , and LAG-3, as well as lower 613 expression of PD-1 + CTLA-4 + LAG-3 + triple-positive subsets ( Supplementary Figur e 11L ) in the 614 CD8+ T cell compartment. These data indicate a reversal of T cell exhaustion and re-engagement 615 of cytotoxic potential. In addition, increased eff ector CD8 + T cells were observed with ADT-030 616 treatment, reflecting improved anti-tumor immunity ( Supplementary Figure 11M ). 617 Based on these results, we further explored the myeloid cell compartment in the TiME after 618 ADT-030 treatment. W e observed an enhanced influx of myeloid cells, as evidenced by 619 increased total macrophages (F4/80 + ) ( Supplementary Figure 11N ). Another characteristic 620 indicating myeloid infiltration was the increased expression of PD-L1 on macrophages after ADT-621 030 treatment, thereby enhancing antigen presentation and potential interaction with eff ector 622 T cells ( Supplementary Figur e 11O ). Phenotypic characterization of macrophages showed an 623 increased M1-type characterized by MHC-II + CD86 + being more frequently expressed, while M2-624 lik e macrophages (CD206 + ) were less frequent from ADT-030 treatment compared to vehicle 625 ( Supplementary Figures 11P-Q ). Additionally , an increased M1/M2 ratio was observed after 626 ADT-030 treatment, signifying an enhanced immune-stimulatory TiME ( Supplementary Figure 627 11R ). Moreover , there was also an increase in the overall frequency of dendritic cells (DCs) post 628 ADT-030 treatment ( Supplementary Figure 11S ). Both conventional subsets of dendritic cells, 629 cDC1, and cDC2, were increased in frequency , suggesting enhanced antigen processing and 630 presentation ( Supplementary Figur es 11T-U ). This further rise in functional antigen-presenting 631 cells, together with T and NK cell infiltration and activation, shows the wide immunomodulatory 632 capacity of ADT-030 in remodeling of the pancreatic TiME toward an anti-tumor immune state. 633 Single cell RNA-seq (scRNA-seq) of orthotopic KPC tumors tr eated with ADT-030 634 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint W e also performed scRNA-seq on the resected tumors from this orthotopic KPC mouse 635 experiment to further investigate the eff ects of ADT-030 on signaling pathways and immune 636 microenvironment. A uniform manifold approximation and projection (UMAP) analysis was 637 performed ( Figure 6A ), including 10 diff erent cellular populations identified as pericytes, 638 gMDSCs, CAFs, endothelial cells, m yocytes, T NK and B cells, dendritic cells, macrophages, 639 acinar to ductal metaplasia (ADM), and PD AC ( Figur e 6A-B ). These populations were identified 640 based on the expression of canonical mark er genes for mature terminal lineages 641 ( Supplementary Figur es 12A-B ). W e then identified 7 PDAC sub-clusters ( Figure 6C-D ) in which 642 pathway analysis revealed that ADT-030-treated mice had significant downregulation in EMT , 643 apical junction, KRAS signaling, and m yogenesis signaling ( Figure 6E ). W e then focused on MAPK 644 signaling, as this pathway plays a major role in driving PDAC. In ADT-030-treated mice, there 645 was a significant reduction in the expression of Raf1, a downstream mediator of activated RAS, 646 suggesting the functional downregulation of the RAS-MAPK pathway in response to ADT-030 647 treatment ( Figure 6F-H ). Although an increase in upstream RAS signaling was observed, the 648 MAPK signaling flux analysis revealed a significant reduction in the expression of Raf1 and 649 Mapk3 ( Figure 6I-J ), suggesting that downstream RAS signaling was completely inhibited. 650 Furthermore, deep analysis revealed a reduction in the expression of several MAPK pathway 651 genes, including Map2k2, Mapk3, Dusp6, and Elk4 ( Figur e 6K ). W e then analyzed the EMT 652 pathway and found the concurrent reduction in the expression of mesenchymal mark ers such as 653 vimentin and fibronectin 1 (FN1) ( Figure 6L-M ). Next, we analyzed the impact of ADT-030 654 treatment on WNT signaling and found that WNT pathway mark ers were also suppressed 655 ( Figure 6N ), including APC, AXIN2, Lrp5 and Lrp6 ( Figure 6O ). Mechanistic investigation 656 confirmed that ADT-030 treatment reduced expression of several MAPK and WNT pathway 657 genes, supporting the similar observations at the protein level ( Figure 6P ). 658 Next, we focused on identifying the role of ADT-030 on the immune microenvironment. T o this 659 end, we sub-clustered the UMAP into four groups, including CD8 T cells, TNK cells, T regs and NK 660 cells ( Figure 6Q ). A concurrent increase in the TNK cells was identified after ADT-030-treated 661 mice, suggesting that ADT-030-treatment may enhance anti-tumor immune responses ( Figure 662 6R). Although the total number of CD8 T cells was reduced, higher numbers of activated CD8 T 663 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint cells were present after ADT-030-treated mice compared to vehicle-treated mice ( Figure 6S ). 664 W e also analyzed several mark ers of CD8 T cell activation, exhaustion, and stem-lik e properties. 665 A significant increase in activation mark ers, including CD69, Prf1, IFNγ, and granzyme A (Gzma) 666 was observed in ADT-030-treated mice. These findings show that ADT-030 increases CD8 T cell 667 activation and enhances cytotoxic potential towards an eff ector state despite an overall 668 reduction in total CD8 T cell numbers ( Figure 6T ). E valuation of the transcriptomic signature of 669 CD8 T cells clearly showed that ADT-030 treatment increased the expression of several early-670 activation genes, eff ector diff erentiation factors, cytotoxicity mediators, and chemokine ligands 671 in ADT-030-treated mice, suggesting increased CD8 T cell functionality ( Figure 6U ). 672 ADT-030 treatment induced a similar but broader remodeling of the TNK compartment, 673 extending the CD8 T-cell-specific effects to encompass both T and NK cells within the TiME 674 ( Figure 7A ). In line with enhanced activation and reduced dysfunction of CD8 T cells, the TNK 675 global state trajectory demonstrated that ADT-030 shifted TNK cells towards higher pan-676 activation scores with relatively lower pan-dysfunction scores compared to vehicle treatment, 677 indicating a coordinated reinforcement of an activated, less dysfunctional state across cytotoxic 678 lymphocytes. This was accompanied by increased expression and prevalence of k ey eff ector and 679 activation genes such as Gzmb, Nkg7, and Prf1 ( Figur e 7B-C ). Additionally , ADT-030 treatment 680 induced a mark ed shift in NK cell functional state toward an activated phenotype compared to 681 vehicle treatment (Supplementary Figure 12C) . NK cell trajectory analysis revealed that ADT-682 030-treated mice showed higher activation and lower dysfunction signatures, indicating 683 coordinated enhancement of activation programs. Concordantly , dot-plot analysis revealed 684 increased expression and prevalence of activation and maturation mark ers, including Zeb2, Bcl2, 685 Klrg1, Itgam, Cd160, Havcr2, Prf1, IFNγ, and Gzmb (Supplementary Figures 12D-E) . T ogether , 686 these results demonstrate that ADT-030 enhances cytotoxic lymphocyte activation within the 687 TiME, driving CD8 T and TNK compartments towards a sustained, less dysfunctional eff ector 688 state to enhance anti-tumor immunity . 689 E ffects of ADT-030 on metastasis 690 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint T o invest igate th e po tent ial o f ADT-030 to bloc k metastasis, we used an established metastatic 691 PDAC cell line, KPCML1, derived from the KPC mouse model 42 . KPCML1 cells have a high 692 propensity for liver and lung metastasis, representative of patients with meta static pancreatic 693 cancer 42 . Using similar tumor inoculation methods and treatment (vehicle vs. AD T-030 at 150 694 mg /kg daily) as described in the Materials and Methods section, we examined the impact of 695 ADT-030 treatment on mice orthotopically implanted with KPCML1 P DAC cells. On day 23 afte r 696 tumor impla ntati on, ADT- 030 treatment decreased the size and weight of the primary 697 orthotopic tumor compared to vehicle treatment ( Figure 5E-F ). Lucif erase levels were measured 698 using ex vivo imaging, which revealed that while mice in the vehicle group implanted 699 orthotopically with KPCML1 cells developed liver and lung metastasis, there was a complete 700 absence of liver ( Supplementary Figures 13A-B ) and lung ( Supplemen tary Figures 13C-D ) 701 metastasis in mice treated with ADT-030. W e also analyzed lung and liver sections histologically 702 by H&E staining, which confirmed metastasis in the vehicle group and supported the obser ved 703 anti-meta stat ic act ivit y of AD T-030 ( Figure 5G - J). 704 ADT-030 enhances the antitumor efficacy of chemotherapy in orthotopic PD AC models 705 KPCML1 cells were implanted orthotopically and treated with vehicle, ADT-030 (150 mg /kg), 706 standard-of-care chemotherapy (a combination of gemcitabine, 50 mg /kg and nab-pacli taxe l, 10 707 mg /kg, GPT x), or a combination of ADT-030 with GPT x. ADT-030 produced a better therapeutic 708 eff ect than chemotherapy as indicated by tumor size and tumor weight measurements ( Figure 709 7D-F ). Interestingly , the combination of ADT-030 with chemotherapy showed better efficacy 710 compared to chemotherapy or ADT-030 alone, demonstrating that ADT-030 has the potential to 711 enhance standard-of-care chemotherapy efficacy for the treatment of PDAC. 712 ADT-030 has increased potency and improv ed therapeutic window compared to other PDE10 713 inhibitors 714 PF-2545920 is a known PDE10 inhibitor developed for CNS disorders such as schizophrenia and 715 Huntington’ s disease. The potency IC 50 values for PF-2545920 to inhibit the prolif eration of MIA 716 PaCa-2 and 2838c cells were measured to be 25.02 and 24.1 µM, respectively ( Supplementary 717 Figure 14A ) compared with AD T-030 having IC 50 values of 3.01 and 1.79 µM, respectively 718 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint ( Figure 1E ). Similarly , colony formation assays revealed that PF-2545920 had IC 50 concentrations 719 exceeding 25 µM ( Supplementary Figure 14B-C ) whereas AD T-030 showed significant inhibition 720 of colony formation at 5 µM ( Figure 1G-H ). 721 W e then compared PF-2545920 with ADT-030 in the KPCML1 orthotopic mouse model of PDAC. 722 Mice were treated with vehicle, ADT-030 (150 mg/kg), or PF-2545920 (10 mg /kg) in which each 723 were once daily by oral gavage. PF- 2545920 failed to show antitumor activity , while ADT-030 724 significantly inhibited tumor growth as evidenced by tumor images and measurement of tumor 725 weight ( Figure 7D-F ). Additionally , we evaluated the eff ects of PF-2545920 and ADT-030 726 treatments on mice using open-field locomotor tests. There were no significant diff erences in 727 behavior or mobility between ADT-030 and vehicle-treated mice, whereas PF-2545920-treated 728 mice displayed significantly reduced mobility throughout the test, reflective of sedation, a 729 known side eff ect of conventional PDE10 inhibitors ( Supplementary Figur e 15A ). These data led 730 us to conclude that ADT-030, but not PF-2545920 displays antitumor activity without causing 731 sedation, which lik ely reflects differences in brain and systemic levels between ADT-030 and 732 known PDE10 inhibitors. 733 ( Supplementary Figur e 15A ). This data led us to conclude that ADT-030 displays superior 734 antitumor activity compared to a known PDE10 inhibitor without causing sedation. 735 ADT-030 suppr esses tumor gr owth and induces prolong ed r esponses in KRAS G12D and 736 KRAS G12C PD AC PD X models 737 T o further evaluate the efficacy of ADT-030 in suppressing PDAC tumor growth in vivo , we used 738 two clinically annotated subcutaneously implanted PDX models of PDAC harboring KRAS G12D and 739 KRAS G12C mutations. ADT-030 was administered orally at a dose of 150 mg/kg for 23 days, once 740 daily , 5x/week. T umor dimensions and body weight were measured twice/week. The results 741 displayed a strong anti-tumor response in both KRAS G12D ( Figure 8A ) and KRAS G12C PDX models 742 ( Figure 8D ) with no apparent systemic toxici ty in terms of reduction in body weight ( Figure 8B 743 and 8 E ). The treatment was stopped after four weeks, and mice were monitored for tumor 744 recurrence and survival. Mice treated with ADT-030 did not develop tumor regrowth over 70 745 days of follow-up ( Figure 8C and 8 F ), while vehicle-treated mice progressively died during the 746 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint post-treatment period. These data demonst rate bo th r obust and durable antitumor activity of 747 ADT-030 in clinically relevant PDX models of PDAC. 748 ADT-030 suppresses the growth of KRAS G12D and KRAS G12C - resistant PD AC cells 749 Finally , we tested the anti-proliferative activity of ADT-030 in KRAS G12D and KRAS G12C resi stan t 750 cell lines grown in vitro . AsPC-1 cells resistant to MRTX1133 (AsPC-1-MRTX-R) and parental cells 751 were treated with ADT-030 or MRTX1133, a KRAS G12D inhibitor 43 . Cell viability measurements 752 using the MTT assay revealed that ADT-030 showed comparable antiprolif erative activity in both 753 parental (IC 50 = 1.75 µM) and AsPC-1-MRTX-R cells (IC 50 = 1.47 µM) ( Figure 8G-I ) , whereas 754 MRTX1133 inhibited the prolif eration of parental AsPC-1 cells (IC 50 = 43.74 nM), but not of the 755 AsPC-1-MRTX-R cells (IC 50 > 25 µM), confirming that these cells are resistant to MRTX1133. W e 756 also determined if this eff ect is sustained with longer treatment durations by performing colony 757 formation assays. The results showed activity of ADT-030 similar as proliferation assays, where 758 ADT-030 inhibited colony formation in both parental and AsPC-1-MRTX-R cells, while MRTX1133 759 inhibited colony formation only in parental cells ( Supplementary Figur es 16A-B ). W e also 760 treated MIA PaCa-2 G12C parental and MIA PaCa-2 resistant to MRTX849 and AMG-510 (MIA-761 AMG-R) cells with ADT-030 and MRTX849, a KRAS G12C inhibitor 44 . ADT-030 inhibited the 762 prolif eration of both MIA PaCa-2 parental and MIA-AMG-R cells with comparable potency . In 763 contrast, MRTX849 inhibited the prolif eration of MIA PaCa-2 parental cells, but not MIA-AMG-R 764 ( Figure 8J-L ) . Similar results were found in colony formation assays, whereas ADT-030 reduced 765 the number and size of colonies in both MIA PaCa-2 parental and MIA-AMG-R cells, while 766 MRTX849 reduced the colony formation only in MIA PaCa-2 parental cells ( Supplementary 767 Figures 16C-D ). T og ether , these results show that ADT-030 exhibits a broad spectrum of RAS 768 inhibitory activity and has the potential to escape acquired resistance that limits the efficacy of 769 mutant-specific KRAS G1 2D and KRAS G12 C inhibitors. 770

771 Aberrant activation of MAPK/ AKT signaling from KRAS mutations, along with the activation of 772 WNT /β-catenin mediated transcription, plays a major role in driving cancer cell prolif eration, 773 survival, and metastasis in PDAC, and other cancers, including colorectal, liver , lung, and breast 774 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint cancer 45 -4 8 . KRAS is mutated in over 90% of PDAC, while mutations or increased expression of β-775 catenin or other pathway components (e.g., W nt ligands, FZD receptors, LRP co-receptor , APC, 776 Axin) are also observed in a high percentage of patients with PDAC and other cancers 49, 50 . 777 Compensatory or cooperative interactions between these signaling pathways lik ely also 778 contribute to aggressiveness of the disease as well as therapy resistance 51 . Phosphodiesterase 779 isozymes have been previously studied in the context of cancer , but no particular isozyme has 780 been targeted by an inhibitor , and no PDE inhibitor has received FDA approval for the treatment 781 of cancer 22 . Recently , several publications have reported that isozyme-specific PDE10 inhibitors 782 or genetic silencing of PDE10 can block RAS and β-catenin by activating PKG 25, 27, 2 9,3 0 . Similarly , 783 cAMP-activated PKA can inhibit signaling downstream of RAS by disrupting interaction with Raf1 784 52 . Our data revealed increased expression of PDE10 in PDAC cells as compared to adjacent 785 normal pancreatic tissue, which provided an initial rationale to targeti this pathway for the 786 treatment of PDAC. AD T-030 is a non-CO X inhibitory sulindac derivative and a second-787 generation analog of ADT-061 (aka MCI-030), previously reported to selectively inhibit PDE10 788 and the prolif eration of colorectal cancer and ovarian cancer cell lines 25 , 30 . In these studies, 789 PDE10 knock down resulted in reduced sensitivity of the cancer cells to ADT-061, as well as 790 known PDE10 inhibitors, which confirmed the selectivity of this class of agents to PDE10 and 791 suggested that PDE10 is a understudied vulnerability of cancer cells. Molecular docking 792 simulations and cellular thermal stability assays presented in this study provide structural 793 insight into the interaction between ADT-030 and PDE10 and confirmation of target 794 engagement, respectively . These findings support an mechanism of action f or ADT-030 involving 795 PDE10 inhibition, elevation of cyclic nucleotides, and protein kinase activation and support 796 future research are needed to further study the oncogenic role of PDE10. 797 Here, we show that ADT-030 inhibits the enzymatic activity of recombinant PDE10 and acti vates 798 cyclic nucleotide signaling in PD AC cells at concentrations that selectively inhibit the 799 prolif eration of KRAS mutant PDAC cells. Of clinical relevance, we found that ADT-030 also 800 inhibits the prolif eration of PDAC cells that develop resistance to KRAS inhibitors and can 801 enhance the efficacy of standard-of-care chemotherapy , suggesting that ADT-030 has the 802 potential to be a front-line treatment for patients with PDAC. ADT-030 is distinct from other 803 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint KRAS inhibitors, FDA-approved or in development, by its capacity to escape resistance, which 804 we attribute to cell cycle arrest and the induction apoptosis, resulting from the dual blockage of 805 β-catenin and RAS signaling. The observed inhibition of MAPK and AKT /PI3K pathways by ADT-806 030 is particularly significant for PDAC treatment, as both pathways are known for their 807 extensive crosstalk and compensatory activation to drive cancer cell prolif eration and survival 808 53, 54 . Compensation from β-catenin may also contribute to resistance to monospecific inhibitors 809 of KRAS or β-catenin where by d ual blockage of RAS or β- caten in pathways through PDE10 810 inhibition may prevent the development of resistance to KRAS inhibitors, FDA-approved or in 811 development 55 . T o corroborate these findings using gene expression profiling, we evaluated 812 tumors excised from mice treated with ADT-030 or vehicle using single-cell transcriptomics. The 813

confirmed suppressive effects of ADT-030 on k ey oncogenic signaling pathways, 814 including RAS-MAPK, EMT , and WNT , as evidenced by reduced expression of Raf1, Mapk3, 815 Map2K2, vimentin, FN1, APC, and Axin2. W e also conducted assays on RAS activation in RAS 816 wild-type and KRAS mutant PDAC cell lines and found that ADT-030 selectively inhibited RAS 817 activation in KRAS mutant PDAC cell lines. This interesting observation needs further study to 818 understand the diff erential effects of PDE10 inhibition and impact of cyclic nucleotide signaling 819 in KRAS mutant versus RAS wild-type PDAC cells. 820 ADT-030 is orally bioavailable with attractive drug-lik e properties and appears to be well 821 tolerated at dosages that exhibit robust and durable antitumor activity . W e found that ADT-030 822 inhibits both primary tumor growth and metastasis without discernible toxicity in several mouse 823 models of PDAC, including PDX and orthotopic models. ADT-030 also potentiated the efficacy of 824 standard-of-care chemotherapy regimens for PD AC. These findings support the rationale for 825 developing ADT-030 as a front-line treatment for PDAC as a monotherapy or in combination 826 with standard-of care chemotherapy . 827 PDE10 inhibitors have been previously developed for the treatment of CNS disorders such as 828 schizophrenia and Huntington’ s disease. W e the refore c ompared ADT-030 to the known PDE10 829 inhibitor , PF-2545920, and found that ADT-030 displayed appreciably greater potency than PF-830 2545920 to inhibit PDAC cell proliferation in vitro . This observation suggests that although 831 PDE10 is a cancer target, PF-2545920 has low potency to inhibit cancer cell prolif eration, which 832 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint may be attributed to compensation by other co-expressed PDE isozymes (e.g., PDE5). When 833 compared to ADT-030 in vivo, PF-2545920 failed to demonstrate anti-tumor activity . In addition, 834 ADT-030 displayed improved tolerance without the side eff ects (sedation) observed with PF-835 2545920 56 . 836 PDAC is also known to be associated with an immunosuppressive TiME, a major factor 837 responsible for resistance to immunotherapy 57 , 58 . The immune infiltration in PDAC is 838 characterized by the abundance of immune suppressive cells and the lack of anti-tumor 839 immune cells 59 . Activation of the immune system by ADT- 030 treatment is another intriguing 840 and clinically relevant finding as we report. ADT-030 treatment increased CD4 + and CD8 + T cells, 841 as well as NK cell infiltration, resulting in a shift towards M1-lik e macrophage polarization. The 842 inhibition of expression of CTLA-4, PD-1, and LAG-3 on CD8 + T cells by ADT-030 treatment 843 suggests that ADT-030 alleviates T cell exhaustion and reestablishes cytotoxic T cell function 844 within the TiME 60 . Aside from maintaining antitumor T cells, ADT-030 also enhanced myeloid 845 cell infiltration by increasing the number of total macrophages (F4/80 + ) within the TiME. In 846 particular , these macrophages displayed enhanced antigen-presenting potential as evidenced by 847 increased expression of PD-L1, crucial for eff ector T cell interaction. In addition, phenotypic 848 analysis revealed a shift in macrophage polarization toward a pro-inflammatory , M1-lik e 849 phenotype (MHCII + CD86 + ), which was accompanied by a decrease in M2-lik e macrophages 850 (CD206 + ) 61 . These results were further supported by the identification of a high M1/M2 ratio. In 851 addition to macrophages, ADT-030 treatment led to an increase in conventional dendritic cells 852 (cDC1 and cDC2), contributing to activated T and NK cells 62 . Our single-cell RNA sequencing 853 data demonstrated that the cytotoxic lymphocyte compartment is reprogrammed, with CD8 T 854 cells and TNK cells exhibiting enhanced activation, eff ector function, and reinvigoration. 855 Activation of cytotoxic genes (IFNγ, Gzma, and Prf1) and activation mark ers (CD69 and Cxcr3), 856 alongside modulation of exhaustion path ways (Lag3 and CTLA4), indicates that ADT-030 not 857 only suppresses tumor cell prolif eration but also potentiates anti-tumor immunity . These 858 convergent tumor-cell and immune-cell eff ects provide a strong mechanistic rationale for 859 evaluating ADT-030 in combination with immune checkpoint blockade or other 860 immunomodulatory strategies, with the goal of converting immunologically cold PDAC into a 861 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint more treatment-responsive state. This will be a future direction for preclinical studies 862 combining ADT-030 with immune checkpoint blockade and possible clinical trials, given the 863

of immunotherapy for the treatment of PDAC. 864 While mutant-selective KRAS G12 C and KRAS G12D inhibitors have demonstrated promising efficacy 865 for KRAS mutant cancers 63 , 64 , acquired resistance remains a major clinical limitation. V arious 866 mechanisms of resistance have been reported, including secondary RAS mutations, activation of 867 co-expressed RAS wild-type isozymes, and compensatory receptor tyrosine kinase mutations, all 868 of which frequently emerge in recurrent tumors and contribute to treatment failure, disease 869 relapse, and death of the patient 65 . In the current study , we investigated whether ADT-030 may 870 be less susceptible to the same mechanisms of resistance that limit the efficacy of KRAS G12C and 871 KRAS G12D inhibitors using PDAC cell lines developed to be resistant to such drugs. Strikingly , ADT-872 030 demonstrated potent anti-proliferative activity in both MRTX1133 and MRTX849 resistant 873 PDAC cell lines, highlighting its broad-spectrum pan-RAS inhibitory activity and its ability to 874 bypass diverse mechanisms of acquired resistance to allele-specific KRAS inhibitors. 875 The therapeutic potential of ADT-030 is supported by the antitumor results observed in 876 clinically and genetically relevant PDAC PD X models harboring KRAS G12D and KRAS G12C mutations 877 as well as several orthotopic mouse models. In these experiments, efficacy and tolerability of 878 ADT-030 were assessed following oral administration at a dose of 150 mg/kg for 4 weeks. This 879 dosage caused no discernible toxicity and is a human equivalent dosage of 12 mg/kg, or 840 mg 880 once daily for a 70 kg human, a mode rate dose for many drugs. T reated mice showed tumor 881 regression and no tumor regrowth for over 70 days after stopping treatment, highlighting the 882 durability in maintaining long-term tumor control. These findings provide compelling evidence 883 of robust and clinically relevant antitumor activity of ADT-030 by inhibiting PDE10 that support 884 IND-enabling studies. This activity and capacity of ADT-030 to simultaneously block RAS and β-885 catenin signaling also supports further mechanistic studies of the oncogenic role of PDE10 886 ( Supplementary Figure 17 ). Future studies focusing on identifying the activity of ADT-030 in 887 genetically engineered mouse models as a monotherapy and in combination with immune 888 checkpoint inhibitors will further help in the translation of this agent to the clinic. In conclusion, 889 the ability of ADT-030 to inhibit PDE10 and block multiple aspects of malignant progression, 890 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint including cancer cell prolif eration, survival, and metastasis, as well as creating a more favorable 891 TiME, while having the potential to escape resistance that limits the efficacy of other RAS 892 inhibitors, mak es ADT-030 a highly desirable drug development candidate for clinical trials in 893 patients with metastatic PDAC and other RAS-driven malignancies. 894 Ref erences: 895 1 Siegel, R. L., Giaquinto, A. N. & Jemal, A. Cancer statistics, 2024. CA: a cancer journal for 896 clinicians 74, 12-49 (2024). 897 2 Balsano, R., T ommasi, C. & Garajova, I. State of the art for metastatic pancreatic cancer 898 treatment: Where are we now? Anticancer research 39, 3405-3412 (2019). 899 3 Eser , S., Schniek e, A., Schneider , G. & Saur , D. Oncogenic KRAS signalling in pancreatic cancer . 900 British journal of cancer 111 , 817-822 (2014). 901 4 Leonhardt, L. & Hebrok, M. KRAS degradation averts PDAC chemoresistance. Nature Cancer 5 , 902 375-377 (2024). 903 5 Zhang, Y . e t a l . β-Catenin mediated T AM phenotype promotes pancreatic cancer metastasis via 904 the OSM/ST A T3/LO XL2 axis. Neoplasia 60 , 101096 (2025). 905 6 Sano, M. e t a l . Activation of WNT /β-catenin signaling enhances pancreatic cancer development 906 and the malignant potential via up-regulation of Cyr61. Neoplasia 18, 785-794 (2016). 907 7 Ram Mak ena, M. e t a l . W nt/β-catenin signaling: the culprit in pancreatic carcinogenesis and 908 therapeutic resistance. International journal of molecular sciences 20, 4242 (2019). 909 8 Manegold, P . e t a l . Diff erentiation therapy targeting the β-catenin/CBP interaction in pancreatic 910 cancer . Cancers 10, 95 (2018). 911 9 Dominguez, A. A. e t a l . Unveiling the Promise: Navigating Clinical T rials 1978–2024 for PD AC. 912 Cancers 16, 3564 (2024). 913 10 Jiménez, D. J., Javed, A., Rubio-T omás, T ., Seye-Loum, N. & Barceló, C. Clinical and preclinical 914 targeting of oncogenic pathways in PDAC: targeted therapeutic approaches for the deadliest 915 cancer . International journal of molecular sciences 25, 2860 (2024). 916 11 Huber , M. e t a l . The immune microenvironment in pancreatic cancer . International journal of 917 molecular sciences 21, 7307 (2020). 918 12 Y ousef , A. e t a l . Impact of KRAS mutations and co-mutations on clinical outcomes in pancreatic 919 ductal adenocarcinoma. NP J Precision Oncology 8 , 27 (2024). 920 13 Zhang, Z., Zhang, H., Liao, X. & T sai, H.-i. KRAS mutation: The booster of pancreatic ductal 921 adenocarcinoma transformation and progression. Frontiers in Cell and Developmental Biology 922 11, 1147676 (2023). 923 14 Li, J., Mizukami, Y ., Zhang, X., Jo, W .-S. & Chung, D. C. Oncogenic K-ras stimulates W nt signaling in 924 colon cancer through inhibition of GSK-3β. Gastroenterology 128 , 1907-1918 (2005). 925 15 Shetu, S. A. & Bandyopadhyay , D. Small-molecule RAS inhibitors as anticancer agents: discovery , 926 development, and mechanistic studies. International Journal of Molecular Sciences 23, 3706 927 (2022). 928 16 Ryu, W .-J., Han, G., Lee, S.-H. & Choi, K.-Y . Suppression of W nt/β-catenin and RAS/ERK pathways 929 provides a therapeutic strategy for gemcitabine-resistant pancreatic cancer . Biochemical and 930 Biophysical Research Communications 549 , 40-46 (2021). 931 17 W ang, W . e t a l . Dual-targeted therapy circumvents non-genetic drug resistance to targeted 932 therapy . Frontiers in Oncology 12, 859455 (2022). 933 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint 18 Raghavendra, N. M., Pingili, D., Kadasi, S., Mettu, A. & Prasad, S. Dual or multi-targeting 934 inhibitors: The next generation anticancer agents. European journal of medicinal chemistry 143 , 935 1277-1300 (2018). 936 19 Lemieux, E., Cagnol, S., Beaudry , K., Carrier , J. & Rivard, N. Oncogenic KRAS signalling promotes 937 the Wnt/β-catenin pathway through LRP6 in colorectal cancer . Oncogene 34, 4914-4927 (2015). 938 20 Samidurai, A. et al. Role of phosphodiesterase 1 in the pathophysiology of diseases and potential 939 therapeutic opportunities. Pharmacology & therapeutics 226 , 107858 (2021). 940 21 Gross, N. E. e t a l . Phosphodiesterase-5 inhibition collaborates with vaccine-based 941 immunotherapy to reprogram myeloid cells in pancreatic ductal adenocarcinoma. JCI insight 9 , 942 e179292 (2024). 943 22 Kelly , M. P . e t a l . Cyclic nucleotide phosphodiesterases as drug targets. Pharmacological reviews 944 77 (2025). 945 23 F aure, M. & Bourne, H. Differential eff ects on cAMP on the MAP kinase cascade: evidence for a 946 cAMP-insensitive step that can bypass Raf-1. Molecular biology of the cell 6 , 1025-1035 (1995). 947 24 Suhasini, M., Li, H., Lohmann, S. M., Boss, G. R. & Pilz, R. B. Cyclic-GMP-dependent protein 948 kinase inhibits the Ras/Mitogen-activated protein kinase pathway . Molecular and cellular biology 949 18, 6983-6994 (1998). 950 25 Borneman, R. M. e t a l . Phosphodiesterase 10A (PDE10A) as a novel target to suppress β-catenin 951 and RAS signaling in epithelial ovarian cancer . Journal of Ovarian Research 15, 120 (2022). 952 26 Jeong, W .-J., Ro, E. J. & Choi, K.-Y . Interaction between Wnt/β-catenin and RAS-ERK pathways and 953 an anti-cancer strategy via degradations of β-catenin and RAS by targeting the W nt/β-catenin 954 pathway . NP J Precision oncology 2 , 5 (2018). 955 27 Zhu, B. e t a l . Phosphodiesterase 10A is overexpressed in lung tumor cells and inhibitors 956 selectively suppress growth by blocking β-catenin and MAPK signaling. Oncotarget 8 , 69264 957 (2017). 958 28 Peng, T . e t a l . Inhibitors of phosphodiesterase as cancer therapeutics. European journal of 959 medici nal chemistry 150 , 742-756 (2018). 960 29 Li, N. e t a l . Phosphodiesterase 10A: a novel target for selective inhibition of colon tumor cell 961 growth and β-catenin-dependent TCF transcriptional activity . Oncogene 34, 1499-1509 (2015). 962 30 Lee, K. J. et al. Suppression of colon tumorigenesis in mutant apc mice by a novel PDE10 inhibitor 963 that reduces oncogenic β-catenin. Cancer Prevention Research 14, 995-1008 (2021). 964 31 Loveless, I. M. e t a l . Human Pancreatic Cancer Single-Cell Atlas Reveals Association of CX CL10+ 965 Fibroblasts and Basal Subtype T umor Cells. Clinical cancer research 31, 756-772 (2025). 966 32 Babic, I. e t a l . MICRO-T AG enzyme complementation enables quantification of cellular drug-967 target engagement in temperature series. SLAS Discovery , 100291 (2025). 968 33 Tinsley , H. N. e t a l . Sulindac sulfide selectively inhibits growth and induces apoptosis of human 969 breast tumor cells by phosphodiesterase 5 inhibition, elevation of cyclic GMP , and activation of 970 protein kinase G. Molecular cancer therapeutics 8 , 3331-3340 (2009). 971 34 Kim, D.-Y ., Park, J.-S., Leem, Y .-H., Park, J.-E. & Kim, H.-S. The potent PDE10A inhibitor MP-10 (PF-972 2545920) suppresses microglial activation in LPS-induced neuroinflammation and MPTP-induced 973 Parkinson’ s disease mouse models. Journal of Neuroimmune Pharmacology 16, 470-482 (2021). 974 35 Seibenhener , M. L. & W ooten, M. C. Use of the open field maze to measure locomotor and 975 anxiety-lik e behavior in mice. Journal of visualized experiments: JoVE, 52434 (2015). 976 36 Bandi, D. S. R. e t a l . ADT-1004: A promising pan-RAS inhibitor for targeting KRAS mutations in 977 pancreatic ductal adenocarcinoma. Cancer Research 84, 5915-5915 (2024). 978 37 Reinhard, M. et al. The 46/50 kDa phosphoprotein V ASP purified from human platelets is a novel 979 protein associated with actin filaments and focal contacts. The EMBO journal 11, 2063-2070 980 (1992). 981 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint 38 Üffing, A., Attridge, E. & T ooze, S. A. T argeting an alternative route: autophagy in RAS-driven 982 cancers. Cell Research 35, 389-390 (2025). 983 39 Gurpinar , E. e t a l . A novel sulindac derivative inhibits lung adenocarcinoma cell growth through 984 suppression of Akt/mT OR signaling and induction of autophagy . Molecular cancer therapeutics 985 12, 663-674 (2013). 986 40 Berven, L. A. & Crouch, M. F . Cellular function of p70S6K: a role in regulating cell motility . 987 Immunology and cell biology 78, 447-451 (2000). 988 41 T u, Z. e t a l . Radiosynthesis and in vivo evaluation of [11C] MP-10 as a PET probe for imaging 989 PDE10A in rodent and non-human primate brain. Bioorganic & medicinal chemistry 19, 1666-990 1673 (2011). 991 42 Spadafora, V . e t a l . Optimization of a mouse model of pancreatic cancer to simulate the human 992 phenotypes of metastasis and cachexia. BMC cancer 24, 414 (2024). 993 43 Hallin, J. e t a l . Anti-tumor efficacy of a potent and selective non-covalent KRASG12D inhibitor . 994 Nature medicine 28, 2171-2182 (2022). 995 44 Hallin, J. e t a l . The KRASG12C inhibitor MRTX849 provides insight toward therapeutic 996 susceptibility of KRAS-mutant cancers in mouse models and patients. Cancer discovery 10 , 54-71 997 (2020). 998 45 Bahar , M. E., Kim, H. J. & Kim, D. R. T argeting the RAS/RAF /MAPK pathway for cancer therapy: 999 from mechanism to clinical studies. Signal transduction and targeted therapy 8 , 455 (2023). 1000 46 Choucair , K. e t a l . T argeting KRAS mutations: orchestrating cancer evolution and therapeutic 1001 challenges. Signal T ransduction and T argeted Therapy 10, 385 (2025). 1002 47 Y u, F . e t a l . W nt/β-catenin signaling in cancers and targeted therapies. Signal T ransduction and 1003 T argeted Therapy 6 , 307 (2021). 1004 48 Xu, H., Ren, S., W ang, Y ., Zhang, T . & Lu, J. Abnormal activation of the Ras/MAPK signaling 1005 pathway in oncogenesis and progression. Cancer Adv 8 , e25002 (2025). 1006 49 Kim, E. e t a l . Promotion of growth factor signaling as a critical function of β-catenin during HCC 1007 progression. Nature communications 10, 1909 (2019). 1008 50 Zeng, G. e t a l . Aberrant Wnt/β-catenin signaling in pancreatic adenocarcinoma. Neoplasia 8 , 1009 279-289 (2006). 1010 51 Lee, S.-K., Hwang, J.-H. & Choi, K.-Y . Interaction of the W nt/β-catenin and RAS-ERK pathways 1011 involving co-stabilization of both β-catenin and RAS plays important roles in the colorectal 1012 tumorigenesis. Advances in biological regulation 68, 46-54 (2018). 1013 52 Dumaz, N. & Marais, R. Protein kinase A blocks Raf-1 activity by stimulating 14-3-3 binding and 1014 blocking Raf-1 interaction with Ras. Journal of Biological Chemistry 278 , 29819-29823 (2003). 1015 53 Bye, B. A. e t a l . Combined PI3K and MAPK inhibition synergizes to suppress PDAC. bioRxiv , 1016 2023.2008. 2015.553438 (2023). 1017 54 Ouissam, A. J., Hind, C., Sami Aziz, B. & Said, A. Inhibition of the PI3K/ AKT /mT OR pathway in 1018 p a n c r e a t i c c a n c e r : i s i t a w o r t h w h i l e e n d e a v o r ? Therapeutic advances in medical onc ology 16 , 1019 17588359241284911 (2024). 1020 55 Awasthi, N. e t a l . Dual inhibition of the PI3K and MAPK pathways enhances nab-1021 paclitax el/gemcitabine chemotherapy response in preclinical models of pancreatic cancer . 1022 Cancer Letters 459 , 41-49 (2019). 1023 56 Uthayathas, S. e t a l . Phosphodiesterase 10A inhibitor MP-10 eff ects in primates: comparison 1024 with risperidone and mechanistic implications. Neuropharmacology 77, 257-267 (2014). 1025 57 Ju, Y . e t a l . Barriers and opportunities in pancreatic cancer immunotherapy . NP J Precision 1026 Oncology 8 , 199 (2024). 1027 58 Y e, N. e t a l . A multi-omic approach reveals utility of CD45 expression in prognosis and novel 1028 target discovery . Frontiers in Genetics 13, 928328 (2022). 1029 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint 59 Pollini, T . et al. The tumour immune microenvironment and microbiome of pancreatic intraductal 1030 papillary mucinous neoplasms. The Lancet Gastroenterology & Hepatology 7 , 1141-1150 (2022). 1031 60 Kabacaoglu, D., Ciecielski, K. J., Ruess, D. A. & Algül, H. Immune checkpoint inhibition for 1032 pancreatic ductal adenocarcinoma: current limitations and future options. Frontiers in 1033 immunology 9 , 1878 (2018). 1034 61 W ang, L. e t a l . PD-L1-expressing tumor-associated macrophages are immunostimulatory and 1035 associate with good clinical outcome in human breast cancer . Cell Reports Medicine 5 (2024). 1036 62 Böttcher , J. P . e t a l . NK cells stimulate recruitment of cDC1 into the tumor microenvironment 1037 promoting cancer immune control. Cell 172 , 1022-1037. e1014 (2018). 1038 63 T anaka, N. e t a l . Clinical acquired resistance to KRASG12C inhibition through a novel KRAS 1039 switch-II pock et mutation and polyclonal alterations converging on RAS–MAPK reactivation. 1040 Cancer discovery 11, 1913-1922 (2021). 1041 64 Li, Y ., Zhao, J. & Li, Y . New exploration of KRASG12D inhibitors and the mechanisms of resistance. 1042 Experimental Hematology & Oncology 14 , 39 (2025). 1043 65 Awad, M. M. et al. Acquired resistance to KRASG12C inhibition in cancer . New England Journal of 1044 Medicine 384 , 2382-2393 (2021). 1045 1046 1047 Conflict of interest: A.B.K, X.C., and G.A.P are affiliated with ADT Pharmaceuticals, LLC. 1048 Data and materials a vailability: All data associated with this study are present in the paper or 1049 the Supplementary Materials. 1050 Funding decl aration: This work was supported by the University of Alabama at Birmingham (UAB), 1051 Birmingham, AL, USA. B.E was also supported by the UAB O’Neal Comprehensive Cancer Center, 1052 National Institutes of Health (NIH) core support grant 5P30CAO13148-47 and 1R01CA294647. This work 1053 was also supported by the NIH grants R01CA254197 (Piazza) and R01CA238514 (Zhou and Piazza) 1054 Contributions 1055 Concept and design: BE and GP; acquisition, analysis, or interpretation of data: DSRB, VRA, GSG, 1056 LC; drafting of the manuscript: DSRB; critical revision of the manuscript for important 1057 intellectual content: all authors; statistical analysis: DSRB; administrative, technical, or material 1058 support: BE; validation: DSRB; supervision: BE. sc-RNA data analysis: SS; Modeling: TH; 1059 Histopath: JBF; Thermal assay: EN and IB; In vitro testing: ABK; All authors have read and 1060 approved the article. 1061 1062 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 Figure leg ends 1075 Figure 1: ADT-030 inhibits PDE10 at concentrations that inhibit proliferation, colony 1076 formation, and motility , while inducing apoptosis and cell cycle arr est of KRAS mutant PD AC 1077 cells . 1078 A. Violin plot of PDE10 expression in human donor , adjacent normal tissue, primary tumor and 1079 metastatic l esion as det ermined by sc-RNAseq analysis of human PD AC. B . Baseline expression 1080 of PDE10 across indicated PDAC cell lines. β-actin was used as a loading control. C . Chemical 1081 structure of ADT-030. D. ADT-030 inhibits the enzymatic activity of recombinant PDE10 using 1082 cAMP and cGMP as substrates. Data are expressed as mean ± SD , n = 2 samples/concentration. 1083 E. PDAC cell lines were treated with various concentrations of ADT-030 f or 3 days followed by 1084 determining viable cell number using MTT assays. Relative percentage cell viability was plotted 1085 with respect to vehicle (DMSO) treated cells. The table lists the IC 50 values for the PD AC cell 1086 lines treated with ADT-030. F - G . The indicated PDAC cell lines were treated with various 1087 concentrations of ADT-030 for 2–4 weeks, and long-term cell survival was measured using 1088 clonogenic assays. Representative images are shown in F and quantification plotted in G. H. 1089 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint Indicated PDAC cell lines were treated with vehicle or the indicated concentrations of ADT-030, 1090 and cell migration was analyzed using cell motility assays. Representative images under 1091 indicated treatment conditions for indicated PDAC cell lines are shown. I. Bar diagrams are 1092 presented to show relative migration (%) from the experiment presented in H. Data represents 1093 the meanLi44±Li44SEM of three biological replicates. nsLi44=Li44not significant, *pLi44<Li440.05, **pLi44<Li440.01, 1094 ***pLi44<Li440.001, ****pLi44<Li440.0001. (one-way ANOV A). 1095 Figure 2. ADT-030 blocks PDE10 and activates PKA/PK G to reduce β-c atenin levels and inhibit 1096 RAS signaling. 1097 A-D. ADT-030 increases cyclic nucleotide levels in a concentration-dependent manner in 2838c3 1098 ( A-B ) and MIA PaCa-2 cells ( C-D). E-F . T reatment with ADT-030 for 4 hrs did not aff ect the 1099 expression of PDE10 but induced V ASP phosphorylation at serine 157 (PKA site) and serine 239 1100 (PKG site) in a concentration-dependent manner in 2838c3 and MIA PaCa-2 cells. β-actin was 1101 used as a loading control. G. ADT-030 decreased phosphorylation of ERK, mT OR, and levels of 1102 active (oncogenic) β-catenin in 2838c3 and MIA PaCa-2 cells in a concentration-dependent 1103 manner . β-actin was used as a loading control. H-J. Indicated cell lines were treated with 1104 increasing concentrations of ADT-030 for 24 hrs and RAS pulldown was performed. Data 1105 represent the meanLi44±Li44SEM of three biological replicates. nsLi44=Li44not significant, *p < 0.05, 1106 **pLi44<Li440.01, ***pLi44<Li440.001, and ****pLi44<Li440.0001 (one -way ANOV A). 1107 Figure 3. Effect of ADT-030 treatment on tumor growth and modulation of TME in 2838c3 cell-1108 implanted C57BL/6J mice. 1109 A-B. Histopathological examination of vital organs (heart, lung, kidney, liver, duodenum, 1110 pancreas, colon, spleen, thymus, testes, and brain) from mice treated with ADT-030 at 150 1111 mg/kg as assessed by H&E staining. C. 2838c3 -f- luc cells were injected orthotopically into the 1112 pancreas of C57BL/6J mice. Representative bioluminescence images at the indicated time 1113 points are shown. D. Relative normalized whole-body bioluminescence intensities in mice under 1114 the indicated conditions (nLi44=Li445). Statistical significance was determined using one-way ANOVA. 1115 E. Images of the pancreatic tumors in the vehicle- and the 150 mg/kg ADT-030-treated groups 1116 at termination. F. Tumor weights at the end of the experiment for the indicated doses. G. 1117 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint Average body weights of mice treated with vehicle and ADT-030 (50, 100, and 150 mg/kg). 1118 Statistical significance was determined using one-way ANOVA. nsLi44=Li44not significant, 1119 **pLi44<Li440.01, and ****pLi44<Li440.0001. 1120 Figure 4. ADT-030 reduces β-catenin levels, inhibits RAS/AKT signaling, and induces 1121 autophagic cell death. 1122 A. Western blots showing levels of pERK, total ERK, pAKT, total AKT, non-phospho-β-catenin, 1123 total β-catenin, LC3A/B, cleaved PARP, and cleaved caspase 3 in 2838c3 -f- luc tumors after 1124 vehicle or ADT-030 treatment. B-G. Bar graphs representing the quantifications of western 1125 blots from panel A : pERK ( B), p-AKT (C ), non-phospho-β-catenin ( D), LC3A/B ( E ), cleaved PARP 1126 ( F), and cleaved caspase 3 ( G) in tumor tissues after ADT-030 vs. vehicle treatments. Welch t-1127 test was used for statistical analysis. H. The inhibitory effect of ADT-030 on activated (GTP-1128 bound) RAS in tumors after vehicle or ADT-030 treatments was assessed by RAS-RBD pull-down 1129 assay. I. Representative Ki-67 IHC results in tumors after vehicle or ADT-030 treatment. J-L. 1130 Representative IF images of LC3A/B (J) , vimentin (K) , and E-Cadherin (L) in tumors after vehicle 1131 or ADT-030 treatment. M. Bar graph representing the quantification of IHC staining for KI-67. N-1132 P. Dot-plot graphs representing the IF quantifications of LC3A/B (N), vimentin (O), and E-1133 Cadherin (P) in tumors after vehicle or ADT-030 treatment. Welch t-test was used for statistical 1134 analysis. ns, non-significant, ∗ pLi44<Li440.05, ∗∗ pLi44<Li440.01, and ∗∗∗∗ pLi44<Li440.0001. 1135 Figure 5. ADT-030 induces tumor growth arrest and regression in KPC and KPCML1 cells-1136 implanted C57BL/6J mice and reduces liver and lung metastasis. 1137 A. KP C -f- luc cells were injected orthotopically into the pancreas of C57BL/6J mice. 1138 Representative bioluminescence images at the indicated time points are shown. B. Relative 1139 normalized whole-body bioluminescence intensities in mice under the indicated conditions 1140 (nLi44=Li445). Statistical significance was determined using one-way ANOVA. C-D. Tumor images 1141 after treatment with vehicle or ADT-030 at 50, 100, and 150 mg/kg ( C ) and bar graph 1142 representing tumor weights ( D) from KPCML1 orthotopic model at termination. Statistical 1143 significance was determined using one-way ANOVA. E-F. Tumor images after treatment with 1144 vehicle or ADT-030 at 150 mg/kg (E) and bar graph representing tumor weights (F) from 1145 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint KPCML1 orthotopic model at termination. Statistical significance was determined using one-1146 way ANOVA. G . Hematoxylin and eosin (H&E) staining of KPC-derived lungs after 4 weeks is 1147 shown. Representative images of H&E stained sections of metastasis in lung are displayed (20x). 1148 H. H&E staining of KPCML1-derived liver metastasis after 4 weeks is shown. Representative 1149 images of H&E staining sections from metastasis in lung are displayed (20x). I-J. Graph 1150 representing the quantification of lung (I) and liver (J) metastatic nodules in lung sections after 1151 vehicle or ADT-030 treatment from KPC and KPCML1 orthotopic experiments. Welch t-test was 1152 used for statistical analysis. ns, non-significant, ∗ pLi44<Li440.05, ∗∗ pLi44<Li440.01, ∗∗∗pLi44<Li440.001, and 1153 ∗∗∗∗ pLi44<Li440.0001 1154 Figure 6. ADT-030 remodels tumor cell states and reinvigorates CD8 T cell in PDAC 1155 A-B. UMAP visualization of all single cells isolated from orthotopic PDAC tumors treated with 1156 either vehicle or ADT-030, colored by major cell types as indicated. C. UMAP reclustering 1157 showing 7 transcriptionally distinct PDAC cell types. D. Split UMAPs of PDAC cells from vehicle 1158 and ADT-030 tumors demonstrating treatment-associated shifts. E. Bar graph showing top 1159 significantly downregulated pathways in ADT-030 treated PDAC cells compared to vehicle as 1160 demonstrated by GSEA. F-H. Violin plots showing expression or module scores for RAF1 1161 suppression (F), upstream RAS (G) , downstream MAPK (H) related signatures in PDAC cells, 1162 including upstream WNT activation score (F), a downstream Wnt target gene score (G) , and 1163 MAPK/ERK pathway score (H), comparing vehicle and ADT-030. I-J. UMAP features plots 1164 demonstrating single-cell MAPK signaling flux for Raf1 (I), and Map2k2 (ERK) (J) with 1165 corresponding dot plots summarizing average pathway activity (color scale) and fraction of 1166 PDAC cells expressing each gene set (dot size) in vehicle and ADT-030 tumors. K-L. Dot plots 1167 summarizing GSEA-derived pathway scores in PDAC cells highlighting RAS/MAPK pathway (K) , 1168 and reduced EMT (L). M-N. Violin plots showing stemness-associated EMT-related module 1169 scores for fibronectin in PDAC cells (M), and WNT/β-catenin-dependent gene signatures (N). O. 1170 Dot plot of canonical WNT/β-catenin pathway genes across PDAC clusters. P. Dot plot of 1171 proposed mechanism-related genes illustrating transcriptional repression. Q. UMAPs of T and 1172 NK cell compartments in vehicle and ADT-030 tumors. R. Stacked bar plot quantifying the 1173 proportion of CD8 T cell, NK cell, T regs and TNK cell states among total T cells in vehicle vs. 1174 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint ADT-030. S. Scatter plot showing CD8 T cell state trajectory scores, with each point representing 1175 a single CD8 T cell positioned according to activation and exhaustion signature scores in vehicle 1176 and ADT-030 treatments. T. Dot plot summarizing the expression of representative activation, 1177 exhaustion, and stem-like genes in CD8 T cells. U. Heat map of CD8 T cell exhaustion-associated 1178 genes across individual CD8 T cells by trajectory state comparing vehicle and ADT-030 treated 1179 tumors depicting broad downregulation of exhaustion markers. 1180 Figure 7. ADT-030 enhances global TNK activation and sensitizes chemotherapy to suppress 1181 PDAC tumor growth in vivo 1182 A. TNK global state trajectory plot displaying pan-activation vs. pan-dysfunction signatures in 1183 vehicle and ADT-030 treated tumors. B. Dot plot showing expression of activation, dysfunction, 1184 and effector genes in TNK cells from vehicle and ADT-030 treated tumors with dot size 1185 representing the percentage of expressing cells. C . Heat map of differentially expressed genes 1186 in TNK cells representing a global transcriptional shift toward an activated, cytotoxic program in 1187 ADT-030 treated tumors compared to vehicle. D. Representative images of tumors harvested 1188 from the mice treated with vehicle, PF-2545920, ADT-030, gemcitabine paclitaxel (GPTx), and 1189 the combination of ADT-030+GPTx. E . Quantification of tumor weights across all treatment 1190 groups showing tumor growth inhibition with ADT-030 and further significant reduction in ADT-1191 030+GPTx combination groups compared to monotherapies and vehicle. F . Comparison of 1192 tumor weights between known PDE10 inhibitor (PF-2545920) and ADT-030 as monotherapy 1193 demonstrating superior efficacy of ADT-030. ANOVA and Welch t-test was used for statistical 1194 analysis. ns, non-significant, ∗ pLi44<Li440.05, ∗∗ pLi44<Li440.01, ∗∗∗pLi44<Li440.001, and ∗∗∗∗ pLi44<Li440.0001. 1195 Figure 8. ADT-030 induces tumor regression and extends survival in KRAS G12D and KRAS G12C 1196 PDX models with potential to escape resistance to KRAS G12D and KRAS G12C inhibitors. A. Tumor 1197 growth in NOD.Cg-Prkdc scid Il2rg tm1 Wjl /SzJ mice implanted with KRAS G12D PDX and treated with 1198 vehicle or 150 mg/kg ADT-030. B. Average body weights of mice treated with vehicle or ADT-1199 030. C. Survival curves of mice after the treatment with the vehicle or ADT-030. D. Tumor 1200 growth in NOD.Cg-Prkdc scid Il2rg tm1 Wjl /SzJ mice implanted with KRAS G12 C PDX and treated with 1201 vehicle or ADT-030. E. Average body weights of mice treated with vehicle or ADT-030. F. 1202 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint Survival curves of mice after the treatment with the vehicle or ADT-030. G-H. The indicated 1203 PDAC cell lines were treated with various concentrations of ADT-030 (G) or MRTX1133 ( H) and 1204 subjected to MTT assays. Relative percentage cell viability was plotted relative to vehicle 1205 treated cells. I. Table showing the IC 50 values for the cell lines used in panels G-H. J-K. The 1206 indicated PDAC cell lines were treated with various concentrations of ADT-030 ( J) or MRTX849 1207 ( K) and subjected to MTT assays. Relative percentage cell viability was plotted relative to 1208 vehicle treated cells. L. Table listing the IC 50 values for each PDAC cell line shown in panels J-K. 1209 Supplementary figure leg ends: 1210 Supplementary Figure 1: A. Docking of ADT-030 to PDE10 (2OUN) resulted in an optimal 1211 docking score of -10.325. Representative surface rendering of 2OUN with electrostatic potential 1212 mapped onto the surface shows ADT-030 bound in the PDE10 cataly t ic pock et. ADT-030 is 1213 rendered as a ball-and-stick representation. Pink and grey spheres illust rate magnesium and 1214 zinc ions in the pock et, respectively . B . Interaction diagram showing molecular interactions of 1215 ADT-030 with PDE10. B inding affinity of ADT-030 to PDE10 as determined by treating HEK293 1216 cells (45 min) expressing PDE10-Micro-T ag with ADT-030. C . W estern blot analysis showing 1217 PDE10 expression in HEK293 cells transf ected with PDE10-Micro-T ag and detected using anti- 1218 Micro-T ag antibody . D . Quantification of Micro-T ag enzyme compleme ntati on in ce ll s 1219 transf ected with PDE10-Micro-T ag c onstruct vs untransf ected HEK293 cells. E . PDE10 thermal 1220 curve yielding a T agg₅₀ of 44°C, providing the fix ed challenge temperature to determine ADT-1221 030 binding to PDE10 in transf ected HEK293 cells . F . AD T-030 binding to PDE10 in transf ected 1222 HEK293 cells. The curve is graphed as the average of two replicates ± SEM. 1223 Supplementary Figure 2: A-B. Representative F ACS analysis showing ADT-030 induced apoptosis 1224 in 2838c3 (A), and MIA PaCa-2 (B) cells after treatment with ADT-030 (2 and 5 µM) or vehicle 1225 for 24 hrs. After double-staining with annexin V and PI, cells were subjected to flow cytometry 1226 analysis. C-D. The indicated PDAC cell lines were treated with ADT-030 at varying concentrations 1227 for 24 hrs, and apoptosis was measured following annexin V /propidium iodide labeling. E-F. 1228 Representative DNA histogram showing cell cycle arrest in 2838c3 (E) , and MIA PaCa-2 (F) cells 1229 after treatment with the indicated concentrations of ADT-030 or vehicle controls for 72 hrs. G-1230 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint H. Flow cytometry analysis of cell cycle distribution in 2838c3 (G) and MIA P aCa-2 (H) cells 1231 treated with vehicle or ADT-030 at the indicated concentrations. Data represents the 1232 meanLi44±Li44SEMr of three biological replicates. nsLi44=Li44not significant, *pLi44<Li440.05, **pLi44<Li440.01, 1233 ***pLi44<Li440.001, ****pLi44<Li440.0001. (one-way ANOV A). 1234 Supplementary Figur e 3: A-C. W estern blot analysis showing the expression of indicated PDE 1235 isozymes, pCREB CREB, VEGF A, and Bcl-2 in 2838c3 and MIA PaCa-2 cells treated with DMSO , 1236 and varied time points/concentrations of ADT-030. 1237 Supplementary Figure 4: A-F . Quan tification of RAS-GTP activation in BxPC-3, Panc 02, KLE, 1238 MKN1, 2838c3, and MIAPaCa-2 cells treated with either DMSO or increasing concentrations of 1239 ADT-030. 1240 Supplementary Figure 5. A. W estern blot analysis showing the expression of LC3 A/B in 2838c3 1241 and MIA PaCa-2 cells treated with DMSO , HCQ , ADT-030, and the combination of HCQ+ADT-1242 030. β-actin was used as a loading control. 1243 Supplementary Figure 6: A-B. Serum biochemical analysis of mice treated with ADT-030. Male 1244 C57BL/6J mice were treated with vehicle or ADT-030 (150 mg /kg) orally , 5 days/week for 2 1245 weeks. A. Serum was collected at the end of the treatment (n=5). Complete blood counts (WBC, 1246 RBC, HGB, HCT , MCV , MCH, MCHC, RDW , PL T , MPV , neutrophils, lymphocytes, monocytes, 1247 eosinophils, and basophils) revealed no diff erence between vehicle and ADT-030 treatment. B. 1248 Biochemical analysis indicated unchanged all measured parameters (total protein, albumin, ALP , 1249 AL T , amylase, urea nitrogen, calcium, creatinine, phosphorus, glucose, sodium, potassium, and 1250 globulin) except for an increase in total bilirubin because of ADT-030 tr eatment compared to 1251 vehicle treatment. ns: not significant and **pLi44<Li440.01. 1252 Supplementary Figure 7: A. Open field and locomotor assay revealed no significant diff erent in 1253 the overall mobility between vehicle and ADT-030 treated mice (n=5). B. ADT-030 plasma 1254 concentrations after daily repeated oral administration of 100 mg /kg. C. Drug concentrations in 1255 lung, kidneys, spleen, heart, liver , brain, ovaries, and colon after oral administration of 100 1256 mg /kg dose. ns: not significant. 1257 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint Supplementary Figur e 8. A-B. T umor images from 2838c3 cell-implanted C57BL/6J mice after 1258 treatment with vehicle or ADT-030 at 50 mg/kg (A) and 100 mg/kg (B) . 1259 Supplementary Figure 9. A. Percentage of CD45 + cells in 2838c3 tumors after treatment with 1260 ADT-030 at 150 mg/kg vs. vehicle. B-H. Increased percentage of total CD3 + T c e l l s ( B), CD4 + T 1261 cells ( C ), CD8 + T c e l l s ( D), CD3 + CTLA4 + cells ( E ), CD3 + PD-1 + cells ( F), CD3 + LAG3 + cells ( G), 1262 CD3 + TIGIT + cells ( H) after treatment with ADT-030 vs. vehicle. I. Increased percentage of NK 1263 (CD3 - NK1.1 + ) cells in vehicle vs. ADT-030 treatment. All quantitative data represent the 1264 meanLi44±Li44SEM. Welch t-test was used for statistical analysis. J. Quantification of RAS-GTP levels 1265 in tumor tissues of 2838c-implanted mice treated with either vehicle or 150 mg/kg dose of ADT-1266 030. Welch t-test was used for statistical analysis. ns, non-significant, ∗ pLi44<Li440.05, ∗∗ pLi44<Li440.01, 1267 ∗∗∗pLi44<Li440.001, and ∗∗∗∗ pLi44<Li440.0001. 1268 Supplementary Figur e 10. A. Representative Ki-67 IHC results in tumors after vehicle or ADT-1269 030 treatment from the KPC- f -luc orthotopic model. B-D. Representative IF images of LC3A/B 1270 (B), E-Cadherin (C) , and vimentin (D) in tumor tissues after ADT-030 vs. vehicle treatments. E . 1271 Bar graph representing the quantification of IHC staining for KI-67. F-H. Dot-plot graphs 1272 representing the immunofluorescence quantifications of LC3A/B (F), E-cadherin (G) , and 1273 vimentin (H) in tumor tissues after ADT-030 vs. vehicle treatment. W elch t-test was used for 1274 stat ist ica l analysis. ∗∗ pLi44<Li440.01, ∗∗∗pLi44<Li440.001, and ∗∗∗∗ pLi44<Li440.0001. 1275 Supplementary Figure 11. ADT-030 modulates tumor immunity in the PD AC TIME in KPC cell-1276 implanted C57BL/6J mice . A. Percentage of CD45 + immune cells in KPC tumors after vehicle or 1277 ADT-030 treatment. B . Percentage/mg tumor of total αβ T cells, C . γδ (T CRγδ + ) T cells, D . TNK 1278 (CD3 + NK1.1 + ) cells, E. NK (CD3 - NK1.1 + ) cells, F. NK1.1 + PD-1 + cells G. NK1.1 + CTLA4 + cells, H. CD4 + 1279 T cells, I. CD4 + PD-1 + cells, CD4 + TIGIT + cells, CD4 + CTLA4 + cells, CD4 + PD-1 + CTLA4 + cells, and 1280 CD4 + F ASr + cells, J. CD4 + T cell subpopulations in tumors from vehicle or ADT-030 treatment. K . 1281 CD8 + T cells, L. CD8 + PD-1 + cells, CD8 + CTLA4 + cells, CD8 + PD-1 + CTLA4 + cells, CD8 + LAG3 + cells, and 1282 CD8 + PD-1 + CTLA4 + LAG3 + cells, M. CD8 + T cell subpopulations in tumors from vehicle or ADT-030 1283 treatment. N. macrophages, O . F4/80+ PD-L1+ macrophages, P. M1 macrophages, Q. M 2 1284 macrophages, R . M1/M2 ratio in KPC tumors after vehicle or ADT-030 treatment. S. Percentage 1285 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint of total dendritic cells/mg tumor , percentage of T. cDC1, and U. cDC2/mg tumor after ADT-030 1286 vs. vehicle treatment. All quantitative data represents the meanLi44±Li44SEM. W elch t-test was used 1287 for stat ist ica l analysis. ns, non-significant, ∗ pLi44<Li440.05, ∗∗ pLi44<Li440.01, ∗∗∗pLi44<Li440.001, and 1288 ∗∗∗∗ pLi44<Li440.0001. 1289 Supplementary Figure 12. A. Dot plot showing the mark ers used to identify various clusters 1290 such as PD AC, ADM, macrophages, dendritic cells, T NK & B cells, myocytes, endothelial cells, 1291 CAFs, pericytes, and gMDCs. B . Dot plot showing the mark ers used to identify NK cells, T regs, 1292 TNK cells, and CD8 T cells. C . NK global state trajectory plot displaying pan-activation vs. pan-1293 dysfunction signatures in vehicle and ADT-030 treated tumors. D. Dot plot showing expression 1294 of activation, dysfunction, and maturation genes in NK cells from vehicle and ADT-030 treated 1295 tumors with dot size representing the percentage of expressing cells. E. Heat map of 1296 diff erentially expressed genes in NK cells representing a global transcriptional shift toward an 1297 activated, cytotoxic program in ADT-030-treated tumors compared to vehicle. 1298 Supplementary Figur e 13: A-B. Ex vivo imaging of livers of KPCML-1-implanted mice treated 1299 with vehicle or ADT-030 ( A) and bar graph representing the bioluminescence quantification (B). 1300 C-D. Ex vivo imaging of lungs in vehicle or ADT-030-treated mice showing reduced metastasis (C) 1301 and bar graph representing number of distant metastases ( D) . W elch t-test was used for 1302 stat ist ica l analysis. pLi44values are listed on bar graphs. 1303 Supplementary Figur e 14. A. Indicated PDAC cell lines were treated with various concentrations 1304 of PF-25465920 for three days followed by determining viable cell number using MTT assays. 1305 Relative percentage cell viability was plotted with respect to vehicle (DMSO) treated cells. B-C. 1306 The indicated PDAC cell lines were treated with various concentrations of PF-2545920 for 2–4 1307 weeks, and long-term cell survival was measured using clonogenic assays. Representative 1308 images are shown in ( B) and quantification plotted in ( C). 1309 Supplementary Figur e 15. A. Open field and locomotor assay revealed that mice treated with 1310 ADT-030 did not show diff erences in their mobility compared to vehicle while PF-2545920 1311 produced significant reduction in mobility revealing CNS toxicity . ns, non-significant, ∗ pLi44<Li440.05, 1312 and ∗∗∗pLi44<Li440.001. 1313 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint Supplementary Figure 16. The indicated PDAC cell lines were treated with various 1314 concentrations of ADT-030 (A & C) , MRTX1133 (B), or MRTX849 (D) for 2–4 weeks, and long-1315 term cell survival was measured using clonogenic assays. 1316 Supplementary Figure 17. Schematic illustration of the proposed mechanism of action of ADT-1317 030. Inhibition of PDE10 by ADT-030 leads to accumulation of cAMP /cGMP and concomitant 1318 activation of PKA/PK G, resulting in both direct antitumor activity and stimulation of antitumor 1319 immunity . Direct eff ects on growth inhibition, induction of apoptosis, and inhibition of 1320 metastasis are mediated by suppression of β-catenin/TCF-LEF transcriptional activity and 1321 inhibition of both ERK1/2 and PI3K signaling downstream of RAS. Anti-tumor immune eff ects of 1322 ADT-030 are characterized by CD8 T cell-mediated cytotoxicity and immunologic cell death. 1323 1324 1325 1326 1327 1328 1329 1330 1331 Mark e r Clone Supplie r -Cat # Dilution PD-1 F ITC 29F .1A12 Bi oLegend, 135214 1-200 CD206 FIT C C068C2 BioLegend, 141703 1-200 TIM-3 PE ID4B BioLegend, 121607 1-1000 F4/80 PE BM8 BioLegend, 123110 1-800 LAG-3 PE Cy7 9D3,1C8 BioLegend, 517008 1-200 L y6C PECy7 HK1.4 BioLegend, 128017 1-3000 CD11b PE CF594 M1/70 BioLegend, 101255 1-3000 CD45 PerCP 30-F11 BioLegend, 103130 1-800 Foxp3 APC F JK-16s Invitrogen, 17-5773-82 1-200 Nuclear CD11c APC N418 BioLegend, 117309 1-400 CTLA4 APC R700 MP6-XT22 BD , 565778 1-1000 Cytoplas m CD172α AF700 P84 BioLegend, 144022 1-1000 CD62L APC Cy7 ME L -14 BioLegend, 104428 1-800 CD86 APC Cy7 GL1 BioLegend, 105045 1-200 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint CD3 BV 421 145-2C11 BioLegend, 100341 1-800 X CR1 BV 421 ZE T BioLegend, 148216 1-1000 Live/Dead Aqua Invitrogen, L34966 1-500 TCR γδ BV 605 GL3 BioLegend, 118219 1-800 PD-L1 BV 605 MIH5 BioLegend, 153606 1-400 CD4 BV 650 GK1.5 BioLegend, 100469 1-800 CD103 BV 711 2E7 BioLegend, 121435 1-800 NK1.1 BV 711 PK136 BioLegend, 108475 1-400 CD8 BV 785 53-6.7 BioLegend, 100750 1-800 CCR7 BV 785 4B12 BioLegend, 120217 1-250 CD69 BUV 395 H1.2F3 BD , 569367 1-400 B220 BUV 395 RA3.6B2 BD , 563793 1-400 CD44 BUV 737 IM7 BD , 612799 1-800 CD16/CD32 93 BioLegend, 101320 1-500 1332 1333 Supplementary T able 1 . Details on antibodies used for multi-parameter flow cytometry . 1334 1335 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted February 14, 2026. ; https://doi.org/10.64898/2026.02.11.705411doi: bioRxiv preprint